{ "complete": true, "report": { "bucket": "vn_S_WHIS_reg", "label": "Strong Whisper-Talk Style", "definition": { "id": "vn_S_WHIS_reg", "category": "talking_style", "head": "vn_S_WHIS_reg", "pole": "high", "label": "Strong Whisper-Talk Style", "floor": 4.0, "canonical_dimension_name": "Whisper-Talk Style" }, "recipe_status": "direct_manual_recipe", "manual_provenance": [ { "id": "rank4/style_s_whis/full-v2-sealed/sealed-selected", "mission": "style_s_whis", "format_status": "CANONICAL", "heldout": true, "qualified": true, "source_cfg": 3, "source_model": { "label": "Humaneness 1.5 + real-speech rank 4", "repo": "laion/Humaneness-Voice-Large-1.5-RealSpeech-LoRAs", "revision": "fed33e554f094fcdc89fc0124aed2f9d42018714", "rank": 4 }, "metrics": { "ce": 5.89743712425232, "genuineness": 4.491656990051269, "wer": 0.014744462204139623, "identity": 0.5729425963759422, "voiceclap": -0.029782484360039232, "fitness": 0.6479893105757475, "steering_v2": 0.6463824013063707, "raw_target": 2.3843189489841463, "target_cdf": 0.6463824013063707, "family_hit_all": null, "nominal_hit_all": null, "n": 100, "valid_fraction": 1, "by_arm": { "reference": { "n": 50, "ce": 5.921265125274658, "genuineness": 4.507347106933594, "wer": 0.012761188607962801, "identity": 0.5729425963759422, "fitness": 0.6561076432237758, "steering_v2": 0.6577473612183379 }, "none": { "n": 50, "ce": 5.87360912322998, "genuineness": 4.475966873168946, "wer": 0.016727735800316446, "identity": null, "fitness": 0.6398709779277192, "steering_v2": 0.6350174413944035 } } } }, { "id": "rank4/whisper/canonical-v1-exploration/repair-canonical_both", "mission": "whisper", "format_status": "CANONICAL", "heldout": false, "qualified": false, "source_cfg": 2 } ], "selected_wording": { "id": "manual-c79d4bcf72d9", "general": "Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.", "inline": [ "Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.", "Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.", "Fade the final words out slowly, letting the slight tremor in the voice remain present until silence." ], "cfg": 3, "recipe": "P04", "strategy": "local-Gemma-mutation", "timing_syntax": "canonical", "canonical_labels": true, "neutral_event": "keep", "parent": "cfg-3", "origin": "exact_frozen_manual_recipe" }, "selection": { "manual_no_adapter": { "config": { "scale": 0.0, "cfg": 4.0 }, "n": 20, "target_z": 0.9416992162959498, "target_raw": 3.272388052940369, "ce": 5.579809713363647, "genuineness": 4.564772891998291, "identity": 0.5142593719065189, "wer": 0.010818713450292398, "valid_fraction": 1, "identity_delta": -0.0398705393075943, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, "manual_adapter": { "config": { "scale": 0.75, "cfg": 5.0 }, "n": 20, "target_z": 0.9724609294460547, "target_raw": 3.321565628051758, "ce": 5.402613615989685, "genuineness": 4.495736527442932, "identity": 0.5043990485370159, "wer": 0.020201023391812865, "valid_fraction": 1, "identity_delta": -0.04973086267709732, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, "grid": [ { "config": { "scale": 0.0, "cfg": 4.0 }, "n": 20, "target_z": 0.9416992162959498, "target_raw": 3.272388052940369, "ce": 5.579809713363647, "genuineness": 4.564772891998291, "identity": 0.5142593719065189, "wer": 0.010818713450292398, "valid_fraction": 1, "identity_delta": -0.0398705393075943, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 3.0 }, "n": 20, "target_z": 0.809667972529383, "target_raw": 3.0613147497177122, "ce": 5.685248684883118, "genuineness": 4.6209434270858765, "identity": 0.5485545992851257, "wer": 0.008350533195734434, "valid_fraction": 1, "identity_delta": -0.005575311928987503, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 2.5 }, "n": 20, "target_z": 0.7416992088391323, "target_raw": 2.9526556849479677, "ce": 5.786716294288635, "genuineness": 4.609646582603455, "identity": 0.571026049554348, "wer": 0.01956140350877193, "valid_fraction": 1, "identity_delta": 0.016896138340234755, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 2.0 }, "n": 20, "target_z": 0.35498047917061104, "target_raw": 2.334423232078552, "ce": 5.950136756896972, "genuineness": 4.619374442100525, "identity": 0.588895545899868, "wer": 0.0, "valid_fraction": 1, "identity_delta": 0.034765634685754776, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 1.5 }, "n": 20, "target_z": 0.16943360326763704, "target_raw": 2.0377965331077577, "ce": 6.061532640457154, "genuineness": 4.5531621932983395, "identity": 0.5723678231239319, "wer": 0.0027777777777777775, "valid_fraction": 1, "identity_delta": 0.01823791190981865, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 1.25 }, "n": 20, "target_z": -0.18339843818451254, "target_raw": 1.473737433552742, "ce": 6.14651927947998, "genuineness": 4.538727283477783, "identity": 0.5401425585150719, "wer": 0.0027777777777777775, "valid_fraction": 1, "identity_delta": -0.013987352699041366, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 1.0 }, "n": 20, "target_z": -0.27021484283974395, "target_raw": 1.3349473655223847, "ce": 6.143739366531372, "genuineness": 4.535589265823364, "identity": 0.5894169166684151, "wer": 0.003125, "valid_fraction": 1, "identity_delta": 0.035287005454301835, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 5.0 }, "n": 20, "target_z": 0.9235351548300006, "target_raw": 3.243349862098694, "ce": 5.424580097198486, "genuineness": 4.5145645380020145, "identity": 0.4980191826820374, "wer": 0.03142909356725146, "valid_fraction": 1, "identity_delta": -0.05611072853207588, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 5.0 }, "n": 20, "target_z": 0.9724609294460547, "target_raw": 3.321565628051758, "ce": 5.402613615989685, "genuineness": 4.495736527442932, "identity": 0.5043990485370159, "wer": 0.020201023391812865, "valid_fraction": 1, "identity_delta": -0.04973086267709732, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 4.0 }, "n": 20, "target_z": 0.9605468740387696, "target_raw": 3.3025190830230713, "ce": 5.486979746818543, "genuineness": 4.534334063529968, "identity": 0.5205555178225041, "wer": 0.01112831097351221, "valid_fraction": 1, "identity_delta": -0.03357439339160919, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 4.0 }, "n": 20, "target_z": 0.957226562048513, "target_raw": 3.297211027145386, "ce": 5.5768311500549315, "genuineness": 4.536844515800476, "identity": 0.5253635004162789, "wer": 0.013906088751289989, "valid_fraction": 0.95, "identity_delta": -0.028766410797834395, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 4.0 }, "n": 20, "target_z": 0.8890625117677899, "target_raw": 3.1882397651672365, "ce": 5.554889583587647, "genuineness": 4.5594382047653195, "identity": 0.5493513569235802, "wer": 0.013943713450292397, "valid_fraction": 1, "identity_delta": -0.0047785542905330654, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 3.0 }, "n": 20, "target_z": 0.766601548766057, "target_raw": 2.9924661040306093, "ce": 5.763285350799561, "genuineness": 4.603684306144714, "identity": 0.5571295686066151, "wer": 0.019499269005847954, "valid_fraction": 0.95, "identity_delta": 0.002999657392501831, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 3.0 }, "n": 20, "target_z": 0.7559570202468251, "target_raw": 2.975449103116989, "ce": 5.747995781898498, "genuineness": 4.603998160362243, "identity": 0.582642811536789, "wer": 0.011165935672514619, "valid_fraction": 1, "identity_delta": 0.028512900322675706, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 3.0 }, "n": 20, "target_z": 0.7377929774229194, "target_raw": 2.9464109420776365, "ce": 5.692198467254639, "genuineness": 4.565086698532104, "identity": 0.5816405564546585, "wer": 0.005902777777777778, "valid_fraction": 1, "identity_delta": 0.02751064524054527, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 2.5 }, "n": 20, "target_z": 0.6708007958213785, "target_raw": 2.839313101768494, "ce": 5.859193205833435, "genuineness": 4.621257257461548, "identity": 0.5294981613755226, "wer": 0.002631578947368421, "valid_fraction": 1, "identity_delta": -0.024631749838590622, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 2.5 }, "n": 20, "target_z": 0.6439453290011213, "target_raw": 2.7963802933692934, "ce": 5.787311983108521, "genuineness": 4.6102741956710815, "identity": 0.5425458692014218, "wer": 0.003125, "valid_fraction": 1, "identity_delta": -0.011584042012691498, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 2.5 }, "n": 20, "target_z": 0.6196289010433355, "target_raw": 2.757506549358368, "ce": 5.802204489707947, "genuineness": 4.630985188484192, "identity": 0.6137341514229775, "wer": 0.013651315789473683, "valid_fraction": 0.95, "identity_delta": 0.05960424020886421, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 2.0 }, "n": 20, "target_z": 0.5707031115136467, "target_raw": 2.679290759563446, "ce": 5.898707938194275, "genuineness": 4.6629928827285765, "identity": 0.5834273673593998, "wer": 0.008333333333333333, "valid_fraction": 1, "identity_delta": 0.02929745614528656, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 2.0 }, "n": 20, "target_z": 0.4915038998782078, "target_raw": 2.5526780068874357, "ce": 5.888978123664856, "genuineness": 4.544061923027039, "identity": 0.5468945533037186, "wer": 0.0061111111111111106, "valid_fraction": 1, "identity_delta": -0.007235357910394669, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 2.0 }, "n": 20, "target_z": 0.4515625089714834, "target_raw": 2.4888252317905426, "ce": 5.943584036827088, "genuineness": 4.625336694717407, "identity": 0.5898355215787887, "wer": 0.0, "valid_fraction": 1, "identity_delta": 0.035705610364675525, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 1.5 }, "n": 20, "target_z": 0.11630860396671366, "target_raw": 1.9528676271438599, "ce": 6.025393557548523, "genuineness": 4.610587954521179, "identity": 0.6227822929620743, "wer": 0.02660818713450292, "valid_fraction": 0.95, "identity_delta": 0.06865238174796104, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 1.5 }, "n": 20, "target_z": 0.09648438650563619, "target_raw": 1.9211754083633423, "ce": 6.074240970611572, "genuineness": 4.6036843180656435, "identity": 0.6020422905683518, "wer": 0.0, "valid_fraction": 1, "identity_delta": 0.04791237935423851, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 1.5 }, "n": 20, "target_z": 0.08916015506302614, "target_raw": 1.9094664394855498, "ce": 6.113557195663452, "genuineness": 4.600232529640198, "identity": 0.5949906706809998, "wer": 0.005902777777777778, "valid_fraction": 1, "identity_delta": 0.04086075946688652, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 1.25 }, "n": 20, "target_z": -0.02890625565086942, "target_raw": 1.7207181811332704, "ce": 6.121102738380432, "genuineness": 4.556614017486572, "identity": 0.6017352595925332, "wer": 0.0, "valid_fraction": 1, "identity_delta": 0.04760534837841988, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 1.25 }, "n": 20, "target_z": -0.09873046751933386, "target_raw": 1.6090928822755814, "ce": 6.103628849983215, "genuineness": 4.564459109306336, "identity": 0.5925638288259506, "wer": 0.003333333333333333, "valid_fraction": 1, "identity_delta": 0.038433917611837384, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 1.25 }, "n": 20, "target_z": -0.1072265695053304, "target_raw": 1.5955104887485505, "ce": 6.114748573303222, "genuineness": 4.55661404132843, "identity": 0.5917562559247017, "wer": 0.0027777777777777775, "valid_fraction": 1, "identity_delta": 0.037626344710588455, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 1.0 }, "n": 20, "target_z": -0.23603515576210277, "target_raw": 1.389589124917984, "ce": 6.146122145652771, "genuineness": 4.5073470830917355, "identity": 0.5861599579453468, "wer": 0.0, "valid_fraction": 1, "identity_delta": 0.032030046731233594, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 1.0 }, "n": 20, "target_z": -0.2720703115242055, "target_raw": 1.331981098651886, "ce": 6.1659788370132445, "genuineness": 4.560065817832947, "identity": 0.5423904977738857, "wer": 0.002631578947368421, "valid_fraction": 1, "identity_delta": -0.011739413440227508, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 1.0 }, "n": 20, "target_z": -0.30507811687323416, "target_raw": 1.2792127817869186, "ce": 6.162404608726502, "genuineness": 4.551279425621033, "identity": 0.5445484340190887, "wer": 0.0, "valid_fraction": 1, "identity_delta": -0.009581477195024491, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 5.0 }, "n": 20, "target_z": 0.9624999860184674, "target_raw": 3.3056414484977723, "ce": 5.468264865875244, "genuineness": 4.472828817367554, "identity": 0.502280056476593, "wer": 0.030134158926728584, "valid_fraction": 0.95, "identity_delta": -0.05184985473752022, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 5.0 }, "n": 20, "target_z": 0.9505859678952693, "target_raw": 3.2865949630737306, "ce": 5.384966015815735, "genuineness": 4.508288502693176, "identity": 0.4990135669708252, "wer": 0.0027777777777777775, "valid_fraction": 1, "identity_delta": -0.05511634424328804, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" } ], "development_standard": [ { "scale": 0.0, "n": 20, "target_z": -0.5654296856794879, "target_raw": 0.8629986792802811, "ce": 6.236668682098388, "genuineness": 4.352642679214478, "identity": 0.5541299112141133, "wer": 0.0, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "scale": 0.5, "n": 20, "target_z": -0.5180664051212825, "target_raw": 0.9387165457010269, "ce": 6.196558141708374, "genuineness": 4.372098445892334, "identity": 0.5782705336809159, "wer": 0.011374269005847953, "valid_fraction": 1, "identity_delta": 0.024140622466802597, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "scale": 0.75, "n": 20, "target_z": -0.5998046929032798, "target_raw": 0.8080446690320968, "ce": 6.2029122591018675, "genuineness": 4.3667638301849365, "identity": 0.5469126231968403, "wer": 0.0027777777777777775, "valid_fraction": 1, "identity_delta": -0.007217288017272949, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "scale": 1.0, "n": 20, "target_z": -0.6624999990678978, "target_raw": 0.7078160792589188, "ce": 6.212443470954895, "genuineness": 4.3178107142448425, "identity": 0.5681150861084461, "wer": 0.0, "valid_fraction": 1, "identity_delta": 0.013985174894332885, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" } ] }, "heldout": { "standard": { "n": 50, "target_z": -0.6456640627708922, "target_raw": 0.7347310471534729, "ce": 6.291711368560791, "genuineness": 4.390236177444458, "identity": 0.5940428373217582, "wer": 0.009192076039902127, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.7236328242240977, "target_raw": 2.9237736463546753, "ce": 5.524906034469605, "genuineness": 4.4301518535614015, "identity": 0.5163092762231827, "wer": 0.02796586327478776, "valid_fraction": 0.98, "identity_delta": -0.0777335610985756, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.8048046794810085, "target_raw": 3.0535399889945984, "ce": 5.401293201446533, "genuineness": 4.389985132217407, "identity": 0.47779575005173686, "wer": 0.024120672233944544, "valid_fraction": 1, "identity_delta": -0.11624708727002143, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.7891015720686113, "target_raw": 3.0284360313415526, "ce": 5.397038917541504, "genuineness": 4.405800771713257, "identity": 0.5218156635761261, "wer": 0.03252567188722795, "valid_fraction": 0.96, "identity_delta": -0.07222717374563217, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.015703107412397138, "ci95_z": [ -0.028164086073446497, 0.04960936641883437 ], "language_deltas": { "en": 0.029492175076825732, "de": -0.06660159350696132, "fr": 0.06718749056246551, "nl": 0.037890616989746946, "es": 0.01054684793990882 }, "languages_improved": 4, "clear_gain": false, "delta_ce": 0.0042542839050292965, "delta_genuineness": -0.01581563949584961, "delta_wer": -0.008404999653283406, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "adapter_vs_best_manual": { "n": 50, "delta_z": 0.08117185525691085, "ci95_z": [ 0.052265602588768376, 0.10554686397206618 ], "language_deltas": { "en": 0.06425779649405801, "de": 0.09804687310666749, "fr": 0.12109374807753927, "nl": 0.08945307749191737, "es": 0.03300778111437212 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.12361283302307129, "delta_genuineness": -0.04016672134399414, "delta_wer": -0.0038451910408432146, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "manual_vs_standard": { "n": 50, "delta_z": 1.3692968869949897, "ci95_z": [ 1.2258203166507673, 1.5401953328227402 ], "language_deltas": { "en": 1.2619140777777371, "de": 1.2414062491261542, "fr": 1.1921875036118958, "nl": 1.6542969052641925, "es": 1.4966796991949693 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.7668053340911866, "delta_genuineness": 0.03991567611694336, "delta_wer": 0.018773787234885632, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "route": { "adapter": false, "scale": 0.0, "cfg": 1.0, "reason": "No new configuration passes the held-out gates; retain standard fallback and inspect audio", "fallback_standard": true }, "reference_only": true, "updated": 1793044567.1731143 }, "rows": [ { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "0012b5d7b0dd2c2296327808a1acc628cfe7c7c1e8e5c1679c5745b1437ff44d", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] La porte est encore ouverte.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b6a07a2e87778a0fc2684366a5402dbe713a2f32a3ef2b598b50bf3700349cc9", "generation_seconds_batch": 17.34514583996497, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9261711835861206, "target_z": 0.09960935970769873, "ce": 6.078212261199951, "genuineness": 4.331617832183838, "identity": 0.5666025876998901, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0012b5d7b0dd2c2296327808a1acc628cfe7c7c1e8e5c1679c5745b1437ff44d.mp3" }, { "id": "heldout-fr-0", "language": "fr", "sentences": [ "Ton message est arrivé après mon départ.", "Je l’ai relu... puis j’ai rangé mon téléphone.", "Dis-moi quand nous pourrons parler." ], "text": "Ton message est arrivé après mon départ. Je l’ai relu... puis j’ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-0", "key": "003ae1ffa30ffbf3f57fde47368a901cdf609591e59971deb8744443485434e5", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Natural clear speech.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4db594391aa81f22103e04681489564b7903358ec5a558dddbe43992007a4a94", "generation_seconds_batch": 22.44589341408573, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.577913522720337, "target_z": 1.132812479610265, "ce": 5.08339262008667, "genuineness": 4.356721878051758, "identity": 0.5865521430969238, "hypothesis": "Ton message est arrivé après mon départ, je l'ai relu puis j'ai rangé mon téléphone. Dis-moi qu'on nous pourrons parler.", "wer": 0.047619047619047616, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/003ae1ffa30ffbf3f57fde47368a901cdf609591e59971deb8744443485434e5.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "0062908e44613056ad9cdc8b8bb7a36bd3cad9eb5f44832fdde62bf7c6d7c4be", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "99968804395f0d5406fe56e388a5e5794dafce1ffda963da8d8d568a4ca716d8", "generation_seconds_batch": 21.77016671700403, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1237179040908813, "target_z": -0.4023437671856338, "ce": 6.141753673553467, "genuineness": 4.570107460021973, "identity": 0.4129529595375061, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen daar samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0062908e44613056ad9cdc8b8bb7a36bd3cad9eb5f44832fdde62bf7c6d7c4be.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "007df26aaad48b892d85505a533c1e8c8577bf0fb570c06db29b479618ab5ab0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "795186718c53e4f3aa3e142bca0ec30605dd09c7e38f660f529c49d3137e0026", "generation_seconds_batch": 17.6202357520815, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1767985820770264, "target_z": -0.36914057271489376, "ce": 6.316492557525635, "genuineness": 4.444586753845215, "identity": 0.4019800126552582, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/007df26aaad48b892d85505a533c1e8c8577bf0fb570c06db29b479618ab5ab0.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "00e506abde78e30e20ae5447ff0263db41df7192ecd80bca6218dd5baf09cef5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "c41e5eee8e20dee7b798df6c3afbffcd06b19548771615ebed588a75af63244f", "generation_seconds_batch": 22.70148966391571, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7794190645217896, "target_z": 0.007812507573330156, "ce": 5.974957466125488, "genuineness": 4.061748504638672, "identity": 0.3700278103351593, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/00e506abde78e30e20ae5447ff0263db41df7192ecd80bca6218dd5baf09cef5.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "0198294d8009f2a8e44e0b2576a8a3cdd4281165eef5c530f0297a7429ce9d41", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "40d669a94967793f886bf5f2eeceb2977afe260ddc92f709e7474971881fd5bc", "generation_seconds_batch": 20.86365207505878, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3156330585479736, "target_z": 0.9687499697066794, "ce": 5.7128496170043945, "genuineness": 4.871357440948486, "identity": 0.6904757022857666, "hypothesis": "La porte est encore ouverte. J'ai j'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0198294d8009f2a8e44e0b2576a8a3cdd4281165eef5c530f0297a7429ce9d41.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "019a4f0967e7bfa371165173894adafb5ffd8a1185541e034aa39108ca4eb649", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2e6b15308de6922d3acbea2ba74b8b2501a03c1a28f1ed082c4f8b747f808373", "generation_seconds_batch": 23.55445547506679, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7481951713562012, "target_z": -0.011718761359995234, "ce": 5.847874641418457, "genuineness": 4.645420074462891, "identity": 0.8005949258804321, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/019a4f0967e7bfa371165173894adafb5ffd8a1185541e034aa39108ca4eb649.mp3" }, { "id": "heldout-fr-4", "language": "fr", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-4", "key": "01f5d5f22d9b51f4e674eec92be80d0710e62f5704093e3409bc811236eec3e5", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "364b3283a52a1abc5e451fbb8957a8f14d4f4a7e8001ce030fcd3ad385d4eb8b", "generation_seconds_batch": 16.800370823941194, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6170549392700195, "target_z": -0.09375001631178803, "ce": 5.959072113037109, "genuineness": 4.582659721374512, "identity": 0.8655979633331299, "hypothesis": "J'ai écrit le numéro sur l'enveloppe. Lis-les bien et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "wer": 0.045454545454545456, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/01f5d5f22d9b51f4e674eec92be80d0710e62f5704093e3409bc811236eec3e5.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "023f0bc50c1d54d29afc953bb7a08bb225fbf17dff90322576399be941e638a7", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c070eeb6bdacbb4cf004177c302d88f21d38c6ec2053f59339bdf7ad88c64444", "generation_seconds_batch": 23.237425912055187, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1345348358154297, "target_z": 0.8554688335979136, "ce": 6.086154937744141, "genuineness": 4.758388519287109, "identity": 0.6578488945960999, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/023f0bc50c1d54d29afc953bb7a08bb225fbf17dff90322576399be941e638a7.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "0245a00ccdcafae81720cd69ef90eaa1410f83260f6c6b53e918d1abffa347c8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "ccf51da257c01f572839ecb82adc4c78d0b71937c3f87799fff084b9f7161bba", "generation_seconds_batch": 16.358972711022943, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1969823837280273, "target_z": 0.8945312223282167, "ce": 5.474569320678711, "genuineness": 4.532451152801514, "identity": 0.3837094008922577, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0245a00ccdcafae81720cd69ef90eaa1410f83260f6c6b53e918d1abffa347c8.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "02dea7dccd08b5ffd28ef1206c0044a3a7ff1fe0247c1d5625c5b4c217e90a99", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f42b7389032654be2e98d5b57a9df50f0d31289f2e96af866af58f5b33386036", "generation_seconds_batch": 22.90019148599822, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.328122615814209, "target_z": 0.9765624772800096, "ce": 5.760505676269531, "genuineness": 4.7458367347717285, "identity": 0.8270131349563599, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/02dea7dccd08b5ffd28ef1206c0044a3a7ff1fe0247c1d5625c5b4c217e90a99.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "033609f84688b8de0d77295fa66b311dc20e6d36cd78b1be0fbec6f1bb8c9318", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ee7d6f3a55dd9514433621e3a96a9fd0743bf5a146bdc67f7307c78e15961f1d", "generation_seconds_batch": 23.682209418038838, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.290654182434082, "target_z": 0.9531251036963667, "ce": 5.200547218322754, "genuineness": 4.65797233581543, "identity": 0.5234811305999756, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/033609f84688b8de0d77295fa66b311dc20e6d36cd78b1be0fbec6f1bb8c9318.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "03665465f1785701005f6a1a11ecbd9bc38eb007f5f72cbdbca86052852151b9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "76871499bbdd36cce6f9af94f0d787107355acd663daa2ae52bf11b7be6f576f", "generation_seconds_batch": 21.715591547079384, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.026087522506714, "target_z": 0.16210934572616612, "ce": 6.1894097328186035, "genuineness": 4.507347106933594, "identity": 0.7930575013160706, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/03665465f1785701005f6a1a11ecbd9bc38eb007f5f72cbdbca86052852151b9.mp3" }, { "id": "heldout-nl-3", "language": "nl", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-3", "key": "04619adb48e3dcd77feb75bb6cc86ad349303d59914a7dadb417478a224fa33d", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Natural clear speech.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Natural clear speech.) [3.2 seconds duration] We zetten het na de lunch terug.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "9b216a84c54a3951486829ae050e4da43cef44c009e303c07117f7cb56c0ec01", "generation_seconds_batch": 17.490973711013794, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6224632263183594, "target_z": 0.5351561722277249, "ce": 5.935244083404541, "genuineness": 4.444586753845215, "identity": 0.35616037249565125, "hypothesis": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier naast de hoge boekkast. We zetten het na de lunch terug.", "wer": 0.047619047619047616, "duration_error_s": 0.120000000000001, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/04619adb48e3dcd77feb75bb6cc86ad349303d59914a7dadb417478a224fa33d.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "046dfdb20673295568c66bdeb808a219f8f895e05ac3d6bd60626969892b1931", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "553c5bd8985105146419c42f742206d272e7676454320d11c1f875ea7650110e", "generation_seconds_batch": 17.07272485306021, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2032270431518555, "target_z": 0.8984374015467079, "ce": 5.903473377227783, "genuineness": 4.457139015197754, "identity": 0.36322030425071716, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/046dfdb20673295568c66bdeb808a219f8f895e05ac3d6bd60626969892b1931.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "048ea3138d6250fcb62dd7e5ab1f411d0136e3bdd100ec8d91da957b6e0da51a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0c66c7bca4c62f7e918c2f778c69f57689ab7052ddc884e7df0b230d9da114fe", "generation_seconds_batch": 22.70480802305974, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.690319538116455, "target_z": 1.2031250477702364, "ce": 5.462655067443848, "genuineness": 4.582659721374512, "identity": 0.7438751459121704, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/048ea3138d6250fcb62dd7e5ab1f411d0136e3bdd100ec8d91da957b6e0da51a.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "05054ba480149ce6aa5b5e7fa868c61cfe09abd2136d8394aaf1b6393f44fbd4", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "759fdc01e242933e03a7d7aad2984e4fa6489bba2803dd15914b475662285a3d", "generation_seconds_batch": 22.39473205001559, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7036454677581787, "target_z": 0.5859375460225448, "ce": 5.9431867599487305, "genuineness": 4.68307638168335, "identity": 0.6233587265014648, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/05054ba480149ce6aa5b5e7fa868c61cfe09abd2136d8394aaf1b6393f44fbd4.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "050f0ed5287e868b86b9495387292c7ce92a43a975a105c64893210bf06e8af3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "78accbf50daed2013ced910274d11bc2f363cd89fb7627f448cdebe746fa54ca", "generation_seconds_batch": 17.020904073957354, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.615382194519043, "target_z": 1.1562500023302553, "ce": 5.577824115753174, "genuineness": 4.7458367347717285, "identity": 0.6408676505088806, "hypothesis": "La porte est encore ouverte, j'ai posé ton manteau sur la chaise Dis-moi comme tu es prêt.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/050f0ed5287e868b86b9495387292c7ce92a43a975a105c64893210bf06e8af3.mp3" }, { "id": "heldout-de-3", "language": "de", "sentences": [ "Jemand hat den kleinen blauen Stuhl umgestellt.", "Er stand hier... neben dem hohen Bücherregal.", "Wir stellen ihn nach dem Essen zurück." ], "text": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier... neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-3", "key": "052b3c129cad88b4d39001740765672f3a4569b593ef084e046d523894d8537b", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Natural clear speech.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Natural clear speech.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "46916b1b842201dced8521fa79e42ae6330f2cc44ec5f6cfbc1957e403e179c2", "generation_seconds_batch": 16.936351155978628, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0073533058166504, "target_z": 0.15039065893434472, "ce": 5.411027908325195, "genuineness": 4.457139015197754, "identity": 0.6579370498657227, "hypothesis": "Ihr Mann hat den kleinen blauen Stuhl umgestellt. Er stand hier neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "wer": 0.09523809523809523, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/052b3c129cad88b4d39001740765672f3a4569b593ef084e046d523894d8537b.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "059057fe536983c39a38e3b9a43162b28cc269bb86842628d2ec8c080044e249", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "c6887a11846fc8d3f559d62f4d369e63af99a46ed3a111a46931c2163aca0386", "generation_seconds_batch": 23.592666225973517, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.2602665424346924, "target_z": 0.3085937508738458, "ce": 6.252951145172119, "genuineness": 4.720732688903809, "identity": 0.33219677209854126, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/059057fe536983c39a38e3b9a43162b28cc269bb86842628d2ec8c080044e249.mp3" }, { "id": "heldout-de-0", "language": "de", "sentences": [ "Deine Nachricht kam, als ich schon weg war.", "Ich las sie zweimal... dann legte ich das Handy weg.", "Sag mir, wann wir reden können." ], "text": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal... dann legte ich das Handy weg. Sag mir, wann wir reden können.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-0", "key": "059867605942d1f4c46c2c44d936e08b7fbe89dd4feef4c05500a7ed426c27a3", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wann wir reden können.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wann wir reden können.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c6fe9235fa72d52c6fe3a19728e3354913868170777a1575ec7ddd1bee3f903f", "generation_seconds_batch": 17.495784764993005, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8177240490913391, "target_z": -0.593750016311788, "ce": 6.316492557525635, "genuineness": 4.7332844734191895, "identity": 0.5738356113433838, "hypothesis": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal, dann schlegte ich das Handy weg. Sag mir, wann wir reden können.", "wer": 0.041666666666666664, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/059867605942d1f4c46c2c44d936e08b7fbe89dd4feef4c05500a7ed426c27a3.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "05f7a8b4be96e81bfdbcceb800273e26f3f5b9f384004c49a94fc12c8ad3a283", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "416466d79e6befe668b84804acb0c048603796e297d68fd1b0aae266a55a15e7", "generation_seconds_batch": 21.251901319948956, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.485914707183838, "target_z": -0.1757812712635808, "ce": 6.086154937744141, "genuineness": 4.545003414154053, "identity": 0.4554651081562042, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/05f7a8b4be96e81bfdbcceb800273e26f3f5b9f384004c49a94fc12c8ad3a283.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "0672b82a23e85d7a95f964bdff4ce6ea938522001c34efd97c04c3e1d4bb2346", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5ad08f1a0d4e4e044def9217e94ee239abb63ba1f9cb316aa0156eafe7b39809", "generation_seconds_batch": 22.209093887009658, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.790235996246338, "target_z": 1.2656251083568777, "ce": 5.839932441711426, "genuineness": 4.758388519287109, "identity": 0.8734922409057617, "hypothesis": "J'ai trouvé la petite boîte elle était près de la vieille fenêtre nous pouvons regarder ensemble ce qu'elle contient", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0672b82a23e85d7a95f964bdff4ce6ea938522001c34efd97c04c3e1d4bb2346.mp3" }, { "id": "heldout-de-2", "language": "de", "sentences": [ "Die Straße führt hinter der Brücke nach links.", "Geh weiter... bis zu dem kleinen Holztor.", "Ich warte auf der anderen Seite." ], "text": "Die Straße führt hinter der Brücke nach links. Geh weiter... bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-2", "key": "06c00a420de8b7d5f9a6a4a95807675d40eada449e6518279bb1c1ef222d8e38", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Natural clear speech.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Natural clear speech.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5f09bc3824da7891a3e20cb6782c547f8aed8614d4793f2261fd47db6802a524", "generation_seconds_batch": 22.961345597053878, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.178248167037964, "target_z": 0.8828125355363954, "ce": 5.43088436126709, "genuineness": 4.457139015197754, "identity": 0.5808612108230591, "hypothesis": "Die Straße führt hinter der Brücke nach links. Geh weiter bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/06c00a420de8b7d5f9a6a4a95807675d40eada449e6518279bb1c1ef222d8e38.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "070e1a7281043355393dc6f5fce1150cda6aa6f53acd36d8f69817655163e2a8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6b194fc9da0d232e7abef9aeab25aa15163b0935a300508313fbf77815353b55", "generation_seconds_batch": 21.77947737602517, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7794190645217896, "target_z": 0.007812507573330156, "ce": 5.689021587371826, "genuineness": 4.595211505889893, "identity": 0.5496580600738525, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/070e1a7281043355393dc6f5fce1150cda6aa6f53acd36d8f69817655163e2a8.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "07944bd4a38cfb4d56f5f1285e2c5b6478ea97cd8d1249c3458f39099d33d64a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "6be4ebfed20348255cfaf3d4b084d6d31da4fa3db822d922b6b8493404eb685d", "generation_seconds_batch": 22.335953415953554, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.259430170059204, "target_z": 0.9335937601948675, "ce": 5.71682071685791, "genuineness": 4.557555198669434, "identity": 0.556840181350708, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/07944bd4a38cfb4d56f5f1285e2c5b6478ea97cd8d1249c3458f39099d33d64a.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "07a8a3a1aed12657b64a7bee2d3d5ff26fc0dff935632f8a30f7b7a7443f8926", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8bf51075a98175914f898e48c6b8598dd5da6fb9c48da49509bcde8b42c49761", "generation_seconds_batch": 22.01190947194118, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.078331708908081, "target_z": 0.8203124749497541, "ce": 5.720792293548584, "genuineness": 4.7834930419921875, "identity": 0.48986390233039856, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/07a8a3a1aed12657b64a7bee2d3d5ff26fc0dff935632f8a30f7b7a7443f8926.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "07cf94499bd4423ace366d4f0cc2c1fd7c0b44ee2a686f1248eb2e1d41a5ece4", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9c63ccc2d808172bb558094ea14884294e27dc78fb12f5786d647d40c72a0de3", "generation_seconds_batch": 22.322724999976344, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7036454677581787, "target_z": 0.5859375460225448, "ce": 6.0146708488464355, "genuineness": 4.720732688903809, "identity": 0.7148261666297913, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/07cf94499bd4423ace366d4f0cc2c1fd7c0b44ee2a686f1248eb2e1d41a5ece4.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "08295c03238188a440ccdfe8849e5415f288cadd8e17847c2ce69145570221aa", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] Die Tür ist noch offen.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "17e6bcd941c1c3b5f63fa6cced44d01c89695b7306e5c13269f53b4516686169", "generation_seconds_batch": 16.016175522003323, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0706373453140259, "target_z": -0.43554688708820005, "ce": 6.24500846862793, "genuineness": 4.607763767242432, "identity": 0.6374292373657227, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/08295c03238188a440ccdfe8849e5415f288cadd8e17847c2ce69145570221aa.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "08319d12dd6384dea2bd472ebf50a8d8e67d0fc89590d89bd121ec8c85bf0474", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6742ce533c280fb6db6216ef79268d936d32a8a56401bb20d492509f8aaac8f8", "generation_seconds_batch": 23.11652394605335, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9034781455993652, "target_z": 0.7109375180594796, "ce": 5.665193557739258, "genuineness": 4.883909225463867, "identity": 0.44355666637420654, "hypothesis": "He encontrado la cajita. estaba junto a la ventana vieja. podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/08319d12dd6384dea2bd472ebf50a8d8e67d0fc89590d89bd121ec8c85bf0474.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "0841d8bd2d7090689571b3fe7e9d7fbe39ef83b3b100d93f797e8a320da36dd9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e94429c373346808afa5c9653a12fe497d116fe4fefd0a50e3166574b6f7b83d", "generation_seconds_batch": 22.853756654076278, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.428039073944092, "target_z": 1.0390625378666507, "ce": 5.569881439208984, "genuineness": 4.68307638168335, "identity": 0.3806963860988617, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0841d8bd2d7090689571b3fe7e9d7fbe39ef83b3b100d93f797e8a320da36dd9.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "0864c706d70c1e78ce31381794341748d1f7c373ac793d2ae45637a88d0a943d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2047e6a6943d62de3c796123036c2ed91d3f94ead1c1148dcf044d687f703afd", "generation_seconds_batch": 23.11652394605335, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7723379135131836, "target_z": 0.6289062631076868, "ce": 5.720792293548584, "genuineness": 4.770940780639648, "identity": 0.6073926687240601, "hypothesis": "He encontrado la cajita. estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0864c706d70c1e78ce31381794341748d1f7c373ac793d2ae45637a88d0a943d.mp3" }, { "id": "heldout-es-2", "language": "es", "sentences": [ "La carretera gira a la izquierda tras el puente.", "Sigue... hasta la pequeña puerta de madera.", "Te esperaré al otro lado." ], "text": "La carretera gira a la izquierda tras el puente. Sigue... hasta la pequeña puerta de madera. Te esperaré al otro lado.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-2", "key": "08db4d97ace48cb78a5f6e134240e675fe0160e18a78738b7653d627efb5fe6c", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Te esperaré al otro lado.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Natural clear speech.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Natural clear speech.) [3.2 seconds duration] Te esperaré al otro lado.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "b983b5e78a7a04945f474b1c85e8dab3c6186bc0294f7a453029d1641944b76b", "generation_seconds_batch": 22.0864660179941, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.040863037109375, "target_z": 0.7968749522297637, "ce": 4.966238498687744, "genuineness": 4.281409740447998, "identity": 0.47785308957099915, "hypothesis": "La carretera gira a la izquierda tras el puente sigue hasta la pequeña puerta de madera. Te esperaré al otro lado.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/08db4d97ace48cb78a5f6e134240e675fe0160e18a78738b7653d627efb5fe6c.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "08e8235c984ba1806eac98fdfd2e30fe7248407313d3334365496fc1b3c1451b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4e9e225bea00f787248784971133a6845c95087aed363e0a1cd8b75421d2c744", "generation_seconds_batch": 22.956270491005853, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7598483562469482, "target_z": 0.6210937555343566, "ce": 6.109982967376709, "genuineness": 4.68307638168335, "identity": 0.21403031051158905, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen daar samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/08e8235c984ba1806eac98fdfd2e30fe7248407313d3334365496fc1b3c1451b.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "0950a2eb3c4894228254c4179bea9451316ac5149aeb913f48344f497f8f2160", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "8f4e09b39285db39b46d175d32e923a2d845d0e5708a2440c55b4f6ae336ed94", "generation_seconds_batch": 21.202916178037412, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4655075073242188, "target_z": 1.0624999114502935, "ce": 5.6413655281066895, "genuineness": 4.369274139404297, "identity": 0.3048524856567383, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0950a2eb3c4894228254c4179bea9451316ac5149aeb913f48344f497f8f2160.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "09b99fd1b2ffe1e487726b8fb80ca20c4470880f417631286e5cdb91029d2035", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0b1cd7b401263eda75524052dfdb539e79c519d85223f79408ed0196713fd26c", "generation_seconds_batch": 21.064869030029513, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.500690221786499, "target_z": 0.4589843352400167, "ce": 5.998785495758057, "genuineness": 4.909013748168945, "identity": 0.8252896070480347, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/09b99fd1b2ffe1e487726b8fb80ca20c4470880f417631286e5cdb91029d2035.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "0a294a8f720b28eb3a1009be0ecfb1b10fd13863998cbae1dfd4cf9ce3d69f0a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "92527e3cc0163ae6f47716bcdb164aee96a55de862175cef94d7a3495c3ce33f", "generation_seconds_batch": 22.93753120000474, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8035616874694824, "target_z": 0.6484374574728383, "ce": 6.054384231567383, "genuineness": 4.758388519287109, "identity": 0.8093175888061523, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0a294a8f720b28eb3a1009be0ecfb1b10fd13863998cbae1dfd4cf9ce3d69f0a.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "0a4ab1e363525356118fb8baf10f626d7a07c85f7dd295be3764dab8c8c8a52e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "26975d54326690a672ed323d73a04acff469c3e19ec426cf489236509a21a149", "generation_seconds_batch": 20.833433876978233, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.066678524017334, "target_z": 0.1874999580554022, "ce": 5.982900142669678, "genuineness": 4.557555198669434, "identity": 0.5151419043540955, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0a4ab1e363525356118fb8baf10f626d7a07c85f7dd295be3764dab8c8c8a52e.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "0adde2c68303d84942052a43cb9fd85d20b8a6f8201b51e0b1c01523f278da86", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] He encontrado la cajita.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "490df6a2204f647caeda48aad72bdccd45822be036adfc3cc3cd9d8d5a5b7185", "generation_seconds_batch": 17.506156646064483, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6428704261779785, "target_z": -0.7031250104861494, "ce": 6.173524379730225, "genuineness": 4.607763767242432, "identity": 0.7225716710090637, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0adde2c68303d84942052a43cb9fd85d20b8a6f8201b51e0b1c01523f278da86.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "0b7f9aff1e08ef0edc1bfc13e3da4dc23bb654f6c562909d00e36e669b8742ad", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "13c64cd345543771632686f59470eb2df3245090d8b4e691a2e11484a87d7c40", "generation_seconds_batch": 22.110985856037587, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6162185668945312, "target_z": 0.5312499930092337, "ce": 5.9431867599487305, "genuineness": 4.632867813110352, "identity": 0.7054890394210815, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0b7f9aff1e08ef0edc1bfc13e3da4dc23bb654f6c562909d00e36e669b8742ad.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "0bde707cb22365fee9675119d25875453c08c377f48abcfe753fd8a14319b3dc", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bff4b7ea06d1c3e5f7a3f70096bc78ad46ef408b9423c6907a7fb064a6ce368b", "generation_seconds_batch": 17.039053219021298, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9863329529762268, "target_z": -0.4882812386400048, "ce": 6.165581703186035, "genuineness": 4.494794845581055, "identity": 0.5760422945022583, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0bde707cb22365fee9675119d25875453c08c377f48abcfe753fd8a14319b3dc.mp3" }, { "id": "heldout-nl-4", "language": "nl", "sentences": [ "Ik schreef het nummer op de envelop.", "Lees het goed... en controleer de laatste twee cijfers.", "Bel me wanneer je even tijd hebt." ], "text": "Ik schreef het nummer op de envelop. Lees het goed... en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-4", "key": "0cb70e86df8754d3bba44fe7f798aac7d0ffe2ca68fc75b92ff1a1ed29986ba7", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "04b8bb8d0ddcefb9c4ff987e292aa25d52c44268eb96000cdd8d69fdc2251a0e", "generation_seconds_batch": 17.376467431080528, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.5367093086242676, "target_z": -0.7695312502912819, "ce": 6.284721851348877, "genuineness": 4.381826400756836, "identity": 0.5465561151504517, "hypothesis": "Ik schreef het nummer op de envelop, lees het goed en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0cb70e86df8754d3bba44fe7f798aac7d0ffe2ca68fc75b92ff1a1ed29986ba7.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "0d36538eaf173c58a17acf800cde6069a50dd302ccea1828d7dac82d51f850da", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "93ce315a4b9f981e33a05bf10407e5084c03df857aa7fbb9a37d16c35c273e1e", "generation_seconds_batch": 23.224551855004393, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5131797790527344, "target_z": 0.4667968428133468, "ce": 5.895530700683594, "genuineness": 4.632867813110352, "identity": 0.7026180028915405, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0d36538eaf173c58a17acf800cde6069a50dd302ccea1828d7dac82d51f850da.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "0de459bb0317d1e1734adcfd6e01e481c916f13544f910c84d71ff3d72094c1c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ce10285f2913c130d270ff859364c96a6ceff6ae983eba94f3763ca79ab925bd", "generation_seconds_batch": 20.299988040933385, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9159677028656006, "target_z": 0.7187500256328098, "ce": 6.09409761428833, "genuineness": 4.7834930419921875, "identity": 0.6723043918609619, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0de459bb0317d1e1734adcfd6e01e481c916f13544f910c84d71ff3d72094c1c.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "0df4c0eb1623df9fcd30eaf3e12aacfe9ff8bd2c99d94798ce82608c1f3db20f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d049e6602e13db3bd54b44244abee39ca0a5b6abceb913c44f47f09dd85e310d", "generation_seconds_batch": 17.455278169945814, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8035616874694824, "target_z": 0.6484374574728383, "ce": 5.764476776123047, "genuineness": 4.582659721374512, "identity": 0.5105161666870117, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast oude raam. We kunnen er samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0df4c0eb1623df9fcd30eaf3e12aacfe9ff8bd2c99d94798ce82608c1f3db20f.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "0e415c7c47492a559a534a9285794b3707625b2d3d60176efa9121fddedb001d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c2a8f4b05f3e1db47d04f64dddb1ccbfe951744c98716db13d3a01b46f8ed5db", "generation_seconds_batch": 22.66546756005846, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2298791408538818, "target_z": -0.3359374528123275, "ce": 6.1020402908325195, "genuineness": 4.457139015197754, "identity": 0.6246387958526611, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0e415c7c47492a559a534a9285794b3707625b2d3d60176efa9121fddedb001d.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "0e88fdb5bb9845b8a9b7ec20f2c9943eddb837db34ed4c170827d4837bed1974", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "38828e50490a77d87aec9307b5721b3eeb3c9f0fb8dccf5ef19ceb32984e053b", "generation_seconds_batch": 16.856875904020853, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2781643867492676, "target_z": 0.9453124469866889, "ce": 5.760505676269531, "genuineness": 4.268857479095459, "identity": 0.5067789554595947, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0e88fdb5bb9845b8a9b7ec20f2c9943eddb837db34ed4c170827d4837bed1974.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "0f21315ba087b16a40278f2702755d33b72e8503dd96073c126fef514f2a8943", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "120abe3560b67e775a9e61089dce6da5b23e754887a8a9d99d1431763832cd6f", "generation_seconds_batch": 16.99920522305183, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7357056140899658, "target_z": -0.01953126893332539, "ce": 6.125868320465088, "genuineness": 4.281409740447998, "identity": 0.6422605514526367, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0f21315ba087b16a40278f2702755d33b72e8503dd96073c126fef514f2a8943.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "0f4264b343b2d0c2d98fdda723df68c702b68487bd2486e032a087b29a0f2b48", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e498becc75dbef68bd73be1fc1b14aa6617e8da1732ccb4cc3efbe43f7ab42fe", "generation_seconds_batch": 22.322724999976344, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8980697393417358, "target_z": 0.0820312549517928, "ce": 5.982900142669678, "genuineness": 4.6956281661987305, "identity": 0.7610126733779907, "hypothesis": "J'ai trouvé la petite boîte elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0f4264b343b2d0c2d98fdda723df68c702b68487bd2486e032a087b29a0f2b48.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "0f7833a66859802383c555ce16b5a54359b81bb881c86a2e4768bf337f694c7f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "eb6150a29b1e449f5b2480fc9676d51b44121fe509d817a13310fad7437c1273", "generation_seconds_batch": 22.96887306100689, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9917412996292114, "target_z": 0.14062498718359512, "ce": 5.982900142669678, "genuineness": 4.65797233581543, "identity": 0.5747431516647339, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0f7833a66859802383c555ce16b5a54359b81bb881c86a2e4768bf337f694c7f.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "0f7ad3ab1710a9e847f1b0c750e9fb5807ee11b28a31625b16adbb1a78c062d1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d7653816a1c02bea44d09b73b907c5c50b3de10bd82155c53324ad69e5a89daf", "generation_seconds_batch": 22.169052313896827, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9909048080444336, "target_z": 0.765624921936443, "ce": 5.355429172515869, "genuineness": 4.356721878051758, "identity": 0.4019739627838135, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/0f7ad3ab1710a9e847f1b0c750e9fb5807ee11b28a31625b16adbb1a78c062d1.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "1038daa1905e99bcbb4969e19ace9989fe5948e9965900a454be45da22ce9c3e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "73deb6274e90c29bf33f8bed279176247074d8437fbd480b270ab6a5861a4fe1", "generation_seconds_batch": 21.505827025976032, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.403059959411621, "target_z": 1.0234375227199906, "ce": 5.589737892150879, "genuineness": 4.281409740447998, "identity": 0.48591065406799316, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1038daa1905e99bcbb4969e19ace9989fe5948e9965900a454be45da22ce9c3e.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "103b52b2368c62822e7ba4b20ec28ed095d0f2887bab3aa9a48c3564b7226d2f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "08b5ecd129a3fee672c029fd11e9c3edde4f8a19923adc172ad7c0f9a96c8bb8", "generation_seconds_batch": 23.529391730902717, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.531914234161377, "target_z": 0.4785156787415159, "ce": 6.125868320465088, "genuineness": 4.482243061065674, "identity": 0.23532436788082123, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/103b52b2368c62822e7ba4b20ec28ed095d0f2887bab3aa9a48c3564b7226d2f.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "10a4e038d9fa655c6c325e6a58c7d4383f598c4db616eed68e7be07ea0310fe2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "eaf0c8858f6e6dd3291c619b7f4263e195f33722697cab5672bddedabe33cc04", "generation_seconds_batch": 23.82416292501148, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7927449941635132, "target_z": -0.6093749941743615, "ce": 6.268836498260498, "genuineness": 4.482243061065674, "identity": 0.6838693618774414, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/10a4e038d9fa655c6c325e6a58c7d4383f598c4db616eed68e7be07ea0310fe2.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "10aa2407519ffbcb40067c75039cfb87a4a7b6a0cf608d8c8394ea92042b22db", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "14625cdaa50b37d06175b108bf117c2ab173e5db92e856ace86e5130ac4c10c4", "generation_seconds_batch": 22.669804868055508, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.456976890563965, "target_z": 0.43164063330153496, "ce": 5.982900142669678, "genuineness": 4.758388519287109, "identity": 0.3896526098251343, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/10aa2407519ffbcb40067c75039cfb87a4a7b6a0cf608d8c8394ea92042b22db.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "114d7ae47fd6b36f555749e98ae6ce51031eadbf9aafbba23d0bf15ce1f2abeb", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a3c14d0cea95bd4e4c306ad79a1d5c4a12e8f681fa18a392f414435c19c6c963", "generation_seconds_batch": 21.77072331495583, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.284409284591675, "target_z": 0.9492187753415279, "ce": 5.49045467376709, "genuineness": 4.65797233581543, "identity": 0.4406743049621582, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/114d7ae47fd6b36f555749e98ae6ce51031eadbf9aafbba23d0bf15ce1f2abeb.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "122129d607665e5da486a7cc4ec5bf43c7cc004f00a5a15805ee5e42a2ca280f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "60f6dfe8f5bb25a9dbcf9099e72261ffd0158256db246d155e71e5be3d623f7f", "generation_seconds_batch": 23.035718755912967, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.502976417541504, "target_z": 1.0859375833066318, "ce": 5.780362129211426, "genuineness": 4.65797233581543, "identity": 0.5777232646942139, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/122129d607665e5da486a7cc4ec5bf43c7cc004f00a5a15805ee5e42a2ca280f.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "1236db6640cb04accdf23158b465ff5cb6c32e3383c77e05faf999541151086e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "98343af767774e5bdd48112cc450f38ee0fd71ad486f68a2ccf15d75606c1616", "generation_seconds_batch": 21.233830463024788, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7028088569641113, "target_z": 1.210937406207219, "ce": 5.550024509429932, "genuineness": 4.7081804275512695, "identity": 0.695256233215332, "hypothesis": "La porte est encore ouverte, j'ai posé ton manteau sur la chaise, dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1236db6640cb04accdf23158b465ff5cb6c32e3383c77e05faf999541151086e.mp3" }, { "id": "heldout-nl-3", "language": "nl", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-3", "key": "1246e08c3a481a542a39fc7d61d489d057a38f0a2cf163bcfbd26ca3e49a50c2", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Natural clear speech.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Natural clear speech.) [3.2 seconds duration] We zetten het na de lunch terug.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "63f4302fae4115618a841ee41f425b97421c818f18561411836c2b93f77804ad", "generation_seconds_batch": 17.490973711013794, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6162185668945312, "target_z": 0.5312499930092337, "ce": 5.919358730316162, "genuineness": 4.381826400756836, "identity": 0.49827858805656433, "hypothesis": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier naast de hoge boekenkast. We zetten het na de lunch terug.", "wer": 0.0, "duration_error_s": 0.120000000000001, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1246e08c3a481a542a39fc7d61d489d057a38f0a2cf163bcfbd26ca3e49a50c2.mp3" }, { "id": "heldout-es-0", "language": "es", "sentences": [ "Tu mensaje llegó después de que me fuera.", "Lo leí dos veces... y guardé el teléfono.", "Dime cuándo podemos hablar." ], "text": "Tu mensaje llegó después de que me fuera. Lo leí dos veces... y guardé el teléfono. Dime cuándo podemos hablar.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-0", "key": "12a7b850c68b1e935d69d2b179f3c0df54c4775061935760902fd9f6cbfc7541", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo podemos hablar.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo podemos hablar.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c1d3cb28e41bf8facdd627aed1bb1abc06aae1e36c5292e8818f090b492033df", "generation_seconds_batch": 17.162874671979807, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1112284660339355, "target_z": -0.41015620019079013, "ce": 6.403861999511719, "genuineness": 4.369274139404297, "identity": 0.6018752455711365, "hypothesis": "Tu mensaje llegó después de que me fuera. Lo leí dos veces y guardé el teléfono. Dime cuándo podemos hablar.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/12a7b850c68b1e935d69d2b179f3c0df54c4775061935760902fd9f6cbfc7541.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "13630549e8e7d4cbf80c22465e26048f56c257a4023f05178e92a77d51291633", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "757eca5a4da647b5f4b5220651ee3a651fea6eb36aa569022d7a988a34f56689", "generation_seconds_batch": 21.32317853393033, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1884517669677734, "target_z": 0.26367194417945816, "ce": 5.780362129211426, "genuineness": 4.7834930419921875, "identity": 0.6823848485946655, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/13630549e8e7d4cbf80c22465e26048f56c257a4023f05178e92a77d51291633.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "13870f49b2557dc276d98a1727d858ab75502066a4e40ed000141496fafe2aef", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c4333e9f0db3826ba2f61393ea4022695360ee8af1b4b2c2440b0a78650a9bf9", "generation_seconds_batch": 23.11871398705989, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.3297953605651855, "target_z": -0.2734375413620339, "ce": 6.078212261199951, "genuineness": 4.425758361816406, "identity": 0.760034441947937, "hypothesis": "Ik heb het klein doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/13870f49b2557dc276d98a1727d858ab75502066a4e40ed000141496fafe2aef.mp3" }, { "id": "heldout-de-2", "language": "de", "sentences": [ "Die Straße führt hinter der Brücke nach links.", "Geh weiter... bis zu dem kleinen Holztor.", "Ich warte auf der anderen Seite." ], "text": "Die Straße führt hinter der Brücke nach links. Geh weiter... bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-2", "key": "139b8f730add44c53edc8f7ea730e8e8885889c5863ebecc2a7ce5cd7e42e7e3", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Natural clear speech.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Natural clear speech.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7e1c0760790a0cafedcae70b50b350c0c4952c78f6e74fed7ec20331d9b113bf", "generation_seconds_batch": 17.04693485388998, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.97217059135437, "target_z": 0.7539062351446216, "ce": 5.665193557739258, "genuineness": 4.519899368286133, "identity": 0.5128390789031982, "hypothesis": "Die Straße führt hinter der Brücke nach links, geh weiter bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/139b8f730add44c53edc8f7ea730e8e8885889c5863ebecc2a7ce5cd7e42e7e3.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "13cc44ed57d15e1bb27497f15aac8076429a74ccba6540071d33b3fee946952c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ab2e765917388f7402bbe9f3a1debe1db44c80d999f86f89849e268c3336e6f6", "generation_seconds_batch": 21.78151240502484, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.716135025024414, "target_z": 0.593750053595875, "ce": 5.649308204650879, "genuineness": 4.607763767242432, "identity": 0.7789683938026428, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/13cc44ed57d15e1bb27497f15aac8076429a74ccba6540071d33b3fee946952c.mp3" }, { "id": "heldout-en-1", "language": "en", "sentences": [ "The keys are on the kitchen table.", "I checked the drawer... then looked beside the window.", "Take them with you when you leave." ], "text": "The keys are on the kitchen table. I checked the drawer... then looked beside the window. Take them with you when you leave.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-1", "key": "146b0c34ed7bd4eb367c375e3d87a154e617f19f4a779c2a7117dd9f4c0411c1", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The keys are on the kitchen table.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Take them with you when you leave.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The keys are on the kitchen table.\n(Natural clear speech.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Natural clear speech.) [3.2 seconds duration] Take them with you when you leave.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b2582c302b2a089f3c79b2caf03bfaced3caf84ea56103d62ba252ec5c922df5", "generation_seconds_batch": 17.014255692018196, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.497567892074585, "target_z": 0.4570312456307711, "ce": 5.637393951416016, "genuineness": 4.570107460021973, "identity": 0.44302186369895935, "hypothesis": "The keys are on the kitchen table. I checked the drawer, then looked beside the window. Take them with you when you leave.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/146b0c34ed7bd4eb367c375e3d87a154e617f19f4a779c2a7117dd9f4c0411c1.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "14ba2cf201825da62ace65853c0aac749658be4d56d7e802071fa7c8ffb04679", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f4ae36f493a1f02de674a9e8dfbe3a39a369e20ce9e7e8309f3068640e2ca8c8", "generation_seconds_batch": 17.804169637034647, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0081896781921387, "target_z": -0.474609350386677, "ce": 6.109982967376709, "genuineness": 4.243753433227539, "identity": 0.6990828514099121, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/14ba2cf201825da62ace65853c0aac749658be4d56d7e802071fa7c8ffb04679.mp3" }, { "id": "heldout-de-2", "language": "de", "sentences": [ "Die Straße führt hinter der Brücke nach links.", "Geh weiter... bis zu dem kleinen Holztor.", "Ich warte auf der anderen Seite." ], "text": "Die Straße führt hinter der Brücke nach links. Geh weiter... bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-2", "key": "15585f3962d31fd159bb39a755593ab24aa0df87e4725d7772f40811258760f5", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "feed588620fe4bd7582855cea5f5fe7766097e07c8b396d4c167918ebabf4c48", "generation_seconds_batch": 17.24085193593055, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8052345514297485, "target_z": -0.6015624866010313, "ce": 6.332377910614014, "genuineness": 4.469690799713135, "identity": 0.556236743927002, "hypothesis": "Die Straße führt hinter der Brücke nach links. Geh weiter bis zum kleinen Holztor. Ich warte auf der anderen Seite.", "wer": 0.09523809523809523, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/15585f3962d31fd159bb39a755593ab24aa0df87e4725d7772f40811258760f5.mp3" }, { "id": "heldout-fr-3", "language": "fr", "sentences": [ "Quelqu’un a déplacé la petite chaise bleue.", "Elle était ici... près de la grande bibliothèque.", "Nous la remettrons après le déjeuner." ], "text": "Quelqu’un a déplacé la petite chaise bleue. Elle était ici... près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-3", "key": "162898e89247911b412ef8aaacab9ab2914f2f7034d20de605d7f00fab70bbb3", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Natural clear speech.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Natural clear speech.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5bfb0cbb226dbd27cdd34e5ab9fab04d56524a398eb2201c1b9dbe13ae7b8854", "generation_seconds_batch": 17.24302426399663, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3468570709228516, "target_z": 0.9882813132081786, "ce": 5.307773113250732, "genuineness": 4.507347106933594, "identity": 0.7688464522361755, "hypothesis": "Quelqu'un a déplacé la petite chaise bleue, elle était ici, près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/162898e89247911b412ef8aaacab9ab2914f2f7034d20de605d7f00fab70bbb3.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "16301ab5afdfec97a47bd839a7ccfeb89f9ef13096fbdb7721c9db5f062213d4", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2f5ded132a0bb9d0bf9b227808e69f00879bae9f6585d327372bc9f3347fd145", "generation_seconds_batch": 22.502366662956774, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5506484508514404, "target_z": 0.4902343655333373, "ce": 5.839932441711426, "genuineness": 4.406930446624756, "identity": 0.32990002632141113, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/16301ab5afdfec97a47bd839a7ccfeb89f9ef13096fbdb7721c9db5f062213d4.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "1662ba6181c6165186c4db3dbc86de6792e7d4b26235bb13ba8ec79bd049592b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "16d22a3e95b8e2294ff43eeba2f6c42d54f12416ca25c39b2063c6ed6d219d0a", "generation_seconds_batch": 17.15911556710489, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6653404235839844, "target_z": 1.187500032623576, "ce": 5.418970584869385, "genuineness": 4.68307638168335, "identity": 0.8076189756393433, "hypothesis": "J'ai trouvé la petite boîte, elle était prête la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.10526315789473684, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1662ba6181c6165186c4db3dbc86de6792e7d4b26235bb13ba8ec79bd049592b.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "1687d77229bc90056d6fe5da5587b15de588a40d7919d58399bc2710c97efcff", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c3d0ebdb00b8a0f0c0b72c06bdf420c93ce2944720fd4760b0139da87529c381", "generation_seconds_batch": 21.251901319948956, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9988224506378174, "target_z": -0.4804687683507615, "ce": 6.165581703186035, "genuineness": 4.419482707977295, "identity": 0.634586751461029, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1687d77229bc90056d6fe5da5587b15de588a40d7919d58399bc2710c97efcff.mp3" }, { "id": "heldout-fr-2", "language": "fr", "sentences": [ "La route tourne à gauche après le pont.", "Continue... jusqu’à la petite barrière en bois.", "Je t’attendrai de l’autre côté." ], "text": "La route tourne à gauche après le pont. Continue... jusqu’à la petite barrière en bois. Je t’attendrai de l’autre côté.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-2", "key": "16cc7a5092e5729cbb009613aa4045788a34cb2d49aa58cddec78c12e4f88b54", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Natural clear speech.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Natural clear speech.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "64426522c282ade44b3e67a47985ea6c0e7424d926026cfcd7e20884862d0b3b", "generation_seconds_batch": 16.1027845320059, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.321877956390381, "target_z": 0.9726562980615183, "ce": 5.0754499435424805, "genuineness": 4.369274139404297, "identity": 0.7966907024383545, "hypothesis": "La route tourne à gauche après le pont continue jusqu'à la petite barrière en bois je t'attendrai de l'autre côté.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/16cc7a5092e5729cbb009613aa4045788a34cb2d49aa58cddec78c12e4f88b54.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "17663568dd027adc5a737bea64e067a61c04357a6de0e6817e1130441e3d75bb", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2b45f598c07c482221005838d58f17f1bd5a3e050fbb3d132f07ea8078cccccc", "generation_seconds_batch": 22.95947557396721, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5015267133712769, "target_z": -0.1660155995128312, "ce": 6.348263263702393, "genuineness": 4.469690799713135, "identity": 0.536811113357544, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/17663568dd027adc5a737bea64e067a61c04357a6de0e6817e1130441e3d75bb.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "17ad061305bb259e7c77ff9e5a913149b6c1d2998d91e07599d56bffcf3a18c5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "64af0d42b343ec536330bf74435b6c836d7f439aa7b41bb42e018fdb2ace133f", "generation_seconds_batch": 20.2958447720157, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5279555320739746, "target_z": 1.101562598453292, "ce": 5.268059730529785, "genuineness": 4.369274139404297, "identity": 0.29966506361961365, "hypothesis": "I found uh the little box. It was beside the old window. We can look inside together.", "wer": 0.0625, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/17ad061305bb259e7c77ff9e5a913149b6c1d2998d91e07599d56bffcf3a18c5.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "17d8f03ed9784356e36d6cbcd7c8806508588ac24f807c709a76629afe5516fd", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7543bb2b89e5d06f66cb0d84d7feb7f2bbe736e92e915395b03333d94c08e0d5", "generation_seconds_batch": 21.657455754000694, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4530181884765625, "target_z": 1.054687553013311, "ce": 5.3514580726623535, "genuineness": 4.469690799713135, "identity": 0.516688346862793, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/17d8f03ed9784356e36d6cbcd7c8806508588ac24f807c709a76629afe5516fd.mp3" }, { "id": "heldout-fr-4", "language": "fr", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-4", "key": "1866e9e4e666d1fbb19458318abc4418afb01d9b081dccddd28d512d903f2a2e", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Natural clear speech.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Natural clear speech.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bc183876250dcfa8fb4624c77ca031010ef002be8ab32358974517db970b6809", "generation_seconds_batch": 23.00563480402343, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3468570709228516, "target_z": 0.9882813132081786, "ce": 5.111192226409912, "genuineness": 4.375550270080566, "identity": 0.6156798005104065, "hypothesis": "Euh j'ai écrit le numéro sur l'enveloppe, lis-le bien et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "wer": 0.045454545454545456, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1866e9e4e666d1fbb19458318abc4418afb01d9b081dccddd28d512d903f2a2e.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "18cdb106834e57952bc3d0dc5bbc76dae43d033abbaafc9479e35934cc436996", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6febba00cbbafed6b90b6ede65f17d85cb459d803c46a37dada59939a0f09250", "generation_seconds_batch": 16.588223517057486, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4444873332977295, "target_z": 0.42382812572820483, "ce": 5.673136234283447, "genuineness": 4.519899368286133, "identity": 0.6294440627098083, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/18cdb106834e57952bc3d0dc5bbc76dae43d033abbaafc9479e35934cc436996.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "1a14b6e279c24d9ad451c5f5328716a473030f457ea77f3e6a44123b65fd5623", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e583b68b44d82f5cdcddf5dc8cd6157f8e36881dafa870355ebe1ba65ec02b54", "generation_seconds_batch": 22.967213697033003, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6607683897018433, "target_z": -0.06640623980513248, "ce": 5.982900142669678, "genuineness": 4.545003414154053, "identity": 0.7989603281021118, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1a14b6e279c24d9ad451c5f5328716a473030f457ea77f3e6a44123b65fd5623.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "1a63618f31e13ba5d0204b51fce63c54ab927d975d2b1f4c7212a0b874d49233", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "492ece0905077d1178acf849bebdccf8242ceea6f33a3d5332ff8e0f2f53a318", "generation_seconds_batch": 23.134083525044844, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7107264995574951, "target_z": -0.0351562840799857, "ce": 5.8875885009765625, "genuineness": 4.833701133728027, "identity": 0.6384759545326233, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1a63618f31e13ba5d0204b51fce63c54ab927d975d2b1f4c7212a0b874d49233.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "1a7b9cf9ac29eb832f0992a7bef8159c39bea6c5bea85f640e64de6b39a1d4b4", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "471a2411c477cdecad78297e78dd24431e4a585119a4a932e6be6fae2f5b7cc2", "generation_seconds_batch": 21.785142901004292, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7028088569641113, "target_z": 1.210937406207219, "ce": 5.557967185974121, "genuineness": 4.65797233581543, "identity": 0.5854675769805908, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1a7b9cf9ac29eb832f0992a7bef8159c39bea6c5bea85f640e64de6b39a1d4b4.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "1ab14593e03d7d6a3de6cf38a92a22af6518a378f3f911bb1ea80d7ae1659828", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3f5b363c825da8876a6203fc17104ff114479f07208d01cce74439d180ca5182", "generation_seconds_batch": 23.82416292501148, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0862493515014648, "target_z": -0.42578121533745045, "ce": 6.062326908111572, "genuineness": 4.369274139404297, "identity": 0.7247989773750305, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1ab14593e03d7d6a3de6cf38a92a22af6518a378f3f911bb1ea80d7ae1659828.mp3" }, { "id": "heldout-nl-2", "language": "nl", "sentences": [ "De weg gaat linksaf na de brug.", "Loop verder... tot aan het kleine houten hek.", "Ik wacht aan de andere kant." ], "text": "De weg gaat linksaf na de brug. Loop verder... tot aan het kleine houten hek. Ik wacht aan de andere kant.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-2", "key": "1ac3f3669a7d17ba0d8ef82c9fe1e8f038eee3d50d114f9345d99a4a9b94d212", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ik wacht aan de andere kant.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Natural clear speech.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Natural clear speech.) [3.2 seconds duration] Ik wacht aan de andere kant.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "20d594a956c4cb18f52f89e3566487bd566b1e9ab7012970a2944cf3fa5586a8", "generation_seconds_batch": 16.42439042206388, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8722543716430664, "target_z": 0.6914063236943281, "ce": 5.6413655281066895, "genuineness": 4.413206577301025, "identity": 0.336036741733551, "hypothesis": "De weg gaat linksaf. Na de brug. Loop verder tot aan het kleine houten hek. Ik wacht aan de andere kant.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1ac3f3669a7d17ba0d8ef82c9fe1e8f038eee3d50d114f9345d99a4a9b94d212.mp3" }, { "id": "heldout-fr-0", "language": "fr", "sentences": [ "Ton message est arrivé après mon départ.", "Je l’ai relu... puis j’ai rangé mon téléphone.", "Dis-moi quand nous pourrons parler." ], "text": "Ton message est arrivé après mon départ. Je l’ai relu... puis j’ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-0", "key": "1c6342414a49079a2b699d19dc339607742de140978cc5c174098e1ad5903064", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Natural clear speech.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d00d24bc8c8880d35c8c92c1766b274a2a0da32a8a542404d70f44d3506a04e0", "generation_seconds_batch": 16.504592689918354, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.615382194519043, "target_z": 1.1562500023302553, "ce": 5.649308204650879, "genuineness": 4.582659721374512, "identity": 0.7673014998435974, "hypothesis": "Ton message est arrivé après mon départ. Je l'ai relu, puis j'ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1c6342414a49079a2b699d19dc339607742de140978cc5c174098e1ad5903064.mp3" }, { "id": "heldout-es-4", "language": "es", "sentences": [ "Escribí el número en el sobre.", "Léelo bien... y comprueba las dos últimas cifras.", "Llámame cuando tengas un momento." ], "text": "Escribí el número en el sobre. Léelo bien... y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-4", "key": "1d559e6da25f8bbacc3b53d5880369445633ce7ab0125ca7b07093108b8c8de3", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Escribí el número en el sobre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Llámame cuando tengas un momento.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Escribí el número en el sobre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Llámame cuando tengas un momento.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bec25b398c174a8d32c730ddd6ada42e01ee8df85c1bdc9a086a94d3e0e2d126", "generation_seconds_batch": 17.347987634013407, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.4805063009262085, "target_z": -0.8046875343712676, "ce": 6.435632705688477, "genuineness": 4.557555198669434, "identity": 0.47852030396461487, "hypothesis": "Escribí el número en el sobre. Léelo bien y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1d559e6da25f8bbacc3b53d5880369445633ce7ab0125ca7b07093108b8c8de3.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "1d5655ccc41dc404f84881ef7081af192b8e2e0d25ddbff6f2a1b887276099cc", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2f6484502c914dd9584cbdba477f12de890525833a869ee9ebcc89ac0541b59d", "generation_seconds_batch": 22.442758857971057, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.328122615814209, "target_z": 0.9765624772800096, "ce": 5.772419452667236, "genuineness": 4.432034492492676, "identity": 0.46681252121925354, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1d5655ccc41dc404f84881ef7081af192b8e2e0d25ddbff6f2a1b887276099cc.mp3" }, { "id": "heldout-de-1", "language": "de", "sentences": [ "Die Schlüssel liegen auf dem Küchentisch.", "Ich suchte in der Schublade... dann neben dem Fenster.", "Nimm sie mit, wenn du gehst." ], "text": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade... dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-1", "key": "1d829402ab2176d2024ae3e918286a39ec094d9a5299caa0e048ba37912e1579", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6456e39ada3e5838d257d322bad485dabb94a5e7fd15ad58caab915aaa8f3973", "generation_seconds_batch": 15.367235821904615, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8926613330841064, "target_z": -0.546875008155894, "ce": 6.324435234069824, "genuineness": 4.432034492492676, "identity": 0.5548044443130493, "hypothesis": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade, dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1d829402ab2176d2024ae3e918286a39ec094d9a5299caa0e048ba37912e1579.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "1e114f1e90c421a63a934980384b21813819cc336675631fc223820d30e8c661", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2b1d0318c34e1a0c80afffa79fd8999c7be52597b065221a00d40c8c64d98b13", "generation_seconds_batch": 23.13597337307874, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0916576385498047, "target_z": 0.20312497320206252, "ce": 5.935244083404541, "genuineness": 4.444586753845215, "identity": 0.6697492003440857, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1e114f1e90c421a63a934980384b21813819cc336675631fc223820d30e8c661.mp3" }, { "id": "heldout-es-1", "language": "es", "sentences": [ "Las llaves están en la mesa de la cocina.", "Miré en el cajón... y luego junto a la ventana.", "Llévatelas cuando te vayas." ], "text": "Las llaves están en la mesa de la cocina. Miré en el cajón... y luego junto a la ventana. Llévatelas cuando te vayas.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-1", "key": "1e4389edade858d49293d7db0a93a54d8b780b1d1589bd3f07ff65898fa4e89e", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Llévatelas cuando te vayas.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Natural clear speech.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Natural clear speech.) [3.2 seconds duration] Llévatelas cuando te vayas.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "9c1e332842b0d0c92264e86c5b360c1298b0e3c8b266ba0065a4c6ce67b53f99", "generation_seconds_batch": 17.60006375994999, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2344510555267334, "target_z": 0.9179687450482072, "ce": 5.0774359703063965, "genuineness": 4.268857479095459, "identity": 0.5889833569526672, "hypothesis": "Las llaves están en la mesa de la cocina. Mira en el cajón y luego junto a la ventana. Llévatelas cuando te vayas.", "wer": 0.043478260869565216, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1e4389edade858d49293d7db0a93a54d8b780b1d1589bd3f07ff65898fa4e89e.mp3" }, { "id": "heldout-es-4", "language": "es", "sentences": [ "Escribí el número en el sobre.", "Léelo bien... y comprueba las dos últimas cifras.", "Llámame cuando tengas un momento." ], "text": "Escribí el número en el sobre. Léelo bien... y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-4", "key": "1eaf0e0e2c966186cf6ada2de41819a315f8b1b1eeeeef5a84a28dd582998db8", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Escribí el número en el sobre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Llámame cuando tengas un momento.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Escribí el número en el sobre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Llámame cuando tengas un momento.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "89f36752c9c1e544cddbd938996fcedcfc854300061e7bb57e1b99723e2d9b5e", "generation_seconds_batch": 17.347987634013407, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6865838766098022, "target_z": -0.6757812339794939, "ce": 6.284721851348877, "genuineness": 4.432034492492676, "identity": 0.6416370868682861, "hypothesis": "Escribí el número en el sobre, léelo bien, y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1eaf0e0e2c966186cf6ada2de41819a315f8b1b1eeeeef5a84a28dd582998db8.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "1eafc706a0b6c860b87a74948c037f122fcef0b252195efc34bdc2c61ac329eb", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2f65d2f72eae710d70af6d3801e8c9e173827a0f0b1e4fc9dc00ce6d253dff48", "generation_seconds_batch": 24.00437844498083, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4507322311401367, "target_z": 0.4277344540830437, "ce": 5.708878040313721, "genuineness": 4.871357440948486, "identity": 0.5484064221382141, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1eafc706a0b6c860b87a74948c037f122fcef0b252195efc34bdc2c61ac329eb.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "1f33d9f585dbd0e496e477935efdf7237b7775a58782ed631ce5d0a3738b6084", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2e7d5f1d38e9c5cdb62001dd1f8a77673d18f5ab58269847679a94719490387d", "generation_seconds_batch": 16.284019420971163, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7606847286224365, "target_z": -0.003906253786665078, "ce": 5.990842819213867, "genuineness": 4.381826400756836, "identity": 0.529251217842102, "hypothesis": "Ik heb het kleine doosje gevonden, het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1f33d9f585dbd0e496e477935efdf7237b7775a58782ed631ce5d0a3738b6084.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "1faf8734db5168daf5d3d2844db95f09bf0497730dbcf5b3884f71ef45da069d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "28107dc2021f41794f1f5ef31b0403473ac5c270e1ca07ab2355af6d7eb227c0", "generation_seconds_batch": 22.09680682606995, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6428704261779785, "target_z": -0.7031250104861494, "ce": 6.125868320465088, "genuineness": 4.306513786315918, "identity": 0.6490633487701416, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1faf8734db5168daf5d3d2844db95f09bf0497730dbcf5b3884f71ef45da069d.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "1fbf6f78502017f9d4edd4a5b5d0ec78060f4e283949e9f269f8bc3896e6f005", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "453315ed38191b7625b1d6d43b1600497c79fd60baab870819c62a7b8729f016", "generation_seconds_batch": 15.957766006002203, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6278717517852783, "target_z": 1.1640625099035855, "ce": 5.486483573913574, "genuineness": 4.65797233581543, "identity": 0.7243961095809937, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1fbf6f78502017f9d4edd4a5b5d0ec78060f4e283949e9f269f8bc3896e6f005.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "1ff78c29c2dd8ec94b4ecd3a95a22ca855b7ce9d154aac14b1f2157a6b7c8d57", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f712b3cd3c20607cd860606724c381b41ef7ab467709e0dc092f8275a6787515", "generation_seconds_batch": 22.318794004037045, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0229651927948, "target_z": 0.1601562561169205, "ce": 6.109982967376709, "genuineness": 4.545003414154053, "identity": 0.8190362453460693, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/1ff78c29c2dd8ec94b4ecd3a95a22ca855b7ce9d154aac14b1f2157a6b7c8d57.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "202fc868fea588a5e63958996f9de330d380a5c27c09350e6c98e14b33e7e953", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e78891467fc842e90576b4304e0fe3adc94a46b87c847763cd895ec1a91c8908", "generation_seconds_batch": 16.837456090957858, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.940110206604004, "target_z": 1.359374900964144, "ce": 5.756534099578857, "genuineness": 4.7834930419921875, "identity": 0.7991219758987427, "hypothesis": "La porte était encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/202fc868fea588a5e63958996f9de330d380a5c27c09350e6c98e14b33e7e953.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "2062a4e7514d96fb119e59cdd4dfda54a6c01e4e13b7a677a21ba296f7e6e8fe", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "87963d8250123e2999fb05b041de99d4cda1ffe21c0654d8125d662bcd9117f7", "generation_seconds_batch": 21.07243009796366, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.965925693511963, "target_z": 0.7499999067897827, "ce": 5.538110733032227, "genuineness": 4.909013748168945, "identity": 0.5779494643211365, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2062a4e7514d96fb119e59cdd4dfda54a6c01e4e13b7a677a21ba296f7e6e8fe.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "20e57638e3d47934efc4c52a2294a6ba0c6cce0d3e224e591720d8b7d3bed592", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "afc50fa30dc93a0d6bcbe064d347440265a7ec613323bfaa12c86d1cfe175cbe", "generation_seconds_batch": 20.833433876978233, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3008575439453125, "target_z": 0.3339843632030819, "ce": 5.95112943649292, "genuineness": 4.8588056564331055, "identity": 0.8431738615036011, "hypothesis": "J'ai trouvé la petite boîte elle était près de la vieille fenêtre nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/20e57638e3d47934efc4c52a2294a6ba0c6cce0d3e224e591720d8b7d3bed592.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "218d055b25d263989a28f5ecd4872de2c97c16eddcaf485a0ea7f7062f1d830c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f6bf5c47771ad824287735fbfc4a2db74dc0e691e9879df43f7695c6b82e321f", "generation_seconds_batch": 16.80239176400937, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.39932429790496826, "target_z": -0.8554687590297398, "ce": 6.1020402908325195, "genuineness": 4.432034492492676, "identity": 0.42484360933303833, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/218d055b25d263989a28f5ecd4872de2c97c16eddcaf485a0ea7f7062f1d830c.mp3" }, { "id": "heldout-en-3", "language": "en", "sentences": [ "Someone moved the little blue chair.", "It used to stand here... beside the tall bookshelf.", "We can put it back after lunch." ], "text": "Someone moved the little blue chair. It used to stand here... beside the tall bookshelf. We can put it back after lunch.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-3", "key": "21e09fe748873199acfeca16d94e944f8f54c868940315a234d8052695bda4cd", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Someone moved the little blue chair.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can put it back after lunch.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Someone moved the little blue chair.\n(Natural clear speech.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Natural clear speech.) [3.2 seconds duration] We can put it back after lunch.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "dfd991c6790e048b9a2b5602bc4c9642be6f3b33c482459e8b732dc6292f1154", "generation_seconds_batch": 16.709362261928618, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.190737724304199, "target_z": 0.8906250431097255, "ce": 5.927301406860352, "genuineness": 4.507347106933594, "identity": 0.28831949830055237, "hypothesis": "Someone moved the little blue chair. Used to stand here beside the top bookshelf. We can put it back after lunch.", "wer": 0.09090909090909091, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/21e09fe748873199acfeca16d94e944f8f54c868940315a234d8052695bda4cd.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "21e4e3b072998506420344c5c3924e3b3602193eb1894501a349f11ec4f62478", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e322f1ca3b89ba1e08d07b6ee6d77372b70825ed3a42883486db54cef410eff8", "generation_seconds_batch": 22.269744602032006, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2032270431518555, "target_z": 0.8984374015467079, "ce": 5.824047088623047, "genuineness": 4.331617832183838, "identity": 0.505315363407135, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/21e4e3b072998506420344c5c3924e3b3602193eb1894501a349f11ec4f62478.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "220c91dbd3abbf3326d16e4cb0f0e6413c2b502f775b557a40064cceb608c04b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7b8f8173994fe9f31bb581d32fbc8cb35403cd6197d4e3869ebcd4f5cc72d049", "generation_seconds_batch": 21.694796059047803, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6661767959594727, "target_z": 0.5625000233025543, "ce": 5.784333229064941, "genuineness": 4.720732688903809, "identity": 0.7424822449684143, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/220c91dbd3abbf3326d16e4cb0f0e6413c2b502f775b557a40064cceb608c04b.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "22694729c25fe4660406bc1a144008154a23b4f8139ff34e1650e8239defe802", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7251fc31e8d0fc801780985d0d446018315dd376a086ed8985cd2d192be6ff3b", "generation_seconds_batch": 22.827060448005795, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8035616874694824, "target_z": 0.6484374574728383, "ce": 5.3792572021484375, "genuineness": 4.532451152801514, "identity": 0.4640847146511078, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/22694729c25fe4660406bc1a144008154a23b4f8139ff34e1650e8239defe802.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "22a7e0c6cd67dc9d0ebfee3861585d823e3bbfb03567d29a8806b67e5fa53be6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "ed32c914ee27923036270c3600ba8520700f03b14fd41345b6f78722c41640f3", "generation_seconds_batch": 22.95977593702264, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1112284660339355, "target_z": -0.41015620019079013, "ce": 6.268836498260498, "genuineness": 4.268857479095459, "identity": 0.17724035680294037, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/22a7e0c6cd67dc9d0ebfee3861585d823e3bbfb03567d29a8806b67e5fa53be6.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "22ee742056ba657dbe81110cc6e259f00b47bca0d2b035846aa73cf00fb7749c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9c0137319ad2bbeeca3190d3e264e999468d4a1d2a0b330f99821fc6db55ec05", "generation_seconds_batch": 21.07243009796366, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0908212661743164, "target_z": 0.8281249825230843, "ce": 5.303802013397217, "genuineness": 4.557555198669434, "identity": 0.38981905579566956, "hypothesis": "he encontrado la cajita. estaba junto a la ventana vieja. podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/22ee742056ba657dbe81110cc6e259f00b47bca0d2b035846aa73cf00fb7749c.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "22f1f1f9aad6dd6fb2b445d54a5feef2ff7d9c6969afd44450d9e91f6b7f374b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "82200cbac49daebfff18b47190584a4d2dc6534365c8d265687093b89c2e78ee", "generation_seconds_batch": 21.356206362019293, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1455745697021484, "target_z": -0.38867191621639297, "ce": 6.300607204437256, "genuineness": 4.6705241203308105, "identity": 0.6047165989875793, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/22f1f1f9aad6dd6fb2b445d54a5feef2ff7d9c6969afd44450d9e91f6b7f374b.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "2325c56b86da9df404f13eae16c806357b0d5049df9387a2cdb789731dbbc758", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4afd86da3ec66bdef25daae209de96b307ec771d8205058a074a8cc45c4053aa", "generation_seconds_batch": 22.060450674965978, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5654239654541016, "target_z": 1.1249999720369348, "ce": 5.411027908325195, "genuineness": 4.482243061065674, "identity": 0.5188605785369873, "hypothesis": "La puerta sigue abierta. he dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2325c56b86da9df404f13eae16c806357b0d5049df9387a2cdb789731dbbc758.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "23c038ba83beac89396bcfd400a6daacb6d33e72b7335c272871148175735e13", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "573c7c8ce5fb0cb90c079f23328d68cf28efc27de846af094f1179139abaa714", "generation_seconds_batch": 17.040694979019463, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6028926372528076, "target_z": 1.1484374947569254, "ce": 5.208489894866943, "genuineness": 4.7081804275512695, "identity": 0.8677860498428345, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/23c038ba83beac89396bcfd400a6daacb6d33e72b7335c272871148175735e13.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "246dc20696473dd52c8a2b1e5996e222867be2d172eee5731240a292fdef5e2d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7c7cd1d7ab5d62d04be4139da3697737f41b536720c978bc264997ca588a369e", "generation_seconds_batch": 21.78151240502484, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.385161876678467, "target_z": 0.3867186774707997, "ce": 6.09409761428833, "genuineness": 4.6956281661987305, "identity": 0.6179889440536499, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/246dc20696473dd52c8a2b1e5996e222867be2d172eee5731240a292fdef5e2d.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "2535f6353f22eaf1b7e5a50cc80030f09019fefe25c57adafb7099d3f6a27d5a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a3c778305022df13df7cf8887b8bddd28f7942fed8f5b01a435921d0d9950dd2", "generation_seconds_batch": 22.954712528036907, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.89098858833313, "target_z": 0.7031250104861494, "ce": 5.331601142883301, "genuineness": 4.356721878051758, "identity": 0.4742790460586548, "hypothesis": "De, de deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2535f6353f22eaf1b7e5a50cc80030f09019fefe25c57adafb7099d3f6a27d5a.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "25a091add53e6c6d3756fb64239fac0d5ab74b1308dcc4a1222322a9d74ec46d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "577cf4c3852abbb3f35861ec8502021bbb3dbc22d0448a18d0d3839f267b7bcc", "generation_seconds_batch": 23.343918393948115, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.207185983657837, "target_z": 0.2753906309712795, "ce": 5.967014789581299, "genuineness": 4.758388519287109, "identity": 0.76223224401474, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/25a091add53e6c6d3756fb64239fac0d5ab74b1308dcc4a1222322a9d74ec46d.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "25ad3f373cc352de50bfe26802e8247102497784a1f10fdf02991079edc3af2c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cf5129d16c794e4dcfb92c865c55d5e5634de7605ca11d60f5a2d85db796cb6e", "generation_seconds_batch": 23.11871398705989, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8168877363204956, "target_z": 0.03125003029332062, "ce": 6.165581703186035, "genuineness": 4.381826400756836, "identity": 0.6800475120544434, "hypothesis": "Ik heb het een kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/25ad3f373cc352de50bfe26802e8247102497784a1f10fdf02991079edc3af2c.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "25aee0b89a22412e597f31784aaa77a63d5ea795db848c093ad07255b06cdd00", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "de96df0ec935c8527090b9db0f195a208b9865488e583dba9e456d8cc35f5e9f", "generation_seconds_batch": 22.41607297293376, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8035616874694824, "target_z": 0.6484374574728383, "ce": 6.165581703186035, "genuineness": 4.883909225463867, "identity": 0.8194754123687744, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/25aee0b89a22412e597f31784aaa77a63d5ea795db848c093ad07255b06cdd00.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "262148d198367b44949ae1ffcc28c4a57cb70f7148b76a8218553de25d932b74", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5c9cae77d94d0830b45f61fbe1edb930c47e158157d8b134ef0711f4bfe8c08d", "generation_seconds_batch": 21.132345052901655, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6116465330123901, "target_z": -0.7226562794194749, "ce": 6.23706579208374, "genuineness": 4.419482707977295, "identity": 0.4172922372817993, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/262148d198367b44949ae1ffcc28c4a57cb70f7148b76a8218553de25d932b74.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "2653570399a63e5bbfbe6e05fd33e50963f2ff12011c8b51d5e1da9036f117f1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2869f4862a8c12631cca1678df1175fa764b859a47c65052372aaa23ba3631e0", "generation_seconds_batch": 28.466669535031542, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.928457260131836, "target_z": 0.72656253320614, "ce": 5.498397350311279, "genuineness": 4.65797233581543, "identity": 0.5934620499610901, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2653570399a63e5bbfbe6e05fd33e50963f2ff12011c8b51d5e1da9036f117f1.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "26923c865a23da6fe0b96cb851b7dc53bc09c2f26da1cbd26e900753e7b849f7", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "06117eceb332b0041948c31ceca513304f60fa2d9cadae0270fbf487769adc69", "generation_seconds_batch": 17.040694979019463, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7402777671813965, "target_z": 1.234375078063557, "ce": 5.637393951416016, "genuineness": 4.758388519287109, "identity": 0.693652868270874, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/26923c865a23da6fe0b96cb851b7dc53bc09c2f26da1cbd26e900753e7b849f7.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "26990322bee50a4126b34128d2345b045db000e40a0d47b4108cc5a27b55e1f3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "17c6760b0d6cc98753724a4c4e2b71a652eb4a26c186dfd40b807d96177abc1c", "generation_seconds_batch": 16.637231742963195, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9534363746643066, "target_z": 0.7421875483528002, "ce": 5.982900142669678, "genuineness": 4.344170093536377, "identity": 0.6236730217933655, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/26990322bee50a4126b34128d2345b045db000e40a0d47b4108cc5a27b55e1f3.mp3" }, { "id": "heldout-en-1", "language": "en", "sentences": [ "The keys are on the kitchen table.", "I checked the drawer... then looked beside the window.", "Take them with you when you leave." ], "text": "The keys are on the kitchen table. I checked the drawer... then looked beside the window. Take them with you when you leave.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-1", "key": "26d12995a5b024b42cbe037fdee48ffb67b79df656c09f059592185aa881aca9", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The keys are on the kitchen table.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Take them with you when you leave.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The keys are on the kitchen table.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Take them with you when you leave.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9046d6a0b10a32a3aa84b44d3ac16d37b5ff30b3b0fac26f1822410da2178774", "generation_seconds_batch": 17.89829301799182, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.42430341243743896, "target_z": -0.8398437438830795, "ce": 6.459460258483887, "genuineness": 4.118232727050781, "identity": 0.2920753061771393, "hypothesis": "The keys are on the kitchen table. I checked the drawer, then looked beside the window. Take them with you when you leave.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/26d12995a5b024b42cbe037fdee48ffb67b79df656c09f059592185aa881aca9.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "26da361c7684029956e3b53def9dd64149227d09b5298c81756937a7812d6402", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "35d955bfb79ed043cfb20a0c4704bfed6cbb0a378cf79194c4b193663d82e83c", "generation_seconds_batch": 22.95977593702264, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6982370615005493, "target_z": -0.042968717085142015, "ce": 6.157639026641846, "genuineness": 4.293961524963379, "identity": 0.6178023815155029, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/26da361c7684029956e3b53def9dd64149227d09b5298c81756937a7812d6402.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "27313882f5e4f8e882c805696e18138b97689d81526ef7d6397aa2aa073dd745", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4d87f963e6bce1eda83bf995d1184cab024c6f39deeca27df57f74259fce4f09", "generation_seconds_batch": 23.79720372101292, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4913229942321777, "target_z": 0.4531249172759321, "ce": 5.919358730316162, "genuineness": 4.6203155517578125, "identity": 0.6300342082977295, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/27313882f5e4f8e882c805696e18138b97689d81526ef7d6397aa2aa073dd745.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "277c38d0a8630df48a3ba4bd79ed696aee2f0089a54b4c15ad95a80d59b774c3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "93aead0ab2f3a6edaf10ba1b3ce3a1b999b09a634c2b2293c79b72bf99e93d7b", "generation_seconds_batch": 21.672672353102826, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4453237056732178, "target_z": -0.2011718835928169, "ce": 6.2767791748046875, "genuineness": 4.519899368286133, "identity": 0.6486901044845581, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/277c38d0a8630df48a3ba4bd79ed696aee2f0089a54b4c15ad95a80d59b774c3.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "28252b21f0bf2d62adfc67294f3890e529e54bdb08340a6f2e03796d24e443d7", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "31986f64e1c293ce3f701b8fd2cb7b3d0b5e9779ee3d24a3382c5a99db1998c5", "generation_seconds_batch": 22.764518029987812, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9596810340881348, "target_z": 0.7460937275712914, "ce": 5.593709468841553, "genuineness": 4.720732688903809, "identity": 0.5717929601669312, "hypothesis": "He encontrado la cajita. estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/28252b21f0bf2d62adfc67294f3890e529e54bdb08340a6f2e03796d24e443d7.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "28b823e67df289d829d60b4449dc5c8797ac1fc81057c3ee7bd324e2a04b5ccb", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "dcb880f49638773b95328b74786b808ca38b3d7f5b2bd7edd2c8ad5af830d957", "generation_seconds_batch": 17.386136341956444, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5889534950256348, "target_z": -0.11132812106769396, "ce": 6.078212261199951, "genuineness": 4.469690799713135, "identity": 0.7308146953582764, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/28b823e67df289d829d60b4449dc5c8797ac1fc81057c3ee7bd324e2a04b5ccb.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "28cba2a07e867adfb8f811b20ceb77222a2f2ab2f7f6ae91f45147b1035cc17c", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "18414f14c3b8e34e6e8af7a8eac605e9209ced09be5002cd3c7e10c39f2804bd", "generation_seconds_batch": 17.09129001107067, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.166594982147217, "target_z": 0.25000001864204346, "ce": 6.316492557525635, "genuineness": 4.394378662109375, "identity": 0.2716919183731079, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/28cba2a07e867adfb8f811b20ceb77222a2f2ab2f7f6ae91f45147b1035cc17c.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "2932897d3e1987f0b97bbee0b41ae39cf92aff9ed7ceafac552c489dafbf1df1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b9594daae69101fe4890bae922b3db1bca144c42bcdb203e007a5a622eefe649", "generation_seconds_batch": 16.017422315082513, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4913229942321777, "target_z": 0.4531249172759321, "ce": 5.8875885009765625, "genuineness": 4.595211505889893, "identity": 0.8080785870552063, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2932897d3e1987f0b97bbee0b41ae39cf92aff9ed7ceafac552c489dafbf1df1.mp3" }, { "id": "heldout-fr-2", "language": "fr", "sentences": [ "La route tourne à gauche après le pont.", "Continue... jusqu’à la petite barrière en bois.", "Je t’attendrai de l’autre côté." ], "text": "La route tourne à gauche après le pont. Continue... jusqu’à la petite barrière en bois. Je t’attendrai de l’autre côté.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-2", "key": "29455a827143b448f4511dfb339f119e26edfd24d35261bbe1e6ea3e22db06a4", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Natural clear speech.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Natural clear speech.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2cfb3c8baa88ed9349336127e837aafc8d85b027372905da3fe2403c5f9219f1", "generation_seconds_batch": 16.705798128037713, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3406121730804443, "target_z": 0.9843749848533397, "ce": 5.089349746704102, "genuineness": 4.381826400756836, "identity": 0.7470126152038574, "hypothesis": "La route tourne à gauche après le pont, continue jusqu'à la petite barrière en bois. Je t'attendrai de l'autre côté.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/29455a827143b448f4511dfb339f119e26edfd24d35261bbe1e6ea3e22db06a4.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "29999ae67713a56ae9ee76d01cf91532d8b2c20afa7afb70135c4bc95ad51acd", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "283360d19763337476bf50ab70d4f9e51e124c5c100c5cfa7cfa108a6a6e49e6", "generation_seconds_batch": 21.726068218937144, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.140779495239258, "target_z": 0.8593750128164048, "ce": 4.943403244018555, "genuineness": 4.375550270080566, "identity": 0.28571921586990356, "hypothesis": "La puerta sigue abierta, he dejado tu abrigo sobre la silla, dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/29999ae67713a56ae9ee76d01cf91532d8b2c20afa7afb70135c4bc95ad51acd.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "29a31e8a983cb110a2490f54cede27ce0d92072ba9df7b83fbe3044a16f7a3e5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7aac018e81e2e8a30b702fd0fd8dd7c8a7d11e0582c5db75c0af4b04d22e6a7a", "generation_seconds_batch": 16.763441997929476, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7325832843780518, "target_z": -0.021484358542571007, "ce": 6.038498878479004, "genuineness": 4.243753433227539, "identity": 0.5728932023048401, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/29a31e8a983cb110a2490f54cede27ce0d92072ba9df7b83fbe3044a16f7a3e5.mp3" }, { "id": "heldout-en-4", "language": "en", "sentences": [ "I wrote the number on the envelope.", "Read it carefully... and check the last two digits.", "Call me when you have a moment." ], "text": "I wrote the number on the envelope. Read it carefully... and check the last two digits. Call me when you have a moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-4", "key": "2a0d905b6aab762ee9561a89fcefac19b8d2df8845afefbb5f240dbc9e23ad65", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I wrote the number on the envelope.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Call me when you have a moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I wrote the number on the envelope.\n(Natural clear speech.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Natural clear speech.) [3.2 seconds duration] Call me when you have a moment.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f91794ce071e84e1f5e01a5d185b807514f40fbd674ba4bf5cd7288268b3eb7b", "generation_seconds_batch": 16.477753099054098, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.060433864593506, "target_z": 0.18359377883691097, "ce": 5.2065043449401855, "genuineness": 3.9676074981689453, "identity": 0.48370909690856934, "hypothesis": "I wrote the number on the envelope, read it carefully, and check the last two digits. Call me when you have a moment.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2a0d905b6aab762ee9561a89fcefac19b8d2df8845afefbb5f240dbc9e23ad65.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "2a20b6068f0f713d8da6771c70262fc659faad2fe3975d8194da68de8d683b76", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "28d2bada369628d7e1bfa710e0150c8f228101e71f71aabc9ebee4c39e959d92", "generation_seconds_batch": 21.136349080945365, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.960517406463623, "target_z": 0.12109371825026974, "ce": 6.125868320465088, "genuineness": 4.457139015197754, "identity": 0.4037590026855469, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2a20b6068f0f713d8da6771c70262fc659faad2fe3975d8194da68de8d683b76.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "2a281f4320c36e6b405a83c00348740ed7d9c6ac02651a1c954cdbff217d206a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1c7556e98bd4da2f25e8889de70c27b7c6f66de2764d8fbf943adb216611a7d5", "generation_seconds_batch": 22.853756654076278, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2344510555267334, "target_z": 0.9179687450482072, "ce": 5.569881439208984, "genuineness": 4.65797233581543, "identity": 0.6868906617164612, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2a281f4320c36e6b405a83c00348740ed7d9c6ac02651a1c954cdbff217d206a.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "2b3ed9b59e05d96cbe9ea834bb6359dcbe4cd6f6217b220cd7af7aaa240cb53c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b57df776329778105c56ad99783f49bd788fe8cb48cafd851ecd53609dd1fa26", "generation_seconds_batch": 16.535797354998067, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3406121730804443, "target_z": 0.9843749848533397, "ce": 4.95531702041626, "genuineness": 4.406930446624756, "identity": 0.5004988312721252, "hypothesis": "La puerta sigue abierta, he dejado tu abrijo sobre la silla, dime cuándo estás lista.", "wer": 0.13333333333333333, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2b3ed9b59e05d96cbe9ea834bb6359dcbe4cd6f6217b220cd7af7aaa240cb53c.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "2b56aae666c72fa00cd2fac965fe2eb95089f67cef82fdef59bf34730368678a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "58f17965393053fb7dfdb3a2b1757a161cb27984ef4c1cd0a6f93fd272720db9", "generation_seconds_batch": 24.147991469944827, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1237179040908813, "target_z": -0.4023437671856338, "ce": 6.181467056274414, "genuineness": 4.8588056564331055, "identity": 0.5452346801757812, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2b56aae666c72fa00cd2fac965fe2eb95089f67cef82fdef59bf34730368678a.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "2b7d690a3ae3bdccba89a8bd0f9446d2b9e83d4b5850aa9412950ca87dfaa30b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "10847b7e3802a1be157bf0f015afcfe8f65abdc9b84bbf5bc4810da2066bfcfb", "generation_seconds_batch": 16.80239176400937, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0175567865371704, "target_z": -0.4687500069907663, "ce": 6.1894097328186035, "genuineness": 4.582659721374512, "identity": 0.5760306715965271, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2b7d690a3ae3bdccba89a8bd0f9446d2b9e83d4b5850aa9412950ca87dfaa30b.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "2bd806fd839fbabb2be934509696e4795c381121f430291c38211b17dcd37979", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5e356e00842bfb0581901405c15f533df77537adefc9883d15be01e097170831", "generation_seconds_batch": 19.738084413926117, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.859764575958252, "target_z": 0.6835936669846502, "ce": 5.609594821929932, "genuineness": 4.432034492492676, "identity": 0.279031366109848, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2bd806fd839fbabb2be934509696e4795c381121f430291c38211b17dcd37979.mp3" }, { "id": "heldout-de-2", "language": "de", "sentences": [ "Die Straße führt hinter der Brücke nach links.", "Geh weiter... bis zu dem kleinen Holztor.", "Ich warte auf der anderen Seite." ], "text": "Die Straße führt hinter der Brücke nach links. Geh weiter... bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-2", "key": "2beb48896a77b3d9ccb77d53d4c87ce97d3d5800d84750d1527b43ffe3b8bc71", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Natural clear speech.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Natural clear speech.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "76c41bb5371b142c6027d3e4451fa2b5ed4c3f3b6075f38c1ca5dcdaca4c5806", "generation_seconds_batch": 17.179478898993693, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1595139503479004, "target_z": 0.871093848744574, "ce": 5.510311126708984, "genuineness": 4.406930446624756, "identity": 0.5888994932174683, "hypothesis": "Die Straße führt hinter der Brücke nach links. Geh weiter bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2beb48896a77b3d9ccb77d53d4c87ce97d3d5800d84750d1527b43ffe3b8bc71.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "2c74a3c9b947dcceb3c50cdb877b2c974f66b3b6d1567439523b853fed780b4d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "a004132b82eb958ef32b3156c10957b196934bf8616e3e6f18cec32af5a4a5e6", "generation_seconds_batch": 23.220006208051927, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.360182762145996, "target_z": 0.3710936623241394, "ce": 6.070269584655762, "genuineness": 4.381826400756836, "identity": 0.23391148447990417, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2c74a3c9b947dcceb3c50cdb877b2c974f66b3b6d1567439523b853fed780b4d.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "2cb8bf285ae9f7e9292eaf9438db9e99f5f4353f095b9312402559a84af4f553", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "89de9478f35aecc0f6f704418b0f802cb936e99fe535febdcdc7c04b571d46bc", "generation_seconds_batch": 23.438289384008385, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6653404235839844, "target_z": 1.187500032623576, "ce": 5.228346347808838, "genuineness": 4.607763767242432, "identity": 0.7694964408874512, "hypothesis": "La porte est encore ouverte, j'ai posé ton manteau sur la chaise, dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2cb8bf285ae9f7e9292eaf9438db9e99f5f4353f095b9312402559a84af4f553.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "2cddb120c63925b3f740d87bea37790b3cadbee790155b9e739537b9c2e0fca8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ca0d8447399c7e752284b06591bcf5cbf65766dd405333313dc07f580138f035", "generation_seconds_batch": 17.412451025098562, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.428039073944092, "target_z": 1.0390625378666507, "ce": 5.673136234283447, "genuineness": 4.400654315948486, "identity": 0.32258155941963196, "hypothesis": "De deur, de deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.10526315789473684, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2cddb120c63925b3f740d87bea37790b3cadbee790155b9e739537b9c2e0fca8.mp3" }, { "id": "heldout-en-4", "language": "en", "sentences": [ "I wrote the number on the envelope.", "Read it carefully... and check the last two digits.", "Call me when you have a moment." ], "text": "I wrote the number on the envelope. Read it carefully... and check the last two digits. Call me when you have a moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-4", "key": "2ce0e743b3a2f0553a405f5156ce0cec80a2449335e5733d62e4ae9523010e1f", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I wrote the number on the envelope.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Call me when you have a moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I wrote the number on the envelope.\n(Natural clear speech.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Natural clear speech.) [3.2 seconds duration] Call me when you have a moment.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6a5240c6b944478d5e2d525d6a4fe42f98782b2e879d6d51db0e60ccf8cde679", "generation_seconds_batch": 17.281083123991266, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7411141395568848, "target_z": 0.6093750687425352, "ce": 5.577824115753174, "genuineness": 4.124508857727051, "identity": 0.3462209701538086, "hypothesis": "I wrote the number on the envelope. Read it carefully and check the last two digits. Call me when you have a moment.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2ce0e743b3a2f0553a405f5156ce0cec80a2449335e5733d62e4ae9523010e1f.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "2d2606dbff748676f1b678ff080e4ee0e2605d44bc0745e184d3cd462b214eab", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1fc5ed5398a212c90c3d278a2132a62736b5977b9fb83401843db2dd30b99fc7", "generation_seconds_batch": 15.674808133975603, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.3681004047393799, "target_z": -0.8750000279630652, "ce": 6.24500846862793, "genuineness": 4.519899368286133, "identity": 0.7123157978057861, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2d2606dbff748676f1b678ff080e4ee0e2605d44bc0745e184d3cd462b214eab.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "2d78674a45132ce52318adc13d2e3b353590edd82132d631cc94ef69603930d2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "548b50f431d31e215635dc67b5450a64bf957a8115d85c258db24e55385ef3ad", "generation_seconds_batch": 22.518019636976533, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.365591287612915, "target_z": 1.0, "ce": 5.367343425750732, "genuineness": 4.3881025314331055, "identity": 0.10426057875156403, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.120000000000001, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2d78674a45132ce52318adc13d2e3b353590edd82132d631cc94ef69603930d2.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "2d79dd5b06e9b58eac29e62c9aa45a86e4918d57d00de1f5200c14eb5fc1b878", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c0fc038be35a24182b46b69f19dc75c9c6ccad1dc0e483525476bcc6cd352116", "generation_seconds_batch": 26.244926927029155, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.615382194519043, "target_z": 1.1562500023302553, "ce": 5.367343425750732, "genuineness": 4.545003414154053, "identity": 0.4474979043006897, "hypothesis": "La puerta sigue abierta. he dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2d79dd5b06e9b58eac29e62c9aa45a86e4918d57d00de1f5200c14eb5fc1b878.mp3" }, { "id": "heldout-fr-2", "language": "fr", "sentences": [ "La route tourne à gauche après le pont.", "Continue... jusqu’à la petite barrière en bois.", "Je t’attendrai de l’autre côté." ], "text": "La route tourne à gauche après le pont. Continue... jusqu’à la petite barrière en bois. Je t’attendrai de l’autre côté.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-2", "key": "2d9a7d1cf23a26da70efc73e1e9bb6034657d67afb7ba4127ebe056ebd35fc91", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Natural clear speech.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Natural clear speech.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cbaf23590e5a93ceee8acfb15033695c2395b61c11600bc7cf9cd0cdd93b8dbe", "generation_seconds_batch": 16.1027845320059, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5904030799865723, "target_z": 1.1406249871835952, "ce": 5.224375247955322, "genuineness": 4.400654315948486, "identity": 0.7362514734268188, "hypothesis": "La route tourne à gauche après le pont, continue jusqu'à la petite arrière en bois. Je t'attendrai de l'autre côté.", "wer": 0.05, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2d9a7d1cf23a26da70efc73e1e9bb6034657d67afb7ba4127ebe056ebd35fc91.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "2dd0ff81278626180c0fd3980d98bc58bbd37cb587e84d64b372e4f97a5e6619", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "52dfcc7c0b26f8a7f518fdca7531be22739bafff87b8557d0ed31890bcee83c1", "generation_seconds_batch": 23.298566226032563, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0643925666809082, "target_z": -0.43945314087486514, "ce": 6.054384231567383, "genuineness": 4.582659721374512, "identity": 0.45149731636047363, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2dd0ff81278626180c0fd3980d98bc58bbd37cb587e84d64b372e4f97a5e6619.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "2e5fb598eb0011d6dab6e2c37c705b04545df66d4ef1d25cd8e2a9c583145492", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7288ee34c1980ff5a4d07fc652f335949a9dc4aee3a56af6bbc03c991ba98438", "generation_seconds_batch": 22.967213697033003, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9854965209960938, "target_z": 0.13671873339693005, "ce": 5.935244083404541, "genuineness": 4.6705241203308105, "identity": 0.724150538444519, "hypothesis": "Ich hab die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0625, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2e5fb598eb0011d6dab6e2c37c705b04545df66d4ef1d25cd8e2a9c583145492.mp3" }, { "id": "heldout-de-1", "language": "de", "sentences": [ "Die Schlüssel liegen auf dem Küchentisch.", "Ich suchte in der Schublade... dann neben dem Fenster.", "Nimm sie mit, wenn du gehst." ], "text": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade... dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-1", "key": "2edd5ecb0c2ba5512a3d84308a52d1ef4bb9b5991409d345d75d549d9aa23418", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Natural clear speech.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Natural clear speech.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "003ca0e60fc0c7a64bd7585c00080758548afe1f66b26b0f7afcb685efc41562", "generation_seconds_batch": 22.76204980793409, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3406121730804443, "target_z": 0.9843749848533397, "ce": 5.7684478759765625, "genuineness": 4.457139015197754, "identity": 0.540678083896637, "hypothesis": "Die Schlüssel liegen auf dem Küchentisch. Ich suche in der Schublade daneben dem Fenster. Nimm sie mit, wenn du gehst.", "wer": 0.14285714285714285, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2edd5ecb0c2ba5512a3d84308a52d1ef4bb9b5991409d345d75d549d9aa23418.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "2ee3c50601a0ba758dd77ae5e17916ef46f8088dfe85472cbe73f8ef8dd5d8e1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "78ff1c817c38eb1590c485ed10efef4315bfcbf950e82a82c8865babbf50ba38", "generation_seconds_batch": 17.18484212597832, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0221288204193115, "target_z": 0.7851562654379423, "ce": 6.181467056274414, "genuineness": 4.68307638168335, "identity": 0.5593397617340088, "hypothesis": "I found the little box. It was beside the old window. We can look inside together,", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2ee3c50601a0ba758dd77ae5e17916ef46f8088dfe85472cbe73f8ef8dd5d8e1.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "2efb4f4d09e0504049aabc0489e9bc1836a593c5d622d84949801384172169e7", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "76adcb5992aaf041cdc1aa46dea381b1007dfc99638a6e1e8c6232b1a03d9bdf", "generation_seconds_batch": 21.478914043982513, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.21571683883667, "target_z": 0.9062500582563858, "ce": 5.593709468841553, "genuineness": 4.7081804275512695, "identity": 0.7633105516433716, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre, nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2efb4f4d09e0504049aabc0489e9bc1836a593c5d622d84949801384172169e7.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "2f1664477d1e30ba2462b3d508aaeea8d7f3bfacef64e95395f587ba63134b16", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1b82c8f4d58e2657f56d51a80b9870a29fb960a1e28f6edd67871d8ef3264d53", "generation_seconds_batch": 22.40528576693032, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7411141395568848, "target_z": 0.6093750687425352, "ce": 5.538110733032227, "genuineness": 4.469690799713135, "identity": 0.35549572110176086, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2f1664477d1e30ba2462b3d508aaeea8d7f3bfacef64e95395f587ba63134b16.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "2fdec7631e461168ec091a8d51a18b243ff32f017b42a2b5098a09c00872de5e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f6fdac2e404f3ea7fa7f0d91cc9021619eb9169b08d444884f0f5f6a46e4e9dd", "generation_seconds_batch": 22.599728844012134, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8722543716430664, "target_z": 0.6914063236943281, "ce": 5.526196479797363, "genuineness": 4.507347106933594, "identity": 0.3653256893157959, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2fdec7631e461168ec091a8d51a18b243ff32f017b42a2b5098a09c00872de5e.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "2ff25e2ada7170c2c2d9cc047ea4fc1d6ad78a85c94a7ed96490510e32c45afc", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "037994f7a3099bd87d7638ee1409b7cf5227669477d3a493c848b25518400236", "generation_seconds_batch": 22.42237393092364, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6403613090515137, "target_z": 1.1718750174769157, "ce": 5.363371849060059, "genuineness": 4.557555198669434, "identity": 0.7165323495864868, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/2ff25e2ada7170c2c2d9cc047ea4fc1d6ad78a85c94a7ed96490510e32c45afc.mp3" }, { "id": "heldout-en-3", "language": "en", "sentences": [ "Someone moved the little blue chair.", "It used to stand here... beside the tall bookshelf.", "We can put it back after lunch." ], "text": "Someone moved the little blue chair. It used to stand here... beside the tall bookshelf. We can put it back after lunch.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-3", "key": "3019e7cbd84954abaf972ba3c0114919e220306910ae470142de372a473f22ea", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Someone moved the little blue chair.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can put it back after lunch.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Someone moved the little blue chair.\n(Natural clear speech.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Natural clear speech.) [3.2 seconds duration] We can put it back after lunch.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "423f403516a54f9141634462b9a24330da4a063e54b2cec594aa140e0acbeef1", "generation_seconds_batch": 16.709362261928618, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.89098858833313, "target_z": 0.7031250104861494, "ce": 5.927301406860352, "genuineness": 4.406930446624756, "identity": 0.44758403301239014, "hypothesis": "Someone moved the little blue chair. It used to stand here beside the tall bookshelf. We can put it back after lunch.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3019e7cbd84954abaf972ba3c0114919e220306910ae470142de372a473f22ea.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "301e83ecc403d8f8bbb50a2652ff880590bd5a0eef048fc289e2bc6801e13bb1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "09e67d25fa0f4f326e9afd137ca0ca0fbeef00d34f1cdc53f7928f4af8719c85", "generation_seconds_batch": 23.018379941000603, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2798372507095337, "target_z": -0.30468749708718074, "ce": 6.046441555023193, "genuineness": 4.432034492492676, "identity": 0.6310700178146362, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/301e83ecc403d8f8bbb50a2652ff880590bd5a0eef048fc289e2bc6801e13bb1.mp3" }, { "id": "heldout-es-0", "language": "es", "sentences": [ "Tu mensaje llegó después de que me fuera.", "Lo leí dos veces... y guardé el teléfono.", "Dime cuándo podemos hablar." ], "text": "Tu mensaje llegó después de que me fuera. Lo leí dos veces... y guardé el teléfono. Dime cuándo podemos hablar.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-0", "key": "30e8839ced04d9f97105e22ae50308de92e79cf94fb00a7a801c7aea5af25bab", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo podemos hablar.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Natural clear speech.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo podemos hablar.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9502c99ff0dce844fe483fbf061ddc5c3e504780244115531dfb89fb77a772f9", "generation_seconds_batch": 17.014550637919456, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2656750679016113, "target_z": 0.9375000885497065, "ce": 5.319687366485596, "genuineness": 4.3127899169921875, "identity": 0.43631893396377563, "hypothesis": "Tu mensaje llegó después de que me fuera, lo leí dos veces y guardé el teléfono, dime cuándo podemos hablar.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/30e8839ced04d9f97105e22ae50308de92e79cf94fb00a7a801c7aea5af25bab.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "313ac4c9b089fe03ee0b762cf4646dd62f42a65a85bba8093ac9b72fa877b271", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "69ebe3e0d4ae3aeb17c72bb713c1fd8a1468c7f0f7f4fca860dd506c46154716", "generation_seconds_batch": 20.215624760952778, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.4992406368255615, "target_z": -0.7929687730112724, "ce": 5.990842819213867, "genuineness": 4.570107460021973, "identity": 0.6514740586280823, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/313ac4c9b089fe03ee0b762cf4646dd62f42a65a85bba8093ac9b72fa877b271.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "3168528777be728bb3f51d374aa49cdd94faa5c23799445c5e3960aeeb5fd0f7", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "05f98eb4229b91b787d0a2bb33a28ff7ea81b9236cb42253d1df2133b7ca5424", "generation_seconds_batch": 17.342588623054326, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0916576385498047, "target_z": 0.20312497320206252, "ce": 6.165581703186035, "genuineness": 4.632867813110352, "identity": 0.8116095066070557, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3168528777be728bb3f51d374aa49cdd94faa5c23799445c5e3960aeeb5fd0f7.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "32609fed369d6d4f69a5b221fe845460a873241b324261513ead07937bdb1483", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "ceae7666fc9b8b58f9436b71db26fae7a4c5f72d14ace401a7dcea81c7156f7b", "generation_seconds_batch": 23.730676439939998, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1595139503479004, "target_z": 0.871093848744574, "ce": 6.109982967376709, "genuineness": 4.519899368286133, "identity": 0.4911937117576599, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/32609fed369d6d4f69a5b221fe845460a873241b324261513ead07937bdb1483.mp3" }, { "id": "heldout-fr-4", "language": "fr", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-4", "key": "32681eff477c55dcb46ab89f62cf34de20501adefe2fbc8de758e3fc343d1081", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Natural clear speech.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Natural clear speech.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8b182949521b134f6ac0e888b1b5a1acdf442e2d530ca8cce29c023d825cfb57", "generation_seconds_batch": 23.00563480402343, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.615382194519043, "target_z": 1.1562500023302553, "ce": 5.2065043449401855, "genuineness": 4.306513786315918, "identity": 0.6953064203262329, "hypothesis": "J'ai écrit le numéro sur l'enveloppe, lis-le bien et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/32681eff477c55dcb46ab89f62cf34de20501adefe2fbc8de758e3fc343d1081.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "3291e8f97a7f8f6a0a40efd0ff014be1ec230e5c277a4b6c961ca35c1924dfb1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "fd4fafb3c981aad21228ee1fc8f1eba39f1acc766b97e190bdffdf533e225527", "generation_seconds_batch": 23.751068977988325, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5670968294143677, "target_z": -0.1249999720369348, "ce": 6.165581703186035, "genuineness": 4.7081804275512695, "identity": 0.6082707643508911, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3291e8f97a7f8f6a0a40efd0ff014be1ec230e5c277a4b6c961ca35c1924dfb1.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "32bdd3e870f435f8afa8f070ffa3b8c78978d0e8ceefbf68ed7fb00e9b0aec28", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cee688e17653d9a7732e5055323ae32e79a26acf560460e8a865b7610355af90", "generation_seconds_batch": 21.623700667987578, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0908212661743164, "target_z": 0.8281249825230843, "ce": 5.665193557739258, "genuineness": 4.7332844734191895, "identity": 0.5532729625701904, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/32bdd3e870f435f8afa8f070ffa3b8c78978d0e8ceefbf68ed7fb00e9b0aec28.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "32c0235c7a6601a1cf7678bffc50c9b767f726e82d14fea041d3c2f13fc10005", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8b642f4f60c829532fb19c2a92f9d14ce2d1d0fc7f59f93ea14488ec1cd3b1a8", "generation_seconds_batch": 21.745318904984742, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5912394523620605, "target_z": 0.5156249778625734, "ce": 5.927301406860352, "genuineness": 4.432034492492676, "identity": 0.4578409790992737, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/32c0235c7a6601a1cf7678bffc50c9b767f726e82d14fea041d3c2f13fc10005.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "32d33190cb26c4d65d3f1c018377a679b9725031b731c26f86ab1f1ad0432341", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "aaadfc44893cd719de131d2014ed92163591a0d93662f84ce0f5a03f3a5ebb88", "generation_seconds_batch": 21.715591547079384, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.504649043083191, "target_z": -0.1640625099035856, "ce": 6.046441555023193, "genuineness": 4.557555198669434, "identity": 0.6705591678619385, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/32d33190cb26c4d65d3f1c018377a679b9725031b731c26f86ab1f1ad0432341.mp3" }, { "id": "heldout-nl-3", "language": "nl", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-3", "key": "331c21acf0ec9d45c07da9d59cb6893cb93cd1b889ef3b039359879663f441af", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Natural clear speech.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Natural clear speech.) [3.2 seconds duration] We zetten het na de lunch terug.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "dd2cc357157c1286311fc5eba62edbb95eeb1ff3dbddefd93523d0e92bf0d3a4", "generation_seconds_batch": 16.979283216991462, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.928457260131836, "target_z": 0.72656253320614, "ce": 6.181467056274414, "genuineness": 4.532451152801514, "identity": 0.484342485666275, "hypothesis": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier, naast de hoge boekenkast. We zetten het na de lunch terug.", "wer": 0.0, "duration_error_s": 0.120000000000001, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/331c21acf0ec9d45c07da9d59cb6893cb93cd1b889ef3b039359879663f441af.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "33b99b9dcb23f4576f6aaa946bc26c49a2fb9846af543bbefd2890788351f5b7", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5c811aca33e0a64d0ebcfa28abd62cb9fc168318f5c10e1c46365f02efef06a9", "generation_seconds_batch": 21.356206362019293, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4453237056732178, "target_z": -0.2011718835928169, "ce": 6.078212261199951, "genuineness": 4.570107460021973, "identity": 0.5137728452682495, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/33b99b9dcb23f4576f6aaa946bc26c49a2fb9846af543bbefd2890788351f5b7.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "33bf7d4f829f69a8a395a4a8a68db8e1b4eea21c5951f0e81957138a804fc7ba", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a85837da8ef01c54eb205c26c6d718cf910c997067b6e8e5ac7b1033228a272e", "generation_seconds_batch": 24.78635242092423, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6108101606369019, "target_z": -0.09765627009845311, "ce": 5.974957466125488, "genuineness": 4.444586753845215, "identity": 0.6058513522148132, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/33bf7d4f829f69a8a395a4a8a68db8e1b4eea21c5951f0e81957138a804fc7ba.mp3" }, { "id": "heldout-nl-2", "language": "nl", "sentences": [ "De weg gaat linksaf na de brug.", "Loop verder... tot aan het kleine houten hek.", "Ik wacht aan de andere kant." ], "text": "De weg gaat linksaf na de brug. Loop verder... tot aan het kleine houten hek. Ik wacht aan de andere kant.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-2", "key": "33f63538ac2c49676cc53462aa2024b47e58cb29b8690774a6580785ecb0e60e", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik wacht aan de andere kant.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik wacht aan de andere kant.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f080972beb316b3c2d9aac287671f541e1589c845ac585aa934f44ea94a5e567", "generation_seconds_batch": 16.910260395961814, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.055861830711364746, "target_z": -1.0703124935917976, "ce": 6.133810997009277, "genuineness": 4.356721878051758, "identity": 0.73698890209198, "hypothesis": "De weg gaat linksaf naar de brug. Loopt verder tot aan het kleine houten hek. Ik wacht aan de andere kant.", "wer": 0.09523809523809523, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/33f63538ac2c49676cc53462aa2024b47e58cb29b8690774a6580785ecb0e60e.mp3" }, { "id": "heldout-en-2", "language": "en", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-2", "key": "347b084eb1c87968a3fc14b855043a2de140e81cf0874b7d65525285bfb31414", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The road turns left behind the bridge.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The road turns left behind the bridge.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I will meet you on the other side.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "31707915f2f248f8db6d4f0961989ebf21aceb4dc4b1a32f79de7bdfb6fd8ffd", "generation_seconds_batch": 17.8032191459788, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2205119132995605, "target_z": -0.341796870776412, "ce": 6.483288764953613, "genuineness": 4.13078498840332, "identity": 0.2921331524848938, "hypothesis": "The road turns left behind the bridge. Keep going until you reach the small wooden gate. I will meet you on the other side.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/347b084eb1c87968a3fc14b855043a2de140e81cf0874b7d65525285bfb31414.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "34a55f56efdeea2e3670402ecaee8bb39442a861eeb7e1da5bc30ff6c97634c8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d39dc39e257bc064bb93ffbe799b2341d74d818323a2486612666944b31ad4fa", "generation_seconds_batch": 22.765458481037058, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.997149705886841, "target_z": 0.7695312502912819, "ce": 6.070269584655762, "genuineness": 4.7834930419921875, "identity": 0.7163709402084351, "hypothesis": "La porte est encore ouverte, j'ai posé ton manteau sur la chaise, dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/34a55f56efdeea2e3670402ecaee8bb39442a861eeb7e1da5bc30ff6c97634c8.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "34c914b9233af5fc530e2f6f5d091de1cc71304970ea0f4543e442c7e3274f2e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "13e9b246c478cd19372632d0862407fbf29f2039d3dbde0a8fb5c3ae2fcf617f", "generation_seconds_batch": 23.035718755912967, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8526835441589355, "target_z": 1.3046874970871807, "ce": 5.959072113037109, "genuineness": 4.871357440948486, "identity": 0.5788796544075012, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi, quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/34c914b9233af5fc530e2f6f5d091de1cc71304970ea0f4543e442c7e3274f2e.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "34e03913ed24934831687a24dd0159d18b19aad782fd173be7e22a49de0b735d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "55958137ba9fcdb1e040f97a9e0de8b19a9834ee9abc8eaec114b688c5788a21", "generation_seconds_batch": 23.336350324912928, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.310224771499634, "target_z": 0.3398437811671664, "ce": 6.117925643920898, "genuineness": 4.532451152801514, "identity": 0.7850397229194641, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/34e03913ed24934831687a24dd0159d18b19aad782fd173be7e22a49de0b735d.mp3" }, { "id": "heldout-fr-3", "language": "fr", "sentences": [ "Quelqu’un a déplacé la petite chaise bleue.", "Elle était ici... près de la grande bibliothèque.", "Nous la remettrons après le déjeuner." ], "text": "Quelqu’un a déplacé la petite chaise bleue. Elle était ici... près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-3", "key": "34e9fc05467b559c7662e00a2a00e38acaeeebc2fab88bd400cc0dced6f9bcce", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Natural clear speech.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Natural clear speech.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "744ca603921e2868541ea28002a65bc44d7f2eac98e5692b3dbfd90d8a894412", "generation_seconds_batch": 16.007171291974373, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.140779495239258, "target_z": 0.8593750128164048, "ce": 5.411027908325195, "genuineness": 4.595211505889893, "identity": 0.5916873812675476, "hypothesis": "Quelqu'un a déplacé la petite chasse bleue. Elle était ici, près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "wer": 0.047619047619047616, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/34e9fc05467b559c7662e00a2a00e38acaeeebc2fab88bd400cc0dced6f9bcce.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "355d802089a5c6fcae167995317344ede017b5f50ed20bd2b4b9bd119c384a28", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2650d1cacb9bcf107ba434ea868baff1789ff4517d3c623219e8e7e5e451b79b", "generation_seconds_batch": 21.785142901004292, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.652850866317749, "target_z": 1.1796875250502459, "ce": 5.49045467376709, "genuineness": 4.607763767242432, "identity": 0.8096796274185181, "hypothesis": "Je j'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/355d802089a5c6fcae167995317344ede017b5f50ed20bd2b4b9bd119c384a28.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "3586d89a4e23d4b093fcfc1e2ed0de384b743783bcbdb9a111acc6fe8d0441e1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "f84c1fe8ade2edf2d1c6bb61dd110fb469546c731ac35b437b904d42beed7d4c", "generation_seconds_batch": 16.688697982928716, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3468570709228516, "target_z": 0.9882813132081786, "ce": 5.502368927001953, "genuineness": 4.494794845581055, "identity": 0.4021896421909332, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3586d89a4e23d4b093fcfc1e2ed0de384b743783bcbdb9a111acc6fe8d0441e1.mp3" }, { "id": "heldout-en-2", "language": "en", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-2", "key": "35ab7aafaf4a93fadec58c7d41f636e145cc24a1c240bcd66c1ad2d17488b1f8", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The road turns left behind the bridge.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The road turns left behind the bridge.\n(Natural clear speech.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Natural clear speech.) [3.2 seconds duration] I will meet you on the other side.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "308f45cc7a89edb197fc2603438971fbec3f27e1721145f7894d2f3acaedc5af", "generation_seconds_batch": 17.2117942039622, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.603729009628296, "target_z": 0.5234374854359035, "ce": 5.283945083618164, "genuineness": 4.293961524963379, "identity": 0.30111318826675415, "hypothesis": "The road turns left behind the bridge. Keep going until you reach the small wooden gate. I will meet you on the other side.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/35ab7aafaf4a93fadec58c7d41f636e145cc24a1c240bcd66c1ad2d17488b1f8.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "361e4b73e4641ecb3173244c8b7e9c0d1e0c05da0e01691be7c0673083f24e39", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "98f1bae3ffdc31e86c07f98ac19d70f6fc41b727e59b681bf09384c3756ee533", "generation_seconds_batch": 22.499105387018062, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8114793300628662, "target_z": -0.5976562328143662, "ce": 6.340320587158203, "genuineness": 4.632867813110352, "identity": 0.7139467000961304, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/361e4b73e4641ecb3173244c8b7e9c0d1e0c05da0e01691be7c0673083f24e39.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "36348432e0cf90230abdc96dd7b288765e79e0e8e1acb4af564bbc393a180970", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "95d72d4c6fb54b265bf4dd9fe9ceabf2df3192e8b9988e335a51dd9994f25bea", "generation_seconds_batch": 20.786258936976083, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.919926404953003, "target_z": 0.09570310592103365, "ce": 5.879645347595215, "genuineness": 4.545003414154053, "identity": 0.4844515323638916, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/36348432e0cf90230abdc96dd7b288765e79e0e8e1acb4af564bbc393a180970.mp3" }, { "id": "heldout-es-1", "language": "es", "sentences": [ "Las llaves están en la mesa de la cocina.", "Miré en el cajón... y luego junto a la ventana.", "Llévatelas cuando te vayas." ], "text": "Las llaves están en la mesa de la cocina. Miré en el cajón... y luego junto a la ventana. Llévatelas cuando te vayas.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-1", "key": "3646013d93c47586e54845603e59a115bb107761836de0b240448b5246ed85b9", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Llévatelas cuando te vayas.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Natural clear speech.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Natural clear speech.) [3.2 seconds duration] Llévatelas cuando te vayas.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "ecd93e2e933d4fdb6998c7e1103461fac82137ba6c29cf3c0bc129bfa19051df", "generation_seconds_batch": 18.38716280902736, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.153269052505493, "target_z": 0.867187520389735, "ce": 5.454712867736816, "genuineness": 4.381826400756836, "identity": 0.5582433938980103, "hypothesis": "Las llaves están en la mesa de la cocina, miren el cajón y luego, junto a la ventana, llévatelas cuando te vayas.", "wer": 0.08695652173913043, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3646013d93c47586e54845603e59a115bb107761836de0b240448b5246ed85b9.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "3659cd1d2e80ccf941a826147fb4c9c0be36b41f425f52503f469101f42bb2af", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c54af3e379a19612d1797f838420154cce45849a1aa6be184408f9b72d3ddf94", "generation_seconds_batch": 22.650421111960895, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7785825729370117, "target_z": 0.6328124423261781, "ce": 5.649308204650879, "genuineness": 4.268857479095459, "identity": 0.5655447244644165, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3659cd1d2e80ccf941a826147fb4c9c0be36b41f425f52503f469101f42bb2af.mp3" }, { "id": "heldout-es-3", "language": "es", "sentences": [ "Alguien ha movido la pequeña silla azul.", "Estaba aquí... junto a la estantería alta.", "La colocaremos después de comer." ], "text": "Alguien ha movido la pequeña silla azul. Estaba aquí... junto a la estantería alta. La colocaremos después de comer.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-3", "key": "36ba5a78d96502e1944a79863327461ca031b2feb0e472be247796e6729df018", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] La colocaremos después de comer.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Natural clear speech.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Natural clear speech.) [3.2 seconds duration] La colocaremos después de comer.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "dadec9cf0f27b4e4832c93419a293c3f909fa9589ff0a8e1284cf5c99e0e5346", "generation_seconds_batch": 20.93221239792183, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.221961498260498, "target_z": 0.9101562374748771, "ce": 5.33954381942749, "genuineness": 4.344170093536377, "identity": 0.49899253249168396, "hypothesis": "Alguien ha movido la pequeña silla azul. Estaba aquí, junto a la estantería alta. La colocaremos después de comer.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/36ba5a78d96502e1944a79863327461ca031b2feb0e472be247796e6729df018.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "36bcf4d926f1e1a993f84e37fd0efd5087a6fb0f6e82a00d877bf33cf4541727", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3592284c61e702a7c4929e69c7951357fe0b476077705dad1b87ce5fde8ffac6", "generation_seconds_batch": 16.465320190996863, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.500690221786499, "target_z": 0.4589843352400167, "ce": 6.054384231567383, "genuineness": 4.821148872375488, "identity": 0.7557622194290161, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/36bcf4d926f1e1a993f84e37fd0efd5087a6fb0f6e82a00d877bf33cf4541727.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "36ece55ddbeaea50f6d78afe693846c7608522783ab7d02838c08d8f1321e62a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2e2504a12413780d16c74a0f4c21c672c7c0b99db7ee32f8136e79e2cacab9e2", "generation_seconds_batch": 17.12950436794199, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.284409284591675, "target_z": 0.9492187753415279, "ce": 5.692992687225342, "genuineness": 4.444586753845215, "identity": 0.21483945846557617, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je je klaar bent.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/36ece55ddbeaea50f6d78afe693846c7608522783ab7d02838c08d8f1321e62a.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "36fbde6b84947e3bf437eea1a9c9fb356e575cb5b5669dcba8e7bbeabb5ca518", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4cbba7e1d1bf5aba29fbb3df15e9c8b093c9b7d7f77ce1127d582dc5662a5b91", "generation_seconds_batch": 17.863581360899843, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8801717758178711, "target_z": -0.5546875157292241, "ce": 6.157639026641846, "genuineness": 4.7458367347717285, "identity": 0.6395004987716675, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/36fbde6b84947e3bf437eea1a9c9fb356e575cb5b5669dcba8e7bbeabb5ca518.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "3732fd015fa4d34c2806f2f0b59596bfde4bff50693dc5a92c5594eb9cde9bba", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cf93b3e82f64918c7a29a521e40f88eda528281f01e545fafa4d709beafce054", "generation_seconds_batch": 22.905993864987977, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.935538411140442, "target_z": 0.10546877767178327, "ce": 6.181467056274414, "genuineness": 4.8588056564331055, "identity": 0.5359929800033569, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3732fd015fa4d34c2806f2f0b59596bfde4bff50693dc5a92c5594eb9cde9bba.mp3" }, { "id": "heldout-nl-4", "language": "nl", "sentences": [ "Ik schreef het nummer op de envelop.", "Lees het goed... en controleer de laatste twee cijfers.", "Bel me wanneer je even tijd hebt." ], "text": "Ik schreef het nummer op de envelop. Lees het goed... en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-4", "key": "3800e26cb090519e995f83080297dae935c115be59e1296b4ca6334c50fe7117", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Natural clear speech.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Natural clear speech.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7f457876855877aaa380296a20b7d5a7e33931f5622830e3a5498fd7b424ee12", "generation_seconds_batch": 17.79212440806441, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7036454677581787, "target_z": 0.5859375460225448, "ce": 5.772419452667236, "genuineness": 4.099404335021973, "identity": 0.5320090055465698, "hypothesis": "Ik schreef het het nummer op de envelop. Lees het goed en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "wer": 0.043478260869565216, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3800e26cb090519e995f83080297dae935c115be59e1296b4ca6334c50fe7117.mp3" }, { "id": "heldout-nl-0", "language": "nl", "sentences": [ "Je bericht kwam nadat ik vertrokken was.", "Ik las het twee keer... en legde mijn telefoon weg.", "Zeg maar wanneer we kunnen praten." ], "text": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer... en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-0", "key": "383d2515aa9242b96038fba1165e0ffec1194ce98246b7083d053e2b292872d4", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Natural clear speech.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "689a3bcb9059b94f2638bc9a471dca2299372e591110e67963a351d90cf4ab4c", "generation_seconds_batch": 15.658426439971663, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.716135025024414, "target_z": 0.593750053595875, "ce": 5.720792293548584, "genuineness": 4.419482707977295, "identity": 0.18181659281253815, "hypothesis": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/383d2515aa9242b96038fba1165e0ffec1194ce98246b7083d053e2b292872d4.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "386b5ea88d8e788fdcad09a6265743f41a84d4e39294f0d34539741d7a070984", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "fed4659f042257338f352560788969ae23999a22eba554ac434d8158a82540ba", "generation_seconds_batch": 22.96887306100689, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.291490316390991, "target_z": 0.3281249452389973, "ce": 5.895530700683594, "genuineness": 4.645420074462891, "identity": 0.6606721878051758, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/386b5ea88d8e788fdcad09a6265743f41a84d4e39294f0d34539741d7a070984.mp3" }, { "id": "heldout-fr-4", "language": "fr", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-4", "key": "387145f0dbd66bee81d1cefebd3bad0ff01a7ab0f2c16f7dd5fe856171ed6097", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Natural clear speech.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Natural clear speech.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bfd0b0805d77c4a165699868b62aca7a5fc7b93ccdb40163d0ec42fe3162b3b7", "generation_seconds_batch": 17.112824529991485, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.552934408187866, "target_z": 1.1171874644636046, "ce": 5.3514580726623535, "genuineness": 4.356721878051758, "identity": 0.7390479445457458, "hypothesis": "J'ai écrit le le numéro sur l'enveloppe, lis-le bien et vérifie les deux derniers chiffres. Appelle-moi quand tu aurais un moment.", "wer": 0.09090909090909091, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/387145f0dbd66bee81d1cefebd3bad0ff01a7ab0f2c16f7dd5fe856171ed6097.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "38b1cf353068b3c938ff065b803a17a0aed796b92df3be3bacc8e233b1571ad2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9b15c598db7ad4d0d53d62901bcb382ae31f891ec986fc4675255ca0ab6f6d12", "generation_seconds_batch": 22.85506317205727, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4904866218566895, "target_z": 1.078124926596954, "ce": 5.4944257736206055, "genuineness": 4.18099308013916, "identity": 0.21382898092269897, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/38b1cf353068b3c938ff065b803a17a0aed796b92df3be3bacc8e233b1571ad2.mp3" }, { "id": "heldout-nl-4", "language": "nl", "sentences": [ "Ik schreef het nummer op de envelop.", "Lees het goed... en controleer de laatste twee cijfers.", "Bel me wanneer je even tijd hebt." ], "text": "Ik schreef het nummer op de envelop. Lees het goed... en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-4", "key": "38c4cf259ad6cc238be4540de0e9af4bedbdb6670c6606e913f6417450e69049", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Natural clear speech.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Natural clear speech.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7c1a8303d8be1c1bdc2ac8d17f350466901ea9bf8533a4f809ee3ab9ba271fea", "generation_seconds_batch": 23.107679182081483, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7286243438720703, "target_z": 0.6015624120328574, "ce": 5.661221981048584, "genuineness": 4.2374773025512695, "identity": 0.5165258646011353, "hypothesis": "Ik schreef het nummer op de envelop. Lees het goed en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/38c4cf259ad6cc238be4540de0e9af4bedbdb6670c6606e913f6417450e69049.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "38d8fb5a83d510f45c901de703ee12a8164e8af72cf8718f1b6ba672e6ff131f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "20be1f2b5083702bdbdbc609f7899335d78e635770371c46b4e0f85862a71bcb", "generation_seconds_batch": 21.435869556036778, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6278717517852783, "target_z": 1.1640625099035855, "ce": 5.538110733032227, "genuineness": 4.7458367347717285, "identity": 0.5417635440826416, "hypothesis": "He encontrado la cajita, estaba junto a la ventana vieja, podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/38d8fb5a83d510f45c901de703ee12a8164e8af72cf8718f1b6ba672e6ff131f.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "38f9da37964a60d650cb23b3520d502a5940f34c9cf79b2daf1903e4d14ec9a1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "557822fcc685430d5e8f19170e16fe2cb726206aa413f93d6d7e5c6b8bff4265", "generation_seconds_batch": 23.241455629002303, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.853519916534424, "target_z": 0.679687487766159, "ce": 5.33954381942749, "genuineness": 4.6203155517578125, "identity": 0.5671309232711792, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/38f9da37964a60d650cb23b3520d502a5940f34c9cf79b2daf1903e4d14ec9a1.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "3a84d6c578d188cc7bc19f86ebd524066b434dbe0a2cb4bdc45b145a4820b8a6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "46211e4a1d4c987a8a15fd0accd0898decfbe6521ce19d7fe061f4ef77e5feaf", "generation_seconds_batch": 16.055952494964004, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.615382194519043, "target_z": 1.1562500023302553, "ce": 5.792276382446289, "genuineness": 4.570107460021973, "identity": 0.38172370195388794, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3a84d6c578d188cc7bc19f86ebd524066b434dbe0a2cb4bdc45b145a4820b8a6.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "3aa91fc8c595698b1525a725744c9a72684fba832a2bc9282ded011b5384e1a0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e7c29bfbc37485cf2e0423afbf8ac06061639eec704fc4f5e5f3a0adf3670443", "generation_seconds_batch": 22.1538234379841, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5475261211395264, "target_z": 0.4882812759240917, "ce": 6.038498878479004, "genuineness": 4.7081804275512695, "identity": 0.6095134019851685, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3aa91fc8c595698b1525a725744c9a72684fba832a2bc9282ded011b5384e1a0.mp3" }, { "id": "heldout-nl-1", "language": "nl", "sentences": [ "De sleutels liggen op de keukentafel.", "Ik keek in de la... en daarna naast het raam.", "Neem ze mee als je weggaat." ], "text": "De sleutels liggen op de keukentafel. Ik keek in de la... en daarna naast het raam. Neem ze mee als je weggaat.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-1", "key": "3abb17f3f6dce6af448c5cc18e1ea7126ee6fce82268b82745489c226a09f131", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Neem ze mee als je weggaat.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Natural clear speech.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Natural clear speech.) [3.2 seconds duration] Neem ze mee als je weggaat.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "46ada196fbde74ba954d263e2864df3d430ceabca90bbeb345e78f92122fe23a", "generation_seconds_batch": 22.47493508597836, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4904866218566895, "target_z": 1.078124926596954, "ce": 5.411027908325195, "genuineness": 4.331617832183838, "identity": 0.16316059231758118, "hypothesis": "De sleutels liggen op de keukentafel. Ik keek in de la en daarna, naast het raam, neem ze mee als je weggaat.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3abb17f3f6dce6af448c5cc18e1ea7126ee6fce82268b82745489c226a09f131.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "3b0eba95de4d3d1e2df7c56470a84b69a78870ca6a7e4f5d211bf3a8856c2720", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] De deur staat nog open.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "21ee6cd7c8f2986121aa241977890c0eb7e5db2c149b879046e4f1b3a8bf3cec", "generation_seconds_batch": 17.59209238004405, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0487806797027588, "target_z": -0.4492187380574409, "ce": 6.229123115539551, "genuineness": 4.557555198669434, "identity": 0.49473145604133606, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3b0eba95de4d3d1e2df7c56470a84b69a78870ca6a7e4f5d211bf3a8856c2720.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "3c0e909c9c8fcc4f124c61e944fa3b48cacf6ad30e3433a4e7c8d59c998613e9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8506f2285b42ca7479655b4f825583ea3364f0a4a309b155f8d8c541c529e12b", "generation_seconds_batch": 23.16459828801453, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6028926372528076, "target_z": 1.1484374947569254, "ce": 5.6413655281066895, "genuineness": 4.796044826507568, "identity": 0.7682808637619019, "hypothesis": "La porte est encore ouverte, j'ai posé ton manteau sur la chaise, dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3c0e909c9c8fcc4f124c61e944fa3b48cacf6ad30e3433a4e7c8d59c998613e9.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "3c1d01ed51d7ed93936472c67b961724f98d724d0d60f24f7f61983dffc76977", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "66a4299d1883c0f13270cf1414ded5fc860067ccea38e26ce58b743f9c5e5a35", "generation_seconds_batch": 25.29638444702141, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0135979652404785, "target_z": 0.15429683815283599, "ce": 6.133810997009277, "genuineness": 4.532451152801514, "identity": 0.6044361591339111, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.42105263157894735, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3c1d01ed51d7ed93936472c67b961724f98d724d0d60f24f7f61983dffc76977.mp3" }, { "id": "heldout-nl-1", "language": "nl", "sentences": [ "De sleutels liggen op de keukentafel.", "Ik keek in de la... en daarna naast het raam.", "Neem ze mee als je weggaat." ], "text": "De sleutels liggen op de keukentafel. Ik keek in de la... en daarna naast het raam. Neem ze mee als je weggaat.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-1", "key": "3c590f9302048ad206a88d3c2440f1028cc9eea1f1662250baf365be21a6ebfa", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Neem ze mee als je weggaat.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Neem ze mee als je weggaat.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "44f0159dd54e1850d8a68ec72f91784ee290154085f5b5a280ed9747498e6a19", "generation_seconds_batch": 16.715598401031457, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.10582005977630615, "target_z": -1.0390624632984768, "ce": 6.109982967376709, "genuineness": 4.231201171875, "identity": 0.6876449584960938, "hypothesis": "De sleutels liggen op de de keukentafel. Ik keek in de la en daarna naast het raam. Neem ze mee als je weggaat.", "wer": 0.045454545454545456, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3c590f9302048ad206a88d3c2440f1028cc9eea1f1662250baf365be21a6ebfa.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "3c751f2ec20a2cb63e7914b349698cd01d97c2d4504e90678c4eece62196ce07", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "15442189debe8fbb251d9600c083ae7c7980abfff8b87fa743c017636e6e16d3", "generation_seconds_batch": 16.470474967965856, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5100574493408203, "target_z": 0.4648437532041012, "ce": 6.1020402908325195, "genuineness": 4.7332844734191895, "identity": 0.560272216796875, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3c751f2ec20a2cb63e7914b349698cd01d97c2d4504e90678c4eece62196ce07.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "3ca2468ec5e9e71743843c5750102711d3ac2e7815e6e4563be20206c770fcf1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "484341a08b37ec826a13a2b58db3e77bbb4407d7d237b57748d6697b98def204", "generation_seconds_batch": 23.550764871994033, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.379753589630127, "target_z": -0.2421875110687133, "ce": 6.1894097328186035, "genuineness": 4.6203155517578125, "identity": 0.7618435025215149, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3ca2468ec5e9e71743843c5750102711d3ac2e7815e6e4563be20206c770fcf1.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "3d0445b116c3eb154649189da014a20efc7b0c04f50e1b525df3d5baa1a5e83e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5238cba8c3be4617a17f8d51a08d785928fbd0f731baec27457abba8c15ffff3", "generation_seconds_batch": 22.60690902802162, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.827704429626465, "target_z": 1.2890624819405203, "ce": 5.281959533691406, "genuineness": 4.7081804275512695, "identity": 0.7933516502380371, "hypothesis": "La porte encore ouverte, j'ai posé ton manteau sur la chaise, dis-moi quand tu es prêt.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3d0445b116c3eb154649189da014a20efc7b0c04f50e1b525df3d5baa1a5e83e.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "3d627f8a46de1ca96fecaeb75ac5e6067e9aa205c0f6023834df5310762c38fd", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a78d0355af85b05c967c9c23d88e9c389d443f3d4370576e8ac9776aafed4c59", "generation_seconds_batch": 23.336350324912928, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.14473819732666, "target_z": 0.23632809310462877, "ce": 6.086154937744141, "genuineness": 4.720732688903809, "identity": 0.5469865798950195, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3d627f8a46de1ca96fecaeb75ac5e6067e9aa205c0f6023834df5310762c38fd.mp3" }, { "id": "heldout-es-0", "language": "es", "sentences": [ "Tu mensaje llegó después de que me fuera.", "Lo leí dos veces... y guardé el teléfono.", "Dime cuándo podemos hablar." ], "text": "Tu mensaje llegó después de que me fuera. Lo leí dos veces... y guardé el teléfono. Dime cuándo podemos hablar.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-0", "key": "3d80ce53997c73a1f1e5c8dfe0f40a385e10953bb1626d5ce43bfad81df56de6", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo podemos hablar.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Natural clear speech.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo podemos hablar.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9ba281f819c326af84e17d07f029bd4040c25c7bd99decf578aa93116fc2f171", "generation_seconds_batch": 17.014550637919456, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.403059959411621, "target_z": 1.0234375227199906, "ce": 5.355429172515869, "genuineness": 4.482243061065674, "identity": 0.6363003253936768, "hypothesis": "Tu mensaje llegó después de que me fuera. Lo leí dos veces y guardé el teléfono. Dime cuándo podemos hablar.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3d80ce53997c73a1f1e5c8dfe0f40a385e10953bb1626d5ce43bfad81df56de6.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "3d9d9a8ad887490dceeeddef456734177df0c5fdd204416cc299d71a81e7c53b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ee2d330f1af9b6e41b5875a87838a4ee8e4d4bcab39a0f502a087cc97ed27f31", "generation_seconds_batch": 20.970334646990523, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4890371561050415, "target_z": -0.17382810708616137, "ce": 6.157639026641846, "genuineness": 4.507347106933594, "identity": 0.732214093208313, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3d9d9a8ad887490dceeeddef456734177df0c5fdd204416cc299d71a81e7c53b.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "3e18eb491e7e209c7d1a2f6bcb20ec6190be0c8aeb3765fd8fef53e8b09712d1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b0183f3d5655234f4024d53449300674d408dbec6de0f040f6efa757360d3bc9", "generation_seconds_batch": 17.020904073957354, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.727787971496582, "target_z": 1.226562421353879, "ce": 5.34748649597168, "genuineness": 4.758388519287109, "identity": 0.564809262752533, "hypothesis": "La porte est encore ouverte, j'ai posé ton manteau sur la chaise, dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3e18eb491e7e209c7d1a2f6bcb20ec6190be0c8aeb3765fd8fef53e8b09712d1.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "3e33215ccd3eda664013941eb695940fbf12202ce48a9245768eadc6f9a4718e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "363b8e79fd9d8e364bb26863f70da1a3fba7a3daf41b9c89b595ed9eb6ab4b78", "generation_seconds_batch": 15.698507847031578, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1720032691955566, "target_z": 0.8789062071815564, "ce": 5.1052350997924805, "genuineness": 4.607763767242432, "identity": 0.49234485626220703, "hypothesis": "He encontrado la cajita, estaba junto a la ventana vieja, podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3e33215ccd3eda664013941eb695940fbf12202ce48a9245768eadc6f9a4718e.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "3e3a409a3560cdba70611061916dbcb485f9028e0eb53ffbdb527521cc76ea8d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0fad20edc36a1f4fd2b3a0758b5b76a9bad41f1209a429d439eb3d548326d73d", "generation_seconds_batch": 20.215624760952778, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2579805850982666, "target_z": -0.3183593480564216, "ce": 6.292664527893066, "genuineness": 4.582659721374512, "identity": 0.6083102226257324, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3e3a409a3560cdba70611061916dbcb485f9028e0eb53ffbdb527521cc76ea8d.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "3e621b8470e127315d1be08f836e52612c6192a9c32d75281e1556f97e430317", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c3e2a57e50f1dafce504fbdcea8eca190e5fc910dc5dd45695206032acf3a9d2", "generation_seconds_batch": 17.70679973694496, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.528791904449463, "target_z": 0.4765625891322703, "ce": 5.7684478759765625, "genuineness": 4.519899368286133, "identity": 0.7647238969802856, "hypothesis": "J'ai trouvé la petite boîte elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3e621b8470e127315d1be08f836e52612c6192a9c32d75281e1556f97e430317.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "3e8435c70737548375a09b9c4c313b6a6bd05b204a82bb38b4a133fa6f27bc7c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "eda52a32d0c54b9691e81f2d05a89b9d62d1e73fd52d938ec0545a01ae259b69", "generation_seconds_batch": 23.629294826998375, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1603500843048096, "target_z": 0.24609369028720454, "ce": 6.213237762451172, "genuineness": 4.582659721374512, "identity": 0.3432689309120178, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3e8435c70737548375a09b9c4c313b6a6bd05b204a82bb38b4a133fa6f27bc7c.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "3f001bec09423015d3e4c3cb4c00b76b1c2e14d58df24c2fd3296d807940563e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5daab81c238a2c18ee63422c2eef9b9b1745b8e46303295ddcbfc88ea0eb7be5", "generation_seconds_batch": 22.567682474968024, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6919922828674316, "target_z": -0.046874970871807094, "ce": 6.38003396987915, "genuineness": 4.720732688903809, "identity": 0.3521975874900818, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3f001bec09423015d3e4c3cb4c00b76b1c2e14d58df24c2fd3296d807940563e.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "3f53b2cc1992112f4b5a57d31f19d73a00997189c9114303295fd81eb68eae8f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d0cfdd0b673a5f55a335f6937ccec01e796a7bc11669bd4ebf51176033c95eb6", "generation_seconds_batch": 23.711149926064536, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.648278832435608, "target_z": -0.07421874737846264, "ce": 6.141753673553467, "genuineness": 4.632867813110352, "identity": 0.4878053069114685, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3f53b2cc1992112f4b5a57d31f19d73a00997189c9114303295fd81eb68eae8f.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "3fe4f16fe719fb2ab6d0df2e1be62d2370547d25f15825b42e9e7947ba72ae26", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "991d25d84eca5f3c2539c6b323674a70eb421fc35be0d2bff4128316a35096ee", "generation_seconds_batch": 21.13660669210367, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.440528392791748, "target_z": 1.0468748963036332, "ce": 5.299830436706543, "genuineness": 4.268857479095459, "identity": 0.15142865478992462, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/3fe4f16fe719fb2ab6d0df2e1be62d2370547d25f15825b42e9e7947ba72ae26.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "4003ba1fb07a2b75a280c6d989fa536799ce897edb14122ea10558729401cb07", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c6c058875a565bb28ef006d39984a5596e887ec9b291849d69919105a1371268", "generation_seconds_batch": 22.233439342002384, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0221288204193115, "target_z": 0.7851562654379423, "ce": 5.542081832885742, "genuineness": 4.394378662109375, "identity": 0.14546439051628113, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4003ba1fb07a2b75a280c6d989fa536799ce897edb14122ea10558729401cb07.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "401fcdf7ab079ba8639b8cc4fa4978c41abc2cdbf626a5696e5e391b8062e12e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e373fb53637391f53b2241ad8bde489a4de239acd8e10564989e18c3c50266ae", "generation_seconds_batch": 22.567682474968024, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2173895835876465, "target_z": -0.34374996038565764, "ce": 6.427689552307129, "genuineness": 4.595211505889893, "identity": 0.25158050656318665, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/401fcdf7ab079ba8639b8cc4fa4978c41abc2cdbf626a5696e5e391b8062e12e.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "40d9e3fe311d349827712959c17b44602c55b5e04266781d5ebf28e8969e475a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "05a39922cc46fb2babe1a5a0020f9e52fe8150000fb102bf23263e19e2cb875c", "generation_seconds_batch": 21.94922745099757, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5654239654541016, "target_z": 1.1249999720369348, "ce": 5.252174377441406, "genuineness": 4.444586753845215, "identity": 0.4883275032043457, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/40d9e3fe311d349827712959c17b44602c55b5e04266781d5ebf28e8969e475a.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "410f0e23f17d1d95c23272a887c1203aab92a791440c5bd8643ebb10ef35c2bb", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d9af1b23db6a85cf84e158876e5103d8bb4d6dc24bf411e8297f1cc8d72b5985", "generation_seconds_batch": 16.4623029519571, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8043981790542603, "target_z": 0.02343752271999047, "ce": 6.133810997009277, "genuineness": 4.65797233581543, "identity": 0.5780158042907715, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/410f0e23f17d1d95c23272a887c1203aab92a791440c5bd8643ebb10ef35c2bb.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "41ab25044df3027937d281fa60eb95b428f23942c22a43c3d7daf1aa11cd4817", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "76907215a21e6848fcaff40050f24320a6a2edb486bbda785058ebad6fe2de41", "generation_seconds_batch": 22.233439342002384, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.259430170059204, "target_z": 0.9335937601948675, "ce": 5.005951881408691, "genuineness": 4.18099308013916, "identity": 0.44518887996673584, "hypothesis": "De deur staat nog open. Ik heb je een jas op de stoel gelegd, zeg maar, wanneer je klaar bent.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/41ab25044df3027937d281fa60eb95b428f23942c22a43c3d7daf1aa11cd4817.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "41bdf115630cc2c8cb221a36dc8e2f0509dfffbe3018889d6832eafd4f23b4d9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "fdbd526fd850a7e0d11c4dcf08c1c2cede19f5cc229a2df4d0fd278050b32a08", "generation_seconds_batch": 17.112483047996648, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4382424354553223, "target_z": 0.4199217973733659, "ce": 5.990842819213867, "genuineness": 4.632867813110352, "identity": 0.6750349402427673, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/41bdf115630cc2c8cb221a36dc8e2f0509dfffbe3018889d6832eafd4f23b4d9.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "42394ae94c1e3642488c39880889c7cb8cfd941eac8f22bbaedce1ab8fcd81d7", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1b2a450008c10b18540429e389b554133ab998be7fee7cf77f29a9de3877a23c", "generation_seconds_batch": 15.77842448791489, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1580641269683838, "target_z": -0.3808594086430628, "ce": 6.221180438995361, "genuineness": 4.381826400756836, "identity": 0.6181568503379822, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/42394ae94c1e3642488c39880889c7cb8cfd941eac8f22bbaedce1ab8fcd81d7.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "4248308979dd42f5135a41fde0a1b4bee9162c14cacf56aedf748109a1d722a5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ba74bec9579a604f8f0f4c3333ceae13fa74198a5ab4af3e1b2efd15ec158001", "generation_seconds_batch": 23.298566226032563, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1174731254577637, "target_z": -0.4062500209722989, "ce": 6.268836498260498, "genuineness": 4.582659721374512, "identity": 0.7202017307281494, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4248308979dd42f5135a41fde0a1b4bee9162c14cacf56aedf748109a1d722a5.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "425ad0d8cf46b85f2386993c1296201bb7e08887297ce5d4609ee2d5797391b3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c6ef5be8c4f6d101df9fb07b87ce68ec5e2c233f2bac993a75920d1113dc7a38", "generation_seconds_batch": 23.89922502299305, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8177240490913391, "target_z": -0.593750016311788, "ce": 6.348263263702393, "genuineness": 4.846253395080566, "identity": 0.6509813070297241, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/425ad0d8cf46b85f2386993c1296201bb7e08887297ce5d4609ee2d5797391b3.mp3" }, { "id": "heldout-es-3", "language": "es", "sentences": [ "Alguien ha movido la pequeña silla azul.", "Estaba aquí... junto a la estantería alta.", "La colocaremos después de comer." ], "text": "Alguien ha movido la pequeña silla azul. Estaba aquí... junto a la estantería alta. La colocaremos después de comer.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-3", "key": "4284e0f7a9f4fdb1004d4a17ef82d5ee8de54ae670f5a2161ac5e25849e0cef3", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] La colocaremos después de comer.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Natural clear speech.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Natural clear speech.) [3.2 seconds duration] La colocaremos después de comer.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c2d1d9725a590c372b7b54260ba8c8b0cfac96aa87ae549506205d956b8112d2", "generation_seconds_batch": 20.93221239792183, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9534363746643066, "target_z": 0.7421875483528002, "ce": 5.609594821929932, "genuineness": 4.519899368286133, "identity": 0.5447659492492676, "hypothesis": "Alguien ha movido la pequeña silla azul. Estaba aquí, junto a la estantería alta. La colocaremos después de comer.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4284e0f7a9f4fdb1004d4a17ef82d5ee8de54ae670f5a2161ac5e25849e0cef3.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "42bed6e198ac5b043c8337ec36abfb6d0e58068337d45bf72b4f947408509616", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "926b3f3b9ca6ee72e1b06584bb37d75ee42c690f22b83b7db8eea39e2bdf938b", "generation_seconds_batch": 20.2958447720157, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1595139503479004, "target_z": 0.871093848744574, "ce": 5.696964263916016, "genuineness": 4.7458367347717285, "identity": 0.4073435962200165, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/42bed6e198ac5b043c8337ec36abfb6d0e58068337d45bf72b4f947408509616.mp3" }, { "id": "heldout-en-0", "language": "en", "sentences": [ "Your message arrived after I had left.", "I read it twice... before putting the phone away.", "Let me know when we can talk." ], "text": "Your message arrived after I had left. I read it twice... before putting the phone away. Let me know when we can talk.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-0", "key": "42bee835549ccbec7b71f92b7296684be23c22c7aef7635114730dc54bdb4744", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Your message arrived after I had left.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Let me know when we can talk.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Your message arrived after I had left.\n(Natural clear speech.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Natural clear speech.) [3.2 seconds duration] Let me know when we can talk.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3317ed83e8d40d4ac6e08b836e0a05599568731adddfc7b2990d23f189eaaa3a", "generation_seconds_batch": 16.519663829007186, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2406959533691406, "target_z": 0.9218750734030461, "ce": 5.530168056488037, "genuineness": 4.369274139404297, "identity": 0.6649121046066284, "hypothesis": "Your message arrived after I had left. I read it twice before putting the phone away. Let me know when we can talk.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/42bee835549ccbec7b71f92b7296684be23c22c7aef7635114730dc54bdb4744.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "43093a1713fc45b85ff70b626f1a38e22dae1b425987f62a1ddbac58f403a498", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d61d01cc545063f25c8a2462c084f0b023111c8b128c12f0a62a21eedaa59a84", "generation_seconds_batch": 20.299988040933385, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4163858890533447, "target_z": 0.4062500209722989, "ce": 6.0146708488464355, "genuineness": 4.758388519287109, "identity": 0.6387066841125488, "hypothesis": "Ich hab die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0625, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/43093a1713fc45b85ff70b626f1a38e22dae1b425987f62a1ddbac58f403a498.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "434235e45f286a8fc17c162176bc26223f08d3ced4d8efc7b78c29c26a862770", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f26f02cdfb4fd48c9400aa40cf045a94cf15477ee3ed7f6fcca5f57f9b4e3c97", "generation_seconds_batch": 22.461986104026437, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9105592966079712, "target_z": 0.08984376252512295, "ce": 6.0146708488464355, "genuineness": 4.400654315948486, "identity": 0.7047907114028931, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/434235e45f286a8fc17c162176bc26223f08d3ced4d8efc7b78c29c26a862770.mp3" }, { "id": "heldout-nl-2", "language": "nl", "sentences": [ "De weg gaat linksaf na de brug.", "Loop verder... tot aan het kleine houten hek.", "Ik wacht aan de andere kant." ], "text": "De weg gaat linksaf na de brug. Loop verder... tot aan het kleine houten hek. Ik wacht aan de andere kant.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-2", "key": "434f953004ee3838b31e4f505ba0cb9202b34ab37936979fff50edb64a6a996f", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik wacht aan de andere kant.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik wacht aan de andere kant.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2f8fb1b2bf6033e653f61feae5e580588b6616531b98fe6c0062d0d2403ce3ed", "generation_seconds_batch": 16.910260395961814, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.48675107955932617, "target_z": -0.8007812805846025, "ce": 6.213237762451172, "genuineness": 4.331617832183838, "identity": 0.5532844662666321, "hypothesis": "De weg gaat linksaf naar de brug. Loop verder tot aan het kleine houten hek. Ik wacht aan de andere kant.", "wer": 0.047619047619047616, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/434f953004ee3838b31e4f505ba0cb9202b34ab37936979fff50edb64a6a996f.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "4364751551feaa41e94c7ec34b1c3148f164d0c35481f77c53c19c3ec812139e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2b8957477160ffcdfd37ebb9b40f30bfbb6bd5cb234aa5b9a00ff91b1fb009d3", "generation_seconds_batch": 22.588062894064933, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.05814790725708, "target_z": -0.4433593200933564, "ce": 6.332377910614014, "genuineness": 4.532451152801514, "identity": 0.3151130676269531, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4364751551feaa41e94c7ec34b1c3148f164d0c35481f77c53c19c3ec812139e.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "436790c66a08ff31fb656ad83b8468f74f92deff6bc88da4d40caf3d12cb9609", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2b75422f1a6522f68dad3b7dd16b8e3b03f00c0421355046dec9c72ce883c725", "generation_seconds_batch": 20.67529825901147, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.897233486175537, "target_z": 0.7070313388409883, "ce": 5.621508598327637, "genuineness": 4.532451152801514, "identity": 0.19573037326335907, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/436790c66a08ff31fb656ad83b8468f74f92deff6bc88da4d40caf3d12cb9609.mp3" }, { "id": "heldout-es-3", "language": "es", "sentences": [ "Alguien ha movido la pequeña silla azul.", "Estaba aquí... junto a la estantería alta.", "La colocaremos después de comer." ], "text": "Alguien ha movido la pequeña silla azul. Estaba aquí... junto a la estantería alta. La colocaremos después de comer.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-3", "key": "43d771a5c055e2aea1186dc99f4cdef2b4c6bc8c6b2779ad35a29ca11c75a91c", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La colocaremos después de comer.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La colocaremos después de comer.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f03444f2b57febd92b9b512ccada5478d783056a759ffb47310346d41f0fa7c5", "generation_seconds_batch": 17.2577074309811, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6740943193435669, "target_z": -0.683593741552824, "ce": 6.268836498260498, "genuineness": 4.457139015197754, "identity": 0.5650266408920288, "hypothesis": "Alguien ha movido la pequeña silla azul. Estaba aquí, junto a la estantería alta. La colocaremos después de comer.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/43d771a5c055e2aea1186dc99f4cdef2b4c6bc8c6b2779ad35a29ca11c75a91c.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "4400821ee10cd10bea4fa35a7ee16832f7313e73f8b380ea80439ca804998a7b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "f64d8cda10577bad8030944f8b264a5810ac19acb95c5facfc526f289245b526", "generation_seconds_batch": 16.67421900201589, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.477997303009033, "target_z": 1.0703125681599714, "ce": 5.395142555236816, "genuineness": 4.281409740447998, "identity": 0.47928428649902344, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4400821ee10cd10bea4fa35a7ee16832f7313e73f8b380ea80439ca804998a7b.mp3" }, { "id": "heldout-en-3", "language": "en", "sentences": [ "Someone moved the little blue chair.", "It used to stand here... beside the tall bookshelf.", "We can put it back after lunch." ], "text": "Someone moved the little blue chair. It used to stand here... beside the tall bookshelf. We can put it back after lunch.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-3", "key": "4404e36b8a7689809428f8b5d411c3d2b7725529799d01f6ee9f2a7850474b5b", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Someone moved the little blue chair.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can put it back after lunch.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Someone moved the little blue chair.\n(Natural clear speech.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Natural clear speech.) [3.2 seconds duration] We can put it back after lunch.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4b9dcdacb315bfa3d1bceb99a198b92b0f1d62b96d61a44664b5895e8fe71f9d", "generation_seconds_batch": 16.58284433791414, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.040863037109375, "target_z": 0.7968749522297637, "ce": 5.561938762664795, "genuineness": 4.362998008728027, "identity": 0.15928253531455994, "hypothesis": "Someone moved a little blue chair. It used to stand here beside the tall bookshelf. We can put it back after lunch.", "wer": 0.045454545454545456, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4404e36b8a7689809428f8b5d411c3d2b7725529799d01f6ee9f2a7850474b5b.mp3" }, { "id": "heldout-en-4", "language": "en", "sentences": [ "I wrote the number on the envelope.", "Read it carefully... and check the last two digits.", "Call me when you have a moment." ], "text": "I wrote the number on the envelope. Read it carefully... and check the last two digits. Call me when you have a moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-4", "key": "4453aa7dd83aa09e6d9aeea3b5c31bc388367e863415ba5ce6bdac9991a029cc", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I wrote the number on the envelope.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Call me when you have a moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I wrote the number on the envelope.\n(Natural clear speech.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Natural clear speech.) [3.2 seconds duration] Call me when you have a moment.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "e9f34ea6a4de5ae64f2e42bbf0e19c3b92fba4c7dac15a2d62e18a6e27726807", "generation_seconds_batch": 16.477753099054098, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8098065853118896, "target_z": 0.6523437858276773, "ce": 5.19657564163208, "genuineness": 4.218649387359619, "identity": 0.3015105128288269, "hypothesis": "I wrote the number on the envelope. Read it carefully and check the last two digits. Call me when you have a moment.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4453aa7dd83aa09e6d9aeea3b5c31bc388367e863415ba5ce6bdac9991a029cc.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "44bee8563d1b71bc34d7999170113ddcd2680aed977b05ffb84b2ba63b066a0d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "172168c5eb799f4bbd43a1cbc9be80836797862e6b7a865cdcd98c15986e6ebe", "generation_seconds_batch": 15.883723756065592, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.360182762145996, "target_z": 0.3710936623241394, "ce": 5.673136234283447, "genuineness": 4.582659721374512, "identity": 0.460191011428833, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/44bee8563d1b71bc34d7999170113ddcd2680aed977b05ffb84b2ba63b066a0d.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "45813127e1d6bc3dbca56b6989a35e0a172d62f162cf57fde8935660bc06c3f2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7a24f43bd9054378618e3f58ef069924cb25470c3277089020692637cb4a4094", "generation_seconds_batch": 16.055952494964004, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.502976417541504, "target_z": 1.0859375833066318, "ce": 5.816103935241699, "genuineness": 4.432034492492676, "identity": 0.38541969656944275, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/45813127e1d6bc3dbca56b6989a35e0a172d62f162cf57fde8935660bc06c3f2.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "4595355e9b594f40cfecc640f0c748f728e064ee225fa7501483d25c02485dca", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8dfa27987d40f80da4177261d8bf1d3a05cd1c883cd6cbdb84f80d88b665b380", "generation_seconds_batch": 22.209093887009658, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8526835441589355, "target_z": 1.3046874970871807, "ce": 5.589737892150879, "genuineness": 4.6705241203308105, "identity": 0.7636678218841553, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4595355e9b594f40cfecc640f0c748f728e064ee225fa7501483d25c02485dca.mp3" }, { "id": "heldout-de-4", "language": "de", "sentences": [ "Ich schrieb die Nummer auf den Umschlag.", "Lies sie genau... und prüfe die letzten beiden Ziffern.", "Ruf an, wenn du Zeit hast." ], "text": "Ich schrieb die Nummer auf den Umschlag. Lies sie genau... und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-4", "key": "45f9ec8497c8691f9e18d794cf98f47c522bfab041b36df5accabf5c5c456412", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1f926d55432786beb6ae03e0ab8cea0956777100eee43a9421aac25f34830c63", "generation_seconds_batch": 17.372806290979497, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": -0.20641863346099854, "target_z": -1.234375003495383, "ce": 6.229123115539551, "genuineness": 4.061748504638672, "identity": 0.4830312430858612, "hypothesis": "Ich schrieb die Nummer auf den Umschlag, lies sie genau und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/45f9ec8497c8691f9e18d794cf98f47c522bfab041b36df5accabf5c5c456412.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "464098b6e829b8d53606a6aba6c97404e957601c3c155893ed153c4f1580f2bf", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2324ab25a771525b258dc550a11563821f3a5a654426d0087bbbf35205cbea5d", "generation_seconds_batch": 23.8592384129297, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9301300048828125, "target_z": -0.5234374854359035, "ce": 6.2767791748046875, "genuineness": 4.808597087860107, "identity": 0.5594840049743652, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/464098b6e829b8d53606a6aba6c97404e957601c3c155893ed153c4f1580f2bf.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "468405d9b32af239d72f0b2db2b3f21a413af790299c2bf5f3123c84cb8412ba", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bcf21a2c683f650038d19868f5f66e8fa88b2291fa3c6e758d90519eba079a2c", "generation_seconds_batch": 23.989452551002614, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.092494010925293, "target_z": -0.4218750361189592, "ce": 6.24500846862793, "genuineness": 4.532451152801514, "identity": 0.4829636216163635, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/468405d9b32af239d72f0b2db2b3f21a413af790299c2bf5f3123c84cb8412ba.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "46973905786ecb6de21c64f1553fea8387c5da6cec2fb8f960d44603298b324c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "43f437aa0edd8bcdb2fd691198dcd1d676346e8a6f9cab64430aba4c7e5120b3", "generation_seconds_batch": 22.95947557396721, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.361019253730774, "target_z": -0.25390627242870856, "ce": 6.348263263702393, "genuineness": 4.595211505889893, "identity": 0.410228967666626, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/46973905786ecb6de21c64f1553fea8387c5da6cec2fb8f960d44603298b324c.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "46b77315272255a990554585599df38efc466a7bc73559b95b7c373e6099cb45", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7dd4d316db590df90ac1647ba7ac40dd8f0b7c4a5204fb997b6c867a73fcfafe", "generation_seconds_batch": 23.391305606928654, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6661767959594727, "target_z": 0.5625000233025543, "ce": 5.665193557739258, "genuineness": 4.469690799713135, "identity": 0.45670804381370544, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/46b77315272255a990554585599df38efc466a7bc73559b95b7c373e6099cb45.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "4729e43d4f5bfeee3e1aecda979ee5103177aa01b7a19d40f201cae9b640a930", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "d4182e23a599e05da6af131bc0c8b48f4d14055405a5be7e0573a0b9062395bd", "generation_seconds_batch": 22.215124303940684, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2656750679016113, "target_z": 0.9375000885497065, "ce": 5.190618991851807, "genuineness": 4.356721878051758, "identity": 0.2652953863143921, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4729e43d4f5bfeee3e1aecda979ee5103177aa01b7a19d40f201cae9b640a930.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "4733490af46403b9941f3a843cd7c8af5ea08fd3da35d85f82b987c9896b78df", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "58c0ff5a791ef282add896715ed0dbd7388c7371ac64c1c6cc459fa2cee5d42f", "generation_seconds_batch": 16.89404582499992, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3789172172546387, "target_z": 0.38281249825230845, "ce": 5.927301406860352, "genuineness": 4.482243061065674, "identity": 0.4609752595424652, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd, zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4733490af46403b9941f3a843cd7c8af5ea08fd3da35d85f82b987c9896b78df.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "47917d63a2a7d6ce2a56beb5ed3807c4ec395d897524147e66f5fc7aa560952f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "24e99c20516fdca222c402b82b746dbacef3471b60f3e932c7f192968f914ebe", "generation_seconds_batch": 22.502366662956774, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4444873332977295, "target_z": 0.42382812572820483, "ce": 5.974957466125488, "genuineness": 4.519899368286133, "identity": 0.4240726828575134, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/47917d63a2a7d6ce2a56beb5ed3807c4ec395d897524147e66f5fc7aa560952f.mp3" }, { "id": "heldout-fr-3", "language": "fr", "sentences": [ "Quelqu’un a déplacé la petite chaise bleue.", "Elle était ici... près de la grande bibliothèque.", "Nous la remettrons après le déjeuner." ], "text": "Quelqu’un a déplacé la petite chaise bleue. Elle était ici... près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-3", "key": "47d00b404f769e82194cfb39f2be7a6bd3cc5bfaac4f38356b0fada2abdb9f4d", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4a6cf8991f483c8cf59e030164bbab36ff8698afc3593037994d520218e26c53", "generation_seconds_batch": 17.045331481960602, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1924104690551758, "target_z": -0.35937497553231795, "ce": 6.062326908111572, "genuineness": 4.507347106933594, "identity": 0.7733650207519531, "hypothesis": "Quelqu'un a déplacé la petite chaise bleue. Elle était ici, près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/47d00b404f769e82194cfb39f2be7a6bd3cc5bfaac4f38356b0fada2abdb9f4d.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "4831f36c8f11aeaed03162706a4c0ec9c244f501d19d915d37f0889ee332384c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "aca18a415b6c6d4d414eb6c82c9d1bea848d2fed76530e58c8b69f5e3240a755", "generation_seconds_batch": 21.787283440004103, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0595974922180176, "target_z": 0.8085937881579327, "ce": 5.462655067443848, "genuineness": 4.15588903427124, "identity": 0.33628636598587036, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4831f36c8f11aeaed03162706a4c0ec9c244f501d19d915d37f0889ee332384c.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "484e60b3a354b2d162bff41889564f9ce1cc5db6460ad09a8dbf715799e98586", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d27104e0997bd2e1f1595b93fb253eb625680d8ea6c867122bd1fa553c48df10", "generation_seconds_batch": 22.01190947194118, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8526835441589355, "target_z": 1.3046874970871807, "ce": 5.879645347595215, "genuineness": 4.68307638168335, "identity": 0.8750615119934082, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/484e60b3a354b2d162bff41889564f9ce1cc5db6460ad09a8dbf715799e98586.mp3" }, { "id": "heldout-nl-0", "language": "nl", "sentences": [ "Je bericht kwam nadat ik vertrokken was.", "Ik las het twee keer... en legde mijn telefoon weg.", "Zeg maar wanneer we kunnen praten." ], "text": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer... en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-0", "key": "48ef4db49dd978bc83f0e754f04af03838ff080cbb822e3d0e7ea9a512f559db", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Natural clear speech.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "067c99433d8b00d14227efbbda40d3824bf411432d6009fd438c62f99563ed5f", "generation_seconds_batch": 15.658426439971663, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.040863037109375, "target_z": 0.7968749522297637, "ce": 5.414999485015869, "genuineness": 4.18099308013916, "identity": 0.5116273164749146, "hypothesis": "Je bericht kwam nadat ik vertrokken was. Ik las twee keer en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "wer": 0.043478260869565216, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/48ef4db49dd978bc83f0e754f04af03838ff080cbb822e3d0e7ea9a512f559db.mp3" }, { "id": "heldout-es-4", "language": "es", "sentences": [ "Escribí el número en el sobre.", "Léelo bien... y comprueba las dos últimas cifras.", "Llámame cuando tengas un momento." ], "text": "Escribí el número en el sobre. Léelo bien... y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-4", "key": "491086f3094a6ae637d3efc824a6ba697ac91c3d07a88e6585057c5760bcea40", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Escribí el número en el sobre.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Llámame cuando tengas un momento.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Escribí el número en el sobre.\n(Natural clear speech.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Natural clear speech.) [3.2 seconds duration] Llámame cuando tengas un momento.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "26d81fb38c73b9dc86fdbdea056f6918290353890771173ecea353d0b90e1481", "generation_seconds_batch": 20.00326439004857, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.221961498260498, "target_z": 0.9101562374748771, "ce": 5.144948482513428, "genuineness": 4.570107460021973, "identity": 0.4847608208656311, "hypothesis": "Escribí el número en el sobre, léelo bien y comprueba las dos últimas cifras. Llámame cuando obtengas un momento.", "wer": 0.05263157894736842, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/491086f3094a6ae637d3efc824a6ba697ac91c3d07a88e6585057c5760bcea40.mp3" }, { "id": "heldout-de-1", "language": "de", "sentences": [ "Die Schlüssel liegen auf dem Küchentisch.", "Ich suchte in der Schublade... dann neben dem Fenster.", "Nimm sie mit, wenn du gehst." ], "text": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade... dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-1", "key": "491182784d1f1cb272e68bf27288d1e9b11ce3cf1f6c4fbc898645afc70cb560", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Natural clear speech.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Natural clear speech.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b23d1e6b9d7a54f58feb105b00b7b508f34504ba0f4fd13dfd6875cf49888551", "generation_seconds_batch": 17.204288207925856, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.431997776031494, "target_z": 0.41601561815487464, "ce": 5.514282703399658, "genuineness": 4.545003414154053, "identity": 0.44935643672943115, "hypothesis": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade, dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/491182784d1f1cb272e68bf27288d1e9b11ce3cf1f6c4fbc898645afc70cb560.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "4935f64f5873a1a4bf4e97c74bb6c6c85aab79ac0eec6b531c225ba16007c3a9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "be39b732a816374059628f49b29728becd5426a87c46cf11d94ac28cc0f13d8d", "generation_seconds_batch": 23.05777097097598, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8918249607086182, "target_z": 0.07812500116512772, "ce": 5.967014789581299, "genuineness": 4.846253395080566, "identity": 0.5310574173927307, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4935f64f5873a1a4bf4e97c74bb6c6c85aab79ac0eec6b531c225ba16007c3a9.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "49c4b9b62bdff5b4fd8fd800968bda0e7e04b69c001e4d93eff0f63b25a5382e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "326bf473fb1c14c5b54f362127ab12626f20a27131f4ae303681be8f53829656", "generation_seconds_batch": 23.3603953019483, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5163021087646484, "target_z": 0.46874993242259244, "ce": 5.839932441711426, "genuineness": 4.432034492492676, "identity": 0.4484406113624573, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/49c4b9b62bdff5b4fd8fd800968bda0e7e04b69c001e4d93eff0f63b25a5382e.mp3" }, { "id": "heldout-es-0", "language": "es", "sentences": [ "Tu mensaje llegó después de que me fuera.", "Lo leí dos veces... y guardé el teléfono.", "Dime cuándo podemos hablar." ], "text": "Tu mensaje llegó después de que me fuera. Lo leí dos veces... y guardé el teléfono. Dime cuándo podemos hablar.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-0", "key": "49d468ae6540cc7c1e5b0e6883c25e2b29599458c9de0a5398000eaca62677ab", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo podemos hablar.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo podemos hablar.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "26c755e4f631db25f8d7fa1591fd9c2d4da1ebc0e98984f1089781fab577ea62", "generation_seconds_batch": 17.162874671979807, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9675986170768738, "target_z": -0.5, "ce": 6.197352409362793, "genuineness": 4.344170093536377, "identity": 0.6037140488624573, "hypothesis": "Tu mensaje llegó después de que me fuera. Lo leí dos veces y guardé el teléfono. Dime cuándo podemos hablar.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/49d468ae6540cc7c1e5b0e6883c25e2b29599458c9de0a5398000eaca62677ab.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "49dda038c4ca5d4f25eb4f08b88659f0400c0026f16fdba58168fe25e040995c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ccdc1751ddee6d3002aea149667b2c08e0cc6bc67883ac58e3c8da661a6379b4", "generation_seconds_batch": 22.29458044108469, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3383262157440186, "target_z": 0.35742188592307234, "ce": 5.657250881195068, "genuineness": 4.720732688903809, "identity": 0.6062345504760742, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/49dda038c4ca5d4f25eb4f08b88659f0400c0026f16fdba58168fe25e040995c.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "4a04d86e602b41b4800c976f75b529a2e053426e02d0e279c711aa20b8f20df3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "956b5b1d163e8fd2614f5dc67ba9772e2ffb6b1d5faa5c821ccaef69c5e14294", "generation_seconds_batch": 22.03403298801277, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1103920936584473, "target_z": 0.2148438091302316, "ce": 5.927301406860352, "genuineness": 4.632867813110352, "identity": 0.7250251173973083, "hypothesis": "La porte est encore ouverte j'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4a04d86e602b41b4800c976f75b529a2e053426e02d0e279c711aa20b8f20df3.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "4a57935150eaab501cb3b50b65fc9fe50cb29c8a8eba6b127c19c2bfe100fcae", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "39044ab2394f72bc7617cef5c565155ec473ece0a3aa450e608db01a245fc92b", "generation_seconds_batch": 22.796636154991575, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.3681004047393799, "target_z": -0.8750000279630652, "ce": 6.38797664642334, "genuineness": 4.469690799713135, "identity": 0.4593758285045624, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4a57935150eaab501cb3b50b65fc9fe50cb29c8a8eba6b127c19c2bfe100fcae.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "4a6e9b3a39ee0ebe30847874b2178db6ee3c3f2fcc9448a887cf7bc07de30701", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bec25b09e9c2ef2746802b36df55f6b716e4cfab8f66f1559ca9c24d85a87ae1", "generation_seconds_batch": 15.698507847031578, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.690319538116455, "target_z": 1.2031250477702364, "ce": 5.236289024353027, "genuineness": 4.519899368286133, "identity": 0.46407175064086914, "hypothesis": "He encontrado la cajita. estaba junto a la ventana vieja. podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4a6e9b3a39ee0ebe30847874b2178db6ee3c3f2fcc9448a887cf7bc07de30701.mp3" }, { "id": "heldout-es-3", "language": "es", "sentences": [ "Alguien ha movido la pequeña silla azul.", "Estaba aquí... junto a la estantería alta.", "La colocaremos después de comer." ], "text": "Alguien ha movido la pequeña silla azul. Estaba aquí... junto a la estantería alta. La colocaremos después de comer.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-3", "key": "4ab76cd2e9a59da01b347403936304c2ef4e8888f38d3cb8e5371fbcf0f2e89f", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] La colocaremos después de comer.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Natural clear speech.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Natural clear speech.) [3.2 seconds duration] La colocaremos después de comer.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b9e47e68b84c71481eec9f27ad866a9bc617634ab28b9a858de73460f594cd93", "generation_seconds_batch": 18.051658822922036, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.722379684448242, "target_z": 0.5976562328143662, "ce": 5.133034706115723, "genuineness": 4.507347106933594, "identity": 0.40466588735580444, "hypothesis": "Alguien ha movido la pequeña silla azul. Estaba aquí, junto a la estantería alta. La colocaremos después de comer.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4ab76cd2e9a59da01b347403936304c2ef4e8888f38d3cb8e5371fbcf0f2e89f.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "4ac13b4222ca6e93d4f9232779c30d393220f8857f63b64cab5e0d09b47284f5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "eaafc853344904c4167fa667dbbd306db0494cd355250f393301a5ccc43b9f52", "generation_seconds_batch": 21.132345052901655, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7053182125091553, "target_z": -0.6640624726194987, "ce": 6.308549880981445, "genuineness": 4.532451152801514, "identity": 0.5753190517425537, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4ac13b4222ca6e93d4f9232779c30d393220f8857f63b64cab5e0d09b47284f5.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "4b1306919fe6bbad9bccce8613a3db7bae5a23ebea7d48d8537c4cf1ad5af06d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cd683452f92fc10050a6c0da48e91898db79b58e674fb4df6d3ae18e07ca893a", "generation_seconds_batch": 21.28497436293401, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8918249607086182, "target_z": 0.07812500116512772, "ce": 6.006728172302246, "genuineness": 4.645420074462891, "identity": 0.7748305797576904, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4b1306919fe6bbad9bccce8613a3db7bae5a23ebea7d48d8537c4cf1ad5af06d.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "4b3295e6047c5ac2662b12cf577710aad0bf047657fc7b693ad0df0c3535d51b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4c1e87d1d84019754d7461b6bcd0f6ba50f8e860399fdf893c1a36032b5c3b3c", "generation_seconds_batch": 21.246781819034368, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0198428630828857, "target_z": 0.15820316650767488, "ce": 6.0305562019348145, "genuineness": 4.65797233581543, "identity": 0.7401496171951294, "hypothesis": "Ich hab die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0625, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4b3295e6047c5ac2662b12cf577710aad0bf047657fc7b693ad0df0c3535d51b.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "4b5f4615d2b8d16544efc1124d0c6ea2c60a2fc07f11b74de3aba61c13f9da85", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3dbffc8f5acff1290f1e60703c6b4d023bf64c8fe02a44a53f11b16e2052d9b8", "generation_seconds_batch": 21.94922745099757, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.477997303009033, "target_z": 1.0703125681599714, "ce": 5.200547218322754, "genuineness": 4.457139015197754, "identity": 0.5412744283676147, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4b5f4615d2b8d16544efc1124d0c6ea2c60a2fc07f11b74de3aba61c13f9da85.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "4ba07669b5e990a743a1a49629bcbc464dcd98ad37d374f9cfedace7f7de2e31", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "38bd6208183c480def680c42715c3732cebb76786d035231d9e32ace902bc5e6", "generation_seconds_batch": 22.55756815499626, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.228206157684326, "target_z": 0.9140624166933683, "ce": 4.954324245452881, "genuineness": 4.406930446624756, "identity": 0.4283885955810547, "hypothesis": "He encontrado la cajita. estaba junto a la ventana vieja. podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4ba07669b5e990a743a1a49629bcbc464dcd98ad37d374f9cfedace7f7de2e31.mp3" }, { "id": "heldout-es-0", "language": "es", "sentences": [ "Tu mensaje llegó después de que me fuera.", "Lo leí dos veces... y guardé el teléfono.", "Dime cuándo podemos hablar." ], "text": "Tu mensaje llegó después de que me fuera. Lo leí dos veces... y guardé el teléfono. Dime cuándo podemos hablar.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-0", "key": "4c02cf707641fe1a41a015a69670634e2da957b98dc75325dc8a058bb4801a23", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo podemos hablar.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Natural clear speech.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo podemos hablar.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3fcfeb46f475f47f4f377dfd0f44942785ebd229aa344b4056263b9ae53bae65", "generation_seconds_batch": 18.141905845026486, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2406959533691406, "target_z": 0.9218750734030461, "ce": 4.876883506774902, "genuineness": 4.262581825256348, "identity": 0.47844141721725464, "hypothesis": "Mi mensaje llegó después de que me fuera, lo leí dos veces y guardé el teléfono, dime cuándo podemos hablar.", "wer": 0.05, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4c02cf707641fe1a41a015a69670634e2da957b98dc75325dc8a058bb4801a23.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "4c047b50e4e9205bf29e1d5ff39adfb88f4975d3d44821113732060867642254", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "aa30f7d6219c306fcac04c1b0c629f9cc449fe13fb2474ddade57715f46d8318", "generation_seconds_batch": 21.615672083105892, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8043981790542603, "target_z": 0.02343752271999047, "ce": 6.181467056274414, "genuineness": 4.482243061065674, "identity": 0.7514363527297974, "hypothesis": "Ik heb het kleine doosje gevonden, het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4c047b50e4e9205bf29e1d5ff39adfb88f4975d3d44821113732060867642254.mp3" }, { "id": "heldout-de-2", "language": "de", "sentences": [ "Die Straße führt hinter der Brücke nach links.", "Geh weiter... bis zu dem kleinen Holztor.", "Ich warte auf der anderen Seite." ], "text": "Die Straße führt hinter der Brücke nach links. Geh weiter... bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-2", "key": "4c0ffe5523c0b2faf2c55bc71d02eda9ee5ff0a840df4ca32b60d429270b724f", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Natural clear speech.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Natural clear speech.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ce6d882286bfdcf481d47c9d38524338cb112fb54a3671843a06d888f79d5bee", "generation_seconds_batch": 17.04693485388998, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1969823837280273, "target_z": 0.8945312223282167, "ce": 5.538110733032227, "genuineness": 4.457139015197754, "identity": 0.516383945941925, "hypothesis": "Die Straße führt hinter der Brücke nach links. Geh weiter bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4c0ffe5523c0b2faf2c55bc71d02eda9ee5ff0a840df4ca32b60d429270b724f.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "4c9d2ba6b78c6d6805ea3bb685fffd7f91d0c2b3c67e611815ad0f8b7627ec66", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8638e9b255b0f1a8b9e57edf9cc0f91bb3944800d3dac35dd3fe7c7015460ec5", "generation_seconds_batch": 21.657455754000694, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6653404235839844, "target_z": 1.187500032623576, "ce": 5.355429172515869, "genuineness": 4.7332844734191895, "identity": 0.5326529741287231, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4c9d2ba6b78c6d6805ea3bb685fffd7f91d0c2b3c67e611815ad0f8b7627ec66.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "4d024bc77c17b9204b06de81a1680f6028e1690aacaa07ccbf7ebb7887163136", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b534325f3d02aee66b2ad52e978581dad1a4577b09a10e189010b897b332f53d", "generation_seconds_batch": 23.35871112300083, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8401942253112793, "target_z": 1.2968751386501982, "ce": 6.125868320465088, "genuineness": 4.770940780639648, "identity": 0.7976149320602417, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4d024bc77c17b9204b06de81a1680f6028e1690aacaa07ccbf7ebb7887163136.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "4d27caf3946f5b161605e8e923fc8bf5cc956a751e1c0e15432587792a0137d3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "052590bcaa1d072d08e3cc3749cbc2ee67f9270fd60ac35b68130bdfb7ba29f8", "generation_seconds_batch": 16.306862293044105, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.10414719581604, "target_z": 0.21093748077539268, "ce": 6.0146708488464355, "genuineness": 4.6203155517578125, "identity": 0.7263913154602051, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4d27caf3946f5b161605e8e923fc8bf5cc956a751e1c0e15432587792a0137d3.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "4d52ebcb381bbe4e0c155e647f371ae76c925b7d6dd8362f4fdd0a8549796fe8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "520297902070779d2590e7096051bb9dd4dcafedb802cd2d44ab11d83b7b446f", "generation_seconds_batch": 21.12398576294072, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.003394603729248, "target_z": 0.7734375786461208, "ce": 5.788304805755615, "genuineness": 4.325342178344727, "identity": 0.4752412736415863, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4d52ebcb381bbe4e0c155e647f371ae76c925b7d6dd8362f4fdd0a8549796fe8.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "4d702052330aad06c0c2c85ea1ffee9b0cbc61b534c46e7769750617b29127c5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "463ede2456d3f5e7ca2a8af6472ad0139c5a4f927cd588cb11ccb85a5bba2b42", "generation_seconds_batch": 22.4401689710794, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.503812789916992, "target_z": 0.46093757398560997, "ce": 5.935244083404541, "genuineness": 4.758388519287109, "identity": 0.61493980884552, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4d702052330aad06c0c2c85ea1ffee9b0cbc61b534c46e7769750617b29127c5.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "4d9ff7d340154f35d410b6b8840239f0611db3c0d6d1321ebcf9aee85da5b6e1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "14605c282c3ba8f3b0e29694451371083a3715fea985f8e24046149b876fe7e2", "generation_seconds_batch": 24.78635242092423, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8262548446655273, "target_z": 0.03710937368923132, "ce": 6.09409761428833, "genuineness": 4.469690799713135, "identity": 0.4361923933029175, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4d9ff7d340154f35d410b6b8840239f0611db3c0d6d1321ebcf9aee85da5b6e1.mp3" }, { "id": "heldout-es-4", "language": "es", "sentences": [ "Escribí el número en el sobre.", "Léelo bien... y comprueba las dos últimas cifras.", "Llámame cuando tengas un momento." ], "text": "Escribí el número en el sobre. Léelo bien... y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-4", "key": "4db6a3d531409536bdc2fab1499ef011836ee967b5f808f9c7984f047cf14853", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Escribí el número en el sobre.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Llámame cuando tengas un momento.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Escribí el número en el sobre.\n(Natural clear speech.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Natural clear speech.) [3.2 seconds duration] Llámame cuando tengas un momento.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "574ffbf0d99c212d16d0572c16db5ada4c43f766c538dbadcb05daf361054cdf", "generation_seconds_batch": 17.98059818998445, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.259430170059204, "target_z": 0.9335937601948675, "ce": 5.403085231781006, "genuineness": 4.482243061065674, "identity": 0.5032714605331421, "hypothesis": "Escribí el número en el sobre, léelo bien y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4db6a3d531409536bdc2fab1499ef011836ee967b5f808f9c7984f047cf14853.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "4dbcb72301ecd78378cd70f7c7456127e6ea8e4890489bb3fee0b7237618788d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "c2c4028c6b78b9d3cce99d12b078303c76330f165c91723d30c6be933597d63d", "generation_seconds_batch": 20.547346844919957, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.115800380706787, "target_z": 0.8437499976697446, "ce": 6.09409761428833, "genuineness": 4.444586753845215, "identity": 0.5123787522315979, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4dbcb72301ecd78378cd70f7c7456127e6ea8e4890489bb3fee0b7237618788d.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "4dcd9c59dfc1cd1be79845487e9492fd4ad1b829cc68da1086b1342f2c7a0d09", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "5b04c9b0b979e87d544103de2da776df0e5ecfd6e3831de257c51718db0998ce", "generation_seconds_batch": 21.77072331495583, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.615382194519043, "target_z": 1.1562500023302553, "ce": 5.2918877601623535, "genuineness": 4.369274139404297, "identity": 0.4866402745246887, "hypothesis": "Hoy he encontrado la cajita, estaba junto a la ventana vieja, podemos mirar juntos lo que hay dentro.", "wer": 0.058823529411764705, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4dcd9c59dfc1cd1be79845487e9492fd4ad1b829cc68da1086b1342f2c7a0d09.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "4eb779cb5c20a8c06af741729660168f3457a852fe2b620f64668f3d47ba19f9", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0bb5243c9a50d0e085bc32d134545158eaeb49af0fae61a7a1b9f3d30f6877bf", "generation_seconds_batch": 17.804169637034647, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0456583499908447, "target_z": -0.4511718276666865, "ce": 6.213237762451172, "genuineness": 4.344170093536377, "identity": 0.6731985807418823, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4eb779cb5c20a8c06af741729660168f3457a852fe2b620f64668f3d47ba19f9.mp3" }, { "id": "heldout-fr-0", "language": "fr", "sentences": [ "Ton message est arrivé après mon départ.", "Je l’ai relu... puis j’ai rangé mon téléphone.", "Dis-moi quand nous pourrons parler." ], "text": "Ton message est arrivé après mon départ. Je l’ai relu... puis j’ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-0", "key": "4edacde119b610215e7c40c81580067d58b5f53a7509335f5906d2a285caf158", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Natural clear speech.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a435537a85f17a5d995737f8ee966400f377fcc7bb4b8c7906ef2bfef7f470dd", "generation_seconds_batch": 17.450143280089833, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.577913522720337, "target_z": 1.132812479610265, "ce": 5.276002407073975, "genuineness": 4.469690799713135, "identity": 0.7892699241638184, "hypothesis": "Ton message est arrivé après mon départ, je l'ai relu, puis j'ai rangé mon téléphone. Dis-moi que nous pourrons parler.", "wer": 0.047619047619047616, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4edacde119b610215e7c40c81580067d58b5f53a7509335f5906d2a285caf158.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "4ee4473c4422ba7e41940ce04e5ddc244d097c4f2312a40ce7a01665f500e7c2", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5f9819ac8ece0f3d035ad09bf1c579b4bf310e7771f463ae7f7aba6d6fd8db73", "generation_seconds_batch": 17.277530936989933, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.776296615600586, "target_z": 0.005859343395910696, "ce": 6.300607204437256, "genuineness": 4.821148872375488, "identity": 0.45247727632522583, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4ee4473c4422ba7e41940ce04e5ddc244d097c4f2312a40ce7a01665f500e7c2.mp3" }, { "id": "heldout-en-4", "language": "en", "sentences": [ "I wrote the number on the envelope.", "Read it carefully... and check the last two digits.", "Call me when you have a moment." ], "text": "I wrote the number on the envelope. Read it carefully... and check the last two digits. Call me when you have a moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-4", "key": "4ef7bf7a30407ab02a53929d6022a9b7457385cee326d5b38a0911738e78721a", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I wrote the number on the envelope.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Call me when you have a moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I wrote the number on the envelope.\n(Natural clear speech.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Natural clear speech.) [3.2 seconds duration] Call me when you have a moment.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5981883c8e68f90ef3d44254c4148eb0ce32730a7b3ec97ace37434ebc16ea47", "generation_seconds_batch": 23.23546012805309, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.578749895095825, "target_z": 0.5078124702892433, "ce": 5.609594821929932, "genuineness": 4.293961524963379, "identity": 0.2942846417427063, "hypothesis": "I wrote the number on the envelope. Read it carefully and check the last two digits. Call me when you have a moment.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4ef7bf7a30407ab02a53929d6022a9b7457385cee326d5b38a0911738e78721a.mp3" }, { "id": "heldout-fr-4", "language": "fr", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-4", "key": "4f4e68b5fb6e560889c4c1c32797fc3256dda30ab4547e51187e5cadb87d24b3", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Natural clear speech.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Natural clear speech.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7387c97caf920c03f677f51e3ce164f69630c274ad3c3091835674b66ca21165", "generation_seconds_batch": 16.197387669002637, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8160512447357178, "target_z": 0.6562499650461685, "ce": 4.576054573059082, "genuineness": 4.005263805389404, "identity": 0.6894909739494324, "hypothesis": "J'ai écrit le numéro sur l'enveloppe, les viens et vérifiez les deux derniers chiffres. Appelle-moi quand il y aura un moment.", "wer": 0.3181818181818182, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4f4e68b5fb6e560889c4c1c32797fc3256dda30ab4547e51187e5cadb87d24b3.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "4ffbc7807692423f0fb8fe6c960f778cf598f6f7e827450734d883199584524c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "88e4cd5eade028b3bb73e0d1063561756c37f0f2c05623c846d3259d9247926c", "generation_seconds_batch": 20.265294632059522, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8472752571105957, "target_z": 0.6757813085476677, "ce": 5.621508598327637, "genuineness": 4.7081804275512695, "identity": 0.44148480892181396, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/4ffbc7807692423f0fb8fe6c960f778cf598f6f7e827450734d883199584524c.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "509168c3fedbd2d03c000fa338952544a8fd4aea8928e5fffea54284a21b670b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "809d27485336e5d107534efa671006c19cb03a26640e3c7ae85becf9c28c46b8", "generation_seconds_batch": 23.030156910070218, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0229651927948, "target_z": 0.1601562561169205, "ce": 5.760505676269531, "genuineness": 4.65797233581543, "identity": 0.5649753212928772, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/509168c3fedbd2d03c000fa338952544a8fd4aea8928e5fffea54284a21b670b.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "5140bebb18deb9a53892d9bb18121d808f8325b7a747a4e7a4de705b7113ec66", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7a268ec7f285f7883a66b0d813ebf45737fa3d18381405b55e5b9aac126afee1", "generation_seconds_batch": 22.575762580963783, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7178077697753906, "target_z": -0.6562499650461685, "ce": 6.308549880981445, "genuineness": 4.7081804275512695, "identity": 0.6419203281402588, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5140bebb18deb9a53892d9bb18121d808f8325b7a747a4e7a4de705b7113ec66.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "51a7c473c15010a7f5fa25dadbded4237b230967d5240e88b36b963ff8079e9f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e6f1107ca8ab45010c3f1c3abab450cab0ddbeadee6cf5b8d99edcb9d30964b3", "generation_seconds_batch": 22.159058392979205, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8784990310668945, "target_z": 0.6953125029128193, "ce": 5.307773113250732, "genuineness": 4.231201171875, "identity": 0.46560028195381165, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/51a7c473c15010a7f5fa25dadbded4237b230967d5240e88b36b963ff8079e9f.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "51f74c05e459a369f5188135c8ab50fd0af064fb473d0d34a9f4f0f32f5b0085", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "15c1f7ea0dfc7facb9fc947f376dcb9a6e7137af7d0bfd149aedfbf28637acfc", "generation_seconds_batch": 22.70480802305974, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4904866218566895, "target_z": 1.078124926596954, "ce": 5.3514580726623535, "genuineness": 4.457139015197754, "identity": 0.7681901454925537, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/51f74c05e459a369f5188135c8ab50fd0af064fb473d0d34a9f4f0f32f5b0085.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "52077416a2ac2484cceeda269923db4f5a584570518ba0018be12e8323f13870", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "26ccb7cd600e6a72d603e34bfa755740d201191580f91be7f5268740a4cac77b", "generation_seconds_batch": 16.406632611062378, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8668458461761475, "target_z": 0.0624999860184674, "ce": 6.181467056274414, "genuineness": 4.7081804275512695, "identity": 0.408711314201355, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/52077416a2ac2484cceeda269923db4f5a584570518ba0018be12e8323f13870.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "5214f7a47f04b48c3966fefd5e140e7916cb75a5db2802ffa1918509450433ab", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "23fde263dedd49084a92292dee0b56d7d8d52568298633e520e48ae8e8507b22", "generation_seconds_batch": 16.23243209591601, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8847436904907227, "target_z": 0.6992186821313106, "ce": 5.855817794799805, "genuineness": 4.256305694580078, "identity": 0.5060286521911621, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd, zeg maar, wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5214f7a47f04b48c3966fefd5e140e7916cb75a5db2802ffa1918509450433ab.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "522aaaa6061bbed9cc242c04d907a44e217452bc30142991ee4356649f4fadf8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6e60668bba8ff31e0414ec4b311f62136b4ad9cec6bb0300f5fc20de88b53c5a", "generation_seconds_batch": 16.262230332009494, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4904866218566895, "target_z": 1.078124926596954, "ce": 5.363371849060059, "genuineness": 4.494794845581055, "identity": 0.6979540586471558, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/522aaaa6061bbed9cc242c04d907a44e217452bc30142991ee4356649f4fadf8.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "5243c3202c639b43788e72bd19d16ee7ecfb5978cecdde8c73767738abe672bf", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "ee7339baf236bd8aee1531d212c1b30b9b50abd4f5950f6cc502c48dc2b56dc6", "generation_seconds_batch": 22.87101107405033, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0845766067504883, "target_z": 0.8242188033045931, "ce": 5.091335296630859, "genuineness": 4.268857479095459, "identity": 0.22952096164226532, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5243c3202c639b43788e72bd19d16ee7ecfb5978cecdde8c73767738abe672bf.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "52860dd49a5e9714ffff16fba66e3e837c4295c7bd353d95fd956ce662f8d1da", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "5a50feacb6ca78dc368db08c8bfd2374f9edfef2774197e648107b051bcd082a", "generation_seconds_batch": 22.87101107405033, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1282899379730225, "target_z": 0.8515625052430748, "ce": 5.708878040313721, "genuineness": 4.469690799713135, "identity": 0.3674822151660919, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/52860dd49a5e9714ffff16fba66e3e837c4295c7bd353d95fd956ce662f8d1da.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "5339270e3e7789abb760c7d81f7d62aaa2b8665ae84df69f717548388d396b1f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cef3313498edbc7bedd2fbee698d9d33510e4643e534435342c6d15f66abcaf9", "generation_seconds_batch": 23.89922502299305, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.5929123163223267, "target_z": -0.7343749662112963, "ce": 6.308549880981445, "genuineness": 4.482243061065674, "identity": 0.5728816986083984, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5339270e3e7789abb760c7d81f7d62aaa2b8665ae84df69f717548388d396b1f.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "53402d51cc42a71c6abcc388ed7c2f77c4f097623f9a97c1363518d4bb28a59e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "235962af082ed35e59d8aeb7bdcb81fab8059607afd540e8cafee631a3a8b024", "generation_seconds_batch": 22.673365589929745, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.791072368621826, "target_z": 0.6406250990358558, "ce": 5.816103935241699, "genuineness": 4.645420074462891, "identity": 0.4909667670726776, "hypothesis": "Die Tür ist noch offen. Ich hab deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/53402d51cc42a71c6abcc388ed7c2f77c4f097623f9a97c1363518d4bb28a59e.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "53c29a39cecd5831d50994c7ef33b99d14d8dc27b2af1902835a1da8564e1729", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a0c0344951b64f7c1c5a8a505b300faf45333afdd449f8c3d2b4e5dfcac6953f", "generation_seconds_batch": 22.512820691918023, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4734251499176025, "target_z": -0.18359377883691097, "ce": 5.974957466125488, "genuineness": 4.413206577301025, "identity": 0.677505612373352, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/53c29a39cecd5831d50994c7ef33b99d14d8dc27b2af1902835a1da8564e1729.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "54498479ffda9e623ed1f4bb918f45bfe6f0cb9cf3ee5f50d9eb707c4952604a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "218546c2113c6b9277ffeb606d1a21a7f3c11aa84cadf4190a768a576c7515aa", "generation_seconds_batch": 23.018379941000603, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.3922431468963623, "target_z": -0.23437500349538315, "ce": 6.284721851348877, "genuineness": 4.632867813110352, "identity": 0.6766159534454346, "hypothesis": "Ik heb het kleine doosje gevonden, het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/54498479ffda9e623ed1f4bb918f45bfe6f0cb9cf3ee5f50d9eb707c4952604a.mp3" }, { "id": "heldout-de-4", "language": "de", "sentences": [ "Ich schrieb die Nummer auf den Umschlag.", "Lies sie genau... und prüfe die letzten beiden Ziffern.", "Ruf an, wenn du Zeit hast." ], "text": "Ich schrieb die Nummer auf den Umschlag. Lies sie genau... und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-4", "key": "546203f143d2dbd908be3097fd76d41d76e93afd1487fa15c64a3d119dff5d49", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "16e712f9a451ac48740fa708c6cecfc0e88c38ddc70eb609e2b0eb145d4a3a0f", "generation_seconds_batch": 17.372806290979497, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.030882716178894043, "target_z": -1.085937508738458, "ce": 6.324435234069824, "genuineness": 4.419482707977295, "identity": 0.619538426399231, "hypothesis": "Ich schrieb die Nummer auf den Umschlag, ließ sie genau und prüfe die letzten beiden Ziffern, ruf an, wenn du Zeit hast.", "wer": 0.045454545454545456, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/546203f143d2dbd908be3097fd76d41d76e93afd1487fa15c64a3d119dff5d49.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "5538df87ec280323f89a9edf7ebcf65240857bb10e838b6c6295b8c00a0e7a70", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3c82b931438bd553239fa5d85dd61141eb431bdea36d431b1a0d03236ba309e7", "generation_seconds_batch": 16.98454019194469, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7973170280456543, "target_z": 0.6445312782543471, "ce": 5.7128496170043945, "genuineness": 4.758388519287109, "identity": 0.7518385052680969, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre, nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5538df87ec280323f89a9edf7ebcf65240857bb10e838b6c6295b8c00a0e7a70.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "557b5e1cfbf8eed3b47bccad0911b0adefe8dd7b0f13b2fe18aa3bd2fd895d99", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "475f4da12619bdc662f7e03be377c9b249bd862dd3460d6f1793945b1a973aa2", "generation_seconds_batch": 17.412451025098562, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3531017303466797, "target_z": 0.9921874924266698, "ce": 5.422942161560059, "genuineness": 4.507347106933594, "identity": 0.31864458322525024, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/557b5e1cfbf8eed3b47bccad0911b0adefe8dd7b0f13b2fe18aa3bd2fd895d99.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "557d40661bb448392aa06c17e80bdd02dae798accb1f979359ad39abc6c22e0c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "74761e5bab70a3670efbaa7511b4e59c2cef84d7c4dc6beec45c8a4f23dbb9f4", "generation_seconds_batch": 16.306862293044105, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9230488538742065, "target_z": 0.09765627009845311, "ce": 5.681078910827637, "genuineness": 4.758388519287109, "identity": 0.7491005063056946, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/557d40661bb448392aa06c17e80bdd02dae798accb1f979359ad39abc6c22e0c.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "55b01837d8afcb7842cc35d2c592c605f43403ef02ccaff0c8fc72a7842c223e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "01eaeb02408bb9129e53763fa3e00117512976c12635a769cf3c84fcc3407e17", "generation_seconds_batch": 23.550764871994033, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6139326095581055, "target_z": -0.09570310592103365, "ce": 6.109982967376709, "genuineness": 4.507347106933594, "identity": 0.6530749797821045, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/55b01837d8afcb7842cc35d2c592c605f43403ef02ccaff0c8fc72a7842c223e.mp3" }, { "id": "heldout-de-0", "language": "de", "sentences": [ "Deine Nachricht kam, als ich schon weg war.", "Ich las sie zweimal... dann legte ich das Handy weg.", "Sag mir, wann wir reden können." ], "text": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal... dann legte ich das Handy weg. Sag mir, wann wir reden können.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-0", "key": "561fd9f7740a5b0832e135ab90419d70861c1426e3827c574917185b77b2b426", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wann wir reden können.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Natural clear speech.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wann wir reden können.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9805348f4406ac8b2cc6e338352bf220fb0f964077f14ffa373f663aea83a59b", "generation_seconds_batch": 17.274377543013543, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.866009473800659, "target_z": 0.6874999953394891, "ce": 5.546053409576416, "genuineness": 4.519899368286133, "identity": 0.6158638000488281, "hypothesis": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal, dann leichte ich das Handy weg. Sag mir, wann wir reden können.", "wer": 0.041666666666666664, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/561fd9f7740a5b0832e135ab90419d70861c1426e3827c574917185b77b2b426.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "5635a8d995f6dd5ad8cb132c270d90e5d97e8c95c5f8fc86d75f039ab6d22e1b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f74d15d9500767098466e867a94e65c13caf47721732554c64c0fd91ea293235", "generation_seconds_batch": 17.567121925996616, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8722543716430664, "target_z": 0.6914063236943281, "ce": 5.919358730316162, "genuineness": 4.444586753845215, "identity": 0.20205149054527283, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5635a8d995f6dd5ad8cb132c270d90e5d97e8c95c5f8fc86d75f039ab6d22e1b.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "563f47fae7b1fda6546617e26f155ab49f847e2267c2756be6bbd6d7d6e4947a", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] La puerta sigue abierta.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f93e4bb751e60286b125ec0ed765473546763bdbdb9c57d22de056f1e03108d7", "generation_seconds_batch": 15.840117499930784, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.3297953605651855, "target_z": -0.2734375413620339, "ce": 5.967014789581299, "genuineness": 4.042920112609863, "identity": 0.41492676734924316, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/563f47fae7b1fda6546617e26f155ab49f847e2267c2756be6bbd6d7d6e4947a.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "56de39eb3d88e41c40549eece06d3e86564c4cc90923449d9efbf28d179ebb0d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "00e2ca6d3eaf17ffaa3bebd9f2fb2014af0b65188ec4c09159d0065b08686ff8", "generation_seconds_batch": 16.577306235092692, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.361019253730774, "target_z": -0.25390627242870856, "ce": 6.213237762451172, "genuineness": 4.18099308013916, "identity": 0.32342633605003357, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/56de39eb3d88e41c40549eece06d3e86564c4cc90923449d9efbf28d179ebb0d.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "5707b9388225aefdd7075c8ce1ac98d9c49c128415ce6ae888a3ef5b7a52beac", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bc632dc114ee0c83fc25392fdc06a3f5a97c57ecd967af937d6f37e8b940f84a", "generation_seconds_batch": 21.623867515008897, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2017775774002075, "target_z": -0.35351563213640724, "ce": 6.292664527893066, "genuineness": 4.494794845581055, "identity": 0.45712119340896606, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5707b9388225aefdd7075c8ce1ac98d9c49c128415ce6ae888a3ef5b7a52beac.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "572a616b2283efd6ebeeff48acea783416fa7ecc35c883ea9fb1483e17a295a1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "143dc603f621409842b3d0bb14bee78ca37e48a7b31566cc82d85980fde341e1", "generation_seconds_batch": 15.957766006002203, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.940110206604004, "target_z": 1.359374900964144, "ce": 5.6413655281066895, "genuineness": 4.6956281661987305, "identity": 0.6683922410011292, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/572a616b2283efd6ebeeff48acea783416fa7ecc35c883ea9fb1483e17a295a1.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "578e798bbc78cfdfe7a7f6f08a324948dc71ef6286d453daf89335b4a3e76bfb", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "18e151b699d9190f0d340d717cb545e120daa9322de1a7eed3a797ef5d4bda9b", "generation_seconds_batch": 22.663575985003263, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.38828444480896, "target_z": 0.38867191621639297, "ce": 5.9431867599487305, "genuineness": 4.494794845581055, "identity": 0.6144887804985046, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/578e798bbc78cfdfe7a7f6f08a324948dc71ef6286d453daf89335b4a3e76bfb.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "58a02e135bbc2f44140f237c2bc202701cb4ab800c983ffb195db1d126a1726f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "87dd43e2f35650aedf2044a39d0c22f0b27c41d6be5797d54eb201ce8c49901d", "generation_seconds_batch": 20.86365207505878, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4530181884765625, "target_z": 1.054687553013311, "ce": 5.34748649597168, "genuineness": 4.607763767242432, "identity": 0.7771047353744507, "hypothesis": "La porte est encore ouverte, j'ai posé ton manteau sur la chaise, dis-moi quand tu es prêt", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/58a02e135bbc2f44140f237c2bc202701cb4ab800c983ffb195db1d126a1726f.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "595b68c395a620b175bf929a5ffa7961c72b50643c77a70a782d0491a20d8649", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4098c20cbda08fc17131c9ba6729108aed9f1515cf8038a0e7afdda6a32d0473", "generation_seconds_batch": 22.692709955968894, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6045653820037842, "target_z": -0.10156252388511819, "ce": 5.847874641418457, "genuineness": 4.507347106933594, "identity": 0.5421059131622314, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/595b68c395a620b175bf929a5ffa7961c72b50643c77a70a782d0491a20d8649.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "596f2f490bf6cab52a617c1956a67cfa34a58636a13d17abbf5c9fd8de8564e4", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e63cee85c862e2c9589fe34e7d743a523df98fd82745f10e3cc255443438e0e4", "generation_seconds_batch": 15.393032015068457, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4163858890533447, "target_z": 0.4062500209722989, "ce": 5.911416053771973, "genuineness": 4.607763767242432, "identity": 0.6189789772033691, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/596f2f490bf6cab52a617c1956a67cfa34a58636a13d17abbf5c9fd8de8564e4.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "59e6e82f4968a2107b0c76c243aff94957bf35f0355560609cce4c033dbac8c0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "98961c009b1b68439aac63695863f16e4f379f4fa9c8b8114cd3f803403b1e3f", "generation_seconds_batch": 23.974819825962186, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.828540802001953, "target_z": 0.6640624726194987, "ce": 5.71682071685791, "genuineness": 4.959221839904785, "identity": 0.7789602279663086, "hypothesis": "J'ai trouvé la petite boîte elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/59e6e82f4968a2107b0c76c243aff94957bf35f0355560609cce4c033dbac8c0.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "5a6512587595e108208815a418500f9d63d7a9de4de7fc463cb895c85466b8bf", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "972a0ecf99a3a614e0754857fb41367f8eaa208c428592d9bcee940ef4e5c458", "generation_seconds_batch": 17.12950436794199, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8035616874694824, "target_z": 0.6484374574728383, "ce": 5.450741291046143, "genuineness": 4.218649387359619, "identity": 0.27127739787101746, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar, wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5a6512587595e108208815a418500f9d63d7a9de4de7fc463cb895c85466b8bf.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "5a7c89fe8b5c40c989f26bc4902c751722f27c2fa7e201b09c7f4017f8c4e342", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "506b2602458430e2cde0bfd6e60c68a0f77cb9e584e9dba00842350c4969d6dc", "generation_seconds_batch": 21.7837157540489, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.422630548477173, "target_z": 0.41015620019079013, "ce": 5.7406487464904785, "genuineness": 4.68307638168335, "identity": 0.7092971801757812, "hypothesis": "Die Tür ist noch offen. Ich hab deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5a7c89fe8b5c40c989f26bc4902c751722f27c2fa7e201b09c7f4017f8c4e342.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "5a804954603f248fb6d4aca3e9ed13f2de2f01b681d4cbeca538341731aadb56", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "16a4c1bc96593e25b352c893634d81e0a295b1f190d7df45e356862f9d26c21b", "generation_seconds_batch": 16.14415013208054, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6653404235839844, "target_z": 1.187500032623576, "ce": 5.34748649597168, "genuineness": 4.356721878051758, "identity": 0.5793885588645935, "hypothesis": "The door is uh still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.05555555555555555, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5a804954603f248fb6d4aca3e9ed13f2de2f01b681d4cbeca538341731aadb56.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "5a914bf4e286b53adc16e303bbe6517b3dda8103ab67e70ba177e3aac37a751e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e0b34aeef9f8202e9be4e5645072bff8080917617a575dbf45deb28859e2ec59", "generation_seconds_batch": 22.70897358295042, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0970659255981445, "target_z": 0.8320311617415755, "ce": 5.557967185974121, "genuineness": 4.7081804275512695, "identity": 0.6152902841567993, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5a914bf4e286b53adc16e303bbe6517b3dda8103ab67e70ba177e3aac37a751e.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "5ab3eb75d8383743c5ad06a166945610c1651b21c421d72c65c111d38b620bda", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "53c387e29d01e308a487952e564b9e0468ffaadf34fafb3a622c10ce23089525", "generation_seconds_batch": 15.871158388908952, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4734251499176025, "target_z": -0.18359377883691097, "ce": 6.300607204437256, "genuineness": 4.595211505889893, "identity": 0.3974800109863281, "hypothesis": "I found the little box. It was beside the old window. We can look inside together,", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5ab3eb75d8383743c5ad06a166945610c1651b21c421d72c65c111d38b620bda.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "5b24ba286c1c662d154db0196562950edcfc8631a295e1053bb222dd000eee13", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5716e171e0a76c2f1fcebb0c7851abccb62227daff1e00d718956931a344acef", "generation_seconds_batch": 21.56337803194765, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7865002155303955, "target_z": -0.6132812479610265, "ce": 6.395919322967529, "genuineness": 4.494794845581055, "identity": 0.6751406192779541, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5b24ba286c1c662d154db0196562950edcfc8631a295e1053bb222dd000eee13.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "5b40874f86ff86f0ddbb6813220b06fbe5a010e778e0295a2f15fea442f6d2ea", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b6f7b7f820e154888536e82633494491ce28a45b285c9f2361ce64b669c7ee6e", "generation_seconds_batch": 20.786258936976083, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7606847286224365, "target_z": -0.003906253786665078, "ce": 6.284721851348877, "genuineness": 4.65797233581543, "identity": 0.7236710786819458, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5b40874f86ff86f0ddbb6813220b06fbe5a010e778e0295a2f15fea442f6d2ea.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "5b5a5db8290dd42684e457fa755c9c57f31940cfec33345321f469541f2944f4", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0fd084ab8dd0b7d349bdbe6ae245654d14d969917009a4c2800ab1fdd0b9db63", "generation_seconds_batch": 20.063229785999283, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7286243438720703, "target_z": 0.6015624120328574, "ce": 5.732706069946289, "genuineness": 4.582659721374512, "identity": 0.5278862118721008, "hypothesis": "De er staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5b5a5db8290dd42684e457fa755c9c57f31940cfec33345321f469541f2944f4.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "5c3542bd8f6192393fdd44d3bb99bc9c89e2fec91fba02c1b095f6bd96945d5d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "b9495e6d696f0f7b4a7187e6f429a02d30eb29e6049271056b7096456c4d8823", "generation_seconds_batch": 22.275458641001023, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.284409284591675, "target_z": 0.9492187753415279, "ce": 5.772419452667236, "genuineness": 4.344170093536377, "identity": 0.295986533164978, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5c3542bd8f6192393fdd44d3bb99bc9c89e2fec91fba02c1b095f6bd96945d5d.mp3" }, { "id": "heldout-fr-3", "language": "fr", "sentences": [ "Quelqu’un a déplacé la petite chaise bleue.", "Elle était ici... près de la grande bibliothèque.", "Nous la remettrons après le déjeuner." ], "text": "Quelqu’un a déplacé la petite chaise bleue. Elle était ici... près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-3", "key": "5d3322435d09b09fdb5d52ee67df2c008cdbe77d684186aff56190f207373b98", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Natural clear speech.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Natural clear speech.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "dbdbb1655a70169174eb278c982f84bb7184af942a277bf3c6ca9281d1ff7450", "generation_seconds_batch": 16.007171291974373, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.328122615814209, "target_z": 0.9765624772800096, "ce": 5.395142555236816, "genuineness": 4.557555198669434, "identity": 0.765573263168335, "hypothesis": "Quelqu'un a déplacé la petite chaise bleue, elle était ici, près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5d3322435d09b09fdb5d52ee67df2c008cdbe77d684186aff56190f207373b98.mp3" }, { "id": "heldout-nl-0", "language": "nl", "sentences": [ "Je bericht kwam nadat ik vertrokken was.", "Ik las het twee keer... en legde mijn telefoon weg.", "Zeg maar wanneer we kunnen praten." ], "text": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer... en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-0", "key": "5d9c05118ab1f8e8cbfe32ff384b80eb8d0bbd618b3ad3c39414608e033ec746", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Natural clear speech.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "48e16d313878e418e543c1386c337161cf952327cf85cffaba04a2b5175d29af", "generation_seconds_batch": 23.283461470971815, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2719197273254395, "target_z": 0.9414062677681977, "ce": 5.589737892150879, "genuineness": 4.268857479095459, "identity": 0.3025926947593689, "hypothesis": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5d9c05118ab1f8e8cbfe32ff384b80eb8d0bbd618b3ad3c39414608e033ec746.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "5da5dcc05454f6b67247ca647218dd8a4da774ca989bc73ef512cc939902d6aa", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d82f9571aac0677521b9d1048c803c09b2bb48d56f957b2bfb75b15799f27806", "generation_seconds_batch": 21.266878924914636, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.716135025024414, "target_z": 0.593750053595875, "ce": 5.446769714355469, "genuineness": 4.557555198669434, "identity": 0.6663205623626709, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5da5dcc05454f6b67247ca647218dd8a4da774ca989bc73ef512cc939902d6aa.mp3" }, { "id": "heldout-en-1", "language": "en", "sentences": [ "The keys are on the kitchen table.", "I checked the drawer... then looked beside the window.", "Take them with you when you leave." ], "text": "The keys are on the kitchen table. I checked the drawer... then looked beside the window. Take them with you when you leave.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-1", "key": "5dabbd2ace997c2b7245e4b4b8d59db43c7d0442592ba98b4383541a9a0bb4f2", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The keys are on the kitchen table.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Take them with you when you leave.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The keys are on the kitchen table.\n(Natural clear speech.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Natural clear speech.) [3.2 seconds duration] Take them with you when you leave.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "002c59cd9e07e3636fbc8395ce48e802509df0492e78eaea6359ae7c659764cc", "generation_seconds_batch": 17.014255692018196, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0221288204193115, "target_z": 0.7851562654379423, "ce": 5.609594821929932, "genuineness": 4.425758361816406, "identity": 0.20901153981685638, "hypothesis": "The keys are on the kitchen table. I checked the drawer, then looked beside the window. Take them with you when you leave.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5dabbd2ace997c2b7245e4b4b8d59db43c7d0442592ba98b4383541a9a0bb4f2.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "5df98cbf3195baf364abb87e6d2f5b18b75caf1eb82767f18c5802e24e1ab03f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4cbfdca405a5a361ff39d67c7aea9b3f01f795467ae569a86b1773291b6015a7", "generation_seconds_batch": 16.805090085952543, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6403613090515137, "target_z": 1.1718750174769157, "ce": 5.454712867736816, "genuineness": 4.507347106933594, "identity": 0.5348498821258545, "hypothesis": "La puerta sigue abierta. he dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5df98cbf3195baf364abb87e6d2f5b18b75caf1eb82767f18c5802e24e1ab03f.mp3" }, { "id": "heldout-fr-2", "language": "fr", "sentences": [ "La route tourne à gauche après le pont.", "Continue... jusqu’à la petite barrière en bois.", "Je t’attendrai de l’autre côté." ], "text": "La route tourne à gauche après le pont. Continue... jusqu’à la petite barrière en bois. Je t’attendrai de l’autre côté.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-2", "key": "5e0af124ad7421917e897f2876d49b25e213a1bc87e01fe43f09b9c255c2a32f", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d114ae8f969098dbd0c667becac1bbb24f639a4829eda2a5dba0e34b7671eff5", "generation_seconds_batch": 17.4834213150898, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9292936325073242, "target_z": 0.10156252388511819, "ce": 6.197352409362793, "genuineness": 4.557555198669434, "identity": 0.9014685750007629, "hypothesis": "La route tourne à gauche après le pont. Continue, jusqu'à la petite barrière en bois. Je t'attendrai de l'autre côté.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5e0af124ad7421917e897f2876d49b25e213a1bc87e01fe43f09b9c255c2a32f.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "5e10b1a3dfbbee118ca4d95b65bd3ac9f3752dcfca333f6cd2fdce75646b3377", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "72047e6715d5dd06f5f14b6ed4d1053bc8fd68a74b26eb77722bfaa184821d65", "generation_seconds_batch": 22.692709955968894, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4796699285507202, "target_z": -0.1796875250502459, "ce": 5.982900142669678, "genuineness": 4.582659721374512, "identity": 0.6324520111083984, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5e10b1a3dfbbee118ca4d95b65bd3ac9f3752dcfca333f6cd2fdce75646b3377.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "5e4f2cf311ff0ee9b1602d3356fc0de96f606fae7f42b91d4d035cfb221e4266", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "fdbe05054ec2b507c88827bc2da0d5ee439b6d28aa6d7a7b64d59b2b986e7777", "generation_seconds_batch": 22.40843909897376, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5568933486938477, "target_z": 0.4941406938881762, "ce": 6.205295085906982, "genuineness": 4.68307638168335, "identity": 0.8363863229751587, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5e4f2cf311ff0ee9b1602d3356fc0de96f606fae7f42b91d4d035cfb221e4266.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "5e7613d0d14e75267dc81a65851be2b177adb8ad0e4fd535fee93b90b4a9f068", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "ba8f24676cc5344d5c69dca7e9b0d3b8be35fac439fc9375ee59b3bd0ad083a2", "generation_seconds_batch": 22.475935191032477, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1720032691955566, "target_z": 0.8789062071815564, "ce": 5.617537498474121, "genuineness": 4.118232727050781, "identity": 0.4435777962207794, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.120000000000001, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5e7613d0d14e75267dc81a65851be2b177adb8ad0e4fd535fee93b90b4a9f068.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "5ee78a6b9a011af45ff6f991927f502962238366b57813948202aa9aa11e308a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d9ecd876e62cb11b01f0d61b680ab99ed2914f5283ca9165f76938166b35668f", "generation_seconds_batch": 21.672672353102826, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1549417972564697, "target_z": -0.38281249825230845, "ce": 6.292664527893066, "genuineness": 4.821148872375488, "identity": 0.6970577239990234, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5ee78a6b9a011af45ff6f991927f502962238366b57813948202aa9aa11e308a.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "5f1b46f7332bc5e92bb84a73696bed3519784177b7b154fa072b5279a7203f9b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "4506a58ca561d72d52752e68c812c9df79a8295214cc00844005593bebc3491e", "generation_seconds_batch": 21.991056522005238, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.140779495239258, "target_z": 0.8593750128164048, "ce": 5.7684478759765625, "genuineness": 4.6203155517578125, "identity": 0.42373090982437134, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5f1b46f7332bc5e92bb84a73696bed3519784177b7b154fa072b5279a7203f9b.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "5f34e171a6ad4ce7bed827939149e5f7f283e93be19c9e7b32515694e8e83b73", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ada8a798a3eccf0107f1ea9d7e62a22d5c06b55be7d96adb3169798b30b625d2", "generation_seconds_batch": 21.820619402104057, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6451563835144043, "target_z": -0.0761719115558821, "ce": 6.1020402908325195, "genuineness": 4.65797233581543, "identity": 0.6387441754341125, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5f34e171a6ad4ce7bed827939149e5f7f283e93be19c9e7b32515694e8e83b73.mp3" }, { "id": "heldout-fr-1", "language": "fr", "sentences": [ "Les clés sont sur la table de la cuisine.", "J’ai regardé dans le tiroir... puis près de la fenêtre.", "Prends-les quand tu partiras." ], "text": "Les clés sont sur la table de la cuisine. J’ai regardé dans le tiroir... puis près de la fenêtre. Prends-les quand tu partiras.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-1", "key": "5f3b2a4df4b8df766802ea181c596b794c39aa6c88602c4ad7f7b477f11e2ad4", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Prends-les quand tu partiras.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Natural clear speech.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Prends-les quand tu partiras.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "28dded3bd648036ab15e877c19f08e7168bdc39fb0d53fb22eff9b7e4ca185e7", "generation_seconds_batch": 16.741675611003302, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.29689884185791, "target_z": 0.957031282914858, "ce": 5.434855937957764, "genuineness": 4.632867813110352, "identity": 0.7440724968910217, "hypothesis": "Les clés sont sur la table de la cuisine. J'ai regardé dans le tiroir, puis près de la fenêtre. Prends-les quand tu partiras.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5f3b2a4df4b8df766802ea181c596b794c39aa6c88602c4ad7f7b477f11e2ad4.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "5fa229c94d2a9e0dbe19ebf97097326a754abafe9af2ce65a957f5367fbd72f0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6b68d96422391d1deebe47e4e4724aa888642b00e78464c9e093cf4d594910c4", "generation_seconds_batch": 16.535797354998067, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2344510555267334, "target_z": 0.9179687450482072, "ce": 5.19657564163208, "genuineness": 4.457139015197754, "identity": 0.4339001774787903, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5fa229c94d2a9e0dbe19ebf97097326a754abafe9af2ce65a957f5367fbd72f0.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "5fc14be7e45174d2a7abcbb76269ffe982b7ff7ef2cfa5a53157440d0f83c855", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4ee8283d2532a229e6af1fb1e13b1c2d5c8b52670ac7333a05c3f7ec35ff664c", "generation_seconds_batch": 17.14468647004105, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0533528327941895, "target_z": 0.8046876089394415, "ce": 5.327630043029785, "genuineness": 4.532451152801514, "identity": 0.36256229877471924, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen heen kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5fc14be7e45174d2a7abcbb76269ffe982b7ff7ef2cfa5a53157440d0f83c855.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "5fd782ce37a1ec6896b7a23455acee7cc640cd6e6298a288b66f0a6ae60730f5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "2940f76b009611a5f77bbc2f0533f1104087d2f1bc6e36a70a4fa06b837d2ef4", "generation_seconds_batch": 26.75805490789935, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5654239654541016, "target_z": 1.1249999720369348, "ce": 5.621508598327637, "genuineness": 4.344170093536377, "identity": 0.3423924446105957, "hypothesis": "The door is still open. I put your goat on the chair. Tell me when you are ready.", "wer": 0.05555555555555555, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/5fd782ce37a1ec6896b7a23455acee7cc640cd6e6298a288b66f0a6ae60730f5.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "603e3c3c50f19547ccf7e725b41a7eda20d0a38a7a088158ddc4e717b37e4081", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "04d01a685f3b63dabe0cb4ab5b3cc2cc1df3b0a6d268e9fa732c5c0b03fab7fe", "generation_seconds_batch": 21.962248128955252, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6028926372528076, "target_z": 1.1484374947569254, "ce": 5.657250881195068, "genuineness": 4.65797233581543, "identity": 0.6491576433181763, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/603e3c3c50f19547ccf7e725b41a7eda20d0a38a7a088158ddc4e717b37e4081.mp3" }, { "id": "heldout-en-3", "language": "en", "sentences": [ "Someone moved the little blue chair.", "It used to stand here... beside the tall bookshelf.", "We can put it back after lunch." ], "text": "Someone moved the little blue chair. It used to stand here... beside the tall bookshelf. We can put it back after lunch.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-3", "key": "60b4cb5589ac8a1eb29257f1b975395f082eed58adbf760b24157784e5805664", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Someone moved the little blue chair.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can put it back after lunch.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Someone moved the little blue chair.\n(Natural clear speech.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Natural clear speech.) [3.2 seconds duration] We can put it back after lunch.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2ee5912614ae62f138f3aec0476eee6f101b2731dea4b964fc15f82aaeb5bc8f", "generation_seconds_batch": 22.48715424502734, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.284409284591675, "target_z": 0.9492187753415279, "ce": 5.538110733032227, "genuineness": 4.2751336097717285, "identity": 0.3729255199432373, "hypothesis": "Someone moved the little blue chair. It used to stand here beside the tall bookshelf. We can put it back after lunch.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/60b4cb5589ac8a1eb29257f1b975395f082eed58adbf760b24157784e5805664.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "60b71baef9bbb037cba2d5db1fd08a4d71a504974c8c8ba1ec708d04ebc62e83", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5345f81e08de922a08f07dcfeb4a1f3efd8d6cb6f4d447ca816109a5a263b112", "generation_seconds_batch": 23.3603953019483, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.34769344329834, "target_z": 0.3632813038871569, "ce": 6.149696350097656, "genuineness": 4.570107460021973, "identity": 0.5701185464859009, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/60b71baef9bbb037cba2d5db1fd08a4d71a504974c8c8ba1ec708d04ebc62e83.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "60e0cc3db995666b35864190f61b165b3d5d0bc3cdcf6359fee7b845460858cf", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6515ebac44cdc38223d61ee71ddeb1004a1712a7e6ce697afe33541f06034bb0", "generation_seconds_batch": 17.137505569960922, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0970659255981445, "target_z": 0.8320311617415755, "ce": 5.736677646636963, "genuineness": 4.846253395080566, "identity": 0.5580739378929138, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/60e0cc3db995666b35864190f61b165b3d5d0bc3cdcf6359fee7b845460858cf.mp3" }, { "id": "heldout-de-2", "language": "de", "sentences": [ "Die Straße führt hinter der Brücke nach links.", "Geh weiter... bis zu dem kleinen Holztor.", "Ich warte auf der anderen Seite." ], "text": "Die Straße führt hinter der Brücke nach links. Geh weiter... bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-2", "key": "611943363d8fba2885feddc858b27bb0469859c63d61fd5cac9387f1cae27b74", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Natural clear speech.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Natural clear speech.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4cb211aa632517c5602fc14629c65886448be9165cc2d0aab693b148413ed553", "generation_seconds_batch": 17.179478898993693, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0158839225769043, "target_z": 0.7812499370831033, "ce": 5.629451274871826, "genuineness": 4.507347106933594, "identity": 0.528414785861969, "hypothesis": "Die Straße führt hinter der Brücke nach links. Geh weiter bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/611943363d8fba2885feddc858b27bb0469859c63d61fd5cac9387f1cae27b74.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "61432b17a1a629625f0bc74ca78b72e6484e7966c28ce0c00b30fbafb0c69e2c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "46d3f232afe99a6137be7c30e32a9680486adfde912beb6dd241cdc2da9657b2", "generation_seconds_batch": 16.31968818604946, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2829596996307373, "target_z": -0.30273433290976126, "ce": 5.974957466125488, "genuineness": 4.632867813110352, "identity": 0.6156497001647949, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/61432b17a1a629625f0bc74ca78b72e6484e7966c28ce0c00b30fbafb0c69e2c.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "61efff4acccbeb92a7f8bb5dba67a56a5091396dab9f05df2e33db6f86f90fa2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "c00169bf3fc83dfc84557a1f0fcb43fd8354a61825d461012baf359186af9c94", "generation_seconds_batch": 23.5427127830917, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4155495166778564, "target_z": 1.0312500302933205, "ce": 6.205295085906982, "genuineness": 4.607763767242432, "identity": 0.16695636510849, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/61efff4acccbeb92a7f8bb5dba67a56a5091396dab9f05df2e33db6f86f90fa2.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "61f62a666bcf765520cff0b35d466cebe65a348d1a5b8f8c257a6fc34b5cb428", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "89d70cc73466be254c3fc897fba7fcb27dc5c9491ab05447c3a065a82e73d412", "generation_seconds_batch": 23.79720372101292, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.244654655456543, "target_z": 0.29882815369127, "ce": 6.165581703186035, "genuineness": 4.557555198669434, "identity": 0.742512047290802, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/61f62a666bcf765520cff0b35d466cebe65a348d1a5b8f8c257a6fc34b5cb428.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "625c86748b55492eab982bc7c89e00457372f2a332d21347c1b603e46fa80eae", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "466e46668f18ada0febbc5a4e0b80e93d798264565a8b4a34b9533fe18ff46c3", "generation_seconds_batch": 25.29638444702141, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0221288204193115, "target_z": 0.7851562654379423, "ce": 6.0305562019348145, "genuineness": 4.6705241203308105, "identity": 0.7332818508148193, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/625c86748b55492eab982bc7c89e00457372f2a332d21347c1b603e46fa80eae.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "627c1107ff2d38393383d52ec7ee2d3ee7d74802f27ff46b368ded91cf1f38df", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e59e0b89b717c3460b4d475cda3156638989da976c6cafbe25b65de2190063fa", "generation_seconds_batch": 21.17429653799627, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.403059959411621, "target_z": 1.0234375227199906, "ce": 5.601652145385742, "genuineness": 4.218649387359619, "identity": 0.3721214532852173, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude laam. We kunnen er samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/627c1107ff2d38393383d52ec7ee2d3ee7d74802f27ff46b368ded91cf1f38df.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "62d9eb0bab5dd213f55851ecc13f074705b3337ea7691cc4bc6de6fa2345e26b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "dd15cbe58cf51abec31ab0078458c9e70c75442b99aebab62db0a8dd7661fb86", "generation_seconds_batch": 15.883723756065592, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.497567892074585, "target_z": 0.4570312456307711, "ce": 5.355429172515869, "genuineness": 4.557555198669434, "identity": 0.6160374283790588, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/62d9eb0bab5dd213f55851ecc13f074705b3337ea7691cc4bc6de6fa2345e26b.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "62fdb8a6d08e1b09cb9a76d9ae555bd68b512a52175621434d09703e17322ff0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "950843b90a65c44703d45590bd381c31806c7719bf693028553352ce158ece2d", "generation_seconds_batch": 23.13597337307874, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7794190645217896, "target_z": 0.007812507573330156, "ce": 6.054384231567383, "genuineness": 4.444586753845215, "identity": 0.3742513656616211, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/62fdb8a6d08e1b09cb9a76d9ae555bd68b512a52175621434d09703e17322ff0.mp3" }, { "id": "heldout-fr-2", "language": "fr", "sentences": [ "La route tourne à gauche après le pont.", "Continue... jusqu’à la petite barrière en bois.", "Je t’attendrai de l’autre côté." ], "text": "La route tourne à gauche après le pont. Continue... jusqu’à la petite barrière en bois. Je t’attendrai de l’autre côté.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-2", "key": "635ccfc34b3ba324f967e4cf7902ee93deb0c6b58f93793ed71b1228cd975412", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Natural clear speech.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Natural clear speech.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "2eefb08db7ffb915a811a20ab4a18a74524a9d999265ac0db4ff8c2c23dd1ab1", "generation_seconds_batch": 16.705798128037713, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4530181884765625, "target_z": 1.054687553013311, "ce": 5.5539960861206055, "genuineness": 4.507347106933594, "identity": 0.8230466842651367, "hypothesis": "La route tourne à gauche après le pont, continue jusqu'à la petite barrière en bois. Je t'attendrai de l'autre côté.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/635ccfc34b3ba324f967e4cf7902ee93deb0c6b58f93793ed71b1228cd975412.mp3" }, { "id": "heldout-es-3", "language": "es", "sentences": [ "Alguien ha movido la pequeña silla azul.", "Estaba aquí... junto a la estantería alta.", "La colocaremos después de comer." ], "text": "Alguien ha movido la pequeña silla azul. Estaba aquí... junto a la estantería alta. La colocaremos después de comer.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-3", "key": "635eee56ec4b871f02c7ae614eef9b8952f25098133e0cb4519e699bc5bb85fe", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] La colocaremos después de comer.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Natural clear speech.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Natural clear speech.) [3.2 seconds duration] La colocaremos después de comer.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7986ac0279cbf660f5dd68fe877e64f065fe42fadeb99f1906b0829c52248e96", "generation_seconds_batch": 17.7595562429633, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.365591287612915, "target_z": 1.0, "ce": 5.383228778839111, "genuineness": 4.607763767242432, "identity": 0.4657232463359833, "hypothesis": "Alguien ha movido la pequeña silla azul. Estaba aquí, junto a la estantería alta. La colocaremos después de comer.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/635eee56ec4b871f02c7ae614eef9b8952f25098133e0cb4519e699bc5bb85fe.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "6365d42308d81dd3cc118341e872b41e147f21bc2b7fdb1ff4631ce7a3a8c9e0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3a4d0e6e06321f17b4fde75c3ce9151dabbb6f54fc4c20fad59c3ed158b4348b", "generation_seconds_batch": 16.465320190996863, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0550254583358765, "target_z": -0.44531248427077585, "ce": 6.09409761428833, "genuineness": 4.6705241203308105, "identity": 0.8105077743530273, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6365d42308d81dd3cc118341e872b41e147f21bc2b7fdb1ff4631ce7a3a8c9e0.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "63cd3c57ebcf5c06d1ada06008cec6c2c058e6c31f82b242acf1f1203d80f5e8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "efa5afc1b2ea111cbe29334356dfafd2ea8a0fa689ed7c4b853484b273a4b190", "generation_seconds_batch": 22.287599133094773, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6545236110687256, "target_z": -0.07031249359179756, "ce": 5.645336627960205, "genuineness": 4.645420074462891, "identity": 0.676892101764679, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/63cd3c57ebcf5c06d1ada06008cec6c2c058e6c31f82b242acf1f1203d80f5e8.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "63d01916a4789b53e83c8f061299cffdbbf6cc11b9bac24e2a2bf9c8956cea68", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "358cf4e0851fafca504b88ae7381f4ba7215148a458de7010c809a06885bab5c", "generation_seconds_batch": 21.43358326505404, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.502976417541504, "target_z": 1.0859375833066318, "ce": 5.202532768249512, "genuineness": 4.557555198669434, "identity": 0.7179434299468994, "hypothesis": "La porte est encore ouverte, j'ai posé ton manteau sur la chaise, dis-moi quand tu es prêt", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/63d01916a4789b53e83c8f061299cffdbbf6cc11b9bac24e2a2bf9c8956cea68.mp3" }, { "id": "heldout-fr-0", "language": "fr", "sentences": [ "Ton message est arrivé après mon départ.", "Je l’ai relu... puis j’ai rangé mon téléphone.", "Dis-moi quand nous pourrons parler." ], "text": "Ton message est arrivé après mon départ. Je l’ai relu... puis j’ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-0", "key": "641694c4acb59421e1878c41bc4206a03bb786df80b48acd06e1fe4523b32a69", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Natural clear speech.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d5ac8702bb8362a284061d9de40b9f29b60b5f849a4f97f4ace096e485bce01a", "generation_seconds_batch": 16.504592689918354, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3905704021453857, "target_z": 1.0156250151466604, "ce": 5.0774359703063965, "genuineness": 4.432034492492676, "identity": 0.6909722685813904, "hypothesis": "Ton message est arrivé après mon départ, je l'ai relu, puis j'ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/641694c4acb59421e1878c41bc4206a03bb786df80b48acd06e1fe4523b32a69.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "645d3155e2bc9d5d0574515d81c01add80fc3e401d9d746b0491c0eb9deb3605", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5de8c66cfd85bc7cd906e200b1e33b0aa8190ddf38dbae86cbdd553bf59ed04e", "generation_seconds_batch": 20.864102661958896, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7357056140899658, "target_z": -0.01953126893332539, "ce": 5.967014789581299, "genuineness": 4.369274139404297, "identity": 0.6516483426094055, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/645d3155e2bc9d5d0574515d81c01add80fc3e401d9d746b0491c0eb9deb3605.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "6486c098379203030889e179aec9e022063996978956df5f19878ed058938e39", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2b5e344f99072cf5669b5053ae980dfb3492c58a9c55bc288d64e8b01a29b752", "generation_seconds_batch": 22.53629216598347, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9667621850967407, "target_z": 0.1249999720369348, "ce": 6.149696350097656, "genuineness": 4.425758361816406, "identity": 0.5038267970085144, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6486c098379203030889e179aec9e022063996978956df5f19878ed058938e39.mp3" }, { "id": "heldout-fr-3", "language": "fr", "sentences": [ "Quelqu’un a déplacé la petite chaise bleue.", "Elle était ici... près de la grande bibliothèque.", "Nous la remettrons après le déjeuner." ], "text": "Quelqu’un a déplacé la petite chaise bleue. Elle était ici... près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-3", "key": "64a4a3e87ba108f8bc61de3524db6113a986bf4a67e9dd1f9db9337b95b0da51", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Natural clear speech.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Natural clear speech.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6ae61797d8f1a9323af90fd8cb7b0ecc10b2569689bff81ea11b58a41f0f2f03", "generation_seconds_batch": 23.34268284298014, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1033108234405518, "target_z": 0.8359374900964144, "ce": 5.252174377441406, "genuineness": 4.494794845581055, "identity": 0.7292512059211731, "hypothesis": "Quelqu'un a déplacé la petite chaise bleue, elle était ici, près de la grande bibliothèque, nous la remettrons après le déjeuner.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/64a4a3e87ba108f8bc61de3524db6113a986bf4a67e9dd1f9db9337b95b0da51.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "652196d3081e8bd42374b991df27afe538002208d3698456a382c3ffc793e56c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "16d0416f118f8b39d77e242234ae052978772fa4e1fbf2f2ed1916b6155d2cd9", "generation_seconds_batch": 22.569874503999017, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.101024866104126, "target_z": 0.20898439116614706, "ce": 6.078212261199951, "genuineness": 4.68307638168335, "identity": 0.6847842931747437, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/652196d3081e8bd42374b991df27afe538002208d3698456a382c3ffc793e56c.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "65834cffa2fbc02862807d00b0ddc6405bf4e3657294ce9338c8e85f0ba914ef", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b4b482416ed482341bbbd34f75c4ecd65baf8dda76cfd047ef354b51f06298bd", "generation_seconds_batch": 22.765458481037058, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7660932540893555, "target_z": 0.6250000838891956, "ce": 5.990842819213867, "genuineness": 4.557555198669434, "identity": 0.7380003333091736, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/65834cffa2fbc02862807d00b0ddc6405bf4e3657294ce9338c8e85f0ba914ef.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "66c1db93303d412cdb23cfa127da36b50df574067692b45836e4c604d56f5cbf", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c9bcd3acf6be4fda0576ae3ac4fecae9e73421ce040924e7ebe10a2fb5bb7fce", "generation_seconds_batch": 21.92380141292233, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4453237056732178, "target_z": -0.2011718835928169, "ce": 6.260893821716309, "genuineness": 4.833701133728027, "identity": 0.6643600463867188, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/66c1db93303d412cdb23cfa127da36b50df574067692b45836e4c604d56f5cbf.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "66c6d4fdf3d5a384fdac674507350502e376253095de9454ebf9109a8700acae", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "21244c110e481f7dc3a8b69d9914211e9170d1189c7a4524ac2a20a72c09ff32", "generation_seconds_batch": 16.764624492963776, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.026087522506714, "target_z": 0.16210934572616612, "ce": 5.9431867599487305, "genuineness": 4.65797233581543, "identity": 0.6490777134895325, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/66c6d4fdf3d5a384fdac674507350502e376253095de9454ebf9109a8700acae.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "66dec2e08064de54ae9d5e2d88f41b629c8ce77c43f08e5b0de7c4eed791dfaa", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "218a3fe7506358b4ebb5f05747a1a0471a14c8eba033a2150e1b5e1b1984e991", "generation_seconds_batch": 23.592666225973517, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3352038860321045, "target_z": 0.3554687963138267, "ce": 6.197352409362793, "genuineness": 4.595211505889893, "identity": 0.4784470200538635, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/66dec2e08064de54ae9d5e2d88f41b629c8ce77c43f08e5b0de7c4eed791dfaa.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "6712de0157261bf0a09f3f84d462f104a28684563254accdcec312a055c0b164", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4c3a27521f42d280565e25bac52836f4d522712923697889774915a97fe29262", "generation_seconds_batch": 22.252671992057003, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4757111072540283, "target_z": 0.4433593200933564, "ce": 6.173524379730225, "genuineness": 4.65797233581543, "identity": 0.5300599932670593, "hypothesis": "Ik heb het kleine doosje gevonden, het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6712de0157261bf0a09f3f84d462f104a28684563254accdcec312a055c0b164.mp3" }, { "id": "heldout-es-2", "language": "es", "sentences": [ "La carretera gira a la izquierda tras el puente.", "Sigue... hasta la pequeña puerta de madera.", "Te esperaré al otro lado." ], "text": "La carretera gira a la izquierda tras el puente. Sigue... hasta la pequeña puerta de madera. Te esperaré al otro lado.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-2", "key": "675ae2e74172df40dbc027e3018bb7a570088d2790d390230acf5a9c69233d97", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Te esperaré al otro lado.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Natural clear speech.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Natural clear speech.) [3.2 seconds duration] Te esperaré al otro lado.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "5e2c62166a06a46625ce7f7eb1c9e7d6a09f63ead04bdc41606da9ea5f41cdfc", "generation_seconds_batch": 17.08183075103443, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.034618377685547, "target_z": 0.7929687730112724, "ce": 5.609594821929932, "genuineness": 4.319066047668457, "identity": 0.46634969115257263, "hypothesis": "La carretera gira a la izquierda tras el puente. sigue hasta la pequeña puerta de madera. Te esperaré al otro lado.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/675ae2e74172df40dbc027e3018bb7a570088d2790d390230acf5a9c69233d97.mp3" }, { "id": "heldout-nl-3", "language": "nl", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-3", "key": "675d7512749becfe70dffb3e0edbab940353d172dae06f44e495c722cf1d01d6", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Natural clear speech.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Natural clear speech.) [3.2 seconds duration] We zetten het na de lunch terug.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "fd0eb0205327c75e538362c690a61b2d3513fa8f3af86e11778c6d5c9c69f694", "generation_seconds_batch": 21.546583125018515, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.935538411140442, "target_z": 0.10546877767178327, "ce": 5.609594821929932, "genuineness": 4.369274139404297, "identity": 0.33945149183273315, "hypothesis": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier naast de hoge boekenkast. We zetten het naar de lunch terug.", "wer": 0.047619047619047616, "duration_error_s": 0.120000000000001, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/675d7512749becfe70dffb3e0edbab940353d172dae06f44e495c722cf1d01d6.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "677f149920495622067a767f660c2786f3d2e2230df35711695d1b0189666191", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "da87410da780c7516dc9b38bb360251e5fd568d18ee2ed74541e78f6de2c6e5f", "generation_seconds_batch": 22.215124303940684, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8152151107788086, "target_z": 1.2812501235035378, "ce": 5.8875885009765625, "genuineness": 4.457139015197754, "identity": 0.3255356550216675, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/677f149920495622067a767f660c2786f3d2e2230df35711695d1b0189666191.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "67bfe5e915f622619aa1264831c4881dee2d662767fc76120376f40f0198dbe9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3b2a82270d188f66da8a5fd12c0b4d5215b467ab155c8dfc1b11f094eae3e1f3", "generation_seconds_batch": 23.25329788296949, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.244654655456543, "target_z": 0.29882815369127, "ce": 5.637393951416016, "genuineness": 4.469690799713135, "identity": 0.6565781831741333, "hypothesis": "De deur staat nog open. ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/67bfe5e915f622619aa1264831c4881dee2d662767fc76120376f40f0198dbe9.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "686ac0e1c6117fadb2645cb234417f39f3c031f76d03b32edd5112a554420bf2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "24cf0a4776f6a7f171f867dec88f10debd3305ed286e14549ba801255fbe621a", "generation_seconds_batch": 22.39473205001559, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.641197681427002, "target_z": 0.546875008155894, "ce": 5.959072113037109, "genuineness": 4.758388519287109, "identity": 0.6355625987052917, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/686ac0e1c6117fadb2645cb234417f39f3c031f76d03b32edd5112a554420bf2.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "68a453f01169fe6ffd6ff3d8e0a1960b0586f054e7126a3383c1d7f0ee20b286", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4806ea2ffe4af699d20fd4112d03c281fbde1865b8c33609af3ea64daa5c48be", "generation_seconds_batch": 16.763441997929476, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5171386003494263, "target_z": -0.15625000233025543, "ce": 5.998785495758057, "genuineness": 4.6956281661987305, "identity": 0.7638112902641296, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/68a453f01169fe6ffd6ff3d8e0a1960b0586f054e7126a3383c1d7f0ee20b286.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "68c8faa566ef97c042ec2434352df3e4518045728b984e4f1df783dddcfb6543", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d54705838de5885cbc716bbb087187bf89be34c45e12e435f159216d86616297", "generation_seconds_batch": 15.50182735407725, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.940946578979492, "target_z": 0.7343748916431224, "ce": 5.903473377227783, "genuineness": 4.720732688903809, "identity": 0.5778281688690186, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/68c8faa566ef97c042ec2434352df3e4518045728b984e4f1df783dddcfb6543.mp3" }, { "id": "heldout-fr-1", "language": "fr", "sentences": [ "Les clés sont sur la table de la cuisine.", "J’ai regardé dans le tiroir... puis près de la fenêtre.", "Prends-les quand tu partiras." ], "text": "Les clés sont sur la table de la cuisine. J’ai regardé dans le tiroir... puis près de la fenêtre. Prends-les quand tu partiras.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-1", "key": "68f3dae896fd8e536f5c8bda7cc8eeeb20853400ead5589b5bb439dc8c2deb2a", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Prends-les quand tu partiras.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Natural clear speech.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Prends-les quand tu partiras.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "537c40a2026ad86db8443544f1b5f72acb2c81429362ef9031114368a39ef8de", "generation_seconds_batch": 16.561459264950827, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.328122615814209, "target_z": 0.9765624772800096, "ce": 5.522225379943848, "genuineness": 4.6705241203308105, "identity": 0.7579017877578735, "hypothesis": "Les clés sont sur la table de la cuisine. J'ai regardé dans le tiroir, puis près de la fenêtre. Prends-les quand tu partiras.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/68f3dae896fd8e536f5c8bda7cc8eeeb20853400ead5589b5bb439dc8c2deb2a.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "69607865db8f2403f5ec001db07608af6b8021b9ddc00b1127ac66de0f292510", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cd892d08a6c4ea2b3f2afbcb91bd3ff4c8cebc775de487662181b1c5a9a48cc4", "generation_seconds_batch": 21.77016671700403, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9230488538742065, "target_z": 0.09765627009845311, "ce": 6.062326908111572, "genuineness": 4.545003414154053, "identity": 0.6756901741027832, "hypothesis": "Ik heb het kleine doosje gevonden, het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/69607865db8f2403f5ec001db07608af6b8021b9ddc00b1127ac66de0f292510.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "69698b2f8b75760e992fe5ab1c9c23934ad31863f1f7b314cd7f97c4a4c2ea03", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "8d2b2f6ffc9d773689fd4770331d8cca93d272f427ba20c1e27ac759a00efd9b", "generation_seconds_batch": 21.99430741602555, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2781643867492676, "target_z": 0.9453124469866889, "ce": 5.613565921783447, "genuineness": 4.444586753845215, "identity": 0.4638374447822571, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/69698b2f8b75760e992fe5ab1c9c23934ad31863f1f7b314cd7f97c4a4c2ea03.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "6979f3d8807a655bc2588f7233dfd010a1117722caafe39bf6d9fdffb37b40d1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0e0f2f089a3820dda8af2c2e7cf743aa154ba390c77e8f8fb01fa240fe837d67", "generation_seconds_batch": 23.787151047959924, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0831269025802612, "target_z": -0.4277343795148699, "ce": 5.982900142669678, "genuineness": 4.557555198669434, "identity": 0.6622947454452515, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6979f3d8807a655bc2588f7233dfd010a1117722caafe39bf6d9fdffb37b40d1.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "6986dca0e456dc7d9916b2c41a4de43e295706a0ca128ce2701e45e4ec6fc8ad", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "94ef96febc4a71f7c3e2316d4f7d69808875e65c84737ee30ed9d144d8d06984", "generation_seconds_batch": 23.629294826998375, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8472752571105957, "target_z": 0.6757813085476677, "ce": 6.022613525390625, "genuineness": 4.758388519287109, "identity": 0.4213563799858093, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6986dca0e456dc7d9916b2c41a4de43e295706a0ca128ce2701e45e4ec6fc8ad.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "69cf87060f8c2d7dd9ed6aa0483dd41afd7effd780b28e91b3119fb0bc54ec34", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "df3d96ce3ddc3a772596948d3d2c99f3ec8ea5f5c8aca9acd7c83e8ba373088e", "generation_seconds_batch": 23.246597405988723, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1112284660339355, "target_z": -0.41015620019079013, "ce": 6.308549880981445, "genuineness": 4.6956281661987305, "identity": 0.720236599445343, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/69cf87060f8c2d7dd9ed6aa0483dd41afd7effd780b28e91b3119fb0bc54ec34.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "6a1ba1beeffa62ae77bffa56a65c36dbde5818e6bbdad88f2cd937ae9d538ba9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "aac77ecbaa52176b88484f90ea52e95b58c643c4550d0a6fc680a88e25db2330", "generation_seconds_batch": 23.220006208051927, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.365591287612915, "target_z": 1.0, "ce": 5.7247633934021, "genuineness": 4.457139015197754, "identity": 0.42083966732025146, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6a1ba1beeffa62ae77bffa56a65c36dbde5818e6bbdad88f2cd937ae9d538ba9.mp3" }, { "id": "heldout-en-4", "language": "en", "sentences": [ "I wrote the number on the envelope.", "Read it carefully... and check the last two digits.", "Call me when you have a moment." ], "text": "I wrote the number on the envelope. Read it carefully... and check the last two digits. Call me when you have a moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-4", "key": "6a61bc20d43b3e97640d8dd45682f7055c5dbaa5aa987cdb55f8a7d7a432a2c2", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I wrote the number on the envelope.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Call me when you have a moment.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I wrote the number on the envelope.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Call me when you have a moment.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "fe4a6d4d8e83ec75ae3b559dbe25f973d8ec6d7cd94472ccc3bce411eab9f53e", "generation_seconds_batch": 16.843197706970386, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.3930795192718506, "target_z": -0.8593750128164048, "ce": 6.3641486167907715, "genuineness": 4.124508857727051, "identity": 0.3344114422798157, "hypothesis": "I wrote the number on the envelope. Read it carefully and check the last two digits. Call me when you have a moment.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6a61bc20d43b3e97640d8dd45682f7055c5dbaa5aa987cdb55f8a7d7a432a2c2.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "6a77752bc6f75f5199bcbd1d1c8d825c5423bfcf857f756fd25416f6394cc345", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "469dc4e0d9593739e4e3fe52352aa0133455a323e1856e498f884f321fac88c5", "generation_seconds_batch": 19.738084413926117, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.032332420349121, "target_z": 0.16601567408100504, "ce": 6.09409761428833, "genuineness": 4.796044826507568, "identity": 0.5911346673965454, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6a77752bc6f75f5199bcbd1d1c8d825c5423bfcf857f756fd25416f6394cc345.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "6ac10db9a219e4922e8e231c62d563db8e8b67107f54a3ecea27b61ff7e797c2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "ba546fa0d13e854c55dae6753ffa9691d02bca6d9d2fae9255fe4f645d26fb2e", "generation_seconds_batch": 22.905993864987977, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8035616874694824, "target_z": 0.6484374574728383, "ce": 5.772419452667236, "genuineness": 4.582659721374512, "identity": 0.5216405391693115, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6ac10db9a219e4922e8e231c62d563db8e8b67107f54a3ecea27b61ff7e797c2.mp3" }, { "id": "heldout-fr-1", "language": "fr", "sentences": [ "Les clés sont sur la table de la cuisine.", "J’ai regardé dans le tiroir... puis près de la fenêtre.", "Prends-les quand tu partiras." ], "text": "Les clés sont sur la table de la cuisine. J’ai regardé dans le tiroir... puis près de la fenêtre. Prends-les quand tu partiras.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-1", "key": "6b451df0b4be21cae4dd2345bc7ce929980a2204c5d9382137332f58e7c9957d", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Prends-les quand tu partiras.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Natural clear speech.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Prends-les quand tu partiras.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "fd2baa64e7316419b355829ec0d506c0507bbe335e7ec25be18ba4aae8ca5ff6", "generation_seconds_batch": 16.741675611003302, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2344510555267334, "target_z": 0.9179687450482072, "ce": 5.200547218322754, "genuineness": 4.645420074462891, "identity": 0.7472876310348511, "hypothesis": "Les clés sont sur la table de la cuisine j'ai regardé dans le tiroir, puis, près de la fenêtre, prends-les quand tu partiras.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6b451df0b4be21cae4dd2345bc7ce929980a2204c5d9382137332f58e7c9957d.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "6b65f2ca73a133b65116e9a2d490eb7b6865f64dc543b47c804942a71cfd5026", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5509fac1736eb80373daa8cddad1eb5627a711be1bc2117f8886a0131c0f1cef", "generation_seconds_batch": 16.23243209591601, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6599318981170654, "target_z": 0.5585936949477154, "ce": 5.510311126708984, "genuineness": 4.68307638168335, "identity": 0.16055390238761902, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd, zeg maar, wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6b65f2ca73a133b65116e9a2d490eb7b6865f64dc543b47c804942a71cfd5026.mp3" }, { "id": "heldout-es-3", "language": "es", "sentences": [ "Alguien ha movido la pequeña silla azul.", "Estaba aquí... junto a la estantería alta.", "La colocaremos después de comer." ], "text": "Alguien ha movido la pequeña silla azul. Estaba aquí... junto a la estantería alta. La colocaremos después de comer.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-3", "key": "6b973c750e3616c70f28663c29496a0c74bea4a87dff4bad65cadc5394b99131", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La colocaremos después de comer.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La colocaremos después de comer.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "061ddfe4ba5754518d47cacdf9d66ddb4f05cc77d2ac4a7afc172319ec36994e", "generation_seconds_batch": 17.2577074309811, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.4180586338043213, "target_z": -0.8437499976697446, "ce": 6.23706579208374, "genuineness": 4.519899368286133, "identity": 0.6486825942993164, "hypothesis": "Alguien ha movido la pequeña silla azul. Estaba aquí, junto a la estantería alta. La colocaremos después de comer.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6b973c750e3616c70f28663c29496a0c74bea4a87dff4bad65cadc5394b99131.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "6c045414013ded1ac9e8a407633032953f169b9dbcab3f34b9492af226a5d1fe", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8eb2feae8993040303aa3fa69bd405add92f7e15d34bf75123e777d0d83f92eb", "generation_seconds_batch": 20.11818587302696, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.078331708908081, "target_z": 0.8203124749497541, "ce": 5.661221981048584, "genuineness": 4.413206577301025, "identity": 0.41422027349472046, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás lista.", "wer": 0.06666666666666667, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6c045414013ded1ac9e8a407633032953f169b9dbcab3f34b9492af226a5d1fe.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "6c858511a668cc14aed3f8b856be0717ca4f9fc6cb99dfc49c6da555c91f6b07", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "82dc7a974370e176c62fe8b9ad514d47fb9e07160b6bf35c41b0c9a96e3c34b3", "generation_seconds_batch": 23.349479842931032, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.503812789916992, "target_z": 0.46093757398560997, "ce": 5.831989288330078, "genuineness": 4.65797233581543, "identity": 0.4923419952392578, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6c858511a668cc14aed3f8b856be0717ca4f9fc6cb99dfc49c6da555c91f6b07.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "6ca905460f41dbc9db93a656791288d07ceba6b25e00ab02dfad2378a8f46047", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bb284e9ea0ce4284ad1262af400363d41564cb180f178b9fd6849096e052b7f9", "generation_seconds_batch": 22.663575985003263, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.795030951499939, "target_z": 0.01757810475590593, "ce": 6.125868320465088, "genuineness": 4.482243061065674, "identity": 0.5425034761428833, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6ca905460f41dbc9db93a656791288d07ceba6b25e00ab02dfad2378a8f46047.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "6ccba601fbe0b4d749eb4f56bb865114e02a3e0ee291a919094013831bcd87c0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "533df0b946aa362efb7f7c486a027967ea6a63a8f89f382f33b897579f1eabc5", "generation_seconds_batch": 22.475935191032477, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.253185272216797, "target_z": 0.9296874318400286, "ce": 5.589737892150879, "genuineness": 4.331617832183838, "identity": 0.4353153705596924, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6ccba601fbe0b4d749eb4f56bb865114e02a3e0ee291a919094013831bcd87c0.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "6d050dcacf6c71a8f7d65377574ea06601e7a985ac65b2f866201f2c52631ff2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e119ee62d74039c2743285a6656ae316a3609be1b07e5349c87f78e5c1fe5402", "generation_seconds_batch": 23.82154154102318, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6224632263183594, "target_z": 0.5351561722277249, "ce": 6.229123115539551, "genuineness": 4.883909225463867, "identity": 0.7610921859741211, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6d050dcacf6c71a8f7d65377574ea06601e7a985ac65b2f866201f2c52631ff2.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "6d0cfd896af60cd427bb4db826e3357b93f0743298ad05acde793a584469e573", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1b32dda7c17583a650b2f74a33cfe730c16ea7ce715e8dcd9d22aa2c8b872c6b", "generation_seconds_batch": 22.441613475908525, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.765256881713867, "target_z": 1.2500000932102173, "ce": 5.232317924499512, "genuineness": 4.65797233581543, "identity": 0.7377129197120667, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6d0cfd896af60cd427bb4db826e3357b93f0743298ad05acde793a584469e573.mp3" }, { "id": "heldout-fr-1", "language": "fr", "sentences": [ "Les clés sont sur la table de la cuisine.", "J’ai regardé dans le tiroir... puis près de la fenêtre.", "Prends-les quand tu partiras." ], "text": "Les clés sont sur la table de la cuisine. J’ai regardé dans le tiroir... puis près de la fenêtre. Prends-les quand tu partiras.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-1", "key": "6d48e98987786c063dcd89d84cf313bfe30cc4571609f3e4eb94bb5f8267e913", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Prends-les quand tu partiras.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Natural clear speech.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Prends-les quand tu partiras.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1973b9735569bf9599213097e011215819687527726081baa93284b2d97341b4", "generation_seconds_batch": 24.369751867023297, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3780808448791504, "target_z": 1.0078125075733302, "ce": 5.295859336853027, "genuineness": 4.432034492492676, "identity": 0.7371368408203125, "hypothesis": "Les clés sont sur la table de la cuisine J'ai regardé dans le tiroir, puis près de la fenêtre prends-les quand tu partiras.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6d48e98987786c063dcd89d84cf313bfe30cc4571609f3e4eb94bb5f8267e913.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "6dab31e302dcd8ee3d1dcf156edd472f908335d8801f20f37a9029439e73b7b0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "66c4ae823a2e805512008bafc40a50565b735de6806ee921ac2f9d6ec282bb64", "generation_seconds_batch": 22.92869902693201, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7402777671813965, "target_z": 1.234375078063557, "ce": 5.395142555236816, "genuineness": 4.532451152801514, "identity": 0.4956883490085602, "hypothesis": "He encontrado la cajita, estaba junto a la ventana vieja, podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6dab31e302dcd8ee3d1dcf156edd472f908335d8801f20f37a9029439e73b7b0.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "6e50595acd20b727a100105e5a09959a0cd83eed8e75b82a2c83c1919b9fd486", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8675ef4e149d9b44897a68408add8542a2adcb1f4421cd53e4204094edb258b8", "generation_seconds_batch": 17.039053219021298, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0331687927246094, "target_z": -0.4589843352400167, "ce": 6.133810997009277, "genuineness": 4.306513786315918, "identity": 0.3689424991607666, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6e50595acd20b727a100105e5a09959a0cd83eed8e75b82a2c83c1919b9fd486.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "6ec144e9d78905dd09a977b1818d40cda020a625bfba1a39a0dc6a991ab8f301", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6f27d2caf0df784afa94d1ebb731f2781353b334666eec91b4363a07acff98ef", "generation_seconds_batch": 21.962248128955252, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.716135025024414, "target_z": 0.593750053595875, "ce": 5.692992687225342, "genuineness": 4.796044826507568, "identity": 0.8500111699104309, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6ec144e9d78905dd09a977b1818d40cda020a625bfba1a39a0dc6a991ab8f301.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "6f1dd57907579924d698feb948185b99118a2de1c26b3197aba560b909ac71b8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4a0b4b203a58e01c37058a820b6fc8403bdd1ec403ebda1ac4fb89e1558f796b", "generation_seconds_batch": 17.09989327006042, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.147860527038574, "target_z": 0.23828118271387438, "ce": 6.252951145172119, "genuineness": 4.532451152801514, "identity": 0.23022505640983582, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6f1dd57907579924d698feb948185b99118a2de1c26b3197aba560b909ac71b8.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "6f49bc08a8804f4580e6f30022ebbb280a17dd9c00f68177566e1e6b77864e04", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "11af6bcd134aa5688f6850b0a7d8f20f3a0076f90fba5bbff7d9222c99261817", "generation_seconds_batch": 21.7837157540489, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.90119206905365, "target_z": 0.08398434456103841, "ce": 5.919358730316162, "genuineness": 4.7332844734191895, "identity": 0.6780257225036621, "hypothesis": "Die Tür ist noch offen, ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6f49bc08a8804f4580e6f30022ebbb280a17dd9c00f68177566e1e6b77864e04.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "6f9a7bb34d69e207308a8847e8a6f866439a5b83057cc3fe3acb304b5bdafb18", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "01f2135b9445e0432d0edeea15f1ddbb88d020da906c85d7aa9ac1309a577792", "generation_seconds_batch": 21.246781819034368, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0947799682617188, "target_z": 0.20507806281130814, "ce": 6.0146708488464355, "genuineness": 4.65797233581543, "identity": 0.739043653011322, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6f9a7bb34d69e207308a8847e8a6f866439a5b83057cc3fe3acb304b5bdafb18.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "6fc98d976be54af917dc3a8a63c37b11beeaef7a406546fe36932de316e55c3e", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c712e779cf4de3a038324bd1043fa1601d8885c64f599103989337504020176d", "generation_seconds_batch": 16.395632397034205, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0760457515716553, "target_z": 0.19335937601948675, "ce": 5.792276382446289, "genuineness": 4.545003414154053, "identity": 0.6900104284286499, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6fc98d976be54af917dc3a8a63c37b11beeaef7a406546fe36932de316e55c3e.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "6fde431775bbbf0deadf99a5e5bc92b72d5c41d3f2a8932c520746447979d64a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "705ba9a7a15dde56d2ea2b5046ee3f5277a2feeceee461018d24e404f918dfad", "generation_seconds_batch": 22.42237393092364, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9650893211364746, "target_z": 1.3749999161108044, "ce": 5.33954381942749, "genuineness": 4.532451152801514, "identity": 0.6386010646820068, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre, nous pouvons garder ensemble ce qu'elle contient.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/6fde431775bbbf0deadf99a5e5bc92b72d5c41d3f2a8932c520746447979d64a.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "70291f26097ea05481c4f8548d1424e7bc4b717dc7677031302db9090c5c5fa0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "dde987824523edbcb401f098e68804727197711412b8415c5b6ec6ba2b5c7757", "generation_seconds_batch": 21.130135593004525, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.853519916534424, "target_z": 0.679687487766159, "ce": 5.792276382446289, "genuineness": 4.231201171875, "identity": 0.5006662607192993, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/70291f26097ea05481c4f8548d1424e7bc4b717dc7677031302db9090c5c5fa0.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "708a04effe6d0ea8bd2af80476d4d81c9eb6e3830dade367b57d33fac22e03f7", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d338479e6e46c73f28f9943b1f6bea91a2e552962c2139ddcfbb99acb5ec5dfc", "generation_seconds_batch": 21.623700667987578, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.89098858833313, "target_z": 0.7031250104861494, "ce": 6.109982967376709, "genuineness": 4.632867813110352, "identity": 0.4674546420574188, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/708a04effe6d0ea8bd2af80476d4d81c9eb6e3830dade367b57d33fac22e03f7.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "7144c10f092f87149d9c93d475e728aca91e3bdb533eedb0e524ac56f5fd6b16", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c09d5acbb8b13525ae7b4ae2a1b5b3c47b29673806571a327f4391b3a3c3752e", "generation_seconds_batch": 21.703463268000633, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8980697393417358, "target_z": 0.0820312549517928, "ce": 5.847874641418457, "genuineness": 4.545003414154053, "identity": 0.45897889137268066, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7144c10f092f87149d9c93d475e728aca91e3bdb533eedb0e524ac56f5fd6b16.mp3" }, { "id": "heldout-fr-2", "language": "fr", "sentences": [ "La route tourne à gauche après le pont.", "Continue... jusqu’à la petite barrière en bois.", "Je t’attendrai de l’autre côté." ], "text": "La route tourne à gauche après le pont. Continue... jusqu’à la petite barrière en bois. Je t’attendrai de l’autre côté.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-2", "key": "71454f94fdba2de752c9f679b46c4280cd9b03b622e45dfc3da8f5ea7351e599", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Natural clear speech.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Natural clear speech.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7ef0f01104d0ff87733556a0201a2f0d8570ed697307bc26d8e87854cd5629f1", "generation_seconds_batch": 22.683973697945476, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5904030799865723, "target_z": 1.1406249871835952, "ce": 4.8488359451293945, "genuineness": 4.293961524963379, "identity": 0.6867527961730957, "hypothesis": "La route tourne à gauche après le pont, continue jusqu'à la petite barrière en bois, je t'attendrai de l'autre côté.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/71454f94fdba2de752c9f679b46c4280cd9b03b622e45dfc3da8f5ea7351e599.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "71533406dd7411ac8787e21c14468139837722a4b5d0297dce03c42d9ad30460", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "330c01c0b952794b1ad3fe5e660d575be6cc9580b869701728d30024f33bcaad", "generation_seconds_batch": 22.110985856037587, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7036454677581787, "target_z": 0.5859375460225448, "ce": 5.542081832885742, "genuineness": 4.68307638168335, "identity": 0.5173324346542358, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/71533406dd7411ac8787e21c14468139837722a4b5d0297dce03c42d9ad30460.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "719d0bf4290aaf8684db45e46dfc80c0c0edd16e7ae614bb94075603d70e4387", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "325936a6e83626e80bf85e805847156c04c4991e000d7918b4a5d5062115eaad", "generation_seconds_batch": 23.751068977988325, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4422013759613037, "target_z": -0.20312497320206252, "ce": 5.998785495758057, "genuineness": 4.808597087860107, "identity": 0.6731770038604736, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/719d0bf4290aaf8684db45e46dfc80c0c0edd16e7ae614bb94075603d70e4387.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "7214d38c04396713625c9385ddbbd52bf23914046a9f52841d967c0604480422", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] Het stond naast het oude raam.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "df7ba8ebb1105b475828a7005486f393f9cf621015f985155d0a3d2ed205e9c2", "generation_seconds_batch": 17.621264959918335, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0550254583358765, "target_z": -0.44531248427077585, "ce": 6.205295085906982, "genuineness": 4.457139015197754, "identity": 0.6539922952651978, "hypothesis": "Ik heb het kleine doosje gevonden, het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7214d38c04396713625c9385ddbbd52bf23914046a9f52841d967c0604480422.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "722152a6a0374a7aaf32adc4679732912dc95d4a2c0551e3f231ff95d2823503", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "87a8c4dbb21b1b1e3cab1923e3bb497b98bf10d96400dac2d67fa42652fe2864", "generation_seconds_batch": 22.06067972292658, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.328122615814209, "target_z": 0.9765624772800096, "ce": 5.736677646636963, "genuineness": 4.632867813110352, "identity": 0.3485603928565979, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/722152a6a0374a7aaf32adc4679732912dc95d4a2c0551e3f231ff95d2823503.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "72e5fcb3c37f22a4fe36ddb10dbb9587b30b4ff5c9a2f052feb2eb8c776de458", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5fc5459712eb8077ce4d163feaeaaf72e565d4c989936d91696905c8eb080cac", "generation_seconds_batch": 22.569874503999017, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.107269525527954, "target_z": 0.2128905703846383, "ce": 5.903473377227783, "genuineness": 4.582659721374512, "identity": 0.6175317764282227, "hypothesis": "Ich hab die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0625, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/72e5fcb3c37f22a4fe36ddb10dbb9587b30b4ff5c9a2f052feb2eb8c776de458.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "733bb1e00df3e3d26a6215a8b5add7f65f61038d8caf40e65a56e599349d2236", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e7a31eeac3b7b89ec2ddd76f59c6112157c8148e281bfef2ffcd906c1903b877", "generation_seconds_batch": 23.3151821839856, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1424522399902344, "target_z": -0.3906250058256386, "ce": 5.9431867599487305, "genuineness": 4.68307638168335, "identity": 0.707152247428894, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana. Vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/733bb1e00df3e3d26a6215a8b5add7f65f61038d8caf40e65a56e599349d2236.mp3" }, { "id": "heldout-en-0", "language": "en", "sentences": [ "Your message arrived after I had left.", "I read it twice... before putting the phone away.", "Let me know when we can talk." ], "text": "Your message arrived after I had left. I read it twice... before putting the phone away. Let me know when we can talk.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-0", "key": "734f10718f8aa881296d28735bf3c811794c9b322c3e8d97cd4554735458ba80", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Your message arrived after I had left.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Let me know when we can talk.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Your message arrived after I had left.\n(Natural clear speech.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Natural clear speech.) [3.2 seconds duration] Let me know when we can talk.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e7313e4e7d9882e10f3c3fa973d52c6fcb614e6bae841342559250cfc1973830", "generation_seconds_batch": 23.622372666024603, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.009639263153076, "target_z": 0.7773437578646121, "ce": 5.474569320678711, "genuineness": 4.419482707977295, "identity": 0.35898324847221375, "hypothesis": "Your message arrived after I had left. I read it twice before putting the phone away. Let me know when we can talk.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/734f10718f8aa881296d28735bf3c811794c9b322c3e8d97cd4554735458ba80.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "743cb1885bd7e5b44e69c03255c438d4633bcca6de0578eea7c7a78968bf1b3f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1afae84543310e344435ed0e4a0aa3e4c4da87133f6a28695eae42c2e0888cd6", "generation_seconds_batch": 21.703463268000633, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1228814125061035, "target_z": 0.22265616756721407, "ce": 6.022613525390625, "genuineness": 4.457139015197754, "identity": 0.5490058064460754, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/743cb1885bd7e5b44e69c03255c438d4633bcca6de0578eea7c7a78968bf1b3f.mp3" }, { "id": "heldout-de-4", "language": "de", "sentences": [ "Ich schrieb die Nummer auf den Umschlag.", "Lies sie genau... und prüfe die letzten beiden Ziffern.", "Ruf an, wenn du Zeit hast." ], "text": "Ich schrieb die Nummer auf den Umschlag. Lies sie genau... und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-4", "key": "74630cb58261b826e9e7150f2f95c7510b50ebd1cee7193087d2cea73892070c", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Natural clear speech.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Natural clear speech.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4acd049228bbfd49e84a7b5f98241814c7c6c42d830e996ca3996dde99e7a540", "generation_seconds_batch": 22.903160133981146, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6287081241607666, "target_z": 0.5390625005825639, "ce": 5.530168056488037, "genuineness": 4.519899368286133, "identity": 0.5951497554779053, "hypothesis": "Ich schrieb die Nummer auf den Umschlag, lies sie genau und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/74630cb58261b826e9e7150f2f95c7510b50ebd1cee7193087d2cea73892070c.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "7517b28f79242fa6e4d7d8c3481dc6e9efc74d1701d5a5d2151f159fe5d1a23c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3caf0b84d6253cd7ebc34bb49a20cdb994adb2b8b0492deb604078ac5f7e14c6", "generation_seconds_batch": 22.599728844012134, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8410303592681885, "target_z": 0.6718749801928289, "ce": 5.808161735534668, "genuineness": 4.3881025314331055, "identity": 0.6118732690811157, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7517b28f79242fa6e4d7d8c3481dc6e9efc74d1701d5a5d2151f159fe5d1a23c.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "757f32674e62dd594329ee2ae9f7ea41adb0cb9fc58b1efa83d12cd4fcda464a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c1a347c4d9891e810441799dbce0218fb302df1f4b6045102d61d5f860af90f8", "generation_seconds_batch": 23.036281498032622, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7544399499893188, "target_z": -0.007812507573330156, "ce": 5.982900142669678, "genuineness": 4.557555198669434, "identity": 0.41507530212402344, "hypothesis": "La puerta sigue abierta. He dejado tu abrijo sobre la silla. Dime cuándo estás listo.", "wer": 0.06666666666666667, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/757f32674e62dd594329ee2ae9f7ea41adb0cb9fc58b1efa83d12cd4fcda464a.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "75be18c59f044ced0e4c5db368461bbddebd87f9f5d53e0c0ae8e97b242c729b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5b5cd45e08a467dcd3dac8d38e647eda1f0a1d901ff8ae92332c22b1a9e9ad89", "generation_seconds_batch": 23.8592384129297, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7669295072555542, "target_z": 0.0, "ce": 6.229123115539551, "genuineness": 4.7458367347717285, "identity": 0.7100394368171692, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/75be18c59f044ced0e4c5db368461bbddebd87f9f5d53e0c0ae8e97b242c729b.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "76c0d0a2038de547f986f353f5a94a1dc30abc6a9be5d220992c0dce92dba898", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c8d4968c586f3007a367e4e0bc7225c6b99ee24a2f1726351346e7e09a3ea215", "generation_seconds_batch": 22.40528576693032, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.552934408187866, "target_z": 1.1171874644636046, "ce": 5.577824115753174, "genuineness": 4.7081804275512695, "identity": 0.47868800163269043, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/76c0d0a2038de547f986f353f5a94a1dc30abc6a9be5d220992c0dce92dba898.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "776e781f6714a45ad370e479ab6d1b49683e63cc0a5d470ffd726ef35df60c58", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "24d8c383f5320ff5c1a383be5db25f0a17a27b338e4d1e808781de9fd729958d", "generation_seconds_batch": 21.478914043982513, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.290654182434082, "target_z": 0.9531251036963667, "ce": 5.700935363769531, "genuineness": 4.720732688903809, "identity": 0.758009135723114, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/776e781f6714a45ad370e479ab6d1b49683e63cc0a5d470ffd726ef35df60c58.mp3" }, { "id": "heldout-fr-3", "language": "fr", "sentences": [ "Quelqu’un a déplacé la petite chaise bleue.", "Elle était ici... près de la grande bibliothèque.", "Nous la remettrons après le déjeuner." ], "text": "Quelqu’un a déplacé la petite chaise bleue. Elle était ici... près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-3", "key": "77962898fb95ddc0e92a153772191fe01482e40f98d2245158cb92652e8a93c1", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "11d5d5081019949554a831bf94b4da04421c4d602e322a49260773bb5128c492", "generation_seconds_batch": 17.045331481960602, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7490315437316895, "target_z": -0.636718770681017, "ce": 6.205295085906982, "genuineness": 4.557555198669434, "identity": 0.7910829782485962, "hypothesis": "Quelqu'un a déplacé la petite chaise bleue. Elle était ici, près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/77962898fb95ddc0e92a153772191fe01482e40f98d2245158cb92652e8a93c1.mp3" }, { "id": "heldout-de-4", "language": "de", "sentences": [ "Ich schrieb die Nummer auf den Umschlag.", "Lies sie genau... und prüfe die letzten beiden Ziffern.", "Ruf an, wenn du Zeit hast." ], "text": "Ich schrieb die Nummer auf den Umschlag. Lies sie genau... und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-4", "key": "77c49b56e5767f3b4b02b51c592f6306efecaaf1c2f0dbed4ce8bacc619fc072", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Natural clear speech.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Natural clear speech.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "26bed5cf4230da4605190921058be37ee7ba080d79ff035e8fa4518dacda84f2", "generation_seconds_batch": 16.56231646600645, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.709890127182007, "target_z": 0.589843725241036, "ce": 5.6254801750183105, "genuineness": 4.7081804275512695, "identity": 0.5176317691802979, "hypothesis": "Ich schrieb die Nummer auf den Umschlag. Lies sie genau und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/77c49b56e5767f3b4b02b51c592f6306efecaaf1c2f0dbed4ce8bacc619fc072.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "7818d59b808ee46e15e3f0c76239015ddefab23b4264d5548c62424032108789", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "a1c14e64081e8072164af7e70d111569b8930a3479cc9747653f21c8de22690b", "generation_seconds_batch": 21.984455551952124, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1720032691955566, "target_z": 0.8789062071815564, "ce": 5.847874641418457, "genuineness": 4.482243061065674, "identity": 0.5163328051567078, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7818d59b808ee46e15e3f0c76239015ddefab23b4264d5548c62424032108789.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "784c201579229a037ca49b2d3fb3d70b47464ff4a8dacdc58cbb9d7087631e05", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "35b8c7096c19ea78a2f60b438a3b535eb3ae4bee5124727ef10aafad5558a1ca", "generation_seconds_batch": 23.55445547506679, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9854965209960938, "target_z": 0.13671873339693005, "ce": 5.8875885009765625, "genuineness": 4.570107460021973, "identity": 0.47237396240234375, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/784c201579229a037ca49b2d3fb3d70b47464ff4a8dacdc58cbb9d7087631e05.mp3" }, { "id": "heldout-nl-4", "language": "nl", "sentences": [ "Ik schreef het nummer op de envelop.", "Lees het goed... en controleer de laatste twee cijfers.", "Bel me wanneer je even tijd hebt." ], "text": "Ik schreef het nummer op de envelop. Lees het goed... en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-4", "key": "788373552ad381e0af73fab0d5c3b63e9a0d481fa4e80a0dc9d7d0cb93a9f020", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Natural clear speech.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Natural clear speech.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "95f97660bf2aeb69b23b381fbb9863d469597c3f6bb217f6aeaf81e4ef9c6c84", "generation_seconds_batch": 23.107679182081483, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9596810340881348, "target_z": 0.7460937275712914, "ce": 5.4825119972229, "genuineness": 4.3127899169921875, "identity": 0.5031636953353882, "hypothesis": "Ik schreef het nummer op de envelop, lees het goed en controleer de laatste twee cijfers wel me wanneer je even tijd hebt.", "wer": 0.043478260869565216, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/788373552ad381e0af73fab0d5c3b63e9a0d481fa4e80a0dc9d7d0cb93a9f020.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "7888178243edf3318056e0ed62e6a9c9b3ca3df2e4ed25f02bf8fb60661eff57", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c78d5b6c2d5d74809c371740e2eee540c26ca46c82faa54eb8a0aae3b0609c29", "generation_seconds_batch": 17.71011035598349, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6919922828674316, "target_z": -0.046874970871807094, "ce": 6.1894097328186035, "genuineness": 4.68307638168335, "identity": 0.5238990187644958, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7888178243edf3318056e0ed62e6a9c9b3ca3df2e4ed25f02bf8fb60661eff57.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "7931d6b4529c89a83e6cecf403abf9057edf20b66a4b31bc56d38a254aee82fb", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3204ead13e187c490dc9639fcdf236d610322668abf53afcb5ceed829cb1a76c", "generation_seconds_batch": 15.87353016808629, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5983206033706665, "target_z": -0.10546877767178327, "ce": 6.229123115539551, "genuineness": 4.519899368286133, "identity": 0.5685691833496094, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7931d6b4529c89a83e6cecf403abf9057edf20b66a4b31bc56d38a254aee82fb.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "79499bbb7608cf18c74de58f8be0e1b9ee4e8e59bceeeb3c77f581071a6d1453", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] The door is still open.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] I put your coat on the chair.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6fef0f6772bfcd4c197a6dd8865bcf26cef26d7591510784df57a38ab13585e0", "generation_seconds_batch": 17.585974140907638, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.089371681213379, "target_z": -0.42382812572820483, "ce": 6.221180438995361, "genuineness": 3.9801597595214844, "identity": 0.33773136138916016, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/79499bbb7608cf18c74de58f8be0e1b9ee4e8e59bceeeb3c77f581071a6d1453.mp3" }, { "id": "heldout-nl-1", "language": "nl", "sentences": [ "De sleutels liggen op de keukentafel.", "Ik keek in de la... en daarna naast het raam.", "Neem ze mee als je weggaat." ], "text": "De sleutels liggen op de keukentafel. Ik keek in de la... en daarna naast het raam. Neem ze mee als je weggaat.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-1", "key": "794d3b9296e3c4be7c1e4182a2fa3cf095f7d593990763ce0f19c1f5e499a989", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Neem ze mee als je weggaat.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Neem ze mee als je weggaat.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "eb62b00cf5afc6d20293d4eb785422f962e31192dae85f3adf9240f9d5abb3cf", "generation_seconds_batch": 16.715598401031457, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.47426164150238037, "target_z": -0.8085937135897588, "ce": 6.3641486167907715, "genuineness": 4.432034492492676, "identity": 0.8421140313148499, "hypothesis": "De sleutels liggen op de keukentafel. Ik keek in de la en daarna naast het raam. Neem ze mee, als je weggaat.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/794d3b9296e3c4be7c1e4182a2fa3cf095f7d593990763ce0f19c1f5e499a989.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "79a7f3391132f663ea6f853c389562ec33fd130046404006efcf2c210a705cf0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "32e593223d021108182c118c6600bf06d1cc3b3a045884847984667a9b93328d", "generation_seconds_batch": 22.307957099052146, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.057311534881592, "target_z": 0.18164068922766535, "ce": 6.0305562019348145, "genuineness": 4.432034492492676, "identity": 0.5210627317428589, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/79a7f3391132f663ea6f853c389562ec33fd130046404006efcf2c210a705cf0.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "79eb7d71dc506e3e3ef765fe509784701fa0dc6ca3254f3afd14ba1957e5efb3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "22ea7e30d394704337d453cd52e7be21457a66bf6da4f64395186a2a2b7c3547", "generation_seconds_batch": 22.287599133094773, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.313347101211548, "target_z": 0.341796870776412, "ce": 5.8875885009765625, "genuineness": 4.7332844734191895, "identity": 0.5344675779342651, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/79eb7d71dc506e3e3ef765fe509784701fa0dc6ca3254f3afd14ba1957e5efb3.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "7a2d54dd0da440379d75f9c55231e4c5d4ef69df535d83701f13283406491297", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "07c00e3be7237614019a9bc961244eca2b927e14f1aae42ad15e4de829767f18", "generation_seconds_batch": 21.064869030029513, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.2259202003479004, "target_z": 0.28710931776310095, "ce": 6.133810997009277, "genuineness": 4.545003414154053, "identity": 0.8534280061721802, "hypothesis": "J'ai trouvé la petite boîte elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7a2d54dd0da440379d75f9c55231e4c5d4ef69df535d83701f13283406491297.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "7a7b9fe164f2ec58cd1637746152004aac24fb4102611a73d13065db8507f76a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "25af8ff75c1c877dac023bd35464034c38ac61215eff9bbcf25c6cfee1dbd37e", "generation_seconds_batch": 16.577306235092692, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.678666353225708, "target_z": 0.5703125308758845, "ce": 6.23706579208374, "genuineness": 4.356721878051758, "identity": 0.2830575406551361, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7a7b9fe164f2ec58cd1637746152004aac24fb4102611a73d13065db8507f76a.mp3" }, { "id": "heldout-en-2", "language": "en", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-2", "key": "7b3e2385dd54918c3ad4eeba02f905ef0166ecf8a5309628bcafb2e022cecfe2", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The road turns left behind the bridge.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The road turns left behind the bridge.\n(Natural clear speech.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Natural clear speech.) [3.2 seconds duration] I will meet you on the other side.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bc778524e8ca9d672edd0ec8dd3c49b5db86ef8af6519907051599905d1e4d3c", "generation_seconds_batch": 22.64156852499582, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.425753116607666, "target_z": 0.4121094389363834, "ce": 5.426913261413574, "genuineness": 4.15588903427124, "identity": 0.39909034967422485, "hypothesis": "The road turns left behind the bridge. I'll keep going until you reach the small wooden gate. I will meet you on the other side.", "wer": 0.041666666666666664, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7b3e2385dd54918c3ad4eeba02f905ef0166ecf8a5309628bcafb2e022cecfe2.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "7ba1263c879f8458b5b8a55f02be59c7ed7b7e5dbf824e97056ca875fa1f2ecf", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5dee843db5b084cdfa7e850f4fdbea09a512cbfddafcd2c9988ec7a47b0002a3", "generation_seconds_batch": 23.036281498032622, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9426195025444031, "target_z": -0.5156250151466603, "ce": 6.062326908111572, "genuineness": 4.256305694580078, "identity": 0.5747169852256775, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7ba1263c879f8458b5b8a55f02be59c7ed7b7e5dbf824e97056ca875fa1f2ecf.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "7ba20cc1b7ff043b3a0f4a064944c0311ac7116c051f930601be2291cf4d7654", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] The door is still open.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] I put your coat on the chair.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f7e95940edb8ea3c598e15cc1bbe44db59fdc8c70abb2b7ba06efdb6d4e62dd0", "generation_seconds_batch": 17.585974140907638, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.835621953010559, "target_z": 0.042968717085142015, "ce": 6.141753673553467, "genuineness": 4.061748504638672, "identity": 0.2900013029575348, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7ba20cc1b7ff043b3a0f4a064944c0311ac7116c051f930601be2291cf4d7654.mp3" }, { "id": "heldout-en-3", "language": "en", "sentences": [ "Someone moved the little blue chair.", "It used to stand here... beside the tall bookshelf.", "We can put it back after lunch." ], "text": "Someone moved the little blue chair. It used to stand here... beside the tall bookshelf. We can put it back after lunch.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-3", "key": "7baebead73a61d4d94a9fba3c1b686e162c72b2f7483ab23be44eba6eafb7e3d", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Someone moved the little blue chair.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can put it back after lunch.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Someone moved the little blue chair.\n(Natural clear speech.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Natural clear speech.) [3.2 seconds duration] We can put it back after lunch.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b87505e55110171d4f080117e44e82c60027afe812ffd01500dccf3df62d1f9b", "generation_seconds_batch": 16.58284433791414, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.153269052505493, "target_z": 0.867187520389735, "ce": 5.585766792297363, "genuineness": 4.394378662109375, "identity": 0.3787574768066406, "hypothesis": "Someone moved the little blue chair. It used to stand here beside the tall bookshelf. We can put it back after lunch.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7baebead73a61d4d94a9fba3c1b686e162c72b2f7483ab23be44eba6eafb7e3d.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "7c4c566f413ed5e3d94f9844854e9be20c8a40a4ba68a659df82909ab7097544", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5911543b1933998ed288b8a207e9b4aee7a751e8dca43e386b91bf17793a61fd", "generation_seconds_batch": 17.341312835924327, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8676822781562805, "target_z": -0.5624999860184674, "ce": 5.982900142669678, "genuineness": 4.444586753845215, "identity": 0.5072431564331055, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7c4c566f413ed5e3d94f9844854e9be20c8a40a4ba68a659df82909ab7097544.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "7c564492a0f0e33e4749e6999ef0c5cd09510406f57ce788da6c9664d3aa50b5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "28d22107da4532da5d280cf52a9ecc650f9c81946c3b52c60ea6266483db8b0e", "generation_seconds_batch": 23.36565126501955, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.310224771499634, "target_z": 0.3398437811671664, "ce": 6.141753673553467, "genuineness": 4.833701133728027, "identity": 0.7985033988952637, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7c564492a0f0e33e4749e6999ef0c5cd09510406f57ce788da6c9664d3aa50b5.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "7cb4f0e45a7fd8f3cf2257411f3a2d2ecf156fe200ce537c78cf790976b1f434", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "46d44d941024749d26443bb8a76117939098c7947749393801e8ba4c45685a6f", "generation_seconds_batch": 22.518019636976533, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.403059959411621, "target_z": 1.0234375227199906, "ce": 5.224375247955322, "genuineness": 4.231201171875, "identity": 0.20056763291358948, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7cb4f0e45a7fd8f3cf2257411f3a2d2ecf156fe200ce537c78cf790976b1f434.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "7d2a1a5d1995457657f0c63de28acd3d8ee4c98268fd22fc1b0aa69962d835eb", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9ed3199aa71ed1c58f655602fbd1cce0be0d386ab67f2085daf62cb3bf569041", "generation_seconds_batch": 23.100148959085345, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.709890127182007, "target_z": 0.589843725241036, "ce": 5.700935363769531, "genuineness": 4.557555198669434, "identity": 0.6273490786552429, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7d2a1a5d1995457657f0c63de28acd3d8ee4c98268fd22fc1b0aa69962d835eb.mp3" }, { "id": "heldout-es-2", "language": "es", "sentences": [ "La carretera gira a la izquierda tras el puente.", "Sigue... hasta la pequeña puerta de madera.", "Te esperaré al otro lado." ], "text": "La carretera gira a la izquierda tras el puente. Sigue... hasta la pequeña puerta de madera. Te esperaré al otro lado.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-2", "key": "7d63d25eca2ec951bceb213a0c86a46e8ad5f7ee565f35edd7ebdfb4d8eacb81", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Te esperaré al otro lado.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Natural clear speech.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Natural clear speech.) [3.2 seconds duration] Te esperaré al otro lado.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "f6bbc4552484af11344732519c6655cef3a55f3e5977f52a58c7a78e7a352c89", "generation_seconds_batch": 22.0864660179941, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0221288204193115, "target_z": 0.7851562654379423, "ce": 5.2045183181762695, "genuineness": 4.381826400756836, "identity": 0.5244097709655762, "hypothesis": "La carretera gira a la izquierda tras el puente. Sigue hasta la pequeña puerta de madera. Te esperaría al otro lado.", "wer": 0.047619047619047616, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7d63d25eca2ec951bceb213a0c86a46e8ad5f7ee565f35edd7ebdfb4d8eacb81.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "7d6691409b67fa2bbb511eb5428da19e019162e134f84288d63af34b5e6d7fd0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "13733587b7375aded90ebb97736456d4faa9c83a8954db32e7a1a191c06ab0db", "generation_seconds_batch": 23.711149926064536, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.2384097576141357, "target_z": 0.2949218253364311, "ce": 6.213237762451172, "genuineness": 4.833701133728027, "identity": 0.3210338056087494, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7d6691409b67fa2bbb511eb5428da19e019162e134f84288d63af34b5e6d7fd0.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "7d88678d812948683b0518aed5c41ce8dcad95624589da5b99410e12b602c963", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9e88806c151ba6cba95bc56b584c66ef78ffdc52394bbfdd89b635d9a136a7b3", "generation_seconds_batch": 22.1538234379841, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5475261211395264, "target_z": 0.4882812759240917, "ce": 5.982900142669678, "genuineness": 4.519899368286133, "identity": 0.735659122467041, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7d88678d812948683b0518aed5c41ce8dcad95624589da5b99410e12b602c963.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "7dd448db81c064eb0cb8010b8607a6b48ea5cc6d92bee40c1e6141541183b77c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "df4824e144a5e8a5d1cb804c602b6b18a8de93b28103ad6783f28161cec6b08a", "generation_seconds_batch": 21.756479145027697, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0595974922180176, "target_z": 0.8085937881579327, "ce": 5.4944257736206055, "genuineness": 4.959221839904785, "identity": 0.5131732821464539, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7dd448db81c064eb0cb8010b8607a6b48ea5cc6d92bee40c1e6141541183b77c.mp3" }, { "id": "heldout-de-4", "language": "de", "sentences": [ "Ich schrieb die Nummer auf den Umschlag.", "Lies sie genau... und prüfe die letzten beiden Ziffern.", "Ruf an, wenn du Zeit hast." ], "text": "Ich schrieb die Nummer auf den Umschlag. Lies sie genau... und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-4", "key": "7f2292ac0e8ffd0be7f16f0a7e0244a6fb6c56bd800a474a122b224088ea2e35", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Natural clear speech.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Natural clear speech.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6f37f38e2a29b695471dd8027bb4fc6931a054051270fb72e46d1cdafc912da6", "generation_seconds_batch": 17.173577815992758, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5256693363189697, "target_z": 0.474609350386677, "ce": 5.6334228515625, "genuineness": 4.6956281661987305, "identity": 0.6330827474594116, "hypothesis": "Ich schrieb die Nummer auf den Umschlag, lies sie genau und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7f2292ac0e8ffd0be7f16f0a7e0244a6fb6c56bd800a474a122b224088ea2e35.mp3" }, { "id": "heldout-en-1", "language": "en", "sentences": [ "The keys are on the kitchen table.", "I checked the drawer... then looked beside the window.", "Take them with you when you leave." ], "text": "The keys are on the kitchen table. I checked the drawer... then looked beside the window. Take them with you when you leave.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-1", "key": "7f6fce1a6f733719f45a363f28fee4e24b2f07a3396aa07b42392913d7c24239", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The keys are on the kitchen table.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Take them with you when you leave.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The keys are on the kitchen table.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Take them with you when you leave.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e12d55d5aac6afbff59d1481266179fd904e68e2e4ed7584a24d10c0ede73b38", "generation_seconds_batch": 17.89829301799182, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.5554436445236206, "target_z": -0.7578124889312867, "ce": 6.300607204437256, "genuineness": 4.149612903594971, "identity": 0.5426070094108582, "hypothesis": "The keys are on the kitchen table. I checked the drawer, then looked beside the window. Take them with you when you leave.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7f6fce1a6f733719f45a363f28fee4e24b2f07a3396aa07b42392913d7c24239.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "7f70647548f002b0774a3c4c77074a7b234ec9df86313f1961b381b3fd6d8f41", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "40376a64971381ff3ea6d98e8233eb945b20864306c982d2febbad5122604688", "generation_seconds_batch": 21.998502444010228, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7660932540893555, "target_z": 0.6250000838891956, "ce": 6.0305562019348145, "genuineness": 4.808597087860107, "identity": 0.792495608329773, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7f70647548f002b0774a3c4c77074a7b234ec9df86313f1961b381b3fd6d8f41.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "7f76634e992510e293f8899252bd30c01e1ae8b7f3b4884c8ef3c28d8c46e3eb", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "488d60c206c37929964ed895bf11c37e40662331240991aee82e7596736146fa", "generation_seconds_batch": 21.136349080945365, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.032332420349121, "target_z": 0.16601567408100504, "ce": 6.006728172302246, "genuineness": 3.986435890197754, "identity": 0.4738043248653412, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7f76634e992510e293f8899252bd30c01e1ae8b7f3b4884c8ef3c28d8c46e3eb.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "7fbd5ea1728a89c8d20f01bad478c4ca7510d71609dbf6a9981240b7f90ee573", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2baaacbb7aeaa57969a19f48511c9a9adc32d27937a1319de5d981770fbf2aa7", "generation_seconds_batch": 21.44586426403839, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0448219776153564, "target_z": 0.1738281816543352, "ce": 5.77639102935791, "genuineness": 4.720732688903809, "identity": 0.5403053760528564, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7fbd5ea1728a89c8d20f01bad478c4ca7510d71609dbf6a9981240b7f90ee573.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "7ff0a61f25d8241322479e0261657f862f85705a91267abf587e73a8b30f979b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "afd7077893900b844b2ca73164842a1edaed91b168fb9422894f78ecd5563597", "generation_seconds_batch": 16.524717071908526, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5849947929382324, "target_z": 0.5117187986440822, "ce": 6.006728172302246, "genuineness": 4.68307638168335, "identity": 0.48706871271133423, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/7ff0a61f25d8241322479e0261657f862f85705a91267abf587e73a8b30f979b.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "800eef4a9a829e5c17fb52a738f3d8dcc66ff5b9f8204ab6f4a4a88bd6582c8a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "31a740d94b98ecf789ae10dea5a56f020f8818dae4930ea72bd6043d8446b92a", "generation_seconds_batch": 22.92869902693201, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4155495166778564, "target_z": 1.0312500302933205, "ce": 5.248203277587891, "genuineness": 4.582659721374512, "identity": 0.42596155405044556, "hypothesis": "Ahí encontrado la cajita, estaba junto a la ventana vieja, podemos mirar juntos lo que hay dentro.", "wer": 0.058823529411764705, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/800eef4a9a829e5c17fb52a738f3d8dcc66ff5b9f8204ab6f4a4a88bd6582c8a.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "801787eec52d0cf5750352786c831370a0203b34b8630f67da1b3ddf36e1714e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "60a8fd81c47433e36f44fb4a3a784ce7ee07bc638b8c31f96ef7019a89bd1560", "generation_seconds_batch": 22.2676793289138, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6919922828674316, "target_z": -0.046874970871807094, "ce": 6.395919322967529, "genuineness": 4.595211505889893, "identity": 0.39778104424476624, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/801787eec52d0cf5750352786c831370a0203b34b8630f67da1b3ddf36e1714e.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "809acfc6a4b83b241f13b250d986723b6196d314beed3f365c6d3c43c9cb53a3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0fb1cf7a357b95f5bb586381a531b909ec570118d66e1b6bc7f4e94ec75b4c88", "generation_seconds_batch": 22.169052313896827, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7286243438720703, "target_z": 0.6015624120328574, "ce": 5.4666266441345215, "genuineness": 4.632867813110352, "identity": 0.5222261548042297, "hypothesis": "la puerta sigue abierta. he dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/809acfc6a4b83b241f13b250d986723b6196d314beed3f365c6d3c43c9cb53a3.mp3" }, { "id": "heldout-es-1", "language": "es", "sentences": [ "Las llaves están en la mesa de la cocina.", "Miré en el cajón... y luego junto a la ventana.", "Llévatelas cuando te vayas." ], "text": "Las llaves están en la mesa de la cocina. Miré en el cajón... y luego junto a la ventana. Llévatelas cuando te vayas.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-1", "key": "80c2597a6b8f43b5338a6b37d654868d679f46ded48514b42837ab18fba327ad", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Llévatelas cuando te vayas.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Llévatelas cuando te vayas.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "772aa68d0cd51462c727093f51bc5b4489ab8069baa6f01d56bedf339e0d18eb", "generation_seconds_batch": 17.306735864956863, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.5554436445236206, "target_z": -0.7578124889312867, "ce": 6.173524379730225, "genuineness": 4.030367851257324, "identity": 0.6309086680412292, "hypothesis": "Las llaves están en la mesa de la cocina. Miré en el cajón y luego, junto a la ventana, llévatelas cuando te vayas.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/80c2597a6b8f43b5338a6b37d654868d679f46ded48514b42837ab18fba327ad.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "812c5b949a0de8459c6eff4c836ae80b0a059234c47ce0cc15175c30a7f6044e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "04e63b94623c0ca75b3372b46eb611a0fc26b925469daf9c30bfddb3974ea3eb", "generation_seconds_batch": 23.82154154102318, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9417831897735596, "target_z": 0.10937503145844833, "ce": 5.911416053771973, "genuineness": 4.482243061065674, "identity": 0.694332480430603, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/812c5b949a0de8459c6eff4c836ae80b0a059234c47ce0cc15175c30a7f6044e.mp3" }, { "id": "heldout-de-0", "language": "de", "sentences": [ "Deine Nachricht kam, als ich schon weg war.", "Ich las sie zweimal... dann legte ich das Handy weg.", "Sag mir, wann wir reden können." ], "text": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal... dann legte ich das Handy weg. Sag mir, wann wir reden können.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-0", "key": "81882f86d35b4c759b8f50da602c9f45d52bf6cad5494ccb3f67bf311378bc75", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wann wir reden können.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Natural clear speech.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wann wir reden können.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2f72d20026bcbe4f9111702b3d81ef3bef345a9d8c545022e833f4ba162e9016", "generation_seconds_batch": 17.274377543013543, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.109555721282959, "target_z": 0.8398438184512533, "ce": 5.343515396118164, "genuineness": 4.432034492492676, "identity": 0.45427075028419495, "hypothesis": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal. Dann legte ich das Handy weg. Sag mir, wann wir reden können.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/81882f86d35b4c759b8f50da602c9f45d52bf6cad5494ccb3f67bf311378bc75.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "81caf973d4f14f9f31dd47eb58daa9a1c313afeacc5ee5856a247ca7ef8db979", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "41c2eba4f935ad8fa0e592dfec4a63f169528f1189489f2af333392d58da1092", "generation_seconds_batch": 23.349479842931032, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7036454677581787, "target_z": 0.5859375460225448, "ce": 5.692992687225342, "genuineness": 4.595211505889893, "identity": 0.48267126083374023, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/81caf973d4f14f9f31dd47eb58daa9a1c313afeacc5ee5856a247ca7ef8db979.mp3" }, { "id": "heldout-fr-1", "language": "fr", "sentences": [ "Les clés sont sur la table de la cuisine.", "J’ai regardé dans le tiroir... puis près de la fenêtre.", "Prends-les quand tu partiras." ], "text": "Les clés sont sur la table de la cuisine. J’ai regardé dans le tiroir... puis près de la fenêtre. Prends-les quand tu partiras.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-1", "key": "82e2e4c54cbda320e42cf7559854e0d8a5fd5011479bf2d0cfda5d608e8c8999", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Prends-les quand tu partiras.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Natural clear speech.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Prends-les quand tu partiras.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "050aae402b19a5771f57c129f66b1f4df1e40c681436690f7d2f305333441331", "generation_seconds_batch": 16.561459264950827, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3406121730804443, "target_z": 0.9843749848533397, "ce": 5.168776512145996, "genuineness": 4.519899368286133, "identity": 0.645530641078949, "hypothesis": "Les galets sont sur la table de la cuisine j'ai regardé dans le tiroir, puis pris la fenêtre prends-les quand tu partiras.", "wer": 0.125, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/82e2e4c54cbda320e42cf7559854e0d8a5fd5011479bf2d0cfda5d608e8c8999.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "8302b3f350fe4c56c8e4b3ace1e8c20d84ef6201542aa07f29678a050272c700", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8afcbe4bd75a9f77f9e8c581708bf560697ee7e676f3850b69bbc3042b799eb6", "generation_seconds_batch": 16.707319209934212, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6403613090515137, "target_z": 1.1718750174769157, "ce": 5.434855937957764, "genuineness": 4.519899368286133, "identity": 0.7764999270439148, "hypothesis": "Je trouvais la petite boîte elle était près de la vieille fenêtre nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.10526315789473684, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8302b3f350fe4c56c8e4b3ace1e8c20d84ef6201542aa07f29678a050272c700.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "834d015a6e3e7ff423e69928b6c6ca962c8a375c54ba7c73afe037b5e2924f5a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7642a815c54dfc83bb5d3486a2be542addb581b35a9e146fb365f90baf51a818", "generation_seconds_batch": 22.178197700995952, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.460099220275879, "target_z": 0.4335937229107806, "ce": 5.669164657592773, "genuineness": 4.6956281661987305, "identity": 0.6604691743850708, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/834d015a6e3e7ff423e69928b6c6ca962c8a375c54ba7c73afe037b5e2924f5a.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "8365bae5d0afaabf7cf2d14ad4635aba0cad79a7d4595cd176e5a0e58c927750", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "abed10adfc04c4702cf148cab3776b506d064b815680226ab865607ad583c4cf", "generation_seconds_batch": 21.233830463024788, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2344510555267334, "target_z": 0.9179687450482072, "ce": 5.6413655281066895, "genuineness": 4.758388519287109, "identity": 0.7760450839996338, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8365bae5d0afaabf7cf2d14ad4635aba0cad79a7d4595cd176e5a0e58c927750.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "83d40fb049ab943a4359c695dc0bc23963d250ef8cea9f524011cd4fc57e2ff0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "96fe088d8ab81e8702dfca2e7d60175647d720e0560cd8d76e89d5890d35f2f4", "generation_seconds_batch": 20.864102661958896, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8668458461761475, "target_z": 0.0624999860184674, "ce": 5.7684478759765625, "genuineness": 4.419482707977295, "identity": 0.5818944573402405, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/83d40fb049ab943a4359c695dc0bc23963d250ef8cea9f524011cd4fc57e2ff0.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "842455def6df2b775927d88e0d86c6414375e117211f51cd342e0e0c0656db02", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "475bedf2b951fd2921eee9bb1d13aa6257cc1cb56e208a8f42c6708736ecd992", "generation_seconds_batch": 17.277530936989933, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9730069637298584, "target_z": 0.1289062258235999, "ce": 6.23706579208374, "genuineness": 4.482243061065674, "identity": 0.42063063383102417, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/842455def6df2b775927d88e0d86c6414375e117211f51cd342e0e0c0656db02.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "84892151ecef9bf78fc113abe34447a9378e063608c9e31139b40e79b719b88e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "46f8ebed170e0beb02f10243f7d64a213b744dc2563ccdd8c84e83b03c2926b9", "generation_seconds_batch": 21.43358326505404, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.827704429626465, "target_z": 1.2890624819405203, "ce": 5.780362129211426, "genuineness": 4.846253395080566, "identity": 0.8537044525146484, "hypothesis": "La portée t'es encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.1111111111111111, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/84892151ecef9bf78fc113abe34447a9378e063608c9e31139b40e79b719b88e.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "8593ce86cf9b53191edd6575feda16fd23e247dff2e39aeb2df5fe539261adb1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d8a3cf9544e685fa82d45fa046502a6a953ae716fe38d217a76895afe8b96cf7", "generation_seconds_batch": 22.557522826013155, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7927449941635132, "target_z": -0.6093749941743615, "ce": 6.213237762451172, "genuineness": 4.65797233581543, "identity": 0.5982300043106079, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8593ce86cf9b53191edd6575feda16fd23e247dff2e39aeb2df5fe539261adb1.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "85a233f61f326f73ee4f2f3c2f05a4a3657843b7184de3fe6df781e05ceba0f6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "b9530578cd1255c59b7e75ae75172ea99a67e9ffd688377951e3f5ab24984c8b", "generation_seconds_batch": 21.99430741602555, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3905704021453857, "target_z": 1.0156250151466604, "ce": 5.669164657592773, "genuineness": 4.545003414154053, "identity": 0.4498240351676941, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/85a233f61f326f73ee4f2f3c2f05a4a3657843b7184de3fe6df781e05ceba0f6.mp3" }, { "id": "heldout-es-0", "language": "es", "sentences": [ "Tu mensaje llegó después de que me fuera.", "Lo leí dos veces... y guardé el teléfono.", "Dime cuándo podemos hablar." ], "text": "Tu mensaje llegó después de que me fuera. Lo leí dos veces... y guardé el teléfono. Dime cuándo podemos hablar.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-0", "key": "85e83afd490ec708a075cb569305eeef9d19c19b762bef5b4d70934c5db5fda6", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo podemos hablar.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Natural clear speech.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo podemos hablar.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "91da1a654fbf12cab35161759aeed09a62b131f41b3df9e8cf8cd4d18e0c8533", "generation_seconds_batch": 18.141905845026486, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1720032691955566, "target_z": 0.8789062071815564, "ce": 5.236289024353027, "genuineness": 4.419482707977295, "identity": 0.6466671228408813, "hypothesis": "Tu mensaje llegó después de que me fuera. Lo leí dos veces y guardé el teléfono. Dime cuándo podemos hablar.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/85e83afd490ec708a075cb569305eeef9d19c19b762bef5b4d70934c5db5fda6.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "85f575b99c30e8b3cece5109da7465ad56dea125b7b5350982a16d7e0fc7e9bc", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f247e05721849d8b0b1688a5dddc8c4033417cbe24812ee6f72008d399f746a4", "generation_seconds_batch": 22.335953415953554, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.365591287612915, "target_z": 1.0, "ce": 5.7406487464904785, "genuineness": 4.607763767242432, "identity": 0.5734754204750061, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/85f575b99c30e8b3cece5109da7465ad56dea125b7b5350982a16d7e0fc7e9bc.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "8655066511886360e2fb90cffaf4c6b2ac6cc4239430fa30aee84235d2457cfa", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a98ec9f16a62da51ee626d59285a722508b5cd5eebf18d9c1f26975ba5516127", "generation_seconds_batch": 15.77842448791489, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9551090598106384, "target_z": -0.5078125075733302, "ce": 6.300607204437256, "genuineness": 4.6956281661987305, "identity": 0.6755889058113098, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8655066511886360e2fb90cffaf4c6b2ac6cc4239430fa30aee84235d2457cfa.mp3" }, { "id": "heldout-nl-2", "language": "nl", "sentences": [ "De weg gaat linksaf na de brug.", "Loop verder... tot aan het kleine houten hek.", "Ik wacht aan de andere kant." ], "text": "De weg gaat linksaf na de brug. Loop verder... tot aan het kleine houten hek. Ik wacht aan de andere kant.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-2", "key": "8661ec363f62df52e352a7a3c1df24a64a7b95319120347abc162b4a7579dde4", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ik wacht aan de andere kant.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Natural clear speech.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Natural clear speech.) [3.2 seconds duration] Ik wacht aan de andere kant.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e805a7ef79f3cc2b9eb7745f114e6c00184b1fe84ced7886c4544896d0b34624", "generation_seconds_batch": 22.34943998802919, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.184492826461792, "target_z": 0.8867187147548866, "ce": 5.669164657592773, "genuineness": 4.193545341491699, "identity": 0.0008642598986625671, "hypothesis": "De weg gaat linksaf naar de brug lok verder tot aan het kleine houten Hek. Ik wacht aan de andere kant.", "wer": 0.09523809523809523, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8661ec363f62df52e352a7a3c1df24a64a7b95319120347abc162b4a7579dde4.mp3" }, { "id": "heldout-de-0", "language": "de", "sentences": [ "Deine Nachricht kam, als ich schon weg war.", "Ich las sie zweimal... dann legte ich das Handy weg.", "Sag mir, wann wir reden können." ], "text": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal... dann legte ich das Handy weg. Sag mir, wann wir reden können.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-0", "key": "86a41e67c97a85d99b7c953aa38a9936ede7109a2d96558f4fe6695556caa26d", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wann wir reden können.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Natural clear speech.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wann wir reden können.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "293b6e2d5c252590cfedbf72dd0833884a52f59d7b2a4c60d23071f6d8e50018", "generation_seconds_batch": 21.96419238904491, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2656750679016113, "target_z": 0.9375000885497065, "ce": 5.285930633544922, "genuineness": 4.457139015197754, "identity": 0.5904591679573059, "hypothesis": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal. Dann legte ich das Handy weg. Sag mir, wann wir reden können.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/86a41e67c97a85d99b7c953aa38a9936ede7109a2d96558f4fe6695556caa26d.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "86d7877e475200f5ee6324c59dc383eb0ac41bcd9e137b1b79c425ed9bc461c3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0cbf288b9fe2e709c587b0787268052706f8c76da24e369e911ed4c7150e30e3", "generation_seconds_batch": 22.600584128056653, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7028088569641113, "target_z": 1.210937406207219, "ce": 5.200547218322754, "genuineness": 4.557555198669434, "identity": 0.6335599422454834, "hypothesis": "La porte la porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.1111111111111111, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/86d7877e475200f5ee6324c59dc383eb0ac41bcd9e137b1b79c425ed9bc461c3.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "87bf2412c48fe9ef55cc6df7ff642e39e815dc22795509cc78deaf2a2281bf16", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f442277ef8173e9a5ebc1a8d61a3f53436e16e253efe016863baf79d4bfe412d", "generation_seconds_batch": 22.51380237203557, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.284409284591675, "target_z": 0.9492187753415279, "ce": 5.919358730316162, "genuineness": 4.545003414154053, "identity": 0.42858651280403137, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/87bf2412c48fe9ef55cc6df7ff642e39e815dc22795509cc78deaf2a2281bf16.mp3" }, { "id": "heldout-es-1", "language": "es", "sentences": [ "Las llaves están en la mesa de la cocina.", "Miré en el cajón... y luego junto a la ventana.", "Llévatelas cuando te vayas." ], "text": "Las llaves están en la mesa de la cocina. Miré en el cajón... y luego junto a la ventana. Llévatelas cuando te vayas.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-1", "key": "87cdd726fa8461670d9bd890269f5dc3c65673718ebbfac11f92a9ffc2b3c9db", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Llévatelas cuando te vayas.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Natural clear speech.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Natural clear speech.) [3.2 seconds duration] Llévatelas cuando te vayas.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "3e08c0c88bf216a7d8e8d63776bd470dbcb585a289b63af4cae76827d7085f56", "generation_seconds_batch": 22.138351614936255, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9846601486206055, "target_z": 0.7617187427179518, "ce": 5.395142555236816, "genuineness": 4.356721878051758, "identity": 0.461786687374115, "hypothesis": "Las llaves están en la mesa de la cocina, miré en el cajón y luego junto a la ventana, llévatelas cuando te vayas.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/87cdd726fa8461670d9bd890269f5dc3c65673718ebbfac11f92a9ffc2b3c9db.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "88010e41f31c90e23f5c26fcd8e0574b72293991189476fc9168b0118121bb16", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d172720e61564f6cb247205cfe73a8e742fdcf267134f260425642ba29482643", "generation_seconds_batch": 22.375198861933313, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4453237056732178, "target_z": -0.2011718835928169, "ce": 5.9431867599487305, "genuineness": 4.720732688903809, "identity": 0.6770074963569641, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/88010e41f31c90e23f5c26fcd8e0574b72293991189476fc9168b0118121bb16.mp3" }, { "id": "heldout-nl-2", "language": "nl", "sentences": [ "De weg gaat linksaf na de brug.", "Loop verder... tot aan het kleine houten hek.", "Ik wacht aan de andere kant." ], "text": "De weg gaat linksaf na de brug. Loop verder... tot aan het kleine houten hek. Ik wacht aan de andere kant.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-2", "key": "88428710e639a2220e6f89a4b983f4febd332a1f617184f6ce0f645d11306c75", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ik wacht aan de andere kant.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Natural clear speech.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Natural clear speech.) [3.2 seconds duration] Ik wacht aan de andere kant.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "fcd272560dad8b73b9e3f5b3b0696b275eb39b68134373d31c970cee24daad65", "generation_seconds_batch": 22.34943998802919, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.078331708908081, "target_z": 0.8203124749497541, "ce": 5.186647415161133, "genuineness": 4.356721878051758, "identity": 0.13803380727767944, "hypothesis": "De weg gaat linksaf naar de brug. Loop verder tot aan het kleine houten. Ik wacht aan de andere kant.", "wer": 0.09523809523809523, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/88428710e639a2220e6f89a4b983f4febd332a1f617184f6ce0f645d11306c75.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "88d6718c3ab8cde08bd3392933b1061088075ee4144def7b4b2d13c91cadd322", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] Die Tür ist noch offen.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b3cf1f9da7cbd183d61ea90f7e0508aa0ad5b988675a5fbf76565aa8d4c9f3ea", "generation_seconds_batch": 16.016175522003323, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.3493661880493164, "target_z": -0.8867187147548866, "ce": 6.332377910614014, "genuineness": 4.582659721374512, "identity": 0.5993884801864624, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du denn seid, Jumatte, das brasst.", "wer": 0.2631578947368421, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/88d6718c3ab8cde08bd3392933b1061088075ee4144def7b4b2d13c91cadd322.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "88fad093ce0ee2350fbc03464a196ae10ec873bcaf21e5ecf86b7ef993ac0fe0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "526c75f58bd65d5e971a1bdb54ed5820248c7cb233b8c9749890625601f9c017", "generation_seconds_batch": 16.262230332009494, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.359346389770508, "target_z": 0.996093671645161, "ce": 5.601652145385742, "genuineness": 4.833701133728027, "identity": 0.8086378574371338, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre, nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/88fad093ce0ee2350fbc03464a196ae10ec873bcaf21e5ecf86b7ef993ac0fe0.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "89d783f29f0687262d2010bca88a5e250ff2735b511e9a70073adcf9644d9560", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f604f171d79e1a174ff58f3562190a6c9b71d7a9a84dd6c78c33ef02358c4a63", "generation_seconds_batch": 17.18484212597832, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.109555721282959, "target_z": 0.8398438184512533, "ce": 5.927301406860352, "genuineness": 4.519899368286133, "identity": 0.6196876168251038, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/89d783f29f0687262d2010bca88a5e250ff2735b511e9a70073adcf9644d9560.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "8a7349eb4fbd8dc3838c1ecc1618645229d89d8ed64e3c42874db1b902206ae3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4a3cd03a343b270bae79584c22f528668745060a3dc1a389dd47e9098244a622", "generation_seconds_batch": 23.58572617196478, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7411141395568848, "target_z": 0.6093750687425352, "ce": 5.605623245239258, "genuineness": 4.595211505889893, "identity": 0.5865250825881958, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8a7349eb4fbd8dc3838c1ecc1618645229d89d8ed64e3c42874db1b902206ae3.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "8aa97b62201c5b3f66d13717e8862bddb653e6660e873205b9eb5b4b517202b1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f43052587e6b97dcd10487ba4cd72c24edf71ca835bbae44171695f05d5616bd", "generation_seconds_batch": 21.32317853393033, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5015267133712769, "target_z": -0.1660155995128312, "ce": 5.871703147888184, "genuineness": 4.7458367347717285, "identity": 0.5049824714660645, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8aa97b62201c5b3f66d13717e8862bddb653e6660e873205b9eb5b4b517202b1.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "8b82c724126323a0edd9e10702d236c7502825b719a41216d43b0d2f40d7fe6f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ba23f93d5886cee0cb88886f14df41dc624bb73f808e11ca6cc62233347b1161", "generation_seconds_batch": 22.499105387018062, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.711562991142273, "target_z": -0.6601562188328336, "ce": 6.38003396987915, "genuineness": 4.406930446624756, "identity": 0.3005557060241699, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8b82c724126323a0edd9e10702d236c7502825b719a41216d43b0d2f40d7fe6f.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "8bb415db6c6fc74090bde30369b282b6e7cacc4dc1aa84395fc79adc508df947", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d0e07b6ecbef27fee4c2a72d033e14d5b286bb73d0369f2655b0aec1696d4f07", "generation_seconds_batch": 23.730676439939998, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0073533058166504, "target_z": 0.15039065893434472, "ce": 6.23706579208374, "genuineness": 4.344170093536377, "identity": 0.2578161358833313, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8bb415db6c6fc74090bde30369b282b6e7cacc4dc1aa84395fc79adc508df947.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "8c170eb42d18deb2243937d600c563746f3fab5f164b5632b3e4a37fe7415292", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "b1691a23424a252a4ffcf8dd2a4ba502ede333df1fc8f8634ecfa3fac79944df", "generation_seconds_batch": 21.991056522005238, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.897233486175537, "target_z": 0.7070313388409883, "ce": 5.7128496170043945, "genuineness": 4.331617832183838, "identity": 0.3788914084434509, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8c170eb42d18deb2243937d600c563746f3fab5f164b5632b3e4a37fe7415292.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "8c5d3b560fcea59cfa858066c6c8fe2b85aa07cc8e6883e1b53ce5e003644ed1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8d0d158615a5c460b0b94652ffa2f3668030821f86f85cbe63348688b8cd61f5", "generation_seconds_batch": 22.375198861933313, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1955327987670898, "target_z": -0.35742188592307234, "ce": 6.038498878479004, "genuineness": 4.758388519287109, "identity": 0.7637258172035217, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8c5d3b560fcea59cfa858066c6c8fe2b85aa07cc8e6883e1b53ce5e003644ed1.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "8c5d9a56ef34db81c2b7b5916aab166b45282ac9748874c27583be77ddc75398", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8aeb71811460bb2386a27effb78ff8b3f38c879158555334753f4e367a38ac85", "generation_seconds_batch": 23.100321167963557, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7513176202774048, "target_z": -0.009765597182575775, "ce": 6.292664527893066, "genuineness": 4.7458367347717285, "identity": 0.5500922799110413, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8c5d9a56ef34db81c2b7b5916aab166b45282ac9748874c27583be77ddc75398.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "8c6e86c7857382a230949697aebe40f383a20db55fd58f099df684899a7cbbab", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e35915b8a918d459d1aee70010ae725333a567f48f55e3e80309c37f67313a3d", "generation_seconds_batch": 24.013636282994412, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.318142294883728, "target_z": -0.906249983688212, "ce": 6.141753673553467, "genuineness": 4.532451152801514, "identity": 0.23685483634471893, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8c6e86c7857382a230949697aebe40f383a20db55fd58f099df684899a7cbbab.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "8c79145a676819a89ba34f786719012079987746a552dc9aadb872224365cc3d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "22f2b9f20b081255acecef54e4a2f430ebaf274e66f3ba21e057b23d97b50190", "generation_seconds_batch": 21.323322594049387, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9230488538742065, "target_z": 0.09765627009845311, "ce": 5.903473377227783, "genuineness": 4.758388519287109, "identity": 0.5182409882545471, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8c79145a676819a89ba34f786719012079987746a552dc9aadb872224365cc3d.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "8cc2e73b7d1abfcacec4da627f05097bfd22b9efd1000d59f8368fdfb03dd245", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7576051c7759ce1e744c4604afd63546ce04bd4e27896bd454f1c39e971f6fa4", "generation_seconds_batch": 22.50890007801354, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7294608354568481, "target_z": -0.02343752271999047, "ce": 5.895530700683594, "genuineness": 4.720732688903809, "identity": 0.5523383021354675, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8cc2e73b7d1abfcacec4da627f05097bfd22b9efd1000d59f8368fdfb03dd245.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "8d5aa38d30987b2cea7342897255a5678f80eedb1d50c716e85ffe41b48262a8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "69ccdcaa136209342c131011ae901040f23936cce9aadae55206281606bddf54", "generation_seconds_batch": 24.161738647031598, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2236342430114746, "target_z": -0.3398437811671664, "ce": 6.24500846862793, "genuineness": 4.7332844734191895, "identity": 0.5827820301055908, "hypothesis": "Die Tür ist noch offen. Ich hab deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8d5aa38d30987b2cea7342897255a5678f80eedb1d50c716e85ffe41b48262a8.mp3" }, { "id": "heldout-en-1", "language": "en", "sentences": [ "The keys are on the kitchen table.", "I checked the drawer... then looked beside the window.", "Take them with you when you leave." ], "text": "The keys are on the kitchen table. I checked the drawer... then looked beside the window. Take them with you when you leave.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-1", "key": "8d7e322c4a331a726f20f6cc40c0ddb2b8f3b80a47a29a80da872281ea78ac4b", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The keys are on the kitchen table.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Take them with you when you leave.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The keys are on the kitchen table.\n(Natural clear speech.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Natural clear speech.) [3.2 seconds duration] Take them with you when you leave.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "08115e36b6dd72c1b30511310c0586e2702e123db142de8a02a0d9aef0110171", "generation_seconds_batch": 17.132835216005333, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.922212600708008, "target_z": 0.7226563539876487, "ce": 5.131048679351807, "genuineness": 4.243753433227539, "identity": 0.24731463193893433, "hypothesis": "The keys are on the kitchen table. I checked the drawer, then looked beside the window. Take them with you when you leave.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8d7e322c4a331a726f20f6cc40c0ddb2b8f3b80a47a29a80da872281ea78ac4b.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "8f2bc4e99db966821a15bc476bba52ff83aabea18f45877012f4b45f034c1962", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c8fe569e367ea89d7150ec978dd81032da654d805a062985e3b9f731a7b4fb97", "generation_seconds_batch": 22.31468472396955, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7794190645217896, "target_z": 0.007812507573330156, "ce": 5.990842819213867, "genuineness": 4.507347106933594, "identity": 0.3679039180278778, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8f2bc4e99db966821a15bc476bba52ff83aabea18f45877012f4b45f034c1962.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "8f3ea13d01dfc06f6dc64834f2c9514ac9cf4ae262beb27a0ffc27b4fb74001d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "8c88e40b1776a36e1d1086bf3f958ac20f71ed320111171d6dd2d066278abd95", "generation_seconds_batch": 16.637231742963195, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.385161876678467, "target_z": 0.3867186774707997, "ce": 5.9431867599487305, "genuineness": 4.21237325668335, "identity": 0.4284359812736511, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8f3ea13d01dfc06f6dc64834f2c9514ac9cf4ae262beb27a0ffc27b4fb74001d.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "8fd884dccfc1825a4319f8b5fb1132810367c69c7ce008b276fbabbb542ec52a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a717fabd11d810f8c3e64538a8a5a5a7ac8b8095fda438616c5d8222f5671f60", "generation_seconds_batch": 16.31968818604946, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6536872386932373, "target_z": 0.5546875157292241, "ce": 5.824047088623047, "genuineness": 4.557555198669434, "identity": 0.584209680557251, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/8fd884dccfc1825a4319f8b5fb1132810367c69c7ce008b276fbabbb542ec52a.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "9082dba13c3368d5b9f18a758dfb5c84ffeaa78096788174779a8fecab533942", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1d2dcb022c571037aea24301fa76a7a4e0d5c0569862e4952b4f7defa2d3991e", "generation_seconds_batch": 22.796636154991575, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.3891208171844482, "target_z": -0.23632809310462877, "ce": 6.348263263702393, "genuineness": 4.65797233581543, "identity": 0.6202579140663147, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9082dba13c3368d5b9f18a758dfb5c84ffeaa78096788174779a8fecab533942.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "90ef5ba3bbf3d76c2c80b8dca363d0644415c00d9758be4b36114224ea309e22", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5a5174c0137b93b2ec68f79e6881383c8ec504c07c8184f4b06e1d158cc38bec", "generation_seconds_batch": 16.470474967965856, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.207185983657837, "target_z": 0.2753906309712795, "ce": 5.982900142669678, "genuineness": 4.770940780639648, "identity": 0.6458172798156738, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/90ef5ba3bbf3d76c2c80b8dca363d0644415c00d9758be4b36114224ea309e22.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "9165d467da47d65c00301f6c77d618991bdd12537cc6a4df82300cf652bf3c49", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "934cc4ed6351d66667dd50d1b1a7d22dfac9a83711af9b8de9169ddac403d368", "generation_seconds_batch": 23.3961015479872, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2860820293426514, "target_z": -0.30078124330051564, "ce": 6.268836498260498, "genuineness": 4.6203155517578125, "identity": 0.4444492757320404, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9165d467da47d65c00301f6c77d618991bdd12537cc6a4df82300cf652bf3c49.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "91b35fc38a207768ab4e072cf10e6ea1dc30b4442b9ff120af96d1acedea6139", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8d90d505feb229a69b8b98ef5d5d7569939fdaf8dd5cf45c72c7b033bb9f5abc", "generation_seconds_batch": 21.726068218937144, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6653404235839844, "target_z": 1.187500032623576, "ce": 5.514282703399658, "genuineness": 4.413206577301025, "identity": 0.4558292031288147, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/91b35fc38a207768ab4e072cf10e6ea1dc30b4442b9ff120af96d1acedea6139.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "91d39cda16e48e778c87318c8ae5be8759481655e5fa3af90cb06a0cc1cc2a8a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f649472a26a4b83bf9b7ceff494c2ba818c19134847496f652b90cdd2c882d2b", "generation_seconds_batch": 22.269744602032006, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0283737182617188, "target_z": 0.7890625937927812, "ce": 5.80021858215332, "genuineness": 4.570107460021973, "identity": 0.48358726501464844, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/91d39cda16e48e778c87318c8ae5be8759481655e5fa3af90cb06a0cc1cc2a8a.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "92036e2243dfc73a606e4f6d66a86863a10778398c7c5bd1517523690caad817", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2d2ac2d29761e523387a6692ed7d1aad9aa0193de82dc017605d6451b868c97c", "generation_seconds_batch": 22.40843909897376, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9854965209960938, "target_z": 0.13671873339693005, "ce": 6.221180438995361, "genuineness": 4.7332844734191895, "identity": 0.7929041385650635, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/92036e2243dfc73a606e4f6d66a86863a10778398c7c5bd1517523690caad817.mp3" }, { "id": "heldout-en-0", "language": "en", "sentences": [ "Your message arrived after I had left.", "I read it twice... before putting the phone away.", "Let me know when we can talk." ], "text": "Your message arrived after I had left. I read it twice... before putting the phone away. Let me know when we can talk.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-0", "key": "92a1e10e153bf874f2fceb03252aafed5a503839eb56ceba1b312e45e54270c5", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Your message arrived after I had left.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Let me know when we can talk.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Your message arrived after I had left.\n(Natural clear speech.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Natural clear speech.) [3.2 seconds duration] Let me know when we can talk.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1512a1d2719053649fa5a0a568f7e67a89bce02f689f02d452cace257d5a0f84", "generation_seconds_batch": 15.159079071017914, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4444873332977295, "target_z": 0.42382812572820483, "ce": 5.327630043029785, "genuineness": 4.193545341491699, "identity": 0.4538123905658722, "hypothesis": "Your message arrived after I had left. I read it twice before putting the phone away. Let me know when we can talk.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/92a1e10e153bf874f2fceb03252aafed5a503839eb56ceba1b312e45e54270c5.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "931d70a2e6a15104512cc185f3476a33c5284941cb32249daf6dbd8860b02ec6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d5db0e472aa9e292db823bd8eeda906da85bff58c214ccc5553b3f114e6532cb", "generation_seconds_batch": 22.575762580963783, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6170549392700195, "target_z": -0.09375001631178803, "ce": 6.038498878479004, "genuineness": 4.770940780639648, "identity": 0.7451181411743164, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/931d70a2e6a15104512cc185f3476a33c5284941cb32249daf6dbd8860b02ec6.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "939ff004c3a0203c4f14eba4e1793d4afe8acbd4887bc34f476a5e470af99bb7", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "a576acd83b6bc6e0d35401e0cfac6776d485d12868baff66e92a2c95bf29cbfe", "generation_seconds_batch": 17.15911556710489, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5904030799865723, "target_z": 1.1406249871835952, "ce": 5.228346347808838, "genuineness": 4.532451152801514, "identity": 0.8254900574684143, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/939ff004c3a0203c4f14eba4e1793d4afe8acbd4887bc34f476a5e470af99bb7.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "94109116b355f826d85f238108572f1dddc1a46a61a1015edf27223289725072", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "4c861b290055558845fcdd6a8f4295474d98bb1e9cef84878bd40cb01e83a736", "generation_seconds_batch": 16.856875904020853, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1969823837280273, "target_z": 0.8945312223282167, "ce": 6.1020402908325195, "genuineness": 4.444586753845215, "identity": 0.4625224471092224, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/94109116b355f826d85f238108572f1dddc1a46a61a1015edf27223289725072.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "946fa06ee97a27f3a6367740f9efcee162261c60d28797acd7a30723f7d03c7f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8aff3e14a393818348f2dc8636908575564606ee935b889029847ca10ccc85fe", "generation_seconds_batch": 20.67637860402465, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.790235996246338, "target_z": 1.2656251083568777, "ce": 5.704906940460205, "genuineness": 4.582659721374512, "identity": 0.7730885744094849, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/946fa06ee97a27f3a6367740f9efcee162261c60d28797acd7a30723f7d03c7f.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "955f9124da792b0ae6e162f639aa590661b11a8ffe808b5e5f15aab9e99bec08", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6886d464df64b54d6791f9d1fd246a3c222cf42608b7139191a7afcba7f869b2", "generation_seconds_batch": 22.956270491005853, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4788336753845215, "target_z": 0.44531255883894966, "ce": 6.046441555023193, "genuineness": 4.595211505889893, "identity": 0.7766463756561279, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/955f9124da792b0ae6e162f639aa590661b11a8ffe808b5e5f15aab9e99bec08.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "9578153b96d2965cc6736f08dd90848d7a28e638ea9afaad0fa550438421ef04", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "083a8096e9fe5e726a1863332f30ab7566a03d63524aaad9e6e12e2f458e138b", "generation_seconds_batch": 20.677407835028134, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.138493537902832, "target_z": 0.23242191388613753, "ce": 5.839932441711426, "genuineness": 4.595211505889893, "identity": 0.5780709981918335, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9578153b96d2965cc6736f08dd90848d7a28e638ea9afaad0fa550438421ef04.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "95bbabe8a0c3723c9e01245b6e710c95c84ba75b1afdb3594a21b6be7bf25e58", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "7d0e12eb68e7d94b00aeccebe8978159f880d369b47cf5461c5630433706240d", "generation_seconds_batch": 23.58572617196478, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.003394603729248, "target_z": 0.7734375786461208, "ce": 5.7128496170043945, "genuineness": 4.68307638168335, "identity": 0.5704792737960815, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/95bbabe8a0c3723c9e01245b6e710c95c84ba75b1afdb3594a21b6be7bf25e58.mp3" }, { "id": "heldout-en-1", "language": "en", "sentences": [ "The keys are on the kitchen table.", "I checked the drawer... then looked beside the window.", "Take them with you when you leave." ], "text": "The keys are on the kitchen table. I checked the drawer... then looked beside the window. Take them with you when you leave.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-1", "key": "95dc37e5dfb0587fcee91dcd94e9967f5e2837731b3dc424f9376cedc8d5ea9e", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The keys are on the kitchen table.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Take them with you when you leave.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The keys are on the kitchen table.\n(Natural clear speech.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Natural clear speech.) [3.2 seconds duration] Take them with you when you leave.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "b3075ec8ef33c4fc1a91cfd8d82cb0ba1f0d8e56c2f6b1a7c3efceb8e7c536df", "generation_seconds_batch": 23.266846791026182, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.853519916534424, "target_z": 0.679687487766159, "ce": 5.228346347808838, "genuineness": 4.482243061065674, "identity": 0.23460358381271362, "hypothesis": "The keys are on the kitchen table. I checked the drawer, then l looked beside the window. Take them with you when you leave.", "wer": 0.043478260869565216, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/95dc37e5dfb0587fcee91dcd94e9967f5e2837731b3dc424f9376cedc8d5ea9e.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "960dd65488bb5d1b6b66e52ee553ec819120c9b00d21e2fedd32269e92bf68a8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c7766c3df082ba4e9d1fd80158474649eafe8eec4f53a94054006ad4151c38da", "generation_seconds_batch": 22.751799242920242, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7669295072555542, "target_z": 0.0, "ce": 6.197352409362793, "genuineness": 4.7332844734191895, "identity": 0.6138550639152527, "hypothesis": "Ich hab die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0625, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/960dd65488bb5d1b6b66e52ee553ec819120c9b00d21e2fedd32269e92bf68a8.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "964a0974f25735bb15ec0470005d6e7db8d6f8a60175780b41e0c7574e13707b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "779deb538abc58078063f373d340e34bd3e47e155f66d270e88412bdcf529067", "generation_seconds_batch": 17.23506281792652, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.428039073944092, "target_z": 1.0390625378666507, "ce": 5.569881439208984, "genuineness": 4.645420074462891, "identity": 0.4552956223487854, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/964a0974f25735bb15ec0470005d6e7db8d6f8a60175780b41e0c7574e13707b.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "96ca8f21f124cdea79d4517254eb8335dc82d74fa98d42d15e172ca7c9d704c4", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "812f45be21a1949e06f122d318eca05ed65666514f6ae7ab845418256b7dc97e", "generation_seconds_batch": 22.06067972292658, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.359346389770508, "target_z": 0.996093671645161, "ce": 5.5539960861206055, "genuineness": 4.65797233581543, "identity": 0.5509849786758423, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/96ca8f21f124cdea79d4517254eb8335dc82d74fa98d42d15e172ca7c9d704c4.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "971cc22cb3b03c6ba279d3637c5abfe2c5fe0e6c1c5f37f928ef7728aceb1767", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2d3bd486c4a008f8ca9a18d5a100642864295e599b3e1e32c119f136f3b2c691", "generation_seconds_batch": 16.273260690970346, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0362911224365234, "target_z": -0.4570312456307711, "ce": 6.149696350097656, "genuineness": 4.532451152801514, "identity": 0.6795802712440491, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/971cc22cb3b03c6ba279d3637c5abfe2c5fe0e6c1c5f37f928ef7728aceb1767.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "9751fa2722606dbaa55d46bbb95a14c4cd3b355a36417c3c05d6f40f25207329", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "70d7f01e38ddcf39bafd733e315031d5c9b74bdbf7c3844b412fb227b125ca0e", "generation_seconds_batch": 16.273260690970346, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0331687927246094, "target_z": -0.4589843352400167, "ce": 6.1020402908325195, "genuineness": 4.444586753845215, "identity": 0.7722195386886597, "hypothesis": "Ik heb het kleine doosje gevonden, het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9751fa2722606dbaa55d46bbb95a14c4cd3b355a36417c3c05d6f40f25207329.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "97a2d32a8c56b10a287b7f18936cf6537ed9f393ad7b3f0a67c51a05ef64f47f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "30e65fb1941cd472605c124655ef67443673ea1415032f173a16b2aa23889d0d", "generation_seconds_batch": 21.56775121903047, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.221961498260498, "target_z": 0.9101562374748771, "ce": 5.367343425750732, "genuineness": 4.105680465698242, "identity": 0.5923640727996826, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/97a2d32a8c56b10a287b7f18936cf6537ed9f393ad7b3f0a67c51a05ef64f47f.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "97e09b38e63bc6a7f3788ae731b2d49c02e0c757efbd2dcde47dc8cd9aeefe2e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8284dc28ec2c3a88dc5f73e7967b71bf0c4da5321fde2d97461455f69a25f4d9", "generation_seconds_batch": 23.391305606928654, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4155495166778564, "target_z": 1.0312500302933205, "ce": 5.4388275146484375, "genuineness": 4.519899368286133, "identity": 0.44955921173095703, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/97e09b38e63bc6a7f3788ae731b2d49c02e0c757efbd2dcde47dc8cd9aeefe2e.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "97ff763d741702a0e71d71a9f0a65fd3f149304bbfeb18c05bc1ce9e9d60db65", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "80fc13d8cc947d124fa4ec41a7388cb406f5a9c5220a6b8ff1282490cb664466", "generation_seconds_batch": 22.318794004037045, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4671804904937744, "target_z": -0.1874999580554022, "ce": 6.070269584655762, "genuineness": 4.557555198669434, "identity": 0.7356393933296204, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/97ff763d741702a0e71d71a9f0a65fd3f149304bbfeb18c05bc1ce9e9d60db65.mp3" }, { "id": "heldout-en-3", "language": "en", "sentences": [ "Someone moved the little blue chair.", "It used to stand here... beside the tall bookshelf.", "We can put it back after lunch." ], "text": "Someone moved the little blue chair. It used to stand here... beside the tall bookshelf. We can put it back after lunch.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-3", "key": "9842b19bd39d665c79ca74b8be917b311640a5d35c78c019f766b71d5808fef5", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Someone moved the little blue chair.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can put it back after lunch.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Someone moved the little blue chair.\n(Natural clear speech.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Natural clear speech.) [3.2 seconds duration] We can put it back after lunch.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5dfd273c05374be3549ab1e80d3dd1f7d4dd5ecef355879848b89c7cf4fdd7a2", "generation_seconds_batch": 22.48715424502734, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3780808448791504, "target_z": 1.0078125075733302, "ce": 5.720792293548584, "genuineness": 4.3127899169921875, "identity": 0.34359443187713623, "hypothesis": "Someone moved the little blue chair. It used to stand here beside the tall bookshelf. We can put it back after lunch.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9842b19bd39d665c79ca74b8be917b311640a5d35c78c019f766b71d5808fef5.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "9865e2a151cd3c642fd8907d9ab45d6f9123dbcc61a80c0ecf34fb383cccba43", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2967c7237d855847371f8f060adb6e59862beb2e47dcb0e5ee2e65597d38a7eb", "generation_seconds_batch": 22.03403298801277, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.577913522720337, "target_z": 1.132812479610265, "ce": 5.732706069946289, "genuineness": 4.65797233581543, "identity": 0.6946552991867065, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9865e2a151cd3c642fd8907d9ab45d6f9123dbcc61a80c0ecf34fb383cccba43.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "9869d1412c2a43fdc736411bd4fbfcc8e3dc3a88f5edc38fcb1af9870254f867", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "11e37f34ab6bd7b8001fc514d3adac694a033c13bf2eb1051d521586e0f41725", "generation_seconds_batch": 24.147991469944827, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.001108407974243, "target_z": 0.14648433057950583, "ce": 6.0146708488464355, "genuineness": 4.833701133728027, "identity": 0.6302626132965088, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9869d1412c2a43fdc736411bd4fbfcc8e3dc3a88f5edc38fcb1af9870254f867.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "98a46ad00383a7ce05a4f8387199ea8f6ba4fae2f9f38f68a3b326162cc18baa", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "0ac1459f23ff9532a743bf7f9725a76de1c9690c9058b2e39bffe4e30fb5aea6", "generation_seconds_batch": 15.199790642946027, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.716135025024414, "target_z": 0.593750053595875, "ce": 6.038498878479004, "genuineness": 4.821148872375488, "identity": 0.42004233598709106, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/98a46ad00383a7ce05a4f8387199ea8f6ba4fae2f9f38f68a3b326162cc18baa.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "98de10d20e0b527ad65b74500962e445545e7597e69087f8cc0603c0edcbf191", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "cf8fb1da25021f65c382870b05602b26bd89f5757e799e50b9f161bec1a19cef", "generation_seconds_batch": 26.75805490789935, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.690319538116455, "target_z": 1.2031250477702364, "ce": 5.164804935455322, "genuineness": 4.293961524963379, "identity": 0.225820854306221, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/98de10d20e0b527ad65b74500962e445545e7597e69087f8cc0603c0edcbf191.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "98e03f5800facfba44786121151e0ddcb1c90346352165242c5c5324d36f532f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d67a09fdad5419db312a603580ea880d2d0855837e5b3023f0e516e7494fcb98", "generation_seconds_batch": 16.999474014970474, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.321877956390381, "target_z": 0.9726562980615183, "ce": 5.593709468841553, "genuineness": 4.68307638168335, "identity": 0.5173414945602417, "hypothesis": "La porte est encore ouverte, j'ai posé ton manteau sur la chaise, dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/98e03f5800facfba44786121151e0ddcb1c90346352165242c5c5324d36f532f.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "994cb158cdbf35d53e0578785174a2e16de4109e9de35e4a0e779ece3a8026c5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a5cd036397f25889dccc8f072a4d624a881212c1448ee6b24cf82d1a5db8a4e3", "generation_seconds_batch": 22.178197700995952, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.221961498260498, "target_z": 0.9101562374748771, "ce": 5.597680568695068, "genuineness": 4.758388519287109, "identity": 0.6272444725036621, "hypothesis": "He encontrado la cajita. estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/994cb158cdbf35d53e0578785174a2e16de4109e9de35e4a0e779ece3a8026c5.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "999b6499f0be20fd7c569726d10a8e71b889e640385fa4edfec3efed21eb4f39", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1f925324b2d3009a02f6bdd4c353caa69a0a6517062b1d6708e1f1cbe325ad07", "generation_seconds_batch": 22.680698742973618, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.310224771499634, "target_z": 0.3398437811671664, "ce": 5.669164657592773, "genuineness": 4.7332844734191895, "identity": 0.5437505841255188, "hypothesis": "He encontrado la cajita. estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/999b6499f0be20fd7c569726d10a8e71b889e640385fa4edfec3efed21eb4f39.mp3" }, { "id": "heldout-de-0", "language": "de", "sentences": [ "Deine Nachricht kam, als ich schon weg war.", "Ich las sie zweimal... dann legte ich das Handy weg.", "Sag mir, wann wir reden können." ], "text": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal... dann legte ich das Handy weg. Sag mir, wann wir reden können.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-0", "key": "99bf1b7d3eed7b885fdae888ab00b4960791c3971d259b46f771306d2801e39b", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wann wir reden können.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wann wir reden können.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a86750362e901daec65f2814d72a73f504fbf7946b2524317e69b127448870fb", "generation_seconds_batch": 17.495784764993005, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.5554436445236206, "target_z": -0.7578124889312867, "ce": 6.332377910614014, "genuineness": 4.607763767242432, "identity": 0.5014426708221436, "hypothesis": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal, dann legte ich das Handy weg. Sag mir, wann wir reden können.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/99bf1b7d3eed7b885fdae888ab00b4960791c3971d259b46f771306d2801e39b.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "99d185451c5db776169e4e25cf5d50e2af3eca07d410af6e99901a92a8b0a30c", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a565c9f950adce35e208750102d81fc6d8b7c97a659515bdd05bb28536a9b764", "generation_seconds_batch": 15.674808133975603, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2236342430114746, "target_z": -0.3398437811671664, "ce": 6.109982967376709, "genuineness": 4.632867813110352, "identity": 0.6802173852920532, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/99d185451c5db776169e4e25cf5d50e2af3eca07d410af6e99901a92a8b0a30c.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "99e356c839afb04547dda312363874d9eb72633ec00d2b1585e908c36b2747b6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a487f5ec3ca9a2dc1474821adfd3e510bc782df065769155c9feb5d6c77ce94b", "generation_seconds_batch": 23.100148959085345, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7973170280456543, "target_z": 0.6445312782543471, "ce": 5.77639102935791, "genuineness": 4.68307638168335, "identity": 0.694378137588501, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/99e356c839afb04547dda312363874d9eb72633ec00d2b1585e908c36b2747b6.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "9a297ef102e8ae85879860a982eb958ff3c66a874b6a048aafce912c36094915", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0f7d6a01ec626bdc7e68035cae5e7cadc580c25ec33d776720e975bf6b3a7504", "generation_seconds_batch": 21.56337803194765, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6108101606369019, "target_z": -0.09765627009845311, "ce": 6.213237762451172, "genuineness": 4.7081804275512695, "identity": 0.509758472442627, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9a297ef102e8ae85879860a982eb958ff3c66a874b6a048aafce912c36094915.mp3" }, { "id": "heldout-nl-1", "language": "nl", "sentences": [ "De sleutels liggen op de keukentafel.", "Ik keek in de la... en daarna naast het raam.", "Neem ze mee als je weggaat." ], "text": "De sleutels liggen op de keukentafel. Ik keek in de la... en daarna naast het raam. Neem ze mee als je weggaat.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-1", "key": "9b54e935cd5d0864ae661bc84ee90d9f88303c7dd3dd3f2fe0fe355706163c08", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Neem ze mee als je weggaat.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Natural clear speech.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Natural clear speech.) [3.2 seconds duration] Neem ze mee als je weggaat.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e5b9242040ca5205ed83951f120281f330ee1122512d2f402e802e68a0c31512", "generation_seconds_batch": 15.741536403889768, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.009639263153076, "target_z": 0.7773437578646121, "ce": 5.418970584869385, "genuineness": 4.20609712600708, "identity": 0.387640118598938, "hypothesis": "De sleutels liggen op de keukentafel. Ik keek in de la en daarna, naast het raam, neem ze mee als je weggaat.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9b54e935cd5d0864ae661bc84ee90d9f88303c7dd3dd3f2fe0fe355706163c08.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "9baf2e946a9213a97985e0179d911760e4b7c611615f092bd9d1371cb57c7620", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e2dabc60e0a567360b16f2aa9abb7ff071586342406b5010e0a8cb7f36814f1d", "generation_seconds_batch": 16.563943021930754, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.147860527038574, "target_z": 0.23828118271387438, "ce": 5.871703147888184, "genuineness": 4.7081804275512695, "identity": 0.6758355498313904, "hypothesis": "Die Tür ist noch offen. Ich hab deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9baf2e946a9213a97985e0179d911760e4b7c611615f092bd9d1371cb57c7620.mp3" }, { "id": "heldout-nl-3", "language": "nl", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-3", "key": "9c09e940c5dab53287ffae2d1a8fa5296f41b2df71714656aaa2bb8830bf5162", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We zetten het na de lunch terug.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7ccb35e799c441a929ec88916674660a5afd1d086e51e62976d84f19a1dafede", "generation_seconds_batch": 16.50808106001932, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": -0.14397084712982178, "target_z": -1.1953124656287324, "ce": 6.38003396987915, "genuineness": 4.293961524963379, "identity": 0.721808671951294, "hypothesis": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier naast de hoge boekenkast. We zetten het na de lunch terug.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9c09e940c5dab53287ffae2d1a8fa5296f41b2df71714656aaa2bb8830bf5162.mp3" }, { "id": "heldout-es-2", "language": "es", "sentences": [ "La carretera gira a la izquierda tras el puente.", "Sigue... hasta la pequeña puerta de madera.", "Te esperaré al otro lado." ], "text": "La carretera gira a la izquierda tras el puente. Sigue... hasta la pequeña puerta de madera. Te esperaré al otro lado.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-2", "key": "9c969db20a6db380e08592f6aed9c9189b3db634d677ebe8f573f593ebdfd2d3", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Te esperaré al otro lado.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Te esperaré al otro lado.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "779b1b4b28633608f8f2df055883fc2f646667f96b5ef5835fbf84460ef68621", "generation_seconds_batch": 16.231410023057833, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9301300048828125, "target_z": -0.5234374854359035, "ce": 6.173524379730225, "genuineness": 4.444586753845215, "identity": 0.6986501812934875, "hypothesis": "La carretera gira a la izquierda tras el puente. Sigue hasta la pequeña puerta de madera. Te esperaré al otro lado.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9c969db20a6db380e08592f6aed9c9189b3db634d677ebe8f573f593ebdfd2d3.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "9d3ea6b571f01297895bf54a4f940d71a1ed928d359237738bacec0b12a7045b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0a8e3f1e074377d9f92d7013c5c0c5957943c15dc668df416fc3309450e8060e", "generation_seconds_batch": 17.09989327006042, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9051508903503418, "target_z": -0.5390625005825639, "ce": 6.173524379730225, "genuineness": 4.582659721374512, "identity": 0.5786643624305725, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oud oude raam. We kunnen er samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9d3ea6b571f01297895bf54a4f940d71a1ed928d359237738bacec0b12a7045b.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "9da6a16f0afafe29b0d2c48b00d1fc29613ff8a9955eb49d6c0d86b8b3d7d350", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7a82aad58c8a5ad5c97c426c0b4f457921bc1e9f86314c79fa45c77685bd6b5c", "generation_seconds_batch": 22.048773981980048, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.578749895095825, "target_z": 0.5078124702892433, "ce": 6.046441555023193, "genuineness": 4.507347106933594, "identity": 0.6570305824279785, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9da6a16f0afafe29b0d2c48b00d1fc29613ff8a9955eb49d6c0d86b8b3d7d350.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "9de7d6e876ac3d7514501d1e7bd25dd4e1a76d0f2ebad7d5592dea0a24ca82cc", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "489c1330786893c0df0f46e216b118de4101d5f40deaee8fa9df598ef1be4113", "generation_seconds_batch": 16.326507619931363, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4788336753845215, "target_z": 0.44531255883894966, "ce": 6.213237762451172, "genuineness": 4.256305694580078, "identity": 0.5593451261520386, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9de7d6e876ac3d7514501d1e7bd25dd4e1a76d0f2ebad7d5592dea0a24ca82cc.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "9e2263bf148b1977258032fb5e61a9cebaf138d4539a554ee4c770a25e499c35", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bcb8099a4a27dac0b26e462fd21e07cb79fb11c0e8e0c01eea3e063003541ab5", "generation_seconds_batch": 15.87353016808629, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0510666370391846, "target_z": 0.17773436087282643, "ce": 6.062326908111572, "genuineness": 4.7332844734191895, "identity": 0.8353061676025391, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9e2263bf148b1977258032fb5e61a9cebaf138d4539a554ee4c770a25e499c35.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "9e71e6fbc6130f9bb1b54bde3adafa34c8e1d08ff23245a01b0e27fb19fc8205", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "400314a8b83690e8dcab0a391b8149bdf7ef746c85a2e5e6a7997708db83626a", "generation_seconds_batch": 23.25329788296949, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.965925693511963, "target_z": 0.7499999067897827, "ce": 5.839932441711426, "genuineness": 4.595211505889893, "identity": 0.5602434277534485, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9e71e6fbc6130f9bb1b54bde3adafa34c8e1d08ff23245a01b0e27fb19fc8205.mp3" }, { "id": "heldout-de-3", "language": "de", "sentences": [ "Jemand hat den kleinen blauen Stuhl umgestellt.", "Er stand hier... neben dem hohen Bücherregal.", "Wir stellen ihn nach dem Essen zurück." ], "text": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier... neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-3", "key": "9e801469135446b53e045cfbcb35112f977300f2f1c9ba00ea5bfe265e05ea1d", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Natural clear speech.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Natural clear speech.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e176908f0721648ef912d7ec49faae2da11cd6b79f70c6c5fe9dae1a9e9b9c2c", "generation_seconds_batch": 24.293834752053954, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5974841117858887, "target_z": 0.5195311570810647, "ce": 5.0754499435424805, "genuineness": 4.406930446624756, "identity": 0.6219208240509033, "hypothesis": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9e801469135446b53e045cfbcb35112f977300f2f1c9ba00ea5bfe265e05ea1d.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "9e857c03f30a269961e363af81b74bd1cfaa7e92be0df01af3343f2f8fa1c582", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9f8f10b5c80ae8f92197b49ecbbe95919ea5c46316587f67690d034d687515e9", "generation_seconds_batch": 16.395632397034205, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1518194675445557, "target_z": -0.38476558786155407, "ce": 6.24500846862793, "genuineness": 4.557555198669434, "identity": 0.7544786930084229, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9e857c03f30a269961e363af81b74bd1cfaa7e92be0df01af3343f2f8fa1c582.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "9e9be894a989d2c2e989a55076151a0e8ba6f2c51d0cd5557bfbb826ba734320", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3420d9110e844d6c9e57ef44e91286713c480efb470dbb2312f516715b549098", "generation_seconds_batch": 23.224551855004393, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.453854560852051, "target_z": 0.42968754369228934, "ce": 5.982900142669678, "genuineness": 4.7834930419921875, "identity": 0.6717789769172668, "hypothesis": "La porte est encore ouverte j'ai posé ton manteau sur la chaise. Dis-moi, quand tu es prêt", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9e9be894a989d2c2e989a55076151a0e8ba6f2c51d0cd5557bfbb826ba734320.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "9f2e979e23c1835007a0ea402a0ca65775c14b55bb2a20bee36d0d37e1676995", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "552b3993547c5c2d86e8dd40656ebfdd16b67a637c4eee2c9f835bb2ae207c61", "generation_seconds_batch": 22.61261413304601, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.776296615600586, "target_z": 0.005859343395910696, "ce": 6.268836498260498, "genuineness": 4.268857479095459, "identity": 0.455219030380249, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9f2e979e23c1835007a0ea402a0ca65775c14b55bb2a20bee36d0d37e1676995.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "9f4c5e09e5d7560f64e68496c82caca525f842c53276baeb4e87c31b4fc225a6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9626b56046f1d20a4355888cfc45768cfd40450436f18a09403250e6ef44342e", "generation_seconds_batch": 23.35871112300083, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.940946578979492, "target_z": 0.7343748916431224, "ce": 5.7128496170043945, "genuineness": 4.8588056564331055, "identity": 0.7521970868110657, "hypothesis": "J'ai trouvé la petite boîte elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9f4c5e09e5d7560f64e68496c82caca525f842c53276baeb4e87c31b4fc225a6.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "9f963f87ec644851bfeb1d1fadb99519df5c93d40f3697763748023ee4345392", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "c1dd6265b7bee112391651f391af090bf2e1000af88d8987ce028c7f22d13f0f", "generation_seconds_batch": 22.728156588971615, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.684911012649536, "target_z": 0.5742187100943758, "ce": 5.967014789581299, "genuineness": 4.306513786315918, "identity": 0.4600263237953186, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/9f963f87ec644851bfeb1d1fadb99519df5c93d40f3697763748023ee4345392.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "a0a4bedffb38ed9dfb575c11f8d21ba94d6f5b69887979c0bb750807497d8e98", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e07828b176d49e41fcb81ab9249d5d88ab9d67412b171b34027678827bf8e44c", "generation_seconds_batch": 23.120403998997062, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7028088569641113, "target_z": 1.210937406207219, "ce": 5.2918877601623535, "genuineness": 4.607763767242432, "identity": 0.6661697030067444, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre, nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a0a4bedffb38ed9dfb575c11f8d21ba94d6f5b69887979c0bb750807497d8e98.mp3" }, { "id": "heldout-fr-0", "language": "fr", "sentences": [ "Ton message est arrivé après mon départ.", "Je l’ai relu... puis j’ai rangé mon téléphone.", "Dis-moi quand nous pourrons parler." ], "text": "Ton message est arrivé après mon départ. Je l’ai relu... puis j’ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-0", "key": "a0deeba0c3b8744b9ef01febb18c5a9eeb805f9afc69d51564b79ccc5f5185cb", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Natural clear speech.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c3ada815b079230014bee85f0797e8b6ca93d06d8f982bd518c963a780e57088", "generation_seconds_batch": 17.450143280089833, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4655075073242188, "target_z": 1.0624999114502935, "ce": 5.297844886779785, "genuineness": 4.6203155517578125, "identity": 0.7238675355911255, "hypothesis": "Attends, le message est arrivé après mon départ, je l'ai relu, puis j'ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "wer": 0.09523809523809523, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a0deeba0c3b8744b9ef01febb18c5a9eeb805f9afc69d51564b79ccc5f5185cb.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "a12d59aee9a41064f2e2f9e0eac3996ba3c6ee9344c48784c98286d6c1d8a789", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2a4091afa4af6810386aa7f886cadde0c725519a5f29fae5b44b250ef845272b", "generation_seconds_batch": 23.989452551002614, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.3297953605651855, "target_z": -0.2734375413620339, "ce": 6.062326908111572, "genuineness": 4.68307638168335, "identity": 0.597082257270813, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a12d59aee9a41064f2e2f9e0eac3996ba3c6ee9344c48784c98286d6c1d8a789.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "a13f651aba1d29e232b3e2385fb385a1df12008a640c45583d181450c2fca67d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0fc7695352793b8d502e6f892d02baf3fbe50bbcc6f27c1b531ffa5646083aad", "generation_seconds_batch": 16.80960233102087, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2719197273254395, "target_z": 0.9414062677681977, "ce": 5.375286102294922, "genuineness": 4.545003414154053, "identity": 0.5658593773841858, "hypothesis": "La puerta la sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.06666666666666667, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a13f651aba1d29e232b3e2385fb385a1df12008a640c45583d181450c2fca67d.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "a14d0123ad3e61520f3a903dcfb95b5abad81a384f7f7808497459af58a7ede5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b73f4e0ab25e29954720717b36cc3d611ecd90279346549b92b204108779394a", "generation_seconds_batch": 23.343918393948115, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.869968295097351, "target_z": 0.06445315019588686, "ce": 6.046441555023193, "genuineness": 4.607763767242432, "identity": 0.6784154176712036, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a14d0123ad3e61520f3a903dcfb95b5abad81a384f7f7808497459af58a7ede5.mp3" }, { "id": "heldout-nl-0", "language": "nl", "sentences": [ "Je bericht kwam nadat ik vertrokken was.", "Ik las het twee keer... en legde mijn telefoon weg.", "Zeg maar wanneer we kunnen praten." ], "text": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer... en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-0", "key": "a1a90286b791ca3ba3c114fc510982061a032a294437a809a09033328eb6170e", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Natural clear speech.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1612a3f98669b1a2b44e66ecbaa9f2835d5348aa3beb85aaf982e7d7631a9ed6", "generation_seconds_batch": 17.088680595043115, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9159677028656006, "target_z": 0.7187500256328098, "ce": 5.601652145385742, "genuineness": 4.356721878051758, "identity": 0.4543686509132385, "hypothesis": "Je bericht kwam nadat ik vertrokken was, ik las het twee keer en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a1a90286b791ca3ba3c114fc510982061a032a294437a809a09033328eb6170e.mp3" }, { "id": "heldout-de-2", "language": "de", "sentences": [ "Die Straße führt hinter der Brücke nach links.", "Geh weiter... bis zu dem kleinen Holztor.", "Ich warte auf der anderen Seite." ], "text": "Die Straße führt hinter der Brücke nach links. Geh weiter... bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-2", "key": "a21f722a3b448b90d979eb7f599a9da414cc40b9b00c8b814963600e6b8c1877", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Natural clear speech.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Natural clear speech.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e3df1f0b275d7e7490a109c08ee845e11d351027f5e1b439d546033f6ff7ffd1", "generation_seconds_batch": 22.961345597053878, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.009639263153076, "target_z": 0.7773437578646121, "ce": 5.311744689941406, "genuineness": 4.432034492492676, "identity": 0.4191397428512573, "hypothesis": "Die Straße führt hinter der Brücke nach links. Geh weiter bis zum kleinen Holztor. Ich warte auf der anderen Seite.", "wer": 0.09523809523809523, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a21f722a3b448b90d979eb7f599a9da414cc40b9b00c8b814963600e6b8c1877.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "a24724cbc839a4de9c9345007e1f9484cf3baf5d2c35cea43633de85aeb8cb78", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c22e2572ee9653519a8e018b4cacd0bb799b3ca04915566dee437456e40e5388", "generation_seconds_batch": 21.241574181010947, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4155495166778564, "target_z": 1.0312500302933205, "ce": 5.355429172515869, "genuineness": 4.469690799713135, "identity": 0.7017637491226196, "hypothesis": "J'ai j'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a24724cbc839a4de9c9345007e1f9484cf3baf5d2c35cea43633de85aeb8cb78.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "a25cc55a2e05b1222161d1b0a15e81399c933b8967f951db420ac47c5adf99dd", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] De deur staat nog open.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e285c38cab63f5c0475f0d86c183b9f9540738d2edbdcca45a7265972c6b67c7", "generation_seconds_batch": 17.59209238004405, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.567933201789856, "target_z": -0.7499999813579565, "ce": 6.141753673553467, "genuineness": 4.557555198669434, "identity": 0.5677171349525452, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a25cc55a2e05b1222161d1b0a15e81399c933b8967f951db420ac47c5adf99dd.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "a28bb5a233a8c6b21c1c9093055e72af80bf1a05bc0b1117840ec8e6db0feeb7", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5a885889df38679dba632310dfbcb00719dc50a7aa0710316b228ca8e98ce679", "generation_seconds_batch": 21.803624239051715, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.790235996246338, "target_z": 1.2656251083568777, "ce": 5.577824115753174, "genuineness": 4.607763767242432, "identity": 0.6439168453216553, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a28bb5a233a8c6b21c1c9093055e72af80bf1a05bc0b1117840ec8e6db0feeb7.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "a430fb7fd54aaed77e19c367493ce5649bd1ec5bdfb314b77ffb464555c53fff", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5f9292c24f040cdbb0e7d54e2b68211bd47aa18a3aa1a8dcafc3b2d34a113cb9", "generation_seconds_batch": 16.823258957010694, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.552934408187866, "target_z": 1.1171874644636046, "ce": 5.371314525604248, "genuineness": 4.645420074462891, "identity": 0.7985765933990479, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a430fb7fd54aaed77e19c367493ce5649bd1ec5bdfb314b77ffb464555c53fff.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "a447b08a33b1d0db46b9458820c3e3e2e18b206cfc18133774701a38f86c174a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bb24d3014c474b41666ef7139efb49de3a1ae4ee0ed5841730bf6771632e3e83", "generation_seconds_batch": 22.96275072300341, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1634726524353027, "target_z": 0.24804692903279785, "ce": 6.348263263702393, "genuineness": 4.68307638168335, "identity": 0.5274926424026489, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a447b08a33b1d0db46b9458820c3e3e2e18b206cfc18133774701a38f86c174a.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "a50fa1bb8391a9c132bb5270f5650567378581df9a462437640cf9357c2e8e87", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "296b72a6526f2dce6e0d15d72ae2b12cee5d9caaa251dc4bded76e47f10bdede", "generation_seconds_batch": 21.998502444010228, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.035454750061035, "target_z": 0.16796876369025066, "ce": 6.062326908111572, "genuineness": 4.595211505889893, "identity": 0.7654612064361572, "hypothesis": "J'ai trouvé la petite boîte elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a50fa1bb8391a9c132bb5270f5650567378581df9a462437640cf9357c2e8e87.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "a523475b02e6a690ea84be692abcdd34b110f3971ba4ca6a455251c9745a6c4d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bd4d00418dc7766d689e63dec8da6c222bb0cf30a531bce2f591a20121e170bf", "generation_seconds_batch": 23.32059576793108, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.2681840658187866, "target_z": -0.9375000139815326, "ce": 6.221180438995361, "genuineness": 4.6705241203308105, "identity": 0.6315075159072876, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a523475b02e6a690ea84be692abcdd34b110f3971ba4ca6a455251c9745a6c4d.mp3" }, { "id": "heldout-de-0", "language": "de", "sentences": [ "Deine Nachricht kam, als ich schon weg war.", "Ich las sie zweimal... dann legte ich das Handy weg.", "Sag mir, wann wir reden können." ], "text": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal... dann legte ich das Handy weg. Sag mir, wann wir reden können.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-0", "key": "a5cd00baf163d303244ca79b7436562cfbdb803d1a26221bef371acc84b60fdc", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wann wir reden können.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Natural clear speech.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wann wir reden können.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5737d8ac84ac5ee49ee5f0342d9dc520f4ac554292dc7de3b1ffbedd3a9a7e02", "generation_seconds_batch": 17.13481791492086, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1595139503479004, "target_z": 0.871093848744574, "ce": 5.569881439208984, "genuineness": 4.582659721374512, "identity": 0.720352828502655, "hypothesis": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal. Dann legte ich das Handy weg. Sag mir, wann wir reden können.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a5cd00baf163d303244ca79b7436562cfbdb803d1a26221bef371acc84b60fdc.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "a5da74c7acf581a3fb6a2fac3be28130aebb16f9fea87b7cb7819d0d7a721338", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8a685968210b56dae6ab3b4c93db57c0ee59c307828037f0f19a41d4f7503815", "generation_seconds_batch": 17.367203232017346, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.716135025024414, "target_z": 0.593750053595875, "ce": 5.871703147888184, "genuineness": 4.934117794036865, "identity": 0.48540106415748596, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a5da74c7acf581a3fb6a2fac3be28130aebb16f9fea87b7cb7819d0d7a721338.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "a6a336f5797c2d9066a99a15520490d3b1a49b1a7120a21048b450e2cf67a79e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6057e6e0565444d40403dae1ddac6a9ab5e72447cf7d9697d543f60982546a5c", "generation_seconds_batch": 28.466669535031542, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5912394523620605, "target_z": 0.5156249778625734, "ce": 5.498397350311279, "genuineness": 4.519899368286133, "identity": 0.5089616775512695, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a6a336f5797c2d9066a99a15520490d3b1a49b1a7120a21048b450e2cf67a79e.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "a6fe3e62fb88d8ff19cd8151432ad623fe66f4485923c4b111e727088e75f90c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5689ba7e60b7dcc35cf48f177c9b4bd86998236a94dff7432a828b343522244e", "generation_seconds_batch": 21.7260549350176, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.709890127182007, "target_z": 0.589843725241036, "ce": 5.383228778839111, "genuineness": 4.582659721374512, "identity": 0.4571259915828705, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a6fe3e62fb88d8ff19cd8151432ad623fe66f4485923c4b111e727088e75f90c.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "a70730a86b599d3848ba9f4abb3c0390272f19dd6434722d8e47636f99caed1c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a700d78e51e5e8a4d180edf75570dc29cfdf64f61a0248d2839b2270ee6f7bd0", "generation_seconds_batch": 21.130135593004525, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2094719409942627, "target_z": 0.9023437299015469, "ce": 5.95112943649292, "genuineness": 4.444586753845215, "identity": 0.5381537675857544, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a70730a86b599d3848ba9f4abb3c0390272f19dd6434722d8e47636f99caed1c.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "a707b8dcc418e1f6ffe70f4907b81e149352cd5352a779e0c8a19e4019022929", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8be477edf764587485fd0a19ef9f41333fcc417d35e1b20e889165604b6d97a2", "generation_seconds_batch": 17.863581360899843, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6232997179031372, "target_z": -0.08984376252512295, "ce": 5.919358730316162, "genuineness": 4.632867813110352, "identity": 0.49879786372184753, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a707b8dcc418e1f6ffe70f4907b81e149352cd5352a779e0c8a19e4019022929.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "a7d2238b6c664a2ad76fdf12442292ab7435254f501370cef3657850c987c9e0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "722556c70f19e8a4dd5ab7166e7dd20fdfc4c119286fa032796438a9b9cc64f3", "generation_seconds_batch": 20.970334646990523, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.835621953010559, "target_z": 0.042968717085142015, "ce": 6.1894097328186035, "genuineness": 4.65797233581543, "identity": 0.5452387928962708, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a7d2238b6c664a2ad76fdf12442292ab7435254f501370cef3657850c987c9e0.mp3" }, { "id": "heldout-de-3", "language": "de", "sentences": [ "Jemand hat den kleinen blauen Stuhl umgestellt.", "Er stand hier... neben dem hohen Bücherregal.", "Wir stellen ihn nach dem Essen zurück." ], "text": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier... neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-3", "key": "a7d78297d4c90b37aa2548e6328db6cb9e9860fdf23b21909c0d2dd6962cd475", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Natural clear speech.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Natural clear speech.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "712d7cf5cb9ccdd0502efb21397a0a67ef1a59ca39c986b4a9523d52634d970a", "generation_seconds_batch": 24.293834752053954, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4382424354553223, "target_z": 0.4199217973733659, "ce": 5.546053409576416, "genuineness": 4.607763767242432, "identity": 0.6200903058052063, "hypothesis": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a7d78297d4c90b37aa2548e6328db6cb9e9860fdf23b21909c0d2dd6962cd475.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "a853a791d6ec734c1c7395d8ddcc026a650ec179730a0c5fb4f67be086d5dbe9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ca79d498cd49575bff2bdcbf532e2092b5f237a403141dac44117a10b851d7a1", "generation_seconds_batch": 21.323322594049387, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6170549392700195, "target_z": -0.09375001631178803, "ce": 6.109982967376709, "genuineness": 4.808597087860107, "identity": 0.7215011119842529, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a853a791d6ec734c1c7395d8ddcc026a650ec179730a0c5fb4f67be086d5dbe9.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "a877388ae7212a5aca30aa177513d201d8c691d4e4a5eb5c519cf6c86385921f", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0d2434b8f2e9ed3e23a76c17b2d886d52db1729b3662efbc5f3a8b7a5016a5ca", "generation_seconds_batch": 16.15291431103833, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8739270567893982, "target_z": -0.5585937322318023, "ce": 6.41180419921875, "genuineness": 4.406930446624756, "identity": 0.5806659460067749, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a877388ae7212a5aca30aa177513d201d8c691d4e4a5eb5c519cf6c86385921f.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "a87c835c6baecfefc210e29ff2d836a54277b8457a045cb7b7ecd213517788a0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ee23d502033f96853a1e5782c739b2c8b8860655d5c048bfb43b55d77b925641", "generation_seconds_batch": 22.048773981980048, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3780808448791504, "target_z": 1.0078125075733302, "ce": 5.935244083404541, "genuineness": 4.507347106933594, "identity": 0.733244776725769, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a87c835c6baecfefc210e29ff2d836a54277b8457a045cb7b7ecd213517788a0.mp3" }, { "id": "heldout-nl-4", "language": "nl", "sentences": [ "Ik schreef het nummer op de envelop.", "Lees het goed... en controleer de laatste twee cijfers.", "Bel me wanneer je even tijd hebt." ], "text": "Ik schreef het nummer op de envelop. Lees het goed... en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-4", "key": "a8bd001b1fcafd2f49b65d73e6ed12567e4c9664909a4d08700c9febfd977540", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1f1019155bc20d76bf174fd57d67f9c1fdb3dde434bfd33b864b61f67cb95ec5", "generation_seconds_batch": 17.376467431080528, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": -0.10650229454040527, "target_z": -1.1718750174769157, "ce": 6.141753673553467, "genuineness": 4.124508857727051, "identity": 0.8276042342185974, "hypothesis": "Ik schreef het nummer op de envelop. Lees het goed en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a8bd001b1fcafd2f49b65d73e6ed12567e4c9664909a4d08700c9febfd977540.mp3" }, { "id": "heldout-en-1", "language": "en", "sentences": [ "The keys are on the kitchen table.", "I checked the drawer... then looked beside the window.", "Take them with you when you leave." ], "text": "The keys are on the kitchen table. I checked the drawer... then looked beside the window. Take them with you when you leave.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-1", "key": "a8d3f81c15124d0163d54c75a9b9d9caf827027a3498e2d8e4f413f5912b6cb4", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The keys are on the kitchen table.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Take them with you when you leave.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The keys are on the kitchen table.\n(Natural clear speech.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Natural clear speech.) [3.2 seconds duration] Take them with you when you leave.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3a71644d118aa00777df07e352759c3e0327ecf953f8b236a58a40e29cfa351d", "generation_seconds_batch": 17.132835216005333, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9909048080444336, "target_z": 0.765624921936443, "ce": 5.260117053985596, "genuineness": 4.3504462242126465, "identity": 0.42770466208457947, "hypothesis": "The keys are on the kitchen table. I check the drawer, then look beside the window. Take them with you when you leave.", "wer": 0.08695652173913043, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a8d3f81c15124d0163d54c75a9b9d9caf827027a3498e2d8e4f413f5912b6cb4.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "a9101feac8d7ced13c7ec8304da39435b28dc855c906b2318b99a4c58c4bc498", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "00463d03e634674e798d4eab834a207a276f8401caa04d1ea8477bf906d16778", "generation_seconds_batch": 22.183518590056337, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9261711835861206, "target_z": 0.09960935970769873, "ce": 6.173524379730225, "genuineness": 4.494794845581055, "identity": 0.5659762620925903, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a9101feac8d7ced13c7ec8304da39435b28dc855c906b2318b99a4c58c4bc498.mp3" }, { "id": "heldout-nl-3", "language": "nl", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-3", "key": "a922779b906b939685d98037c294176d4f0a5a0c46208cbf04c918673a53bc3e", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Natural clear speech.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Natural clear speech.) [3.2 seconds duration] We zetten het na de lunch terug.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "aa62219b246522a375aac0826f95c804c0351eb807ef8eb11d1f6a75b942ad72", "generation_seconds_batch": 16.979283216991462, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.56626033782959, "target_z": 0.49999996271591307, "ce": 5.756534099578857, "genuineness": 4.394378662109375, "identity": 0.3682903051376343, "hypothesis": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier naast de hoge boekenkast. We zetten het naar de lunch terug.", "wer": 0.047619047619047616, "duration_error_s": 0.120000000000001, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a922779b906b939685d98037c294176d4f0a5a0c46208cbf04c918673a53bc3e.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "a949598a09c2e6c53e881e68216382364b283aa9483ff7943cd63529f39a92cd", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c1579b9131214afe157d73a17dd10189f60190b8d7564733f054bacbdaa302fb", "generation_seconds_batch": 22.673365589929745, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5912394523620605, "target_z": 0.5156249778625734, "ce": 5.903473377227783, "genuineness": 4.846253395080566, "identity": 0.6291944980621338, "hypothesis": "Die Tür ist noch offen. Ich hab deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a949598a09c2e6c53e881e68216382364b283aa9483ff7943cd63529f39a92cd.mp3" }, { "id": "heldout-fr-0", "language": "fr", "sentences": [ "Ton message est arrivé après mon départ.", "Je l’ai relu... puis j’ai rangé mon téléphone.", "Dis-moi quand nous pourrons parler." ], "text": "Ton message est arrivé après mon départ. Je l’ai relu... puis j’ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-0", "key": "a94b7d5da78ff217f8e5a4f7ae4520900bef0ad764d0e4ee898e69de10c28aea", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "94d62564293c8ad4cc7ff15582a4473cd3fe9eb1b3fc3d82a8c08224db3eac66", "generation_seconds_batch": 16.71202097297646, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.295449137687683, "target_z": -0.29492189990460493, "ce": 6.09409761428833, "genuineness": 4.406930446624756, "identity": 0.8263388872146606, "hypothesis": "Ton message est arrivé après mon départ, je l'ai relu, puis j'ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a94b7d5da78ff217f8e5a4f7ae4520900bef0ad764d0e4ee898e69de10c28aea.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "a9e03b1400f4955448b7c1c7ec8c9c68a3487d12a60b2495d3b3c95146000399", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "87f2d049d6f56c0a8d6e148df9e41418b91110c42054a0d508dc0f4f4a902f6c", "generation_seconds_batch": 22.55756815499626, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4530181884765625, "target_z": 1.054687553013311, "ce": 5.307773113250732, "genuineness": 4.519899368286133, "identity": 0.5856072902679443, "hypothesis": "He encontrado la cajita, estaba junto a la ventana vieja. podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a9e03b1400f4955448b7c1c7ec8c9c68a3487d12a60b2495d3b3c95146000399.mp3" }, { "id": "heldout-fr-4", "language": "fr", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-4", "key": "a9f27066ad97f6d67e85856f422869852e744b79355d9d02610da0a5aec8a8e3", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Natural clear speech.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Natural clear speech.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0e75eb3b6d170ca57d4b912f5082cb9b9e6388ab305132087d81bd4f21ff1bc2", "generation_seconds_batch": 16.197387669002637, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.115800380706787, "target_z": 0.8437499976697446, "ce": 4.971202373504639, "genuineness": 4.281409740447998, "identity": 0.6551336050033569, "hypothesis": "J'ai écrit le numéro sur l'enveloppe, lis-le bien et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/a9f27066ad97f6d67e85856f422869852e744b79355d9d02610da0a5aec8a8e3.mp3" }, { "id": "heldout-es-4", "language": "es", "sentences": [ "Escribí el número en el sobre.", "Léelo bien... y comprueba las dos últimas cifras.", "Llámame cuando tengas un momento." ], "text": "Escribí el número en el sobre. Léelo bien... y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-4", "key": "aac2ac5e2a6e5df320afbca9f7284cd386d3f0f5aa6e360abfc0986cbd7cc747", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Escribí el número en el sobre.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Llámame cuando tengas un momento.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Escribí el número en el sobre.\n(Natural clear speech.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Natural clear speech.) [3.2 seconds duration] Llámame cuando tengas un momento.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "deb8b54ea61b22c0daad82889ca5cf83b29825bf16174428d07e5974744f36bf", "generation_seconds_batch": 17.98059818998445, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8160512447357178, "target_z": 0.6562499650461685, "ce": 5.168776512145996, "genuineness": 4.65797233581543, "identity": 0.4463723599910736, "hypothesis": "Ah, escribí el número en el sobre. Léelo bien y comprueba las dos últimas cifras. Llámame cuando tenga su momento.", "wer": 0.15789473684210525, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/aac2ac5e2a6e5df320afbca9f7284cd386d3f0f5aa6e360abfc0986cbd7cc747.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "ab013dbe07bd536c153abb8369ba40c0fa0ae1599c77401a80c14d751a5b4faf", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6b16f53d452a022743dbea6016c81e8db69d819d3a766752083f9fb21cc9f871", "generation_seconds_batch": 21.179101658053696, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6653404235839844, "target_z": 1.187500032623576, "ce": 5.319687366485596, "genuineness": 4.6705241203308105, "identity": 0.7083823084831238, "hypothesis": "J'ai j'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ab013dbe07bd536c153abb8369ba40c0fa0ae1599c77401a80c14d751a5b4faf.mp3" }, { "id": "heldout-en-0", "language": "en", "sentences": [ "Your message arrived after I had left.", "I read it twice... before putting the phone away.", "Let me know when we can talk." ], "text": "Your message arrived after I had left. I read it twice... before putting the phone away. Let me know when we can talk.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-0", "key": "ab7a94299f968552a7b966fa3391724b652d8e20483888a7767196f29a7f4a69", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Your message arrived after I had left.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Let me know when we can talk.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Your message arrived after I had left.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Let me know when we can talk.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6b4f0845d707d75bf930be95b1a28b9e784157a03f0a256a204a90d308ca77c2", "generation_seconds_batch": 17.125993411056697, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.129962682723999, "target_z": -0.3984375133989687, "ce": 6.435632705688477, "genuineness": 4.15588903427124, "identity": 0.14684253931045532, "hypothesis": "Your message arrived after I had left. I read it twice, before putting the phone away. Let me know when we can talk.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ab7a94299f968552a7b966fa3391724b652d8e20483888a7767196f29a7f4a69.mp3" }, { "id": "heldout-es-2", "language": "es", "sentences": [ "La carretera gira a la izquierda tras el puente.", "Sigue... hasta la pequeña puerta de madera.", "Te esperaré al otro lado." ], "text": "La carretera gira a la izquierda tras el puente. Sigue... hasta la pequeña puerta de madera. Te esperaré al otro lado.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-2", "key": "ac02df05e588ebe73e31fe896ff83980ac5d73eecaa32dce186d425f789fdc0e", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Te esperaré al otro lado.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Natural clear speech.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Natural clear speech.) [3.2 seconds duration] Te esperaré al otro lado.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "9f895ab5d3408eae7ed8b4ae6b1c777b84ebb154b6f1fdab0ec62116b6f7e1a3", "generation_seconds_batch": 17.70909604500048, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1345348358154297, "target_z": 0.8554688335979136, "ce": 5.387199878692627, "genuineness": 4.319066047668457, "identity": 0.46244245767593384, "hypothesis": "La carretera gira a la izquierda tras el puente. Sigue hasta la pequeña puerta de madera te esperaré al otro lado.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ac02df05e588ebe73e31fe896ff83980ac5d73eecaa32dce186d425f789fdc0e.mp3" }, { "id": "heldout-en-0", "language": "en", "sentences": [ "Your message arrived after I had left.", "I read it twice... before putting the phone away.", "Let me know when we can talk." ], "text": "Your message arrived after I had left. I read it twice... before putting the phone away. Let me know when we can talk.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-0", "key": "ac3c9aec46c16dea12b7d8180a61cf704719dedc9b9a7f1e7e7e9fe84ab7200d", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Your message arrived after I had left.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Let me know when we can talk.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Your message arrived after I had left.\n(Natural clear speech.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Natural clear speech.) [3.2 seconds duration] Let me know when we can talk.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3b1712af425e2eabc185aba2c8b5f7abdc8b4a7453bca044e8c00f0b17de145b", "generation_seconds_batch": 23.622372666024603, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0908212661743164, "target_z": 0.8281249825230843, "ce": 5.518253803253174, "genuineness": 4.331617832183838, "identity": 0.42441511154174805, "hypothesis": "Your message arrived after I had left. I read it twice before putting the phone away. Let me know and we can talk.", "wer": 0.043478260869565216, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ac3c9aec46c16dea12b7d8180a61cf704719dedc9b9a7f1e7e7e9fe84ab7200d.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "ac44548a5d6c952b96064c3317e288bb8524635aca440a0e45e99d0686f95761", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b30af89ebf4fd2d615ddab3b53b7fd3a17ba816eabc4ee17b4050ae57b97acac", "generation_seconds_batch": 22.237728627049364, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4904866218566895, "target_z": 1.078124926596954, "ce": 5.681078910827637, "genuineness": 4.532451152801514, "identity": 0.31809741258621216, "hypothesis": "Ik heb het kleine doosje gevonden, het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ac44548a5d6c952b96064c3317e288bb8524635aca440a0e45e99d0686f95761.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "ac495e29f395e5ecf9dead267a4e5c3e1fd840ca7a2687a35a0984616fb8c123", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2fc29fb8bed7cab3c2ec7f28c20464520aa48f82b73156707c9e12d8c66e061f", "generation_seconds_batch": 16.156822989927605, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6241360902786255, "target_z": -0.7148437718461447, "ce": 6.300607204437256, "genuineness": 4.507347106933594, "identity": 0.6278847455978394, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ac495e29f395e5ecf9dead267a4e5c3e1fd840ca7a2687a35a0984616fb8c123.mp3" }, { "id": "heldout-es-3", "language": "es", "sentences": [ "Alguien ha movido la pequeña silla azul.", "Estaba aquí... junto a la estantería alta.", "La colocaremos después de comer." ], "text": "Alguien ha movido la pequeña silla azul. Estaba aquí... junto a la estantería alta. La colocaremos después de comer.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-3", "key": "ac941d03f134efe18231b47fe5c08768c2be2db0082e9a87177d481b722bebcf", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] La colocaremos después de comer.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Natural clear speech.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Natural clear speech.) [3.2 seconds duration] La colocaremos después de comer.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ed43ef555c2d1704263d950d526b7a25ce24c26326f7ee13e7fe3f5ee540dcf5", "generation_seconds_batch": 17.7595562429633, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7036454677581787, "target_z": 0.5859375460225448, "ce": 5.6413655281066895, "genuineness": 4.7834930419921875, "identity": 0.515130877494812, "hypothesis": "Alguien ha movido la pequeña silla azul. Estaba aquí, junto a la estantería alta. La colocaremos después de comer.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ac941d03f134efe18231b47fe5c08768c2be2db0082e9a87177d481b722bebcf.mp3" }, { "id": "heldout-nl-2", "language": "nl", "sentences": [ "De weg gaat linksaf na de brug.", "Loop verder... tot aan het kleine houten hek.", "Ik wacht aan de andere kant." ], "text": "De weg gaat linksaf na de brug. Loop verder... tot aan het kleine houten hek. Ik wacht aan de andere kant.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-2", "key": "ad3dfc1636c0bb4c230954fbb21d9bebc2bc38402b176c2197e5a57a36461161", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ik wacht aan de andere kant.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Natural clear speech.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Natural clear speech.) [3.2 seconds duration] Ik wacht aan de andere kant.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d94d65267cca9d60e721cce03ddfc31cb9e5df23b97dffe6733a5b9a56dc81b5", "generation_seconds_batch": 15.843171798973344, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3406121730804443, "target_z": 0.9843749848533397, "ce": 5.71682071685791, "genuineness": 4.425758361816406, "identity": 0.19032198190689087, "hypothesis": "De weg gaat linksaf naar de brug, loopt verder tot aan het kleine houten hek. Ik wacht aan de andere kant.", "wer": 0.09523809523809523, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ad3dfc1636c0bb4c230954fbb21d9bebc2bc38402b176c2197e5a57a36461161.mp3" }, { "id": "heldout-es-1", "language": "es", "sentences": [ "Las llaves están en la mesa de la cocina.", "Miré en el cajón... y luego junto a la ventana.", "Llévatelas cuando te vayas." ], "text": "Las llaves están en la mesa de la cocina. Miré en el cajón... y luego junto a la ventana. Llévatelas cuando te vayas.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-1", "key": "adb1941ab9a7cf95a202555a9cf0bfa45db75199a10f4a53e48f18b2663485da", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Llévatelas cuando te vayas.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Natural clear speech.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Natural clear speech.) [3.2 seconds duration] Llévatelas cuando te vayas.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "7b56127addfa23c853c744fdccb1887e9b8a9462642cabb0d39eeb3ee42b7f02", "generation_seconds_batch": 22.138351614936255, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9909048080444336, "target_z": 0.765624921936443, "ce": 5.188632965087891, "genuineness": 4.344170093536377, "identity": 0.5383808612823486, "hypothesis": "Las llaves están en la mesa de la cocina, miren el cajón y luego junto a la ventana, llévatelas cuando te vayas.", "wer": 0.08695652173913043, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/adb1941ab9a7cf95a202555a9cf0bfa45db75199a10f4a53e48f18b2663485da.mp3" }, { "id": "heldout-nl-0", "language": "nl", "sentences": [ "Je bericht kwam nadat ik vertrokken was.", "Ik las het twee keer... en legde mijn telefoon weg.", "Zeg maar wanneer we kunnen praten." ], "text": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer... en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-0", "key": "adcc0121d6c11de6953f8243f84ad890832f75283c96342c1db246fddc443438", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2ba0d0350c24bbacb139c7bb399776ec0385259bda503193587bd58f94e8fbc8", "generation_seconds_batch": 17.006464859005064, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.055861830711364746, "target_z": -1.0703124935917976, "ce": 6.1894097328186035, "genuineness": 4.375550270080566, "identity": 0.6791343092918396, "hypothesis": "Je bericht kwam nadat ik vertrokken was, ik las het twee keer en dachte mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "wer": 0.043478260869565216, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/adcc0121d6c11de6953f8243f84ad890832f75283c96342c1db246fddc443438.mp3" }, { "id": "heldout-en-2", "language": "en", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-2", "key": "ae3130a5db36c2790a31e23135fde257111c56fa9c5b8061c1f3e51a8cb61258", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The road turns left behind the bridge.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The road turns left behind the bridge.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I will meet you on the other side.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "47c047f0a5ee453f76d4a4eaa6e03c58d99f8196dc1e770fa75c9e159cac971c", "generation_seconds_batch": 17.8032191459788, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8427031636238098, "target_z": -0.5781250011651277, "ce": 6.435632705688477, "genuineness": 4.18099308013916, "identity": 0.4614142179489136, "hypothesis": "The road turns left behind the bridge. Keep going until you reach the small wooden gate. I will meet you on the other side.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ae3130a5db36c2790a31e23135fde257111c56fa9c5b8061c1f3e51a8cb61258.mp3" }, { "id": "heldout-de-1", "language": "de", "sentences": [ "Die Schlüssel liegen auf dem Küchentisch.", "Ich suchte in der Schublade... dann neben dem Fenster.", "Nimm sie mit, wenn du gehst." ], "text": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade... dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-1", "key": "ae603b8700cac226270099569e5aa2c3368e4d6274b624d5521e91c0f97dcf05", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Natural clear speech.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Natural clear speech.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "932553ee4108c1b501d7fc139bb6ad51030652e8d55de02631073fe2b394cd49", "generation_seconds_batch": 17.200157932937145, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.853519916534424, "target_z": 0.679687487766159, "ce": 5.593709468841553, "genuineness": 4.532451152801514, "identity": 0.6008461117744446, "hypothesis": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade, dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ae603b8700cac226270099569e5aa2c3368e4d6274b624d5521e91c0f97dcf05.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "aeb599189f56a825145340cc5b546808fcd3e351d4705edab01518db11907946", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c5976457706580d83bf2e58da42ba1e75183ccf57cf98f9c5685ed7c8e1f0e0b", "generation_seconds_batch": 17.6202357520815, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6201773881912231, "target_z": -0.09179685213436857, "ce": 6.316492557525635, "genuineness": 4.6203155517578125, "identity": 0.40961551666259766, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/aeb599189f56a825145340cc5b546808fcd3e351d4705edab01518db11907946.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "aebc4fbee7de5b46a558983591cba06c392c7b68aba46ac00ac7fd44af084e32", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4003263f18bc15c00a033bd98c8a5138a254b5a729449888b88421f72e6d2b96", "generation_seconds_batch": 22.59513175790198, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5279555320739746, "target_z": 1.101562598453292, "ce": 5.7128496170043945, "genuineness": 4.8588056564331055, "identity": 0.6226717233657837, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/aebc4fbee7de5b46a558983591cba06c392c7b68aba46ac00ac7fd44af084e32.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "b094d62a76ea30c700cd9ff3e26b6d2ec4936a165d2b43de24c38b706caf4cf5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b9a85cd831e8394c3e08498d3a4aa8f43470cf4c6f0060ad58cb944d75d34d2a", "generation_seconds_batch": 24.013636282994412, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0612702369689941, "target_z": -0.44140623048411076, "ce": 6.300607204437256, "genuineness": 4.632867813110352, "identity": 0.5374684929847717, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b094d62a76ea30c700cd9ff3e26b6d2ec4936a165d2b43de24c38b706caf4cf5.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "b17aa644465f84d971e1b4037bf43ff593f425ae3e8a5e8a892df6eac4dfb4fb", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "65baec93be733fd87628c023f2c47c31e0b87eb5f66dde2596682a41fc7c21d6", "generation_seconds_batch": 23.094897332950495, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7527670860290527, "target_z": 1.2421874365005394, "ce": 5.458683967590332, "genuineness": 4.758388519287109, "identity": 0.6631914377212524, "hypothesis": "Euh j'ai trouvé la petite boîte, elle était près de la vieille fenêtre, nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b17aa644465f84d971e1b4037bf43ff593f425ae3e8a5e8a892df6eac4dfb4fb.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "b1d79e63dd498bc17f47fca0a6ec9b23991dcde0391db667161717abff7c8442", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "027636ec53a5a73c2723ba8ab5d19f5d150dc7cad756e3a7b2230f6a2c510150", "generation_seconds_batch": 22.096404760028236, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1033108234405518, "target_z": 0.8359374900964144, "ce": 5.863759994506836, "genuineness": 4.557555198669434, "identity": 0.5008194446563721, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b1d79e63dd498bc17f47fca0a6ec9b23991dcde0391db667161717abff7c8442.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "b21149f4a1ae9fba0264526a40a20368ca42003b9b8f238e061074037c957e83", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "52034f8f5e9261aae76b1b82bbbc825c64919c3946adebe44fe87e1b596a190e", "generation_seconds_batch": 22.588062894064933, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2798372507095337, "target_z": -0.30468749708718074, "ce": 6.213237762451172, "genuineness": 4.570107460021973, "identity": 0.4357515573501587, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b21149f4a1ae9fba0264526a40a20368ca42003b9b8f238e061074037c957e83.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "b2121e50268d0459fd359397ee678b4e19e234c9dfa73b14723d25c99107bb80", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b766c4b6f16d40166978b19397d834a3d90760fb0d999fa9f16fc068a65c0a8f", "generation_seconds_batch": 21.56775121903047, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2094719409942627, "target_z": 0.9023437299015469, "ce": 5.486483573913574, "genuineness": 4.356721878051758, "identity": 0.4060172140598297, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude graam. We kunnen er samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b2121e50268d0459fd359397ee678b4e19e234c9dfa73b14723d25c99107bb80.mp3" }, { "id": "heldout-es-2", "language": "es", "sentences": [ "La carretera gira a la izquierda tras el puente.", "Sigue... hasta la pequeña puerta de madera.", "Te esperaré al otro lado." ], "text": "La carretera gira a la izquierda tras el puente. Sigue... hasta la pequeña puerta de madera. Te esperaré al otro lado.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-2", "key": "b25fb6ff43f600b151a1d4027d8e0ac824d7fdf9e6d78d7367f90ce1bd4d5482", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Te esperaré al otro lado.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Natural clear speech.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Natural clear speech.) [3.2 seconds duration] Te esperaré al otro lado.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "b2dbcc850c26312ee2b3a18022f909a0aa881cfceffcdacb2dd08483b44704dd", "generation_seconds_batch": 17.08183075103443, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.078331708908081, "target_z": 0.8203124749497541, "ce": 4.846105575561523, "genuineness": 4.319066047668457, "identity": 0.3841286301612854, "hypothesis": "La carretera gira a la izquierda tras el puente. Sigue hasta la pequeña puerta de madera. Te esperaré al otro lado.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b25fb6ff43f600b151a1d4027d8e0ac824d7fdf9e6d78d7367f90ce1bd4d5482.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "b2954d72f5bcefe679a5706a4202d82cfe33a8e55ee6b6e68fc767f3b79de65a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e5a681f54d66377b36c4e457bddd5f0d8e8f7a3e5b31c61abc28fdf30f30eb37", "generation_seconds_batch": 16.764624492963776, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.097902536392212, "target_z": 0.20703130155690144, "ce": 6.109982967376709, "genuineness": 4.645420074462891, "identity": 0.6791948080062866, "hypothesis": "La porte est encore ouverte , j'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b2954d72f5bcefe679a5706a4202d82cfe33a8e55ee6b6e68fc767f3b79de65a.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "b2d369879f7c3b63c88ca57dadf30eb2669e72e8254ccfa20e1722dc1b1ea9a9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cd22a19ba1a555f834dff3a154a6df293ab732229c881df6f29209e40321ab53", "generation_seconds_batch": 15.393032015068457, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7794190645217896, "target_z": 0.007812507573330156, "ce": 5.903473377227783, "genuineness": 4.1998209953308105, "identity": 0.5095645189285278, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b2d369879f7c3b63c88ca57dadf30eb2669e72e8254ccfa20e1722dc1b1ea9a9.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "b33e9ecab4d94103241183a0f332b47f467c4c2149fd0c0671fac36222e28cee", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a0ae6d734f2aea4d3d2153bd1de15af0aad6717906d504e05cc232b58156105e", "generation_seconds_batch": 23.18066223000642, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0158839225769043, "target_z": 0.7812499370831033, "ce": 5.8875885009765625, "genuineness": 4.532451152801514, "identity": 0.5269418954849243, "hypothesis": "Ik. Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b33e9ecab4d94103241183a0f332b47f467c4c2149fd0c0671fac36222e28cee.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "b3f52a68b637683d8fbbf87b2141dfe5af38a83774da4edaf9d7432a5b2fa001", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ec517a24497b65b71b751bd72768917def089250bc18e8cd3e2c3c5075e63909", "generation_seconds_batch": 16.60055090696551, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.34769344329834, "target_z": 0.3632813038871569, "ce": 5.895530700683594, "genuineness": 4.7458367347717285, "identity": 0.6297984719276428, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b3f52a68b637683d8fbbf87b2141dfe5af38a83774da4edaf9d7432a5b2fa001.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "b41346288617dc3c4195242151985214e5f67e8b5460caf5a85ab6ebe0a760aa", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4a7b0572c32b5137dff6565e50c5ccaf236ceade36456cccbc01b431e21bd40c", "generation_seconds_batch": 21.17429653799627, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2406959533691406, "target_z": 0.9218750734030461, "ce": 5.998785495758057, "genuineness": 4.406930446624756, "identity": 0.5424465537071228, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b41346288617dc3c4195242151985214e5f67e8b5460caf5a85ab6ebe0a760aa.mp3" }, { "id": "heldout-de-1", "language": "de", "sentences": [ "Die Schlüssel liegen auf dem Küchentisch.", "Ich suchte in der Schublade... dann neben dem Fenster.", "Nimm sie mit, wenn du gehst." ], "text": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade... dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-1", "key": "b429ad869b4f073af736e4e6510807542fc4f5324b0eb7f096c719d9002f4f71", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e022a15528c673ed3efe5b9c96af6e1641dbe51b20c224e95ea5f3b597946868", "generation_seconds_batch": 15.367235821904615, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.254858136177063, "target_z": -0.320312512233841, "ce": 6.3641486167907715, "genuineness": 4.519899368286133, "identity": 0.5716516375541687, "hypothesis": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade, dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b429ad869b4f073af736e4e6510807542fc4f5324b0eb7f096c719d9002f4f71.mp3" }, { "id": "heldout-es-4", "language": "es", "sentences": [ "Escribí el número en el sobre.", "Léelo bien... y comprueba las dos últimas cifras.", "Llámame cuando tengas un momento." ], "text": "Escribí el número en el sobre. Léelo bien... y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-4", "key": "b42ce0a44f667d143e6fb86c315e0a9a7a168d3a180716f883572764cc6d55ad", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Escribí el número en el sobre.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Llámame cuando tengas un momento.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Escribí el número en el sobre.\n(Natural clear speech.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Natural clear speech.) [3.2 seconds duration] Llámame cuando tengas un momento.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "43c10f3aea6cfae36620ad368dd583288b99a2fc93479a15280a7f0f1509fcd5", "generation_seconds_batch": 17.486024309066124, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5912394523620605, "target_z": 0.5156249778625734, "ce": 5.522225379943848, "genuineness": 4.632867813110352, "identity": 0.420767605304718, "hypothesis": "Escribí el número en el sobre, léalo bien y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b42ce0a44f667d143e6fb86c315e0a9a7a168d3a180716f883572764cc6d55ad.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "b4a7bcebb53f89fe556b3534bbcf00c30d230493d16ad7c382e19750f328901a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "575e472368b9e6ca150b58b3668887c7da217e17a453121f6371bce8f94c3ddc", "generation_seconds_batch": 21.12398576294072, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8410303592681885, "target_z": 0.6718749801928289, "ce": 5.39911413192749, "genuineness": 4.262581825256348, "identity": 0.4478387236595154, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b4a7bcebb53f89fe556b3534bbcf00c30d230493d16ad7c382e19750f328901a.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "b51d261b9988a197eb9160596ca6634b922fe0836f0c60b7565aa4d712bf8096", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "fffb9a20354969afe01d924bf5bbf83fa8f90df1ea7580e2f468dbf7a78011d3", "generation_seconds_batch": 21.623867515008897, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8824578523635864, "target_z": 0.07226565776921702, "ce": 6.435632705688477, "genuineness": 4.7081804275512695, "identity": 0.49362993240356445, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b51d261b9988a197eb9160596ca6634b922fe0836f0c60b7565aa4d712bf8096.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "b531228f140e5b7fce7ed38a0d8b776e16345bf08974571a9ee6154bc2732659", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5dd6793a7e8c572271bffc0fea6c6f4ce40483c6231b4984ee6138588adb986d", "generation_seconds_batch": 22.600584128056653, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6653404235839844, "target_z": 1.187500032623576, "ce": 5.506340026855469, "genuineness": 4.557555198669434, "identity": 0.6716949343681335, "hypothesis": "La porte est encore ouverte, j'ai poussé ton manteau sur la chaise, dis-moi quand tu es prêt.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b531228f140e5b7fce7ed38a0d8b776e16345bf08974571a9ee6154bc2732659.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "b558f4af4c2cce6bfb811a44490aa2489b77b190f946019cc022fd9fc84e77db", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "22446cd455043895d4002a484966256dfce6f5d9ffe6099da5579ba93ce83370", "generation_seconds_batch": 21.803624239051715, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7028088569641113, "target_z": 1.210937406207219, "ce": 5.673136234283447, "genuineness": 4.770940780639648, "identity": 0.8248999714851379, "hypothesis": "La porte est encore ouverte. j'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b558f4af4c2cce6bfb811a44490aa2489b77b190f946019cc022fd9fc84e77db.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "b5dae44ddc429ca054dd66780968ef7a3652d7a481c05037eb2fdf00dd6c56e1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2839c02bea40f29908246e6c87bb948eb190261f36bbf06997d4eb93a500fd99", "generation_seconds_batch": 22.650421111960895, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0595974922180176, "target_z": 0.8085937881579327, "ce": 5.6254801750183105, "genuineness": 4.419482707977295, "identity": 0.20346654951572418, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b5dae44ddc429ca054dd66780968ef7a3652d7a481c05037eb2fdf00dd6c56e1.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "b6014bfee8397d32bfc20b1017f0a9eaf745660c6ccf391b2f8bce3fb105193a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f6eb01aecb85db663f17ee0a79d26c49d37e2375ec74feda2298a25c4d994748", "generation_seconds_batch": 21.1595835769549, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.494445562362671, "target_z": 0.45507815602152546, "ce": 5.804190158843994, "genuineness": 4.65797233581543, "identity": 0.580016553401947, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b6014bfee8397d32bfc20b1017f0a9eaf745660c6ccf391b2f8bce3fb105193a.mp3" }, { "id": "heldout-fr-3", "language": "fr", "sentences": [ "Quelqu’un a déplacé la petite chaise bleue.", "Elle était ici... près de la grande bibliothèque.", "Nous la remettrons après le déjeuner." ], "text": "Quelqu’un a déplacé la petite chaise bleue. Elle était ici... près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-3", "key": "b6062007756ef1148402bbc05f14b59f04d3004d57accd651d9befe715f8cbf1", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Natural clear speech.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Natural clear speech.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "172e28c5e6689615877f424622375cfef7e3c52e790d6c954bb032c096927eb0", "generation_seconds_batch": 17.24302426399663, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.928457260131836, "target_z": 0.72656253320614, "ce": 5.272031307220459, "genuineness": 4.482243061065674, "identity": 0.7020612955093384, "hypothesis": "Quelqu'un a déplacé la petite chaise bleue elle était ici, près de la grande bibliothèque, nous la remettrons après le déjeuner.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b6062007756ef1148402bbc05f14b59f04d3004d57accd651d9befe715f8cbf1.mp3" }, { "id": "heldout-en-2", "language": "en", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-2", "key": "b66c5abcc3cc22ecb3af8203377f57717ffe14b9421d307a0efb966b63d8d85e", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The road turns left behind the bridge.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The road turns left behind the bridge.\n(Natural clear speech.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Natural clear speech.) [3.2 seconds duration] I will meet you on the other side.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4e7839fe54222fa3ce869cbd003b8f0f76bdeffa0a8d5a8b40b0935764c4ff0a", "generation_seconds_batch": 16.441124774049968, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9596810340881348, "target_z": 0.7460937275712914, "ce": 6.1020402908325195, "genuineness": 4.2374773025512695, "identity": 0.28170305490493774, "hypothesis": "The road turns left behind the bridge. Keep going until you reach the small wooden gate. I will meet you on the other side.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b66c5abcc3cc22ecb3af8203377f57717ffe14b9421d307a0efb966b63d8d85e.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "b68b2559f4be535fc3bc40f0841bb7058577e4763cb2b132f7bbb833fbcf1d73", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "708a9ccaa4ed4fd362c701b10eb9df4d9afa19975b5b77bec7cb5d64fc37f188", "generation_seconds_batch": 22.96275072300341, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.3173059225082397, "target_z": -0.28124997436719024, "ce": 6.3641486167907715, "genuineness": 4.369274139404297, "identity": 0.3927912712097168, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b68b2559f4be535fc3bc40f0841bb7058577e4763cb2b132f7bbb833fbcf1d73.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "b6d951d0f1f273251b9dbb1f5443eb059d3bf2fd6d2ae74c467f04ce08bee56e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "895f314ad5de66eeeaab1fb53360e54a83790f45c16acdba91695f2c493cc1f9", "generation_seconds_batch": 17.07272485306021, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.328122615814209, "target_z": 0.9765624772800096, "ce": 6.006728172302246, "genuineness": 4.432034492492676, "identity": 0.4323132634162903, "hypothesis": "Ik heb het een kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b6d951d0f1f273251b9dbb1f5443eb059d3bf2fd6d2ae74c467f04ce08bee56e.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "b722537f354ff21442949daad1425e74e67155d5e51a3e9163bef2076d13079a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d4e5fb8cf67fd378805b3672e1f676f2ac86cb47dd60016e31e427303f2cec3f", "generation_seconds_batch": 23.5427127830917, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0658421516418457, "target_z": 0.8124999673764239, "ce": 6.0305562019348145, "genuineness": 4.482243061065674, "identity": 0.3779272437095642, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b722537f354ff21442949daad1425e74e67155d5e51a3e9163bef2076d13079a.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "b7de978e1ed7772805c898dc86f0900f06eb8e7f323283ca9665cb977835f7aa", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b55dbce6be60f81a11fbdfd11a24aed502a5a61b384704f1e1a1d601e2b6a4ce", "generation_seconds_batch": 21.13660669210367, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1282899379730225, "target_z": 0.8515625052430748, "ce": 5.605623245239258, "genuineness": 4.595211505889893, "identity": 0.15029951930046082, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b7de978e1ed7772805c898dc86f0900f06eb8e7f323283ca9665cb977835f7aa.mp3" }, { "id": "heldout-es-3", "language": "es", "sentences": [ "Alguien ha movido la pequeña silla azul.", "Estaba aquí... junto a la estantería alta.", "La colocaremos después de comer." ], "text": "Alguien ha movido la pequeña silla azul. Estaba aquí... junto a la estantería alta. La colocaremos después de comer.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-3", "key": "b855aea7552bdd9051e90adc850424560c2eeb8b17b8388f6626c3c0c9639fe6", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] La colocaremos después de comer.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Natural clear speech.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Natural clear speech.) [3.2 seconds duration] La colocaremos después de comer.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c9641c600467b5a78223827a724855c90b795a48c95af5f8dae530dd1c08900a", "generation_seconds_batch": 18.051658822922036, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1657586097717285, "target_z": 0.8750000279630652, "ce": 5.260117053985596, "genuineness": 4.469690799713135, "identity": 0.5411046743392944, "hypothesis": "Alguien ha movido la pequeña silla azul. Estaba aquí, junto a la estantería alta. La colocaremos después de comer.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b855aea7552bdd9051e90adc850424560c2eeb8b17b8388f6626c3c0c9639fe6.mp3" }, { "id": "heldout-es-4", "language": "es", "sentences": [ "Escribí el número en el sobre.", "Léelo bien... y comprueba las dos últimas cifras.", "Llámame cuando tengas un momento." ], "text": "Escribí el número en el sobre. Léelo bien... y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-4", "key": "b8b23707e4af73a8f91bbeb58b27fec14575aa5424df5f6fdd629f654c47260f", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Escribí el número en el sobre.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Llámame cuando tengas un momento.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Escribí el número en el sobre.\n(Natural clear speech.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Natural clear speech.) [3.2 seconds duration] Llámame cuando tengas un momento.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cc69eb2db3f6981fb960d99856617e213b3d4689c2f847e470adce23f213d9bf", "generation_seconds_batch": 17.486024309066124, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5506484508514404, "target_z": 0.4902343655333373, "ce": 5.736677646636963, "genuineness": 4.6203155517578125, "identity": 0.5284570455551147, "hypothesis": "Escribí el número en el sobre. Léelo bien y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b8b23707e4af73a8f91bbeb58b27fec14575aa5424df5f6fdd629f654c47260f.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "b8c5c47e0eab2c4035fc7dad02f3800301eca63121a3a3516faded770e3f8cff", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2297fd5aac07a6b08cfb4b800e8de532791d6510465391b8f1cd81418283d347", "generation_seconds_batch": 22.50890007801354, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.960517406463623, "target_z": 0.12109371825026974, "ce": 5.649308204650879, "genuineness": 4.6705241203308105, "identity": 0.6685783863067627, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b8c5c47e0eab2c4035fc7dad02f3800301eca63121a3a3516faded770e3f8cff.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "b8ef7101a08624231ab7efb6a4f7f2e8ebe01770094ed334cceb8aa4a71957d0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ffbc0f5b62dbcb5ff9fbfc590ddf6c3a96d8b988d373bfdf816ffc49760d0c70", "generation_seconds_batch": 16.837456090957858, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7527670860290527, "target_z": 1.2421874365005394, "ce": 5.665193557739258, "genuineness": 4.720732688903809, "identity": 0.7166337966918945, "hypothesis": "La porte est encore ouverte, j'ai posé ton manteau sur la chaise, dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b8ef7101a08624231ab7efb6a4f7f2e8ebe01770094ed334cceb8aa4a71957d0.mp3" }, { "id": "heldout-es-1", "language": "es", "sentences": [ "Las llaves están en la mesa de la cocina.", "Miré en el cajón... y luego junto a la ventana.", "Llévatelas cuando te vayas." ], "text": "Las llaves están en la mesa de la cocina. Miré en el cajón... y luego junto a la ventana. Llévatelas cuando te vayas.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-1", "key": "b91ce932f33e7683d433d56e1213b838da1326a9bcee8885652244f316e442ec", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Llévatelas cuando te vayas.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Natural clear speech.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Natural clear speech.) [3.2 seconds duration] Llévatelas cuando te vayas.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "da161bc5e3b27ad435ecf5c267ddc4a07bfc815a5fc3742547196ae69f927b1d", "generation_seconds_batch": 18.38716280902736, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.29689884185791, "target_z": 0.957031282914858, "ce": 5.411027908325195, "genuineness": 4.262581825256348, "identity": 0.4473658800125122, "hypothesis": "A las llaves están en la mesa de la cocina, miren el cajón y luego junto a la ventana, llévatelas cuando te vayas.", "wer": 0.13043478260869565, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/b91ce932f33e7683d433d56e1213b838da1326a9bcee8885652244f316e442ec.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "ba3a0a2eeeaf21d4de80f0999bbef66b1204a711c8792841dd80bd99e0c3ff2e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c1b5e3ecbc2da3624d4d28d929060571f6279ad1b0de22cc90a7f16c913cb31e", "generation_seconds_batch": 22.680698742973618, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.365591287612915, "target_z": 1.0, "ce": 5.411027908325195, "genuineness": 4.68307638168335, "identity": 0.4352918267250061, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ba3a0a2eeeaf21d4de80f0999bbef66b1204a711c8792841dd80bd99e0c3ff2e.mp3" }, { "id": "heldout-fr-1", "language": "fr", "sentences": [ "Les clés sont sur la table de la cuisine.", "J’ai regardé dans le tiroir... puis près de la fenêtre.", "Prends-les quand tu partiras." ], "text": "Les clés sont sur la table de la cuisine. J’ai regardé dans le tiroir... puis près de la fenêtre. Prends-les quand tu partiras.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-1", "key": "ba70e7e208a6b15a9236ccd5a69e0e5fd403df2a937202d690effa2cdc45d9f2", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Prends-les quand tu partiras.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Prends-les quand tu partiras.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "fbec98e96f4f22d4a580a433c277c6f108d5e008e5403f590e399ffc1f9374e1", "generation_seconds_batch": 17.377614044002257, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9800881743431091, "target_z": -0.49218749242666987, "ce": 6.348263263702393, "genuineness": 4.482243061065674, "identity": 0.7471933364868164, "hypothesis": "Les clés sont sur la table de la cuisine. J'ai regardé dans le tiroir, puis près de la fenêtre. Prends-les quand tu partiras.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ba70e7e208a6b15a9236ccd5a69e0e5fd403df2a937202d690effa2cdc45d9f2.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "ba88932728d51b682fbaa63b583b8c13f12cefb8029b97197558ac3562992d56", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] He encontrado la cajita.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a6a4d0cd071c116d5a3bb0ba575f632f73075a9c969f3aaaf745a315a66e6ea1", "generation_seconds_batch": 17.506156646064483, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9105592966079712, "target_z": 0.08984376252512295, "ce": 6.316492557525635, "genuineness": 4.6956281661987305, "identity": 0.5859376192092896, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ba88932728d51b682fbaa63b583b8c13f12cefb8029b97197558ac3562992d56.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "bb52a606f7b05451e8062eda5ab9a8c7df3fc2e290c732888a7decb7b41c7bd0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "08f36ab26ec68c9c188a0ac74291afe0b41d0e8aebb8d6f72610ff7eb9d0ed21", "generation_seconds_batch": 22.237728627049364, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.321877956390381, "target_z": 0.9726562980615183, "ce": 6.078212261199951, "genuineness": 4.570107460021973, "identity": 0.40121227502822876, "hypothesis": "Ik heb het... Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.16666666666666666, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bb52a606f7b05451e8062eda5ab9a8c7df3fc2e290c732888a7decb7b41c7bd0.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "bb8ce2b6b507c2a950528b235cb23524c66597a521ef1ff5ae3b49702f216792", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] I found the little box.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] It was beside the old window.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7581f4974113581469d3058dd73700973a289144efa23df537c54ae0fadb9507", "generation_seconds_batch": 18.140566439018585, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5546072721481323, "target_z": -0.13281247961026496, "ce": 6.141753673553467, "genuineness": 4.337893962860107, "identity": 0.4262309670448303, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bb8ce2b6b507c2a950528b235cb23524c66597a521ef1ff5ae3b49702f216792.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "bbaeaf8e752ffe50e645949ce8ee94cdd7090d2502968e7ab9f89063381d3af5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6d1edf431d8f3f98208912c9d01dbcd084efefde4d459108cc56ddefaf177684", "generation_seconds_batch": 16.805090085952543, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.715298652648926, "target_z": 1.2187500629168966, "ce": 5.395142555236816, "genuineness": 4.419482707977295, "identity": 0.5408597588539124, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bbaeaf8e752ffe50e645949ce8ee94cdd7090d2502968e7ab9f89063381d3af5.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "bc1165d658d1274203892b4a12ab7544a7561ea118f9a611e929cdc2bec2b7dc", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "9dded9ce6f96d1512fe707b595c4b96cfc03f861f66d3879927710e70d61f1cb", "generation_seconds_batch": 22.827060448005795, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4476096630096436, "target_z": 0.42578121533745045, "ce": 5.681078910827637, "genuineness": 4.796044826507568, "identity": 0.45574188232421875, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bc1165d658d1274203892b4a12ab7544a7561ea118f9a611e929cdc2bec2b7dc.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "bc552d0372942542e6a3cac099be4bf60c928cd787f6c5a37b33f4257becfa01", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "50f9cdeabf139531d28288252e060e85bee1ac76946494701d4e8107ae91c505", "generation_seconds_batch": 23.787151047959924, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0519030094146729, "target_z": -0.4472656484481953, "ce": 6.260893821716309, "genuineness": 4.507347106933594, "identity": 0.6375567317008972, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bc552d0372942542e6a3cac099be4bf60c928cd787f6c5a37b33f4257becfa01.mp3" }, { "id": "heldout-en-1", "language": "en", "sentences": [ "The keys are on the kitchen table.", "I checked the drawer... then looked beside the window.", "Take them with you when you leave." ], "text": "The keys are on the kitchen table. I checked the drawer... then looked beside the window. Take them with you when you leave.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-1", "key": "bc588cc972f84095f13f7d4a9889376eb56469b74008721ab034cc5938ccee82", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The keys are on the kitchen table.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Take them with you when you leave.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The keys are on the kitchen table.\n(Natural clear speech.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Natural clear speech.) [3.2 seconds duration] Take them with you when you leave.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d8639eeecd423d03aa5181ced5d57921df9287baf0698b0f4d28c316f6ad8eb3", "generation_seconds_batch": 23.266846791026182, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9909048080444336, "target_z": 0.765624921936443, "ce": 5.736677646636963, "genuineness": 4.344170093536377, "identity": 0.27803629636764526, "hypothesis": "The keys are on the kitchen table. I checked the drawer, then looked beside the window. Take them with you when you leave.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bc588cc972f84095f13f7d4a9889376eb56469b74008721ab034cc5938ccee82.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "bc7b08460e28da102a3a88a926322091a286d02dfc9d41030a7b3324287ad1b2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5a3804f38ff7e62e46a1922fdb777bf697475850c7e9481ecd2743194a3f343c", "generation_seconds_batch": 22.307957099052146, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4725887775421143, "target_z": 0.44140623048411076, "ce": 5.935244083404541, "genuineness": 4.632867813110352, "identity": 0.6700050830841064, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bc7b08460e28da102a3a88a926322091a286d02dfc9d41030a7b3324287ad1b2.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "bc85bf3786be8c36df993fe2ac89f73708e6f4f5fe2f6df83e7c16685adf3bf3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5a40061b3707288535631bf88ecf69f2f8e43f5fbb7397185a4326f8b04cc6cd", "generation_seconds_batch": 23.682209418038838, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.253185272216797, "target_z": 0.9296874318400286, "ce": 4.959288597106934, "genuineness": 4.218649387359619, "identity": 0.3252139091491699, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bc85bf3786be8c36df993fe2ac89f73708e6f4f5fe2f6df83e7c16685adf3bf3.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "bceda968b6d5fd7520330cf05a4488fa174f1b580d6b144cc028a2e1601d8f57", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "90919ff2a4e950de9ab53155fb38f19112a7105683b66de1e0c677876741ceed", "generation_seconds_batch": 20.063229785999283, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.2009410858154297, "target_z": 0.27148430261644063, "ce": 5.919358730316162, "genuineness": 4.369274139404297, "identity": 0.5095064043998718, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bceda968b6d5fd7520330cf05a4488fa174f1b580d6b144cc028a2e1601d8f57.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "bd49a45fc54c0eb66c7820770e3ac3adf3a811bdf52942b1cab436c61985fe8f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "fd978a05be72ce00f55f977afaaca1d0c4429568c2335f1f7807aa6fe75ee410", "generation_seconds_batch": 22.4401689710794, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1603500843048096, "target_z": 0.24609369028720454, "ce": 5.982900142669678, "genuineness": 4.482243061065674, "identity": 0.6344577074050903, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bd49a45fc54c0eb66c7820770e3ac3adf3a811bdf52942b1cab436c61985fe8f.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "bd49c01e3f827a51123055b4f860f2e2f8f29be1747435856d7c75958630fdd6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0b6bb953ebb1f5a4efd529b607ad77bcce7352b938a7a1ec282bddbaac83e2b8", "generation_seconds_batch": 21.92380141292233, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9886189699172974, "target_z": 0.1386718975743495, "ce": 6.125868320465088, "genuineness": 4.7332844734191895, "identity": 0.7648621797561646, "hypothesis": "Ich habe die kleine Schacht gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0625, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bd49c01e3f827a51123055b4f860f2e2f8f29be1747435856d7c75958630fdd6.mp3" }, { "id": "heldout-nl-0", "language": "nl", "sentences": [ "Je bericht kwam nadat ik vertrokken was.", "Ik las het twee keer... en legde mijn telefoon weg.", "Zeg maar wanneer we kunnen praten." ], "text": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer... en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-0", "key": "bd4cd02f0de879aa0a0edf3750e24a28d0def37270b93eacd0356cf05fc8a819", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Natural clear speech.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ca0ba052959b08862af8bbd44182df8d9ff8a4d27c061aad910db7447c6bc943", "generation_seconds_batch": 23.283461470971815, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2656750679016113, "target_z": 0.9375000885497065, "ce": 5.5539960861206055, "genuineness": 4.362998008728027, "identity": 0.24375595152378082, "hypothesis": "Je bericht kwam nadat ik vertrokken was. Ik las twee keer en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "wer": 0.043478260869565216, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bd4cd02f0de879aa0a0edf3750e24a28d0def37270b93eacd0356cf05fc8a819.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "bd5074045c12559e71a4a6a61c212e348b546e9a6fa859f456cdcb0bac926141", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "828d7f862d63fc1eef13620142ee3a56297a5585ee1da486a4be5878a604034b", "generation_seconds_batch": 22.267045760992914, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.365591287612915, "target_z": 1.0, "ce": 5.6334228515625, "genuineness": 4.532451152801514, "identity": 0.4906514883041382, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bd5074045c12559e71a4a6a61c212e348b546e9a6fa859f456cdcb0bac926141.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "bd5a0afd62549e2bc4a1c11ce0127b6af206166095cd9e64f2c649d07a0eb565", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1c17b1e26a6bc3b48be88505af6f5f5be0a38f66467b632a84c4415b351aa32d", "generation_seconds_batch": 21.436049563926645, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.965925693511963, "target_z": 0.7499999067897827, "ce": 5.7247633934021, "genuineness": 4.557555198669434, "identity": 0.44173532724380493, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bd5a0afd62549e2bc4a1c11ce0127b6af206166095cd9e64f2c649d07a0eb565.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "bd8ea5486e922ac720385dd9871a0b5186a1cdef78289795cbb5f2441a1b68de", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "00157cebc88422309e92164f9186d39ee82d97ddd5f1f16139130192ae5031c1", "generation_seconds_batch": 17.137505569960922, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5225470066070557, "target_z": 0.4726562607774314, "ce": 5.895530700683594, "genuineness": 4.808597087860107, "identity": 0.5922227501869202, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bd8ea5486e922ac720385dd9871a0b5186a1cdef78289795cbb5f2441a1b68de.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "be679a891152c1d74896d4e66aeb803e49c163508733cd7af1438c474ec83a18", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "39b50e4fa0ebf8eda83c3a46c3b089521c9b9ed5ebea5c04af3774ed102859c2", "generation_seconds_batch": 22.512820691918023, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.545240044593811, "target_z": -0.1386718975743495, "ce": 5.974957466125488, "genuineness": 4.65797233581543, "identity": 0.45356321334838867, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/be679a891152c1d74896d4e66aeb803e49c163508733cd7af1438c474ec83a18.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "be74112a23c3084e3935c1710c57edc32803812f6b7b5d602a6b85c5f1efd302", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3af6025eea301d6cd4bbb1cd249a88e9d4972eff58dbe82dd0f14a9d3ebb3a39", "generation_seconds_batch": 23.438289384008385, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.0275373458862305, "target_z": 1.4140626031138028, "ce": 5.478540420532227, "genuineness": 4.7834930419921875, "identity": 0.7523610591888428, "hypothesis": "La porte est encore ouverte, j'ai poussé ton manteau sur la chaise, dis-moi quand tu es prêt.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/be74112a23c3084e3935c1710c57edc32803812f6b7b5d602a6b85c5f1efd302.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "bf39eee1298b30e4564b9ea3416f30b454541275ffc7fb28ac3db0db584a84d6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "b92f29ac4105dfb91d502dec0af19adef5a08e7a1ee783b384739b11f397972e", "generation_seconds_batch": 16.406632611062378, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9043145179748535, "target_z": 0.08593750873845787, "ce": 6.435632705688477, "genuineness": 4.758388519287109, "identity": 0.47899341583251953, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bf39eee1298b30e4564b9ea3416f30b454541275ffc7fb28ac3db0db584a84d6.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "bfa09b3f776d430b2b6e42efb4fab728d2936044c96e882e0ab99e95729be1e0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1702b325763f2509166109d663e73080acb40f22b3678ff5df301c1bba649690", "generation_seconds_batch": 16.707319209934212, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.934701919555664, "target_z": 0.7304687124246312, "ce": 5.855817794799805, "genuineness": 4.6956281661987305, "identity": 0.7930183410644531, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bfa09b3f776d430b2b6e42efb4fab728d2936044c96e882e0ab99e95729be1e0.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "bfad9d0124024654397dadc71ff04f7fd275c2739ecfc2edd9ebcea4ce3054f1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b00a0327fd2edda0b026264d9e00a2a76a47a631972c219ff68007ab0bf5ac3d", "generation_seconds_batch": 23.974819825962186, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6974005699157715, "target_z": 0.5820312176677058, "ce": 5.808161735534668, "genuineness": 4.8588056564331055, "identity": 0.8459395170211792, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bfad9d0124024654397dadc71ff04f7fd275c2739ecfc2edd9ebcea4ce3054f1.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "bfc960898de28e8bcab6971bec6299874eb2b1fb7b7ea59cbb41743b63ae844a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7975ae9328565749353858064cb0cfcc0a09e1283fd001c386d5a4edf2f4a70e", "generation_seconds_batch": 21.179101658053696, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.902641773223877, "target_z": 1.3359375273805014, "ce": 5.287916660308838, "genuineness": 4.362998008728027, "identity": 0.8250266909599304, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bfc960898de28e8bcab6971bec6299874eb2b1fb7b7ea59cbb41743b63ae844a.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "bfd753118c8c0cbc4e848aa5f28b2397570442a90ce400f9735a8530d6530b0c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5c77e1dab7c3b98141f3bc04a03e1ee26fa487485b7234afa7841dc5e13d6477", "generation_seconds_batch": 17.455278169945814, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8098065853118896, "target_z": 0.6523437858276773, "ce": 5.911416053771973, "genuineness": 4.582659721374512, "identity": 0.5265432596206665, "hypothesis": "Ik heb het kleine doosje gevonden, het stond naast het oude ramen. We kunnen er samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/bfd753118c8c0cbc4e848aa5f28b2397570442a90ce400f9735a8530d6530b0c.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "c001658f1cae86da32e9baab4095f3c0398b9c88c4116c3c03160050a90fa402", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "994eed997ab776710cc54b62a52bea3be59db34bc207a19bf826e3cc97bba699", "generation_seconds_batch": 22.1596444620518, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.350815773010254, "target_z": 0.3652343934964025, "ce": 5.903473377227783, "genuineness": 4.394378662109375, "identity": 0.6413227319717407, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c001658f1cae86da32e9baab4095f3c0398b9c88c4116c3c03160050a90fa402.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "c01c8680c3d1655fd5c7010d88698eb694274a50fe5296c2cc381c2781400c8d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "743da25dc97aecc9d83b48ddde9fd50372cf2e5fadb7ca38b8d05921dadeddd8", "generation_seconds_batch": 22.65826919605024, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.51546573638916, "target_z": 1.0937499417436143, "ce": 5.689021587371826, "genuineness": 4.356721878051758, "identity": 0.3578408658504486, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c01c8680c3d1655fd5c7010d88698eb694274a50fe5296c2cc381c2781400c8d.mp3" }, { "id": "heldout-de-3", "language": "de", "sentences": [ "Jemand hat den kleinen blauen Stuhl umgestellt.", "Er stand hier... neben dem hohen Bücherregal.", "Wir stellen ihn nach dem Essen zurück." ], "text": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier... neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-3", "key": "c10e0a7e12ebf9a905f1d366f0199504ace61e748a17932f0704c4642aa2207d", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Natural clear speech.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Natural clear speech.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "549cd4e3379a80e1cd5d24b4fc68f6dfdcbe6f51b66a6f1886279484d5be6ac1", "generation_seconds_batch": 16.936351155978628, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.422630548477173, "target_z": 0.41015620019079013, "ce": 5.577824115753174, "genuineness": 4.582659721374512, "identity": 0.4996691942214966, "hypothesis": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c10e0a7e12ebf9a905f1d366f0199504ace61e748a17932f0704c4642aa2207d.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "c13802b52b56672fcd7421fa387bdad3377e370fb9eb9ccb3dc211196a08735e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "5bd6f5cce0915a10de9903af62bbb4b5375629e39cee323de01073b190660533", "generation_seconds_batch": 24.090297435992397, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.485914707183838, "target_z": -0.1757812712635808, "ce": 6.070269584655762, "genuineness": 3.9927117824554443, "identity": 0.27900269627571106, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c13802b52b56672fcd7421fa387bdad3377e370fb9eb9ccb3dc211196a08735e.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "c18e6362a6bf1864c7a67c51a93b0985325a6527c5209ce00437d069ea7fbeb9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8b16b37a0ea64f47122a525ab3542d14c988e5161fd4fbfa942994ceffc7f027", "generation_seconds_batch": 17.70679973694496, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.784827470779419, "target_z": 0.636718770681017, "ce": 6.062326908111572, "genuineness": 4.7834930419921875, "identity": 0.7127952575683594, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c18e6362a6bf1864c7a67c51a93b0985325a6527c5209ce00437d069ea7fbeb9.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "c1dd5b7383a6dfab0a4c489e0b53ac0fe31afe0171bfcaf4f2fc1d8b631344be", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d4eac0b92c7a92b2420b054ffb7820b4f1201aa06d8e90abfc0197f8d526ad7f", "generation_seconds_batch": 22.669804868055508, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0541889667510986, "target_z": 0.17968745048207205, "ce": 6.125868320465088, "genuineness": 4.65797233581543, "identity": 0.6269503831863403, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c1dd5b7383a6dfab0a4c489e0b53ac0fe31afe0171bfcaf4f2fc1d8b631344be.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "c1efb5bafb7adbb7aa764ce284ac27577a1544c8bbfb7300815e4927985f2275", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6191efacabbc7b5c9979f89f5e16eb3643fc06e9a37ca64347413905c63e7364", "generation_seconds_batch": 22.59513175790198, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6653404235839844, "target_z": 1.187500032623576, "ce": 5.685050010681152, "genuineness": 4.7332844734191895, "identity": 0.8661136627197266, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c1efb5bafb7adbb7aa764ce284ac27577a1544c8bbfb7300815e4927985f2275.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "c22fc747e2d6cfd2111db6eb70753722fee0ef77c9a7a65140c9e712f4668c95", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "2d3b06a68a4fab738afbd6ebf5762b2cf9a57a98aae2d05a943c140c33e5ac82", "generation_seconds_batch": 27.172643971978687, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3905704021453857, "target_z": 1.0156250151466604, "ce": 5.982900142669678, "genuineness": 4.65797233581543, "identity": 0.5444421768188477, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c22fc747e2d6cfd2111db6eb70753722fee0ef77c9a7a65140c9e712f4668c95.mp3" }, { "id": "heldout-es-2", "language": "es", "sentences": [ "La carretera gira a la izquierda tras el puente.", "Sigue... hasta la pequeña puerta de madera.", "Te esperaré al otro lado." ], "text": "La carretera gira a la izquierda tras el puente. Sigue... hasta la pequeña puerta de madera. Te esperaré al otro lado.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-2", "key": "c2967acb3c8fe94e847606436fdd717f00f520a587143154e4e699c7e6ed8717", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Te esperaré al otro lado.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Natural clear speech.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Natural clear speech.) [3.2 seconds duration] Te esperaré al otro lado.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "086e48fb2668998b83f83f711a5db342bbbbd19516bed1b506c6091c04e526e9", "generation_seconds_batch": 17.70909604500048, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9097228050231934, "target_z": 0.7148436972779708, "ce": 5.053607940673828, "genuineness": 4.325342178344727, "identity": 0.5185028910636902, "hypothesis": "La carretera gira a la izquierda tras el puente hasta la pequeña puerta de madera de esperar el otro lado.", "wer": 0.19047619047619047, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c2967acb3c8fe94e847606436fdd717f00f520a587143154e4e699c7e6ed8717.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "c2b8dd4917b310960483720ae62c6ef70dffa54f66f6f390334a64167de03c80", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b9e8bbda11a879d4a98bdddf9fe50c383dbe29f92dfbf2fba364dd2f6bdefdb7", "generation_seconds_batch": 22.076076486031525, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.072086811065674, "target_z": 0.8164061465949152, "ce": 5.550024509429932, "genuineness": 4.7458367347717285, "identity": 0.5493096113204956, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c2b8dd4917b310960483720ae62c6ef70dffa54f66f6f390334a64167de03c80.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "c319612f18fc7f4a1194a1a26a50ef64d60a9bc3ad62caca3f39eaec76b77ff5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3fb4081ec62334b5f53cb469cff78a3c61eedbe028989178816f6b76c1d42b19", "generation_seconds_batch": 19.81498774397187, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.578749895095825, "target_z": 0.5078124702892433, "ce": 6.1020402908325195, "genuineness": 4.6705241203308105, "identity": 0.7115404605865479, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c319612f18fc7f4a1194a1a26a50ef64d60a9bc3ad62caca3f39eaec76b77ff5.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "c32769168afe42125bf3ecdefe636add5062847db80bfea8f5d3f78f42a4f176", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8171d3d15c458129d87ed1d983578042f6974b965adbced01fbdd0ffb24dd7d6", "generation_seconds_batch": 17.342588623054326, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5077714920043945, "target_z": -0.16210934572616612, "ce": 6.205295085906982, "genuineness": 4.6956281661987305, "identity": 0.8318143486976624, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c32769168afe42125bf3ecdefe636add5062847db80bfea8f5d3f78f42a4f176.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "c3969da6af8f47ec3619aec328ba2f3bbe028e77afc4a16c044f8be869767d9c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e5eb1e953486447189077bed970da8975500d6a2c0cbdc4854d85f24aad66317", "generation_seconds_batch": 23.013191931997426, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2344510555267334, "target_z": 0.9179687450482072, "ce": 5.6254801750183105, "genuineness": 4.344170093536377, "identity": 0.3486492335796356, "hypothesis": "Ik heb het kleine doosje gevonden, het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c3969da6af8f47ec3619aec328ba2f3bbe028e77afc4a16c044f8be869767d9c.mp3" }, { "id": "heldout-es-2", "language": "es", "sentences": [ "La carretera gira a la izquierda tras el puente.", "Sigue... hasta la pequeña puerta de madera.", "Te esperaré al otro lado." ], "text": "La carretera gira a la izquierda tras el puente. Sigue... hasta la pequeña puerta de madera. Te esperaré al otro lado.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-2", "key": "c3977db0634952a7f9e4083e6135fbbc87199d34c7cb8c3338515d9ae38c18f4", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Te esperaré al otro lado.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Te esperaré al otro lado.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4f36f3bd0767c3ed40747c3b15ed889513d0b44ca983a6155663a57e96fea338", "generation_seconds_batch": 16.231410023057833, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.48675107955932617, "target_z": -0.8007812805846025, "ce": 6.395919322967529, "genuineness": 4.306513786315918, "identity": 0.416146844625473, "hypothesis": "La carretera gira a la izquierda tras el puente. Sigue hasta la pequeña puerta de madera. Te esperaré al otro lado.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c3977db0634952a7f9e4083e6135fbbc87199d34c7cb8c3338515d9ae38c18f4.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "c3b3d5eec3985d887ffabbbd779ffeaaf65d0f9494cad9dd0984edf6010ef444", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "0924109b9dee0381c7d053718c2f452d098c3604327024af7fe06dadfa940849", "generation_seconds_batch": 21.984455551952124, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.866009473800659, "target_z": 0.6874999953394891, "ce": 5.8875885009765625, "genuineness": 4.432034492492676, "identity": 0.40686821937561035, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c3b3d5eec3985d887ffabbbd779ffeaaf65d0f9494cad9dd0984edf6010ef444.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "c3c123938576b87b575f06765238fde5f075123da95ba9d5fb6da7b9f3bea615", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "975334699ff034f25650034bfcb7407de9454873522f4c3454993c383c1a7cc0", "generation_seconds_batch": 22.159058392979205, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0845766067504883, "target_z": 0.8242188033045931, "ce": 5.5659098625183105, "genuineness": 4.425758361816406, "identity": 0.5035514235496521, "hypothesis": "De deur staat nog open. Ik heb je jas op een stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c3c123938576b87b575f06765238fde5f075123da95ba9d5fb6da7b9f3bea615.mp3" }, { "id": "heldout-nl-4", "language": "nl", "sentences": [ "Ik schreef het nummer op de envelop.", "Lees het goed... en controleer de laatste twee cijfers.", "Bel me wanneer je even tijd hebt." ], "text": "Ik schreef het nummer op de envelop. Lees het goed... en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-4", "key": "c454a26483dbb8d89d464d80365781d6b94db85838a99126faa4b0939b88537f", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Natural clear speech.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Natural clear speech.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "09e28e83111f911c8f7d055aed8fdcda0c10b6e2b63ed9d5824a45d506109e2e", "generation_seconds_batch": 17.79212440806441, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1541054248809814, "target_z": 0.2421875110687133, "ce": 5.700935363769531, "genuineness": 4.13078498840332, "identity": 0.6135193705558777, "hypothesis": "Ik s- ik schreef het nummer op aan de envelop, lees het goed en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "wer": 0.13043478260869565, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c454a26483dbb8d89d464d80365781d6b94db85838a99126faa4b0939b88537f.mp3" }, { "id": "heldout-nl-3", "language": "nl", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-3", "key": "c493619abd8899c9b3ce96f2a00a6bcdd900e874ef21436707b986365a8437ee", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We zetten het na de lunch terug.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "51d1fe86afb55a73249fefe4495b18c1f74b340a22ab6ffcfae07fb886a4bb79", "generation_seconds_batch": 16.50808106001932, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.11830949783325195, "target_z": -1.0312500302933205, "ce": 6.3641486167907715, "genuineness": 4.507347106933594, "identity": 0.764272928237915, "hypothesis": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier naast de hoge boekenkast. We zetten het na de lunch terug.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c493619abd8899c9b3ce96f2a00a6bcdd900e874ef21436707b986365a8437ee.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "c4e3218d163406ab57661bd58542bcfb057502b55ab17a1bd6c7950d02cd67a1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6ad0ca07a74c9b10de0a8116b8733f723ceb523fcd0969a60430e567cc48d3ef", "generation_seconds_batch": 16.60055090696551, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9471914768218994, "target_z": 0.7382812199979613, "ce": 5.40705680847168, "genuineness": 4.607763767242432, "identity": 0.7847495675086975, "hypothesis": "Die Tür ist noch offen. Ich hab deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c4e3218d163406ab57661bd58542bcfb057502b55ab17a1bd6c7950d02cd67a1.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "c4f628537d49167fca05d4a7c15929404ae346fd9da270c35b3b89b9d51d2e00", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0356c0b74d3873459e85ed53e7aae587d873af8ce20f1d5de4e67a49bd40e8da", "generation_seconds_batch": 17.251369801000692, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.003394603729248, "target_z": 0.7734375786461208, "ce": 5.450741291046143, "genuineness": 4.68307638168335, "identity": 0.49686485528945923, "hypothesis": "He encontrado la cajita, estaba junto a la ventana, vieja, podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c4f628537d49167fca05d4a7c15929404ae346fd9da270c35b3b89b9d51d2e00.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "c4fb4a741be6cfcba9029f48e3030a65089543cef52dbaadb389edd9b61ebb67", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "19e996c0b9fbf1c03b6b15466392f32e2207e1bdab39c4af9d04c6a418eb2607", "generation_seconds_batch": 21.241574181010947, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.915131092071533, "target_z": 1.343749885817484, "ce": 5.411027908325195, "genuineness": 4.557555198669434, "identity": 0.7726324796676636, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c4fb4a741be6cfcba9029f48e3030a65089543cef52dbaadb389edd9b61ebb67.mp3" }, { "id": "heldout-nl-4", "language": "nl", "sentences": [ "Ik schreef het nummer op de envelop.", "Lees het goed... en controleer de laatste twee cijfers.", "Bel me wanneer je even tijd hebt." ], "text": "Ik schreef het nummer op de envelop. Lees het goed... en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-4", "key": "c57efcec4253b6a7987ef58c3ed4e6ebc0c3e8322806ab634a0efc56862a683e", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Natural clear speech.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Natural clear speech.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "329725367268f2850a0f5211b300daf5a543251f7df44b4851d34085dbd9f48d", "generation_seconds_batch": 16.850471340003423, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.003394603729248, "target_z": 0.7734375786461208, "ce": 5.764476776123047, "genuineness": 4.331617832183838, "identity": 0.22807541489601135, "hypothesis": "Ik schreef het nummer op de envelop, lees het goed en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c57efcec4253b6a7987ef58c3ed4e6ebc0c3e8322806ab634a0efc56862a683e.mp3" }, { "id": "heldout-de-1", "language": "de", "sentences": [ "Die Schlüssel liegen auf dem Küchentisch.", "Ich suchte in der Schublade... dann neben dem Fenster.", "Nimm sie mit, wenn du gehst." ], "text": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade... dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-1", "key": "c5c4717fc5ab5cbbfc6edfda1fa75e01bceae00e04a3ca2b3465e4b387a05284", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Natural clear speech.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Natural clear speech.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "08c4c7e9b3cbaf2691213987e606ace31bd673fe19f1cc67b2e91c8fd83133ec", "generation_seconds_batch": 17.204288207925856, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.535036563873291, "target_z": 0.4804687683507615, "ce": 5.792276382446289, "genuineness": 4.6705241203308105, "identity": 0.6033540964126587, "hypothesis": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade. Dann nehm dem Fenster. Nimm sie mit, wenn du gehst.", "wer": 0.047619047619047616, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c5c4717fc5ab5cbbfc6edfda1fa75e01bceae00e04a3ca2b3465e4b387a05284.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "c5c8c11ac322f4c4e03ed7dc2ee6b392b0bc87d50d58cb59d7e48efd2a225ee0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "c05960de321901863e1e0e4b4e3f7012387b7e8418a9e181bc280e8f0238e7e8", "generation_seconds_batch": 16.14415013208054, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.153269052505493, "target_z": 0.867187520389735, "ce": 5.557967185974121, "genuineness": 4.400654315948486, "identity": 0.3031851053237915, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c5c8c11ac322f4c4e03ed7dc2ee6b392b0bc87d50d58cb59d7e48efd2a225ee0.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "c5ff974f2583b16293fc4d24efeb86345f745d3ce8baef1f4f2c2aa0fb429f82", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bf938615465d57fad0c814f0f94773730b434d692168f2e552fdedd2e03c1f75", "generation_seconds_batch": 16.961241595097817, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4921594858169556, "target_z": -0.17187501747691575, "ce": 5.974957466125488, "genuineness": 4.7458367347717285, "identity": 0.573531985282898, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c5ff974f2583b16293fc4d24efeb86345f745d3ce8baef1f4f2c2aa0fb429f82.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "c6302408469e469ef7020481ccec0325f979bf5927a8a5181e330fc859f4d0a4", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "78841b40372f10f27a9914dc4caa94813f4e34c50e866a446e6c7e8ecfe09ff1", "generation_seconds_batch": 15.50182735407725, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.935538411140442, "target_z": 0.10546877767178327, "ce": 5.847874641418457, "genuineness": 4.8588056564331055, "identity": 0.6479216814041138, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c6302408469e469ef7020481ccec0325f979bf5927a8a5181e330fc859f4d0a4.mp3" }, { "id": "heldout-nl-2", "language": "nl", "sentences": [ "De weg gaat linksaf na de brug.", "Loop verder... tot aan het kleine houten hek.", "Ik wacht aan de andere kant." ], "text": "De weg gaat linksaf na de brug. Loop verder... tot aan het kleine houten hek. Ik wacht aan de andere kant.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-2", "key": "c66923cefd64fdd1e1cf78c7f10af322ed7198d3a19e10f31108f1640ac147a9", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ik wacht aan de andere kant.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Natural clear speech.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Natural clear speech.) [3.2 seconds duration] Ik wacht aan de andere kant.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "801cd2468c2992150c8e0b051cb8a24fdc0123065756621fb4356196a240c19f", "generation_seconds_batch": 16.42439042206388, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1595139503479004, "target_z": 0.871093848744574, "ce": 5.526196479797363, "genuineness": 4.507347106933594, "identity": 0.3604172468185425, "hypothesis": "De weg gaat linksaf naar de brug. Loop verder. Tot aan het kleine houten hek. Ik wacht aan de andere kant.", "wer": 0.047619047619047616, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c66923cefd64fdd1e1cf78c7f10af322ed7198d3a19e10f31108f1640ac147a9.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "c6d06a6eaa77103ac1f21b9973e7cf7da5b65298fd504d12a3d3c92e0b484d4f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "df8684785d354293cee303981a4b3565ca6dd1dcdafd9a9753b596b211f8f68d", "generation_seconds_batch": 22.764518029987812, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0158839225769043, "target_z": 0.7812499370831033, "ce": 5.629451274871826, "genuineness": 4.909013748168945, "identity": 0.5998225212097168, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c6d06a6eaa77103ac1f21b9973e7cf7da5b65298fd504d12a3d3c92e0b484d4f.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "c77262620bae00d82a132d788a265dc7940978b44620ab6387ef8de8381e9f38", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7a8c8f9bda55305b73e6a3e9aa4396ecc96e70cbe9460911df3e290a1dd6f489", "generation_seconds_batch": 16.524717071908526, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.285245656967163, "target_z": 0.3242187660205061, "ce": 5.748591423034668, "genuineness": 4.582659721374512, "identity": 0.4912145137786865, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c77262620bae00d82a132d788a265dc7940978b44620ab6387ef8de8381e9f38.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "c780f7258a78c443769c190d09d364c39b12c0ea5f38fc1b6038d38956c9f731", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "ea8eac3e87114b1bdef8fe87211d5850f838cfc3cf4de912a872e1a83bffa704", "generation_seconds_batch": 16.326507619931363, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.172839641571045, "target_z": 0.2539061978605347, "ce": 6.284721851348877, "genuineness": 4.362998008728027, "identity": 0.33388203382492065, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c780f7258a78c443769c190d09d364c39b12c0ea5f38fc1b6038d38956c9f731.mp3" }, { "id": "heldout-de-0", "language": "de", "sentences": [ "Deine Nachricht kam, als ich schon weg war.", "Ich las sie zweimal... dann legte ich das Handy weg.", "Sag mir, wann wir reden können." ], "text": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal... dann legte ich das Handy weg. Sag mir, wann wir reden können.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-0", "key": "c7914bbeee43f4dbf8dad4c3ed282c23a6b399beb47c6da4a1f03baf109d9009", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wann wir reden können.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Natural clear speech.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wann wir reden können.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9632514337b9f828c499ef5af95f835176fc37d5f038ff46413c8d5b2a208135", "generation_seconds_batch": 17.13481791492086, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.178248167037964, "target_z": 0.8828125355363954, "ce": 5.168776512145996, "genuineness": 4.381826400756836, "identity": 0.6618747711181641, "hypothesis": "Deine Nachri kam, als ich schon weg war. Ich las sie zweimal. Dann lagte ich das Handy weg. Sag mir, wann wir reden können.", "wer": 0.08333333333333333, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c7914bbeee43f4dbf8dad4c3ed282c23a6b399beb47c6da4a1f03baf109d9009.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "c7920d72e580580cc01fde62ea4d45c5c12d058750a807c98165a48b16a457ec", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6561cd70c4a559fc6e75036b80bdc291d5b1c46fe67fd9d9e5ce3684d1803f53", "generation_seconds_batch": 23.100321167963557, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2892043590545654, "target_z": -0.29882815369127, "ce": 6.300607204437256, "genuineness": 4.532451152801514, "identity": 0.4501883387565613, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c7920d72e580580cc01fde62ea4d45c5c12d058750a807c98165a48b16a457ec.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "c7993cc3bddb1163de8e0a354f17c496a04a7e63c35dcf394f943fe333456068", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "84114cc37ec2c3fc781f502742b6e61df196bcb0b7065c84a42e52fb27f26354", "generation_seconds_batch": 22.60690902802162, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.715298652648926, "target_z": 1.2187500629168966, "ce": 5.4944257736206055, "genuineness": 4.65797233581543, "identity": 0.619090735912323, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c7993cc3bddb1163de8e0a354f17c496a04a7e63c35dcf394f943fe333456068.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "c7ffb93ab472097706c32a016d71806851dc454ab0cfa07036f06d668d3d3bb7", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e720b37820ccbfcf778fd01e9c54d5f27cd59d999419c2d41ad0e80eef3638dc", "generation_seconds_batch": 22.267045760992914, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6287081241607666, "target_z": 0.5390625005825639, "ce": 5.677107334136963, "genuineness": 4.7458367347717285, "identity": 0.6802245378494263, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c7ffb93ab472097706c32a016d71806851dc454ab0cfa07036f06d668d3d3bb7.mp3" }, { "id": "heldout-nl-0", "language": "nl", "sentences": [ "Je bericht kwam nadat ik vertrokken was.", "Ik las het twee keer... en legde mijn telefoon weg.", "Zeg maar wanneer we kunnen praten." ], "text": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer... en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-0", "key": "c82131bfaf89850d394bf3a21117deb94436b65306054c3e1cbac0c65e2f913d", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7a208d60b79679e2d8a746ef91f49d05c127ff6722f1330698406ee2708a5c0c", "generation_seconds_batch": 17.006464859005064, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.17451250553131104, "target_z": -0.9960937462133349, "ce": 6.41180419921875, "genuineness": 4.293961524963379, "identity": 0.6340515613555908, "hypothesis": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c82131bfaf89850d394bf3a21117deb94436b65306054c3e1cbac0c65e2f913d.mp3" }, { "id": "heldout-en-3", "language": "en", "sentences": [ "Someone moved the little blue chair.", "It used to stand here... beside the tall bookshelf.", "We can put it back after lunch." ], "text": "Someone moved the little blue chair. It used to stand here... beside the tall bookshelf. We can put it back after lunch.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-3", "key": "c8907269f81c0a33f09ba3b9d9f76b6acb2a8250e6881dbc11786cea01719e9d", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Someone moved the little blue chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can put it back after lunch.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Someone moved the little blue chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can put it back after lunch.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cdae0be177b740688ebf9ce7901ed7006bb1f56f82f3653bd91617457a39d921", "generation_seconds_batch": 17.12878136395011, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0487806797027588, "target_z": -0.4492187380574409, "ce": 6.435632705688477, "genuineness": 4.6705241203308105, "identity": 0.36806201934814453, "hypothesis": "Someone moved the little blue chair. It used to stand here beside the tall bookshelf. We can put it back after lunch.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c8907269f81c0a33f09ba3b9d9f76b6acb2a8250e6881dbc11786cea01719e9d.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "c8af0a98225c24087e3ff43a04976ca27248ba5902781fcc0556c998a06a4f3e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6a8134986081bf833ae85529ace9e30d61053ee3c17482c41bc2f7b569a9ffa6", "generation_seconds_batch": 22.954712528036907, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.034618377685547, "target_z": 0.7929687730112724, "ce": 5.208489894866943, "genuineness": 4.218649387359619, "identity": 0.4099881052970886, "hypothesis": "De deur is staat nog open. Ik heb je meer jas op de stoel gelegd, zeg maar wanneer je klaar bent.", "wer": 0.10526315789473684, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c8af0a98225c24087e3ff43a04976ca27248ba5902781fcc0556c998a06a4f3e.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "c8c3e9a088dedb9a516b4e9db32adb1163ee32f52278c98d6f1158171cef7e3f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2f487370ac2a00e20f5cce3ac57098b51d263f576d6e3c3130f12871b97fed73", "generation_seconds_batch": 22.41607297293376, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.29689884185791, "target_z": 0.957031282914858, "ce": 5.609594821929932, "genuineness": 4.532451152801514, "identity": 0.7932213544845581, "hypothesis": "J' j'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c8c3e9a088dedb9a516b4e9db32adb1163ee32f52278c98d6f1158171cef7e3f.mp3" }, { "id": "heldout-de-4", "language": "de", "sentences": [ "Ich schrieb die Nummer auf den Umschlag.", "Lies sie genau... und prüfe die letzten beiden Ziffern.", "Ruf an, wenn du Zeit hast." ], "text": "Ich schrieb die Nummer auf den Umschlag. Lies sie genau... und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-4", "key": "c976d09f1c0101f6cf04cd188fd429b14843e3403852f1ff6aa041c804c49489", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Natural clear speech.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Natural clear speech.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3ad817db6d6c9ffef3e6aac013a0b84ca4970e6274983b2517fd6dd929418f56", "generation_seconds_batch": 17.173577815992758, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7411141395568848, "target_z": 0.6093750687425352, "ce": 5.550024509429932, "genuineness": 4.65797233581543, "identity": 0.6223987340927124, "hypothesis": "Ich schrieb die Nummer auf den Umschlag, lies sie genau und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c976d09f1c0101f6cf04cd188fd429b14843e3403852f1ff6aa041c804c49489.mp3" }, { "id": "heldout-de-4", "language": "de", "sentences": [ "Ich schrieb die Nummer auf den Umschlag.", "Lies sie genau... und prüfe die letzten beiden Ziffern.", "Ruf an, wenn du Zeit hast." ], "text": "Ich schrieb die Nummer auf den Umschlag. Lies sie genau... und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-4", "key": "c97fec8793919b8fb26ff3c2656397fe9e57545ed71199e1ee88af6195a9f94c", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Natural clear speech.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Natural clear speech.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3f4bc1599214bbbf199e821801defc9e93382bb8f4a368bdc4e5074236c32588", "generation_seconds_batch": 16.56231646600645, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.316469430923462, "target_z": 0.34374996038565764, "ce": 5.847874641418457, "genuineness": 4.582659721374512, "identity": 0.7552926540374756, "hypothesis": "Ich schrieb die Nummer auf den Umschlag, ließ sie genau und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "wer": 0.045454545454545456, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c97fec8793919b8fb26ff3c2656397fe9e57545ed71199e1ee88af6195a9f94c.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "c9d68a7dc044f78275a07680e85375210517444dc826616763e6b40b9a994525", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "defe26157a7351aad1857bfdb70e06db6fa247b8a2182196e5d26534a84ee8c0", "generation_seconds_batch": 22.2676793289138, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.753603458404541, "target_z": 0.6171874271795177, "ce": 5.959072113037109, "genuineness": 4.494794845581055, "identity": 0.5003564357757568, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c9d68a7dc044f78275a07680e85375210517444dc826616763e6b40b9a994525.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "c9f2eae2e5860294d5d7bdd585e3529522e712527cc70f2029d7c223be0a3d1a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d735819cc86a9e374da0dfc0ea13cd524791cf849d293b05ba87f2100fabb0d6", "generation_seconds_batch": 22.61261413304601, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6451563835144043, "target_z": -0.0761719115558821, "ce": 6.1894097328186035, "genuineness": 4.231201171875, "identity": 0.30072280764579773, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/c9f2eae2e5860294d5d7bdd585e3529522e712527cc70f2029d7c223be0a3d1a.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "ca0417a4ed43bff73c7bd043042db5d190dd2829f237cc701af6a7eb72e1551e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b0f49c95ff44e23306418c66169bf08ac5c1242d74d7703e1b40da248f161101", "generation_seconds_batch": 17.14468647004105, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.153269052505493, "target_z": 0.867187520389735, "ce": 5.5817952156066895, "genuineness": 4.482243061065674, "identity": 0.3336872160434723, "hypothesis": "Ik heb een klein doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.1111111111111111, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ca0417a4ed43bff73c7bd043042db5d190dd2829f237cc701af6a7eb72e1551e.mp3" }, { "id": "heldout-en-2", "language": "en", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-2", "key": "ca1e1c2ec75b79931dc6a0a9512052d19b28f3977e3d7f0cb5ae7d016565232f", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The road turns left behind the bridge.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The road turns left behind the bridge.\n(Natural clear speech.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Natural clear speech.) [3.2 seconds duration] I will meet you on the other side.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6e4bade9f42423fd6e0a662c3a181b409089033c7dff3618a46c60e68b86c41d", "generation_seconds_batch": 22.64156852499582, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7411141395568848, "target_z": 0.6093750687425352, "ce": 5.454712867736816, "genuineness": 4.231201171875, "identity": 0.35114341974258423, "hypothesis": "The road turns left behind the bridge. Keep going until you reach the small wooden gate. I will meet you on the other side.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ca1e1c2ec75b79931dc6a0a9512052d19b28f3977e3d7f0cb5ae7d016565232f.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "ca7cf1e3b22b1d628dee1eb92c850361e66e8be79699904a6c573ec5589f719c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e42134819ee0ff18b0de007d7aa6a9cc90c1986d725ef000de8f3c73470ad7b6", "generation_seconds_batch": 20.67637860402465, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3905704021453857, "target_z": 1.0156250151466604, "ce": 5.677107334136963, "genuineness": 4.65797233581543, "identity": 0.7820543050765991, "hypothesis": "La porte est encore ouverte, j'ai posé ton manteau sur la chaise, dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ca7cf1e3b22b1d628dee1eb92c850361e66e8be79699904a6c573ec5589f719c.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "ca91b63542b9f7e8d804f8250b8ad028ddfb2fdb99eb0e67197a1dde1b6b9983", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d5b0ffdcdea742d083f70f1c20c61b6192a23f1a9ca57b65b5d4284ba17bc98c", "generation_seconds_batch": 21.77947737602517, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.035454750061035, "target_z": 0.16796876369025066, "ce": 5.967014789581299, "genuineness": 4.7458367347717285, "identity": 0.6927143335342407, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ca91b63542b9f7e8d804f8250b8ad028ddfb2fdb99eb0e67197a1dde1b6b9983.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "cac88715c9fe9e2ab88217a60fec0ce1b77f23bf2f370257334eadc4a2875a45", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "7a61ba55bd285ac307d067e71f4fe4161eb0c2b86d38ac57d235a159e177f069", "generation_seconds_batch": 20.547346844919957, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4016103744506836, "target_z": -0.2285155855312986, "ce": 6.1894097328186035, "genuineness": 4.2249250411987305, "identity": 0.2951389253139496, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cac88715c9fe9e2ab88217a60fec0ce1b77f23bf2f370257334eadc4a2875a45.mp3" }, { "id": "heldout-de-1", "language": "de", "sentences": [ "Die Schlüssel liegen auf dem Küchentisch.", "Ich suchte in der Schublade... dann neben dem Fenster.", "Nimm sie mit, wenn du gehst." ], "text": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade... dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-1", "key": "cb67a76fc9114e487040375f70b0271dec8a88e701c20557291a36cd8405bf1f", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Natural clear speech.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Natural clear speech.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "493fb712ae9e8bbfc8178df2873b374f751ed47b2f51f7ee706e38706a4c7d90", "generation_seconds_batch": 17.200157932937145, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.866009473800659, "target_z": 0.6874999953394891, "ce": 5.4666266441345215, "genuineness": 4.469690799713135, "identity": 0.39760491251945496, "hypothesis": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade. Dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cb67a76fc9114e487040375f70b0271dec8a88e701c20557291a36cd8405bf1f.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "cbee243f89a0074fa9881a5fd92dcb90fe8c0d592bee1862225ea9be4a3f2009", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "58fad98b18d375524181989bac09997b651977cf2e4a4312272571a1202c76ce", "generation_seconds_batch": 24.91639245103579, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1799209117889404, "target_z": -0.36718748310564814, "ce": 5.990842819213867, "genuineness": 4.331617832183838, "identity": 0.6884741187095642, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cbee243f89a0074fa9881a5fd92dcb90fe8c0d592bee1862225ea9be4a3f2009.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "cc62e9d9f1864032f27818112fb61fc93488834cf17ad297bd66f555e0b2ea2c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4766ecb4197d9da37ed762c3a2f9c0b88b82781de8dcf05ac463e5f81eef9223", "generation_seconds_batch": 23.134083525044844, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5546072721481323, "target_z": -0.13281247961026496, "ce": 6.149696350097656, "genuineness": 4.7834930419921875, "identity": 0.608375072479248, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cc62e9d9f1864032f27818112fb61fc93488834cf17ad297bd66f555e0b2ea2c.mp3" }, { "id": "heldout-fr-3", "language": "fr", "sentences": [ "Quelqu’un a déplacé la petite chaise bleue.", "Elle était ici... près de la grande bibliothèque.", "Nous la remettrons après le déjeuner." ], "text": "Quelqu’un a déplacé la petite chaise bleue. Elle était ici... près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-3", "key": "cc7984979d9025dd78c462ee9aa592e9078ade6bb595fbfef02d5ce8f42a3aec", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Natural clear speech.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Natural clear speech.) [3.2 seconds duration] Nous la remettrons après le déjeuner.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "a0ea798781e3ff4df49fc26584d083825e4adec626e2a5dd6dc8983faea87990", "generation_seconds_batch": 23.34268284298014, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5654239654541016, "target_z": 1.1249999720369348, "ce": 5.391171455383301, "genuineness": 4.632867813110352, "identity": 0.8802014589309692, "hypothesis": "Quelqu'un a déplacé la petite chaise bleue, elle était ici, près de la grande bibliothèque. Nous la remettrons après le déjeuner.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cc7984979d9025dd78c462ee9aa592e9078ade6bb595fbfef02d5ce8f42a3aec.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "ccd06fdcd60ebab7a5f13f82e964f492587ab3c35c0ea4e7bb1af0f02ddcfcca", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1c37f57279e32bb94518e005e4be3490bd2b65ca27893cfe40c873b2452b3e84", "generation_seconds_batch": 17.567121925996616, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6287081241607666, "target_z": 0.5390625005825639, "ce": 5.824047088623047, "genuineness": 4.293961524963379, "identity": 0.6363843083381653, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ccd06fdcd60ebab7a5f13f82e964f492587ab3c35c0ea4e7bb1af0f02ddcfcca.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "ccd78c9616b6d7bbf9f43603333cdea8a6ec7289cb82d6bdac4b705514df4ce8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "dd3f30f5383c04977cc882e585c67db067014f7f8d4b9e46fcac7c021511434e", "generation_seconds_batch": 23.094897332950495, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.827704429626465, "target_z": 1.2890624819405203, "ce": 5.613565921783447, "genuineness": 4.7332844734191895, "identity": 0.8034563064575195, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ccd78c9616b6d7bbf9f43603333cdea8a6ec7289cb82d6bdac4b705514df4ce8.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "cd33e4571f8066199cb0f70e5a22dc77154e0539207d390024bada67ad24e144", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "ca2a583018b781b0ec14b17da9c3f4b69b35a229c2f55f1321dd04c528facce8", "generation_seconds_batch": 22.53629216598347, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.132248640060425, "target_z": 0.2285155855312986, "ce": 6.062326908111572, "genuineness": 4.507347106933594, "identity": 0.33559516072273254, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cd33e4571f8066199cb0f70e5a22dc77154e0539207d390024bada67ad24e144.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "cd67b4b6627e00aadc4d3c44ce5b3b3475c37b5c0719846f0c4913d96bbbf449", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "35e24925a29689ee27dcaf1930fe7eb2fa766c33baf7af37d8a109f0dff67afa", "generation_seconds_batch": 23.14045072393492, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.057311534881592, "target_z": 0.18164068922766535, "ce": 6.173524379730225, "genuineness": 4.306513786315918, "identity": 0.40403178334236145, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cd67b4b6627e00aadc4d3c44ce5b3b3475c37b5c0719846f0c4913d96bbbf449.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "cd749373d5ec79d66e6a0c061df3602ba2ad890b9a1e21d5682a9cfaa85b5fdc", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "fc828d1e5b8f8fd4ce615b46c922005a612b31075c9c7691cbc9fd0d321f9d1a", "generation_seconds_batch": 22.461986104026437, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9324159622192383, "target_z": 0.1035156134943638, "ce": 6.054384231567383, "genuineness": 4.231201171875, "identity": 0.7615901231765747, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cd749373d5ec79d66e6a0c061df3602ba2ad890b9a1e21d5682a9cfaa85b5fdc.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "cd81eec9d1e95c4e727147b8c3aa76f33bb7cbe3cf34961b630ef7f8655e8033", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "38f63d56d4ec58683b922c60f93fd33ad7fab93f019f10ed6fc7eab000067132", "generation_seconds_batch": 22.318154538050294, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3780808448791504, "target_z": 1.0078125075733302, "ce": 6.0146708488464355, "genuineness": 4.469690799713135, "identity": 0.3665774464607239, "hypothesis": "The door is still open. Put your coat on the chair. Tell me when you are ready.", "wer": 0.05555555555555555, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cd81eec9d1e95c4e727147b8c3aa76f33bb7cbe3cf34961b630ef7f8655e8033.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "cd8f847ce154ea2c5af806c9aefdc9d5a3b7a40c97a652dabbb07bc2066e4390", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a93a34077c27b128227adab6323b08e00665a7ffd32157066793b342a488a66b", "generation_seconds_batch": 16.999474014970474, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.153269052505493, "target_z": 0.867187520389735, "ce": 6.070269584655762, "genuineness": 4.645420074462891, "identity": 0.7524203062057495, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cd8f847ce154ea2c5af806c9aefdc9d5a3b7a40c97a652dabbb07bc2066e4390.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "cdc41813de1eb1bd97321f11c994b7549e322af4a4ebad6ae62596f7cd5f5971", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "535ff3e4eb4e9f2f7590318eabdf69bfb5c43d332ae7445446c122eade7c2599", "generation_seconds_batch": 20.265294632059522, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.89098858833313, "target_z": 0.7031250104861494, "ce": 5.990842819213867, "genuineness": 4.871357440948486, "identity": 0.5044954419136047, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cdc41813de1eb1bd97321f11c994b7549e322af4a4ebad6ae62596f7cd5f5971.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "cdd76b323fe3695608501ce1ac0ed4c7fa777c22a8c5ca2cd994b2f8070ad822", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ba4284648166092810397d23cabf5f21a852624e8d8405b35d599f6d7bda4784", "generation_seconds_batch": 21.29929030605126, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4765475988388062, "target_z": -0.18164061465949152, "ce": 6.300607204437256, "genuineness": 4.507347106933594, "identity": 0.6389406323432922, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cdd76b323fe3695608501ce1ac0ed4c7fa777c22a8c5ca2cd994b2f8070ad822.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "ce00f0fa034de2cb0f107d3f7e7adb3af78e7ac33346e36ce160d634adc73c09", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "6263d0aa42dacb7567157108b001bd34fdd40f8d0fe73b9786d51e1c5988c207", "generation_seconds_batch": 16.67421900201589, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2469406127929688, "target_z": 0.9257812526215373, "ce": 5.510311126708984, "genuineness": 4.469690799713135, "identity": 0.3430248498916626, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ce00f0fa034de2cb0f107d3f7e7adb3af78e7ac33346e36ce160d634adc73c09.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "ce0c4ffb99795accafb899c50aceb116d7c6bc9b01571351fbf33d1a0be6b4aa", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6cea85ecb2e03a0b20230a70f576a9e5af8acc76270302a67d0da51f3b646cba", "generation_seconds_batch": 19.81498774397187, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0221288204193115, "target_z": 0.7851562654379423, "ce": 5.998785495758057, "genuineness": 4.796044826507568, "identity": 0.706091046333313, "hypothesis": "La porte est encore ouverte, j'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ce0c4ffb99795accafb899c50aceb116d7c6bc9b01571351fbf33d1a0be6b4aa.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "ce55e0fe5c5cd09977bcac0f1ba132f80216c101c4f26fb15811eae74ec787f2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2aa87d8cc8ba00ff8ac4d71815d41cb99de14772dc78703a6e3ee7b68c72f507", "generation_seconds_batch": 22.31468472396955, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.385161876678467, "target_z": 0.3867186774707997, "ce": 5.839932441711426, "genuineness": 4.356721878051758, "identity": 0.6639195084571838, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ce55e0fe5c5cd09977bcac0f1ba132f80216c101c4f26fb15811eae74ec787f2.mp3" }, { "id": "heldout-de-3", "language": "de", "sentences": [ "Jemand hat den kleinen blauen Stuhl umgestellt.", "Er stand hier... neben dem hohen Bücherregal.", "Wir stellen ihn nach dem Essen zurück." ], "text": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier... neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-3", "key": "ce642be42d2b952443c3fb267eb33ba687a543a14f99d2dc0fea4cddd5124488", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Natural clear speech.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Natural clear speech.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7cbd272b5231c64ed3ecbd4c289f1e30ff69f5165b8e7cc26e945a087f5821c3", "generation_seconds_batch": 16.903751563979313, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8472752571105957, "target_z": 0.6757813085476677, "ce": 5.577824115753174, "genuineness": 4.570107460021973, "identity": 0.6662791967391968, "hypothesis": "Jemand hat einen kleinen blauen Stuhl umgestellt. Er stand hier neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "wer": 0.047619047619047616, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ce642be42d2b952443c3fb267eb33ba687a543a14f99d2dc0fea4cddd5124488.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "ce6ff23aba110781b80a9e078061a031a8ea44db89a83b3b5ac824008c95f6e8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9bd7fa88f86438900df8c5f2e6a66c89bcb6325c8bc7215e50e9a1ecccafd3f9", "generation_seconds_batch": 22.70897358295042, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3780808448791504, "target_z": 1.0078125075733302, "ce": 5.871703147888184, "genuineness": 4.883909225463867, "identity": 0.6951742172241211, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ce6ff23aba110781b80a9e078061a031a8ea44db89a83b3b5ac824008c95f6e8.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "ceb7cb061fc93e735a99cea200190c901f9db055af4b667200218873bece21b8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "47728a5a2b6e895bf9c389caaa19dcac0fabaa89e1a5075f01fbff7df756d388", "generation_seconds_batch": 22.096404760028236, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.538158893585205, "target_z": 0.48242185796000714, "ce": 6.24500846862793, "genuineness": 4.6203155517578125, "identity": 0.42837613821029663, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ceb7cb061fc93e735a99cea200190c901f9db055af4b667200218873bece21b8.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "ced8cf47abe8eebba1d06ca2e0783a39aea38d73386999d9bfca68da9592c5dd", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] La porte est encore ouverte.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "41fe29e67128d740a28c0107cceebdd9ad7f10a0d49cb6dab79d9d4269f4117b", "generation_seconds_batch": 17.34514583996497, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.578749895095825, "target_z": 0.5078124702892433, "ce": 6.086154937744141, "genuineness": 4.595211505889893, "identity": 0.82554692029953, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ced8cf47abe8eebba1d06ca2e0783a39aea38d73386999d9bfca68da9592c5dd.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "cee11ff6df2625969831dd41c5f2255d5f6a760ff371befff00c8ea5cda6eb48", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "39e65a1ac0340389044e75a04cef0c46e24c5e0ceffa9f109b38ade971e4a0e0", "generation_seconds_batch": 23.241455629002303, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8472752571105957, "target_z": 0.6757813085476677, "ce": 5.039708137512207, "genuineness": 4.557555198669434, "identity": 0.6805063486099243, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cee11ff6df2625969831dd41c5f2255d5f6a760ff371befff00c8ea5cda6eb48.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "cf2214203e0828712a839b29ac2ee0b55a5cb16c5c6a20e4ba77ae9ae431c00d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8a2d87dae109ff79989421dd1cea75e6d0f357ea5a86ee1161ffd3add189d74e", "generation_seconds_batch": 16.89404582499992, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0854129791259766, "target_z": 0.19921879398357129, "ce": 5.855817794799805, "genuineness": 4.582659721374512, "identity": 0.731996476650238, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd, zeg maar, wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cf2214203e0828712a839b29ac2ee0b55a5cb16c5c6a20e4ba77ae9ae431c00d.mp3" }, { "id": "heldout-de-3", "language": "de", "sentences": [ "Jemand hat den kleinen blauen Stuhl umgestellt.", "Er stand hier... neben dem hohen Bücherregal.", "Wir stellen ihn nach dem Essen zurück." ], "text": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier... neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-3", "key": "cf68e6506caaaad7e7371e72de9d5417fef89211300487ff83acd2484eaab510", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "93d03a33a199c14a98065a7c7bbceebd52ea51ebec9535810584cc8b5adde850", "generation_seconds_batch": 17.56127679697238, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.3306318521499634, "target_z": -0.8984374761148818, "ce": 6.395919322967529, "genuineness": 4.6203155517578125, "identity": 0.6226609945297241, "hypothesis": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/cf68e6506caaaad7e7371e72de9d5417fef89211300487ff83acd2484eaab510.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "d0f2950c57ae130a8e53bb1afca825b7e7433535fc48f98893127eab2274e746", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7185995fca59383f2ea64225a47c24d7c1216b9b337c3b4045eb5cb96daff8be", "generation_seconds_batch": 22.318154538050294, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7036454677581787, "target_z": 0.5859375460225448, "ce": 6.022613525390625, "genuineness": 4.337893962860107, "identity": 0.2339472472667694, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d0f2950c57ae130a8e53bb1afca825b7e7433535fc48f98893127eab2274e746.mp3" }, { "id": "heldout-en-4", "language": "en", "sentences": [ "I wrote the number on the envelope.", "Read it carefully... and check the last two digits.", "Call me when you have a moment." ], "text": "I wrote the number on the envelope. Read it carefully... and check the last two digits. Call me when you have a moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-4", "key": "d11caac31fa3732e2517803823f1b039213fb31e9438619fd83d8b232c894b7d", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I wrote the number on the envelope.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Call me when you have a moment.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I wrote the number on the envelope.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Call me when you have a moment.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7a5447c02ed868f0332eb4c7621d235201194bd54e4d9b45caced1f45b1404cd", "generation_seconds_batch": 16.843197706970386, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9613538384437561, "target_z": -0.5039062537866651, "ce": 6.459460258483887, "genuineness": 3.9676074981689453, "identity": 0.5767395496368408, "hypothesis": "I wrote the number on the envelope. Read it carefully and check the last two digits. Call me when you have a moment.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d11caac31fa3732e2517803823f1b039213fb31e9438619fd83d8b232c894b7d.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "d13202582145839a4a80b9275e4c84c9e664cf9b80c7ad750d2e34fa5b118c75", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1925f4703c9dc1b39330203a1ea339e0e2ef6d9110c6e96491f0850f55c0923a", "generation_seconds_batch": 21.505827025976032, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0158839225769043, "target_z": 0.7812499370831033, "ce": 5.152891159057617, "genuineness": 4.457139015197754, "identity": 0.488894522190094, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het augaan. We kunnen er samen in kijken.", "wer": 0.1111111111111111, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d13202582145839a4a80b9275e4c84c9e664cf9b80c7ad750d2e34fa5b118c75.mp3" }, { "id": "heldout-nl-2", "language": "nl", "sentences": [ "De weg gaat linksaf na de brug.", "Loop verder... tot aan het kleine houten hek.", "Ik wacht aan de andere kant." ], "text": "De weg gaat linksaf na de brug. Loop verder... tot aan het kleine houten hek. Ik wacht aan de andere kant.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-2", "key": "d19bdc8e01e9f26b708f908bc3a52d62ce566b4d4cd674500b15790d875120a4", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ik wacht aan de andere kant.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Natural clear speech.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Natural clear speech.) [3.2 seconds duration] Ik wacht aan de andere kant.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3d1c4fd82d212a58c62ff60500fa93efe9bad7e875beccf10f890a6a1fe6a917", "generation_seconds_batch": 15.843171798973344, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.578749895095825, "target_z": 0.5078124702892433, "ce": 5.387199878692627, "genuineness": 4.344170093536377, "identity": 0.24903294444084167, "hypothesis": "De weg gaat linksaf naar de brug. Lok verder tot aan het kleine houten hek. Ik wacht aan de andere kant.", "wer": 0.09523809523809523, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d19bdc8e01e9f26b708f908bc3a52d62ce566b4d4cd674500b15790d875120a4.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "d1e0457d91aacc4834ddd43e836fe3d61cd24a32a4295c074a44045b3d0955d6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bb553ae4da9fa9f0ddabfefcdb2a622cdba70fccf367efa5a7c5a4de9a79d513", "generation_seconds_batch": 22.66546756005846, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6678495407104492, "target_z": -0.6874999953394891, "ce": 6.149696350097656, "genuineness": 4.432034492492676, "identity": 0.6866102814674377, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d1e0457d91aacc4834ddd43e836fe3d61cd24a32a4295c074a44045b3d0955d6.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "d1fd7ef562698801e91b705900291195ea935edfce91d85a5061c4cbcac56dcf", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "444f5d1c60920c03f92c975e7334f29ab7480571cf82286e9c34f2c29e1359df", "generation_seconds_batch": 21.675431331968866, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1282899379730225, "target_z": 0.8515625052430748, "ce": 6.197352409362793, "genuineness": 4.896461486816406, "identity": 0.5878786444664001, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi, quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d1fd7ef562698801e91b705900291195ea935edfce91d85a5061c4cbcac56dcf.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "d22e22f3d7ea08e04bf3a2f30f5a49dc75d8672b2584099b9a9f677d9294dedc", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "b70038141465b8209a2912756fc6259697f7dab55a44953b96c169c7ccda8be0", "generation_seconds_batch": 22.70148966391571, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6919922828674316, "target_z": -0.046874970871807094, "ce": 6.23706579208374, "genuineness": 4.149612903594971, "identity": 0.28384390473365784, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d22e22f3d7ea08e04bf3a2f30f5a49dc75d8672b2584099b9a9f677d9294dedc.mp3" }, { "id": "heldout-es-1", "language": "es", "sentences": [ "Las llaves están en la mesa de la cocina.", "Miré en el cajón... y luego junto a la ventana.", "Llévatelas cuando te vayas." ], "text": "Las llaves están en la mesa de la cocina. Miré en el cajón... y luego junto a la ventana. Llévatelas cuando te vayas.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-1", "key": "d2319f0ba4d4a13f099c61fd2524d2cd47bb09bca092815ed8a29e7f5b0c527c", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Llévatelas cuando te vayas.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Llévatelas cuando te vayas.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e3b95d2454d38692e8e082a8da74cb86699dda545a7fabc608ce11ada2449680", "generation_seconds_batch": 17.306735864956863, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.2744288444519043, "target_z": -0.9335937601948675, "ce": 6.1894097328186035, "genuineness": 4.174716949462891, "identity": 0.5363321304321289, "hypothesis": "Las llaves están en la mesa de la cocina, miré en el cajón y luego junto a la ventana, llévatelas cuando te vayas.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d2319f0ba4d4a13f099c61fd2524d2cd47bb09bca092815ed8a29e7f5b0c527c.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "d26fbb7ffa9f0baf01605de0fd1b17a66763b59951019423e6318835375af0f6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6640af4ff682812b515686b1aefdd5e9e4b2ecb34f6da4ef7ded2f970eb0029e", "generation_seconds_batch": 16.27042228600476, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.350815773010254, "target_z": 0.3652343934964025, "ce": 5.6334228515625, "genuineness": 4.909013748168945, "identity": 0.5920608043670654, "hypothesis": "He encontrado la cajita. Estaba junto la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.058823529411764705, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d26fbb7ffa9f0baf01605de0fd1b17a66763b59951019423e6318835375af0f6.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "d27c54f781d097bfebf4fb659de303539a2dca6dd365633894919133b8f186e8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "8132c57490a669a6e091ad62a3a5cad47c757f68c93783b9e43e0a71caa6f732", "generation_seconds_batch": 22.275458641001023, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.147024154663086, "target_z": 0.8632811920348961, "ce": 5.403085231781006, "genuineness": 4.319066047668457, "identity": 0.5054882168769836, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d27c54f781d097bfebf4fb659de303539a2dca6dd365633894919133b8f186e8.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "d2d32deb772b5629942db6268a5c2cd54b8ee79df3cc34ef2f20d41836b151e0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "12057271d6cd4c71b336ad3ea0705e6f042be091e28bf62712ef4d03d4f6117e", "generation_seconds_batch": 23.18066223000642, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6162185668945312, "target_z": 0.5312499930092337, "ce": 5.919358730316162, "genuineness": 4.545003414154053, "identity": 0.26897183060646057, "hypothesis": "Ik heb het kleine doosje gevonden, het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d2d32deb772b5629942db6268a5c2cd54b8ee79df3cc34ef2f20d41836b151e0.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "d365d8a790b5f48bb74152929c80842a2b7e644b49e4c255801f57974f96822e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6ee4d5b0cb00b076b05a2461030c15e5b06159e5e9b54a989151319a437ee345", "generation_seconds_batch": 23.3961015479872, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.435956597328186, "target_z": -0.2070312269887276, "ce": 6.356205940246582, "genuineness": 4.494794845581055, "identity": 0.3988189399242401, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d365d8a790b5f48bb74152929c80842a2b7e644b49e4c255801f57974f96822e.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "d38a6271bf9de1d3eb806dda523cf359099e50604639aef7d20b6b0dd0a398d6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "10fa2693c5208a7e2b08aed9f14d66dbc45f09fdf251731cd85ef6d084e5e675", "generation_seconds_batch": 21.7260549350176, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5279555320739746, "target_z": 1.101562598453292, "ce": 5.49045467376709, "genuineness": 4.720732688903809, "identity": 0.45472419261932373, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d38a6271bf9de1d3eb806dda523cf359099e50604639aef7d20b6b0dd0a398d6.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "d38b1c8768bd52f14280f4b1b185f3348c116d273ef3abbdc1ecb525db6b43cd", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ddd3ada1a5eeb75be50b4c88741d3ca347d5b4392b820be3ede8de2cdc5e53ea", "generation_seconds_batch": 21.29929030605126, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6170549392700195, "target_z": -0.09375001631178803, "ce": 6.284721851348877, "genuineness": 4.545003414154053, "identity": 0.29811251163482666, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d38b1c8768bd52f14280f4b1b185f3348c116d273ef3abbdc1ecb525db6b43cd.mp3" }, { "id": "heldout-fr-4", "language": "fr", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-4", "key": "d3fab0ead1fc9dcf899f59bddb0c9d4b54373ba1829b6dc49f77fc4027ce1740", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a6b4b3a5caee1e45d91a7c60bcb5b8bd097f2914128a5842b7159edf9892b265", "generation_seconds_batch": 16.800370823941194, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7802554368972778, "target_z": -0.6171875017476915, "ce": 6.086154937744141, "genuineness": 4.582659721374512, "identity": 0.7347681522369385, "hypothesis": "J'ai écrit le numéro sur l'enveloppe. Lis-le bien et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d3fab0ead1fc9dcf899f59bddb0c9d4b54373ba1829b6dc49f77fc4027ce1740.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "d42e519a06852ae179467a9b773b7513b7c45ed257f9ee932a60fb14a84b2d54", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d508b3201df8560aa339df6d3191c7f0d169b7a0da7fc187bf832cdea6d22487", "generation_seconds_batch": 23.246597405988723, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7856638431549072, "target_z": 0.011718761359995234, "ce": 6.117925643920898, "genuineness": 4.7458367347717285, "identity": 0.7632225155830383, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d42e519a06852ae179467a9b773b7513b7c45ed257f9ee932a60fb14a84b2d54.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "d446c9a309722dfa516d7811753dba17497cacc4750e8b3ab60e741e20dbd3f4", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e8dac48d4899b3da08d6c4b1044f82adf510756beaa00b63e1f64d961ea5d2f5", "generation_seconds_batch": 22.441613475908525, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.652850866317749, "target_z": 1.1796875250502459, "ce": 5.043679237365723, "genuineness": 4.356721878051758, "identity": 0.5980861186981201, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre, nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d446c9a309722dfa516d7811753dba17497cacc4750e8b3ab60e741e20dbd3f4.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "d4d3ed6c4901f373408674a09720b70aac9af63b8cec57ea786667cca8d39280", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "16eefa05cd752d6bc1b478f917e818392482c88f36090fbbe0e38474e3284aee", "generation_seconds_batch": 17.341312835924327, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7731742858886719, "target_z": 0.003906253786665078, "ce": 5.95112943649292, "genuineness": 4.356721878051758, "identity": 0.4193156361579895, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d4d3ed6c4901f373408674a09720b70aac9af63b8cec57ea786667cca8d39280.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "d55db00d37c91b3d03357e9f2caa772b11253483cade714a798dd519a53244f1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1326fec323a273c9ec16b7bddb0fe021c5943ccc4715319ae110d18dbd6c2013", "generation_seconds_batch": 22.751799242920242, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6201773881912231, "target_z": -0.09179685213436857, "ce": 6.1894097328186035, "genuineness": 4.65797233581543, "identity": 0.6273448467254639, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d55db00d37c91b3d03357e9f2caa772b11253483cade714a798dd519a53244f1.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "d5fcc232bcc9d1bd47cd8a6cbee5ee25d81df1f541e385bf0d42882b08ad452e", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a6d7958af72ed6983e52df77f43ab214a0b8e5cc4a8c59ca4f267480d54d394d", "generation_seconds_batch": 16.15291431103833, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8551927208900452, "target_z": -0.5703124935917976, "ce": 6.308549880981445, "genuineness": 4.3504462242126465, "identity": 0.6407703161239624, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d5fcc232bcc9d1bd47cd8a6cbee5ee25d81df1f541e385bf0d42882b08ad452e.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "d6004fa916ac6e02ee63c11cfe8511e14f4b8c537f8a3b1494bae04429838429", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "033ce39754127c0c1d313363ea8a1296c1a18e6373aed68c217cc91846e77c16", "generation_seconds_batch": 23.380126139032654, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.244654655456543, "target_z": 0.29882815369127, "ce": 5.959072113037109, "genuineness": 4.68307638168335, "identity": 0.6227479577064514, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d6004fa916ac6e02ee63c11cfe8511e14f4b8c537f8a3b1494bae04429838429.mp3" }, { "id": "heldout-nl-1", "language": "nl", "sentences": [ "De sleutels liggen op de keukentafel.", "Ik keek in de la... en daarna naast het raam.", "Neem ze mee als je weggaat." ], "text": "De sleutels liggen op de keukentafel. Ik keek in de la... en daarna naast het raam. Neem ze mee als je weggaat.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-1", "key": "d60d0e7adbfe5bf1820351a8db8cca77275ac10a13e0c1dbabf9c53f1c89978b", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Neem ze mee als je weggaat.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Natural clear speech.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Natural clear speech.) [3.2 seconds duration] Neem ze mee als je weggaat.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f45dc1e89129856557ee661313c1582cdd727f237cee62a1c65cf30e689e361a", "generation_seconds_batch": 22.47493508597836, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.853519916534424, "target_z": 0.679687487766159, "ce": 5.299830436706543, "genuineness": 4.20609712600708, "identity": 0.2851698696613312, "hypothesis": "De sleutels liggen op de keukentafel. Ik keek je aan de la en daarna naast het raam neem ze mee als je weggaat.", "wer": 0.09090909090909091, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d60d0e7adbfe5bf1820351a8db8cca77275ac10a13e0c1dbabf9c53f1c89978b.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "d6ba78c5c96a2eb0e7a188e9d114155db49efa12303c25e316a08e5f82e4f57e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.5, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1200f6a72f8210dd73e7e0002e5abd7cfaf77fdb4bb98a2de7f68efc3197cbe4", "generation_seconds_batch": 20.11818587302696, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1282899379730225, "target_z": 0.8515625052430748, "ce": 5.2640886306762695, "genuineness": 4.256305694580078, "identity": 0.4001038074493408, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d6ba78c5c96a2eb0e7a188e9d114155db49efa12303c25e316a08e5f82e4f57e.mp3" }, { "id": "heldout-nl-4", "language": "nl", "sentences": [ "Ik schreef het nummer op de envelop.", "Lees het goed... en controleer de laatste twee cijfers.", "Bel me wanneer je even tijd hebt." ], "text": "Ik schreef het nummer op de envelop. Lees het goed... en controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-4", "key": "d734fdf6d11e9803c5f23b7f7124541b188e93815370291c2ee4934bd0c4c8f5", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Natural clear speech.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Natural clear speech.) [3.2 seconds duration] Bel me wanneer je even tijd hebt.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "a8c180271d1573ee556476ce55ae64540ce47b85b9f2127987cf43dfdb65e303", "generation_seconds_batch": 16.850471340003423, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.940946578979492, "target_z": 0.7343748916431224, "ce": 5.617537498474121, "genuineness": 4.2751336097717285, "identity": 0.3738037943840027, "hypothesis": "Ik schreef het nummer op de envelop. Lees het goed. Controleer de laatste twee cijfers. Bel me wanneer je even tijd hebt.", "wer": 0.043478260869565216, "duration_error_s": 0.120000000000001, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d734fdf6d11e9803c5f23b7f7124541b188e93815370291c2ee4934bd0c4c8f5.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "d84f813f4c14331187f3d0a8dfe396fd25a8423380c7d4f96ce808b11ce19304", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8747c4b945e933fc2c17898fa4b417c69dc06e366d14edd51a70773cc9d42477", "generation_seconds_batch": 22.060450674965978, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.615382194519043, "target_z": 1.1562500023302553, "ce": 5.557967185974121, "genuineness": 4.65797233581543, "identity": 0.4511042535305023, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d84f813f4c14331187f3d0a8dfe396fd25a8423380c7d4f96ce808b11ce19304.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "d87e8463a7319fd460c54b73c5766079fc8d0f8162270ba35853ee085291a357", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "440076e801e0c376018e20a938d251d395d0f1ac5a8d10598d07f9d64e2a9e67", "generation_seconds_batch": 23.237425912055187, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.228206157684326, "target_z": 0.9140624166933683, "ce": 5.831989288330078, "genuineness": 4.545003414154053, "identity": 0.6238635182380676, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d87e8463a7319fd460c54b73c5766079fc8d0f8162270ba35853ee085291a357.mp3" }, { "id": "heldout-en-0", "language": "en", "sentences": [ "Your message arrived after I had left.", "I read it twice... before putting the phone away.", "Let me know when we can talk." ], "text": "Your message arrived after I had left. I read it twice... before putting the phone away. Let me know when we can talk.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-0", "key": "d8d76eb2dcab601ce2f53c1d7e381f019c3b4c31ab96d897fe00683e4eff69e8", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Your message arrived after I had left.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Let me know when we can talk.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Your message arrived after I had left.\n(Natural clear speech.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Natural clear speech.) [3.2 seconds duration] Let me know when we can talk.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e8b9507a5b3b9c9433f8ceea07acfc357d3bfc1571199fab939dd57f9854983c", "generation_seconds_batch": 16.519663829007186, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.140779495239258, "target_z": 0.8593750128164048, "ce": 5.458683967590332, "genuineness": 4.413206577301025, "identity": 0.45622581243515015, "hypothesis": "Your message arrived after I had left. I read it twice before putting the phone away. Let me know when we can talk.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d8d76eb2dcab601ce2f53c1d7e381f019c3b4c31ab96d897fe00683e4eff69e8.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "d930399d982590accc1f54e1d0a58a8869edeeb68a755b5d65077fb99401712a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "64910966cd85d5926ca833a036aa5f75d7262db6d566ee8aace0316e4b02d31a", "generation_seconds_batch": 22.251747611910105, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.228206157684326, "target_z": 0.9140624166933683, "ce": 5.649308204650879, "genuineness": 4.7332844734191895, "identity": 0.6076258420944214, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d930399d982590accc1f54e1d0a58a8869edeeb68a755b5d65077fb99401712a.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "d97fa714369427d32b5f96f160ad8f9d89cafc1859b624a53ab63bf6a1276771", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cfde91452a44090d575723c000ea36cb5a92da14686fdb63bc856e9ef7b1439b", "generation_seconds_batch": 21.787283440004103, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0908212661743164, "target_z": 0.8281249825230843, "ce": 5.2918877601623535, "genuineness": 4.268857479095459, "identity": 0.4060979187488556, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d97fa714369427d32b5f96f160ad8f9d89cafc1859b624a53ab63bf6a1276771.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "d983d406d2c9d0a9aab5fe8e8995420e4abb6065804446e769b9dd78e1910aba", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "d165051067447e33a59b7efef1c53d278cf777037fc769c0284c5e314dc4542d", "generation_seconds_batch": 21.840928632067516, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6732579469680786, "target_z": -0.05859373223180233, "ce": 6.292664527893066, "genuineness": 4.231201171875, "identity": 0.4317215085029602, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d983d406d2c9d0a9aab5fe8e8995420e4abb6065804446e769b9dd78e1910aba.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "d9da05677c318c612f7ad583830187ff86199b8855114143e3560c58bade2342", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "80f058db03af3527c64bd2938ddeae798a3b6b2bd7c9a6a8b18f1661a6e4a7c4", "generation_seconds_batch": 22.85506317205727, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.97217059135437, "target_z": 0.7539062351446216, "ce": 5.649308204650879, "genuineness": 4.18099308013916, "identity": 0.49566978216171265, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d9da05677c318c612f7ad583830187ff86199b8855114143e3560c58bade2342.mp3" }, { "id": "heldout-en-2", "language": "en", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-2", "key": "d9f64e10289241b04c4106df928112d729ab6c75a652d91361297f6cf24b6787", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The road turns left behind the bridge.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The road turns left behind the bridge.\n(Natural clear speech.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Natural clear speech.) [3.2 seconds duration] I will meet you on the other side.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a0d96122976101ffb19c20931cabec7b3a2171e739b2c3dea8580c9b9fc635d4", "generation_seconds_batch": 17.2117942039622, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6287081241607666, "target_z": 0.5390625005825639, "ce": 5.673136234283447, "genuineness": 4.331617832183838, "identity": 0.44611096382141113, "hypothesis": "The road turns left behind the bridge. Keep going until you reach the small wooden gate. I will meet you on the other side.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/d9f64e10289241b04c4106df928112d729ab6c75a652d91361297f6cf24b6787.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "da104d56129dd5fd22231c70cbfd50060bec198435ef6cf65ae3b50370bc5c13", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c1a853dc8f701aa382252a798d91be41f8672a80844d2a72e9a49a3320480039", "generation_seconds_batch": 21.694796059047803, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7411141395568848, "target_z": 0.6093750687425352, "ce": 6.0146708488464355, "genuineness": 4.482243061065674, "identity": 0.7492952346801758, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/da104d56129dd5fd22231c70cbfd50060bec198435ef6cf65ae3b50370bc5c13.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "daadd287bfc8b5a1aa1fdef52691e1dbe69270a1c72f9d2b996035fea1409ad2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "255de75e134a656c8a70d55e396d92bea54fbfa7ef014102a50773b2e82b3535", "generation_seconds_batch": 23.32059576793108, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.092494010925293, "target_z": -0.4218750361189592, "ce": 6.141753673553467, "genuineness": 4.419482707977295, "identity": 0.6040700674057007, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/daadd287bfc8b5a1aa1fdef52691e1dbe69270a1c72f9d2b996035fea1409ad2.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "dac6daef73b0151a8100e8e092e70fa164a87c1886c76561546bf3a4c6e4eac6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "29aca737cd5bbea12e73b02876667d7c3cf197f4cea1d2ac3656dc8763cf6ac5", "generation_seconds_batch": 16.4623029519571, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.5929123163223267, "target_z": -0.7343749662112963, "ce": 6.300607204437256, "genuineness": 4.557555198669434, "identity": 0.4043724536895752, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dac6daef73b0151a8100e8e092e70fa164a87c1886c76561546bf3a4c6e4eac6.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "dacc5a0b9021f956bc01758abd2f274dd8b2cd23dac3aa59e666f6c5c00cce36", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c4b8e5bed38db7dba549fe9ee67baeb212f97a4c1cfe253618fe53ac39dc8d67", "generation_seconds_batch": 23.3151821839856, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9238852262496948, "target_z": -0.5273437392225686, "ce": 6.395919322967529, "genuineness": 4.770940780639648, "identity": 0.5472122430801392, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dacc5a0b9021f956bc01758abd2f274dd8b2cd23dac3aa59e666f6c5c00cce36.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "db07fb289ba4f2aec8b5ed45c589179b58ac3227b74f0789aa7c60c13b95d51d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "271680018adcf50d3ffd435141ee88a1888bc4fe871fbef6c6a7b27de5e7c6bc", "generation_seconds_batch": 16.80960233102087, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.552934408187866, "target_z": 1.1171874644636046, "ce": 5.4944257736206055, "genuineness": 4.68307638168335, "identity": 0.44018471240997314, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/db07fb289ba4f2aec8b5ed45c589179b58ac3227b74f0789aa7c60c13b95d51d.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "db153c9c5ade02ff016ea8216a04bdc2708a9c24042d3d0cf4bb0b62806dc83b", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] Het stond naast het oude raam.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9fe1a2dcc96f8b829a305cd7cc8a814388d1c153689b758685b52e5654e2baa1", "generation_seconds_batch": 17.621264959918335, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9301300048828125, "target_z": -0.5234374854359035, "ce": 6.252951145172119, "genuineness": 4.482243061065674, "identity": 0.7627896666526794, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/db153c9c5ade02ff016ea8216a04bdc2708a9c24042d3d0cf4bb0b62806dc83b.mp3" }, { "id": "heldout-en-0", "language": "en", "sentences": [ "Your message arrived after I had left.", "I read it twice... before putting the phone away.", "Let me know when we can talk." ], "text": "Your message arrived after I had left. I read it twice... before putting the phone away. Let me know when we can talk.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-0", "key": "db192910e9ae67de251fb2a990940eab881f66f627c78b50b638ce61f5a63679", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Your message arrived after I had left.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Let me know when we can talk.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Your message arrived after I had left.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Let me know when we can talk.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "142e71a99999b936204605c6efa1feb13ced2c230754e456263f8cc9da9d9250", "generation_seconds_batch": 17.125993411056697, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8676822781562805, "target_z": -0.5624999860184674, "ce": 6.427689552307129, "genuineness": 4.2374773025512695, "identity": 0.3454686403274536, "hypothesis": "Your message arrived after I had left. I read it twice, before putting the phone away. Let me know when we can talk.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/db192910e9ae67de251fb2a990940eab881f66f627c78b50b638ce61f5a63679.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "dba0b66394d4283a964887ec95a66169e6287110e8f12ed3735c7bbb59ed9fca", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "efb65bff82a9e771d0224eaafcc05020545dbe8ddda3a6b2f4ce192c1d6db835", "generation_seconds_batch": 21.266878924914636, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.034618377685547, "target_z": 0.7929687730112724, "ce": 5.474569320678711, "genuineness": 4.570107460021973, "identity": 0.6517766714096069, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dba0b66394d4283a964887ec95a66169e6287110e8f12ed3735c7bbb59ed9fca.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "dc086df64ecb2d2e422d7fb0d3654c9998c73d89aa5e60e48c12a27187a51c4b", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] I found the little box.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] It was beside the old window.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7aaf24fb89c87ca9224e9412783833e8fde7ccd5f4a113b01f3555d34910fb40", "generation_seconds_batch": 18.140566439018585, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.092494010925293, "target_z": -0.4218750361189592, "ce": 6.332377910614014, "genuineness": 4.632867813110352, "identity": 0.31163787841796875, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dc086df64ecb2d2e422d7fb0d3654c9998c73d89aa5e60e48c12a27187a51c4b.mp3" }, { "id": "heldout-en-4", "language": "en", "sentences": [ "I wrote the number on the envelope.", "Read it carefully... and check the last two digits.", "Call me when you have a moment." ], "text": "I wrote the number on the envelope. Read it carefully... and check the last two digits. Call me when you have a moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-4", "key": "dc243e7761cd9b508e1dd227c1b9351ff809986bbc9d72ee791e801f88619687", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I wrote the number on the envelope.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Call me when you have a moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I wrote the number on the envelope.\n(Natural clear speech.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Natural clear speech.) [3.2 seconds duration] Call me when you have a moment.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e4566efb37185d94d19c640787da0cc15ee6f7dc0597ce975f2cea9947f74b82", "generation_seconds_batch": 17.281083123991266, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.003394603729248, "target_z": 0.7734375786461208, "ce": 5.935244083404541, "genuineness": 4.3504462242126465, "identity": 0.6687805652618408, "hypothesis": "I wrote the number on the envelope. Read it carefully and check the last two digits. Call me when you have a moment.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dc243e7761cd9b508e1dd227c1b9351ff809986bbc9d72ee791e801f88619687.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "dc3064be3d19f460905617d42ed965c0e7c35bc583afd19e815e2fb8feae06a9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 1.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4146ef473d6519cf9cff18665c8793bb5bd099ac0a145ef249aa053394410c1b", "generation_seconds_batch": 21.756479145027697, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.502976417541504, "target_z": 1.0859375833066318, "ce": 5.371314525604248, "genuineness": 4.582659721374512, "identity": 0.6144590973854065, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dc3064be3d19f460905617d42ed965c0e7c35bc583afd19e815e2fb8feae06a9.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "dc584ea09b230d6aab068df3f6d6220e6b05447eb14189c7d66b04285d9d0154", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 1.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6066ec9ea4efd109caf8c40d37fe99462a14c18a87c247e6e8d7b9ac4b9e716a", "generation_seconds_batch": 21.28497436293401, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7107264995574951, "target_z": -0.0351562840799857, "ce": 6.062326908111572, "genuineness": 4.645420074462891, "identity": 0.7395254373550415, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dc584ea09b230d6aab068df3f6d6220e6b05447eb14189c7d66b04285d9d0154.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "dc7685a5a80419e9e1f414b982b355431118995c118eec7817becc040e614df6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2486fad734847ca9c52604e6b582abc261fce237f90bf7cd9c631cc0853af24c", "generation_seconds_batch": 16.588223517057486, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.828540802001953, "target_z": 0.6640624726194987, "ce": 5.446769714355469, "genuineness": 4.507347106933594, "identity": 0.6307690739631653, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dc7685a5a80419e9e1f414b982b355431118995c118eec7817becc040e614df6.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "dca2cb513bc0b83261dedb092cd9e39d29b772aec09b041d34f8bde02b201ea2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7dfe56e844ece24266c5c5dd24e7b65578854d6e07a86720b6a61e1ac3ab1fef", "generation_seconds_batch": 22.252671992057003, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.198655128479004, "target_z": -0.3554687963138267, "ce": 6.252951145172119, "genuineness": 4.6705241203308105, "identity": 0.5593185424804688, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen daar samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dca2cb513bc0b83261dedb092cd9e39d29b772aec09b041d34f8bde02b201ea2.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "dcae66ce55e7e760eac2149b25708cbea32d5e368ea2621e5301a03ff54e029e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1003d21b78d1d0a4bfa0f452c219d686bd5abcfdce8151bfd45e0d0e162e3bfb", "generation_seconds_batch": 22.29458044108469, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4507322311401367, "target_z": 0.4277344540830437, "ce": 5.557967185974121, "genuineness": 4.6705241203308105, "identity": 0.5517285466194153, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dcae66ce55e7e760eac2149b25708cbea32d5e368ea2621e5301a03ff54e029e.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "dcdc8acb5f728682a6a4104d84f10eadee7f0ee7e40b9955ac30d68d818e7b5b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.75, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a0e9d697481292e6a62fb3fb5218176022edbc1295474b08d7ba6165d79d9346", "generation_seconds_batch": 21.44586426403839, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7785825729370117, "target_z": 0.6328124423261781, "ce": 5.371314525604248, "genuineness": 4.582659721374512, "identity": 0.6266835331916809, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dcdc8acb5f728682a6a4104d84f10eadee7f0ee7e40b9955ac30d68d818e7b5b.mp3" }, { "id": "heldout-de-0", "language": "de", "sentences": [ "Deine Nachricht kam, als ich schon weg war.", "Ich las sie zweimal... dann legte ich das Handy weg.", "Sag mir, wann wir reden können." ], "text": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal... dann legte ich das Handy weg. Sag mir, wann wir reden können.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-0", "key": "dce97c8ef816d74c5dea1a53bcfdf97474a9a63a44f8c33a5e5dfe80678a2d8a", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wann wir reden können.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Natural clear speech.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wann wir reden können.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "84831ea38a511e23fc1506a8cc0c7e19d9803d7e0be1a5cb49e2ffb8d902f6ee", "generation_seconds_batch": 21.96419238904491, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7598483562469482, "target_z": 0.6210937555343566, "ce": 5.117149353027344, "genuineness": 4.400654315948486, "identity": 0.4618930220603943, "hypothesis": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal. Dann legte ich das Handy weg. Sag mir, wann wir reden können.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dce97c8ef816d74c5dea1a53bcfdf97474a9a63a44f8c33a5e5dfe80678a2d8a.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "dd6bb6426a7709c796109ef804e7c7cec48a5cc1e062da61957c6b31e00ab4ee", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 1.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "080bcb627f2f7e4ca99cd9e2e26ff60c7371d78829a999dcdcdf4c1c0ac975df", "generation_seconds_batch": 20.67529825901147, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.360182762145996, "target_z": 0.3710936623241394, "ce": 5.911416053771973, "genuineness": 4.68307638168335, "identity": 0.6675888299942017, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dd6bb6426a7709c796109ef804e7c7cec48a5cc1e062da61957c6b31e00ab4ee.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "dd9370edfc07ebc73981b04994fbf4a8a794d86d002bd538eb43c470120bd960", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3f13f300b83b2a8851ba4bafac1690e8f23de9ca883f7b4e649afe80f1da3ba8", "generation_seconds_batch": 16.27042228600476, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.003394603729248, "target_z": 0.7734375786461208, "ce": 5.518253803253174, "genuineness": 4.65797233581543, "identity": 0.5756145715713501, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dd9370edfc07ebc73981b04994fbf4a8a794d86d002bd538eb43c470120bd960.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "dd99c3ec6dd1ea668d37d9936eacc96451d2c1541e3a4829544605658022a4c3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "fed90c5484ac8cee740fd0b0c4b1b2b55b226c0e7ecce7ee7a91839a9434e1c8", "generation_seconds_batch": 22.442758857971057, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5904030799865723, "target_z": 1.1406249871835952, "ce": 5.792276382446289, "genuineness": 4.457139015197754, "identity": 0.5488362908363342, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dd99c3ec6dd1ea668d37d9936eacc96451d2c1541e3a4829544605658022a4c3.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "de1606e6c6b00d54735de7df25588e19c68b4d19c4e8752d6160daa24f9f56ee", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d72c37f989063295ef91a51ced8df9105f2700c7e3e37e05600eeca72825a0f0", "generation_seconds_batch": 16.961241595097817, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1892880201339722, "target_z": -0.36132813970973743, "ce": 5.9431867599487305, "genuineness": 4.582659721374512, "identity": 0.685520350933075, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/de1606e6c6b00d54735de7df25588e19c68b4d19c4e8752d6160daa24f9f56ee.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "de202a1c62a463d2cb7018961fa1b37f2739e7cc0827993e994989b28741b855", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "499926de1b11c71ad70f70d73473e58fa43288468c0f5808ec2a4735f3729483", "generation_seconds_batch": 27.090455877012573, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6661767959594727, "target_z": 0.5625000233025543, "ce": 5.895530700683594, "genuineness": 4.545003414154053, "identity": 0.6802157163619995, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/de202a1c62a463d2cb7018961fa1b37f2739e7cc0827993e994989b28741b855.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "df1cdc9ad58940f4b1cce9b066863ed4f6bd378512c01b0ee19e6d0feb89ef03", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "45bf813a96d085efabcfd8c65b17e830c15bcb834aa1515b9843672d376dc77b", "generation_seconds_batch": 16.823258957010694, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.715298652648926, "target_z": 1.2187500629168966, "ce": 5.6334228515625, "genuineness": 4.758388519287109, "identity": 0.6084152460098267, "hypothesis": "La porte est encore ouverte j'ai posé ton manteau sur la chaise, dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/df1cdc9ad58940f4b1cce9b066863ed4f6bd378512c01b0ee19e6d0feb89ef03.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "df5911e6c596645ce1d516ba223d11b8e8ae7fdc7e9f198a6db6c08f27fca95c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0c66227dda0661191664d2a452c05168979baf34ba373332130caaf48dc1e331", "generation_seconds_batch": 16.358972711022943, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5654239654541016, "target_z": 1.1249999720369348, "ce": 5.530168056488037, "genuineness": 4.557555198669434, "identity": 0.5232107043266296, "hypothesis": "We found the little box. It was beside the old window. We can look inside together.", "wer": 0.0625, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/df5911e6c596645ce1d516ba223d11b8e8ae7fdc7e9f198a6db6c08f27fca95c.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "df8c69a2cb85094da5005facb46a395da5c33d827d7645f8cdd438598031fd43", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ebdcf4adfbb31922c52f239aeed51b3e627829037e1a2f6f6482deeb5317a236", "generation_seconds_batch": 17.680610417970456, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8410303592681885, "target_z": 0.6718749801928289, "ce": 5.871703147888184, "genuineness": 4.557555198669434, "identity": 0.2518325746059418, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/df8c69a2cb85094da5005facb46a395da5c33d827d7645f8cdd438598031fd43.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "dfb4b31d16f601c2fce12cfd5f20b2b175f87c7e6fd0eeb3dc7f76da4266b1ba", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.75, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "07f1052216ac854e90d66cc28e7612d28b3f5eda0fe3dad98047632677ac69f0", "generation_seconds_batch": 21.202916178037412, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3905704021453857, "target_z": 1.0156250151466604, "ce": 5.6334228515625, "genuineness": 4.406930446624756, "identity": 0.41940099000930786, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/dfb4b31d16f601c2fce12cfd5f20b2b175f87c7e6fd0eeb3dc7f76da4266b1ba.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "e0c9bfcad52d37d90683e58d2733641288fe2faf453187dde5166423f1de0ac8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b7fd8fb674c97906da207d193d9076becd80cb0b3f833f7b29e4811caa759628", "generation_seconds_batch": 21.675431331968866, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.259430170059204, "target_z": 0.9335937601948675, "ce": 5.760505676269531, "genuineness": 4.7458367347717285, "identity": 0.8501188158988953, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur le sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.1111111111111111, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e0c9bfcad52d37d90683e58d2733641288fe2faf453187dde5166423f1de0ac8.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "e0fb805d1456c66a1d4249d6891953c37c0d576762c46e283e1ae02f0b3c66e6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 1.0, "cfg": 1.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "9b87f6796dd50083075c13c392f3f9615b1c2a86c5e85cd907bb43371bec539e", "generation_seconds_batch": 21.840928632067516, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3820395469665527, "target_z": 0.38476558786155407, "ce": 6.157639026641846, "genuineness": 4.400654315948486, "identity": 0.34906715154647827, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e0fb805d1456c66a1d4249d6891953c37c0d576762c46e283e1ae02f0b3c66e6.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "e118627d711ff695d713b77052b8de93cbfa8ac4d1cebfb1aaf52692551b2cf8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1702, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e48bc0bd5d4df33618decaef718090067594308cd466af30c9071732574ec86c", "generation_seconds_batch": 21.465164726017974, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.342284917831421, "target_z": -0.2656250337887038, "ce": 6.0305562019348145, "genuineness": 4.482243061065674, "identity": 0.6133525371551514, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e118627d711ff695d713b77052b8de93cbfa8ac4d1cebfb1aaf52692551b2cf8.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "e236ebb3b979cfe672f107460e206ca0cb58c73194bf5c0332a93997156a0fec", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "32af69ed64d1acf8c9230fb033e1f3f57815bdebeb602df35961af2a916ed583", "generation_seconds_batch": 23.380126139032654, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5920758247375488, "target_z": -0.10937503145844833, "ce": 6.1020402908325195, "genuineness": 4.7458367347717285, "identity": 0.5999668836593628, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e236ebb3b979cfe672f107460e206ca0cb58c73194bf5c0332a93997156a0fec.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "e2c888a780e6885be06abc86379be20d79433787990010f78ac9036158214e88", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-f7af130df2f4", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Adopt a warm, conspiratorial whisper, as if you are sharing a beautiful, fragile truth with a trusted confidant.\nSCRIPT:\n(A warm, low register whisper, inviting the listener in with gentle intimacy.) [3.2 seconds duration] La puerta sigue abierta.\n(The whisper should be smooth and flowing, like silk, with a gentle, confidential cadence.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Sustain the warmth while keeping the volume at the absolute minimum, a hushed confidence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "40b95c0f2618217b4b0fe1180808608d15568b661bb49c3e22574636624c4aa1", "generation_seconds_batch": 15.840117499930784, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.39848792552948, "target_z": -0.23046874970871808, "ce": 5.903473377227783, "genuineness": 4.2751336097717285, "identity": 0.3390194773674011, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e2c888a780e6885be06abc86379be20d79433787990010f78ac9036158214e88.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "e2d60e1caba39de09d68145f9d69ad567c5778953bf2ca130af20ce674f0d41d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "d12d5f605c30e07ce920c4c3011798b52cfe2fed1094e6f961e38ce4bb391aa4", "generation_seconds_batch": 16.688697982928716, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4155495166778564, "target_z": 1.0312500302933205, "ce": 5.478540420532227, "genuineness": 4.432034492492676, "identity": 0.21164552867412567, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e2d60e1caba39de09d68145f9d69ad567c5778953bf2ca130af20ce674f0d41d.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "e3593e3ad7ed3701a329e89b757bcc215881d26062ca194180b53cd75345add0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "69b4d1c54624f649f1afc3d7e5de6cb72461e53278e572f5c4d85bd71061babc", "generation_seconds_batch": 22.35039490298368, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6287081241607666, "target_z": 0.5390625005825639, "ce": 5.550024509429932, "genuineness": 4.6956281661987305, "identity": 0.4506320357322693, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e3593e3ad7ed3701a329e89b757bcc215881d26062ca194180b53cd75345add0.mp3" }, { "id": "heldout-es-0", "language": "es", "sentences": [ "Tu mensaje llegó después de que me fuera.", "Lo leí dos veces... y guardé el teléfono.", "Dime cuándo podemos hablar." ], "text": "Tu mensaje llegó después de que me fuera. Lo leí dos veces... y guardé el teléfono. Dime cuándo podemos hablar.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-0", "key": "e3bfbd4eb28bbec5285a1673909e7532c36de838c1bb7200af7b68d9e504659c", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo podemos hablar.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Natural clear speech.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo podemos hablar.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cc4d01b346d234dff60aeca107b8a22bb44caa103a139f429c73f3d0c10e4715", "generation_seconds_batch": 22.732390693039633, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2469406127929688, "target_z": 0.9257812526215373, "ce": 5.115163326263428, "genuineness": 4.362998008728027, "identity": 0.45328983664512634, "hypothesis": "Tu mensaje llegó después de que me fuera. Lo leí dos veces y guardé el teléfono. Dime cuándo podemos hablar.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e3bfbd4eb28bbec5285a1673909e7532c36de838c1bb7200af7b68d9e504659c.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "e3d26139ebc4dd7deba40eeb07c8d129c105fc8f9e15086967cf9bf904a637e5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6be50d633036199dd96281d36140b147d831681efbded75e790101402d3679d2", "generation_seconds_batch": 22.183518590056337, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0456583499908447, "target_z": -0.4511718276666865, "ce": 6.133810997009277, "genuineness": 4.444586753845215, "identity": 0.5583418607711792, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e3d26139ebc4dd7deba40eeb07c8d129c105fc8f9e15086967cf9bf904a637e5.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "e3f50e03610b0770a2c283842f821b163b8941a2ddd69675bd6cc51356ee794f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "dfb760f15cb004b016b18a05dce945092e937281b711c43ce92f848e3867227e", "generation_seconds_batch": 17.112483047996648, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6911559104919434, "target_z": 0.5781250384492146, "ce": 5.855817794799805, "genuineness": 4.582659721374512, "identity": 0.6344817876815796, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e3f50e03610b0770a2c283842f821b163b8941a2ddd69675bd6cc51356ee794f.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "e42d2eefc33d8ee3cd63a3c2c944c34391cf8fd05b033a3396e952d03527ebb8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0380d88352977b7e6d5d99b417596c7ed334cd4ea446e84b1b046eca9691ec6d", "generation_seconds_batch": 16.98454019194469, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3905704021453857, "target_z": 1.0156250151466604, "ce": 5.903473377227783, "genuineness": 4.7081804275512695, "identity": 0.6966730356216431, "hypothesis": "J'ai trouvé la petite boîte, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e42d2eefc33d8ee3cd63a3c2c944c34391cf8fd05b033a3396e952d03527ebb8.mp3" }, { "id": "heldout-fr-0", "language": "fr", "sentences": [ "Ton message est arrivé après mon départ.", "Je l’ai relu... puis j’ai rangé mon téléphone.", "Dis-moi quand nous pourrons parler." ], "text": "Ton message est arrivé après mon départ. Je l’ai relu... puis j’ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-0", "key": "e4722d9700580945fc554ef6c5befd7e2c4bd351f69104f569e079dc784a7202", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Natural clear speech.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3173b291f46c504867f7e0ab5e83f995926fa7e48e01ce41f930888a61195bef", "generation_seconds_batch": 22.44589341408573, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.51546573638916, "target_z": 1.0937499417436143, "ce": 5.027793884277344, "genuineness": 4.344170093536377, "identity": 0.6118291616439819, "hypothesis": "Ton message est arrivé après mon départ. Je l'ai relu, puis j'ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e4722d9700580945fc554ef6c5befd7e2c4bd351f69104f569e079dc784a7202.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "e4dfc7dc4e9e3065b392614f294c20d6dbc11e8a99325d85cf3eb3282b502caf", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1c2ab3a4609531b34f64adba4a3a5a05b4914d6f1858d547fce92585e795c3fd", "generation_seconds_batch": 17.251369801000692, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5654239654541016, "target_z": 1.1249999720369348, "ce": 5.248203277587891, "genuineness": 4.65797233581543, "identity": 0.510909914970398, "hypothesis": "He encontrado la cajita. estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e4dfc7dc4e9e3065b392614f294c20d6dbc11e8a99325d85cf3eb3282b502caf.mp3" }, { "id": "heldout-es-4", "language": "es", "sentences": [ "Escribí el número en el sobre.", "Léelo bien... y comprueba las dos últimas cifras.", "Llámame cuando tengas un momento." ], "text": "Escribí el número en el sobre. Léelo bien... y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-4", "key": "e55ab48a96a29b8b3f45e8faca2f0d5d1191a867f0b010213f2d13ab0689467f", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Escribí el número en el sobre.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Llámame cuando tengas un momento.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Escribí el número en el sobre.\n(Natural clear speech.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Natural clear speech.) [3.2 seconds duration] Llámame cuando tengas un momento.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "2e08172ad287f50140db5c5eed46ed8a7779fa84177f57f57ffd0ae72b0a811a", "generation_seconds_batch": 20.00326439004857, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0221288204193115, "target_z": 0.7851562654379423, "ce": 5.367343425750732, "genuineness": 4.319066047668457, "identity": 0.570122241973877, "hypothesis": "Escribí el número en el sobre, léelo bien y comprueba las dos últimas cifras. Llámame cuando tengas un momento.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e55ab48a96a29b8b3f45e8faca2f0d5d1191a867f0b010213f2d13ab0689467f.mp3" }, { "id": "heldout-nl-1", "language": "nl", "sentences": [ "De sleutels liggen op de keukentafel.", "Ik keek in de la... en daarna naast het raam.", "Neem ze mee als je weggaat." ], "text": "De sleutels liggen op de keukentafel. Ik keek in de la... en daarna naast het raam. Neem ze mee als je weggaat.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-1", "key": "e6569775d22df3978a15673b5af52aeca8e3f85e456584447bb0b87476ad38b4", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Neem ze mee als je weggaat.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Natural clear speech.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Natural clear speech.) [3.2 seconds duration] Neem ze mee als je weggaat.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b06672741bb8d2907cc9b97915453a1689680c237d1508494cb62b6d504300a1", "generation_seconds_batch": 16.90157121606171, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.109555721282959, "target_z": 0.8398438184512533, "ce": 5.371314525604248, "genuineness": 4.243753433227539, "identity": 0.5040663480758667, "hypothesis": "De sleutels liggen op de keukentafel. Je kik in de la en daarna naast het raam neem ze mee als je weggaat.", "wer": 0.09090909090909091, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e6569775d22df3978a15673b5af52aeca8e3f85e456584447bb0b87476ad38b4.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "e66b924509dfe7ad4ec6240e5aebbe62789d96db71ce363e78e1c5f93f8d97ca", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "52108fdcb21db3edf087306babc5655edaa566aa79ee2287baaa5894f764bb36", "generation_seconds_batch": 17.680610417970456, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.866009473800659, "target_z": 0.6874999953394891, "ce": 5.95112943649292, "genuineness": 4.545003414154053, "identity": 0.41631996631622314, "hypothesis": "Ik heb een kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e66b924509dfe7ad4ec6240e5aebbe62789d96db71ce363e78e1c5f93f8d97ca.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "e68a2ea8fdb5f612dbdaae468982128dfe3f39d19ff3898933f53e18e2a16f8e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "dab6925e14760c51b8fbc7c073485d9ec236f835a1885ac3bac0106b1529a949", "generation_seconds_batch": 23.699324600049295, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9917412996292114, "target_z": 0.14062498718359512, "ce": 6.0146708488464355, "genuineness": 4.921566009521484, "identity": 0.6699421405792236, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e68a2ea8fdb5f612dbdaae468982128dfe3f39d19ff3898933f53e18e2a16f8e.mp3" }, { "id": "heldout-en-3", "language": "en", "sentences": [ "Someone moved the little blue chair.", "It used to stand here... beside the tall bookshelf.", "We can put it back after lunch." ], "text": "Someone moved the little blue chair. It used to stand here... beside the tall bookshelf. We can put it back after lunch.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-3", "key": "e6f1e80cc6606ef00fdc322d5829b94fd65ad80a74ab015f88321e5dc194804c", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Someone moved the little blue chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can put it back after lunch.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Someone moved the little blue chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can put it back after lunch.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f1f6996e45cce767b5fd84e736e2fe70c6c9d562a6bcb13f23b73d018f5a3318", "generation_seconds_batch": 17.12878136395011, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.4180586338043213, "target_z": -0.8437499976697446, "ce": 6.403861999511719, "genuineness": 4.168440818786621, "identity": 0.1601288765668869, "hypothesis": "Someone moved the little blue chair. It used to stand here, beside the tall bookshelf. We can put it back after lunch.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e6f1e80cc6606ef00fdc322d5829b94fd65ad80a74ab015f88321e5dc194804c.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "e783e3f8612df9769565b536ebc90c0d1d2d40bfcfeae57b6e934a56fb75bbd5", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "57415da2a30620b83e7002e9f323a35399ba115e8a16dc81b61fabd6ca985de6", "generation_seconds_batch": 23.013191931997426, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.072086811065674, "target_z": 0.8164061465949152, "ce": 5.287916660308838, "genuineness": 4.143336772918701, "identity": 0.41039425134658813, "hypothesis": "Ik heb een klein met kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.16666666666666666, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e783e3f8612df9769565b536ebc90c0d1d2d40bfcfeae57b6e934a56fb75bbd5.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "e79d19bdb556fb6737c7c2e504598b24173136fbf0693cefc2dbd3e3241c803c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "337f73bdc4cfb7057395067d05571e3fafe4ee868932db5b45c8d41cc1681cb4", "generation_seconds_batch": 22.90019148599822, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.652850866317749, "target_z": 1.1796875250502459, "ce": 5.617537498474121, "genuineness": 4.595211505889893, "identity": 0.8131540417671204, "hypothesis": "J'ai trouvé des petites boîtes, elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.15789473684210525, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e79d19bdb556fb6737c7c2e504598b24173136fbf0693cefc2dbd3e3241c803c.mp3" }, { "id": "heldout-fr-1", "language": "fr", "sentences": [ "Les clés sont sur la table de la cuisine.", "J’ai regardé dans le tiroir... puis près de la fenêtre.", "Prends-les quand tu partiras." ], "text": "Les clés sont sur la table de la cuisine. J’ai regardé dans le tiroir... puis près de la fenêtre. Prends-les quand tu partiras.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-1", "key": "e7b286bf7756944fc54b33d551285c234e99d5e36ad04254244ac8da48a08c9b", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Prends-les quand tu partiras.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Natural clear speech.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Prends-les quand tu partiras.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5f860e1532e73c58528c31b121e5e6bd7723a945c813b728e370ef5888161b05", "generation_seconds_batch": 24.369751867023297, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.284409284591675, "target_z": 0.9492187753415279, "ce": 5.315715789794922, "genuineness": 4.532451152801514, "identity": 0.7170402407646179, "hypothesis": "Les clés sont sur la table de la cuisine. J'ai regardé dans le tiroir, puis près de la fenêtre. Prends-les quand tu partiras.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e7b286bf7756944fc54b33d551285c234e99d5e36ad04254244ac8da48a08c9b.mp3" }, { "id": "heldout-fr-2", "language": "fr", "sentences": [ "La route tourne à gauche après le pont.", "Continue... jusqu’à la petite barrière en bois.", "Je t’attendrai de l’autre côté." ], "text": "La route tourne à gauche après le pont. Continue... jusqu’à la petite barrière en bois. Je t’attendrai de l’autre côté.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-2", "key": "e8555738a7c5c709e889b433de0827e95048203dabc6469e5ccb5695ccd2d4d4", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Natural clear speech.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Natural clear speech.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "98a26dd9156e765fe0bbbe1b217c4b050a7c7aa1f4dc84df36e27ed34f7bdc52", "generation_seconds_batch": 22.683973697945476, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.827704429626465, "target_z": 1.2890624819405203, "ce": 5.589737892150879, "genuineness": 4.494794845581055, "identity": 0.6744328737258911, "hypothesis": "En route tourne à gauche après le pont. Continue jusqu'à la petite barrière en bois. Je t'attendrai de l'autre côté.", "wer": 0.05, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e8555738a7c5c709e889b433de0827e95048203dabc6469e5ccb5695ccd2d4d4.mp3" }, { "id": "heldout-nl-1", "language": "nl", "sentences": [ "De sleutels liggen op de keukentafel.", "Ik keek in de la... en daarna naast het raam.", "Neem ze mee als je weggaat." ], "text": "De sleutels liggen op de keukentafel. Ik keek in de la... en daarna naast het raam. Neem ze mee als je weggaat.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-1", "key": "e86962d30123c91fdf1e6ea5bcb4753541557e5ab4e6c32f294957f04221c8e9", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Neem ze mee als je weggaat.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Natural clear speech.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Natural clear speech.) [3.2 seconds duration] Neem ze mee als je weggaat.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "dc518bf6397a34942dfa5049c9e11d15ff7959f7b0cecc0326ce6e15a3d16d17", "generation_seconds_batch": 15.741536403889768, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0533528327941895, "target_z": 0.8046876089394415, "ce": 5.414999485015869, "genuineness": 4.262581825256348, "identity": 0.2309018075466156, "hypothesis": "De sleutels liggen op de keukentafel. Je kikken in de la en daarna naast het raam. Neem ze mee als je weggaat.", "wer": 0.09090909090909091, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e86962d30123c91fdf1e6ea5bcb4753541557e5ab4e6c32f294957f04221c8e9.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "e92dd369bf39166cc0047b88c0535ef748879f85486777d7212c843bcc33635e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8edace0cf372d47df005df6993fdcd77b4864127b676771ab4bd3e615ef8c401", "generation_seconds_batch": 21.436049563926645, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6974005699157715, "target_z": 0.5820312176677058, "ce": 5.95112943649292, "genuineness": 4.6956281661987305, "identity": 0.5139604210853577, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e92dd369bf39166cc0047b88c0535ef748879f85486777d7212c843bcc33635e.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "e9b94e62940e9919ffeaca3ae0e309de24b5781b52092559ebbe256786af8237", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ad75a34e61c48816cb7dd7d30cc237e40a1543c1d35621d708b8c299da23ea9d", "generation_seconds_batch": 23.699324600049295, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.175962209701538, "target_z": 0.255859436606128, "ce": 5.8875885009765625, "genuineness": 4.8588056564331055, "identity": 0.6354474425315857, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e9b94e62940e9919ffeaca3ae0e309de24b5781b52092559ebbe256786af8237.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "e9c6eedbb289ffd6e2186b6aff97f7d40e76af9d1a1eebe12ed517797dcdcbf8", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3b4ea8dfac5deb82d84e90153bed08e7c8b561ccf49b21f4a07d7083339463b0", "generation_seconds_batch": 16.156822989927605, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.417222261428833, "target_z": -0.21874998834872283, "ce": 6.252951145172119, "genuineness": 4.758388519287109, "identity": 0.6597793102264404, "hypothesis": "Ich habe die kleine Schacht gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0625, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e9c6eedbb289ffd6e2186b6aff97f7d40e76af9d1a1eebe12ed517797dcdcbf8.mp3" }, { "id": "heldout-de-3", "language": "de", "sentences": [ "Jemand hat den kleinen blauen Stuhl umgestellt.", "Er stand hier... neben dem hohen Bücherregal.", "Wir stellen ihn nach dem Essen zurück." ], "text": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier... neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-3", "key": "e9d266c3907611a53d0c5f7688a7e231bba29ead6f87b8ae8d29cb37ab2fae41", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8107e5a6a290c141924435272ce9529f62635f738a3436b73ed7b7799f377037", "generation_seconds_batch": 17.56127679697238, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.19949162006378174, "target_z": -0.9804687310666746, "ce": 6.483288764953613, "genuineness": 4.582659721374512, "identity": 0.34335726499557495, "hypothesis": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/e9d266c3907611a53d0c5f7688a7e231bba29ead6f87b8ae8d29cb37ab2fae41.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "ea7fefbd942805bd3891d747084d721d36b12cc0df5313818fe816c7c742fb09", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.5, "cfg": 5.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "6800aa7f8003f597d215fb871569d3b50dd9408a80f87c21aadd1e04b5d19f62", "generation_seconds_batch": 22.65826919605024, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7527670860290527, "target_z": 1.2421874365005394, "ce": 5.609594821929932, "genuineness": 4.469690799713135, "identity": 0.4636603891849518, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ea7fefbd942805bd3891d747084d721d36b12cc0df5313818fe816c7c742fb09.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "ea87da2c6996d2847b4c84cac1ce52fb2421ac336a83fb5ab7e491d8c0d065b6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.0, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "68095629a2b2e62abfc89e237465ffe54ebab2e485156a161bbdbe0d4eb18608", "generation_seconds_batch": 17.71011035598349, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9301300048828125, "target_z": -0.5234374854359035, "ce": 6.1020402908325195, "genuineness": 4.168440818786621, "identity": 0.387817919254303, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ea87da2c6996d2847b4c84cac1ce52fb2421ac336a83fb5ab7e491d8c0d065b6.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "eb066507f7d7cf2b37680dee91b620ef51540c285f6213767b98533bc6ce6d8e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.0, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "07ce9394579457c8fa0965613a507b6233299fcae522cf54b41c65cabc779ec1", "generation_seconds_batch": 16.563943021930754, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.10414719581604, "target_z": 0.21093748077539268, "ce": 5.669164657592773, "genuineness": 4.557555198669434, "identity": 0.5832359790802002, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/eb066507f7d7cf2b37680dee91b620ef51540c285f6213767b98533bc6ce6d8e.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "eb6c52766731523d96500139d65ac5242f7fa3fc9b6ce2244ce4816f96b146fa", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a971ab294df368e9aa5446b5e2d05d2cfad50698b670d8f9b549317bb8420ea9", "generation_seconds_batch": 26.244926927029155, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.577913522720337, "target_z": 1.132812479610265, "ce": 5.081407070159912, "genuineness": 4.3504462242126465, "identity": 0.48779863119125366, "hypothesis": "La puerta sigue abierta. He dejado tu abrijo sobre la silla. Dime cuándo estás listo.", "wer": 0.06666666666666667, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/eb6c52766731523d96500139d65ac5242f7fa3fc9b6ce2244ce4816f96b146fa.mp3" }, { "id": "heldout-fr-2", "language": "fr", "sentences": [ "La route tourne à gauche après le pont.", "Continue... jusqu’à la petite barrière en bois.", "Je t’attendrai de l’autre côté." ], "text": "La route tourne à gauche après le pont. Continue... jusqu’à la petite barrière en bois. Je t’attendrai de l’autre côté.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-2", "key": "ebe89981a89e9506d02bacdfa583f9bfe1f04faa699193749ab6419de85c588f", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Je t’attendrai de l’autre côté.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7530f5577c3cf2ac79275d92bb84101abe05d8c30c7c5c8c11b04bbab1ed1ebc", "generation_seconds_batch": 17.4834213150898, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.979251742362976, "target_z": 0.13281247961026496, "ce": 6.046441555023193, "genuineness": 4.444586753845215, "identity": 0.8205846548080444, "hypothesis": "La route tourne à gauche après le pont, continue jusqu'à la petite barrière en bois. Je t'attendrai de l'autre côté.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ebe89981a89e9506d02bacdfa583f9bfe1f04faa699193749ab6419de85c588f.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "ec056f8ffbca4301e9135627c746b49a61c019d3809ab57403894368aac56857", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f65823209e64cc8def683966ccc55ed514fb5ac7e51164f41fd1c6553c3a3560", "generation_seconds_batch": 22.1596444620518, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.179084539413452, "target_z": 0.2578125262153736, "ce": 6.006728172302246, "genuineness": 4.369274139404297, "identity": 0.5439201593399048, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ec056f8ffbca4301e9135627c746b49a61c019d3809ab57403894368aac56857.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "ec123085306b78f53f61922df2ac2606f74ebdd6b9f5c8869c36dd6a3b4aaa3b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "23bcc64fa0997c1ad97d075307a58dfdcfd86ca7717fb00952c8d5d1c250e293", "generation_seconds_batch": 21.820619402104057, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.6014430522918701, "target_z": -0.1035156134943638, "ce": 5.974957466125488, "genuineness": 4.632867813110352, "identity": 0.7602643966674805, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ec123085306b78f53f61922df2ac2606f74ebdd6b9f5c8869c36dd6a3b4aaa3b.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "ecd8846e23d9a0439842f3bc5e37ba11f8196da5359ea889d68d8e4d0a0ab9a8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "68ab3b3f76c7a8acbd256e89e7835e2b76401eb2195c3a2983a5b665d52ed06f", "generation_seconds_batch": 21.465164726017974, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4827923774719238, "target_z": -0.17773436087282643, "ce": 6.24500846862793, "genuineness": 4.582659721374512, "identity": 0.46003323793411255, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ecd8846e23d9a0439842f3bc5e37ba11f8196da5359ea889d68d8e4d0a0ab9a8.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "ecf16f14441347e403d26a71081d9959e618c436cba3a5cfe2452f5ec3cdca7c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1702, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "54b25247db776b7d1060014570538431b4845d068e8b39775530c6dc6684a98a", "generation_seconds_batch": 22.51380237203557, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1853291988372803, "target_z": 0.26171870543386483, "ce": 5.879645347595215, "genuineness": 4.457139015197754, "identity": 0.646136999130249, "hypothesis": "Ik heb het kleine, ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.2222222222222222, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ecf16f14441347e403d26a71081d9959e618c436cba3a5cfe2452f5ec3cdca7c.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "edcfcc00cdd6ce79e0f70a654273b90dc98b5a35d4f883558f2f7b6d0e99a936", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "33709c8d5cfff34a0299048867ac734f54f8db350ac666427f009256ff199c5b", "generation_seconds_batch": 16.558907449943945, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.641197681427002, "target_z": 0.546875008155894, "ce": 5.577824115753174, "genuineness": 4.557555198669434, "identity": 0.5406486988067627, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/edcfcc00cdd6ce79e0f70a654273b90dc98b5a35d4f883558f2f7b6d0e99a936.mp3" }, { "id": "heldout-es-0", "language": "es", "sentences": [ "Tu mensaje llegó después de que me fuera.", "Lo leí dos veces... y guardé el teléfono.", "Dime cuándo podemos hablar." ], "text": "Tu mensaje llegó después de que me fuera. Lo leí dos veces... y guardé el teléfono. Dime cuándo podemos hablar.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-0", "key": "ee5a6b4d52be7d63a2e9a7483699a73b8c7ef4709d42b55e64aef9ea74c89e96", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo podemos hablar.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Natural clear speech.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo podemos hablar.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ec940e62acd96897a6ca9ae00b1fc1d1ffa7837916fb004802ce91bd1c68d3e0", "generation_seconds_batch": 22.732390693039633, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.334367275238037, "target_z": 0.9804686564985008, "ce": 5.331601142883301, "genuineness": 4.300237655639648, "identity": 0.5520340204238892, "hypothesis": "Tu mensaje llegó después de que me fuera. Lo leí dos veces y guardé el teléfono. Dime cuándo podemos hablar.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ee5a6b4d52be7d63a2e9a7483699a73b8c7ef4709d42b55e64aef9ea74c89e96.mp3" }, { "id": "heldout-nl-3", "language": "nl", "sentences": [ "Iemand heeft het kleine blauwe stoeltje verplaatst.", "Het stond hier... naast de hoge boekenkast.", "We zetten het na de lunch terug." ], "text": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier... naast de hoge boekenkast. We zetten het na de lunch terug.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-3", "key": "ee97f6e98a1b762c7e8768850f002ee0f1a8818b48f33eee0e5e8052966ad142", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We zetten het na de lunch terug.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Natural clear speech.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Natural clear speech.) [3.2 seconds duration] We zetten het na de lunch terug.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "8f231b15a6c6dfa5b9a0099f012fda3e0f1aeff7cf7033fab00ce118596fbe88", "generation_seconds_batch": 21.546583125018515, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.866009473800659, "target_z": 0.6874999953394891, "ce": 5.9431867599487305, "genuineness": 4.65797233581543, "identity": 0.3303645849227905, "hypothesis": "Iemand heeft het kleine blauwe stoeltje verplaatst. Het stond hier naast de hoge boekenkast. We zetten het naar de lunch terug.", "wer": 0.047619047619047616, "duration_error_s": 0.120000000000001, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ee97f6e98a1b762c7e8768850f002ee0f1a8818b48f33eee0e5e8052966ad142.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "ee9e7b513c251c89fa1affa554eb21327175aaf2f42fe1ca55d6e61953517280", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 1.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bc4bc255a88b6024f194897624b623407cc48e01633fc9ef1d33a3a4e71d8d58", "generation_seconds_batch": 22.09680682606995, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0456583499908447, "target_z": -0.4511718276666865, "ce": 6.062326908111572, "genuineness": 4.413206577301025, "identity": 0.5103887915611267, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ee9e7b513c251c89fa1affa554eb21327175aaf2f42fe1ca55d6e61953517280.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "eea683c75bd9d09c48e157b66385a4267d5856ec2c8a60fee59048e09397c80a", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "39d3a783de1ee8d8f42bec28c6bf1a1f0ff9b37d8e4fb0b65df5f050b8bfc666", "generation_seconds_batch": 24.090297435992397, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.791072368621826, "target_z": 0.6406250990358558, "ce": 6.141753673553467, "genuineness": 4.3881025314331055, "identity": 0.4720958471298218, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/eea683c75bd9d09c48e157b66385a4267d5856ec2c8a60fee59048e09397c80a.mp3" }, { "id": "heldout-de-3", "language": "de", "sentences": [ "Jemand hat den kleinen blauen Stuhl umgestellt.", "Er stand hier... neben dem hohen Bücherregal.", "Wir stellen ihn nach dem Essen zurück." ], "text": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier... neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-3", "key": "ef37c51b58911f7c63c8f3b6c7ec04d6c60e77bf38c897a684c957acc2e81964", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Natural clear speech.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Natural clear speech.) [3.2 seconds duration] Wir stellen ihn nach dem Essen zurück.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d8c0d1df57b161fb0c711474a52a4426d11fec3fc0888b0b51e5cd2497c293d7", "generation_seconds_batch": 16.903751563979313, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.853519916534424, "target_z": 0.679687487766159, "ce": 5.331601142883301, "genuineness": 4.331617832183838, "identity": 0.5945907831192017, "hypothesis": "Jemand hat den kleinen blauen Stuhl umgestellt. Er stand hier neben dem hohen Bücherregal. Wir stellen ihn nach dem Essen zurück.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ef37c51b58911f7c63c8f3b6c7ec04d6c60e77bf38c897a684c957acc2e81964.mp3" }, { "id": "heldout-nl-0", "language": "nl", "sentences": [ "Je bericht kwam nadat ik vertrokken was.", "Ik las het twee keer... en legde mijn telefoon weg.", "Zeg maar wanneer we kunnen praten." ], "text": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer... en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-0", "key": "ef466669943d989886ebbe9c03557f9104dfa5e94c726674c3a4c1fd40941424", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Natural clear speech.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer we kunnen praten.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cb0629d1a6a92507758877d12b9c7f9ece59cc585c97f21d2397c1af235edf29", "generation_seconds_batch": 17.088680595043115, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8160512447357178, "target_z": 0.6562499650461685, "ce": 5.569881439208984, "genuineness": 4.356721878051758, "identity": 0.6012935638427734, "hypothesis": "Je bericht kwam nadat ik vertrokken was, ik las het twee keer en legde mijn telefoon weg. Zeg maar wanneer we kunnen praten.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ef466669943d989886ebbe9c03557f9104dfa5e94c726674c3a4c1fd40941424.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "efa9387be5688ec0da60422ab4212ac679c771e6098149d4fba3766ffad323f0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1701, "scale": 0.5, "cfg": 2.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "da7c873ee7d54ef3c8986fcb54ea639d98f8eb9c86c50e49e3520569bdf5d0db", "generation_seconds_batch": 23.030156910070218, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8449891805648804, "target_z": 0.048828135049226555, "ce": 5.847874641418457, "genuineness": 4.65797233581543, "identity": 0.714322566986084, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/efa9387be5688ec0da60422ab4212ac679c771e6098149d4fba3766ffad323f0.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "efae6c1a588bc70b1c3f3d842f2475ad43edeaea952ed8f3cb1f0ad5d89ca438", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "31bc43fefb7f33d72ff6f48ab042461a5d6e86baafee2916f7e9e96246429033", "generation_seconds_batch": 27.090455877012573, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8347854614257812, "target_z": 0.6679686518379899, "ce": 5.605623245239258, "genuineness": 4.6705241203308105, "identity": 0.5325648188591003, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/efae6c1a588bc70b1c3f3d842f2475ad43edeaea952ed8f3cb1f0ad5d89ca438.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "efcf7b7482423f4df6c8eb7480b2b175e68115e734efde624e2c47872654ff90", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1701, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "df2224fc97bfdf112a238cd18fa7887293bc2f9cbee5b599cd0991d0022a5f08", "generation_seconds_batch": 20.677407835028134, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.7357056140899658, "target_z": -0.01953126893332539, "ce": 6.117925643920898, "genuineness": 4.607763767242432, "identity": 0.5817959308624268, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/efcf7b7482423f4df6c8eb7480b2b175e68115e734efde624e2c47872654ff90.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "f002502b15b223cb13ce8fd4910659fcd30cb9733216f1a19a981c21bce623f0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a6a9586cfe1719d82931cf16b2a591d19232608723e9174331250640f64bb0ea", "generation_seconds_batch": 24.00437844498083, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9417831897735596, "target_z": 0.10937503145844833, "ce": 6.09409761428833, "genuineness": 4.7834930419921875, "identity": 0.617527961730957, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f002502b15b223cb13ce8fd4910659fcd30cb9733216f1a19a981c21bce623f0.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "f007afcefa87b939056a003eb45bcb7d3c23aa3a3b9a135714e2031bbbb4f9d0", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "59e3b876d55d0b70555ee73ceb463f27f899a007fb3e35e517d0f10fe890da90", "generation_seconds_batch": 22.41554347402416, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.552934408187866, "target_z": 1.1171874644636046, "ce": 5.224375247955322, "genuineness": 4.570107460021973, "identity": 0.40345144271850586, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f007afcefa87b939056a003eb45bcb7d3c23aa3a3b9a135714e2031bbbb4f9d0.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "f0a66164e1af162efff409919f8211f9a1dd1d5f7a666264c927ccd66f60e4e9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.75, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "0a3af27788910418fea886091e8f634770477b3b39e35ff89c5396209fe26fe8", "generation_seconds_batch": 22.728156588971615, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.440528392791748, "target_z": 1.0468748963036332, "ce": 5.665193557739258, "genuineness": 4.325342178344727, "identity": 0.4733204245567322, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f0a66164e1af162efff409919f8211f9a1dd1d5f7a666264c927ccd66f60e4e9.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "f151beec70f31f0b60f7b974ca5e2317d75f597304101370f539e234f518e871", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ccd9d9ecf4d244af9154be2a8305b9cdb22290d6e243a0c3fc5cb2b9f1fa2111", "generation_seconds_batch": 17.386136341956444, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.425753116607666, "target_z": 0.4121094389363834, "ce": 5.959072113037109, "genuineness": 4.432034492492676, "identity": 0.35903504490852356, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f151beec70f31f0b60f7b974ca5e2317d75f597304101370f539e234f518e871.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "f17fa12cf51a5754b4fa0c955885143ac94b6d0e90d8f1ea274e6b92248ea9f2", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.75, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "10f0dc6a33c307478f50c144b17bcf669ce57a4ed2c99f44095034f698b7287a", "generation_seconds_batch": 23.529391730902717, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.270470142364502, "target_z": -0.3105468404830914, "ce": 6.133810997009277, "genuineness": 3.9048471450805664, "identity": 0.29081982374191284, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f17fa12cf51a5754b4fa0c955885143ac94b6d0e90d8f1ea274e6b92248ea9f2.mp3" }, { "id": "heldout-fr-4", "language": "fr", "sentences": [ "J’ai écrit le numéro sur l’enveloppe.", "Lis-le bien... et vérifie les deux derniers chiffres.", "Appelle-moi quand tu auras un moment." ], "text": "J’ai écrit le numéro sur l’enveloppe. Lis-le bien... et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-4", "key": "f18bb6742a264707cf5af0f9f1a46d27238cba57f4a544756e1d67a04c9ba4a2", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Natural clear speech.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Natural clear speech.) [3.2 seconds duration] Appelle-moi quand tu auras un moment.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4193ee9c1ecb3ff4dad86f03302f421ff7098a184b32b39466ce605ac46db6e3", "generation_seconds_batch": 17.112824529991485, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4155495166778564, "target_z": 1.0312500302933205, "ce": 4.8803582191467285, "genuineness": 4.243753433227539, "identity": 0.7211533784866333, "hypothesis": "J'ai écrit le numéro sur l'enveloppe, lis-le bien et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f18bb6742a264707cf5af0f9f1a46d27238cba57f4a544756e1d67a04c9ba4a2.mp3" }, { "id": "heldout-fr-0", "language": "fr", "sentences": [ "Ton message est arrivé après mon départ.", "Je l’ai relu... puis j’ai rangé mon téléphone.", "Dis-moi quand nous pourrons parler." ], "text": "Ton message est arrivé après mon départ. Je l’ai relu... puis j’ai rangé mon téléphone. Dis-moi quand nous pourrons parler.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-0", "key": "f21bc0adf2df6aeda4aa5bc68657c60b57559ffed3b647118e970ebfd0ba6ad8", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand nous pourrons parler.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "878ea77c4e77e7f31a05ab5f4f23b09a4a718a96bdcdb44177b7b61b122b765a", "generation_seconds_batch": 16.71202097297646, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.182206869125366, "target_z": 0.2597656158246192, "ce": 6.221180438995361, "genuineness": 4.758388519287109, "identity": 0.7321841716766357, "hypothesis": "Ton message est arrivé après mon départ. Je l'ai relu, puis j'ai rangé mon téléphone. Dis-moi, quand nous pourrons parler.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f21bc0adf2df6aeda4aa5bc68657c60b57559ffed3b647118e970ebfd0ba6ad8.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "f2296ea2bace8b96b81e59322e36251816a286d77af19aa7ab8a93aa5014717e", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a8fff25aa49fa79d2f6cc661a2b44bd17e6aaa8fed13e83257adb4b3087949c6", "generation_seconds_batch": 27.172643971978687, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4530181884765625, "target_z": 1.054687553013311, "ce": 5.824047088623047, "genuineness": 4.595211505889893, "identity": 0.4200194478034973, "hypothesis": "I found uh the little box. It was beside the old window. We can look inside together.", "wer": 0.0625, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f2296ea2bace8b96b81e59322e36251816a286d77af19aa7ab8a93aa5014717e.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "f2548c1da7c69729859b8d409ed109e454fae4777fdb52dd27ab1b04bd1606e1", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1702, "scale": 0.75, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "be86f2787a29eb1f30501481f76f271547b7aec8bfdae25cba00d40ee30b8f4a", "generation_seconds_batch": 22.93753120000474, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7660932540893555, "target_z": 0.6250000838891956, "ce": 6.006728172302246, "genuineness": 4.532451152801514, "identity": 0.6845899820327759, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f2548c1da7c69729859b8d409ed109e454fae4777fdb52dd27ab1b04bd1606e1.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "f2bc7bb41758bd8c133a4258d87b03916aa1a684014d607d6d0b323c8400c89d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c568b83e9a2395b3031f43d79dba00ffb084c160acd6ca4f2c1f361b63156a61", "generation_seconds_batch": 16.99920522305183, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.3516521453857422, "target_z": -0.2597656158246192, "ce": 6.109982967376709, "genuineness": 4.432034492492676, "identity": 0.652609646320343, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f2bc7bb41758bd8c133a4258d87b03916aa1a684014d607d6d0b323c8400c89d.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "f3be8e7ab096d8dd2705a92a266b555213b935f9564c3fabc5555566a47938a9", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.75, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "24a615770d21c6ec3a114cbbb2ba264cbbc52d780d5ac0c267ab898a06c10901", "generation_seconds_batch": 22.076076486031525, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.313347101211548, "target_z": 0.341796870776412, "ce": 5.649308204650879, "genuineness": 4.494794845581055, "identity": 0.6000731587409973, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f3be8e7ab096d8dd2705a92a266b555213b935f9564c3fabc5555566a47938a9.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "f3db43146ebea18f3e9ab2acb44bf745e0a32435afb47f57611676b95c68d90b", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "608a6aa514e7a6235d554139cc37b49464c0e4940dca9710a83ca288c0f46362", "generation_seconds_batch": 24.91639245103579, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1362074613571167, "target_z": -0.3945312596123037, "ce": 6.086154937744141, "genuineness": 4.557555198669434, "identity": 0.5650063753128052, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f3db43146ebea18f3e9ab2acb44bf745e0a32435afb47f57611676b95c68d90b.mp3" }, { "id": "heldout-en-0", "language": "en", "sentences": [ "Your message arrived after I had left.", "I read it twice... before putting the phone away.", "Let me know when we can talk." ], "text": "Your message arrived after I had left. I read it twice... before putting the phone away. Let me know when we can talk.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-0", "key": "f3f997aadfdfa5fd64e6f5b9de0b479a0d8a68b0a967bf27758c14eb4b407d8c", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Your message arrived after I had left.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Let me know when we can talk.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Your message arrived after I had left.\n(Natural clear speech.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Natural clear speech.) [3.2 seconds duration] Let me know when we can talk.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7d1ed77b979f8aa3953f4c9ce4311901656faf03e3d0b82f35cae82ea03a3f38", "generation_seconds_batch": 15.159079071017914, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7348692417144775, "target_z": 0.6054687403876964, "ce": 5.720792293548584, "genuineness": 4.381826400756836, "identity": 0.5068739056587219, "hypothesis": "Your message arrived after I had left. I read it twice, before putting the phone away. Let me know when we can talk.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f3f997aadfdfa5fd64e6f5b9de0b479a0d8a68b0a967bf27758c14eb4b407d8c.mp3" }, { "id": "heldout-es-1", "language": "es", "sentences": [ "Las llaves están en la mesa de la cocina.", "Miré en el cajón... y luego junto a la ventana.", "Llévatelas cuando te vayas." ], "text": "Las llaves están en la mesa de la cocina. Miré en el cajón... y luego junto a la ventana. Llévatelas cuando te vayas.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-es-1", "key": "f4bfed3bef03a521f5d080d4d8214eab6c2c36165aae034411feb884f93615e2", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Llévatelas cuando te vayas.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Natural clear speech.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Natural clear speech.) [3.2 seconds duration] Llévatelas cuando te vayas.", "seed": 1701, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "c4a107ac5608b9db6ba783598b69a8ec1a6abcc855020bb6dd27e0f5f0aed1bc", "generation_seconds_batch": 17.60006375994999, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.284409284591675, "target_z": 0.9492187753415279, "ce": 5.5539960861206055, "genuineness": 4.469690799713135, "identity": 0.5466010570526123, "hypothesis": "Las llaves están en la mesa de la cocina, miren el cajón y luego junto a la ventana, llévatelas cuando te vayas.", "wer": 0.08695652173913043, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f4bfed3bef03a521f5d080d4d8214eab6c2c36165aae034411feb884f93615e2.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "f53da1d9e1bd85d4467df77202e20829be516e1b3aeb5462d2bd2d60606ddcdd", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 1.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "871cef49d7ba60b1645eea92f00840694bef7b109233c92d217627dc03bbf387", "generation_seconds_batch": 22.251747611910105, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7660932540893555, "target_z": 0.6250000838891956, "ce": 5.752562999725342, "genuineness": 4.770940780639648, "identity": 0.6263378858566284, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f53da1d9e1bd85d4467df77202e20829be516e1b3aeb5462d2bd2d60606ddcdd.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "f543bbe2120d47e341358f8dc310a69a1cd80a719fa569dd5b3db0b16ca412a3", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "387fa9d35d523e07675b5ece725687ad73996c15a8bd1e7e6a8c63130ace684c", "generation_seconds_batch": 23.120403998997062, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4904866218566895, "target_z": 1.078124926596954, "ce": 4.891279220581055, "genuineness": 4.519899368286133, "identity": 0.6467388272285461, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f543bbe2120d47e341358f8dc310a69a1cd80a719fa569dd5b3db0b16ca412a3.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "f5ed249d44dd6a318eb3d6ec0f2f6e6fd910c03f7f12f49378a651258343cf1f", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.5, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "241d1578ae2d34ab1a35338008af0376a95313cc7aeb56f914f346de2ed45fbb", "generation_seconds_batch": 22.35039490298368, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.040863037109375, "target_z": 0.7968749522297637, "ce": 5.323658466339111, "genuineness": 4.545003414154053, "identity": 0.780266523361206, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f5ed249d44dd6a318eb3d6ec0f2f6e6fd910c03f7f12f49378a651258343cf1f.mp3" }, { "id": "heldout-nl-1", "language": "nl", "sentences": [ "De sleutels liggen op de keukentafel.", "Ik keek in de la... en daarna naast het raam.", "Neem ze mee als je weggaat." ], "text": "De sleutels liggen op de keukentafel. Ik keek in de la... en daarna naast het raam. Neem ze mee als je weggaat.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-nl-1", "key": "f61052750def7339096ed6a19f821b2d23969713588d127cdab92e4f6c309200", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Neem ze mee als je weggaat.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Natural clear speech.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Natural clear speech.) [3.2 seconds duration] Neem ze mee als je weggaat.", "seed": 1701, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1d34eff42914af60c7cf8edf344377b310e9e2b348f2328db98544e640c6cce6", "generation_seconds_batch": 16.90157121606171, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0845766067504883, "target_z": 0.8242188033045931, "ce": 5.343515396118164, "genuineness": 4.231201171875, "identity": 0.2787697911262512, "hypothesis": "De sleutels liggen op de kuikentafel. Ik keek in de laan en daarna naast het raam neem ze mee als je weggaat.", "wer": 0.09090909090909091, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f61052750def7339096ed6a19f821b2d23969713588d127cdab92e4f6c309200.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "f6fd0f860828b865f374992cd0849c1d711ae464ea3849f988569df057c69301", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1701, "scale": 0.5, "cfg": 1.5, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "860b7d4e67ee599797fc7e0b652fe34f2532f3928369a4c6b4c3469ba2bf2558", "generation_seconds_batch": 23.14045072393492, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8793354034423828, "target_z": 0.07031249359179756, "ce": 6.205295085906982, "genuineness": 4.469690799713135, "identity": 0.38024094700813293, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f6fd0f860828b865f374992cd0849c1d711ae464ea3849f988569df057c69301.mp3" }, { "id": "heldout-en-4", "language": "en", "sentences": [ "I wrote the number on the envelope.", "Read it carefully... and check the last two digits.", "Call me when you have a moment." ], "text": "I wrote the number on the envelope. Read it carefully... and check the last two digits. Call me when you have a moment.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-4", "key": "f76a33266a4495ba9cb990c1226744b8d12fd193226e7a7828e39ca67338d3c5", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I wrote the number on the envelope.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Call me when you have a moment.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I wrote the number on the envelope.\n(Natural clear speech.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Natural clear speech.) [3.2 seconds duration] Call me when you have a moment.", "seed": 1702, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d539c3c8034ab0682f8a89b8965a1785bef2fb73d6d6191a30fcc43e94cb8d41", "generation_seconds_batch": 23.23546012805309, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.709890127182007, "target_z": 0.589843725241036, "ce": 5.593709468841553, "genuineness": 4.306513786315918, "identity": 0.44857877492904663, "hypothesis": "I wrote the number on the envelope. Read it carefully and check the last two digits. Call me when you have a moment.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f76a33266a4495ba9cb990c1226744b8d12fd193226e7a7828e39ca67338d3c5.mp3" }, { "id": "heldout-fr-1", "language": "fr", "sentences": [ "Les clés sont sur la table de la cuisine.", "J’ai regardé dans le tiroir... puis près de la fenêtre.", "Prends-les quand tu partiras." ], "text": "Les clés sont sur la table de la cuisine. J’ai regardé dans le tiroir... puis près de la fenêtre. Prends-les quand tu partiras.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-fr-1", "key": "f7a65885d0e107603bdce9d92b5fc3b68a68524db7b0b06f79b6538794e8bf5c", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Prends-les quand tu partiras.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Prends-les quand tu partiras.", "seed": 1701, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7f3434765db5e99ca1f8c8d7db44a78ee970ac822f3e0a67944f7a3b3ed5bf7c", "generation_seconds_batch": 17.377614044002257, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1049835681915283, "target_z": -0.414062528545629, "ce": 6.38797664642334, "genuineness": 4.6956281661987305, "identity": 0.7442787885665894, "hypothesis": "Les clés sont sur la table de la cuisine. J'ai regardé dans le tiroir, puis près de la fenêtre. Prends les quand tu partiras.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f7a65885d0e107603bdce9d92b5fc3b68a68524db7b0b06f79b6538794e8bf5c.mp3" }, { "id": "heldout-de-2", "language": "de", "sentences": [ "Die Straße führt hinter der Brücke nach links.", "Geh weiter... bis zu dem kleinen Holztor.", "Ich warte auf der anderen Seite." ], "text": "Die Straße führt hinter der Brücke nach links. Geh weiter... bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-2", "key": "f86feddc7af334e5c3d6691e715ece3168f8e7618bd3b63ca7d0171f412fdf9d", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich warte auf der anderen Seite.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0a58df2f03df7a81518cac17821520c879f3556dcff7cfcd8c14f4abf9b5c150", "generation_seconds_batch": 17.24085193593055, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0416994094848633, "target_z": 0.1718749429087419, "ce": 6.316492557525635, "genuineness": 4.6203155517578125, "identity": 0.6038836240768433, "hypothesis": "Die Straße führt hinter der Brücke nach links. Geh weiter bis zu dem kleinen Holztor. Ich warte auf der anderen Seite.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f86feddc7af334e5c3d6691e715ece3168f8e7618bd3b63ca7d0171f412fdf9d.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "f88ba4e5b8d50a81a281b059583f4bf7f6ca894ead5d5aad3f1839238a655ef8", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1701, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a60abb46224310b70538da368957e12beb4cbec43418b771b3fbe80477c1b0c6", "generation_seconds_batch": 22.41554347402416, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1220450401306152, "target_z": 0.8476561768882358, "ce": 5.470598220825195, "genuineness": 4.444586753845215, "identity": 0.5429278016090393, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f88ba4e5b8d50a81a281b059583f4bf7f6ca894ead5d5aad3f1839238a655ef8.mp3" }, { "id": "vn_S_WHIS_reg-fr-0", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-0", "key": "f9311259131a3e0d8742e5b3968c20ea6dee795db27d026374c44b20e43fa217", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La porte est encore ouverte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La porte est encore ouverte.\n(Natural clear speech.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Natural clear speech.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "seed": 1701, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "99f123e9a1b3e728752cf1b48feb18e6c51ed4b628f29adfb79f111622a4a6c2", "generation_seconds_batch": 23.16459828801453, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.715298652648926, "target_z": 1.2187500629168966, "ce": 5.403085231781006, "genuineness": 4.6705241203308105, "identity": 0.7573660016059875, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f9311259131a3e0d8742e5b3968c20ea6dee795db27d026374c44b20e43fa217.mp3" }, { "id": "heldout-en-2", "language": "en", "sentences": [ "The road turns left behind the bridge.", "Keep going... until you reach the small wooden gate.", "I will meet you on the other side." ], "text": "The road turns left behind the bridge. Keep going... until you reach the small wooden gate. I will meet you on the other side.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-en-2", "key": "f95c878c8ead4eff6afbec087e483958bf3a60a5bb445e3365f183d77a806464", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The road turns left behind the bridge.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] I will meet you on the other side.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The road turns left behind the bridge.\n(Natural clear speech.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Natural clear speech.) [3.2 seconds duration] I will meet you on the other side.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "eced57cee749f7a143faefa838d61860411eafbcb3f6ff6c223bc3497d1bd0c3", "generation_seconds_batch": 16.441124774049968, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5506484508514404, "target_z": 0.4902343655333373, "ce": 5.609594821929932, "genuineness": 4.337893962860107, "identity": 0.4585247337818146, "hypothesis": "The road turns left behind the bridge. Keep going until you reach the small wooden gate. I will meet you on the other side.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f95c878c8ead4eff6afbec087e483958bf3a60a5bb445e3365f183d77a806464.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "f9beb49762b7131c9edfc3e4fefce36685bbf2ad0649b29eac4bd41b51c73fad", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ed64b02b835f6f161130318899594ab6877a80ee3a0dbc908cc2ff97ac07a368", "generation_seconds_batch": 16.284019420971163, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.954272747039795, "target_z": 0.11718753903177849, "ce": 5.919358730316162, "genuineness": 4.457139015197754, "identity": 0.4732952415943146, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f9beb49762b7131c9edfc3e4fefce36685bbf2ad0649b29eac4bd41b51c73fad.mp3" }, { "id": "vn_S_WHIS_reg-en-0", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-0", "key": "f9cd6420dbeafd372215bc97ce5c51fb77fbad8a3749573df7267287c8e4c9c5", "bucket": "vn_S_WHIS_reg", "phase": "wording", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] The door is still open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] I put your coat on the chair.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Tell me when you are ready.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] The door is still open.\n(Natural clear speech.) [3.8 seconds duration] I put your coat on the chair.\n(Natural clear speech.) [3.2 seconds duration] Tell me when you are ready.", "seed": 1702, "scale": 0.0, "cfg": 1.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "69bba440811ee7249600ceb56689b027b4306f7b5856959c8bcdc526462c0e60", "generation_seconds_batch": 17.09129001107067, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9113956689834595, "target_z": -0.5351562467958988, "ce": 6.300607204437256, "genuineness": 4.256305694580078, "identity": 0.29303255677223206, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f9cd6420dbeafd372215bc97ce5c51fb77fbad8a3749573df7267287c8e4c9c5.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "f9e5081db2c7126a585af6584e44c68aba499fbbc75086a6b46e05594b92afdf", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "33d6c312b94d2c479384158fe241ff5b8235a0b3c642f5b0ca53caca63c1f341", "generation_seconds_batch": 21.745318904984742, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1103920936584473, "target_z": 0.2148438091302316, "ce": 5.760505676269531, "genuineness": 4.432034492492676, "identity": 0.5551475286483765, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/f9e5081db2c7126a585af6584e44c68aba499fbbc75086a6b46e05594b92afdf.mp3" }, { "id": "vn_S_WHIS_reg-de-1", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-1", "key": "fa78e963a59e9d654c853473435f4b197a19e84d37c8bf49e86df3ecc174207d", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Natural clear speech.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Natural clear speech.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "seed": 1702, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "119ec3d7dd69cb0268036337df180e4dcc921324ba675cd1aa5026bb5f31f36f", "generation_seconds_batch": 16.017422315082513, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.2352874279022217, "target_z": 0.29296873572718546, "ce": 5.959072113037109, "genuineness": 4.796044826507568, "identity": 0.6603482961654663, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fa78e963a59e9d654c853473435f4b197a19e84d37c8bf49e86df3ecc174207d.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "fab5f8c23922f79744e238b0369b587932b4a957bed180de4ebe76254092590c", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.0, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d0798f825b654e988280f1accd2215bc1fae8bf5dec8102ad9a141aa6f9ddb7f", "generation_seconds_batch": 16.558907449943945, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8160512447357178, "target_z": 0.6562499650461685, "ce": 5.561938762664795, "genuineness": 4.6203155517578125, "identity": 0.2709486782550812, "hypothesis": "Die Tür ist noch offen. Ich hab deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.05263157894736842, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fab5f8c23922f79744e238b0369b587932b4a957bed180de4ebe76254092590c.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "fb9d6664f7f4724e0cedfc204a6109eef8f45e711506e9ec0d1ca1c38df2f299", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.0, "cfg": 1.5, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a4eadc76d002ecc2d143d697c1fa17791a56d06ef7a7e6fec5d9145a064caf52", "generation_seconds_batch": 15.871158388908952, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2080223560333252, "target_z": -0.3496093783497422, "ce": 6.427689552307129, "genuineness": 4.6705241203308105, "identity": 0.34023770689964294, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fb9d6664f7f4724e0cedfc204a6109eef8f45e711506e9ec0d1ca1c38df2f299.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "fbb91d82ce896ffadf56ff2f798f3f80b8da49700c5c8aa4ee2278538fbd2772", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 1.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "450a74513806895ae929e560c49a9264ee0c9f9ace114ab0c1d580e9e2e51351", "generation_seconds_batch": 23.05777097097598, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2032270431518555, "target_z": 0.8984374015467079, "ce": 5.546053409576416, "genuineness": 4.7332844734191895, "identity": 0.6548981666564941, "hypothesis": "He encontrado la cajita. estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fbb91d82ce896ffadf56ff2f798f3f80b8da49700c5c8aa4ee2278538fbd2772.mp3" }, { "id": "heldout-de-4", "language": "de", "sentences": [ "Ich schrieb die Nummer auf den Umschlag.", "Lies sie genau... und prüfe die letzten beiden Ziffern.", "Ruf an, wenn du Zeit hast." ], "text": "Ich schrieb die Nummer auf den Umschlag. Lies sie genau... und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-4", "key": "fc824fceeb9c8cab27e61d57d05c0cae6cf0d12f10e3af5e1785229e2dcb8ed1", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Natural clear speech.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Natural clear speech.) [3.2 seconds duration] Ruf an, wenn du Zeit hast.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3f6f2c7d2ff425c7c0501bce008de597fa0a7972f30fc176eaa12dd38b991fe8", "generation_seconds_batch": 22.903160133981146, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.138493537902832, "target_z": 0.23242191388613753, "ce": 5.808161735534668, "genuineness": 4.7081804275512695, "identity": 0.5704197287559509, "hypothesis": "Ich schrieb die Nummer auf den Umschlag. Les sie genau und prüfe die letzten beiden Ziffern. Ruf an, wenn du Zeit hast.", "wer": 0.045454545454545456, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fc824fceeb9c8cab27e61d57d05c0cae6cf0d12f10e3af5e1785229e2dcb8ed1.mp3" }, { "id": "vn_S_WHIS_reg-de-0", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-de-0", "key": "fcf2cbf9f7635147d017095db615132ffdd4b873bc94c43d5b9b47a0e65706de", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Tür ist noch offen.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Tür ist noch offen.\n(Natural clear speech.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Natural clear speech.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "seed": 1702, "scale": 0.5, "cfg": 1.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cd89a6b87bd594c2c8307f7ab8d3f484d8ccf88dbc5ac5cddea064314e730140", "generation_seconds_batch": 24.161738647031598, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7740106582641602, "target_z": -0.6210937555343566, "ce": 6.157639026641846, "genuineness": 4.545003414154053, "identity": 0.5029393434524536, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fcf2cbf9f7635147d017095db615132ffdd4b873bc94c43d5b9b47a0e65706de.mp3" }, { "id": "vn_S_WHIS_reg-en-1", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-en-1", "key": "fd2b234e1de4b9e761f8ee77436dc23f89516560494582ddacfc52ffeec137ec", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] I found the little box.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] It was beside the old window.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We can look inside together.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] I found the little box.\n(Natural clear speech.) [3.8 seconds duration] It was beside the old window.\n(Natural clear speech.) [3.2 seconds duration] We can look inside together.", "seed": 1702, "scale": 0.0, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "0513d7b28107af3d8bcafd5ef8458f22f81f395926b5f84c57e998a3fe428263", "generation_seconds_batch": 15.199790642946027, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.290654182434082, "target_z": 0.9531251036963667, "ce": 5.748591423034668, "genuineness": 4.582659721374512, "identity": 0.4092274606227875, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fd2b234e1de4b9e761f8ee77436dc23f89516560494582ddacfc52ffeec137ec.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "fd6610e98da7b90789a6c318e40b8f8d52ae857ad28204e5b302ec087b829d47", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1701, "scale": 0.0, "cfg": 2.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5c68463fc281df7e6a844b8dd81696e665c250472ce983e2aaf19ac4d817634f", "generation_seconds_batch": 17.367203232017346, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4734251499176025, "target_z": -0.18359377883691097, "ce": 5.871703147888184, "genuineness": 4.833701133728027, "identity": 0.5812541842460632, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fd6610e98da7b90789a6c318e40b8f8d52ae857ad28204e5b302ec087b829d47.mp3" }, { "id": "heldout-de-1", "language": "de", "sentences": [ "Die Schlüssel liegen auf dem Küchentisch.", "Ich suchte in der Schublade... dann neben dem Fenster.", "Nimm sie mit, wenn du gehst." ], "text": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade... dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "heldout-de-1", "key": "fd70325264b020a20abd2ee6697c21e0ee2453c265e4048a4b45fbe102a089b2", "bucket": "vn_S_WHIS_reg", "phase": "heldout", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Natural clear speech.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Natural clear speech.) [3.2 seconds duration] Nimm sie mit, wenn du gehst.", "seed": 1701, "scale": 0.75, "cfg": 5.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b0e8d10923802f4d9e8cdeffd4a275c508e553375cc9b144d520211af87aa1ba", "generation_seconds_batch": 22.76204980793409, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7785825729370117, "target_z": 0.6328124423261781, "ce": 5.696964263916016, "genuineness": 4.557555198669434, "identity": 0.519858717918396, "hypothesis": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fd70325264b020a20abd2ee6697c21e0ee2453c265e4048a4b45fbe102a089b2.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "fda4470538badc25a9a498dd1cf11a66711042f31a62a76a2a777c2a04c3a516", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e12df2aa18e75eabd22acf30c5fd2d11dfd259e127ff04321ab9deb6e28b3ed3", "generation_seconds_batch": 22.557522826013155, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9511502981185913, "target_z": 0.11523437485435904, "ce": 5.728734970092773, "genuineness": 4.482243061065674, "identity": 0.6884427070617676, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fda4470538badc25a9a498dd1cf11a66711042f31a62a76a2a777c2a04c3a516.mp3" }, { "id": "vn_S_WHIS_reg-nl-1", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-1", "key": "fda64a4d9c1acc8acaae289deec3f9a39325c90ce9207a6bb2e9545677afacc7", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Het stond naast het oude raam.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] We kunnen er samen in kijken.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Natural clear speech.) [3.8 seconds duration] Het stond naast het oude raam.\n(Natural clear speech.) [3.2 seconds duration] We kunnen er samen in kijken.", "seed": 1701, "scale": 0.75, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "be6fd4476c4ae4247cc61472afff0caf9d2208a8c0d4a4d01210fda796fb03c2", "generation_seconds_batch": 21.615672083105892, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0175567865371704, "target_z": -0.4687500069907663, "ce": 6.086154937744141, "genuineness": 4.406930446624756, "identity": 0.5222664475440979, "hypothesis": "Ik heb het kleine doosje gevonden, het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fda64a4d9c1acc8acaae289deec3f9a39325c90ce9207a6bb2e9545677afacc7.mp3" }, { "id": "vn_S_WHIS_reg-es-0", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-0", "key": "fe0314b6c34266faeec7958e7129508f19f91167f5431fd78f3945953ec82a50", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] La puerta sigue abierta.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Dime cuándo estás listo.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] La puerta sigue abierta.\n(Natural clear speech.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Natural clear speech.) [3.2 seconds duration] Dime cuándo estás listo.", "seed": 1702, "scale": 0.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "58b14b886c0ae377a8f889fe4023de6f5e82ff5dfe52c2defa9d37ea007bc798", "generation_seconds_batch": 17.23506281792652, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6278717517852783, "target_z": 1.1640625099035855, "ce": 5.2918877601623535, "genuineness": 4.507347106933594, "identity": 0.4862731695175171, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fe0314b6c34266faeec7958e7129508f19f91167f5431fd78f3945953ec82a50.mp3" }, { "id": "vn_S_WHIS_reg-nl-0", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-nl-0", "key": "fe055ef2dea0209cceff017986324342b48df5b10088d215de07c58144b4de29", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] De deur staat nog open.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] De deur staat nog open.\n(Natural clear speech.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Natural clear speech.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "seed": 1701, "scale": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "40a4f96524ff0f5c1fe97f7dfdb3c3438f61c618f938204f1f8d3d4a7e5a5eda", "generation_seconds_batch": 21.1595835769549, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.609973907470703, "target_z": 0.5273438137907425, "ce": 5.609594821929932, "genuineness": 4.306513786315918, "identity": 0.3322318196296692, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd, zeg maar, wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fe055ef2dea0209cceff017986324342b48df5b10088d215de07c58144b4de29.mp3" }, { "id": "vn_S_WHIS_reg-fr-1", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-fr-1", "key": "fe80165a8e0af2d7ab8613a1613a52d36f5f87267a5009b3ad65ffdc4cc957a6", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Natural clear speech.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Natural clear speech.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "seed": 1702, "scale": 0.5, "cfg": 1.25, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1a9a7055c2d4b0891c32f3c500e5b5e240e7a31b51bea4aff87b42a0e515c035", "generation_seconds_batch": 23.36565126501955, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.10414719581604, "target_z": 0.21093748077539268, "ce": 6.09409761428833, "genuineness": 4.6705241203308105, "identity": 0.7118610739707947, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/fe80165a8e0af2d7ab8613a1613a52d36f5f87267a5009b3ad65ffdc4cc957a6.mp3" }, { "id": "vn_S_WHIS_reg-es-1", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "sentence_seconds": [ 3.2, 3.8, 3.2 ], "seconds": 10.2, "frames": 128, "max_frames": 188, "reference": true, "task_id": "vn_S_WHIS_reg-es-1", "key": "ffc1f1b32f6c4fd73ecbfe86794ee484ce6c8ca3607ffcfc86aa229fd6ff5d73", "bucket": "vn_S_WHIS_reg", "phase": "grid", "candidate": "manual-c79d4bcf72d9", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Explore a more vulnerable, hesitant quality in the whisper. The speech should sound like it is being carefully offered.\nSCRIPT:\n(Begin with a noticeable hesitation before the first word, as if gathering immense courage to speak softly. Keep the tone breathy.) [3.2 seconds duration] He encontrado la cajita.\n(Deliver the core of the passage with slight, soft stutters embedded in the airflow, emphasizing the fragile nature of the sound.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Fade the final words out slowly, letting the slight tremor in the voice remain present until silence.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "neutral_prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects.\nSCRIPT:\n(Natural clear speech.) [3.2 seconds duration] He encontrado la cajita.\n(Natural clear speech.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Natural clear speech.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "seed": 1702, "scale": 1.0, "cfg": 4.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7b0e1ea6f17f20fd14a5c73c672cd0233c643ec7e2584fa871593200a92dfc9a", "generation_seconds_batch": 21.435869556036778, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.440528392791748, "target_z": 1.0468748963036332, "ce": 5.387199878692627, "genuineness": 4.65797233581543, "identity": 0.5663636326789856, "hypothesis": "He encontrado la cajita, estaba junto a la ventana vieja. pudemos mirar juntos lo que hay dentro.", "wer": 0.058823529411764705, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_WHIS_reg/ffc1f1b32f6c4fd73ecbfe86794ee484ce6c8ca3607ffcfc86aa229fd6ff5d73.mp3" }, { "key": "original-vn_S_WHIS_reg-de-0-seed1701-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "ac6b136c3c832f61cd0c408957f12ce3058139e3c06eb8a05e3afaa22c34ef61", "task_id": "vn_S_WHIS_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.35561084747314453, "target_z": -0.8828125355363954, "ce": 6.324435234069824, "genuineness": 4.444586753845215, "identity": 0.3705710470676422, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/9b3397f751ab0e3691b08e2d.mp3" }, { "key": "original-vn_S_WHIS_reg-de-0-seed1701-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "a6418a131d7ecccd5d73aa52be2e0f5c47d56ec77f86b362bdaf35882fd4109e", "task_id": "vn_S_WHIS_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7615211009979248, "target_z": -0.6289062631076868, "ce": 6.3641486167907715, "genuineness": 4.6705241203308105, "identity": 0.4992162585258484, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/3d46ff2673aed3812d3a8b3c.mp3" }, { "key": "original-vn_S_WHIS_reg-de-0-seed1701-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "f6d1ce8c89b4ccb820c10553b71d16d6ee779031c45769ea5162c63c142d6e5b", "task_id": "vn_S_WHIS_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7302972078323364, "target_z": -0.6484375320410122, "ce": 6.38797664642334, "genuineness": 4.545003414154053, "identity": 0.6075156927108765, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/e009f8a8e963fa76f95a74d2.mp3" }, { "key": "original-vn_S_WHIS_reg-de-0-seed1701-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "fbb650cb77acaaad8cbc405bdcde87bc1a9844a4ffa566c74827ad3743993b19", "task_id": "vn_S_WHIS_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.155778169631958, "target_z": -1.0078125075733302, "ce": 6.252951145172119, "genuineness": 4.369274139404297, "identity": 0.6650416851043701, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/51a7e5ba9975b02524b4acac.mp3" }, { "key": "original-vn_S_WHIS_reg-de-0-seed1702-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "222dbfad4ad274bd70976b1eee11021850c46d38b057a2a4b99bab290426e34e", "task_id": "vn_S_WHIS_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.10582005977630615, "target_z": -1.0390624632984768, "ce": 6.284721851348877, "genuineness": 4.532451152801514, "identity": 0.5895543694496155, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/b21515e6b1d386d6239f3b8c.mp3" }, { "key": "original-vn_S_WHIS_reg-de-0-seed1702-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "0c16dd646538ed6f4d6328fc73170deb3a54eb0bf88e04d6deb95cbd7fbf37be", "task_id": "vn_S_WHIS_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0175567865371704, "target_z": -0.4687500069907663, "ce": 6.252951145172119, "genuineness": 4.7458367347717285, "identity": 0.7395527362823486, "hypothesis": "Die Tür ist doch offen. Ich hab deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.10526315789473684, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/1fba4d898dab614a16aaf939.mp3" }, { "key": "original-vn_S_WHIS_reg-de-0-seed1702-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "e9a7703b53f8fee1ff83f6ef45130f129073e07a6c686e9c1d17189d7a671e22", "task_id": "vn_S_WHIS_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.3306318521499634, "target_z": -0.8984374761148818, "ce": 6.252951145172119, "genuineness": 4.369274139404297, "identity": 0.47122955322265625, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/3302128fbbdbeedbb3cd2d12.mp3" }, { "key": "original-vn_S_WHIS_reg-de-0-seed1702-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Die Tür ist noch offen.", "Ich habe deinen Mantel auf den Stuhl gelegt.", "Sag mir, wenn du bereit bist." ], "text": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "8ec73fc1bbc0015649ca0742c00ef05a43b8e00e3c8325c8a84279097cc388b5", "task_id": "vn_S_WHIS_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9426195025444031, "target_z": -0.5156250151466603, "ce": 6.141753673553467, "genuineness": 4.482243061065674, "identity": 0.6534345149993896, "hypothesis": "Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/ff5591e8e795d24a4d62463e.mp3" }, { "key": "original-vn_S_WHIS_reg-de-1-seed1701-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "38f9f3173f2d84136dabc5861cac37ce3ea4631b77400bace08f76a3d682fda8", "task_id": "vn_S_WHIS_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8177240490913391, "target_z": -0.593750016311788, "ce": 6.340320587158203, "genuineness": 4.425758361816406, "identity": 0.6464388370513916, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/e76141dd554b3d5e9d61876f.mp3" }, { "key": "original-vn_S_WHIS_reg-de-1-seed1701-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "ae71e2324ac1b6cdbdb423471d77251427022673d1ed72d000b414724a8afa2e", "task_id": "vn_S_WHIS_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.5616884231567383, "target_z": -0.7539062351446216, "ce": 6.348263263702393, "genuineness": 4.2751336097717285, "identity": 0.5694175362586975, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/13318527b587c0b071341917.mp3" }, { "key": "original-vn_S_WHIS_reg-de-1-seed1701-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "1eb67b7a94018873455e1de40686970e8695487dc3a1e7a9152a942dc5f43519", "task_id": "vn_S_WHIS_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.23071539402008057, "target_z": -0.960937536701523, "ce": 6.356205940246582, "genuineness": 4.457139015197754, "identity": 0.48427727818489075, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/545a924250d47b1500a40071.mp3" }, { "key": "original-vn_S_WHIS_reg-de-1-seed1701-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "83604a3b433051377f7873baba60ed05435b182e70f510d04e1d0658af9a14fe", "task_id": "vn_S_WHIS_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.16826772689819336, "target_z": -1.0, "ce": 6.332377910614014, "genuineness": 4.557555198669434, "identity": 0.6497645974159241, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/bab85121543c6451fa384015.mp3" }, { "key": "original-vn_S_WHIS_reg-de-1-seed1702-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "faded32fc31dfb6eac970cb45d0bb66e394268e682a23558889fe0cf760592e8", "task_id": "vn_S_WHIS_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9551090598106384, "target_z": -0.5078125075733302, "ce": 6.332377910614014, "genuineness": 4.570107460021973, "identity": 0.49292176961898804, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/5ad6bb5cc74f11cd554b9636.mp3" }, { "key": "original-vn_S_WHIS_reg-de-1-seed1702-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "623684c050c6ab7991bde6bd65c8b999580c5605ca44b4d7b6d827b39de099f0", "task_id": "vn_S_WHIS_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.26193928718566895, "target_z": -0.9414062677681977, "ce": 6.308549880981445, "genuineness": 4.281409740447998, "identity": 0.6173437833786011, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/833033cb57f47f0621a7dfc5.mp3" }, { "key": "original-vn_S_WHIS_reg-de-1-seed1702-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "c3fe7d262482b858aec42982eccf931fbf7f49df0722e554200f04ef15298f81", "task_id": "vn_S_WHIS_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.155778169631958, "target_z": -1.0078125075733302, "ce": 6.443574905395508, "genuineness": 4.469690799713135, "identity": 0.5003708600997925, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/4a2bcc527f8217ad98ca1452.mp3" }, { "key": "original-vn_S_WHIS_reg-de-1-seed1702-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "de", "sentences": [ "Ich habe die kleine Schachtel gefunden.", "Sie stand neben dem alten Fenster.", "Wir können gemeinsam hineinschauen." ], "text": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 1.0, "frames_generated": 127, "seconds_generated": 10.16, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "12f5de5613cc829214999620a96b8f09f177c3401d973c7fad1cbf914af05e30", "task_id": "vn_S_WHIS_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.1870020627975464, "target_z": -0.9882812386400047, "ce": 6.459460258483887, "genuineness": 4.482243061065674, "identity": 0.44803130626678467, "hypothesis": "Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/de/9eea3a8a71490135b78ceeb1.mp3" }, { "key": "original-vn_S_WHIS_reg-en-0-seed1701-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "387b96b28ee1ae65b8cbe998c4f440daae38006aa2d8d1114305279b8e9182b1", "task_id": "vn_S_WHIS_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9261711835861206, "target_z": 0.09960935970769873, "ce": 6.181467056274414, "genuineness": 4.030367851257324, "identity": 0.178518608212471, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/ad53c288e8f23e752da37dab.mp3" }, { "key": "original-vn_S_WHIS_reg-en-0-seed1701-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.5, "frames_generated": 127, "seconds_generated": 10.16, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "ad25733f64e8f8f74454a67d464831158cc89621a629c9d9bcba376e8d89dda9", "task_id": "vn_S_WHIS_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5108938217163086, "target_z": -0.1601562561169205, "ce": 6.260893821716309, "genuineness": 3.8797430992126465, "identity": 0.31384193897247314, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/82ad83db7cc55245036cfbf1.mp3" }, { "key": "original-vn_S_WHIS_reg-en-0-seed1701-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "11cd23a45feb7ff0c5178a7ca8e338ea8558a754e8bc666da025983053cb7580", "task_id": "vn_S_WHIS_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.233001470565796, "target_z": -0.3339843632030819, "ce": 6.23706579208374, "genuineness": 3.961331605911255, "identity": 0.3337501287460327, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/79db98118afcce03d04e7e6f.mp3" }, { "key": "original-vn_S_WHIS_reg-en-0-seed1701-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "3f5abc301e943bd7d2063b4e832a13865c40d573367ec44f8ce86ecb5005fbf7", "task_id": "vn_S_WHIS_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0862493515014648, "target_z": -0.42578121533745045, "ce": 6.165581703186035, "genuineness": 3.7291181087493896, "identity": 0.1797390729188919, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/e6aa42781a44577be13dc9f2.mp3" }, { "key": "original-vn_S_WHIS_reg-en-0-seed1702-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "f41f77ebc7a4cf4326dfbc1f012aed35cd1d6c0bdcc84dc4d76c3acdce231fa5", "task_id": "vn_S_WHIS_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.3703864812850952, "target_z": -0.248046854464624, "ce": 6.213237762451172, "genuineness": 4.168440818786621, "identity": 0.2607855796813965, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/57f96a930c24e38b49795061.mp3" }, { "key": "original-vn_S_WHIS_reg-en-0-seed1702-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.5, "frames_generated": 127, "seconds_generated": 10.16, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "43374449dc42caeed554d8345ed489ee8aa7a5338d5c2b9936789db75871d9b1", "task_id": "vn_S_WHIS_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.523383378982544, "target_z": -0.15234374854359037, "ce": 6.109982967376709, "genuineness": 4.319066047668457, "identity": 0.35584211349487305, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/8191039ac6f66cb9d4892b94.mp3" }, { "key": "original-vn_S_WHIS_reg-en-0-seed1702-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "86256e2fa9f42538532c3747d0dbe34cb0d19341e725a5fb7d190c0d8838997a", "task_id": "vn_S_WHIS_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.983210563659668, "target_z": -0.4902343655333373, "ce": 6.229123115539551, "genuineness": 3.7667744159698486, "identity": 0.19571669399738312, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/0e31a483b9cbefe6ef4063ea.mp3" }, { "key": "original-vn_S_WHIS_reg-en-0-seed1702-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "The door is still open.", "I put your coat on the chair.", "Tell me when you are ready." ], "text": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "32a6f0e7d342de875965619a56359df519ced4b89785c9def98106d539605022", "task_id": "vn_S_WHIS_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1767985820770264, "target_z": -0.36914057271489376, "ce": 6.173524379730225, "genuineness": 3.8922953605651855, "identity": 0.33357536792755127, "hypothesis": "The door is still open. I put your coat on the chair. Tell me when you are ready.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/75fb8eee9c0c9f654f037c4b.mp3" }, { "key": "original-vn_S_WHIS_reg-en-1-seed1701-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "093be44bad48b1e8605b3ac2c5ed32a532f66f1f890739515de221032a09d0b3", "task_id": "vn_S_WHIS_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7677658796310425, "target_z": -0.6250000093210217, "ce": 6.284721851348877, "genuineness": 4.080576419830322, "identity": 0.3561449646949768, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/0a8767363c12c5c07e093543.mp3" }, { "key": "original-vn_S_WHIS_reg-en-1-seed1701-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "eed0651cd3809f3992ca7c29e348d137a4c7f3301d3816efde1bc4712dba3575", "task_id": "vn_S_WHIS_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8418667316436768, "target_z": 0.046874970871807094, "ce": 6.316492557525635, "genuineness": 4.457139015197754, "identity": 0.4985172748565674, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/8330eb14ece64c72c1f8282d.mp3" }, { "key": "original-vn_S_WHIS_reg-en-1-seed1701-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "0aa28988fec5eea8ceedc7c727d558fac0c0616346f206f4442e03ea5686a1c6", "task_id": "vn_S_WHIS_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.0737597942352295, "target_z": -0.4335937229107806, "ce": 6.38003396987915, "genuineness": 4.331617832183838, "identity": 0.37370628118515015, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/934e1c9d97757794194af86c.mp3" }, { "key": "original-vn_S_WHIS_reg-en-1-seed1701-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "6e1ad4c8915df55bdaf855ae35b8b0ca624391fd760f3b2c204aff50f8ef0350", "task_id": "vn_S_WHIS_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9551090598106384, "target_z": -0.5078125075733302, "ce": 6.53094482421875, "genuineness": 4.545003414154053, "identity": 0.38562414050102234, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/cb14a3a470418cef58e042de.mp3" }, { "key": "original-vn_S_WHIS_reg-en-1-seed1702-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "5b6b4a6c3f860a477d38b82ae3973c4b581d62ef83ff315461125a29735d5469", "task_id": "vn_S_WHIS_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6491152048110962, "target_z": -0.6992187566994844, "ce": 6.483288764953613, "genuineness": 4.457139015197754, "identity": 0.3242148756980896, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/31618b70e44d5de99a835a54.mp3" }, { "key": "original-vn_S_WHIS_reg-en-1-seed1702-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.5, "frames_generated": 127, "seconds_generated": 10.16, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "f4326495e9e071e6d2011ca6e8f5dc8dc845ef32b8b829dd2bb4571b4ab7c816", "task_id": "vn_S_WHIS_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6928286552429199, "target_z": -0.6718749801928289, "ce": 6.284721851348877, "genuineness": 4.2249250411987305, "identity": 0.371037095785141, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.03999999999999915, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/55fe583c981c76c2bf8e6f3d.mp3" }, { "key": "original-vn_S_WHIS_reg-en-1-seed1702-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "fecc9f0c40848745bc40b4708cd5f7d9be726b23b1ce1d5d4f8d5208eea55581", "task_id": "vn_S_WHIS_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1861655712127686, "target_z": -0.3632813038871569, "ce": 6.300607204437256, "genuineness": 4.413206577301025, "identity": 0.34379974007606506, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/1fa485496c218ca64d0b3bc2.mp3" }, { "key": "original-vn_S_WHIS_reg-en-1-seed1702-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "en", "sentences": [ "I found the little box.", "It was beside the old window.", "We can look inside together." ], "text": "I found the little box. It was beside the old window. We can look inside together.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "2237c956020cf8cde2dee642e58f42f2932e31c5dc6e07bfefa77370d6f96a63", "task_id": "vn_S_WHIS_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6990734338760376, "target_z": -0.6679687264061638, "ce": 6.3641486167907715, "genuineness": 4.457139015197754, "identity": 0.5165914297103882, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/en/e0b53056e07dcbe7b54b285f.mp3" }, { "key": "original-vn_S_WHIS_reg-es-0-seed1701-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "aebb8938c28ee178ca25a5cf2ee9923678182f401538a2bb871eceaef99127a6", "task_id": "vn_S_WHIS_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9301300048828125, "target_z": -0.5234374854359035, "ce": 5.8875885009765625, "genuineness": 3.835810661315918, "identity": 0.4694689214229584, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/8ab63b631400804ab6430187.mp3" }, { "key": "original-vn_S_WHIS_reg-es-0-seed1701-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "1e19c976350e51499e505e4e43c067b90b32b5adc7732bcc5482115c95dd06c4", "task_id": "vn_S_WHIS_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8489479422569275, "target_z": -0.5742187473784627, "ce": 6.09409761428833, "genuineness": 4.080576419830322, "identity": 0.5683618783950806, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás lista.", "wer": 0.06666666666666667, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/88d43ccdf4a10dff3eb2c30f.mp3" }, { "key": "original-vn_S_WHIS_reg-es-0-seed1701-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "e2e2434f514ba49a2bb13c914672da3667069c48aaca491de1a7234a88c8a97d", "task_id": "vn_S_WHIS_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8676822781562805, "target_z": -0.5624999860184674, "ce": 6.09409761428833, "genuineness": 4.218649387359619, "identity": 0.5316578149795532, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/32214cc48ab1f68799908516.mp3" }, { "key": "original-vn_S_WHIS_reg-es-0-seed1701-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "b43b271d24a7bfeb0cc78fff3f4c692ac45ae5c852a029676dea76024e285c19", "task_id": "vn_S_WHIS_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7865002155303955, "target_z": -0.6132812479610265, "ce": 6.054384231567383, "genuineness": 4.162164688110352, "identity": 0.59152752161026, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/ff42af439f028c88f6342254.mp3" }, { "key": "original-vn_S_WHIS_reg-es-0-seed1702-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "b6309c57fac50506bfb2b2a2507bb9e83275058ba41ec2c0523260e9c6eb0d2f", "task_id": "vn_S_WHIS_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.089371681213379, "target_z": -0.42382812572820483, "ce": 6.062326908111572, "genuineness": 4.017816066741943, "identity": 0.5334363579750061, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/8dc136384e48dc35a6fff206.mp3" }, { "key": "original-vn_S_WHIS_reg-es-0-seed1702-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "87e923c971ba9a9f944293092741dc8c4b0b2b2e1f6aaebb057a501160001428", "task_id": "vn_S_WHIS_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7490315437316895, "target_z": -0.636718770681017, "ce": 5.935244083404541, "genuineness": 4.030367851257324, "identity": 0.38945382833480835, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/682cdf5ed66753122810bd3b.mp3" }, { "key": "original-vn_S_WHIS_reg-es-0-seed1702-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "258c0695b965e2be0356795cea571247b73dc7103d571a13d30ec35b41db740e", "task_id": "vn_S_WHIS_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6803390979766846, "target_z": -0.679687487766159, "ce": 6.109982967376709, "genuineness": 3.986435890197754, "identity": 0.514769434928894, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/d3a6a26f474256550f62f6a7.mp3" }, { "key": "original-vn_S_WHIS_reg-es-0-seed1702-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "La puerta sigue abierta.", "He dejado tu abrigo sobre la silla.", "Dime cuándo estás listo." ], "text": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "9a62b62a5096f457b80421a227a33fdb308779a4a9714f66541125cd95a38405", "task_id": "vn_S_WHIS_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8239688277244568, "target_z": -0.5898437625251229, "ce": 6.173524379730225, "genuineness": 4.005263805389404, "identity": 0.5810372829437256, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/d1f0f9d08d0d5a938864f8fa.mp3" }, { "key": "original-vn_S_WHIS_reg-es-1-seed1701-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "8dc7fc7ed4d6199cf5fc57988f5dab782521934bf2e14a5832f4fcf8a8b5a182", "task_id": "vn_S_WHIS_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7677658796310425, "target_z": -0.6250000093210217, "ce": 6.356205940246582, "genuineness": 4.7458367347717285, "identity": 0.6149548292160034, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/77249cabc728441544c2b06d.mp3" }, { "key": "original-vn_S_WHIS_reg-es-1-seed1701-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "ad347b0dca40ee327585535ea9f937cefc23460e1fe8bdbdae376f60abbfe6b1", "task_id": "vn_S_WHIS_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.164309024810791, "target_z": -0.3769530802882239, "ce": 6.086154937744141, "genuineness": 4.331617832183838, "identity": 0.5128825902938843, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/1d4fdc5ae17e6dcbec77e7dc.mp3" }, { "key": "original-vn_S_WHIS_reg-es-1-seed1701-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "5c347b9687cbbea5c8eb9c6faab2133383b120dc919a2610b4eb471a85066be6", "task_id": "vn_S_WHIS_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6178913116455078, "target_z": -0.7187500256328098, "ce": 6.109982967376709, "genuineness": 4.607763767242432, "identity": 0.5413133502006531, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/d62b41da974b57ba4fa82100.mp3" }, { "key": "original-vn_S_WHIS_reg-es-1-seed1701-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "aac766f2b0bb3a305b4e93921c213ebbbb668b4e952523ae9c310d4c81413ebe", "task_id": "vn_S_WHIS_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.133085012435913, "target_z": -0.3964844237897231, "ce": 6.268836498260498, "genuineness": 4.607763767242432, "identity": 0.41241025924682617, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/69f216b02fc02599002ac969.mp3" }, { "key": "original-vn_S_WHIS_reg-es-1-seed1702-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "af6d696ec6795c1a3946f95b626a7b1c94c78bb9438eedbda81b30a992a0e91a", "task_id": "vn_S_WHIS_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8614374995231628, "target_z": -0.5664062398051325, "ce": 6.268836498260498, "genuineness": 4.545003414154053, "identity": 0.5609867572784424, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/772168eda34cfa9dc55adc0d.mp3" }, { "key": "original-vn_S_WHIS_reg-es-1-seed1702-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "33bb8cb7c78631f12746c3726a126c9238f05c84eceb53cc086c92b81338fcf7", "task_id": "vn_S_WHIS_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.7427867650985718, "target_z": -0.640625024467682, "ce": 6.213237762451172, "genuineness": 4.545003414154053, "identity": 0.6894427537918091, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/05fc3ca337903fbdec0776af.mp3" }, { "key": "original-vn_S_WHIS_reg-es-1-seed1702-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "cb4d741897ea3a08dc617983c7c719dcaeb7359c02a2d6d91c470789b5002b18", "task_id": "vn_S_WHIS_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.8864165544509888, "target_z": -0.550781261942559, "ce": 5.982900142669678, "genuineness": 4.758388519287109, "identity": 0.7907114028930664, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/7280ec8cba84696067a88875.mp3" }, { "key": "original-vn_S_WHIS_reg-es-1-seed1702-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "es", "sentences": [ "He encontrado la cajita.", "Estaba junto a la ventana vieja.", "Podemos mirar juntos lo que hay dentro." ], "text": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "54550881fbc28adf2d1938d7fe1b2bc093b630246cc39a6dc165a6134b65026c", "task_id": "vn_S_WHIS_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6491152048110962, "target_z": -0.6992187566994844, "ce": 6.23706579208374, "genuineness": 4.381826400756836, "identity": 0.5196483731269836, "hypothesis": "He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/es/ccb5d4b617e4cc09f51601b0.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-0-seed1701-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "91397c86c4e2d20f9f17fce4152d2b5f31346c2ac6f13da549053e6a3e6ea149", "task_id": "vn_S_WHIS_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.504649043083191, "target_z": -0.1640625099035856, "ce": 6.141753673553467, "genuineness": 4.494794845581055, "identity": 0.7651387453079224, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/4f3a8603ea1ea13939b3027a.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-0-seed1701-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "1b19d064112dd163d23035240e06ef212b139940e14c8db0f25ac4bc109392f6", "task_id": "vn_S_WHIS_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.1205955743789673, "target_z": -0.4042968567948794, "ce": 6.260893821716309, "genuineness": 4.457139015197754, "identity": 0.7461879253387451, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/daf11885970585f684975b29.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-0-seed1701-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "9e4796dc59295e4fbcd7dcfc5e32407a98d25034d397d444836030ded1dbbc00", "task_id": "vn_S_WHIS_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.2665114402770996, "target_z": 0.3125000792286847, "ce": 6.165581703186035, "genuineness": 4.582659721374512, "identity": 0.7530038952827454, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/1f082b7c7e7ab6e89d84dca9.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-0-seed1701-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "60bf0accfecebb9c382f5e3b6048de2fdb8a33bd4b2af0d7d5e6aafff67c451a", "task_id": "vn_S_WHIS_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.3641417026519775, "target_z": -0.2519531082512891, "ce": 6.260893821716309, "genuineness": 4.344170093536377, "identity": 0.7906546592712402, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/08b3c0b364c3188bfcfa5d10.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-0-seed1702-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "5287f8b5863fef8f8d26f84521ef78eb1f52b99332b593370d9e4b4c9dbc97dd", "task_id": "vn_S_WHIS_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.311061143875122, "target_z": -0.28515622815385533, "ce": 6.229123115539551, "genuineness": 4.231201171875, "identity": 0.656397819519043, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/9619d90b5642140b9db5fce9.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-0-seed1702-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "71b437d24ec8426d4a1e19d01d39166d0c667c3ea59ea8c2db93f06afd5ae188", "task_id": "vn_S_WHIS_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4734251499176025, "target_z": -0.18359377883691097, "ce": 6.260893821716309, "genuineness": 4.7081804275512695, "identity": 0.8034999370574951, "hypothesis": "La porte est encore ouverte , j'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/cc61adea4b3afbda6514ebf2.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-0-seed1702-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "8d7202fa73da593bd3514beabe0315bb1a344dd102428bfb8fb740744e952e5b", "task_id": "vn_S_WHIS_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.27671480178833, "target_z": -0.30664066126460016, "ce": 6.197352409362793, "genuineness": 4.268857479095459, "identity": 0.6414709687232971, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/ddca5101e994585b68814c1b.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-0-seed1702-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "La porte est encore ouverte.", "J’ai posé ton manteau sur la chaise.", "Dis-moi quand tu es prêt." ], "text": "La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "3102c14a31e81ee9c8a61ec3fedc613212b4bea4e1e57a4168cc159d7e4e2d6d", "task_id": "vn_S_WHIS_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4453237056732178, "target_z": -0.2011718835928169, "ce": 6.086154937744141, "genuineness": 4.319066047668457, "identity": 0.8339059352874756, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/0fae1e31137edebdaa41d4c6.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-1-seed1701-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "559480759646c16202c70a60cec8a4727a5864dffba9b521936ad26ee7c78666", "task_id": "vn_S_WHIS_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5577296018600464, "target_z": -0.13085939000101934, "ce": 6.1894097328186035, "genuineness": 4.532451152801514, "identity": 0.8286293745040894, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/d1ab601f937fdea638024f82.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-1-seed1701-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "0fdb3e0df3425b33b673e9c9e41f3de997ed908150cb183e61934e2d67fd9f78", "task_id": "vn_S_WHIS_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.5358729362487793, "target_z": -0.1445312409702602, "ce": 6.157639026641846, "genuineness": 4.507347106933594, "identity": 0.7331330180168152, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/27763fc8f528942b13641111.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-1-seed1701-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "e1dacfa97dbb1bb35def76c053f216832bf7a190787d9de1f43953f341cfa3d0", "task_id": "vn_S_WHIS_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.813765287399292, "target_z": 0.029296866115901165, "ce": 6.070269584655762, "genuineness": 4.65797233581543, "identity": 0.7429172992706299, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/ca7fc194f02d55c794a7141c.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-1-seed1701-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "afa6445b24926c61f31a9cede9794acbc834b444c094dc18c60d40b642c5f0c8", "task_id": "vn_S_WHIS_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9051508903503418, "target_z": -0.5390625005825639, "ce": 6.157639026641846, "genuineness": 4.6203155517578125, "identity": 0.8038060665130615, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/8a827cc98ee55ff7c04db517.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-1-seed1702-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "3cde602df2d2a2b270ce41d0b92b2e1b45c13f08db72f66d4b7e3adcdfc58088", "task_id": "vn_S_WHIS_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9363747835159302, "target_z": -0.5195312316492384, "ce": 6.316492557525635, "genuineness": 4.607763767242432, "identity": 0.7536294460296631, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/1a2aa5a2c12540898f9f04f3.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-1-seed1702-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "607199c66c039620d99becdcac6ac49aa995a859c70b594a0139fac76616a03c", "task_id": "vn_S_WHIS_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.4890371561050415, "target_z": -0.17382810708616137, "ce": 6.157639026641846, "genuineness": 4.425758361816406, "identity": 0.7502268552780151, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/2089629d5b0952580b183f81.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-1-seed1702-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "3ba30c10112e271b7c4e071105e53eb49d33c30bef597fa17ffb1d5d404ce1b9", "task_id": "vn_S_WHIS_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.2486133575439453, "target_z": -0.3242187660205061, "ce": 6.062326908111572, "genuineness": 4.394378662109375, "identity": 0.6623623371124268, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/09ad0452cdf8f63d1a41abe9.mp3" }, { "key": "original-vn_S_WHIS_reg-fr-1-seed1702-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "fr", "sentences": [ "J’ai trouvé la petite boîte.", "Elle était près de la vieille fenêtre.", "Nous pouvons regarder ensemble ce qu’elle contient." ], "text": "J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "cf0ce4ec4bb0e11b8d62ca3f56d7eea89bef4f7587860b3a192675bea285b658", "task_id": "vn_S_WHIS_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.9051508903503418, "target_z": -0.5390625005825639, "ce": 5.95112943649292, "genuineness": 4.394378662109375, "identity": 0.7284080982208252, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/fr/8c6fa01efdf2bdb0f82a9eb9.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-0-seed1701-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "d79ac8f894fab2df8e9f95fe2ebb086c1e333546a4b9f25e2a6e35ab523176f9", "task_id": "vn_S_WHIS_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.3493661880493164, "target_z": -0.8867187147548866, "ce": 6.157639026641846, "genuineness": 4.400654315948486, "identity": 0.7391611337661743, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/8f1dccedbcbbdc0422555c60.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-0-seed1701-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "b92ca8dd82b52233317fdf01f6a653707f37e1a55d6e604e2aa6377d1d26a0ed", "task_id": "vn_S_WHIS_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.44928252696990967, "target_z": -0.8242187287364192, "ce": 6.173524379730225, "genuineness": 4.507347106933594, "identity": 0.42696741223335266, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/60618f941c40cd004f76603f.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-0-seed1701-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "1725bcf5cf6ea551c9196bec8529045027ea957b052515033c344af7468016f7", "task_id": "vn_S_WHIS_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.2494497299194336, "target_z": -0.9492187753415279, "ce": 6.252951145172119, "genuineness": 4.482243061065674, "identity": 0.6445729732513428, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/56cf04616fc040a9261a62c8.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-0-seed1701-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "6297abb55d1284dce9e5d82d9bca28d098a36de4b5f4a7e41e02a8276f46bee7", "task_id": "vn_S_WHIS_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.299407958984375, "target_z": -0.9179687450482072, "ce": 6.157639026641846, "genuineness": 4.2374773025512695, "identity": 0.42564448714256287, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/d19254d3c6211da32779fb04.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-0-seed1702-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "57fe2d5c2e9107501b4c54c6aea0f9debe123dbd6bffe3c8e0bfabfe53858bd7", "task_id": "vn_S_WHIS_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.43054819107055664, "target_z": -0.8359374900964144, "ce": 6.221180438995361, "genuineness": 4.519899368286133, "identity": 0.701641857624054, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/82bdf1756e1a00820695d762.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-0-seed1702-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "c787e90e08f9ed43d1cbdc190ed1559b7e14c872a6afbbad7e98e5c8669577ae", "task_id": "vn_S_WHIS_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.20573639869689941, "target_z": -0.9765624772800096, "ce": 6.086154937744141, "genuineness": 4.3127899169921875, "identity": 0.6520899534225464, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/40ac5b716e41d2111d8ec91a.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-0-seed1702-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "85672c7b1ca0f6eb61cfe1ecd731a58ee882bd2707037fe6e9157bb64c005248", "task_id": "vn_S_WHIS_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.01839315891265869, "target_z": -1.093750016311788, "ce": 6.1020402908325195, "genuineness": 4.362998008728027, "identity": 0.5355874300003052, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/d849c8e0e6f2a7ffaec2d155.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-0-seed1702-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "De deur staat nog open.", "Ik heb je jas op de stoel gelegd.", "Zeg maar wanneer je klaar bent." ], "text": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "ea6cd41bb0accfdc7855444d2d66a94ce55ae3eeba53fb3c771016235b58a646", "task_id": "vn_S_WHIS_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.39932429790496826, "target_z": -0.8554687590297398, "ce": 6.133810997009277, "genuineness": 4.369274139404297, "identity": 0.6643363237380981, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/38ca9eff913001be506186c7.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-1-seed1701-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "2efa47276b8185d6d779b552d0a7b66964e1281652338370704827e5ff6afa28", "task_id": "vn_S_WHIS_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.1307990550994873, "target_z": -1.0234375227199906, "ce": 6.197352409362793, "genuineness": 4.042920112609863, "identity": 0.6204383373260498, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/d9a309bb6ffb8b7d3e30b873.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-1-seed1701-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "c86d99f133c651a76ed628dc053d06cf1efa0d93a3f942c864fb9bb5f04245e7", "task_id": "vn_S_WHIS_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.6803390979766846, "target_z": -0.679687487766159, "ce": 6.24500846862793, "genuineness": 4.281409740447998, "identity": 0.6856727004051208, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen het er samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/bc9a4d155ac2cebd6481fb01.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-1-seed1701-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "0ac6f962582474f18554365cba31b8ba6797596581bb975fff2f62e238d9ae93", "task_id": "vn_S_WHIS_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.14328861236572266, "target_z": -1.0156250151466604, "ce": 6.229123115539551, "genuineness": 4.432034492492676, "identity": 0.819326639175415, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het Oude Daagraam. We kunnen er samen in kijken.", "wer": 0.05555555555555555, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/260c75987885979d2ed16ee3.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-1-seed1701-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1701, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "b648389a0721352aa68249a3a8abcef7c6145194dd7457b7bc4d46ea5f7ac443", "task_id": "vn_S_WHIS_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": -0.13148140907287598, "target_z": -1.187500032623576, "ce": 6.24500846862793, "genuineness": 4.281409740447998, "identity": 0.49700599908828735, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/0c60ec333eb11643398fd118.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-1-seed1702-scale0.0", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "2614bd89fcd13500bbe2a13bd8ed0c7d1d0a97156bfa8df5a7c92fcb8734cd03", "task_id": "vn_S_WHIS_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.443037748336792, "target_z": -0.8281249825230843, "ce": 6.260893821716309, "genuineness": 4.369274139404297, "identity": 0.6195645928382874, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/a40bb38280dfe27de8e4283a.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-1-seed1702-scale0.5", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.5, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "ccba806575b87d32350ba67aa493e4b6e78f65e88cce14722f8b0b02141ad3eb", "task_id": "vn_S_WHIS_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.14328861236572266, "target_z": -1.0156250151466604, "ce": 6.0146708488464355, "genuineness": 4.400654315948486, "identity": 0.6427230834960938, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/cbc83b70ed49f271658be549.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-1-seed1702-scale0.75", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 0.75, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "0c54e0ff2c92a720d51b4552ff5d103557ebf34cf8279cf6e94fd5bb7a470352", "task_id": "vn_S_WHIS_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.16826772689819336, "target_z": -1.0, "ce": 6.09409761428833, "genuineness": 4.268857479095459, "identity": 0.4501926898956299, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/c452ec01baf469d885d64c2d.mp3" }, { "key": "original-vn_S_WHIS_reg-nl-1-seed1702-scale1.0", "bucket": "vn_S_WHIS_reg", "language": "nl", "sentences": [ "Ik heb het kleine doosje gevonden.", "Het stond naast het oude raam.", "We kunnen er samen in kijken." ], "text": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "directions": [ "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word." ], "seconds": 10.2, "frames": 128, "seeds": [ 1701, 1702 ], "scales": [ 0.0, 0.5, 0.75, 1.0 ], "cfg": 1.0, "reference": true, "case_type": "Neutral wording, target emotion/style varies only through acoustic performance", "seed": 1702, "scale": 1.0, "frames_generated": 128, "seconds_generated": 10.24, "ended": true, "contract_sha256": "21397118883a5516dfd6b25aa06fd72b751379eac6386cba5f533cd4f6b48b2a", "mp3_sha256": "7b5bf8c676b05fd30b10a3356fd0dcfd00378a9d6eadf97711c6d3a3fe205621", "task_id": "vn_S_WHIS_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam.\n(Strong Whisper-Talk Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken.", "original_audio_reused": true, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 0.20573639869689941, "target_z": -0.9765624772800096, "ce": 6.1020402908325195, "genuineness": 4.118232727050781, "identity": 0.6821146011352539, "hypothesis": "Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken.", "wer": 0.0, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/audio/vn_S_WHIS_reg/nl/9f672063bd6f3c9b4e0884ba.mp3" } ], "source_standard_audio_unchanged": true }