{ "complete": true, "report": { "bucket": "vn_S_CONV_reg", "label": "Strong Conversational Style", "definition": { "id": "vn_S_CONV_reg", "category": "talking_style", "head": "vn_S_CONV_reg", "pole": "high", "label": "Strong Conversational Style", "floor": 4.0, "canonical_dimension_name": "Conversational Style" }, "recipe_status": "direct_manual_recipe", "manual_provenance": [ { "id": "rank4/style_s_conv/full-v2-sealed/sealed-selected", "mission": "style_s_conv", "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": 6.332298398017883, "genuineness": 4.484188513755798, "wer": 0.011363345200038748, "identity": 0.6087002092599869, "voiceclap": -0.2229728887602687, "fitness": 0.775530948296041, "steering_v2": 0.8787492774924832, "raw_target": 4.101361718177795, "target_cdf": 0.8787492774924832, "family_hit_all": null, "nominal_hit_all": null, "n": 100, "valid_fraction": 1, "by_arm": { "reference": { "n": 50, "ce": 6.349375143051147, "genuineness": 4.476594505310058, "wer": 0.011141645677936001, "identity": 0.6087002092599869, "fitness": 0.7772720464085884, "steering_v2": 0.8815285556316911 }, "none": { "n": 50, "ce": 6.3152216529846195, "genuineness": 4.491782522201538, "wer": 0.011585044722141497, "identity": null, "fitness": 0.7737898501834936, "steering_v2": 0.8759699993532754 } } } } ], "selected_wording": { "id": "manual-ea97be76622a", "general": "Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.", "inline": [ "Employ slight vocal breathiness during transitions to mimic natural thought retrieval.", "Vary the projection level slightly, moving between confidential softness and moderate conversational volume.", "Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response." ], "cfg": 3, "recipe": "P06", "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": 5.0 }, "n": 20, "target_z": 0.8892578159067981, "target_raw": 3.6941535472869873, "ce": 6.009905362129212, "genuineness": 4.496050250530243, "identity": 0.6101109743118286, "wer": 0.04391124871001032, "valid_fraction": 0.9, "identity_delta": 0.0017529085278511047, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, "manual_adapter": { "config": { "scale": 0.5, "cfg": 3.0 }, "n": 20, "target_z": 0.8164062650541234, "target_raw": 3.5990567445755004, "ce": 6.095686149597168, "genuineness": 4.4850671768188475, "identity": 0.5666230410337448, "wer": 0.008742690058479531, "valid_fraction": 1, "identity_delta": -0.0417350247502327, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, "grid": [ { "config": { "scale": 0.0, "cfg": 5.0 }, "n": 20, "target_z": 0.8892578159067981, "target_raw": 3.6941535472869873, "ce": 6.009905362129212, "genuineness": 4.496050250530243, "identity": 0.6101109743118286, "wer": 0.04391124871001032, "valid_fraction": 0.9, "identity_delta": 0.0017529085278511047, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 3.0 }, "n": 20, "target_z": 0.8606200890167581, "target_raw": 3.6567712783813477, "ce": 6.097274684906006, "genuineness": 4.491970849037171, "identity": 0.5885374203324318, "wer": 0.01388888888888889, "valid_fraction": 0.95, "identity_delta": -0.019820645451545715, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 4.0 }, "n": 20, "target_z": 0.8651855562972046, "target_raw": 3.6627308130264282, "ce": 6.024003553390503, "genuineness": 4.49981586933136, "identity": 0.5380832232534886, "wer": 0.0254609563123495, "valid_fraction": 0.95, "identity_delta": -0.07027484253048896, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 2.5 }, "n": 20, "target_z": 0.5231933522269668, "target_raw": 3.216311144828796, "ce": 6.131031036376953, "genuineness": 4.450548934936523, "identity": 0.5133890949189663, "wer": 0.002631578947368421, "valid_fraction": 1, "identity_delta": -0.09496897086501122, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 2.0 }, "n": 20, "target_z": 0.5000976797140987, "target_raw": 3.1861631989479067, "ce": 6.114351487159729, "genuineness": 4.40755809545517, "identity": 0.5362320303916931, "wer": 0.0027777777777777775, "valid_fraction": 1, "identity_delta": -0.07212603539228439, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 1.5 }, "n": 20, "target_z": 0.4809081729409475, "target_raw": 3.1611141681671144, "ce": 6.154461979866028, "genuineness": 4.355153179168701, "identity": 0.5371626794338227, "wer": 0.005263157894736842, "valid_fraction": 1, "identity_delta": -0.07119538635015488, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 1.0 }, "n": 20, "target_z": 0.42089844515191815, "target_raw": 3.0827804446220397, "ce": 6.166375970840454, "genuineness": 4.3774330615997314, "identity": 0.5226569667458534, "wer": 0.03172514619883041, "valid_fraction": 0.95, "identity_delta": -0.08570109903812409, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.0, "cfg": 1.25 }, "n": 20, "target_z": 0.41762696011250333, "target_raw": 3.078510010242462, "ce": 6.158830404281616, "genuineness": 4.423562002182007, "identity": 0.5561861038208008, "wer": 0.013651315789473683, "valid_fraction": 1, "identity_delta": -0.05217196196317673, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 3.0 }, "n": 20, "target_z": 0.8164062650541234, "target_raw": 3.5990567445755004, "ce": 6.095686149597168, "genuineness": 4.4850671768188475, "identity": 0.5666230410337448, "wer": 0.008742690058479531, "valid_fraction": 1, "identity_delta": -0.0417350247502327, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 1.5 }, "n": 20, "target_z": 0.6274413878336702, "target_raw": 3.3523913621902466, "ce": 6.169553017616272, "genuineness": 4.411323714256286, "identity": 0.5670784123241901, "wer": 0.011374269005847953, "valid_fraction": 1, "identity_delta": -0.04127965345978737, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 2.0 }, "n": 20, "target_z": 0.5586913986783467, "target_raw": 3.2626485347747805, "ce": 6.1288468599319454, "genuineness": 4.434545040130615, "identity": 0.5665004506707192, "wer": 0.005409356725146199, "valid_fraction": 1, "identity_delta": -0.04185761511325836, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 1.25 }, "n": 20, "target_z": 0.5448242399845188, "target_raw": 3.2445470333099364, "ce": 6.178687119483948, "genuineness": 4.411323666572571, "identity": 0.5600721567869187, "wer": 0.009444444444444445, "valid_fraction": 1, "identity_delta": -0.04828590899705887, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 4.0 }, "n": 20, "target_z": 0.8510741983181299, "target_raw": 3.644310545921326, "ce": 6.041278886795044, "genuineness": 4.480360126495361, "identity": 0.45310133751481774, "wer": 0.02719298245614035, "valid_fraction": 0.9, "identity_delta": -0.1552567282691598, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 5.0 }, "n": 20, "target_z": 0.791162095449044, "target_raw": 3.566104257106781, "ce": 5.959270787239075, "genuineness": 4.469377064704895, "identity": 0.5042166091501713, "wer": 0.02304093567251462, "valid_fraction": 1, "identity_delta": -0.10414145663380622, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 4.0 }, "n": 20, "target_z": 0.755810559436453, "target_raw": 3.5199581146240235, "ce": 5.982900238037109, "genuineness": 4.46749427318573, "identity": 0.5116010467521847, "wer": 0.0116593567251462, "valid_fraction": 1, "identity_delta": -0.09675701903179287, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 5.0 }, "n": 20, "target_z": 0.7447997652618389, "target_raw": 3.5055851697921754, "ce": 5.954703783988952, "genuineness": 4.45651113986969, "identity": 0.5119926284998655, "wer": 0.01918644650842793, "valid_fraction": 1, "identity_delta": -0.09636543728411198, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 5.0 }, "n": 20, "target_z": 0.7243896730186182, "target_raw": 3.478942847251892, "ce": 5.916777539253235, "genuineness": 4.415403151512146, "identity": 0.5280558466911316, "wer": 0.0331922944616443, "valid_fraction": 1, "identity_delta": -0.08030221909284592, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 4.0 }, "n": 20, "target_z": 0.7000488398570494, "target_raw": 3.447169530391693, "ce": 6.005933952331543, "genuineness": 4.409754729270935, "identity": 0.5203762128949165, "wer": 0.008388157894736842, "valid_fraction": 1, "identity_delta": -0.08798185288906098, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 2.0 }, "n": 20, "target_z": 0.6923339963121707, "target_raw": 3.4370989561080934, "ce": 6.093303394317627, "genuineness": 4.39939923286438, "identity": 0.5401479966938496, "wer": 0.005902777777777778, "valid_fraction": 1, "identity_delta": -0.06821006909012794, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 2.5 }, "n": 20, "target_z": 0.6605468657606209, "target_raw": 3.3956056118011473, "ce": 6.1222941160202025, "genuineness": 4.445528054237366, "identity": 0.5531423911452293, "wer": 0.008333333333333333, "valid_fraction": 0.95, "identity_delta": -0.05521567463874817, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 3.0 }, "n": 20, "target_z": 0.6577880896319139, "target_raw": 3.392004442214966, "ce": 6.0736452579498295, "genuineness": 4.371470820903778, "identity": 0.5472578026354313, "wer": 0.01, "valid_fraction": 0.95, "identity_delta": -0.06110026314854622, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 2.5 }, "n": 20, "target_z": 0.6529785052574535, "target_raw": 3.385726249217987, "ce": 6.121102690696716, "genuineness": 4.42199296951294, "identity": 0.5499628573656082, "wer": 0.01915204678362573, "valid_fraction": 0.95, "identity_delta": -0.058395208418369295, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 2.0 }, "n": 20, "target_z": 0.6083495790393361, "target_raw": 3.327469861507416, "ce": 6.09389910697937, "genuineness": 4.443331456184387, "identity": 0.5214393571019172, "wer": 0.032736498108015134, "valid_fraction": 0.95, "identity_delta": -0.08691870868206024, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 3.0 }, "n": 20, "target_z": 0.6048584059667913, "target_raw": 3.3229126572608947, "ce": 6.073645305633545, "genuineness": 4.38653335571289, "identity": 0.5225342605262995, "wer": 0.051228070175438595, "valid_fraction": 0.95, "identity_delta": -0.08582380525767803, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 1.5 }, "n": 20, "target_z": 0.5328613243035001, "target_raw": 3.2289312362670897, "ce": 6.185835528373718, "genuineness": 4.388102352619171, "identity": 0.5518124878406525, "wer": 0.023830409356725146, "valid_fraction": 0.95, "identity_delta": -0.05654557794332504, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 1.0 }, "n": 20, "target_z": 0.5261474512544345, "target_raw": 3.2201672792434692, "ce": 6.121499872207641, "genuineness": 4.336325025558471, "identity": 0.5475277192890644, "wer": 0.002631578947368421, "valid_fraction": 1, "identity_delta": -0.060830346494913104, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 1.5 }, "n": 20, "target_z": 0.5058105640026327, "target_raw": 3.19362051486969, "ce": 6.159227585792541, "genuineness": 4.371470856666565, "identity": 0.5506190851330757, "wer": 0.01111111111111111, "valid_fraction": 0.95, "identity_delta": -0.05773898065090179, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 1.25 }, "n": 20, "target_z": 0.5049316383552876, "target_raw": 3.1924732089042664, "ce": 6.182261300086975, "genuineness": 4.443645334243774, "identity": 0.53924600481987, "wer": 0.008333333333333333, "valid_fraction": 0.95, "identity_delta": -0.06911206096410752, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.75, "cfg": 2.5 }, "n": 20, "target_z": 0.49062498944070965, "target_raw": 3.173798018693924, "ce": 6.083772206306458, "genuineness": 4.404733920097351, "identity": 0.5290160275995731, "wer": 0.00804093567251462, "valid_fraction": 1, "identity_delta": -0.07934203818440437, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 1.0 }, "n": 20, "target_z": 0.4349365479115984, "target_raw": 3.101105087995529, "ce": 6.18464412689209, "genuineness": 4.380571103096008, "identity": 0.526631398126483, "wer": 0.023830409356725146, "valid_fraction": 0.95, "identity_delta": -0.08172666765749455, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 0.5, "cfg": 1.25 }, "n": 20, "target_z": 0.4079895083804464, "target_raw": 3.0659297585487364, "ce": 6.186629819869995, "genuineness": 4.448979926109314, "identity": 0.525599243491888, "wer": 0.03698830409356725, "valid_fraction": 0.95, "identity_delta": -0.08275882229208946, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" }, { "config": { "scale": 1.0, "cfg": 1.0 }, "n": 20, "target_z": 0.40104980122050327, "target_raw": 3.056871008872986, "ce": 6.129839658737183, "genuineness": 4.3388353943824765, "identity": 0.5516514360904694, "wer": 0.005263157894736842, "valid_fraction": 1, "identity_delta": -0.05670662969350815, "identity_pairs": 20, "gates_pass": false, "gate_scope": "development_provisional_20_pairs" } ], "development_standard": [ { "scale": 0.0, "n": 20, "target_z": 0.06953125335328815, "target_raw": 2.6241231322288514, "ce": 6.192983913421631, "genuineness": 4.2967860221862795, "identity": 0.6083580657839776, "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": 0.31643066671792963, "target_raw": 2.9464133858680723, "ce": 6.173127245903015, "genuineness": 4.377433049678802, "identity": 0.560355444252491, "wer": 0.0, "valid_fraction": 1, "identity_delta": -0.04800262153148651, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "scale": 0.75, "n": 20, "target_z": 0.20075685340179925, "target_raw": 2.7954185247421264, "ce": 6.19377818107605, "genuineness": 4.319693493843078, "identity": 0.5934447690844535, "wer": 0.008742690058479531, "valid_fraction": 1, "identity_delta": -0.014913296699523926, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" }, { "scale": 1.0, "n": 20, "target_z": 0.06489256738288411, "target_raw": 2.6180680215358736, "ce": 6.1369880676269535, "genuineness": 4.246577668190002, "identity": 0.6004750847816467, "wer": 0.005263157894736842, "valid_fraction": 1, "identity_delta": -0.00788298100233078, "identity_pairs": 20, "gates_pass": true, "gate_scope": "development_provisional_20_pairs" } ] }, "heldout": { "standard": { "n": 50, "target_z": 0.26108397268219125, "target_raw": 2.874166555404663, "ce": 6.240560531616211, "genuineness": 4.325969552993774, "identity": 0.5837892112135887, "wer": 0.010263504611330698, "valid_fraction": 1, "identity_delta": 0.0, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "manual": { "n": 50, "target_z": 0.7879101627906956, "target_raw": 3.5618593454360963, "ce": 6.070905027389526, "genuineness": 4.5003179264068605, "identity": 0.5285264265537262, "wer": 0.02275270686597918, "valid_fraction": 1, "identity_delta": -0.05526278465986252, "identity_pairs": 50, "gates_pass": false, "gate_scope": "heldout_50_reference_pairs" }, "adapter": { "n": 50, "target_z": 0.6863623100960149, "target_raw": 3.4293038129806517, "ce": 6.172174110412597, "genuineness": 4.479983587265014, "identity": 0.5721769249439239, "wer": 0.015780726519856954, "valid_fraction": 1, "identity_delta": -0.011612286269664764, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" }, "matched_no_adapter": { "n": 50, "target_z": 0.625039065949606, "target_raw": 3.3492554903030394, "ce": 6.174477519989014, "genuineness": 4.490652866363526, "identity": 0.5690372335910797, "wer": 0.020416619612271786, "valid_fraction": 0.98, "identity_delta": -0.014751977622509002, "identity_pairs": 50, "gates_pass": true, "gate_scope": "heldout_50_reference_pairs" } }, "adapter_vs_matched_cfg": { "n": 50, "delta_z": 0.061323244146408884, "ci95_z": [ -0.03576171592646003, 0.13406738749595548 ], "language_deltas": { "en": 0.08447266059322164, "de": -0.11611327157362329, "fr": 0.08984374807364297, "nl": 0.07968748701492676, "es": 0.16872559662387634 }, "languages_improved": 4, "clear_gain": false, "delta_ce": -0.002303409576416016, "delta_genuineness": -0.010669279098510741, "delta_wer": -0.004635893092414831, "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": -0.10154785269468067, "ci95_z": [ -0.11517577280078406, -0.08536134507961728 ], "language_deltas": { "en": -0.11953127618418616, "de": -0.10712889311331532, "fr": -0.09433595556851537, "nl": -0.11484370926111635, "es": -0.07189942934627015 }, "languages_improved": 0, "clear_gain": false, "delta_ce": 0.10126908302307129, "delta_genuineness": -0.020334339141845702, "delta_wer": -0.006971980346122223, "note": "Exploratory language-cluster bootstrap; five voices, one fixed reference voice per language. No familywise claim across72targets." }, "manual_vs_standard": { "n": 50, "delta_z": 0.5268261901085043, "ci95_z": [ 0.4328515764232804, 0.6227148366956092 ], "language_deltas": { "en": 0.4999512058047467, "de": 0.3661133035368805, "fr": 0.5611816834463113, "nl": 0.6998046679688802, "es": 0.5070800897857028 }, "languages_improved": 5, "clear_gain": true, "delta_ce": -0.16965550422668457, "delta_genuineness": 0.17434837341308593, "delta_wer": 0.012489202254648479, "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": 1793025074.136193 }, "rows": [ { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "0012087888999fddc927b6ca4fd15d75d737366bce00016e08f79d9aa93ac333", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "af7380fca79a1b5a88b6caaff0b72cdbe2ad5b53f7d04156fc760443a593b27b", "generation_seconds_batch": 25.305237293010578, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4409868717193604, "target_z": 0.6953124422092893, "ce": 6.133810997009277, "genuineness": 4.080576419830322, "identity": 0.35906803607940674, "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_CONV_reg/0012087888999fddc927b6ca4fd15d75d737366bce00016e08f79d9aa93ac333.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": "001d280403122c0b9f599818036e858368629e1ba56bd4a79e43cc0351060e9a", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The road turns left behind the bridge.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "99d32af5b7c1a602591127473b93e850e7793c43d0026f5cd4e0065e13f667d7", "generation_seconds_batch": 17.555115088936873, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5021753311157227, "target_z": 0.7421875634984352, "ce": 6.2767791748046875, "genuineness": 4.425758361816406, "identity": 0.32837051153182983, "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_CONV_reg/001d280403122c0b9f599818036e858368629e1ba56bd4a79e43cc0351060e9a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "0071b2022818e6996d802c3fe8161179622bb7a70b6cc499b812602c8a8b68b1", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7e7a8afee5b41dd3f81df27f750b71f614437de7eb90322c1d61fd1d3d2e7f9e", "generation_seconds_batch": 17.377995159942657, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.721433162689209, "target_z": 0.9101561240733281, "ce": 6.213237762451172, "genuineness": 4.7081804275512695, "identity": 0.5482209920883179, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la 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_CONV_reg/0071b2022818e6996d802c3fe8161179622bb7a70b6cc499b812602c8a8b68b1.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "00bb82ace8612aabf8ced3ef039c9df2f51fed80c7d1e1aa52951b83bdd76f55", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ac3ca61409b95f7d2e96baef0aa5a8e1dea7ef9ae02be579b18773b7664bc4d6", "generation_seconds_batch": 26.46138009498827, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3956868648529053, "target_z": -0.10546870326800561, "ce": 6.070269584655762, "genuineness": 4.218649387359619, "identity": 0.42190349102020264, "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_CONV_reg/00bb82ace8612aabf8ced3ef039c9df2f51fed80c7d1e1aa52951b83bdd76f55.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "00f49aa8ab73900bca197b2cb45b1c46dcb0be3097d5b791974f31a9de3b3e91", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "164838a9671f875ff856cb418e6b3a32c30d00e1b0210f80a03a1892fc83eaa7", "generation_seconds_batch": 23.298908620956354, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.590724468231201, "target_z": 0.043945277807736986, "ce": 6.173524379730225, "genuineness": 4.457139015197754, "identity": 0.6091731786727905, "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_CONV_reg/00f49aa8ab73900bca197b2cb45b1c46dcb0be3097d5b791974f31a9de3b3e91.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "011452c8561006fcd6c49b471460e602e917d068742fb68f2466a1b966b74be5", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d1f75d199f65176bc910a4a721107d46ab15153bae0f1a0eddb6573c58f10282", "generation_seconds_batch": 25.053604001062922, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.06306791305542, "target_z": 1.1718749272265125, "ce": 5.903473377227783, "genuineness": 4.469690799713135, "identity": 0.6583083868026733, "hypothesis": "Ik heb het een klein doosje gevonden. Het stond naast het, eh, daar 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_CONV_reg/011452c8561006fcd6c49b471460e602e917d068742fb68f2466a1b966b74be5.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": "011ec60068dee043c6f1636ed4610a5f91cbe06e960ba700b019dc841e5daf33", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d507846db961fa359bda39561b7234c04e13fff61c1ac6e38c7d3e36c8e35588", "generation_seconds_batch": 23.513895066920668, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8489089012145996, "target_z": 1.0078124593324629, "ce": 6.197352409362793, "genuineness": 4.507347106933594, "identity": 0.5085954666137695, "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_CONV_reg/011ec60068dee043c6f1636ed4610a5f91cbe06e960ba700b019dc841e5daf33.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "0156fe55ff415610c64cc8a67474f2a8ecd63b13e281480799a54f9f26b656ca", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "2f534a1d3e6cba6b559d903b9f2e300d7b45fccb9c298b4dbfb60aeb6cf2c42e", "generation_seconds_batch": 17.20448352897074, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.783212661743164, "target_z": 0.19140621254305792, "ce": 6.0146708488464355, "genuineness": 4.557555198669434, "identity": 0.7784184217453003, "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": 0.03999999999999915, "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_CONV_reg/0156fe55ff415610c64cc8a67474f2a8ecd63b13e281480799a54f9f26b656ca.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "01682206af29e2a11118e49b0aa3e56313f0f13c09bcf395c2a58281dab7e271", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "fa71674f4f899dfd724b7b180ca760fa30b4c2d053bb211eb95ed5df5f075cb3", "generation_seconds_batch": 18.01774843293242, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.527670383453369, "target_z": 0.7617187574913884, "ce": 6.125868320465088, "genuineness": 4.331617832183838, "identity": 0.5591764450073242, "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_CONV_reg/01682206af29e2a11118e49b0aa3e56313f0f13c09bcf395c2a58281dab7e271.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "51febdc7b7449bcd9c106588429f5e1911daf8f04cefda01fc0ef684840f19a8", "generation_seconds_batch": 17.683924596989527, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5716030597686768, "target_z": 0.029296790989429933, "ce": 5.974957466125488, "genuineness": 4.15588903427124, "identity": 0.5783830881118774, "hypothesis": "J'ai écrit le numéro sur l'enveloppe, l'IND1, et vérifie les deux derniers chiffres. Appelle-moi quand tu auras un moment.", "wer": 0.13636363636363635, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/01f5d5f22d9b51f4e674eec92be80d0710e62f5704093e3409bc811236eec3e5.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "029d30f858a2d311bf0ff7ef1ec69f197957e76256dc0141a5bfea8bc604e854", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3c6237c951b1c9c1117cc30f6fef574f80df9c75a99612524d66f26ed184f2cd", "generation_seconds_batch": 22.946933374973014, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.609254837036133, "target_z": 0.82421879744546, "ce": 6.022613525390625, "genuineness": 4.7332844734191895, "identity": 0.561586856842041, "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_CONV_reg/029d30f858a2d311bf0ff7ef1ec69f197957e76256dc0141a5bfea8bc604e854.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": "0403ad2b43390e8cee7b2e2004591ce5c74cca246e4ec5e01bc3c9e6829a7aa5", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The road turns left behind the bridge.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1f490b42cc6241eb7c8fe20d5c20fde2e9bbafd40fe4b1ab2c5634abb1817587", "generation_seconds_batch": 21.548421447048895, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2574219703674316, "target_z": 0.5546874436362205, "ce": 6.435632705688477, "genuineness": 4.507347106933594, "identity": 0.4249153733253479, "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_CONV_reg/0403ad2b43390e8cee7b2e2004591ce5c74cca246e4ec5e01bc3c9e6829a7aa5.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": "045408144378f9b34026d7f3c9611178b1189ad0403d457c3fae4becf2d375a0", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "275e3610c75e521ed86147964d9df418010da9922b8d0e16512fc227938cc471", "generation_seconds_batch": 15.992530610994436, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1350455284118652, "target_z": 0.4609375663522974, "ce": 6.165581703186035, "genuineness": 4.770940780639648, "identity": 0.6216147541999817, "hypothesis": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade daneben dem Fenster. Nimm sie mit, wenn du gehst.", "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_CONV_reg/045408144378f9b34026d7f3c9611178b1189ad0403d457c3fae4becf2d375a0.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": "04587464ba6315a66101a2199725f2a0c764771252abfe4a5ad56b105a81bebb", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e6019e73d42dde96e200dcaf90b4ae342ce876211d49cfc6203bece2a4961727", "generation_seconds_batch": 21.580728361965157, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4052937030792236, "target_z": 0.6679686975602811, "ce": 6.054384231567383, "genuineness": 4.469690799713135, "identity": 0.7587078809738159, "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_CONV_reg/04587464ba6315a66101a2199725f2a0c764771252abfe4a5ad56b105a81bebb.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "045cea4513fc91aa9498588eb6b64fb5344a705b856dca3f9dbd685adaf6f9b2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a6578f57f3995241e147b48b62b056f45a744fb88c2df52680f92ddcae83b7d3", "generation_seconds_batch": 17.130912492983043, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1120998859405518, "target_z": 0.44335941869976575, "ce": 6.324435234069824, "genuineness": 4.582659721374512, "identity": 0.4996601343154907, "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_CONV_reg/045cea4513fc91aa9498588eb6b64fb5344a705b856dca3f9dbd685adaf6f9b2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "04c2fab79bc29d6538ade411dd781b252b0515e906ffe66ea6aec995c499c711", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "f932450674a3c8d1c8789833b20850cd4307766f45fe4db735f8dde542fbf806", "generation_seconds_batch": 17.133493465022184, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.323709487915039, "target_z": 0.6054688402533938, "ce": 6.0305562019348145, "genuineness": 4.444586753845215, "identity": 0.687849223613739, "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": 0.03999999999999915, "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_CONV_reg/04c2fab79bc29d6538ade411dd781b252b0515e906ffe66ea6aec995c499c711.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "18cd79d4b243e860b63ed2ef8f2bc9f55d64bfd85f543a141f2d42488388c54a", "generation_seconds_batch": 17.796819278039038, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3390064239501953, "target_z": 0.6171874835902921, "ce": 6.324435234069824, "genuineness": 4.758388519287109, "identity": 0.631322979927063, "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_CONV_reg/059867605942d1f4c46c2c44d936e08b7fbe89dd4feef4c05500a7ed426c27a3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "05d3182c7305c51ea23bbac42f25915080ce58e2454c0b306889d449cedb46bf", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "449c7af285c4bd6427594ee85052e83b7b8c75b81545f10c55d33ac951707b67", "generation_seconds_batch": 25.67923100106418, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.318610429763794, "target_z": 0.6015625649253663, "ce": 6.23706579208374, "genuineness": 4.344170093536377, "identity": 0.6674972772598267, "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_CONV_reg/05d3182c7305c51ea23bbac42f25915080ce58e2454c0b306889d449cedb46bf.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "0641c6693781bb3be8eb656bf008f7a74a1a369a281d34c282ef5808407ef847", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "fa1d7769cb05f99bb91fc0d7c632ab6e7e07d2268da386e6c078d1fe8e2a3f61", "generation_seconds_batch": 23.298908620956354, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.170738697052002, "target_z": 0.4882813110013057, "ce": 6.109982967376709, "genuineness": 4.645420074462891, "identity": 0.5328992009162903, "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_CONV_reg/0641c6693781bb3be8eb656bf008f7a74a1a369a281d34c282ef5808407ef847.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "067be3fb23f31f6e5882513b84a756f5e39a39c8a7df9a2d3f0b334a5ed06f2c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e869e11c02cdd40533d092d6d58d314f10698ba1afa4bb855ef320c305943c95", "generation_seconds_batch": 24.975468205986544, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6806411743164062, "target_z": 0.8789062867434765, "ce": 6.141753673553467, "genuineness": 4.720732688903809, "identity": 0.7102342247962952, "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_CONV_reg/067be3fb23f31f6e5882513b84a756f5e39a39c8a7df9a2d3f0b334a5ed06f2c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "06b81dad978cb2e701048b318bac4e99854d2a69c03f559c7e5b76c1e176c944", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "348f54bf60591ce79fd199207041b7b121076779caacf81092cf7586c655d95d", "generation_seconds_batch": 17.68834655894898, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.624551773071289, "target_z": 0.8359374407823582, "ce": 6.133810997009277, "genuineness": 4.457139015197754, "identity": 0.6958392858505249, "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_CONV_reg/06b81dad978cb2e701048b318bac4e99854d2a69c03f559c7e5b76c1e176c944.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "06d5d2b16e85f4deb1732bcdaea2d0e632f5b29003bd6c09ff88a2b2074a27de", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6b9d99e0c4474749294964b63862bfeaf0afc848b14912d4c5e88b6c482ce39c", "generation_seconds_batch": 17.976073677069508, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.078956127166748, "target_z": 0.4179687203911791, "ce": 6.09409761428833, "genuineness": 4.570107460021973, "identity": 0.6564352512359619, "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_CONV_reg/06d5d2b16e85f4deb1732bcdaea2d0e632f5b29003bd6c09ff88a2b2074a27de.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "071304f42e83bc9d2cce4931f555453d81656a97d4c0de4589b6705bb8ab0c23", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "401ace4973c45627abc95db734e424b7d728f1f9619f4b6af369b1689c14e8f8", "generation_seconds_batch": 24.24352619005367, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.788311719894409, "target_z": 0.19531248787108543, "ce": 6.38003396987915, "genuineness": 4.570107460021973, "identity": 0.5016577839851379, "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_CONV_reg/071304f42e83bc9d2cce4931f555453d81656a97d4c0de4589b6705bb8ab0c23.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "071f01e668748a2f0852ec09e834ff00bce45b719e7e105ba5a0bb7705c336b5", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "a6b599dd04cb320ea513338323691d54b0476fa67666dbb4a371892cbaefc354", "generation_seconds_batch": 19.312209739000537, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.961087465286255, "target_z": 1.0937499686075152, "ce": 6.221180438995361, "genuineness": 4.507347106933594, "identity": 0.8311889171600342, "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_CONV_reg/071f01e668748a2f0852ec09e834ff00bce45b719e7e105ba5a0bb7705c336b5.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "077ab6ece96c69e1f614d55d03a4df8860ac21bd5b21e1f7c218095f883fb6de", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "33712d37f72b242baedd4d5c08deb520d919e91772322793e913a7a38bd1fbdd", "generation_seconds_batch": 19.307599302963354, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9112794399261475, "target_z": -0.47656248501722315, "ce": 5.375286102294922, "genuineness": 3.747946262359619, "identity": 0.5573304891586304, "hypothesis": "La puerta sigue abría, sigue, sigue, sigue, cállate cuando descansan.", "wer": 0.8, "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_CONV_reg/077ab6ece96c69e1f614d55d03a4df8860ac21bd5b21e1f7c218095f883fb6de.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "077ba5b34117ccb8342739916ef7b5de30d8d6a24edd70524b0dd7a7f5312dae", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a5126fd34736affb5d348d386cd2ef74d967d4e7541e28e01f682f4363c291ec", "generation_seconds_batch": 15.964217353030108, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5116896629333496, "target_z": -0.016601533158728712, "ce": 6.38003396987915, "genuineness": 4.457139015197754, "identity": 0.5838509798049927, "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_CONV_reg/077ba5b34117ccb8342739916ef7b5de30d8d6a24edd70524b0dd7a7f5312dae.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "07b03461c9202db91214c44fd8d0c030bb8348c581d671aa829bc4133b3e8403", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "09f3bee6dc53b058398ae677c1b257661b2c562236e6ac380c22a726a71b7ff3", "generation_seconds_batch": 23.444204721949063, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2574219703674316, "target_z": 0.5546874436362205, "ce": 5.804190158843994, "genuineness": 4.7081804275512695, "identity": 0.6608791351318359, "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_CONV_reg/07b03461c9202db91214c44fd8d0c030bb8348c581d671aa829bc4133b3e8403.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "07f9f24bca67f86ab9e2e86f2bb290aa4c4e584b447a089214c8f35ee03d61bf", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "766efa4af2e3192d9635cf766921fef40f79f9700527ca959ec136310577312b", "generation_seconds_batch": 22.247472380055115, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1254384517669678, "target_z": -0.3125000171231735, "ce": 5.919358730316162, "genuineness": 4.13078498840332, "identity": 0.3690357506275177, "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_CONV_reg/07f9f24bca67f86ab9e2e86f2bb290aa4c4e584b447a089214c8f35ee03d61bf.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "0846f14ae710c973f562f0278a98ca7869a49a64ff69db39e5a78bfe27939648", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "67d55ecf2cc454e73ba1d839987dd59bb2b58c2f2db766caa8dafc8ec41e29f2", "generation_seconds_batch": 18.126038942020386, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1101412773132324, "target_z": -0.324218843107256, "ce": 5.967014789581299, "genuineness": 4.030367851257324, "identity": 0.8723829388618469, "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_CONV_reg/0846f14ae710c973f562f0278a98ca7869a49a64ff69db39e5a78bfe27939648.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "0902f75201a291ed2937b620ac061bc96f67de195bea6add2bb4ee7554bbdd7a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b8eedb371f9997ecbc50e1af2336bd9b3a1358e0d782bc75815f861ace04395b", "generation_seconds_batch": 24.254187489044853, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.619452953338623, "target_z": 0.8320313481015149, "ce": 6.141753673553467, "genuineness": 4.545003414154053, "identity": 0.43763303756713867, "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_CONV_reg/0902f75201a291ed2937b620ac061bc96f67de195bea6add2bb4ee7554bbdd7a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "0906cae9e3e4252e3ebfb47473c929e149877921f5e6d5b96f7c00365957c586", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "fb7c355deb2c1bc2639194a559caa331cf1fdd9c4b7d230d20bb8363d9d0f828", "generation_seconds_batch": 23.823927422985435, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.563363552093506, "target_z": 0.7890625021403966, "ce": 6.23706579208374, "genuineness": 4.086852550506592, "identity": 0.47710686922073364, "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_CONV_reg/0906cae9e3e4252e3ebfb47473c929e149877921f5e6d5b96f7c00365957c586.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "0916f5f79560a8db6cf87a6324ee93c6217ac4d305d8319cb0e8b629801f6e3d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "ecc6115123f84935f5cb6154a2cae0f718a803d9e1a45d6f001507fe7264122a", "generation_seconds_batch": 18.11529825697653, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9304933547973633, "target_z": 1.0703124992865345, "ce": 6.149696350097656, "genuineness": 4.375550270080566, "identity": 0.4437902569770813, "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_CONV_reg/0916f5f79560a8db6cf87a6324ee93c6217ac4d305d8319cb0e8b629801f6e3d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "0a4ab9ae6b6fba689564a327ec1195ea611009a32431f26ff433db8a1f65e674", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "74192f70de6eef9fe546dcd4d524862c9c9c1f154cd0cd3abca3b5fc735ec148", "generation_seconds_batch": 15.517977745039389, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.042672157287598, "target_z": 1.156250191208771, "ce": 5.895530700683594, "genuineness": 4.645420074462891, "identity": 0.8403730392456055, "hypothesis": "Non, j'ai trouvé j'ai trouvé à la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu'elle contient.", "wer": 0.21052631578947367, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/0a4ab9ae6b6fba689564a327ec1195ea611009a32431f26ff433db8a1f65e674.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": "0a7d5248c28b7b751fab45e38481ad0fd36332f8c87ee33ffb4356dcb92902f3", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "8ea4fee036edfdf05dc94785cdd3b9d9bb411603194425fdcf9f8a4c78bc8eae", "generation_seconds_batch": 17.35063890099991, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.012077808380127, "target_z": 1.1328125392406059, "ce": 6.0146708488464355, "genuineness": 4.250029563903809, "identity": 0.642880916595459, "hypothesis": "Alguien ha movido la pequeña silla azul. Estaba aquí junto a la estantería alta. ¿La colocaríamos después de comer?", "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_CONV_reg/0a7d5248c28b7b751fab45e38481ad0fd36332f8c87ee33ffb4356dcb92902f3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "0b15e8cb8c9cb16db9bb318377d981269519d6a1e0c8244d3296edc410af5d15", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5554d7be347a5831b8bb6fbf3136c7de78d7cecddd7dea3d54135bded47a3447", "generation_seconds_batch": 23.903037442010827, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.344696521759033, "target_z": -0.14453127390109638, "ce": 5.911416053771973, "genuineness": 4.118232727050781, "identity": 0.558691143989563, "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_CONV_reg/0b15e8cb8c9cb16db9bb318377d981269519d6a1e0c8244d3296edc410af5d15.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "0b245748e938e065b004fbf35ca34ff6fd9af691342dc900c66568f49e3af45e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0ba981fde4eef3bbea27a226ff2cd2d866c12776839a040948afa3e2d4b4bc33", "generation_seconds_batch": 23.932636426063254, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7469284534454346, "target_z": 0.9296875007134655, "ce": 6.125868320465088, "genuineness": 4.281409740447998, "identity": 0.5513801574707031, "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_CONV_reg/0b245748e938e065b004fbf35ca34ff6fd9af691342dc900c66568f49e3af45e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "0bb9eaa1808eac176ac50813adfeb5a536217835a0b062f956cfa0bdfd3e2bba", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c8f56e67f7b0ec70c72aef332394e34d5b77ffbe0617dfe1fb461cfda83bf5aa", "generation_seconds_batch": 19.312209739000537, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.420590877532959, "target_z": 0.6796875235443636, "ce": 6.078212261199951, "genuineness": 4.532451152801514, "identity": 0.23343606293201447, "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_CONV_reg/0bb9eaa1808eac176ac50813adfeb5a536217835a0b062f956cfa0bdfd3e2bba.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "0bd581393c619de57e9b712fcc0a652cdc0521981738a74843d048d77dee9429", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "24e8de1a7728f3af27d3fceaa0361badf355060dffd576ae5dc5785b437472e8", "generation_seconds_batch": 23.66805281001143, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.609254837036133, "target_z": 0.82421879744546, "ce": 5.974957466125488, "genuineness": 4.344170093536377, "identity": 0.7660118341445923, "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_CONV_reg/0bd581393c619de57e9b712fcc0a652cdc0521981738a74843d048d77dee9429.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "0bf1b279bb8e632cfc307bdb62173cafc4b7c5778e86ec68d68e8a1849a7a555", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4cf6c2d99700713ef0a19a97d58eefa0d4caf98c0357755e83ae425ee39a05b0", "generation_seconds_batch": 22.13093923998531, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9406914710998535, "target_z": 1.0781250499425894, "ce": 5.998785495758057, "genuineness": 4.607763767242432, "identity": 0.7627279162406921, "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_CONV_reg/0bf1b279bb8e632cfc307bdb62173cafc4b7c5778e86ec68d68e8a1849a7a555.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "0c0d68f00d67c59e494f27ffe4cc52b6193416f576b372e5931e57201426c937", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ccf0d1b4c4b339b63c99c17b7f633915f12acb06f62154d137b7135a7aa9e5ba", "generation_seconds_batch": 23.63493035105057, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5378684997558594, "target_z": 0.7695313081474434, "ce": 6.125868320465088, "genuineness": 4.6705241203308105, "identity": 0.5758787393569946, "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_CONV_reg/0c0d68f00d67c59e494f27ffe4cc52b6193416f576b372e5931e57201426c937.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "58deac5c43be163300a482c8908df2d3f26b7423e9f1e673664ed4923d567d0d", "generation_seconds_batch": 16.69407768000383, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7462446689605713, "target_z": 0.16308585340024673, "ce": 6.165581703186035, "genuineness": 4.293961524963379, "identity": 0.7459996342658997, "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_CONV_reg/0cb70e86df8754d3bba44fe7f798aac7d0ffe2ca68fc75b92ff1a1ed29986ba7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "0d5ee14fe5aa3f808cc4cd6b762450a3563155a0afbae5424cab4bac5146292b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "215552dea50d8bf69b0885248463d01c45e30b3ad4141ee1e2253a95e6ebbafa", "generation_seconds_batch": 24.693376626004465, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2115306854248047, "target_z": 0.5195311483311572, "ce": 5.982900142669678, "genuineness": 4.469690799713135, "identity": 0.6747298836708069, "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_CONV_reg/0d5ee14fe5aa3f808cc4cd6b762450a3563155a0afbae5424cab4bac5146292b.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": "0db45f38d24865f98158a75a7e7e7c2ec1a44a652295fcae7724a4b477cad256", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "62b5fda2fd856ef5ec0616a369db8f0eeb83ac75a398c1e5dbfea5203b4197ad", "generation_seconds_batch": 15.029004493029788, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.675542116165161, "target_z": 0.875000011415449, "ce": 6.133810997009277, "genuineness": 4.645420074462891, "identity": 0.5593069791793823, "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_CONV_reg/0db45f38d24865f98158a75a7e7e7c2ec1a44a652295fcae7724a4b477cad256.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "0e30d63542baa1f78c881892b455df9d563f3ba0a870ffc8ef48e9ac74ae9594", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "0dd2951950daf4b6a9175e601f17834172abad5d48bfa24ed5a72242544a19ee", "generation_seconds_batch": 16.11871103802696, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.06306791305542, "target_z": 1.1718749272265125, "ce": 6.046441555023193, "genuineness": 4.168440818786621, "identity": 0.4427807927131653, "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_CONV_reg/0e30d63542baa1f78c881892b455df9d563f3ba0a870ffc8ef48e9ac74ae9594.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "0f060256b47f19d2f457e3df23a1908a1fe52d83368d9f6d1d1d1eef0f99d9d0", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2b80356ae387ac1170534bb5fab1f78fe32fda9c7a82bec966270177c0613a63", "generation_seconds_batch": 22.104711909079924, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.961087465286255, "target_z": 1.0937499686075152, "ce": 6.372091293334961, "genuineness": 4.482243061065674, "identity": 0.2093391716480255, "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_CONV_reg/0f060256b47f19d2f457e3df23a1908a1fe52d83368d9f6d1d1d1eef0f99d9d0.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "0f108849b290e08c14a22e5a949fa12846957dadb4f73be80580a01be868fdc2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f615816c771fa3fec96de2dc7847550c13e4c38e2bb7152e99aede2442e4681c", "generation_seconds_batch": 17.133493465022184, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.26702880859375, "target_z": 1.3281248444645073, "ce": 5.959072113037109, "genuineness": 4.469690799713135, "identity": 0.5372358560562134, "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_CONV_reg/0f108849b290e08c14a22e5a949fa12846957dadb4f73be80580a01be868fdc2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "0fe7c923c080b3c422cd29eabf51aacf90e834ea169b8db56bed605b3adc8afd", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7af8f57ed61a8bc113916b3db260227e6a9d9cabae804f315a5e2adae0aad9ba", "generation_seconds_batch": 24.673354918020777, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.420590877532959, "target_z": 0.6796875235443636, "ce": 6.252951145172119, "genuineness": 4.65797233581543, "identity": 0.6579872965812683, "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_CONV_reg/0fe7c923c080b3c422cd29eabf51aacf90e834ea169b8db56bed605b3adc8afd.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": "107d9726ebb66d5d730d8afca6f628aaa8431590033de8581da421a95754d7f8", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c68a4f76852faaef3a846f20a54bf8237ab04f7b453fc1d8fb7c185dcd0ff83f", "generation_seconds_batch": 15.145282767014578, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.969327211380005, "target_z": 0.3339844401037326, "ce": 6.205295085906982, "genuineness": 4.68307638168335, "identity": 0.38474220037460327, "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_CONV_reg/107d9726ebb66d5d730d8afca6f628aaa8431590033de8581da421a95754d7f8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "10f9968d8a70d1e56f2e7aa56303830e36467f9751ad0198c99f6cb47630b4cd", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "dda8b5279eef8907f08a1120b98756602b9a5d488ef27e725b494ddb32c9b965", "generation_seconds_batch": 23.29842176300008, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9406914710998535, "target_z": 1.0781250499425894, "ce": 6.157639026641846, "genuineness": 4.6705241203308105, "identity": 0.6404514312744141, "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": 0.03999999999999915, "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_CONV_reg/10f9968d8a70d1e56f2e7aa56303830e36467f9751ad0198c99f6cb47630b4cd.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "10facf9d4566d4fa33abcbb194dc06ebfad1512ebdc0e7312ca54fc6295ee756", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cba4dfd5ffd40dfd69cff6ae4f712762b153f40dd7ebdd7601150739db4b629d", "generation_seconds_batch": 18.415036966907792, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.170738697052002, "target_z": 0.4882813110013057, "ce": 6.149696350097656, "genuineness": 4.469690799713135, "identity": 0.4697655439376831, "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_CONV_reg/10facf9d4566d4fa33abcbb194dc06ebfad1512ebdc0e7312ca54fc6295ee756.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "11d8ea72d29039afd750487b8d49fd82164ba6f2eb2a44aace6540c3d129ff5d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8a4b5b972fde208bf498b8ff91e76b5f53b5451815794e271ab8182e08b518e5", "generation_seconds_batch": 18.695571545977145, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.186035633087158, "target_z": 0.49999995433820393, "ce": 6.141753673553467, "genuineness": 4.595211505889893, "identity": 0.544238805770874, "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_CONV_reg/11d8ea72d29039afd750487b8d49fd82164ba6f2eb2a44aace6540c3d129ff5d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "11fde491d7abed613dbf0e1d1744151070bcecea082e6043b23e491694eebe1b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3fdf3f8ec78156ee6b475460896aa7ee07758a66c9f4e6753892ee05e8f2dcbf", "generation_seconds_batch": 24.83776940801181, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6385278701782227, "target_z": 0.08056640352991254, "ce": 6.141753673553467, "genuineness": 4.444586753845215, "identity": 0.6853256225585938, "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_CONV_reg/11fde491d7abed613dbf0e1d1744151070bcecea082e6043b23e491694eebe1b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "120df1cd8179196ed7408873ada383fdc99b980b35070c77c3c8d5345969cf96", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b3efba481ad4df68c1763c233272427e464cf5d5715da9689d9c47bff462c676", "generation_seconds_batch": 22.83357784000691, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.205840587615967, "target_z": 1.2812499058225457, "ce": 6.324435234069824, "genuineness": 4.582659721374512, "identity": 0.333222895860672, "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_CONV_reg/120df1cd8179196ed7408873ada383fdc99b980b35070c77c3c8d5345969cf96.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "121cf8de1fb3a95696a745d392715f1a210f8e9ad78d3d0260d11c4af8fab5e3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0effd029322aad43e66009e23e40b4a6904cd3402fc63077c773f7c4808b4e16", "generation_seconds_batch": 17.78694170701783, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.12994647026062, "target_z": 0.45703129102426987, "ce": 5.824047088623047, "genuineness": 3.998987913131714, "identity": 0.6591799259185791, "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_CONV_reg/121cf8de1fb3a95696a745d392715f1a210f8e9ad78d3d0260d11c4af8fab5e3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "1226943490dd763f75c5abdd4d9977a7c44ebdfcca84afb4bf45c3eb5b928e6f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c48d049126ce2e39d9e1a8c30f4638bbe2716a251b142c5c218e37da4acadadd", "generation_seconds_batch": 23.978681421955116, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4970762729644775, "target_z": 0.7382812881704076, "ce": 6.109982967376709, "genuineness": 4.444586753845215, "identity": 0.5405009984970093, "hypothesis": "Ta porte est 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_CONV_reg/1226943490dd763f75c5abdd4d9977a7c44ebdfcca84afb4bf45c3eb5b928e6f.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "9639560fb41ab6f61e64881a723ebd534f7f396c1d6b6989d295997290183f48", "generation_seconds_batch": 17.58974752004724, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4243688583374023, "target_z": -0.08349606436413712, "ce": 6.24500846862793, "genuineness": 4.287685871124268, "identity": 0.6879021525382996, "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_CONV_reg/12a7b850c68b1e935d69d2b179f3c0df54c4775061935760902fd9f6cbfc7541.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": "132db96e4cf38da68aee358b9bdb459d222e89f6e57add80501bc873fdae8071", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4c24c996655ef4764b60437f518fee5ca930378ca2d020e5e770887515b63562", "generation_seconds_batch": 21.594552156049758, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4970762729644775, "target_z": 0.7382812881704076, "ce": 6.038498878479004, "genuineness": 4.582659721374512, "identity": 0.6716334819793701, "hypothesis": "De sleutels liggen op de keukentafel. Ik keek in de la en 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_CONV_reg/132db96e4cf38da68aee358b9bdb459d222e89f6e57add80501bc873fdae8071.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "139b178ce44ce91c05205a6776bcda9814585de2dbe08c61cad6bcb70c64c268", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7d72aa8e8a4c459aa1d8d322f11c3f08eb3475f294158fb502cc234a528885db", "generation_seconds_batch": 17.300805610022508, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.180936813354492, "target_z": 0.49609386165736064, "ce": 6.268836498260498, "genuineness": 4.7458367347717285, "identity": 0.7362003326416016, "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_CONV_reg/139b178ce44ce91c05205a6776bcda9814585de2dbe08c61cad6bcb70c64c268.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "139b7c88e07728e52692a0a1cb59bac2e49412d3443e1e876e40281d2ac6a340", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d487d09a619dddbbb6665a55b2a2573cc28dcb20414383055391c2de12501b17", "generation_seconds_batch": 22.577713961945847, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3033132553100586, "target_z": 0.5898437389412837, "ce": 5.704906940460205, "genuineness": 4.808597087860107, "identity": 0.7450491189956665, "hypothesis": "J'ai trouvé la petite boîte, elle était 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_CONV_reg/139b7c88e07728e52692a0a1cb59bac2e49412d3443e1e876e40281d2ac6a340.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "14968bade40ee911f2142dd5aaca539d11ba8508b6748bb3bb5a00214039ae1b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2682cbcbebd4eaeffec35417ca494b7d3c956ea9e3c0cabf0f278b33e8ca421b", "generation_seconds_batch": 17.300805610022508, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.662748336791992, "target_z": 0.09912116567624708, "ce": 6.356205940246582, "genuineness": 4.243753433227539, "identity": 0.6574444770812988, "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_CONV_reg/14968bade40ee911f2142dd5aaca539d11ba8508b6748bb3bb5a00214039ae1b.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": "1502717474f50ae02bdb1e2eed74a2c3857fd665fbd08606a1a08b92785ebc3d", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d17e3ea468af27c6bb9aaa9f0f3510a958328d6b567da59dea82f608d8c3e4b5", "generation_seconds_batch": 21.359757354017347, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.971876621246338, "target_z": 0.3359374864441543, "ce": 6.23706579208374, "genuineness": 4.18099308013916, "identity": 0.7048746347427368, "hypothesis": "Las llaves están en la mesa de la cocina. Mire en el cajón y luego, junto a la ventana, llévatelas cuando te vayas.", "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_CONV_reg/1502717474f50ae02bdb1e2eed74a2c3857fd665fbd08606a1a08b92785ebc3d.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "f58b48318919e5e858e9f2c4277cc717cc2223431fa8bf4bf3f7523f2473e2f6", "generation_seconds_batch": 17.614973227959126, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3090033531188965, "target_z": -0.17187501855010465, "ce": 6.260893821716309, "genuineness": 4.1998209953308105, "identity": 0.5186073184013367, "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_CONV_reg/15585f3962d31fd159bb39a755593ab24aa0df87e4725d7772f40811258760f5.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "155f826bdf508efd2dc474a0813b0de506c077b09f4f3ee3215eb28e68b95023", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "7f8902007ac01e67694115f107b6adfbe4dd68bf6cac7bf6fdfc61e93dc256b4", "generation_seconds_batch": 17.376710313023068, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2880163192749023, "target_z": 0.5781250956043855, "ce": 6.260893821716309, "genuineness": 4.337893962860107, "identity": 0.3800598382949829, "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_CONV_reg/155f826bdf508efd2dc474a0813b0de506c077b09f4f3ee3215eb28e68b95023.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": "161b50de9da14d7706aef44a037d53f90c15784c5865f12dea0c61f3aba62f80", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "e8bcf3cb2e530e8d19ffbc53b2555b2e389de2737bb46dad4ee3eaac38a4a7f5", "generation_seconds_batch": 28.373287641094066, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5837597846984863, "target_z": 0.8046876034525067, "ce": 6.157639026641846, "genuineness": 4.632867813110352, "identity": 0.4777790904045105, "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": 0.120000000000001, "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_CONV_reg/161b50de9da14d7706aef44a037d53f90c15784c5865f12dea0c61f3aba62f80.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "164c4b850ea302c8385872b4b8721dd333abfadf51cd2565f9f708bac361bda6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "567e10a910f22d5b6185e7837b632281d2e61e4108e6c337a7a08add2f7bd2f1", "generation_seconds_batch": 24.10360326897353, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.134454250335693, "target_z": 1.226562416524529, "ce": 6.141753673553467, "genuineness": 4.306513786315918, "identity": 0.5462163686752319, "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_CONV_reg/164c4b850ea302c8385872b4b8721dd333abfadf51cd2565f9f708bac361bda6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "1678129ece998c4552f484e92eeb9710077d2c7eb0918e5779b6fe18cb7ca967", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c491b45b81ea25552e000a23641ebfd254787c42ce15c60a12764550ce0efeca", "generation_seconds_batch": 23.483065218897536, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.730947732925415, "target_z": 0.15136721006334844, "ce": 5.879645347595215, "genuineness": 4.15588903427124, "identity": 0.6730512380599976, "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_CONV_reg/1678129ece998c4552f484e92eeb9710077d2c7eb0918e5779b6fe18cb7ca967.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "179ba4d6cc37903f54ef7f858ab0bbf0df6a701e16407775ffdc0caa6bf57e24", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0c0d5d23f3b1e479ef8d3b394f043e512ef25bd0c677679ce79f6e8277b294e6", "generation_seconds_batch": 21.756293547921814, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2880163192749023, "target_z": 0.5781250956043855, "ce": 5.700935363769531, "genuineness": 3.773050308227539, "identity": 0.6434228420257568, "hypothesis": "La puerta sigue a de fe. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "wer": 0.2, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/179ba4d6cc37903f54ef7f858ab0bbf0df6a701e16407775ffdc0caa6bf57e24.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": "17f1357e7e560364552f6a03ad1a6d332683fd897d323fd73c6511b389ec375f", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2aa54c97c7b087ff52ad275273f6bbff3c3b7f055e0aafbf524ed3fdcbb8d9fb", "generation_seconds_batch": 15.11226050904952, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.53590989112854, "target_z": 0.0019530463404216574, "ce": 5.935244083404541, "genuineness": 4.2751336097717285, "identity": 0.8504011631011963, "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_CONV_reg/17f1357e7e560364552f6a03ad1a6d332683fd897d323fd73c6511b389ec375f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "18175dd640bc4f225015216e0a5c2dae2232e63cc8bc63965f3e61fb1fe95962", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c941946ffd398574c554cc6f1868597f00a40da80ad59a0ee745aa143a0d34f0", "generation_seconds_batch": 18.479565732064657, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.563363552093506, "target_z": 0.7890625021403966, "ce": 5.442798614501953, "genuineness": 4.457139015197754, "identity": 0.7078575491905212, "hypothesis": "Puis si j'ai trouvé 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_CONV_reg/18175dd640bc4f225015216e0a5c2dae2232e63cc8bc63965f3e61fb1fe95962.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "18434951646945b1f63f70f3d810f2ade485751537c04fb7c16b3728d0482219", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a2ff1da773f49ac5d30ba122b9ea9c555dab9cdfb1f68f3a9691c6b82872d584", "generation_seconds_batch": 23.185613811016083, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.17838716506958, "target_z": 0.49414063266975483, "ce": 6.213237762451172, "genuineness": 4.507347106933594, "identity": 0.5779286026954651, "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_CONV_reg/18434951646945b1f63f70f3d810f2ade485751537c04fb7c16b3728d0482219.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "18d485211cd8b423d60ecf209730c8b11d10dcd5971505b421a766026bd74630", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9852338ee1bac866e6288f59d280bcb01a6f89e97aaab8e7656cbe1b36aa4857", "generation_seconds_batch": 27.02163700305391, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5531654357910156, "target_z": 0.7812499514843417, "ce": 6.38797664642334, "genuineness": 4.507347106933594, "identity": 0.49785295128822327, "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_CONV_reg/18d485211cd8b423d60ecf209730c8b11d10dcd5971505b421a766026bd74630.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "18e9c11e3d7a3f0710632a11ebe656bae6d5439e4e80458f6e8b061230708586", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "04ad031a79f7fd41794b40a59d0a694ef19a05fd61864df1029c77811e44419d", "generation_seconds_batch": 24.425943535985425, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5116896629333496, "target_z": -0.016601533158728712, "ce": 6.157639026641846, "genuineness": 4.306513786315918, "identity": 0.7231242656707764, "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_CONV_reg/18e9c11e3d7a3f0710632a11ebe656bae6d5439e4e80458f6e8b061230708586.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "18f6b15269f73a84f20bb355164b5a50f6ddd915666fecc11bf7adb410276000", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "983fc2737b8051582159b4cf5324db9b66e9dbdf624e231d9221b54ac0e24c1a", "generation_seconds_batch": 25.22124816698488, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7826216220855713, "target_z": 0.9570312453624739, "ce": 5.998785495758057, "genuineness": 4.6705241203308105, "identity": 0.31344977021217346, "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_CONV_reg/18f6b15269f73a84f20bb355164b5a50f6ddd915666fecc11bf7adb410276000.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "197c13f22cbcf1b61bcd6f11591ad6d78d16a6cf94a660d9d698753554df6584", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8312f09e2d2dd08a8563805a364f367e5378bff6ea7fbf9be896b5435536e7d9", "generation_seconds_batch": 22.302212193026207, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.979525089263916, "target_z": 0.34179680811260343, "ce": 6.141753673553467, "genuineness": 4.419482707977295, "identity": 0.5834946632385254, "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_CONV_reg/197c13f22cbcf1b61bcd6f11591ad6d78d16a6cf94a660d9d698753554df6584.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": "198eeefd6f5e2c9a7d8b58f15fdd3c0600cd297d787f9d624b518b018905b56d", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Your message arrived after I had left.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "81d0f2b4d15b63ae3b4ec0064b2ded781c61614ce574b7e427096e6039aaaaec", "generation_seconds_batch": 23.61702426301781, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.318610429763794, "target_z": 0.6015625649253663, "ce": 6.41180419921875, "genuineness": 4.287685871124268, "identity": 0.530829668045044, "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_CONV_reg/198eeefd6f5e2c9a7d8b58f15fdd3c0600cd297d787f9d624b518b018905b56d.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": "198f6785e42d7d31c104ae9b0f03750a459a7b9bff178d0b3820cbd6bb37f620", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Escribí el número en el sobre.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "29ac69b2d7c5a0de22c215bd0fbcb632db698efe3cd686998b453f026090c4c0", "generation_seconds_batch": 16.82258327503223, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.776839017868042, "target_z": 0.1865235053684117, "ce": 5.982900142669678, "genuineness": 4.281409740447998, "identity": 0.5649862289428711, "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_CONV_reg/198f6785e42d7d31c104ae9b0f03750a459a7b9bff178d0b3820cbd6bb37f620.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "199407c9b07fc548d33d27b051d42662816794f9d29be67090720a7627280e85", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "911d3075ec670113afeb033e29dd4c607e89c49994d2d0e1889bf284b5fd1afa", "generation_seconds_batch": 19.299460766022094, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.277818202972412, "target_z": 0.5703125449483305, "ce": 6.24500846862793, "genuineness": 4.017816066741943, "identity": 0.43035370111465454, "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_CONV_reg/199407c9b07fc548d33d27b051d42662816794f9d29be67090720a7627280e85.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "1995890963b34250e5e72fd9a7dec7bc0a16c8a26bd8e3ffabe97a1d8396e783", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e68cfdd227ff8f791c0dd6fbedd9c1ff7782e30bc3ae78aafa48c96f64492d6a", "generation_seconds_batch": 23.283609776059166, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.433929443359375, "target_z": -0.07617191227857567, "ce": 5.8875885009765625, "genuineness": 4.13078498840332, "identity": 0.4067617952823639, "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_CONV_reg/1995890963b34250e5e72fd9a7dec7bc0a16c8a26bd8e3ffabe97a1d8396e783.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "19bfa0f0c744433199dde768adef7f35a8f978563ef53f9765cdbd25caa21f26", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "638154697e4921382d4cd4ac72b30b13b555fff238518eb2136c4967c6dabca5", "generation_seconds_batch": 24.836885498953052, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4511849880218506, "target_z": 0.7031249928653444, "ce": 6.046441555023193, "genuineness": 3.9425034523010254, "identity": 0.3901747763156891, "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_CONV_reg/19bfa0f0c744433199dde768adef7f35a8f978563ef53f9765cdbd25caa21f26.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "1a43ca622c1641ebeb52a3291601337145d837ae84f4493c16d1ef1c91412357", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4d4e0637323d84de429874620c4e9d45050104dc0d3253ffa9160a6418aeec54", "generation_seconds_batch": 23.444204721949063, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8285129070281982, "target_z": 0.9921875406675371, "ce": 5.982900142669678, "genuineness": 4.331617832183838, "identity": 0.6257456541061401, "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_CONV_reg/1a43ca622c1641ebeb52a3291601337145d837ae84f4493c16d1ef1c91412357.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": "1a5041cf8969c9fecf97ed2c1a950391f3ecabfe77ad3d1f6e931c25aed1df57", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d54470a362fc3abc8831cb81b7a8caca2dcdbe65981197e23b16f01fbdd0d146", "generation_seconds_batch": 15.133737206924707, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.716334342956543, "target_z": 0.9062500313924848, "ce": 6.268836498260498, "genuineness": 4.796044826507568, "identity": 0.6171703338623047, "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_CONV_reg/1a5041cf8969c9fecf97ed2c1a950391f3ecabfe77ad3d1f6e931c25aed1df57.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "1b22dad54886569780f6f2c08edb5e7cfd2e18c88a54634e9b1eec872868841d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1fef68b8d680069697bfbdee14f6c9f484c030e33bf30bba13c99795bf5051f5", "generation_seconds_batch": 26.88887528993655, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4052937030792236, "target_z": 0.6679686975602811, "ce": 5.967014789581299, "genuineness": 4.300237655639648, "identity": 0.48062142729759216, "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_CONV_reg/1b22dad54886569780f6f2c08edb5e7cfd2e18c88a54634e9b1eec872868841d.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": "1b51ca491b9e80831f5725ac9a255b40f7c4e20be42e045a5bca2badaf7870e5", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6ddd66a3a310ad64e4b08583c9c4d6b936cf032c69fc5c976042804d0f22dda1", "generation_seconds_batch": 15.145282767014578, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0942533016204834, "target_z": 0.42968754637526163, "ce": 6.1894097328186035, "genuineness": 4.7332844734191895, "identity": 0.7076421976089478, "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_CONV_reg/1b51ca491b9e80831f5725ac9a255b40f7c4e20be42e045a5bca2badaf7870e5.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "1b570a233c4715a7d11b7433a60ac8eef6b2f73b4e5f02b83896ff01c5883e22", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "82d6ea2e711b0f70c067ce4a55ddd52dafc4dd01084bf6c8b62f23a264353a50", "generation_seconds_batch": 26.196921192924492, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.086604595184326, "target_z": 0.42382804205962826, "ce": 6.308549880981445, "genuineness": 4.482243061065674, "identity": 0.5903575420379639, "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_CONV_reg/1b570a233c4715a7d11b7433a60ac8eef6b2f73b4e5f02b83896ff01c5883e22.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": "1b5997c17a765800b0a36fec1fed528c60bf946ebe861fe8c3a62b24a7821213", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Escribí el número en el sobre.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "07b495956b45d9d8ce85f4bfbf87f58f3eac48d50691673bca37a8f6c2861225", "generation_seconds_batch": 21.20019926398527, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0942533016204834, "target_z": 0.42968754637526163, "ce": 6.117925643920898, "genuineness": 4.457139015197754, "identity": 0.6753761768341064, "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_CONV_reg/1b5997c17a765800b0a36fec1fed528c60bf946ebe861fe8c3a62b24a7821213.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "1b8a0a8795b79a2acd0ed4a7926716731509d65e1808d344205690e34be75df6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2423f1335ffc581be0e78383459fef5d4599a2c04c6abbdad9158b0331738fa1", "generation_seconds_batch": 17.294604912051, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7571263313293457, "target_z": 0.9374998687223364, "ce": 6.316492557525635, "genuineness": 4.570107460021973, "identity": 0.6036677360534668, "hypothesis": "I found the little box. It was beside the 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_CONV_reg/1b8a0a8795b79a2acd0ed4a7926716731509d65e1808d344205690e34be75df6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "1be8fa25e34dedf8bb469b04a05f1a8ad324cbd21a2d070f3c4399d47af9008f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "30c3e007569bda72cbe6558402c6f8cfba7ed73f05c43f41ff9fa4c9b76feaaf", "generation_seconds_batch": 26.02799546904862, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.181527853012085, "target_z": -0.26953117116205527, "ce": 6.006728172302246, "genuineness": 4.049196243286133, "identity": 0.6132941246032715, "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_CONV_reg/1be8fa25e34dedf8bb469b04a05f1a8ad324cbd21a2d070f3c4399d47af9008f.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": "1c3d0c5273a04f75f19c5f9a59593e750fc99f6994390f912101fc972f67f2ab", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c301f4d61e8f8e6a82a7c64593f64e6b7bb6ea8a65291582248ff436cd71c5c2", "generation_seconds_batch": 15.169848765945062, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5378684997558594, "target_z": 0.7695313081474434, "ce": 6.109982967376709, "genuineness": 4.720732688903809, "identity": 0.7700859904289246, "hypothesis": "La route tourna à gauche après le pont, continue jusqu'à la petite barrière en bois. Je je t'attendrai de l'autre côté.", "wer": 0.1, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/1c3d0c5273a04f75f19c5f9a59593e750fc99f6994390f912101fc972f67f2ab.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "1c5043f45ec7f5e39bb0641d78b3f01dcaf55c1db30e209c92c76080852f2b71", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ea74ca015c828235023d0b08170a035fb22365eb6a1111b6c09eee856135f5bb", "generation_seconds_batch": 17.061552187078632, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.737321376800537, "target_z": 0.15624991723799467, "ce": 6.427689552307129, "genuineness": 4.2249250411987305, "identity": 0.519788920879364, "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_CONV_reg/1c5043f45ec7f5e39bb0641d78b3f01dcaf55c1db30e209c92c76080852f2b71.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "1d0e1ce98db503f846a65dbc6dcc98ad984d1fe4c04e5ef92486a6f5604e5884", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "8673c5a3612f16cec4b280b7dc36323298fcb3d7537bd68ee89aa3d84026e86f", "generation_seconds_batch": 18.91927820595447, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6347498893737793, "target_z": 0.8437499914384132, "ce": 6.133810997009277, "genuineness": 4.394378662109375, "identity": 0.39725130796432495, "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_CONV_reg/1d0e1ce98db503f846a65dbc6dcc98ad984d1fe4c04e5ef92486a6f5604e5884.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "9ae93d8ad0cda6ad1762c3882ec9f18825673df79dfcb3e3d710efa5b1827d51", "generation_seconds_batch": 17.843989157001488, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.62896728515625, "target_z": 0.0732422514443511, "ce": 6.284721851348877, "genuineness": 4.099404335021973, "identity": 0.753247857093811, "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_CONV_reg/1d559e6da25f8bbacc3b53d5880369445633ce7ab0125ca7b07093108b8c8de3.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "9d2b172d3402cc22a34c0747ff93e2fd18d0cfe487dbe9956b207bfb23708baa", "generation_seconds_batch": 17.8303987220861, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4460859298706055, "target_z": 0.6992187175373169, "ce": 6.38003396987915, "genuineness": 4.381826400756836, "identity": 0.5717245936393738, "hypothesis": "Die Schlüsse liegen auf dem Küchentisch. Ich suchte in der Schublade dann neben 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_CONV_reg/1d829402ab2176d2024ae3e918286a39ec094d9a5299caa0e048ba37912e1579.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "1e8065d041e5ff4cea6d8698f8774521d10ec4e9e2a751a1a72113761f4a9f97", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0acb9361d6df077d09d4d76987c7034d21c94f1c06c1c16abab4420d4ae84e19", "generation_seconds_batch": 16.095190137042664, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.683781623840332, "target_z": 0.11523430026448225, "ce": 6.213237762451172, "genuineness": 4.306513786315918, "identity": 0.5581436157226562, "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_CONV_reg/1e8065d041e5ff4cea6d8698f8774521d10ec4e9e2a751a1a72113761f4a9f97.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "1ea087dfe66caa6b0f88d4a845f7eed09ed91a33d239fd94100901714691103b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "fc84949ef9d1b2d235777ba9dfc8581c86d186ca7468a56257ef903667b381f8", "generation_seconds_batch": 26.34194855997339, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.379798650741577, "target_z": 0.6484375035673278, "ce": 6.09409761428833, "genuineness": 4.557555198669434, "identity": 0.11134643852710724, "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_CONV_reg/1ea087dfe66caa6b0f88d4a845f7eed09ed91a33d239fd94100901714691103b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "1ea7469af48f160b55beb6be9570fe34e68f6fac7618bc1ea6b4fbcea65899d2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2892f5f77a68ee04e6b26a7eca0ce1a74382693ad1edbbc5e002174c46ac79d8", "generation_seconds_batch": 24.329753921017982, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.524437189102173, "target_z": -0.006835936162252069, "ce": 5.935244083404541, "genuineness": 4.444586753845215, "identity": 0.6439195871353149, "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_CONV_reg/1ea7469af48f160b55beb6be9570fe34e68f6fac7618bc1ea6b4fbcea65899d2.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "e22b1d7935b4684ab5e24ce548a764ea2bc2bd66d067eab4783ed5693d41284e", "generation_seconds_batch": 17.843989157001488, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3039042949676514, "target_z": -0.17578129387813213, "ce": 6.23706579208374, "genuineness": 4.143336772918701, "identity": 0.6423805952072144, "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_CONV_reg/1eaf0e0e2c966186cf6ada2de41819a315f8b1b1eeeeef5a84a28dd582998db8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "1f0fb5b7b6387f62daf2baa94a4da3ebe1c6a2d4aedb8b0a8618570af2ff688f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f94601ede9d077be71252b2c44ff14f34883a330411b4746b80a6c3708de4c2b", "generation_seconds_batch": 16.600682582939044, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.563363552093506, "target_z": 0.7890625021403966, "ce": 6.324435234069824, "genuineness": 4.532451152801514, "identity": 0.5797468423843384, "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_CONV_reg/1f0fb5b7b6387f62daf2baa94a4da3ebe1c6a2d4aedb8b0a8618570af2ff688f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "1f0fc9a99bcb61e9e915f9a874a2c50710fd5c0bdff1535031ed355193261255", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0d1098e754c2c0a0f5ecd3b82ff8907b4ddd52e6aff05c7ac4df21053959e614", "generation_seconds_batch": 22.055877617909573, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0509114265441895, "target_z": 0.39648429741062, "ce": 6.443574905395508, "genuineness": 4.65797233581543, "identity": 0.6827150583267212, "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_CONV_reg/1f0fc9a99bcb61e9e915f9a874a2c50710fd5c0bdff1535031ed355193261255.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "1f1b2e998c16d1b9c48c353aa89d39f41a881e4674d4692608967f7d3f5ca10a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cae0cc1ad9dd72d15549c5d3b605af8690e99b4f400eff19813ba10441c93a40", "generation_seconds_batch": 24.09584124397952, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5837597846984863, "target_z": 0.8046876034525067, "ce": 6.0146708488464355, "genuineness": 4.68307638168335, "identity": 0.6640453338623047, "hypothesis": "Et 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_CONV_reg/1f1b2e998c16d1b9c48c353aa89d39f41a881e4674d4692608967f7d3f5ca10a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "1f34b7ece12cc2d53655b2714658892b3682bb1b03983813495094abc124aae3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "577c6ba9c2fc1009aeb17f23e47d30d3c826119291c8183d832e1d6b8d40f05c", "generation_seconds_batch": 25.454617141978815, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.001879692077637, "target_z": 1.124999988584551, "ce": 6.038498878479004, "genuineness": 4.645420074462891, "identity": 0.7617467641830444, "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_CONV_reg/1f34b7ece12cc2d53655b2714658892b3682bb1b03983813495094abc124aae3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "1f538ed159e366234e657c1d6bc152e53ec80fbed490ec0cd214911627455f1a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3f294ee9b03d1f8897879a1c670e6f65f6d53e72dc95234fff62eaeee75d9683", "generation_seconds_batch": 26.560341875068843, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.134454250335693, "target_z": 1.226562416524529, "ce": 5.95112943649292, "genuineness": 4.545003414154053, "identity": 0.2875756025314331, "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_CONV_reg/1f538ed159e366234e657c1d6bc152e53ec80fbed490ec0cd214911627455f1a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "1f9d3b491f084eaa25c7dabb266cb7fb1506ac4801a3d8780f4f8870c62126be", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "feed9e4f681d84f2d299f6802fe79fc1b0237e97eff8e051e1471c202ff589f1", "generation_seconds_batch": 16.828201264957897, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1605405807495117, "target_z": 0.48046876034525066, "ce": 5.95112943649292, "genuineness": 4.432034492492676, "identity": 0.15851491689682007, "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_CONV_reg/1f9d3b491f084eaa25c7dabb266cb7fb1506ac4801a3d8780f4f8870c62126be.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "1ff7b256a963bbd0306be13c834b0da56995cc242085a16f605a43b95b552b41", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "26a8d2a365307cdbf6b1ffbd865005216659db0f7c184aff08e9741c499f17f0", "generation_seconds_batch": 22.577713961945847, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7469284534454346, "target_z": 0.9296875007134655, "ce": 6.046441555023193, "genuineness": 4.833701133728027, "identity": 0.743070662021637, "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_CONV_reg/1ff7b256a963bbd0306be13c834b0da56995cc242085a16f605a43b95b552b41.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "2042a62532aeca8e89824e79cc9b653156e1bf3a15d29cff9184be8d87fa8833", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6642a3009c57629188475f1f9b79798f9b0a437c54385e3bf7e8a6fb12b4314e", "generation_seconds_batch": 25.583739039953798, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.716334342956543, "target_z": 0.9062500313924848, "ce": 5.974957466125488, "genuineness": 4.582659721374512, "identity": 0.6358917951583862, "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_CONV_reg/2042a62532aeca8e89824e79cc9b653156e1bf3a15d29cff9184be8d87fa8833.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "210b0014efc5f7a148d317705bb38326413c12d9133e6f95b559d526b58a6a55", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "181580c47582c58a22cdef125041bb332c7cdf7c49115752f472ebbc17133630", "generation_seconds_batch": 18.375780744943768, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.2364349365234375, "target_z": 1.3046875577907107, "ce": 6.0305562019348145, "genuineness": 4.570107460021973, "identity": 0.5180609226226807, "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_CONV_reg/210b0014efc5f7a148d317705bb38326413c12d9133e6f95b559d526b58a6a55.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "2163854ba6f0cc62ee3cc05609524aec871b95f6be62750674142be4c75162a1", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2e20766026d5be73df80393d2acbfa2e0649bc86b63116dcff5f41ec6ab482e5", "generation_seconds_batch": 17.05012504302431, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.629650831222534, "target_z": 0.8398437161103858, "ce": 6.157639026641846, "genuineness": 4.570107460021973, "identity": 0.668298602104187, "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_CONV_reg/2163854ba6f0cc62ee3cc05609524aec871b95f6be62750674142be4c75162a1.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": "21e750c8dca396f77f25914e0680980c3f55f2619e62f718554be49fd5307a8d", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Escribí el número en el sobre.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9c68781eab078649f41f5c8a09771105234ca434cda64b3e5813e7101810df55", "generation_seconds_batch": 16.82258327503223, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2880163192749023, "target_z": 0.5781250956043855, "ce": 6.125868320465088, "genuineness": 4.494794845581055, "identity": 0.7010172605514526, "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_CONV_reg/21e750c8dca396f77f25914e0680980c3f55f2619e62f718554be49fd5307a8d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "21ff175f1dd016d9530902cd5d7aa19144ff0f9b1bba22f7e911a05b900ae5bb", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ac378a7dbb72ea3f99f82bc7614ea39575d4ce613219158187a8f5daffa5d20c", "generation_seconds_batch": 23.111561674042605, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3492045402526855, "target_z": 0.625000034246347, "ce": 6.372091293334961, "genuineness": 4.494794845581055, "identity": 0.3866073489189148, "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_CONV_reg/21ff175f1dd016d9530902cd5d7aa19144ff0f9b1bba22f7e911a05b900ae5bb.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "220be349a49b115938b69c2a9729edbf2b12bdee302600925be31a8b68450669", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "be936a64577bb2e29f02aa1e723f535b93312df2a4dd2a18a5c2749689d8fb2d", "generation_seconds_batch": 26.198932545026764, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4919772148132324, "target_z": 0.7343750128423802, "ce": 6.229123115539551, "genuineness": 4.582659721374512, "identity": 0.8259865045547485, "hypothesis": "Jij ben ik 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_CONV_reg/220be349a49b115938b69c2a9729edbf2b12bdee302600925be31a8b68450669.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "2223700f74c4d4600073e1e63560856b1f6889ef180e76118e94d29e2a79ada0", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "42a7a2b81adee31f3337816fc6dddb0348d9415bf830b5524bedd9d1e9dc1e01", "generation_seconds_batch": 24.9696722519584, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8285129070281982, "target_z": 0.9921875406675371, "ce": 5.903473377227783, "genuineness": 4.68307638168335, "identity": 0.7116900682449341, "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": 0.03999999999999915, "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_CONV_reg/2223700f74c4d4600073e1e63560856b1f6889ef180e76118e94d29e2a79ada0.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "224c1007ac2db4f5a09da281b87ed32332f324b6d14cfa253807d0a8b1dfe061", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1b635ad071f2888c2083180fbb2b54271b4894f1f2749827c73cc31295bdeb21", "generation_seconds_batch": 25.406823559082113, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.632791519165039, "target_z": 0.07617191227857567, "ce": 6.0305562019348145, "genuineness": 4.193545341491699, "identity": 0.5692223310470581, "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_CONV_reg/224c1007ac2db4f5a09da281b87ed32332f324b6d14cfa253807d0a8b1dfe061.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "22da5196aef9c8ca3e65a4fb0f5db023518b408956edbfc4499d6369546d576f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9d179607a2328d293bad6dbd2f4c8b665f01d5401d8f3024fef95b218fe3ff87", "generation_seconds_batch": 23.24755011498928, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2574219703674316, "target_z": 0.5546874436362205, "ce": 5.839932441711426, "genuineness": 4.607763767242432, "identity": 0.5339040756225586, "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_CONV_reg/22da5196aef9c8ca3e65a4fb0f5db023518b408956edbfc4499d6369546d576f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "22f0b7a5dafe1fbd85841ecfcb0b71aede9e000081065fc22a4a431fb52de859", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e8869b01cb3880ad5b8fc2a947eb5b801be15f850c8f5b5de7ce77e20f37e1d9", "generation_seconds_batch": 23.66536195005756, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.06306791305542, "target_z": 1.1718749272265125, "ce": 6.24500846862793, "genuineness": 4.519899368286133, "identity": 0.5560545921325684, "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_CONV_reg/22f0b7a5dafe1fbd85841ecfcb0b71aede9e000081065fc22a4a431fb52de859.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "2331475a2eb5f1e0a1599d879a6ec7d40de918d449c948c043acbefe55223780", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b4e8970b58f24ad7f7f6b2c036428433d31cc28300cf4af7ee16427328224ca5", "generation_seconds_batch": 18.151362388045527, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.685149192810059, "target_z": 1.64843759489092, "ce": 5.95112943649292, "genuineness": 4.7458367347717285, "identity": 0.7117319107055664, "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_CONV_reg/2331475a2eb5f1e0a1599d879a6ec7d40de918d449c948c043acbefe55223780.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "23552de7512089674b680f7948ef565338814f767daeff75d02304b59496b451", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "8ed99954a886bd61597f8d6750e4e4435bb51649da511c39002cdb6e64517006", "generation_seconds_batch": 22.576245508971624, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5429673194885254, "target_z": 0.7734374008282867, "ce": 6.1894097328186035, "genuineness": 4.093128204345703, "identity": 0.3566054105758667, "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_CONV_reg/23552de7512089674b680f7948ef565338814f767daeff75d02304b59496b451.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": "235ba6df72c24ee31ea65e26246b9915bd062fe99642f1a9dd21b231577101a4", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Escribí el número en el sobre.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8424db868a1eb2c0ab0637a615c13153acd99a5069d560bbfc37bd9ce47b00f4", "generation_seconds_batch": 21.20019926398527, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0534610748291016, "target_z": 0.3984375263982258, "ce": 6.348263263702393, "genuineness": 4.632867813110352, "identity": 0.5615893602371216, "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_CONV_reg/235ba6df72c24ee31ea65e26246b9915bd062fe99642f1a9dd21b231577101a4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "238e7227cba9d6706977060726ec874a67713f1e7e379b02ef261c7ec311bbc3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5048d91a2237f36aa45a8522b79a65cfb94d33973e66517dd876794bc876128d", "generation_seconds_batch": 22.946933374973014, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.471580982208252, "target_z": 0.7187499115302701, "ce": 6.046441555023193, "genuineness": 4.645420074462891, "identity": 0.5930730104446411, "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_CONV_reg/238e7227cba9d6706977060726ec874a67713f1e7e379b02ef261c7ec311bbc3.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": "23c7996d5532eeb86d70f14bc85cd01e1c4bd89237106604f817baf1e64bcf25", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Someone moved the little blue chair.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "960bd90eefc3e7657b1f2f7bbfbae59d76e918f2ce94874947ad1ad31bdbe8a8", "generation_seconds_batch": 15.214639994082972, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9406914710998535, "target_z": 1.0781250499425894, "ce": 6.356205940246582, "genuineness": 4.444586753845215, "identity": 0.5457929372787476, "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_CONV_reg/23c7996d5532eeb86d70f14bc85cd01e1c4bd89237106604f817baf1e64bcf25.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "23ea67a2e599e6f22049938075a328f203520feaaaa2e600e021d5daa7da16d1", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8b0479f2c6a12b2d287433df492c130f91522c05c5f9e5f08837239aaf40897e", "generation_seconds_batch": 22.104711909079924, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9508895874023438, "target_z": 1.0859376005986443, "ce": 6.221180438995361, "genuineness": 4.482243061065674, "identity": 0.4594377279281616, "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_CONV_reg/23ea67a2e599e6f22049938075a328f203520feaaaa2e600e021d5daa7da16d1.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "24a57f9a0acd0cb4e4e038811de17ed208f0d73ca75bfeef848a4458f8fb013f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6e0d7201475a53b5fed4a3ce7afffefb3dc44dba184d0154ff6a5fe5140063c8", "generation_seconds_batch": 22.34058797894977, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.170738697052002, "target_z": 0.4882813110013057, "ce": 6.197352409362793, "genuineness": 4.287685871124268, "identity": 0.4443066120147705, "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_CONV_reg/24a57f9a0acd0cb4e4e038811de17ed208f0d73ca75bfeef848a4458f8fb013f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "24f7fa6d37d1fcceaefdabaeb4d570f2679d26aeb2280d517e78e49c5ef6dafe", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a8479a9181201e48b60b885ec0f35c44c19505799cf32b67cfbdfc8ce95fbc04", "generation_seconds_batch": 19.0542571679689, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.512373447418213, "target_z": 0.7500001141544901, "ce": 6.038498878479004, "genuineness": 4.469690799713135, "identity": 0.6848940849304199, "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_CONV_reg/24f7fa6d37d1fcceaefdabaeb4d570f2679d26aeb2280d517e78e49c5ef6dafe.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "25427feb7352addcd0cb3588a47ef2e30ff088bca04eac97f87c3ab1443ff1c9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "93d9545f2b44d5a447e383c882e07cbcd5ee4b1661bcd5a494e1b6948b7092b6", "generation_seconds_batch": 24.684078398975544, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.723299264907837, "target_z": 0.14550788839489928, "ce": 5.696964263916016, "genuineness": 4.413206577301025, "identity": 0.4450589418411255, "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_CONV_reg/25427feb7352addcd0cb3588a47ef2e30ff088bca04eac97f87c3ab1443ff1c9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "258766abb35456569369eaa46e752b3ea72e880b9773902923ccdde0b066d16b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "106defe4d14038abf93f83c2879a71faa769561e245387412ee63c50645917eb", "generation_seconds_batch": 23.477816154947504, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.282917022705078, "target_z": 0.5742186376291738, "ce": 6.173524379730225, "genuineness": 4.231201171875, "identity": 0.520118236541748, "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_CONV_reg/258766abb35456569369eaa46e752b3ea72e880b9773902923ccdde0b066d16b.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": "258d41c77d6f04a784b22eb7be6bb27e7bfcef328512dfea658adcf83989e316", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6302947fe9094021830a6a5152a10b647be54dc08938fb2a4423a2933264e70a", "generation_seconds_batch": 21.38254901196342, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.563363552093506, "target_z": 0.7890625021403966, "ce": 6.173524379730225, "genuineness": 4.557555198669434, "identity": 0.5066068172454834, "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_CONV_reg/258d41c77d6f04a784b22eb7be6bb27e7bfcef328512dfea658adcf83989e316.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "261f8ca8c049600df319db7f7456079bf17024c6e897a2cab91216c66fbebb54", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5e9f9096e478ed3e81ea91505ddd50f93835febd62cf08b710cd6a1f5fb7c866", "generation_seconds_batch": 23.51673040597234, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.17838716506958, "target_z": 0.49414063266975483, "ce": 6.1020402908325195, "genuineness": 4.381826400756836, "identity": 0.6173672676086426, "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_CONV_reg/261f8ca8c049600df319db7f7456079bf17024c6e897a2cab91216c66fbebb54.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "26613ee9a2a7c8fb7bb60c15edf7b3f8133d12237e4d18030c93daac571dc3b2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "2368f4f4992aa4728656a5439f99f03ad9c8377e76328025c2a58e88d44448e5", "generation_seconds_batch": 16.329503054963425, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.001879692077637, "target_z": 1.124999988584551, "ce": 6.252951145172119, "genuineness": 4.570107460021973, "identity": 0.2901945114135742, "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_CONV_reg/26613ee9a2a7c8fb7bb60c15edf7b3f8133d12237e4d18030c93daac571dc3b2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "2679a0e21cbcb1146ef2f730a5242deb176a5deb183f3790719815e0569ce063", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e13accdc0973410ccf37532d6e252c89494e37f3d3d20a7fde6449a854999065", "generation_seconds_batch": 23.254128345055506, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.277818202972412, "target_z": 0.5703125449483305, "ce": 6.181467056274414, "genuineness": 4.457139015197754, "identity": 0.7676262259483337, "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_CONV_reg/2679a0e21cbcb1146ef2f730a5242deb176a5deb183f3790719815e0569ce063.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "26accd10b2c76149a5b93e73783dfc01d85c0cc9707c4e8373123c8b172b76d7", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a4d4cd0933848bdf24047f7157a88e378708e39e38f03606e2eee4f6c9501d21", "generation_seconds_batch": 23.418251048075035, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.409801483154297, "target_z": 1.4374998230605402, "ce": 5.77639102935791, "genuineness": 4.20609712600708, "identity": 0.03344427049160004, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. 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_CONV_reg/26accd10b2c76149a5b93e73783dfc01d85c0cc9707c4e8373123c8b172b76d7.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "809d1196d73d3d24b730713722c7deb1d0d3d01ce4a200ce6626ecb6f5c83ae5", "generation_seconds_batch": 19.46237588895019, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.084055185317993, "target_z": 0.4218749957192066, "ce": 6.395919322967529, "genuineness": 4.344170093536377, "identity": 0.18616779148578644, "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_CONV_reg/26d12995a5b024b42cbe037fdee48ffb67b79df656c09f059592185aa881aca9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "26dc92758024ac5b482170f8d0b7773f3da0365c2d8f3e97a0cbc02fff1a44c0", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0139e7477cdb713331c82a14b7e236cb0c00e337fd24841e4da34c92d5e1358f", "generation_seconds_batch": 23.165857334039174, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8795032501220703, "target_z": 1.0312501113006278, "ce": 6.09409761428833, "genuineness": 4.607763767242432, "identity": 0.4834092855453491, "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_CONV_reg/26dc92758024ac5b482170f8d0b7773f3da0365c2d8f3e97a0cbc02fff1a44c0.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "27d576306041706538df12bf331038b6ccc6844f110c066a4ed2067426fee90e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "317695cf4c55b4b167d915a95f3686b9c413d18b4b221b598b56d6aff8cb76c1", "generation_seconds_batch": 22.247472380055115, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.205840587615967, "target_z": 1.2812499058225457, "ce": 5.998785495758057, "genuineness": 4.444586753845215, "identity": 0.613669753074646, "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_CONV_reg/27d576306041706538df12bf331038b6ccc6844f110c066a4ed2067426fee90e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "28302220b6f2238bf58d25764278c867082e4a4dc078c1b0bdca350d69eff41a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "cf0e3b02175d93748cfa772212013a9bc3963e55fa263360dd68fb1f5a5246cf", "generation_seconds_batch": 23.418251048075035, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.26702880859375, "target_z": 1.3281248444645073, "ce": 6.086154937744141, "genuineness": 4.369274139404297, "identity": 0.679306149482727, "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_CONV_reg/28302220b6f2238bf58d25764278c867082e4a4dc078c1b0bdca350d69eff41a.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": "28d7736f5b90c4b56fd84949872cb92ebeb0753bac7fb081a310b332931b8070", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "811dcd6188b2367f4351d1bd4d59a7100b85885fcf94c90b189bd995943257c2", "generation_seconds_batch": 17.555278534069657, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.117198944091797, "target_z": 0.44726569402779326, "ce": 6.046441555023193, "genuineness": 4.507347106933594, "identity": 0.5260723233222961, "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_CONV_reg/28d7736f5b90c4b56fd84949872cb92ebeb0753bac7fb081a310b332931b8070.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "292fbf4b6444f8db3ee7adc6ac31f000547271cb45c1436366ccf471fc0621f1", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "2f3e6e1a61d899faf2ce49cda1ef6c2ca72aa41f7b129e4d8f74ad2defe21748", "generation_seconds_batch": 24.113650102051906, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2064318656921387, "target_z": 0.5156250556503139, "ce": 6.09409761428833, "genuineness": 3.8922953605651855, "identity": 0.42226484417915344, "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_CONV_reg/292fbf4b6444f8db3ee7adc6ac31f000547271cb45c1436366ccf471fc0621f1.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": "297842069ba3dc88c00263f05a27a7601c151cf389b5187a814c761178716ed4", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "892f28e179d35568d9aa4ba8bcd05c3a0d640e34cf46715361c1d79674cf918b", "generation_seconds_batch": 25.05367086699698, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6041557788848877, "target_z": 0.8203125221174324, "ce": 6.165581703186035, "genuineness": 4.720732688903809, "identity": 0.8411597609519958, "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_CONV_reg/297842069ba3dc88c00263f05a27a7601c151cf389b5187a814c761178716ed4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "29d1df68828f0dacb0e4b824ad48b9a9a6fe75a5715ddbdf925d668620ad3855", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3efe072c7b08180b4dac4d77307d67095bf0db1d39ba51ec9066fb85a2ca9c3c", "generation_seconds_batch": 24.693376626004465, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.124256134033203, "target_z": 1.2187498658684741, "ce": 6.1894097328186035, "genuineness": 4.557555198669434, "identity": 0.5522009134292603, "hypothesis": "Ja, 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_CONV_reg/29d1df68828f0dacb0e4b824ad48b9a9a6fe75a5715ddbdf925d668620ad3855.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "29da42b62ebe5d82eb581ff9785ae0bb685331483df0e9431f6bc27cbb5d032d", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3652f35a36e60ff2611c6224ccef405c12ffc155b307485452aa3c16c3db33b3", "generation_seconds_batch": 15.585673489025794, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.614353656768799, "target_z": 0.8281248901263033, "ce": 6.292664527893066, "genuineness": 4.3127899169921875, "identity": 0.1623455137014389, "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_CONV_reg/29da42b62ebe5d82eb581ff9785ae0bb685331483df0e9431f6bc27cbb5d032d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "2a8cf78b195194354df8290b6119bc29319ba46c19e530586cbd7e493d75d6c2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "24cb55aad47743227f78aff207fa38808b1df9420f228fff00c3553f46ac8a32", "generation_seconds_batch": 17.25354571407661, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.777522563934326, "target_z": 0.9531249700344463, "ce": 5.998785495758057, "genuineness": 4.632867813110352, "identity": 0.6188931465148926, "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_CONV_reg/2a8cf78b195194354df8290b6119bc29319ba46c19e530586cbd7e493d75d6c2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "2a8e1b5393387588363b38a890d0a3180eb847a106c30757e0a6f405b92c7297", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "382ec92ba10881d1be23b5a1dff80305a9449497aa1b306484ac9856459fdfac", "generation_seconds_batch": 24.78121139202267, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8622474670410156, "target_z": 0.25195302350952364, "ce": 6.038498878479004, "genuineness": 4.6705241203308105, "identity": 0.6527366638183594, "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_CONV_reg/2a8e1b5393387588363b38a890d0a3180eb847a106c30757e0a6f405b92c7297.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "2ad8c49467d4fa37f5fe0cb87da997ca62ee153365e78cad54048580c7afc3ee", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e48a75a69e151f75b4fffbc3d4716902a4a94213aa974046d82e9c5ad325dcca", "generation_seconds_batch": 25.990586593048647, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.83292818069458, "target_z": 0.22949216868234576, "ce": 6.340320587158203, "genuineness": 4.20609712600708, "identity": 0.5385712385177612, "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_CONV_reg/2ad8c49467d4fa37f5fe0cb87da997ca62ee153365e78cad54048580c7afc3ee.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "2aff0c4cdf385ec0bfaeef6e4556c8678f9ff2d28346743e469c75a226fbca06", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9e7efd5abd9ef435d728d39b1d83287a359554d85d17601653cb65b337a04742", "generation_seconds_batch": 22.2809402680723, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8545989990234375, "target_z": 0.24609370184107449, "ce": 5.8875885009765625, "genuineness": 4.444586753845215, "identity": 0.5951476097106934, "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_CONV_reg/2aff0c4cdf385ec0bfaeef6e4556c8678f9ff2d28346743e469c75a226fbca06.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "2b19909493154fd7beb32850241d5c4abc97d45929c05ee08ab13ff343c917fe", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2505a0ab087152c8d6337c7715741cfa5c0f84211818ec0fe554ca6455a33967", "generation_seconds_batch": 17.973642447032034, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2523231506347656, "target_z": 0.5507813509553772, "ce": 6.157639026641846, "genuineness": 4.582659721374512, "identity": 0.5708023309707642, "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_CONV_reg/2b19909493154fd7beb32850241d5c4abc97d45929c05ee08ab13ff343c917fe.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "2bbabb2104e743e526f338f36cc001fc4f04b7ce3c42c409339f1a66f42f1f22", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "0297f9cee5a357f4be570f30838aad9441d8ba47fa3977848334d8aadb1b490d", "generation_seconds_batch": 25.67923100106418, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9412825107574463, "target_z": 0.3125000171231735, "ce": 6.284721851348877, "genuineness": 4.444586753845215, "identity": 0.632843554019928, "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_CONV_reg/2bbabb2104e743e526f338f36cc001fc4f04b7ce3c42c409339f1a66f42f1f22.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": "2c17d97d5674f33930ec51eccc733609ad6ff4785856b9771b3f2188d87bf914", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7e4d9f5ed2b940b07a37b01b1cb134b9aa17d858f9d69884fdd4c945328fb63d", "generation_seconds_batch": 21.552040781942196, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.577658176422119, "target_z": 0.03393548152601357, "ce": 5.974957466125488, "genuineness": 4.193545341491699, "identity": 0.6649079918861389, "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_CONV_reg/2c17d97d5674f33930ec51eccc733609ad6ff4785856b9771b3f2188d87bf914.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "2c348eb42b77329cb640bc2dfc6358cde46876a3b0c48c098858c6b6c4df14d7", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f91e2d78102ca40def4723c9bbc6616d5d5d19271b7e4e4fdb474feb74f6a166", "generation_seconds_batch": 23.659395167953335, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.665343999862671, "target_z": 0.867187460759394, "ce": 5.8875885009765625, "genuineness": 4.482243061065674, "identity": 0.2960100769996643, "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_CONV_reg/2c348eb42b77329cb640bc2dfc6358cde46876a3b0c48c098858c6b6c4df14d7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "2c60360627b042852aac8e8c58e5ffa4c10af3d3e4392ddfba9fb3cd5535f881", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "885ad03b57c3ecb0b76a37b32eb1d78217a3d7e771a6af254136b28ac608afd0", "generation_seconds_batch": 25.494953381014057, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2880163192749023, "target_z": 0.5781250956043855, "ce": 5.95112943649292, "genuineness": 4.607763767242432, "identity": 0.6424272060394287, "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_CONV_reg/2c60360627b042852aac8e8c58e5ffa4c10af3d3e4392ddfba9fb3cd5535f881.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "2ca950581fb6e610e82c07c9f2830822cbbe5770e68400587c0363551a3a9427", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "92068ed5f0e27ff960ebada74a23526cd0bb5c2c6bc47acf58055d4f155489ab", "generation_seconds_batch": 21.258632850949652, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6856937408447266, "target_z": 0.11669913068159454, "ce": 6.181467056274414, "genuineness": 4.482243061065674, "identity": 0.7485325932502747, "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_CONV_reg/2ca950581fb6e610e82c07c9f2830822cbbe5770e68400587c0363551a3a9427.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "2d42962d6d60c8c057883af7e545053730dfc63c98a3d9c0fa2af876f129f318", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5fe6721e939e374acae7c0d04b168664123d4db0ad7dfbeb537c6260d8ed6621", "generation_seconds_batch": 23.51210604305379, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.379798650741577, "target_z": 0.6484375035673278, "ce": 6.229123115539551, "genuineness": 4.595211505889893, "identity": 0.5786536931991577, "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_CONV_reg/2d42962d6d60c8c057883af7e545053730dfc63c98a3d9c0fa2af876f129f318.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": "2d8159b6c9515d80fc95dec2ae40931fad243d4a0b4ea887f1d0a2d617137dfc", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ec563c9307972c15be1a0e3bc6e95533b66c964bf2892473cd2434afffede4c6", "generation_seconds_batch": 28.373287641094066, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1911346912384033, "target_z": 0.5039062296662314, "ce": 6.149696350097656, "genuineness": 4.469690799713135, "identity": 0.3557009696960449, "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_CONV_reg/2d8159b6c9515d80fc95dec2ae40931fad243d4a0b4ea887f1d0a2d617137dfc.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "2e801cbea35512baff1fd8b219a7f1efab400bc08db9ad1ef1644e4ec5d9a911", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1f8daa840ab8de4894114a12eb20ddb17ac055ea69b7a1386135a37403dfc468", "generation_seconds_batch": 24.329753921017982, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3587188720703125, "target_z": -0.1337890624108168, "ce": 6.054384231567383, "genuineness": 4.15588903427124, "identity": 0.40632519125938416, "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_CONV_reg/2e801cbea35512baff1fd8b219a7f1efab400bc08db9ad1ef1644e4ec5d9a911.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "2ef60f2d27d1f7ee9d62351abf63f9342814ebae1662e37648c9d5a85db3d587", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ec883ca37e55c0340092edf5280dc1004f7d4944b862f8f7740c13105538e236", "generation_seconds_batch": 17.377999630058184, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5531654357910156, "target_z": 0.7812499514843417, "ce": 6.292664527893066, "genuineness": 4.381826400756836, "identity": 0.28278958797454834, "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_CONV_reg/2ef60f2d27d1f7ee9d62351abf63f9342814ebae1662e37648c9d5a85db3d587.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "2f3d0b3a858301d536bf9a5d6613be9149c4489b43d2564bff4bd53bdfdd3edc", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2b43a6483d5458e88d2b5c8d264489fe553967c3d59f20cff1a66f6034e6dc64", "generation_seconds_batch": 23.165857334039174, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.430197715759277, "target_z": 1.4531249243726503, "ce": 6.268836498260498, "genuineness": 4.582659721374512, "identity": 0.5998109579086304, "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_CONV_reg/2f3d0b3a858301d536bf9a5d6613be9149c4489b43d2564bff4bd53bdfdd3edc.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "2f739a37b9be10d22cfe40ca6f564ca71d3d058629c0e640e5fc8fb565323059", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "29c38c2293bc0a2c2f240339223eb49f69045363b159b94f9f98ce668f4cab43", "generation_seconds_batch": 24.12591344199609, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1528921127319336, "target_z": 0.4746094386768015, "ce": 6.0146708488464355, "genuineness": 4.381826400756836, "identity": 0.5960010886192322, "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_CONV_reg/2f739a37b9be10d22cfe40ca6f564ca71d3d058629c0e640e5fc8fb565323059.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "2f987fd20f9f5eff94c207c9d955818a6ce85a05fcdbc497b3c8cc27ca5431eb", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "07ddc53ee2294cd691c04efe54e37bbe3f91edd9a0a8d1ed95515d7e4bb3abf0", "generation_seconds_batch": 25.68976305401884, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.96226966381073, "target_z": -0.4375000057077245, "ce": 6.086154937744141, "genuineness": 4.049196243286133, "identity": 0.5798274874687195, "hypothesis": "Ta porte est 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_CONV_reg/2f987fd20f9f5eff94c207c9d955818a6ce85a05fcdbc497b3c8cc27ca5431eb.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "3001fbb177b81e6cd9fc92839ca216e8cfb912ea1dc1f08c66e89205c073d87a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "aa7d1afe6c7e0cf2c832736bbe060c23e7d2759dc6cf2542047047f2b9d94588", "generation_seconds_batch": 24.33279059000779, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6041557788848877, "target_z": 0.8203125221174324, "ce": 6.062326908111572, "genuineness": 4.080576419830322, "identity": 0.61891108751297, "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_CONV_reg/3001fbb177b81e6cd9fc92839ca216e8cfb912ea1dc1f08c66e89205c073d87a.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": "306f208683dce4e36e0ed5ff31e86f563f5ea4b9ae90c0ced453e91da0621873", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7f0e70fea21f4463d2ed545e2caf2a8525af44d01275d6f7ef752bccf0335a30", "generation_seconds_batch": 17.35063890099991, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.205840587615967, "target_z": 1.2812499058225457, "ce": 6.157639026641846, "genuineness": 4.720732688903809, "identity": 0.7087207436561584, "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_CONV_reg/306f208683dce4e36e0ed5ff31e86f563f5ea4b9ae90c0ced453e91da0621873.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": "307124c4af14a8816b76beebc9217bcb229a310ab30ff1131aa6e504d6b09e63", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c2763d1b73417b18cf5537e725d95e87fad2d32226abb815d8a2012ce24c5082", "generation_seconds_batch": 21.55607965704985, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.716334342956543, "target_z": 0.9062500313924848, "ce": 6.109982967376709, "genuineness": 4.457139015197754, "identity": 0.4588984549045563, "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_CONV_reg/307124c4af14a8816b76beebc9217bcb229a310ab30ff1131aa6e504d6b09e63.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "307583918fc4b49e6700efbd0075c24ea812466e9bfa9c25f660ff6b1a31b52c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "226eff6b3685d8fc8294e97e2100b09ed6608ce1ae6c26c85228b6a7ad2f8a01", "generation_seconds_batch": 24.45010371808894, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.388038158416748, "target_z": -0.11132820758363894, "ce": 5.7684478759765625, "genuineness": 4.174716949462891, "identity": 0.5376040935516357, "hypothesis": "La puerta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "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_CONV_reg/307583918fc4b49e6700efbd0075c24ea812466e9bfa9c25f660ff6b1a31b52c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "30c1849f67e52642ded684ecd2d6eb416ec4abe076ae70ae9c16ea85b4e6d68c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ff5543b6f156bb8d1796eedefd48009fe84bec2551d92ce2c86b2a8a2caff856", "generation_seconds_batch": 17.973642447032034, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.97442626953125, "target_z": 0.33789071543176014, "ce": 6.173524379730225, "genuineness": 4.419482707977295, "identity": 0.5365384817123413, "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_CONV_reg/30c1849f67e52642ded684ecd2d6eb416ec4abe076ae70ae9c16ea85b4e6d68c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "31451552421092add4beb2a2106a1ede52ae0b6946b3f2271c41d393d2bca535", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "34615a03a15721d1df230efed0bd0787201bf6901e41f2f24815d1c5039705b3", "generation_seconds_batch": 24.52233663003426, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3950958251953125, "target_z": 0.6601563295514103, "ce": 6.308549880981445, "genuineness": 4.419482707977295, "identity": 0.42826372385025024, "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_CONV_reg/31451552421092add4beb2a2106a1ede52ae0b6946b3f2271c41d393d2bca535.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": "31c32f49997cfbfbc324601576e4d55833d62030c0d754f11e3a27d9b95c119a", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b8d6402181fe562fb65439ccd4a7683ed4204dae95f5e23ebf9b3da8391851dc", "generation_seconds_batch": 17.52500246197451, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6347498893737793, "target_z": 0.8437499914384132, "ce": 5.772419452667236, "genuineness": 4.758388519287109, "identity": 0.678996205329895, "hypothesis": "J'ai écrit le numéro sur l'enveloppe, lis-le bien and verifie les deux derniers chiffres. Appelle-moi quand tu auras 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_CONV_reg/31c32f49997cfbfbc324601576e4d55833d62030c0d754f11e3a27d9b95c119a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "3265cb187d76b63b02bbb43dedcc0efd9fd37f981cfa99b341b961227ee0c1fe", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2efacd8accaa256eff4630896b48ae4bba6f42d1352a99f8338b4a546d471bef", "generation_seconds_batch": 24.113650102051906, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5072741508483887, "target_z": 0.7460936561792785, "ce": 6.109982967376709, "genuineness": 4.174716949462891, "identity": 0.3441583514213562, "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_CONV_reg/3265cb187d76b63b02bbb43dedcc0efd9fd37f981cfa99b341b961227ee0c1fe.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "336341e069f1a22161b983ab80995bded285f2fb989b0c7fe6652e7e34c32ddb", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "451227dbbe10e997915bad6129109de0781f87ffba8e57bb81d2a677bd7f4f5e", "generation_seconds_batch": 16.329503054963425, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2523231506347656, "target_z": 0.5507813509553772, "ce": 6.356205940246582, "genuineness": 4.570107460021973, "identity": 0.43330642580986023, "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_CONV_reg/336341e069f1a22161b983ab80995bded285f2fb989b0c7fe6652e7e34c32ddb.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": "336582d51e82b4874e287bd46a59f8bd6bac6e680bbe8a21ee59b026a4366f45", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "62f1bc3fe3b762cac3a94d4cd7b7a218a9adfc5a06cd854666b0a2e02a9e569f", "generation_seconds_batch": 17.464369576075114, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.670443058013916, "target_z": 0.8710937360874215, "ce": 5.998785495758057, "genuineness": 4.607763767242432, "identity": 0.7245506644248962, "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_CONV_reg/336582d51e82b4874e287bd46a59f8bd6bac6e680bbe8a21ee59b026a4366f45.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": "3382466e5fdaf9e15d1a5539105f76db13c86891058a5479f0d76825cbb060f8", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ae745adac151b4486cd26cca00321d52bf50cf34cbb025d460d71831e611ecff", "generation_seconds_batch": 17.235907964990474, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.226236820220947, "target_z": 1.2968750071346555, "ce": 5.974957466125488, "genuineness": 4.419482707977295, "identity": 0.6113578677177429, "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_CONV_reg/3382466e5fdaf9e15d1a5539105f76db13c86891058a5479f0d76825cbb060f8.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "0be8de9d35b56444743ac19562074aa5c2d971651b1b30c0088fd80443f9120e", "generation_seconds_batch": 18.383798646973446, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.328808307647705, "target_z": 0.6093749329342371, "ce": 6.1894097328186035, "genuineness": 4.519899368286133, "identity": 0.7246178388595581, "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_CONV_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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "3afc1749c3e8c5ce55245b7f73a8044b0993550cee1b5f8869ed8aaa790f9993", "generation_seconds_batch": 16.748898196965456, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0075697898864746, "target_z": 0.36328123109316257, "ce": 6.475345611572266, "genuineness": 4.13078498840332, "identity": 0.5882409811019897, "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_CONV_reg/347b084eb1c87968a3fc14b855043a2de140e81cf0874b7d65525285bfb31414.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "349caa8a8b839b6d8fd6f3d27c38923cfc6cb7c0ba5484a027234c641cdb596f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "f4849e08fd37d4c4a884b60d9e1c86b0a3527b9c9cee8469870135069e62b18e", "generation_seconds_batch": 24.07227897492703, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.675542116165161, "target_z": 0.875000011415449, "ce": 6.157639026641846, "genuineness": 4.65797233581543, "identity": 0.6202733516693115, "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": 0.03999999999999915, "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_CONV_reg/349caa8a8b839b6d8fd6f3d27c38923cfc6cb7c0ba5484a027234c641cdb596f.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": "34ed38101e56c054a3391d07b0fde8f0fe47fcc03b34f37dbf01f06a49bfaa77", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8bbb81ac56f5f93c9811748f977a99f8c51b3796da94d1af7510a9c92f491926", "generation_seconds_batch": 21.424831039039418, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2064318656921387, "target_z": 0.5156250556503139, "ce": 6.006728172302246, "genuineness": 4.306513786315918, "identity": 0.47288379073143005, "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_CONV_reg/34ed38101e56c054a3391d07b0fde8f0fe47fcc03b34f37dbf01f06a49bfaa77.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "34f89c0660cda5595c105f443c5b514efacca24172792143cdea42faa548d130", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "985719339c5f4494b97986f95facb9d772c6ed88d900be6d93febfc6248fabb2", "generation_seconds_batch": 24.572221812908538, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3179266452789307, "target_z": -0.1650390823878526, "ce": 6.38003396987915, "genuineness": 4.18099308013916, "identity": 0.5772494077682495, "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_CONV_reg/34f89c0660cda5595c105f443c5b514efacca24172792143cdea42faa548d130.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "34fd63b5c7b55eb9733f978e1ad95c361ee7d85477cde5f322de66c6aef59510", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "61738319cac01c83cd1ba96cee67b62b78191bfccc5f163ab092ed9c2de64465", "generation_seconds_batch": 25.454617141978815, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.473446846008301, "target_z": -0.045898506795342825, "ce": 6.165581703186035, "genuineness": 4.444586753845215, "identity": 0.6240602731704712, "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_CONV_reg/34fd63b5c7b55eb9733f978e1ad95c361ee7d85477cde5f322de66c6aef59510.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "35f65a79a11d602f5afde9b9cf9015a3801cd02fe6e7edc900dbd75567f538b6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f57a542c6ae89164ed279610f59168059013009fd4661f4c94900b634651123b", "generation_seconds_batch": 23.79511560301762, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.415491819381714, "target_z": 0.6757812482163361, "ce": 6.284721851348877, "genuineness": 4.293961524963379, "identity": 0.34147411584854126, "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_CONV_reg/35f65a79a11d602f5afde9b9cf9015a3801cd02fe6e7edc900dbd75567f538b6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "360021633c7c334f33349e32ed8bc2f2d26e1cfbe849ce4987dfb951dac13095", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f4b11697d8c73d38baf0119183b5fe234cf2587b9f96c277c190fc76ecf3e3c0", "generation_seconds_batch": 23.935674431966618, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5626797676086426, "target_z": 0.022460854827177863, "ce": 6.443574905395508, "genuineness": 4.268857479095459, "identity": 0.6252460479736328, "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_CONV_reg/360021633c7c334f33349e32ed8bc2f2d26e1cfbe849ce4987dfb951dac13095.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "36ae1623858bd951f5cf98a08b258f224dfa214c17fa54a88a1d9cf7568aaefd", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "d36079239a5fa060b0a291b6e405de6914f60ec3da37e2f62283bff4d7e96c8f", "generation_seconds_batch": 23.595628913957626, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.777522563934326, "target_z": 0.9531249700344463, "ce": 6.09409761428833, "genuineness": 4.532451152801514, "identity": 0.01320798322558403, "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": 0.03999999999999915, "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_CONV_reg/36ae1623858bd951f5cf98a08b258f224dfa214c17fa54a88a1d9cf7568aaefd.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": "36c6337576106fd5c7e9bf969f472ee699ccf0acfea868e9bd352baaf9ec9919", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The road turns left behind the bridge.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "53e514485b6ca3ad1b27e98eaae76862a52d6abc186c641768b7d72fa81c2ad0", "generation_seconds_batch": 17.555115088936873, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4409868717193604, "target_z": 0.6953124422092893, "ce": 6.419747352600098, "genuineness": 4.425758361816406, "identity": 0.39489102363586426, "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_CONV_reg/36c6337576106fd5c7e9bf969f472ee699ccf0acfea868e9bd352baaf9ec9919.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "37083c3261ce94cf6aa270a3265d1bb5cf78dfb36469ec4f6cd23503ecc790d8", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "20f2a8f1cfff5366bd09c662a17962db52542a839cb02e039afd2f2427f113f2", "generation_seconds_batch": 23.520976540981792, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.665343999862671, "target_z": 0.867187460759394, "ce": 5.974957466125488, "genuineness": 4.6705241203308105, "identity": 0.7098636627197266, "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_CONV_reg/37083c3261ce94cf6aa270a3265d1bb5cf78dfb36469ec4f6cd23503ecc790d8.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": "371240a78d8977cc2c4c3206c9214e85c68398e16d139e1a055227108b38af18", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "bb813f58d01eb343f739d016ed19d105c2e8c743d8c28251d41328738e711fac", "generation_seconds_batch": 15.133737206924707, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.461383104324341, "target_z": 0.7109375435213994, "ce": 6.006728172302246, "genuineness": 4.65797233581543, "identity": 0.7901447415351868, "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_CONV_reg/371240a78d8977cc2c4c3206c9214e85c68398e16d139e1a055227108b38af18.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "37194dbe183d136fe5434c77cff0fd00111d5af2abd497b5d548cd3b9958f54a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "33b3ab9e4b41d3f4d6bafb69e59ea3faefe55f17ed6b68152bf005d706d07e46", "generation_seconds_batch": 22.84915072203148, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1560325622558594, "target_z": -0.28906254780219276, "ce": 6.300607204437256, "genuineness": 4.268857479095459, "identity": 0.4754220247268677, "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_CONV_reg/37194dbe183d136fe5434c77cff0fd00111d5af2abd497b5d548cd3b9958f54a.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": "377517bcf198aab0fb2d32920df97e8dc73543228c29a3d2bee61fe0331635a8", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "62c63453230c7bbfb01447c69260fa1878a358259e619e1cadb6c4cfd2057a16", "generation_seconds_batch": 15.34686818299815, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4409868717193604, "target_z": 0.6953124422092893, "ce": 6.252951145172119, "genuineness": 4.494794845581055, "identity": 0.7109746336936951, "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": 0.120000000000001, "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_CONV_reg/377517bcf198aab0fb2d32920df97e8dc73543228c29a3d2bee61fe0331635a8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "379951c5a0e40759925708f72426385b2e86f12ba56360d6b765d6f17983d669", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "56033b11eef55d751ae9618453b63487cca2cb0e5cefe012fc7e2262480889ab", "generation_seconds_batch": 22.3940667869756, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1095502376556396, "target_z": 0.4414061897121599, "ce": 6.229123115539551, "genuineness": 3.9801597595214844, "identity": 0.2971372604370117, "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_CONV_reg/379951c5a0e40759925708f72426385b2e86f12ba56360d6b765d6f17983d669.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "38512eb5d77e30f38c7192f5faa4bc112185f88aeac3bba5f73df04db0fae84a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "35eeb0eaba87d2326a3283378ad4d616bca1ff4b05012535d3871477c863141c", "generation_seconds_batch": 24.707960744970478, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2472238540649414, "target_z": 0.5468748929801656, "ce": 5.982900142669678, "genuineness": 4.607763767242432, "identity": 0.7879606485366821, "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_CONV_reg/38512eb5d77e30f38c7192f5faa4bc112185f88aeac3bba5f73df04db0fae84a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "38d8f7e48fe9881f7d81743ebdde871bd7b5770fe1bedcec2a32c3e83829e927", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "809fb71a15ca79fd67ea3c3355a304236e36e3f1ad89fb719b667cdc3c2c1196", "generation_seconds_batch": 17.40324608690571, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.971876621246338, "target_z": 0.3359374864441543, "ce": 6.197352409362793, "genuineness": 4.557555198669434, "identity": 0.6661766767501831, "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_CONV_reg/38d8f7e48fe9881f7d81743ebdde871bd7b5770fe1bedcec2a32c3e83829e927.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "393983991bafa615cd9a1c239637051cbd3aeaad25d37f06a06201856434b505", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6af437a97814572fc2600b5ac75a6c0693c61402d12f845ef5ccecb2692ad491", "generation_seconds_batch": 25.685742989066057, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.624551773071289, "target_z": 0.8359374407823582, "ce": 5.8875885009765625, "genuineness": 4.444586753845215, "identity": 0.16747257113456726, "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_CONV_reg/393983991bafa615cd9a1c239637051cbd3aeaad25d37f06a06201856434b505.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "39406fcdff1e1473d673fd960cb9a60b322861b98c4112d8a7b5e97112f0ebee", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2d805619be4fb7e563c2140a003c3e1cf5777d613e371846b86a662eeeaf207b", "generation_seconds_batch": 22.79114615393337, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.481095552444458, "target_z": -0.0400390024797095, "ce": 6.054384231567383, "genuineness": 4.293961524963379, "identity": 0.27690964937210083, "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_CONV_reg/39406fcdff1e1473d673fd960cb9a60b322861b98c4112d8a7b5e97112f0ebee.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "395a8842f26332fe07fa5e658adae9a6d7c0f6c3ba13ebf8080ffe9e88327c4f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9c5cab866fb0c8c34d9319122bed3194c8a6bc78ac6c4f376d09be5f95845127", "generation_seconds_batch": 22.218144059064798, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8494999408721924, "target_z": 0.24218742651304698, "ce": 6.213237762451172, "genuineness": 4.143336772918701, "identity": 0.4612548351287842, "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_CONV_reg/395a8842f26332fe07fa5e658adae9a6d7c0f6c3ba13ebf8080ffe9e88327c4f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "3990a67eea2727a156684c5d5149e33d61d3ae575f21b63ed0cd0f94c9ff1639", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e45895af7cab5b0710421f68df01cc916326e322d6b7baa244be150251ee924e", "generation_seconds_batch": 25.0208580349572, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.588175058364868, "target_z": 0.04199223146731533, "ce": 6.1020402908325195, "genuineness": 4.369274139404297, "identity": 0.6572389602661133, "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_CONV_reg/3990a67eea2727a156684c5d5149e33d61d3ae575f21b63ed0cd0f94c9ff1639.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "39c91b75d3a949b462936b805a5662ecffcf225818a10881b370c8396f581dcc", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7a8e73ea248d2395c45273fc599684481fd00ebf11d13956b834402231296ae2", "generation_seconds_batch": 24.986891867010854, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3485207557678223, "target_z": -0.1416016130668718, "ce": 6.443574905395508, "genuineness": 4.262581825256348, "identity": 0.49011659622192383, "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_CONV_reg/39c91b75d3a949b462936b805a5662ecffcf225818a10881b370c8396f581dcc.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": "39e8081943aba0c471bcc79698e62a232952a15f7ca35a2ed2aa54972bfca730", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "fa5f9133e035b7bc911fdd931ad117611e795e5025c1486135bc8192d3fd7aa3", "generation_seconds_batch": 15.426218340988271, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.481779098510742, "target_z": 0.7265624621863251, "ce": 6.062326908111572, "genuineness": 4.432034492492676, "identity": 0.4591771066188812, "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_CONV_reg/39e8081943aba0c471bcc79698e62a232952a15f7ca35a2ed2aa54972bfca730.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "3a5f8f27fe598cccc184b39d6fe62351c9d79dfa779ce5ff458b45d2d9a0de75", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "0ede0475e520f2fa2daf7426e2caf48d0ddc086f04291ee05ca12de146098dda", "generation_seconds_batch": 16.11406943597831, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1044511795043945, "target_z": 0.43749991438413244, "ce": 6.157639026641846, "genuineness": 4.394378662109375, "identity": 0.644694983959198, "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": 0.03999999999999915, "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_CONV_reg/3a5f8f27fe598cccc184b39d6fe62351c9d79dfa779ce5ff458b45d2d9a0de75.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": "3af7458938402d97e4ab63a6431df42276367dd9624037c4facacc12f46b8292", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I wrote the number on the envelope.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9fb6e2df7e479d11c58243783d85d52a699508afc9d5920add2cce5301a7c968", "generation_seconds_batch": 17.457468157052062, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.808116912841797, "target_z": 0.9765626220026113, "ce": 6.268836498260498, "genuineness": 4.337893962860107, "identity": 0.33172571659088135, "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_CONV_reg/3af7458938402d97e4ab63a6431df42276367dd9624037c4facacc12f46b8292.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "3b14fdb11c0a3bfb6347a87b55f811b095fd2854245bb24e3977a68f623b3ded", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "257990500574250afd50446f395fe619d641a64b4f5ba7d6c2c73d95b127802b", "generation_seconds_batch": 23.387461010948755, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8489089012145996, "target_z": 1.0078124593324629, "ce": 6.197352409362793, "genuineness": 4.381826400756836, "identity": 0.6276332139968872, "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_CONV_reg/3b14fdb11c0a3bfb6347a87b55f811b095fd2854245bb24e3977a68f623b3ded.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": "3b8816e9e7b5ef01fb27f3d00cd5bb4cbdeb1b9f83ea276da31813749179616a", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b10e8a266e7307b53d7b2454a5d2f054a26e7e1d5195229af8f087c26f49da3b", "generation_seconds_batch": 15.125952775939368, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0993523597717285, "target_z": 0.4335938217032891, "ce": 6.149696350097656, "genuineness": 4.632867813110352, "identity": 0.6893109083175659, "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_CONV_reg/3b8816e9e7b5ef01fb27f3d00cd5bb4cbdeb1b9f83ea276da31813749179616a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "3b99d0a16cb3e48cd3eb1095c8c0aa96db82adcf35c9dfa703e390138aecfcdf", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "adb4cfa6c340cabcb25251bf62028630fe33187ef07817442a30205ced02d696", "generation_seconds_batch": 17.061552187078632, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.890292167663574, "target_z": 0.2734374464900828, "ce": 6.356205940246582, "genuineness": 4.406930446624756, "identity": 0.6199033856391907, "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_CONV_reg/3b99d0a16cb3e48cd3eb1095c8c0aa96db82adcf35c9dfa703e390138aecfcdf.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "3bf0aca9d1f72e33f4105cd4dc5539c3203683a064589c9c710339a1b463c930", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "abae3eb3f1a558dc154874294ca1a3c21b8ef8072e1a6981979bff74706cd09a", "generation_seconds_batch": 22.68453354993835, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.481779098510742, "target_z": 0.7265624621863251, "ce": 5.998785495758057, "genuineness": 4.6705241203308105, "identity": 0.6907225251197815, "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_CONV_reg/3bf0aca9d1f72e33f4105cd4dc5539c3203683a064589c9c710339a1b463c930.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "bfdf43f64570ef183628339d996da84b5082a2f14cf0642ce58b16af39e2efcd", "generation_seconds_batch": 18.527643244946375, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3548946380615234, "target_z": -0.1367187232450414, "ce": 6.125868320465088, "genuineness": 4.15588903427124, "identity": 0.6216987371444702, "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_CONV_reg/3c590f9302048ad206a88d3c2440f1028cc9eea1f1662250baf365be21a6ebfa.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "3c5ef55149385854c9a2ba78faa21bd1997161e68c48d9b335e045f541054cdf", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cf15cd008266b40d9332510372d20a2cd4b6a56bd23b1018a29bb814999084c8", "generation_seconds_batch": 23.119907727930695, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.379798650741577, "target_z": 0.6484375035673278, "ce": 5.792276382446289, "genuineness": 4.645420074462891, "identity": 0.6147006154060364, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaîne. 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_CONV_reg/3c5ef55149385854c9a2ba78faa21bd1997161e68c48d9b335e045f541054cdf.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": "3c680ace032a7795786bf790ff9fc59973c426ac5fbcfce9c9706cba1de77b4a", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "370c7dd9d4a5df9f19104eab0a45f784bee6b3bf15873b777c48f04011d0fadd", "generation_seconds_batch": 15.029004493029788, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8998992443084717, "target_z": 1.0468750299655536, "ce": 6.149696350097656, "genuineness": 4.582659721374512, "identity": 0.3712776303291321, "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": 0.03999999999999915, "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_CONV_reg/3c680ace032a7795786bf790ff9fc59973c426ac5fbcfce9c9706cba1de77b4a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "3d2568aed3d6a6865f419522bbd054b4b402e73ab3c300948665764f51c8d44b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3f1958b9a3efa59ef808f394fb00c286806cb2bf34f29721c016748eb1e0fcc1", "generation_seconds_batch": 15.517977745039389, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6041557788848877, "target_z": 0.8203125221174324, "ce": 5.8875885009765625, "genuineness": 4.68307638168335, "identity": 0.708986222743988, "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_CONV_reg/3d2568aed3d6a6865f419522bbd054b4b402e73ab3c300948665764f51c8d44b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "3d7543b60534b13076ff3f90ab3e16c83f4e6d8a9414f8ae5a877b02c3e639a4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8e6a99d22f8933b87a604645f592f1c4743eab06a82ee728ddb9fcb2c68805ff", "generation_seconds_batch": 25.053604001062922, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3950958251953125, "target_z": 0.6601563295514103, "ce": 6.078212261199951, "genuineness": 4.306513786315918, "identity": 0.6898417472839355, "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_CONV_reg/3d7543b60534b13076ff3f90ab3e16c83f4e6d8a9414f8ae5a877b02c3e639a4.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": "3e48451c35b9f510ed1e7e327ed3f7628fe5a4a78644b58f32144ae3fcd5c375", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The road turns left behind the bridge.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e623ceb9f546a1f7be6b0bb56b188ff13fbcabe9cbe585b734649b04f06db418", "generation_seconds_batch": 21.548421447048895, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6449480056762695, "target_z": 0.8515625420944682, "ce": 6.4515180587768555, "genuineness": 4.381826400756836, "identity": 0.34635695815086365, "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_CONV_reg/3e48451c35b9f510ed1e7e327ed3f7628fe5a4a78644b58f32144ae3fcd5c375.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": "3fb893b94ea540d6f2f18a3b730d099c6e7cf8de5e68ac5a6fb6ed4610cb456c", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "65ccc9be9e79d81d4d7433e495e337d5a31c139f9357ee6f25e2b626b15d2100", "generation_seconds_batch": 15.111979463952594, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4976673126220703, "target_z": -0.027343744649008277, "ce": 5.982900142669678, "genuineness": 4.582659721374512, "identity": 0.3579709529876709, "hypothesis": "De sleutels liggen op de keukentafel. Keek in de laag 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_CONV_reg/3fb893b94ea540d6f2f18a3b730d099c6e7cf8de5e68ac5a6fb6ed4610cb456c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "4010aab6b5473f7cac42b6862eead7dbb7ff88e5b0f5a22bc9e92e2c9e1acdd8", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "ed501dc9bfc55a1594f1995485872699cbdcc7a42bc361b071835dd1116e74fd", "generation_seconds_batch": 24.06128426408395, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.711235284805298, "target_z": 0.9023437560644573, "ce": 6.133810997009277, "genuineness": 4.419482707977295, "identity": 0.40050598978996277, "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_CONV_reg/4010aab6b5473f7cac42b6862eead7dbb7ff88e5b0f5a22bc9e92e2c9e1acdd8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "4045d16bd5d34303a5caecd058b983efa13376797567868fe74d31063c52754f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d09059b037bc43aaf0f032d9c4a15586eddb8c82078ef22d37332634a083fe0b", "generation_seconds_batch": 22.16825778305065, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6551461219787598, "target_z": 0.8593750927505233, "ce": 6.0146708488464355, "genuineness": 4.532451152801514, "identity": 0.4978952407836914, "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_CONV_reg/4045d16bd5d34303a5caecd058b983efa13376797567868fe74d31063c52754f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "404c23357e1e61ba2e684f3ff266e4f909c4f8b495563245d3389de36a7ad31e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d799459e8a050b679597216f2ff53962f711c0eaa32b452818906eefc7002ddd", "generation_seconds_batch": 20.255277213989757, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.711235284805298, "target_z": 0.9023437560644573, "ce": 6.157639026641846, "genuineness": 4.7081804275512695, "identity": 0.5278104543685913, "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_CONV_reg/404c23357e1e61ba2e684f3ff266e4f909c4f8b495563245d3389de36a7ad31e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "405374bcde3185845d1f1db3b628c6935e04dc149980737d21769d045747d8eb", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "779f74563a3a80662c6d6e55ae060feca69b2083442da217004e276001bd83be", "generation_seconds_batch": 24.521114674047567, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.776839017868042, "target_z": 0.1865235053684117, "ce": 6.403861999511719, "genuineness": 4.394378662109375, "identity": 0.5104648470878601, "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_CONV_reg/405374bcde3185845d1f1db3b628c6935e04dc149980737d21769d045747d8eb.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "40630cd325720887ffdc72a857c9943254283dff3756ca8db4fbe23a0eccbfb1", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8ac8c8337aa4697d85bcacada852b9a62a0b60fe1f4a9aa294316e1d6914db25", "generation_seconds_batch": 26.323338488931768, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.318610429763794, "target_z": 0.6015625649253663, "ce": 6.1894097328186035, "genuineness": 4.607763767242432, "identity": 0.4933808445930481, "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_CONV_reg/40630cd325720887ffdc72a857c9943254283dff3756ca8db4fbe23a0eccbfb1.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "40e8e83d4c22fac1cb77030e435d4cc1bb24835d30d117ec89c324690339ef3d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0d4cb0d361ed9089adafdca57475d66a9330f8408a7db3088e6ad31a885bdc32", "generation_seconds_batch": 21.258632850949652, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7653660774230957, "target_z": 0.1777343402185538, "ce": 6.125868320465088, "genuineness": 4.337893962860107, "identity": 0.6257667541503906, "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_CONV_reg/40e8e83d4c22fac1cb77030e435d4cc1bb24835d30d117ec89c324690339ef3d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "41519baf17efde820cc019c1343c8f2fc75d78a0409b825b4d36e6dd504a5d77", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7b7e42117ea6cb53a7070da11d790da4ba3424932b283d0ddd4e81d921be3f33", "generation_seconds_batch": 23.932636426063254, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9540300369262695, "target_z": 0.32226561411965016, "ce": 5.855817794799805, "genuineness": 4.118232727050781, "identity": 0.6500214338302612, "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_CONV_reg/41519baf17efde820cc019c1343c8f2fc75d78a0409b825b4d36e6dd504a5d77.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "42531e5ad83ae6d62bd84e92bf4da52e5d2119328cfd700ae922a0efa39a64bd", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "87be5faddaee6c14d26a8ba71330c47b73ab815b0e0fb237db2b42dde98b382d", "generation_seconds_batch": 24.980389069998637, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6449480056762695, "target_z": 0.8515625420944682, "ce": 6.332377910614014, "genuineness": 4.545003414154053, "identity": 0.35981398820877075, "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_CONV_reg/42531e5ad83ae6d62bd84e92bf4da52e5d2119328cfd700ae922a0efa39a64bd.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": "42ee610bc4c8697c69ba2016bd6b5faf847f763615effcc18876ace69ca2b19c", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "53cd53a6213ebcb9350476b2566436a98bf9d287059f17dcde4ae1f8aba949bd", "generation_seconds_batch": 21.359757354017347, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8998992443084717, "target_z": 1.0468750299655536, "ce": 6.213237762451172, "genuineness": 4.256305694580078, "identity": 0.6124602556228638, "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_CONV_reg/42ee610bc4c8697c69ba2016bd6b5faf847f763615effcc18876ace69ca2b19c.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "2a232498a93bb56affc7180506ef950a110dd1fe8b1a49193c6156ee92e40765", "generation_seconds_batch": 18.383798646973446, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8456757068634033, "target_z": 0.23925776567882243, "ce": 6.09409761428833, "genuineness": 4.074300289154053, "identity": 0.7055721879005432, "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_CONV_reg/434f953004ee3838b31e4f505ba0cb9202b34ab37936979fff50edb64a6a996f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "4356d41d9695638f3021cf85f2f2de8f64f3959a3cd355a59841110d491f8427", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "adbecba8686a1476fab153350198ffc3ada3266938044793c56e63f603a964e0", "generation_seconds_batch": 22.31962956895586, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5142390727996826, "target_z": -0.014648486818307056, "ce": 6.38797664642334, "genuineness": 4.268857479095459, "identity": 0.7113256454467773, "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_CONV_reg/4356d41d9695638f3021cf85f2f2de8f64f3959a3cd355a59841110d491f8427.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "4358ac1da1a8152713498b418058c7463c56c584b00861cbaea999159962c398", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f61f771f8b2052b447e36aacb6c7815c5e54cc4b18312f5aba0f748c25b1adf7", "generation_seconds_batch": 24.79566278902348, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.821455478668213, "target_z": 0.22070318617967205, "ce": 6.157639026641846, "genuineness": 4.469690799713135, "identity": 0.7544587850570679, "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_CONV_reg/4358ac1da1a8152713498b418058c7463c56c584b00861cbaea999159962c398.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": "437aac43ae919f5cc310d69ee397df204a693c8f15518f34d3080245b97558c3", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "209a4c5b4dafb37ef484385d4897a663b72ff8ba2c090198dbc1477119d26831", "generation_seconds_batch": 17.4617559799226, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.124256134033203, "target_z": 1.2187498658684741, "ce": 6.086154937744141, "genuineness": 4.65797233581543, "identity": 0.4453507959842682, "hypothesis": "Die bericht kwam nadat ik vertrokken was. Ik las het 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_CONV_reg/437aac43ae919f5cc310d69ee397df204a693c8f15518f34d3080245b97558c3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "43896f2d803c29c8d04f198e822b502f3e3685c01efe47e0d0992575fb4f4a41", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "27f9a8e7246ce983974121d52b5efc5e7189bfa16edabd185b70daf9d3c989e7", "generation_seconds_batch": 21.928342066938058, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.124256134033203, "target_z": 1.2187498658684741, "ce": 6.173524379730225, "genuineness": 4.6203155517578125, "identity": 0.7297868728637695, "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_CONV_reg/43896f2d803c29c8d04f198e822b502f3e3685c01efe47e0d0992575fb4f4a41.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "43c1be6a9fda92ac2b88284c784e209540929227d178ed6783bfa10d5e9d7cab", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "87b8e03204bc72608e4139ac6a41719e1f9ee346b55d54be3fe5635897cc25db", "generation_seconds_batch": 19.53781936597079, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5429673194885254, "target_z": 0.7734374008282867, "ce": 6.260893821716309, "genuineness": 4.68307638168335, "identity": 0.6895621418952942, "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_CONV_reg/43c1be6a9fda92ac2b88284c784e209540929227d178ed6783bfa10d5e9d7cab.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": "43d05aa28786098f3fc1891c0cb611f469f1e3f5445cc65a72e876fbc56b7014", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a0925a76955b5a4021ed61f78ff1433800cbc41c4f16ecf119cfbe298026c439", "generation_seconds_batch": 15.254141589975916, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.226236820220947, "target_z": 1.2968750071346555, "ce": 6.348263263702393, "genuineness": 4.645420074462891, "identity": 0.6343686580657959, "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_CONV_reg/43d05aa28786098f3fc1891c0cb611f469f1e3f5445cc65a72e876fbc56b7014.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "3615f21f07a1f07aa7671b0b1b53cbb02998b9896c167b6f320300a62429f805", "generation_seconds_batch": 17.593891810043715, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6347498893737793, "target_z": 0.8437499914384132, "ce": 6.221180438995361, "genuineness": 4.6203155517578125, "identity": 0.5789223909378052, "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_CONV_reg/43d771a5c055e2aea1186dc99f4cdef2b4c6bc8c6b2779ad35a29ca11c75a91c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "4471635bc1f1c70e35ab6c929eaab70bc0d294dd45a04fe1bbccc739641220be", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "f969f559641d8f90a0fae5b368e56cdce9f50f309b27b060c19934d861009727", "generation_seconds_batch": 22.72915472392924, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7143759727478027, "target_z": 0.13867195223264722, "ce": 6.300607204437256, "genuineness": 4.149612903594971, "identity": 0.4286254644393921, "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_CONV_reg/4471635bc1f1c70e35ab6c929eaab70bc0d294dd45a04fe1bbccc739641220be.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "44fef6ffd030f0e16320f948bfa36dfa1f24e971ed0021b21ec668c2728394cd", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "09319d3efe5fe35f94078ca4437d8bd88818ba102daa8eaf178aa00308bd48be", "generation_seconds_batch": 17.20448352897074, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9565796852111816, "target_z": 0.324218843107256, "ce": 6.125868320465088, "genuineness": 4.482243061065674, "identity": 0.582045316696167, "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_CONV_reg/44fef6ffd030f0e16320f948bfa36dfa1f24e971ed0021b21ec668c2728394cd.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": "455aa291a9bd7eac4562d9917916bfcab4b20d4106d44e9010301b16b214e938", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7503e784b79ad828cc6eac94abc27a94113578c7a0741d84cf66b6b84518423c", "generation_seconds_batch": 21.832873945008032, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7959601879119873, "target_z": 0.20117180953953456, "ce": 5.824047088623047, "genuineness": 4.65797233581543, "identity": 0.691159725189209, "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_CONV_reg/455aa291a9bd7eac4562d9917916bfcab4b20d4106d44e9010301b16b214e938.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "456b0fd3c63bf257079f2cca8a7dde05d11af24e29cc5315f96c2b2573b310a3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "941a7dc30dc689e18ccadf61f49fdb2888079e57f942477cbeed298c98706d97", "generation_seconds_batch": 23.185450160992332, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7194747924804688, "target_z": 0.14257804491349052, "ce": 5.903473377227783, "genuineness": 4.099404335021973, "identity": 0.7025663256645203, "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_CONV_reg/456b0fd3c63bf257079f2cca8a7dde05d11af24e29cc5315f96c2b2573b310a3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "457955c98bc68bb28f1f7755ce74ac427a0dfd0522c5decbaf2f7a592b5cfe38", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3e4159d5552001686232276bf35207ea35efc773e3a0bdf4d4d8a5cdb7ac476f", "generation_seconds_batch": 25.953460013959557, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6544623374938965, "target_z": 0.0927734454373044, "ce": 6.260893821716309, "genuineness": 4.607763767242432, "identity": 0.6896160840988159, "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_CONV_reg/457955c98bc68bb28f1f7755ce74ac427a0dfd0522c5decbaf2f7a592b5cfe38.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "45c23269368abc18b638e9a524d288eb4bdef21208dda14e0ca13c256eb74c49", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "bd9075adf8bb1a5c2ca2ddadc2dd11aea163d900db50d72b15955022e5f7871f", "generation_seconds_batch": 23.234237976022996, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9406914710998535, "target_z": 1.0781250499425894, "ce": 6.0305562019348145, "genuineness": 3.9801597595214844, "identity": 0.531145453453064, "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_CONV_reg/45c23269368abc18b638e9a524d288eb4bdef21208dda14e0ca13c256eb74c49.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "45dec36a089a1e7283e75fb5636b3d8f736d414ce528863df1ae2a85c06746fa", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a2e890852be681fb1a337526aea87d706dfedab1bd5cdfe87a60534bbe411f66", "generation_seconds_batch": 23.972920401021838, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8647971153259277, "target_z": 0.25390625249712945, "ce": 6.125868320465088, "genuineness": 4.381826400756836, "identity": 0.7864410877227783, "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_CONV_reg/45dec36a089a1e7283e75fb5636b3d8f736d414ce528863df1ae2a85c06746fa.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": "45f1b0875ba9b1770f2dde195ef3372f481136317b3858dd95017c84863a8abf", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The keys are on the kitchen table.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6849f014521b73a86aea11f0c00d7dec81525ce1fb480d02ff3871ac7bcf79a3", "generation_seconds_batch": 15.288394376984797, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4511849880218506, "target_z": 0.7031249928653444, "ce": 6.173524379730225, "genuineness": 4.457139015197754, "identity": 0.5015850067138672, "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_CONV_reg/45f1b0875ba9b1770f2dde195ef3372f481136317b3858dd95017c84863a8abf.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "7d4d7c948f8693c98dc9b95d0d747049b2df0838ed444c14edfa8ef4d540e6fd", "generation_seconds_batch": 17.508170478977263, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.178978204727173, "target_z": -0.2714844001496611, "ce": 6.229123115539551, "genuineness": 4.149612903594971, "identity": 0.4881887137889862, "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_CONV_reg/45f9ec8497c8691f9e18d794cf98f47c522bfab041b36df5accabf5c5c456412.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "45ff116fcb80f5b8128a54d846ea76fccec815e1007135e0f00db453d3a60b75", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "d33550b1bcb7c6f3f37be2e185195fb9231473140ba5142a6b2dfd5482a80d4b", "generation_seconds_batch": 24.01211730495561, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.767324447631836, "target_z": 0.9453124193783914, "ce": 6.165581703186035, "genuineness": 4.770940780639648, "identity": 0.6252008676528931, "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_CONV_reg/45ff116fcb80f5b8128a54d846ea76fccec815e1007135e0f00db453d3a60b75.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "46456452dc1453fbd7104f17dd74889451bc6cb0b3e3f81055b9912b2979f08e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "77d170f17b9ba893fa9680edd85a5bbf8455a775943625d1239fd49538236c37", "generation_seconds_batch": 23.659395167953335, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7469284534454346, "target_z": 0.9296875007134655, "ce": 6.038498878479004, "genuineness": 4.300237655639648, "identity": 0.11956155300140381, "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_CONV_reg/46456452dc1453fbd7104f17dd74889451bc6cb0b3e3f81055b9912b2979f08e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "46a5940c3825c6b1a1adc77f9520ffcb8bdc00b572ad4452f9ad76f622454275", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c32b28147213d967d22a568006008a543a0f9e6213d32589ac00368ec886d82a", "generation_seconds_batch": 16.788822289090604, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.632791519165039, "target_z": 0.07617191227857567, "ce": 6.372091293334961, "genuineness": 4.256305694580078, "identity": 0.5951979160308838, "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_CONV_reg/46a5940c3825c6b1a1adc77f9520ffcb8bdc00b572ad4452f9ad76f622454275.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "46cd44794b297dd68f5b49fd2c532b70f87f50583555b2095186037a32f4c974", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8d5daa237182ee0a0161f7a1bef1ed9b4cc433c47d0b4fd39df955cdbe3618c3", "generation_seconds_batch": 22.79114615393337, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9081387519836426, "target_z": 0.2871093188145869, "ce": 6.23706579208374, "genuineness": 4.243753433227539, "identity": 0.4776083528995514, "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_CONV_reg/46cd44794b297dd68f5b49fd2c532b70f87f50583555b2095186037a32f4c974.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "46e5f181ef7a84c03a817541dd186e8e9cd25e50b35819ac382c4b6992a71ad7", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "803241d35b630d6f090df2ffcf4070e7ed9daddcb3b28c59beb294abd8d00ed7", "generation_seconds_batch": 24.59336571698077, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.716334342956543, "target_z": 0.9062500313924848, "ce": 5.863759994506836, "genuineness": 4.381826400756836, "identity": 0.5646299123764038, "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_CONV_reg/46e5f181ef7a84c03a817541dd186e8e9cd25e50b35819ac382c4b6992a71ad7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "46ff46088f6d7fdd917a7ce55f0e83ba7e8e5be9dc76213f254d1db2774a2db0", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4d50c001fa45f8f66dd14645420da04eecacc071f28cfb92cca0feedb91d90f7", "generation_seconds_batch": 18.151362388045527, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.297623157501221, "target_z": 1.3515624964326722, "ce": 5.990842819213867, "genuineness": 4.6705241203308105, "identity": 0.7849128246307373, "hypothesis": "He encontrado la cajita. Estaba junto a la estaba junto la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.17647058823529413, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/46ff46088f6d7fdd917a7ce55f0e83ba7e8e5be9dc76213f254d1db2774a2db0.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "47249d09162e2a3e456796e07aa1783ecccf604f4e3ff6184e7dfca6159519c7", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5362895f52acc3cdfa8225ecf8134ba3961eb5dfce9f24b43953a8ff1bac2aef", "generation_seconds_batch": 23.119907727930695, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5378684997558594, "target_z": 0.7695313081474434, "ce": 5.935244083404541, "genuineness": 4.758388519287109, "identity": 0.6984764933586121, "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_CONV_reg/47249d09162e2a3e456796e07aa1783ecccf604f4e3ff6184e7dfca6159519c7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "473cbe6f251af23a02909ca36c5446202910f27ee8ee827cef9b11408e389230", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "665d156abe7776e375dc826804a797c3e51b3780ef7044cd1e5e730f96811525", "generation_seconds_batch": 23.542454715934582, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4919772148132324, "target_z": 0.7343750128423802, "ce": 6.0146708488464355, "genuineness": 4.262581825256348, "identity": 0.7095566391944885, "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_CONV_reg/473cbe6f251af23a02909ca36c5446202910f27ee8ee827cef9b11408e389230.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": "4745ea8aaaa7e2a4826c8a732d962d898e4f9000d568992a5f3689b708eed0c3", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b399fec140052168962033f489e392a12e6cec1d7ae16731da5da67fc985b8df", "generation_seconds_batch": 19.099420040030964, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3899967670440674, "target_z": 0.6562500542233828, "ce": 5.804190158843994, "genuineness": 4.582659721374512, "identity": 0.7266949415206909, "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_CONV_reg/4745ea8aaaa7e2a4826c8a732d962d898e4f9000d568992a5f3689b708eed0c3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "474b5b13be4eae1c0772bffe971d2e8ee70b10eef242d7371c01e234afb9b305", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "5073eb3630f87c37e6dbf032b8a98c66cb27cbecc8896b78b23994bc6ea0ad47", "generation_seconds_batch": 24.314591440954246, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.808116912841797, "target_z": 0.9765626220026113, "ce": 5.903473377227783, "genuineness": 4.545003414154053, "identity": 0.2908477783203125, "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_CONV_reg/474b5b13be4eae1c0772bffe971d2e8ee70b10eef242d7371c01e234afb9b305.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "47696e66de39bee7e139f0ac10d41df9ae40dc0761cefb54890b0ac71ea1b96f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "17cf537dd1bf6a61a9f2f1df00be95a70ec60de8c1f3dd0d6664f422d580b27b", "generation_seconds_batch": 16.600682582939044, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.825279712677002, "target_z": 0.22363284701389663, "ce": 5.863759994506836, "genuineness": 4.3881025314331055, "identity": 0.6938639283180237, "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_CONV_reg/47696e66de39bee7e139f0ac10d41df9ae40dc0761cefb54890b0ac71ea1b96f.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "fc0ec8da3c4372658512448f2cc8f6e7a6db206f0a419de6129b332390d8c90e", "generation_seconds_batch": 16.90704535599798, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.739871025085449, "target_z": 0.1582031462256005, "ce": 6.157639026641846, "genuineness": 4.444586753845215, "identity": 0.6945229768753052, "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_CONV_reg/47d00b404f769e82194cfb39f2be7a6bd3cc5bfaac4f38356b0fada2abdb9f4d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "4807ab42b70b3bf352423fbd2289643111214d1484bf2cf50a128b6d29c1c0bc", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8915cf19c3398cca4493ab71c97c892d110a9fdf61651c5dd0b47d9ef88a2f84", "generation_seconds_batch": 24.741782058030367, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4103927612304688, "target_z": 0.6718749728883086, "ce": 6.205295085906982, "genuineness": 4.595211505889893, "identity": 0.5995923280715942, "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_CONV_reg/4807ab42b70b3bf352423fbd2289643111214d1484bf2cf50a128b6d29c1c0bc.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "480ddae1783124a0d5ade2e85fd80135ab2a755f23dd76580ddc4639a39af16c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7ec3f86569b49fdb13cd097a71cebe6f317cf76f8a0ee710182d94b6ae8f920f", "generation_seconds_batch": 24.59336571698077, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.205840587615967, "target_z": 1.2812499058225457, "ce": 5.871703147888184, "genuineness": 4.356721878051758, "identity": 0.6771162152290344, "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_CONV_reg/480ddae1783124a0d5ade2e85fd80135ab2a755f23dd76580ddc4639a39af16c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "486694b2f604eaef46be76e48d0701bf5a5b4f75af052d790cd969937bce2454", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9fc779331997e92714f43aafff0eb90d037a1e61853e5954a3a9ec3f04d81fd2", "generation_seconds_batch": 21.998323668027297, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8489089012145996, "target_z": 1.0078124593324629, "ce": 6.221180438995361, "genuineness": 4.469690799713135, "identity": 0.4098832309246063, "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_CONV_reg/486694b2f604eaef46be76e48d0701bf5a5b4f75af052d790cd969937bce2454.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "48a1cb44765dbd7cb8f22db72e7ca21b5adb793ee1a08a5322aa84ab273b62e5", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1bb08f68fb9ad25a647e85cc45217c527e4c1d65975f038cbe894195de5170d5", "generation_seconds_batch": 18.0278738590423, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8998992443084717, "target_z": 1.0468750299655536, "ce": 6.086154937744141, "genuineness": 4.582659721374512, "identity": 0.729117214679718, "hypothesis": "Ik heb het 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_CONV_reg/48a1cb44765dbd7cb8f22db72e7ca21b5adb793ee1a08a5322aa84ab273b62e5.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": "48c8955a3714db3a227d819eacb15b3d1beb6b9bb1142cfba96338c06d22d972", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2e8854e08c08ef65cb12fdb9b6b9c4f9224da9274bb9450b80eaccd5c5060650", "generation_seconds_batch": 15.343215593020432, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.512373447418213, "target_z": 0.7500001141544901, "ce": 6.046441555023193, "genuineness": 4.632867813110352, "identity": 0.6638468503952026, "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_CONV_reg/48c8955a3714db3a227d819eacb15b3d1beb6b9bb1142cfba96338c06d22d972.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "492f15058847195fb59207b28551e156d621f2b6904018f3d68042940b693aa4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a0cd3452b57c042d1e4875a0d9b1b42dbb1413ed4a078b93b9dfaf276bd2ec1d", "generation_seconds_batch": 24.33279059000779, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6959381103515625, "target_z": 0.8906249300803748, "ce": 6.173524379730225, "genuineness": 4.645420074462891, "identity": 0.7512992024421692, "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_CONV_reg/492f15058847195fb59207b28551e156d621f2b6904018f3d68042940b693aa4.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": "499e2aaa0927f3c90ffe4ca7cd4cbe7b4d4562c0e8f10417ea66d223987e6321", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "0b900fee8f089b3f4349d061654766ef95974227c695edc1b42b9c8ec331228c", "generation_seconds_batch": 18.394700975040905, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.716334342956543, "target_z": 0.9062500313924848, "ce": 5.935244083404541, "genuineness": 4.519899368286133, "identity": 0.24792931973934174, "hypothesis": "Ik schreef het nummer op de envelop, lees het goed, en controleer de laatste twee cijfers. Bel meer wanneer je even tijd hebt.", "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_CONV_reg/499e2aaa0927f3c90ffe4ca7cd4cbe7b4d4562c0e8f10417ea66d223987e6321.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "49b5496b30e6934468a90def51715e97bdde50f608ea50eed7e36297ec44fb9f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "602a45590686a95ef59b416f4c4443e43e5f689a8d8e311df9b1457c8870473f", "generation_seconds_batch": 23.815440216916613, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.670443058013916, "target_z": 0.8710937360874215, "ce": 6.022613525390625, "genuineness": 4.381826400756836, "identity": 0.7858824729919434, "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": 0.03999999999999915, "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_CONV_reg/49b5496b30e6934468a90def51715e97bdde50f608ea50eed7e36297ec44fb9f.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "5f0d2469c51d1d8bc532f01a1b058a3568fb5310c6e189d429bf20a1eb78988e", "generation_seconds_batch": 17.58974752004724, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.479820728302002, "target_z": -0.041015616973512414, "ce": 6.165581703186035, "genuineness": 4.1998209953308105, "identity": 0.6715476512908936, "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_CONV_reg/49d468ae6540cc7c1e5b0e6883c25e2b29599458c9de0a5398000eaca62677ab.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "49dea3c5422be87895763127f1430422626891c305c56837999f3c16cc57e750", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "434a96cb6be44878f9487d7a09506cbc0bf751ffdf130305def4c4cc6e8efcad", "generation_seconds_batch": 17.493057081010193, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.216629981994629, "target_z": 0.5234376063063689, "ce": 6.022613525390625, "genuineness": 4.796044826507568, "identity": 0.5340437293052673, "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_CONV_reg/49dea3c5422be87895763127f1430422626891c305c56837999f3c16cc57e750.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": "49e99884baefd34c1bf1c229fef681eaba2983a86b6ed23d99baab91e50493ca", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "26a2024626c32cd2bf0f4ca615ed7135f219858811bf44d38efeb5627b65b058", "generation_seconds_batch": 17.574974497081712, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5582644939422607, "target_z": 0.7851562268123692, "ce": 6.149696350097656, "genuineness": 4.6203155517578125, "identity": 0.7397096157073975, "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": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/49e99884baefd34c1bf1c229fef681eaba2983a86b6ed23d99baab91e50493ca.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "4a0a5b5a258384e14ec15374313fe802709f12c6fed68d9dd56c1063492eecae", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "84cc0bb6d8ec7eefff9c425e8055c262b1236a504e89765697e22860e7e7a968", "generation_seconds_batch": 18.730553666013293, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.563363552093506, "target_z": 0.7890625021403966, "ce": 6.054384231567383, "genuineness": 4.507347106933594, "identity": 0.6403580904006958, "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_CONV_reg/4a0a5b5a258384e14ec15374313fe802709f12c6fed68d9dd56c1063492eecae.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "4a2bfcdbcf23af8291ad43dab6ce6f5797c958ce7dd24bd0cb51a4b6ef1e488b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5385b34e24da3a1340b757c5daf24b990ff023fc992e8fb07424b0cfa225d663", "generation_seconds_batch": 22.07339332101401, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5531654357910156, "target_z": 0.7812499514843417, "ce": 6.268836498260498, "genuineness": 4.406930446624756, "identity": 0.34904754161834717, "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_CONV_reg/4a2bfcdbcf23af8291ad43dab6ce6f5797c958ce7dd24bd0cb51a4b6ef1e488b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "4a7b7218f489060a4a61b53e7ff4b4088bf143e60f01de9689c9d1dc41fbd28b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4cb2f7b87a182faa914135b77ddf28eca64230cb8a7077425aa8af6de84ca5b2", "generation_seconds_batch": 22.836119381943718, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6500468254089355, "target_z": 0.8554686347753115, "ce": 5.855817794799805, "genuineness": 4.293961524963379, "identity": 0.6224621534347534, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la vieille fenêtre.", "wer": 0.3684210526315789, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/4a7b7218f489060a4a61b53e7ff4b4088bf143e60f01de9689c9d1dc41fbd28b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "4aac13753cf7573fbd3bdbbc3b0822773b527cdea9f0b49976cfa247556c3260", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "aef9320b32a7f54039a08d15122fc4c432dd5cfdf26ebba7b5b2221c4a64447f", "generation_seconds_batch": 23.402635614969768, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3561694622039795, "target_z": -0.13574210875123846, "ce": 6.157639026641846, "genuineness": 4.218649387359619, "identity": 0.79597407579422, "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_CONV_reg/4aac13753cf7573fbd3bdbbc3b0822773b527cdea9f0b49976cfa247556c3260.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "4b636d8292de0115ca537968c9c7593c8d5d5e03a8a10b62a09e6f7283f20e02", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "29a82a912ac45fc21452e909e7132b8d3879346112f71075715c26cbcf85d4ff", "generation_seconds_batch": 22.819248728919774, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.563363552093506, "target_z": 0.7890625021403966, "ce": 6.109982967376709, "genuineness": 4.8588056564331055, "identity": 0.5978338122367859, "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": 0.03999999999999915, "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_CONV_reg/4b636d8292de0115ca537968c9c7593c8d5d5e03a8a10b62a09e6f7283f20e02.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "4bfb8d5eca4caca0eca1986c90973171f6c40791b510c34e4f8c8e782bf31b9e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "af89a84c5aaeaea5c9ec04ac5cdbf50c5e0e111a6a26fe690647c49cd2544a10", "generation_seconds_batch": 24.718449481064454, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.835477828979492, "target_z": 0.23144539766995162, "ce": 5.982900142669678, "genuineness": 4.243753433227539, "identity": 0.6495055556297302, "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_CONV_reg/4bfb8d5eca4caca0eca1986c90973171f6c40791b510c34e4f8c8e782bf31b9e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "4d07ea89e02dcdb5008bb94a82dfe34d0595395bf06a731beec6bf87f16ed4a1", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "b0d6ae7e3a9c47406bb38135c9498c03542bcd19b300901158f8744d418a44e3", "generation_seconds_batch": 24.423196024028584, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7979187965393066, "target_z": 0.9687500713465563, "ce": 5.903473377227783, "genuineness": 4.344170093536377, "identity": 0.35186266899108887, "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_CONV_reg/4d07ea89e02dcdb5008bb94a82dfe34d0595395bf06a731beec6bf87f16ed4a1.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "4d0849263b8f74f38e79211835fbe9848d9189d84578c66d34b897d7c7e63f9a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0c4143d01ed354a2c4e52466e58e9584c3ffa0fd3ae7db6fd748b8989743a99f", "generation_seconds_batch": 23.02964445995167, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.675542116165161, "target_z": 0.875000011415449, "ce": 6.086154937744141, "genuineness": 4.482243061065674, "identity": 0.4600657820701599, "hypothesis": "Ik heb het 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_CONV_reg/4d0849263b8f74f38e79211835fbe9848d9189d84578c66d34b897d7c7e63f9a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "4d16046766088ffeb3100ef9d44b9968c1f5b5b4762ed8207147f4a0cd1e7589", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "58c2bf792a09244c124bf252a93aee81b366d6c5fc9f758b5750d0a0f332dd40", "generation_seconds_batch": 24.33969635493122, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.0834641456604, "target_z": 1.1875000285386226, "ce": 6.24500846862793, "genuineness": 4.519899368286133, "identity": 0.40894895792007446, "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_CONV_reg/4d16046766088ffeb3100ef9d44b9968c1f5b5b4762ed8207147f4a0cd1e7589.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "4d5cdc7b3f5c5aa6c34299c7fadcbdd06c051665463b701df89f998263bb5913", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c1150f7e6da9e7330b07b458a6c06cc59ed36635a1c03c8715b32de1632deada", "generation_seconds_batch": 18.980337023036554, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.001879692077637, "target_z": 1.124999988584551, "ce": 6.125868320465088, "genuineness": 4.231201171875, "identity": 0.39153677225112915, "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_CONV_reg/4d5cdc7b3f5c5aa6c34299c7fadcbdd06c051665463b701df89f998263bb5913.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "4d7040278624110ce2cc0fc3315e7441040f547344738d38d3395007b00ed22d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "352c949b100832572793fc084f912a5fdbf40ae3af0f7d50141ea8ea97f11837", "generation_seconds_batch": 25.22124816698488, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.0834641456604, "target_z": 1.1875000285386226, "ce": 6.165581703186035, "genuineness": 4.2751336097717285, "identity": 0.47840064764022827, "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_CONV_reg/4d7040278624110ce2cc0fc3315e7441040f547344738d38d3395007b00ed22d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "4da0f051972918be905bf9da9736c92f86e2ea2cb449743ed42c0e210378f303", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8cd9120f840391039e5b43a25c927c3bd75e3d5d523515e037dfb6a86310def6", "generation_seconds_batch": 24.1996202589944, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.8857842683792114, "target_z": -0.49609377033376856, "ce": 5.927301406860352, "genuineness": 4.18099308013916, "identity": 0.516007125377655, "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_CONV_reg/4da0f051972918be905bf9da9736c92f86e2ea2cb449743ed42c0e210378f303.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "4db018d0e9063b77ab93306258a4a53c6b7ab0969522b771c44b3652c7482c32", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "98a7300be5cb7b932ad47bf2867bd82956ef9a6523d1d0a2c7ef21575b5c192b", "generation_seconds_batch": 23.665498134098016, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4001946449279785, "target_z": 0.6640624222322536, "ce": 6.316492557525635, "genuineness": 4.482243061065674, "identity": 0.2269904464483261, "hypothesis": "Nee, 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_CONV_reg/4db018d0e9063b77ab93306258a4a53c6b7ab0969522b771c44b3652c7482c32.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "4e0c67703f0ed64f90c88ed831b370fb2e4760730cc38aca592a1851de9b77f2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d8d440a6f7df6b053f1e69c2e85a34c6f9d5b6e521918b63ab301d5e4517d618", "generation_seconds_batch": 27.02163700305391, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2370259761810303, "target_z": 0.5390625249712947, "ce": 6.300607204437256, "genuineness": 4.319066047668457, "identity": 0.3898414969444275, "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_CONV_reg/4e0c67703f0ed64f90c88ed831b370fb2e4760730cc38aca592a1851de9b77f2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "4e505acd93532ee7df38ee354c83658b5cc404e993181397f0794bdb9cc3e624", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9afb509a25255a2dc6685ce89b37a0ce2f851a2532982aa86d5555484d0b6d66", "generation_seconds_batch": 22.34058797894977, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1554415225982666, "target_z": 0.47656248501722315, "ce": 6.165581703186035, "genuineness": 4.394378662109375, "identity": 0.6104364395141602, "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_CONV_reg/4e505acd93532ee7df38ee354c83658b5cc404e993181397f0794bdb9cc3e624.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "4eb786c48b297a8a17da2fc964e1064aa0f70d7b05aecafa5f26c407ca6b148f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6c373ae3ee594354a93dbeb49814d91aad359be37699a0e7b068278ef47c0a1c", "generation_seconds_batch": 24.039389958023094, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.282233476638794, "target_z": -0.19238282703686085, "ce": 5.911416053771973, "genuineness": 4.306513786315918, "identity": 0.6585224270820618, "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_CONV_reg/4eb786c48b297a8a17da2fc964e1064aa0f70d7b05aecafa5f26c407ca6b148f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "4f3aa99f265e6d4910e0995c6763dec287bb5988786ff1bfe229cd2c391510c4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "e36992de24a8579df30f708eabf274d8af4355cab7284a9c4e8d8078866989ba", "generation_seconds_batch": 23.793451146106236, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.920295238494873, "target_z": 1.0624999486304794, "ce": 5.839932441711426, "genuineness": 4.482243061065674, "identity": 0.6581970453262329, "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": 0.03999999999999915, "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_CONV_reg/4f3aa99f265e6d4910e0995c6763dec287bb5988786ff1bfe229cd2c391510c4.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": "4f5e6871df88b91f96a655e9ba682293b2191c4c951d06a7124e5594f8f4861c", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b8a0eae1214e116bbbc21cefa707e36313ab3334122604b94e4b32c278d408fe", "generation_seconds_batch": 15.109293124987744, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.441577911376953, "target_z": -0.07031259061012653, "ce": 6.0146708488464355, "genuineness": 4.036643981933594, "identity": 0.6805331707000732, "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": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/4f5e6871df88b91f96a655e9ba682293b2191c4c951d06a7124e5594f8f4861c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "4f9803894cc83b9e6718e77eb0c14bb9eb8c626395bb28dd1b5ec2ec3469e24f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "93f88371d19ff63f854025dc7cffc612de18693b712b0f16496993b6a12bf58d", "generation_seconds_batch": 21.928342066938058, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7118263244628906, "target_z": 0.1367187232450414, "ce": 5.903473377227783, "genuineness": 4.231201171875, "identity": 0.733444094657898, "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_CONV_reg/4f9803894cc83b9e6718e77eb0c14bb9eb8c626395bb28dd1b5ec2ec3469e24f.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": "4fa278641121286dced609ee9d8f26ca3fb79ffc8a41b4d2bea5d0dd73dc4bff", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f1ad7128da7aea94af2ad811b4a22fdd278d033424e77f0bbe2941b90d0f2ed6", "generation_seconds_batch": 15.058864225982688, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.328808307647705, "target_z": 0.6093749329342371, "ce": 6.205295085906982, "genuineness": 4.432034492492676, "identity": 0.6415742635726929, "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_CONV_reg/4fa278641121286dced609ee9d8f26ca3fb79ffc8a41b4d2bea5d0dd73dc4bff.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "4fd0cffc4f0b90c9851ffc8abd8a71459a5c0cded1cf8ac0135169ce43fe13a8", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "84aefe73936594c85638507f2a86dbf58fadc720e3e0b6118c815bb5979667ba", "generation_seconds_batch": 24.09584124397952, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6959381103515625, "target_z": 0.8906249300803748, "ce": 6.1894097328186035, "genuineness": 4.846253395080566, "identity": 0.6095753908157349, "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_CONV_reg/4fd0cffc4f0b90c9851ffc8abd8a71459a5c0cded1cf8ac0135169ce43fe13a8.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": "4fec91ea357baf88c8ff59f9544b9b3254bc145d036b7169dee561bd33fb8db6", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Your message arrived after I had left.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "79d40882c23877c08a06b48694dcb8134f5bdf0d8118d929a89462ca31af4c90", "generation_seconds_batch": 15.352203065995127, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.33390736579895, "target_z": 0.6132812082622645, "ce": 6.324435234069824, "genuineness": 4.331617832183838, "identity": 0.3050141930580139, "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_CONV_reg/4fec91ea357baf88c8ff59f9544b9b3254bc145d036b7169dee561bd33fb8db6.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": "5009efd33b28e25a5e78f86a478aeadf74ed6a47f814009dc75076fa3365e415", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c222cee0eeceefbbb7a6e6cc1a5d78687a00134f8c1ac3d06c0b84ea6dfd6895", "generation_seconds_batch": 18.13911257009022, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9814834594726562, "target_z": 1.109374887272441, "ce": 5.982900142669678, "genuineness": 4.582659721374512, "identity": 0.302836537361145, "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_CONV_reg/5009efd33b28e25a5e78f86a478aeadf74ed6a47f814009dc75076fa3365e415.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "500c52ae05106bd549274e2faa7d672ee2ef5580b361e4f799505b4accc10f26", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "dcb43933f7543ca1d8024ba9820a5ff75b400451f2d07dd7371b9f7b10350fc1", "generation_seconds_batch": 15.854248579009436, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.2874250411987305, "target_z": 1.343749945776617, "ce": 6.268836498260498, "genuineness": 4.545003414154053, "identity": 0.21438515186309814, "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_CONV_reg/500c52ae05106bd549274e2faa7d672ee2ef5580b361e4f799505b4accc10f26.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "5025eb074ed24c2e8ef233a7b287de0a017f368312b9e09752377267611b9af4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "62f4d763fae378e473b751c307b7c779df3f7b2bc21994f2216db042640d9cd2", "generation_seconds_batch": 24.294068397022784, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.379798650741577, "target_z": 0.6484375035673278, "ce": 5.903473377227783, "genuineness": 4.494794845581055, "identity": 0.567966103553772, "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_CONV_reg/5025eb074ed24c2e8ef233a7b287de0a017f368312b9e09752377267611b9af4.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": "50546f8608e09bf8d8a20677d629cce17b3fc59b28f7706c40d023da11e52510", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "5bab76e1ce850c2c9a8e1ba14987ea817b01babe54b5bb68c50e38c184b7ccea", "generation_seconds_batch": 17.464369576075114, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0891542434692383, "target_z": 0.4257812710472341, "ce": 5.863759994506836, "genuineness": 4.595211505889893, "identity": 0.8258156776428223, "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_CONV_reg/50546f8608e09bf8d8a20677d629cce17b3fc59b28f7706c40d023da11e52510.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "50c840de689ae07a837944ff9d20193879b3baa9a06dccd1774f10c71599b51d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "c7e17732aa301222e2866ccb7c20a16db3b6b813e3a4e1e419bdc3defc3d4e0e", "generation_seconds_batch": 21.996297345962375, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.430788993835449, "target_z": 0.6875000742004186, "ce": 6.125868320465088, "genuineness": 3.9927117824554443, "identity": 0.3434426188468933, "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_CONV_reg/50c840de689ae07a837944ff9d20193879b3baa9a06dccd1774f10c71599b51d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "51468b84963f41a7646ba189fa9fd376e6c8924ca12846c292362a813ff37aac", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f576736a9c8f512857d577621590970ed23dd5f491d543b98f45bf2cedfb2e19", "generation_seconds_batch": 24.986368063022383, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9948222637176514, "target_z": 0.3535156340966859, "ce": 6.284721851348877, "genuineness": 4.18099308013916, "identity": 0.35801056027412415, "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_CONV_reg/51468b84963f41a7646ba189fa9fd376e6c8924ca12846c292362a813ff37aac.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "518a1b6f67a5e96bc946329b84aa959fbe2a57f85f9b99ce0ac5468d8a3d0b06", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5c824e806ccb829ec6602d230d431c91d178ca49b320ccee9dd45127a2957087", "generation_seconds_batch": 16.08214338903781, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.730947732925415, "target_z": 0.15136721006334844, "ce": 6.1894097328186035, "genuineness": 4.218649387359619, "identity": 0.763334333896637, "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_CONV_reg/518a1b6f67a5e96bc946329b84aa959fbe2a57f85f9b99ce0ac5468d8a3d0b06.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "52064d394733eaf107ab25b75be2aef305c026bb7686cad3e9b9260bfaea9ec9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "94346c8046a316d2358a93265eb3a80b42ce9cb5fdc2f5236a47747fd2367f8e", "generation_seconds_batch": 21.583940683980472, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.527670383453369, "target_z": 0.7617187574913884, "ce": 6.086154937744141, "genuineness": 4.595211505889893, "identity": 0.09940273314714432, "hypothesis": "De deur staat nog open. Ik heb je je has 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_CONV_reg/52064d394733eaf107ab25b75be2aef305c026bb7686cad3e9b9260bfaea9ec9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "52798df596ead004e37bbaef2d0e742f81a92274b43a1bb476c8da2989b35c11", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "de5b7539552959ccb3b43059f68d00eee82c686fd3c7009722c4a81011eae548", "generation_seconds_batch": 21.425960997934453, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.369009494781494, "target_z": 1.4062499857306887, "ce": 5.967014789581299, "genuineness": 4.425758361816406, "identity": 0.2686818838119507, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. 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_CONV_reg/52798df596ead004e37bbaef2d0e742f81a92274b43a1bb476c8da2989b35c11.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "5283cf2730894c4b3f2ada35285cbd6e829d92418169de86b2240b033085ed4f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "54c9bb8d708fcb3ae5914eb4992d1587538fb65d92bbf8b04a7fd2ad2cc812b7", "generation_seconds_batch": 25.480015222914517, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.627692461013794, "target_z": 0.07226563695054819, "ce": 6.221180438995361, "genuineness": 4.369274139404297, "identity": 0.5538575649261475, "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": 0.03999999999999915, "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_CONV_reg/5283cf2730894c4b3f2ada35285cbd6e829d92418169de86b2240b033085ed4f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "52fa579042ff2241fe0ace3fc21bf587ef62ef59b51511402b98b03d6e7112e7", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "ef52c46132771f386e0b73d5c4e1fda596185cf22e1e7a2def36b81e7844cb15", "generation_seconds_batch": 22.3940667869756, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0534610748291016, "target_z": 0.3984375263982258, "ce": 6.173524379730225, "genuineness": 4.080576419830322, "identity": 0.4110475778579712, "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_CONV_reg/52fa579042ff2241fe0ace3fc21bf587ef62ef59b51511402b98b03d6e7112e7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "532c6b5f9b54645dfed084020f76cd3dc4ffabcd0902470704101e9e4f94840e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "db6e869c4fd08a4a47403e470fff7c9bbdd378149fe0ec9ea44070c4e0aa2fec", "generation_seconds_batch": 24.298144386033528, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6347498893737793, "target_z": 0.8437499914384132, "ce": 6.300607204437256, "genuineness": 4.105680465698242, "identity": 0.06849458813667297, "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_CONV_reg/532c6b5f9b54645dfed084020f76cd3dc4ffabcd0902470704101e9e4f94840e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "53430b0f2c7576702c1f145774229aee04bf4af37c17f49ec4aa7b3ea8b688b3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "66b31aee3a355290efabe4e84ffae47adb423c3ad75107d58a55ab27896d3753", "generation_seconds_batch": 24.150943858898245, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.890292167663574, "target_z": 0.2734374464900828, "ce": 5.967014789581299, "genuineness": 4.432034492492676, "identity": 0.6215664148330688, "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_CONV_reg/53430b0f2c7576702c1f145774229aee04bf4af37c17f49ec4aa7b3ea8b688b3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "5399728d49be60799fc42b74a9926a227fcf6968c08ff15a1d1cb3acd32fdbad", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1a300327802d12ead1c95f38d3b6f0dc2cf4ae72fb6e513ac8cfa74019a303de", "generation_seconds_batch": 17.33382919197902, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2064318656921387, "target_z": 0.5156250556503139, "ce": 6.229123115539551, "genuineness": 4.482243061065674, "identity": 0.6560806035995483, "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_CONV_reg/5399728d49be60799fc42b74a9926a227fcf6968c08ff15a1d1cb3acd32fdbad.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "53a6437ad8179943202c05f1c8b571e0c3fa8ec36ed2b51107a5eb564484d51d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "a7312471ea4b5764723bb8955268d732dd89d794c48824ba8e31b99da70c7e73", "generation_seconds_batch": 23.41782316903118, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9055893421173096, "target_z": 0.28515627247416525, "ce": 5.824047088623047, "genuineness": 4.6956281661987305, "identity": 0.6072704792022705, "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_CONV_reg/53a6437ad8179943202c05f1c8b571e0c3fa8ec36ed2b51107a5eb564484d51d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "53aba2f1c3cb7f7bc8530154fe08770a650a9d69475e8394fd47890cef90a427", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "fe18fa1675afe9e3b4723e08834ecaadc35f419f67b3fe1a333150f05e3795e6", "generation_seconds_batch": 23.66805281001143, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7105515003204346, "target_z": 0.13574210875123846, "ce": 6.133810997009277, "genuineness": 4.595211505889893, "identity": 0.6499342322349548, "hypothesis": "Ta porte est 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_CONV_reg/53aba2f1c3cb7f7bc8530154fe08770a650a9d69475e8394fd47890cef90a427.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": "53bcc7e5cdff8ebe92f543ffb3a6f07ac706e119202b647531b4788d6dfc9377", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5c0cc0053aa88e3fe332e0a4d5d1bb13a760330ada96e3409a45371f7af5b581", "generation_seconds_batch": 21.552040781942196, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8489089012145996, "target_z": 1.0078124593324629, "ce": 6.062326908111572, "genuineness": 4.545003414154053, "identity": 0.8015841841697693, "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_CONV_reg/53bcc7e5cdff8ebe92f543ffb3a6f07ac706e119202b647531b4788d6dfc9377.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "543e8ecd025972124ee1e14a6a0eb762fc41fe6e7bee6733cd0135fd109fd1c3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "bd5942c212afd190b086350a748a9822e5f39be412a96bc727a3b1a5aa4b1496", "generation_seconds_batch": 26.117432302911766, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3510704040527344, "target_z": -0.13964838407926597, "ce": 6.23706579208374, "genuineness": 4.093128204345703, "identity": 0.3948226869106293, "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_CONV_reg/543e8ecd025972124ee1e14a6a0eb762fc41fe6e7bee6733cd0135fd109fd1c3.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "c58a571a05107864567ff4f6686bbe57204e71c9637522d003421219e5014a55", "generation_seconds_batch": 17.508170478977263, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0056114196777344, "target_z": -0.404296848066675, "ce": 6.2767791748046875, "genuineness": 4.017816066741943, "identity": 0.5580976009368896, "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_CONV_reg/546203f143d2dbd908be3097fd76d41d76e93afd1487fa15c64a3d119dff5d49.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": "547d9949ea1e86ecea2386dddd082b7e28bed39a9c809c018321f006071c800b", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "96bf4f67185d17cbacaaae7bc32e42b3296edaaed2a2f46e3d0ddc0995373070", "generation_seconds_batch": 21.424831039039418, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3950958251953125, "target_z": 0.6601563295514103, "ce": 6.125868320465088, "genuineness": 4.545003414154053, "identity": 0.7233802080154419, "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": 0.03999999999999915, "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_CONV_reg/547d9949ea1e86ecea2386dddd082b7e28bed39a9c809c018321f006071c800b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "5502f62fd7e62b9f5ca8f8737900c3c8cca3714cc310c9a7ea3a8c0eb67654be", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "16512f2666e3f65fc616f5e4d34911d83bd13ccd7ee22fd968f8760ab91f86dd", "generation_seconds_batch": 24.792980027967133, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.379798650741577, "target_z": 0.6484375035673278, "ce": 5.673136234283447, "genuineness": 4.494794845581055, "identity": 0.6561270356178284, "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_CONV_reg/5502f62fd7e62b9f5ca8f8737900c3c8cca3714cc310c9a7ea3a8c0eb67654be.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "550c45f3f2f706f6f22f0c5c211224e69e16785b56df3b53eaaa4abe224ae7fd", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4fc790ca665b5918e10f5ae088b6829b38d01f1a75cb3f71ec1d08045d4315c3", "generation_seconds_batch": 17.37831445701886, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.038163900375366, "target_z": 0.38671870041414336, "ce": 5.982900142669678, "genuineness": 4.243753433227539, "identity": 0.5831414461135864, "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_CONV_reg/550c45f3f2f706f6f22f0c5c211224e69e16785b56df3b53eaaa4abe224ae7fd.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "553006a413370fab3bcbf7245779cad4096d9b44358854d67f1bd0178ae93aa6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7692f94e67859bcf3745935650e6908f0c4ea2e25629788541ac504b30db46ec", "generation_seconds_batch": 23.29842176300008, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.017767906188965, "target_z": 0.3710937817492176, "ce": 6.078212261199951, "genuineness": 4.557555198669434, "identity": 0.6653568744659424, "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_CONV_reg/553006a413370fab3bcbf7245779cad4096d9b44358854d67f1bd0178ae93aa6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "553fe008cdffda37fc4117b1c7dde21ef686702b22e4f596c89dfab848858190", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4c5311cb3f176a96eec04149c8694c0c884c1d706f2b272214e2f652a1f429ab", "generation_seconds_batch": 25.774161521927454, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6500468254089355, "target_z": 0.8554686347753115, "ce": 6.324435234069824, "genuineness": 4.557555198669434, "identity": 0.3599501848220825, "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_CONV_reg/553fe008cdffda37fc4117b1c7dde21ef686702b22e4f596c89dfab848858190.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": "555cb61140279189ee89bf5ac33bd131a2903cbf2f9660583346ddcdc9402cca", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "124a27c68da55ea4603364887a4471a04f5ff64fe078ed20dc250fc611b1c777", "generation_seconds_batch": 15.125952775939368, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.665343999862671, "target_z": 0.867187460759394, "ce": 6.0146708488464355, "genuineness": 4.607763767242432, "identity": 0.7526290416717529, "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_CONV_reg/555cb61140279189ee89bf5ac33bd131a2903cbf2f9660583346ddcdc9402cca.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "5567ced2cd5bad91fbd067366f401fdbeb104534265f297c07cc50f0220e40dd", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "192deb551e597d82524ad25427b1b7c905a58a4f3c131d3498a6296b10d404af", "generation_seconds_batch": 23.790160959004425, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.808116912841797, "target_z": 0.9765626220026113, "ce": 6.0146708488464355, "genuineness": 4.821148872375488, "identity": 0.6647733449935913, "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_CONV_reg/5567ced2cd5bad91fbd067366f401fdbeb104534265f297c07cc50f0220e40dd.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "556a0157fde26130a3f7c1a8f48c98792fae18f9855da8010be4388d90419480", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "4d6b4a89fba26fb795f7a3cad25355eb9516de8c1ec9b38a3391f3216b092a5a", "generation_seconds_batch": 25.497869760962203, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.165048599243164, "target_z": 1.250000068492694, "ce": 6.2767791748046875, "genuineness": 4.381826400756836, "identity": 0.24045529961585999, "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_CONV_reg/556a0157fde26130a3f7c1a8f48c98792fae18f9855da8010be4388d90419480.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": "55a713c49392be0358c13ed21ef906266e77f2af5c77aee78549bc23eb473a12", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "11a0a7d16d3093fe31a27aa652a348169ce6119cab3a2282092562e1286e849e", "generation_seconds_batch": 21.434167910018004, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8998992443084717, "target_z": 1.0468750299655536, "ce": 6.070269584655762, "genuineness": 4.457139015197754, "identity": 0.6276254653930664, "hypothesis": "Les clés are sur la table de la cuisine. J'ai regardé dans le tiroir, puis près de la fenêtre. Prends-les quand tu partiras.", "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_CONV_reg/55a713c49392be0358c13ed21ef906266e77f2af5c77aee78549bc23eb473a12.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "5629afac60f6917f18998b03195e043e494e66be08614e253a5d0d21ee59bebd", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1a4ac1e4873a3c6e45745b819a965d32dc35970682ac073a6c15dda7b2b75997", "generation_seconds_batch": 17.037154870922677, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.2108471393585205, "target_z": -0.24707031633487742, "ce": 5.911416053771973, "genuineness": 4.111956596374512, "identity": 0.4535030424594879, "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_CONV_reg/5629afac60f6917f18998b03195e043e494e66be08614e253a5d0d21ee59bebd.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "567a46ff65de5fdfd6c07b56935e86d29a42068eaa0104dafe6a9d7365719c16", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "53ef34ee00927e7ef50ec0728572acceef23ca6bbf115933c82b843130b4ddb1", "generation_seconds_batch": 22.441024884930812, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5837597846984863, "target_z": 0.8046876034525067, "ce": 6.252951145172119, "genuineness": 4.494794845581055, "identity": 0.44649502635002136, "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_CONV_reg/567a46ff65de5fdfd6c07b56935e86d29a42068eaa0104dafe6a9d7365719c16.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "56de38dbd0e58f137b5c8598b3843d7b7e6235fd3d8244a2a65496f96e95c6e9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6d3c5175699c10a166be70bd865a6fb7c1c18362b24714cdecc7c6c7914ff66c", "generation_seconds_batch": 24.289057996938936, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.076406717300415, "target_z": 0.41601567405075746, "ce": 5.967014789581299, "genuineness": 4.557555198669434, "identity": -0.005713621154427528, "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_CONV_reg/56de38dbd0e58f137b5c8598b3843d7b7e6235fd3d8244a2a65496f96e95c6e9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "56ed8d00ac03f6e7ffa151dbc6557bc39e695967bf085536ac547fee0de2d620", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cfb88520ff01b38936d075d460185e376185aea6bdde84281c3d7d793fae0e6d", "generation_seconds_batch": 22.95933572098147, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.636615753173828, "target_z": 0.07910157311280025, "ce": 6.173524379730225, "genuineness": 4.356721878051758, "identity": 0.72508704662323, "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_CONV_reg/56ed8d00ac03f6e7ffa151dbc6557bc39e695967bf085536ac547fee0de2d620.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "571df0be729b983352586ee6a58820be82155f12fdef021b45bf61d2248879df", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1d1d2595c1965109812a5decbbcb0acb51688c5b1399246607bf72f3eac737a6", "generation_seconds_batch": 25.258093671989627, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9157874584198, "target_z": 0.2929688231302202, "ce": 6.205295085906982, "genuineness": 4.557555198669434, "identity": 0.5751338005065918, "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_CONV_reg/571df0be729b983352586ee6a58820be82155f12fdef021b45bf61d2248879df.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "5735a71de60cd9572933f1d59493ea799d785615d446bc9d0f1f791c88b34c69", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2b8ce0de510c2f9da049efeb0628e51a7107993dc2bd6dbb9a1d9df6866630ca", "generation_seconds_batch": 22.83357784000691, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6398489475250244, "target_z": 0.8476562667664407, "ce": 6.300607204437256, "genuineness": 4.570107460021973, "identity": 0.3364061117172241, "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_CONV_reg/5735a71de60cd9572933f1d59493ea799d785615d446bc9d0f1f791c88b34c69.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "5761e285ccc0a69240c9fed765a7a89d931d0d46db66040020a066c468b1f54e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "44a72eed79c402ebbd259152dff3a25c08b02895de32f475e98b55e5c3b797d9", "generation_seconds_batch": 23.520976540981792, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.415491819381714, "target_z": 0.6757812482163361, "ce": 6.109982967376709, "genuineness": 4.7332844734191895, "identity": 0.7806072235107422, "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_CONV_reg/5761e285ccc0a69240c9fed765a7a89d931d0d46db66040020a066c468b1f54e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "5773b300ded1d7e19e9a576fd428e427a146992b2ffd40dee24d7fef0ab03f47", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "dca6c9ff0325a18334835afd2c0acf68373095322db2b075742b8bab6724f8c3", "generation_seconds_batch": 16.74724169599358, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.824004888534546, "target_z": 0.2226562325200937, "ce": 6.070269584655762, "genuineness": 4.381826400756836, "identity": 0.2608402669429779, "hypothesis": "De deuren staan 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_CONV_reg/5773b300ded1d7e19e9a576fd428e427a146992b2ffd40dee24d7fef0ab03f47.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": "57932b089ae034cc278a8c6622ef15336e8ab1df57d207bd5d42a95a6df5dcc7", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The road turns left behind the bridge.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cb05100856f0530db969011ac43940d10e185afc8027b941e96f831e46f6eeab", "generation_seconds_batch": 15.30347288097255, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7895865440368652, "target_z": 0.19628910236488833, "ce": 6.292664527893066, "genuineness": 4.118232727050781, "identity": 0.44635140895843506, "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_CONV_reg/57932b089ae034cc278a8c6622ef15336e8ab1df57d207bd5d42a95a6df5dcc7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "57ea801b0ad7d851532b0f73922881662664d3eab4bba1671911bd3ce66cac1f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "70d6d93ee42281ac03a7ebca2b63ea5be242a242e1394016e6e553ca12210fd8", "generation_seconds_batch": 18.491748607950285, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.318610429763794, "target_z": 0.6015625649253663, "ce": 6.133810997009277, "genuineness": 4.250029563903809, "identity": 0.3683607578277588, "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_CONV_reg/57ea801b0ad7d851532b0f73922881662664d3eab4bba1671911bd3ce66cac1f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "58af23183bd838c776f0c570c847bc2f6584e9a2108638df89223df5537b8a26", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "bcd6d423c707a464ac685315ff35a7c9e9edca0193fd824a68344a882a9982f1", "generation_seconds_batch": 22.322644317988306, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0591511726379395, "target_z": -0.36328123109316257, "ce": 5.808161735534668, "genuineness": 3.986435890197754, "identity": 0.6562162041664124, "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_CONV_reg/58af23183bd838c776f0c570c847bc2f6584e9a2108638df89223df5537b8a26.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "58cc59890d726364d9cba1ff24cea5370ac6fcbc67077896185638f54ca28010", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "a9d4580a2afd1e0260dd1bc17dd027d88fefafcf094f7e73bdd1e4307d88813f", "generation_seconds_batch": 22.876277455943637, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.727123498916626, "target_z": 0.14843754922912386, "ce": 5.5539960861206055, "genuineness": 3.8169827461242676, "identity": 0.6497504711151123, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás lista.", "wer": 0.06666666666666667, "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_CONV_reg/58cc59890d726364d9cba1ff24cea5370ac6fcbc67077896185638f54ca28010.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "5946ebad4895fbf3e3e032de2b0e6e3e2a19a36a7c3a3296dd8166fb026869e7", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2897cdc92eb8c111e6e13f19a2d74afa639bac1ab283b19a369415c0d8a29d5b", "generation_seconds_batch": 23.24755011498928, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4001946449279785, "target_z": 0.6640624222322536, "ce": 5.95112943649292, "genuineness": 4.7834930419921875, "identity": 0.510814905166626, "hypothesis": "Marie, 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_CONV_reg/5946ebad4895fbf3e3e032de2b0e6e3e2a19a36a7c3a3296dd8166fb026869e7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "59cda91ccb7f63da9eeda213a785fe350beb4896954c63fa3129c620b685f448", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "db1a31e46f2dfd319b95f2553464b34f2403339ea99505e809b37299c87a6deb", "generation_seconds_batch": 24.07771485997364, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6908392906188965, "target_z": 0.8867188373995315, "ce": 6.038498878479004, "genuineness": 4.7081804275512695, "identity": 0.6979237794876099, "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_CONV_reg/59cda91ccb7f63da9eeda213a785fe350beb4896954c63fa3129c620b685f448.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "59d1ca7b3ada6e3045e24647e3bf61670880ba9f408f5f199e5cd92436411408", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "30cabf9e6328c40ce25ca257eff34a096a1fa397ff744f383ace2d11b6154910", "generation_seconds_batch": 24.45010371808894, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9642281532287598, "target_z": 0.3300781647757052, "ce": 5.589737892150879, "genuineness": 4.369274139404297, "identity": 0.45240432024002075, "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_CONV_reg/59d1ca7b3ada6e3045e24647e3bf61670880ba9f408f5f199e5cd92436411408.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "59ff9f85e859a7458657693123de0e447e8795e1c61a2b33b9df94a96f7ed7ce", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6e909adf83c464d034ad58249225c47f4627c1c0dd1108abd657194fb8ad1b6b", "generation_seconds_batch": 26.21769967500586, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 1.9571707248687744, "target_z": -0.4414061897121599, "ce": 6.078212261199951, "genuineness": 4.281409740447998, "identity": 0.6036958694458008, "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_CONV_reg/59ff9f85e859a7458657693123de0e447e8795e1c61a2b33b9df94a96f7ed7ce.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "5a7beb4e327fb47fa62073d1d0b8220ef5b59c730360cd9456f89d4c7e1a238b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0c139d421231b917da8c2841584454db4df3dc183d8b91e7f9a95a00f1bea53b", "generation_seconds_batch": 17.05012504302431, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5078654289245605, "target_z": -0.019531193992953288, "ce": 6.109982967376709, "genuineness": 4.174716949462891, "identity": 0.755513072013855, "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_CONV_reg/5a7beb4e327fb47fa62073d1d0b8220ef5b59c730360cd9456f89d4c7e1a238b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "5b3d8892aaad53a79d8400fb11d15e05920877359a504c0c0678ec084d6fcd9d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cade77294e6e78fa03bfb7cf755da162dde42f0632c5c4bcc6714f1272044f5f", "generation_seconds_batch": 23.8044083340792, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.461383104324341, "target_z": 0.7109375435213994, "ce": 5.839932441711426, "genuineness": 4.7834930419921875, "identity": 0.7808873653411865, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt. Bah.", "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_CONV_reg/5b3d8892aaad53a79d8400fb11d15e05920877359a504c0c0678ec084d6fcd9d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "5ba18755deed2d5c241b9d635f6a838a773af502024eeac0ca1cae7322f00d32", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9bb444ccab1c5e0fd04f395f95e34f777cd672e9908cf02c4926dbbdb5b9d9cb", "generation_seconds_batch": 21.756293547921814, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7921359539031982, "target_z": 0.19824214870531, "ce": 5.720792293548584, "genuineness": 4.457139015197754, "identity": 0.5545452833175659, "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_CONV_reg/5ba18755deed2d5c241b9d635f6a838a773af502024eeac0ca1cae7322f00d32.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "5beb3cf19320b2ee82449ddf7b54b3b59b8aaefe99d514897b57a3b5c757f558", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9292e0dd444f7a5f0801792c20fd99a27e035166f315d6fb54dba1975a404b66", "generation_seconds_batch": 18.09734173596371, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.477271318435669, "target_z": -0.04296866331393407, "ce": 6.38003396987915, "genuineness": 4.406930446624756, "identity": 0.5100095868110657, "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_CONV_reg/5beb3cf19320b2ee82449ddf7b54b3b59b8aaefe99d514897b57a3b5c757f558.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "5c10885613393b3fb392107779831c90a3f5a6c9543887ccc930a24cc53e7d18", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cd702903b99e94bbbc6535d46f90b4f0ffeb9938f46895d97d846886b51e8a11", "generation_seconds_batch": 23.02964445995167, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9814834594726562, "target_z": 1.109374887272441, "ce": 6.157639026641846, "genuineness": 4.68307638168335, "identity": 0.52637779712677, "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_CONV_reg/5c10885613393b3fb392107779831c90a3f5a6c9543887ccc930a24cc53e7d18.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "5c4d828c8b0e4c7ae343c7f7e3037f1fce059c36e98f893955ce8dd63261ec4e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1261b8fdb80780fafb59fe2dc10ac2da0334affc84834f506f416dac480afc08", "generation_seconds_batch": 21.68326976604294, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.568462610244751, "target_z": 0.7929687774684242, "ce": 6.006728172302246, "genuineness": 4.419482707977295, "identity": 0.4305996298789978, "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_CONV_reg/5c4d828c8b0e4c7ae343c7f7e3037f1fce059c36e98f893955ce8dd63261ec4e.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": "5cef00576310135862c235d88258ea4d9a6715a09a3cff8c71e73e4e3d320e46", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Your message arrived after I had left.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f4ec51f426e9af989bc951991d5e97900b31b69f3b3394350a7f16a8c2f57c39", "generation_seconds_batch": 17.394741942058317, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0075697898864746, "target_z": 0.36328123109316257, "ce": 6.356205940246582, "genuineness": 4.293961524963379, "identity": 0.5610464215278625, "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_CONV_reg/5cef00576310135862c235d88258ea4d9a6715a09a3cff8c71e73e4e3d320e46.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "5cf34911f6d55f9455e396ca8473465c47c7d24facfd6ac8d548b96119a07e80", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e84b8751968a244a9c3bba138af26b53b9790e2fa815ae3073ac5311e2a7ecd8", "generation_seconds_batch": 18.913523713010363, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4103927612304688, "target_z": 0.6718749728883086, "ce": 5.998785495758057, "genuineness": 4.7458367347717285, "identity": 0.7493798732757568, "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_CONV_reg/5cf34911f6d55f9455e396ca8473465c47c7d24facfd6ac8d548b96119a07e80.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "5d18c9adc3f4e009bd143464cc421e8ee9b2d432443e1c00a2498423cc1652ce", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cd91ed423ddffe7ac80e9c4f061b1e55b7c1943e61fb2e23aab740a30c1bf22c", "generation_seconds_batch": 24.94570687203668, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.573561668395996, "target_z": 0.7968750527964517, "ce": 6.0146708488464355, "genuineness": 4.331617832183838, "identity": 0.6337448358535767, "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_CONV_reg/5d18c9adc3f4e009bd143464cc421e8ee9b2d432443e1c00a2498423cc1652ce.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "5d1fa6019ebe8db7823cfef34f22404d441fb45a2031715fb541ded9b5c8bdf1", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "574e4744bbde81a231710acbc6bc5c5cff4b2130094bbfbbe1aebdb055d2a09e", "generation_seconds_batch": 22.84915072203148, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7895865440368652, "target_z": 0.19628910236488833, "ce": 6.395919322967529, "genuineness": 4.457139015197754, "identity": 0.5757369995117188, "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_CONV_reg/5d1fa6019ebe8db7823cfef34f22404d441fb45a2031715fb541ded9b5c8bdf1.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "5d2809adf893c27323e08bcf323969c89c9b62785390ff6e4a9a6ab48e13dd8a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "76822721349cf4ef79ce974650938951b4f73f42d0318de1549fb6d2ed44b8d8", "generation_seconds_batch": 22.40016520395875, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.348613262176514, "target_z": 1.3906248844185787, "ce": 5.506340026855469, "genuineness": 4.394378662109375, "identity": 0.6556112766265869, "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_CONV_reg/5d2809adf893c27323e08bcf323969c89c9b62785390ff6e4a9a6ab48e13dd8a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "5d31fea3cd0dac505bb684108fa62a265548bfc1e3f19794c288f9a3cf690bd0", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c2fb3b7880db46a4293b6bde3a4729328450d3f53821a35add28ee4b946197e2", "generation_seconds_batch": 23.790160959004425, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7418293952941895, "target_z": 0.9257812253854381, "ce": 5.998785495758057, "genuineness": 4.758388519287109, "identity": 0.6487622261047363, "hypothesis": "Que he 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_CONV_reg/5d31fea3cd0dac505bb684108fa62a265548bfc1e3f19794c288f9a3cf690bd0.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": "5dc5cec6d1a2d24fd6450571e9928f62cb6131db31512144619cdafbe6595765", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "5a21c05a6323f859ed079f6739beb702a89b30004c4a0cacb345876a359bc0d6", "generation_seconds_batch": 15.12361750099808, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7010371685028076, "target_z": 0.8945312054084023, "ce": 6.181467056274414, "genuineness": 4.582659721374512, "identity": 0.7434254884719849, "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": 0.03999999999999915, "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_CONV_reg/5dc5cec6d1a2d24fd6450571e9928f62cb6131db31512144619cdafbe6595765.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "a14f763ed5f426927186ee44fbeafe6c346403bef4d2db0477cf5bfbcac5ae59", "generation_seconds_batch": 16.938916768995114, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.122297763824463, "target_z": 0.45117178670863656, "ce": 6.09409761428833, "genuineness": 4.406930446624756, "identity": 0.7560688257217407, "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_CONV_reg/5e0af124ad7421917e897f2876d49b25e213a1bc87e01fe43f09b9c255c2a32f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "5e9fa8a493cb78caef1a9ec21ca59d2f4c08ce8f1e8c7502b328254361d70a60", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "282b36bbc5e7ab2279b6740c67ed64ec729c67c9a9830ac6dc1a211432bbf29e", "generation_seconds_batch": 24.01211730495561, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.573561668395996, "target_z": 0.7968750527964517, "ce": 6.141753673553467, "genuineness": 4.821148872375488, "identity": 0.5528767704963684, "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_CONV_reg/5e9fa8a493cb78caef1a9ec21ca59d2f4c08ce8f1e8c7502b328254361d70a60.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "5ec2c15d3b3d4419fd81a1ef6c9ff7a75efb7cea26c59af6664a56e3e9c21698", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1e3c0b62eddb38983bdeb464537e0953a7f035ec9783ca73579876ab8b405fbe", "generation_seconds_batch": 22.814648712985218, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9814834594726562, "target_z": 1.109374887272441, "ce": 5.95112943649292, "genuineness": 4.519899368286133, "identity": 0.5364912748336792, "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_CONV_reg/5ec2c15d3b3d4419fd81a1ef6c9ff7a75efb7cea26c59af6664a56e3e9c21698.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "5edade7606f78bcf7d58e9e8965fcc704546b9a0735e6adc2f3b4675f7a70daf", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "798326354a36ad2196e292e8262cd92922b63e2fd99e23b3b1bfd4910fc89357", "generation_seconds_batch": 22.07339332101401, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3848977088928223, "target_z": 0.6523437788953553, "ce": 6.157639026641846, "genuineness": 4.331617832183838, "identity": 0.3339056968688965, "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_CONV_reg/5edade7606f78bcf7d58e9e8965fcc704546b9a0735e6adc2f3b4675f7a70daf.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": "5f2c00883a2887d0ebf76b5c329c47654c78109849c9710a7bb5de0c31b34d40", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Someone moved the little blue chair.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e193f907107f07e399f42bee86c7ceccfdebbab21b39730685d320044e9f6616", "generation_seconds_batch": 15.214639994082972, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5939576625823975, "target_z": 0.8124999714613774, "ce": 6.38003396987915, "genuineness": 4.243753433227539, "identity": 0.31424880027770996, "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_CONV_reg/5f2c00883a2887d0ebf76b5c329c47654c78109849c9710a7bb5de0c31b34d40.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "607d33c8bd31fe82272bc668a5cb5e72d8a98ba5308f2b58019937a93726d200", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2f31887666d7b2810c29e1e69a9b0f2c0f4dc32690a2e8b0842aa1091effb5e3", "generation_seconds_batch": 21.574564099078998, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.042672157287598, "target_z": 1.156250191208771, "ce": 6.09409761428833, "genuineness": 4.532451152801514, "identity": 0.5830365419387817, "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_CONV_reg/607d33c8bd31fe82272bc668a5cb5e72d8a98ba5308f2b58019937a93726d200.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "6087a994dfdbe4128be130a899036e953eb4ed35ea89d2007caf8155d8ef7b1d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "ba03e7b727c88518b3fb9ed0a1e02e64023b48e885b9986af24170ddfa91e756", "generation_seconds_batch": 26.21769967500586, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.971876621246338, "target_z": 0.3359374864441543, "ce": 5.855817794799805, "genuineness": 4.243753433227539, "identity": 0.5260342359542847, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás lista.", "wer": 0.06666666666666667, "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_CONV_reg/6087a994dfdbe4128be130a899036e953eb4ed35ea89d2007caf8155d8ef7b1d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "6090d7ec5b8f06d785c0c0fbb596adbb17ebf031eafbfb7c148c98e9a2ee3b90", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f2e78b8a60476e3f4e02ee672c6be3e204c2ccdb86ca83ec8249f388b1192b9f", "generation_seconds_batch": 25.823580799042247, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2880163192749023, "target_z": 0.5781250956043855, "ce": 6.332377910614014, "genuineness": 4.168440818786621, "identity": 0.48751601576805115, "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_CONV_reg/6090d7ec5b8f06d785c0c0fbb596adbb17ebf031eafbfb7c148c98e9a2ee3b90.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "60e13c5a7fdb65de512155574a24fc62365d270d8d7b6ef52e5136af598e5dd0", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5199f7acfc99a93955347e34c37204747c0c20568966341879477cc45e93e4c0", "generation_seconds_batch": 26.198932545026764, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5582644939422607, "target_z": 0.7851562268123692, "ce": 6.133810997009277, "genuineness": 4.494794845581055, "identity": 0.6313420534133911, "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_CONV_reg/60e13c5a7fdb65de512155574a24fc62365d270d8d7b6ef52e5136af598e5dd0.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "610dfbf5fbe1d1478cd17eccdcd53ba0e9c252e67eb20138598c59337de04095", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ac85028cf80dfed99ac6bc04271523107feb176255377f6ecd44474e6f144a31", "generation_seconds_batch": 23.111561674042605, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8285129070281982, "target_z": 0.9921875406675371, "ce": 6.109982967376709, "genuineness": 4.20609712600708, "identity": 0.23839503526687622, "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_CONV_reg/610dfbf5fbe1d1478cd17eccdcd53ba0e9c252e67eb20138598c59337de04095.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": "616484945cd57817627763c5ee62b59bd4c00bff3e3cd35f4024953cdb614819", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I wrote the number on the envelope.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "82720de7c54d8dd4a09386eae6b0236e8f91c6027f7bcea19cbfa0e85685042b", "generation_seconds_batch": 15.262272540014237, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6347498893737793, "target_z": 0.8437499914384132, "ce": 6.403861999511719, "genuineness": 4.256305694580078, "identity": 0.22247567772865295, "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_CONV_reg/616484945cd57817627763c5ee62b59bd4c00bff3e3cd35f4024953cdb614819.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "620fd8a3a9c8cb107612a1df51e28bc67cda78fccb90b0a8a06e90ed9608528a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "913979e3c8deb7f1361124f199d18ae9fa1b9e073acc962aedaca53973162bd1", "generation_seconds_batch": 18.034853058983572, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7571263313293457, "target_z": 0.9374998687223364, "ce": 5.967014789581299, "genuineness": 4.645420074462891, "identity": 0.6987536549568176, "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_CONV_reg/620fd8a3a9c8cb107612a1df51e28bc67cda78fccb90b0a8a06e90ed9608528a.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": "624c7d410b275130883a333bf8cf8999c071ee29f3f89854e45772bebe363407", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "fd8d9a1f4155c9b9a284a9be3476bf8b9f8f332f23a0d512b3dc0425af96777d", "generation_seconds_batch": 15.12361750099808, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.318610429763794, "target_z": 0.6015625649253663, "ce": 6.173524379730225, "genuineness": 4.344170093536377, "identity": 0.7506991028785706, "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_CONV_reg/624c7d410b275130883a333bf8cf8999c071ee29f3f89854e45772bebe363407.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "6373295117fa28906eae45031fdcb1988c9988508e6b6b537d6d8a3bb97f6cf5", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7c859624b2a708acf0676f7836de40354929eb4e365cb8c0b9f1c31efea5618b", "generation_seconds_batch": 25.990586593048647, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0234580039978027, "target_z": -0.39062497574217087, "ce": 6.260893821716309, "genuineness": 4.030367851257324, "identity": 0.590356707572937, "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_CONV_reg/6373295117fa28906eae45031fdcb1988c9988508e6b6b537d6d8a3bb97f6cf5.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "63a567a72cc6c4ce301ba5df6d6f815e84d5092af5297013aa8ad448be3921a0", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7d4f06cab7003e76ff5c8453c901b91bb170e02918112e717a700edcb1807b1d", "generation_seconds_batch": 23.112058863043785, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.1854448318481445, "target_z": 1.265625169804804, "ce": 5.974957466125488, "genuineness": 4.557555198669434, "identity": 0.5280705690383911, "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_CONV_reg/63a567a72cc6c4ce301ba5df6d6f815e84d5092af5297013aa8ad448be3921a0.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": "642a4d8d2f5ddd4e20f6c584d3dc85ebbee869b08113bf8635d6d6c464c0ba58", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "3f4d7a47836fabe613f614b113cdae3b98993eb67658c964c74643b5770c6f3a", "generation_seconds_batch": 22.032627566950396, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.665343999862671, "target_z": 0.867187460759394, "ce": 6.078212261199951, "genuineness": 4.607763767242432, "identity": 0.7771166563034058, "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_CONV_reg/642a4d8d2f5ddd4e20f6c584d3dc85ebbee869b08113bf8635d6d6c464c0ba58.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "65887f9702ea7f3bd9b4a980a0dbe564a67fbd908d8725fb143fcc6145aede19", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b0e534302a983e5f4c9d864feb452218e14def94e72665b716bb53bef3f69578", "generation_seconds_batch": 18.65115834691096, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.674858331680298, "target_z": 0.10839836410223018, "ce": 6.117925643920898, "genuineness": 4.469690799713135, "identity": 0.7478443384170532, "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_CONV_reg/65887f9702ea7f3bd9b4a980a0dbe564a67fbd908d8725fb143fcc6145aede19.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "658a2f37ae717851a8161d0b20524fc13106011c9d7fea66b7e72353c8ab8869", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "c03c26d9e1e78eb052eb0445550628b81b2c6efda7ae46aa21862d1981bd9904", "generation_seconds_batch": 17.376710313023068, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.910097122192383, "target_z": 1.0546873979744245, "ce": 6.181467056274414, "genuineness": 4.381826400756836, "identity": 0.4168248176574707, "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_CONV_reg/658a2f37ae717851a8161d0b20524fc13106011c9d7fea66b7e72353c8ab8869.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "65a4cf14f882f302dd47f9ca8dfb2141d5b01dcc48cb25a7a8286c729f3ea889", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "755e7a2ea7379d00b8f5712d10851d0362c8f55cbd3ef70198a4958bb58707d4", "generation_seconds_batch": 24.494189434917644, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4919772148132324, "target_z": 0.7343750128423802, "ce": 6.141753673553467, "genuineness": 4.055472373962402, "identity": 0.34614884853363037, "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_CONV_reg/65a4cf14f882f302dd47f9ca8dfb2141d5b01dcc48cb25a7a8286c729f3ea889.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "66661b8c2a4fcc803d7c344e754b41e43c279cb714ebc7d3c8c6cc906fc0c8dc", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "baf8e64ab1b400fee2bff8f568e34abc63720bfdf75e9501bdc370d6a421397c", "generation_seconds_batch": 23.4627615690697, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.925985336303711, "target_z": 0.300781191139091, "ce": 6.1894097328186035, "genuineness": 4.607763767242432, "identity": 0.7477453351020813, "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_CONV_reg/66661b8c2a4fcc803d7c344e754b41e43c279cb714ebc7d3c8c6cc906fc0c8dc.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "676ec87d148a0d2e84b13fd4ca6004f39654a04a1ba75c1f3cbef2c7d8c7d3c2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7a94024fdcd59ecdaec178772a0b7d028c6980de56f999a04a51c15b82b1c402", "generation_seconds_batch": 26.68118863797281, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9406914710998535, "target_z": 1.0781250499425894, "ce": 6.070269584655762, "genuineness": 4.1998209953308105, "identity": 0.49827703833580017, "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_CONV_reg/676ec87d148a0d2e84b13fd4ca6004f39654a04a1ba75c1f3cbef2c7d8c7d3c2.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": "67e7644dfb38930db46e7f05b40296329777e83ff6898c7435757f7fb4802331", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "5be27d739d757727c0bc4d2b62e8e4c4aab5ecd5a73d709e2923e7b6687fcc43", "generation_seconds_batch": 17.386320078046992, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.716334342956543, "target_z": 0.9062500313924848, "ce": 6.229123115539551, "genuineness": 4.532451152801514, "identity": 0.5822159051895142, "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_CONV_reg/67e7644dfb38930db46e7f05b40296329777e83ff6898c7435757f7fb4802331.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "6837d45f3e13301e8bddd3aa8468e86d0739bd2f8eee32635e2e7ef4ca9257ce", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2a3c85598eb3ac7eaae0f216b0691f6f504b4ab0cbf33d1f102ce219795184c5", "generation_seconds_batch": 24.001647231983952, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.920295238494873, "target_z": 1.0624999486304794, "ce": 5.990842819213867, "genuineness": 4.344170093536377, "identity": 0.6732670664787292, "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_CONV_reg/6837d45f3e13301e8bddd3aa8468e86d0739bd2f8eee32635e2e7ef4ca9257ce.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": "6868ed13db8ad96567ad2e2cc0601f1907f811068e04a68f1d788f2b84499cca", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The keys are on the kitchen table.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "374d007cdcc514bfa7b44e5227e5b005d3c30fc90fa0bea416034aa188020a7a", "generation_seconds_batch": 16.802226728061214, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8795032501220703, "target_z": 1.0312501113006278, "ce": 6.149696350097656, "genuineness": 4.118232727050781, "identity": 0.37003207206726074, "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_CONV_reg/6868ed13db8ad96567ad2e2cc0601f1907f811068e04a68f1d788f2b84499cca.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "68b37997bb6379833cd73d6931c2d18984288bc7f7a506a3dcd9a4b911d5491a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8ca738a3f1f6c998564634134f07c3a43f7dbd6828e9d155befbfa463cb04291", "generation_seconds_batch": 24.375515970983543, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.563363552093506, "target_z": 0.7890625021403966, "ce": 6.1020402908325195, "genuineness": 4.770940780639648, "identity": 0.6133385300636292, "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_CONV_reg/68b37997bb6379833cd73d6931c2d18984288bc7f7a506a3dcd9a4b911d5491a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "68b98f7760ae1f74a6c8e6b5f7c13d0c231f194dcb8d92ef9837cd7ab86ecb3c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4dfead0dbeea4a296224f1f0ca89386d5d560ae4c6ea8fe765cf2884eb5fed9e", "generation_seconds_batch": 22.322644317988306, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3701915740966797, "target_z": -0.12500007990814307, "ce": 5.959072113037109, "genuineness": 4.2374773025512695, "identity": 0.4733527600765228, "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_CONV_reg/68b98f7760ae1f74a6c8e6b5f7c13d0c231f194dcb8d92ef9837cd7ab86ecb3c.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": "68d61f02e078641970b9b6256b13414b2678af67e098cd20068af4a6f275000a", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9be3d38f7eab42694b06440aa946a10976b000776e58da5e2de84782141b9c8f", "generation_seconds_batch": 16.04912235995289, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6646604537963867, "target_z": 0.10058599609335937, "ce": 6.229123115539551, "genuineness": 4.356721878051758, "identity": 0.2815161943435669, "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_CONV_reg/68d61f02e078641970b9b6256b13414b2678af67e098cd20068af4a6f275000a.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": "68e4f1814372defdfa61cf5534118d4d8fc567c0e910e3d02852daf1a743645a", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Your message arrived after I had left.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b8b7c0568f1e841b81e70842efb1f34793cf3fe10ae74d4f1818f5f3ef5e6ec3", "generation_seconds_batch": 17.394741942058317, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.042672157287598, "target_z": 1.156250191208771, "ce": 6.515059471130371, "genuineness": 4.645420074462891, "identity": 0.35426396131515503, "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_CONV_reg/68e4f1814372defdfa61cf5534118d4d8fc567c0e910e3d02852daf1a743645a.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": "699b75ec5426b8c7fb679fb7d4a37d2f4a2ef1785d28070e5df3542c24deb50a", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "4bd6a80308d97e123650ffbc4c01b95f4824c940ca81a11649a26b70162d4a32", "generation_seconds_batch": 15.169848765945062, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3492045402526855, "target_z": 0.625000034246347, "ce": 5.990842819213867, "genuineness": 4.394378662109375, "identity": 0.5081873536109924, "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_CONV_reg/699b75ec5426b8c7fb679fb7d4a37d2f4a2ef1785d28070e5df3542c24deb50a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "69a2a7cb546b71afd97a87f37decca0f003b903b1c7d7d01d048d5096d010805", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "fc7761a6a825b8d39a84be0c3233a0c3f6ea02c44fbd640128047d35b398782c", "generation_seconds_batch": 18.375780744943768, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3848977088928223, "target_z": 0.6523437788953553, "ce": 5.974957466125488, "genuineness": 4.519899368286133, "identity": 0.04409356415271759, "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": 0.03999999999999915, "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_CONV_reg/69a2a7cb546b71afd97a87f37decca0f003b903b1c7d7d01d048d5096d010805.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": "69c8ece6be1954dbcc2747dadcaf05512aa2280479bfd4d78765b7a89b159a85", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I wrote the number on the envelope.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4ce411f858e2dc3968f773e277dfde370937cb007adc739e252fa2e096fb0e8b", "generation_seconds_batch": 23.703053252073005, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7571263313293457, "target_z": 0.9374998687223364, "ce": 6.292664527893066, "genuineness": 4.174716949462891, "identity": 0.5953548550605774, "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_CONV_reg/69c8ece6be1954dbcc2747dadcaf05512aa2280479bfd4d78765b7a89b159a85.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": "69f3ac753e525bca8af2d6dc4ec82d351be37d766f661a1d91db580766d6fce7", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "eb11c94b6dab233edaa96845251c84bcee4053182ffbcb9b2ab0feeb060f1016", "generation_seconds_batch": 15.068975197034888, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.221728801727295, "target_z": 0.5273436989872122, "ce": 6.229123115539551, "genuineness": 4.532451152801514, "identity": 0.30725806951522827, "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_CONV_reg/69f3ac753e525bca8af2d6dc4ec82d351be37d766f661a1d91db580766d6fce7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "6a40696da214ffbd0f2564d2c279db80991d49e252bb6c7df243fa4e34ee1f16", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "635012edf4b3421889f3c9edfb1b7c9a903099b174df3311393b70754f3cb941", "generation_seconds_batch": 23.26624842500314, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.573561668395996, "target_z": 0.7968750527964517, "ce": 6.046441555023193, "genuineness": 4.325342178344727, "identity": 0.43041402101516724, "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_CONV_reg/6a40696da214ffbd0f2564d2c279db80991d49e252bb6c7df243fa4e34ee1f16.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "0b857ddef5d758b15320495bdb5aed8f77060ac81d80cf5351069edd73da5ccc", "generation_seconds_batch": 18.48193127696868, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.328808307647705, "target_z": 0.6093749329342371, "ce": 6.3641486167907715, "genuineness": 4.293961524963379, "identity": 0.26220837235450745, "hypothesis": "I wrote the number on the envelope. Write it carefully and check the last two digits. Call me when you have a moment.", "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_CONV_reg/6a61bc20d43b3e97640d8dd45682f7055c5dbaa5aa987cdb55f8a7d7a432a2c2.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": "6a7711d15831af79d4d7ec271d12a9b9c4c4f4295f741cbb442de4af5499ff95", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b2a3762864e11610c279feb23a14a30260c438b9750bcec6b70737aaaec52ba7", "generation_seconds_batch": 15.29140407196246, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.85910701751709, "target_z": 1.0156250099885178, "ce": 6.149696350097656, "genuineness": 4.306513786315918, "identity": 0.31344741582870483, "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_CONV_reg/6a7711d15831af79d4d7ec271d12a9b9c4c4f4295f741cbb442de4af5499ff95.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "6acc33508817a0f27589e0b4ab11755c2d83e95a70eb821d381edea546bd608b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "9552a96b2e177aa7ebcfdaadc552fcc987f44effa33a2a0fa458f0cb38915500", "generation_seconds_batch": 23.128393740043975, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4919772148132324, "target_z": 0.7343750128423802, "ce": 6.213237762451172, "genuineness": 4.419482707977295, "identity": 0.24458174407482147, "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_CONV_reg/6acc33508817a0f27589e0b4ab11755c2d83e95a70eb821d381edea546bd608b.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": "6ad59f312bc939153d88cfcdc54f39a01691d282f4d1c63e1649f9b512ac9f31", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4cbf22074c6edc9d039dd773da9d59738e93fd9ca4e7bcdf01a9a1ffc529d623", "generation_seconds_batch": 15.992530610994436, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.522571325302124, "target_z": 0.757812482163361, "ce": 6.157639026641846, "genuineness": 4.532451152801514, "identity": 0.6272099614143372, "hypothesis": "Die Schlüssel liegen auf dem Küchentisch. Ich suchte in der Schublade daneben dem Fenster. Nimm sie mit, wenn du gehst.", "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_CONV_reg/6ad59f312bc939153d88cfcdc54f39a01691d282f4d1c63e1649f9b512ac9f31.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "6ae7f8eb85fa82e3ecacfc97b29bc8b4d573e801ca58808129314a0ac164d422", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3991c56f4b458a753e82abe71c4df1b1b8bb1d164360b63ce2d3e6cff3c4bb36", "generation_seconds_batch": 16.76785416295752, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.882643699645996, "target_z": 0.2675781248216336, "ce": 5.919358730316162, "genuineness": 4.381826400756836, "identity": 0.6359502077102661, "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_CONV_reg/6ae7f8eb85fa82e3ecacfc97b29bc8b4d573e801ca58808129314a0ac164d422.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "6b4f308bd0556028290f17dea534d01d32ab7e8fd7ee1507dc11d1b8a7954075", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ec1448ddfb7e471aa933988558347c53539f5a53a4f74c571724c264c51f181c", "generation_seconds_batch": 25.627598464954644, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9081387519836426, "target_z": 0.2871093188145869, "ce": 6.268836498260498, "genuineness": 4.080576419830322, "identity": 0.3345772624015808, "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_CONV_reg/6b4f308bd0556028290f17dea534d01d32ab7e8fd7ee1507dc11d1b8a7954075.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "210d258e5dcabed0622d3e23ad840f87bf371755a2a8117724a8e68b916aa03f", "generation_seconds_batch": 17.593891810043715, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.430788993835449, "target_z": 0.6875000742004186, "ce": 5.982900142669678, "genuineness": 4.532451152801514, "identity": 0.7182673215866089, "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_CONV_reg/6b973c750e3616c70f28663c29496a0c74bea4a87dff4bad65cadc5394b99131.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "6ba8269f8b957282cad550fb87707667712be19b415e74e3841f03fac336d18b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "59de88e4ca284f7e1fbef6cbbd5d3bdbd75fc2079f56497f53abae34d2f1381e", "generation_seconds_batch": 20.214222824084572, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.517472267150879, "target_z": 0.7539062068353334, "ce": 6.086154937744141, "genuineness": 4.582659721374512, "identity": 0.6650700569152832, "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_CONV_reg/6ba8269f8b957282cad550fb87707667712be19b415e74e3841f03fac336d18b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "6bf824f2eb4168eddcacd441d4d570aae2eb160ca5089a6cee3d060997920d72", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "8543ca6eaa06e72db2801bb2078f3d9212e4a6c0be587d78c7bede7db46d3fa9", "generation_seconds_batch": 25.043361529009417, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.22682785987854, "target_z": 0.5312499743152397, "ce": 6.078212261199951, "genuineness": 3.9048471450805664, "identity": 0.4658389091491699, "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_CONV_reg/6bf824f2eb4168eddcacd441d4d570aae2eb160ca5089a6cee3d060997920d72.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "6d37cfe4827ecf07a09aca71e7cdfadc13c8d1ef7e14710d25538e683e80783b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "5c5aaa21b8f26676f957f9ff2fb5690e34102facf590cb123c2892b44d7ba4ce", "generation_seconds_batch": 24.294068397022784, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.415491819381714, "target_z": 0.6757812482163361, "ce": 6.213237762451172, "genuineness": 4.6705241203308105, "identity": 0.4980385899543762, "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_CONV_reg/6d37cfe4827ecf07a09aca71e7cdfadc13c8d1ef7e14710d25538e683e80783b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "6d3b6925b9ebd25a66fcd555f16588f10822999972d31bc2d22088088871db4d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2e8f2073fdd03ddedcca70f3c3c7fd7fe5a7dbbb4e4391c9b88e253cacd8ecb6", "generation_seconds_batch": 25.914695473969914, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4919772148132324, "target_z": 0.7343750128423802, "ce": 6.0146708488464355, "genuineness": 4.68307638168335, "identity": 0.5984871983528137, "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_CONV_reg/6d3b6925b9ebd25a66fcd555f16588f10822999972d31bc2d22088088871db4d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "6d97db414b9157849e6756edf33dd5dad0c1211097d590397113e769b1e42b84", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1e73c00e75ceedd448baae1fa9db00d875c9034417eaebcba7f274b0e3aad80e", "generation_seconds_batch": 23.815440216916613, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.808116912841797, "target_z": 0.9765626220026113, "ce": 6.173524379730225, "genuineness": 4.582659721374512, "identity": 0.6672946810722351, "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_CONV_reg/6d97db414b9157849e6756edf33dd5dad0c1211097d590397113e769b1e42b84.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "6da1f68dcc6a5a8ed94d92175028430f17bfca2728c547f48b2bb9309661e73d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4fabb5aeec0c7bb2532f6a1c7f16070899d505077e0b39c9c5c7965e6616aee2", "generation_seconds_batch": 24.289057996938936, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.920295238494873, "target_z": 1.0624999486304794, "ce": 6.149696350097656, "genuineness": 4.68307638168335, "identity": 0.4805224537849426, "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_CONV_reg/6da1f68dcc6a5a8ed94d92175028430f17bfca2728c547f48b2bb9309661e73d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "6e5a11aff41c87fea15d4f3e1ec98b466ce1a4d6a7800607b8b4898fe423980e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "01e7437a405aee9ec47dd45ff6a89ab34d036b9b11a8b199198c6d3b9b114f60", "generation_seconds_batch": 18.11529825697653, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.001879692077637, "target_z": 1.124999988584551, "ce": 6.268836498260498, "genuineness": 4.362998008728027, "identity": 0.44365084171295166, "hypothesis": "The door's still open. I put your coat on the chair. Tell me when you are ready.", "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_CONV_reg/6e5a11aff41c87fea15d4f3e1ec98b466ce1a4d6a7800607b8b4898fe423980e.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": "6e954637c80fa283f3a0b279215f12705d76bf81ad065e066a6cd6b7853e4f08", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik schreef het nummer op de envelop.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lees het goed... en controleer de laatste twee cijfers.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "9384eda923a41cb3713a7471c75fa274545083887320a763824d896172441cb6", "generation_seconds_batch": 18.394700975040905, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7418293952941895, "target_z": 0.9257812253854381, "ce": 6.213237762451172, "genuineness": 4.469690799713135, "identity": 0.6242390871047974, "hypothesis": "Ik schreef het nummer op de envelop. Lees het goed en controleer de laatste twee cijfers. Mel me, wanneer je even tijd hebt.", "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_CONV_reg/6e954637c80fa283f3a0b279215f12705d76bf81ad065e066a6cd6b7853e4f08.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "6ee6cfa3b595f87bb27d4cd21a4634f0513487e8d3a218a2b330705c6791d238", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "16d66b758ad2c6d0872be38c7f2b79c256a2376861302e8adde206c3fe6e503b", "generation_seconds_batch": 24.118739309953526, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.144652366638184, "target_z": 1.2343749671805841, "ce": 6.046441555023193, "genuineness": 4.293961524963379, "identity": 0.5697402954101562, "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_CONV_reg/6ee6cfa3b595f87bb27d4cd21a4634f0513487e8d3a218a2b330705c6791d238.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "6f6493a34591c041165170638f7d1b45367ee8583e2fc34f1b63b4281780ff0b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c9222d28ffe725c9678660855897aadac0d6bd3122e678ed73bcce1a0b2e481c", "generation_seconds_batch": 24.825870117987506, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.984624147415161, "target_z": 0.34570308344063094, "ce": 5.967014789581299, "genuineness": 4.262581825256348, "identity": 0.3796982765197754, "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_CONV_reg/6f6493a34591c041165170638f7d1b45367ee8583e2fc34f1b63b4281780ff0b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "7059ae5856e2959456feb1a297bb0034366e94da16a87f10a20fc6cae9f080a2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "70d7b468cc58769ae04e2ddf7719a1b270d959ae2e18ca19fff3d044bd691efa", "generation_seconds_batch": 24.039389958023094, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.420590877532959, "target_z": 0.6796875235443636, "ce": 6.2767791748046875, "genuineness": 4.6203155517578125, "identity": 0.7765803337097168, "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_CONV_reg/7059ae5856e2959456feb1a297bb0034366e94da16a87f10a20fc6cae9f080a2.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": "709a8b64b41626f69fafed3780d712a9b5ebdcf5a389081cfff8718ebce1534b", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "105a680ec2a91b0a114a77dbd33af9e255a9096134b87712c99e3aec12357488", "generation_seconds_batch": 25.643560266005807, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.492568254470825, "target_z": -0.03125001997703577, "ce": 6.062326908111572, "genuineness": 4.281409740447998, "identity": 0.47163134813308716, "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_CONV_reg/709a8b64b41626f69fafed3780d712a9b5ebdcf5a389081cfff8718ebce1534b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "70ba13fc3159adc3c72618ef068bbc6e7d67f10d017d63b1b8e5c398eca563b6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c51ef5da8c10e1b1c24fb1ba6bd3ff6c51b7a1bb0f94f07d1e313526da0890c5", "generation_seconds_batch": 17.377995159942657, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.022275924682617, "target_z": 1.140625089896661, "ce": 6.109982967376709, "genuineness": 4.6705241203308105, "identity": 0.798688530921936, "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_CONV_reg/70ba13fc3159adc3c72618ef068bbc6e7d67f10d017d63b1b8e5c398eca563b6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "70d464f5d95e088fd01bfdd43f8cef2399bdcacf061ac8ea1e402ab63d94c6ee", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1835f0fffb374b4c3dcf26562596a16edb9ee5ac8d04a3d6f374beb0f18885a4", "generation_seconds_batch": 25.406823559082113, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3797523975372314, "target_z": -0.11767574517539746, "ce": 6.125868320465088, "genuineness": 4.3127899169921875, "identity": 0.3432658314704895, "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_CONV_reg/70d464f5d95e088fd01bfdd43f8cef2399bdcacf061ac8ea1e402ab63d94c6ee.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "70dbab45e0bd272ab770d292370d8ddebc917f22b360c6a6b5a222f45b1fe773", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5ae4d375819ea8633352a933d635974c61f42ebf4ae4c073f858a67bfbc14791", "generation_seconds_batch": 26.42996378301177, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5277833938598633, "target_z": -0.004272482932305565, "ce": 6.078212261199951, "genuineness": 4.632867813110352, "identity": 0.617588996887207, "hypothesis": "Ta porte est 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_CONV_reg/70dbab45e0bd272ab770d292370d8ddebc917f22b360c6a6b5a222f45b1fe773.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": "70ded04d430d54a9870b721947f39d110b7677a300ac453ee839b1ca5908feb5", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6b3de05d93129a1ddb316c19e22c04125bc522911eb69e1a76aea387ffcd827e", "generation_seconds_batch": 21.350276642944664, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7449700832366943, "target_z": 0.162109421553628, "ce": 6.117925643920898, "genuineness": 4.557555198669434, "identity": 0.5998709797859192, "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_CONV_reg/70ded04d430d54a9870b721947f39d110b7677a300ac453ee839b1ca5908feb5.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "70eb6d82ff5d79641bfd91e9daebe6dd51bcab7f83ac54ec311690ad859cb4ab", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d149c664f0daa6a4b68457962bb4a8110a730f803b2b1af1afb31b3df1984a35", "generation_seconds_batch": 23.32950896001421, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.189176321029663, "target_z": -0.2636718494936061, "ce": 6.3641486167907715, "genuineness": 4.256305694580078, "identity": 0.5337912440299988, "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_CONV_reg/70eb6d82ff5d79641bfd91e9daebe6dd51bcab7f83ac54ec311690ad859cb4ab.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": "721ab2cd3569fc7e1edb0977bad65273fcd1831da0a01e33c9dd5b248f20640d", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The keys are on the kitchen table.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c769791567d583ade96a12a10bbcf8860f21ddc3bbc43ab142ab260f6fa486ca", "generation_seconds_batch": 21.409407417988405, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.144652366638184, "target_z": 1.2343749671805841, "ce": 6.475345611572266, "genuineness": 4.519899368286133, "identity": 0.41854649782180786, "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_CONV_reg/721ab2cd3569fc7e1edb0977bad65273fcd1831da0a01e33c9dd5b248f20640d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "7257f847c94c5f143ce65b8b473ea797f53edee9a4803c593554f132a382ef95", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a3cd403fa770f8c5eb6442dc4bdf2c11909135913bcef3f912d9c5b8a7cbb34d", "generation_seconds_batch": 23.544018284068443, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4109838008880615, "target_z": -0.09375005993110731, "ce": 6.284721851348877, "genuineness": 4.2751336097717285, "identity": 0.5712742805480957, "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_CONV_reg/7257f847c94c5f143ce65b8b473ea797f53edee9a4803c593554f132a382ef95.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": "730247a68002abe6f21bac61a2e15b677862b80bd4e8a896f72abf31effe7e69", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The keys are on the kitchen table.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2cc04ca875a10197dcf7053f57f62d90b567726f5b1427c544d8060b93118e6f", "generation_seconds_batch": 21.409407417988405, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.420590877532959, "target_z": 0.6796875235443636, "ce": 6.427689552307129, "genuineness": 4.758388519287109, "identity": 0.43700629472732544, "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_CONV_reg/730247a68002abe6f21bac61a2e15b677862b80bd4e8a896f72abf31effe7e69.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "7304c4c869b5f23bb3bf4804e9a6da42ae20929b660a6f481b2c211d7ac45f34", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8fb6fdbb6fb393e93f7bdb59e9f998098ab0851b15952e43c2742458d8ff7762", "generation_seconds_batch": 26.55085543403402, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.85910701751709, "target_z": 1.0156250099885178, "ce": 6.125868320465088, "genuineness": 4.607763767242432, "identity": 0.7364281415939331, "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_CONV_reg/7304c4c869b5f23bb3bf4804e9a6da42ae20929b660a6f481b2c211d7ac45f34.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "7363d3c51640680d6f2d0a11c2a5126a73ecdff0e31437cfc51f3b37d13ec741", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f5632f59651afa2c4e1547f9578393d1b9641fcd5e593af0575687ae5200548c", "generation_seconds_batch": 24.06128426408395, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.675542116165161, "target_z": 0.875000011415449, "ce": 6.213237762451172, "genuineness": 4.13078498840332, "identity": 0.5160038471221924, "hypothesis": "The door is till open. I put your coat on the chair. Tell me when you are ready.", "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_CONV_reg/7363d3c51640680d6f2d0a11c2a5126a73ecdff0e31437cfc51f3b37d13ec741.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "7389b49d69219608a7fa3e33c9c457006d755f080b071287ea05e049982b2a49", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d22b950c1aea61ff190adf3b1333638c0ab8c48b20f85e03048397048ecbc52c", "generation_seconds_batch": 18.415036966907792, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5531654357910156, "target_z": 0.7812499514843417, "ce": 6.125868320465088, "genuineness": 4.532451152801514, "identity": 0.5688014626502991, "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_CONV_reg/7389b49d69219608a7fa3e33c9c457006d755f080b071287ea05e049982b2a49.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "73bc1ada84e0d1c276350e036d521893e10fd7f95353d98bebd7f1830dae47d9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3f6c86e9844b99771ea5c5e88acc116e267f47e7f6d43a13da3cd75f0debbe29", "generation_seconds_batch": 25.491545805009082, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.130537509918213, "target_z": -0.30859374179514604, "ce": 6.062326908111572, "genuineness": 3.9299514293670654, "identity": 0.43542519211769104, "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_CONV_reg/73bc1ada84e0d1c276350e036d521893e10fd7f95353d98bebd7f1830dae47d9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "74038634e1291c9b0e9a99d6581485a9957e9199e783fd96302e52e43215ca33", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f9f41463d38a660dc5d28a092037ee32e7600639263a16d0baa12b9baffe7cff", "generation_seconds_batch": 23.758423281949945, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.26702880859375, "target_z": 1.3281248444645073, "ce": 6.109982967376709, "genuineness": 4.344170093536377, "identity": 0.5675057172775269, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. 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_CONV_reg/74038634e1291c9b0e9a99d6581485a9957e9199e783fd96302e52e43215ca33.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "74205913f966981fd1c7e668dfd6fffb463c238992257a06fb9fc4170f4fbe0a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "1f86eafe779f69aada7e0dd12352fae9c6fafcdfc73e579baf5cc36decaf8c88", "generation_seconds_batch": 25.10923226608429, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8897011280059814, "target_z": 1.0390624793094987, "ce": 6.1020402908325195, "genuineness": 4.049196243286133, "identity": 0.39293795824050903, "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_CONV_reg/74205913f966981fd1c7e668dfd6fffb463c238992257a06fb9fc4170f4fbe0a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "7530d1b4f5243668cf1f7fc24c4e5e826e92914f13f842d5ca8fdc105e00b037", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "79dda426edd4570759bb9805ec0eac13afb3779caf46785185328212b79908d5", "generation_seconds_batch": 19.630737451952882, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8494999408721924, "target_z": 0.24218742651304698, "ce": 6.372091293334961, "genuineness": 4.256305694580078, "identity": 0.6136967539787292, "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_CONV_reg/7530d1b4f5243668cf1f7fc24c4e5e826e92914f13f842d5ca8fdc105e00b037.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "7545180def5a087dd022fdbb42eb3acef4acdfff2e2b1d2b89a5256c707af8d5", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4448b84970d4e9d47772d10eb0a184dafe90554b7ede6aa82943baa06857f700", "generation_seconds_batch": 25.480015222914517, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7571263313293457, "target_z": 0.9374998687223364, "ce": 6.038498878479004, "genuineness": 4.369274139404297, "identity": 0.44456928968429565, "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_CONV_reg/7545180def5a087dd022fdbb42eb3acef4acdfff2e2b1d2b89a5256c707af8d5.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "75a75ad85dbb5f05b61e47db23563833ae48178f70b4ae6137166ffef8c32bc3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "bf1a3149ca0bf0df305665ea50a7ff7e1625756d142e9b306fbc729fadb34779", "generation_seconds_batch": 24.94570687203668, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8897011280059814, "target_z": 1.0390624793094987, "ce": 6.006728172302246, "genuineness": 4.400654315948486, "identity": 0.3959740996360779, "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_CONV_reg/75a75ad85dbb5f05b61e47db23563833ae48178f70b4ae6137166ffef8c32bc3.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": "764f4efb43b549962920f6d8e142ba4b4dd9d37bc2e28bc6f18b76f4e51972a2", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "66e2b8b0c9c5f911c0eb45846e3843d8cedacacf6b919a4ce2986162f2853ef9", "generation_seconds_batch": 17.4617559799226, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.05286979675293, "target_z": 1.1640623765704576, "ce": 5.927301406860352, "genuineness": 4.281409740447998, "identity": 0.23109155893325806, "hypothesis": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer en slechtde 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_CONV_reg/764f4efb43b549962920f6d8e142ba4b4dd9d37bc2e28bc6f18b76f4e51972a2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "76ba43c12db8a67078141f6eb1b964a69617b5ec4c1651f5c08f5e8fa26a20bb", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "921a177b1a4524a7ab314844d342f01187f51b2816c30f0678c208f9ecca00a0", "generation_seconds_batch": 22.836119381943718, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.12994647026062, "target_z": 0.45703129102426987, "ce": 6.133810997009277, "genuineness": 4.570107460021973, "identity": 0.7880556583404541, "hypothesis": "J'ai trouvé la petite boîte. Elle était près de la d'un vieil 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_CONV_reg/76ba43c12db8a67078141f6eb1b964a69617b5ec4c1651f5c08f5e8fa26a20bb.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "76ce4280777a0f5b7c1d5c2fae04b0bf92cb70882a0a0994fe1ee864dcaad0b8", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a0541e438e5dcf28ce8ee1d2fdb9532e6de6ad8e92fa123face8cf58a718864b", "generation_seconds_batch": 27.278512349002995, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5429673194885254, "target_z": 0.7734374008282867, "ce": 5.728734970092773, "genuineness": 4.494794845581055, "identity": 0.49161040782928467, "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_CONV_reg/76ce4280777a0f5b7c1d5c2fae04b0bf92cb70882a0a0994fe1ee864dcaad0b8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "76f980de1f4afaa515961bf36e67c978ae291d6950478281624bedd9f8967df5", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e8a7782a6080f929771777a8214cac78f06bc379e59de714ee37a40b81cee9dd", "generation_seconds_batch": 23.185450160992332, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.267620086669922, "target_z": 0.5624999942922755, "ce": 5.935244083404541, "genuineness": 4.7081804275512695, "identity": 0.4721952974796295, "hypothesis": "La puerta sigue 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_CONV_reg/76f980de1f4afaa515961bf36e67c978ae291d6950478281624bedd9f8967df5.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "77893e02aa5ffe0ad76496d657bc51f72a2cbfca71acc3ab8fa9c22e92374da9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "bbed65c78280e5995654aa9553788de481f08dbb56ab563b754bef1d168c0e33", "generation_seconds_batch": 22.68453354993835, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.313511371612549, "target_z": 0.5976562895973387, "ce": 6.125868320465088, "genuineness": 4.595211505889893, "identity": 0.5447519421577454, "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_CONV_reg/77893e02aa5ffe0ad76496d657bc51f72a2cbfca71acc3ab8fa9c22e92374da9.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "13eaf0684ca25cdd084cdcb65df6acb31c70628e19e0c4c7a6801c745b5d488d", "generation_seconds_batch": 16.90704535599798, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.0540521144866943, "target_z": -0.3671875064211901, "ce": 6.1894097328186035, "genuineness": 4.344170093536377, "identity": 0.6272411942481995, "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_CONV_reg/77962898fb95ddc0e92a153772191fe01482e40f98d2245158cb92652e8a93c1.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "77d22ea4c15784946e09c4cd73a42e3bd9393224be5560308a1abca2ee05b9d1", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "68255f0e618b668ab5013dcd44b74d547fc26379b8dc3c4fd402dd7138b2b9ec", "generation_seconds_batch": 24.10360326897353, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.430788993835449, "target_z": 0.6875000742004186, "ce": 6.173524379730225, "genuineness": 4.105680465698242, "identity": 0.3462141454219818, "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_CONV_reg/77d22ea4c15784946e09c4cd73a42e3bd9393224be5560308a1abca2ee05b9d1.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "7820eb42b3af2b5421c5ebf0c747f4212f4482233146204f49935563f7dc5ae1", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d0f5aa559a2fea57e7aacba0970984bde41d6ca71b587585907c292387102bc7", "generation_seconds_batch": 22.441024884930812, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3950958251953125, "target_z": 0.6601563295514103, "ce": 6.252951145172119, "genuineness": 4.607763767242432, "identity": 0.6439687013626099, "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_CONV_reg/7820eb42b3af2b5421c5ebf0c747f4212f4482233146204f49935563f7dc5ae1.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "78c0e166afa8074e529e08bb4615346d9d434de1a801c97c9f54defbdd1a86b0", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "52f4064d34e4eae6437a37d5c20e958b3d55a524022bbecbcff047690287e039", "generation_seconds_batch": 24.254187489044853, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.711235284805298, "target_z": 0.9023437560644573, "ce": 5.839932441711426, "genuineness": 4.720732688903809, "identity": 0.7202380895614624, "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_CONV_reg/78c0e166afa8074e529e08bb4615346d9d434de1a801c97c9f54defbdd1a86b0.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "7917a5765b1c1a4ad2d1f2e2d70c4cce716ca14b1a717b282c0489b04472a899", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "579882ee481bb259ddefb144d0f109d122b713abf75bf014bca3e55fff0545f8", "generation_seconds_batch": 17.900141792953946, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.144652366638184, "target_z": 1.2343749671805841, "ce": 6.086154937744141, "genuineness": 4.846253395080566, "identity": 0.5194001197814941, "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_CONV_reg/7917a5765b1c1a4ad2d1f2e2d70c4cce716ca14b1a717b282c0489b04472a899.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": "792a09e7f5dbc7f60087887639e797b8f8131ee029ca423846dcb8740e0ace17", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "98f379886b5e7fffd0f65fdcbccafad5727efbb9044e9a635a46a1e59ffa9445", "generation_seconds_batch": 15.109293124987744, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6398489475250244, "target_z": 0.8476562667664407, "ce": 6.117925643920898, "genuineness": 4.356721878051758, "identity": 0.616695761680603, "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_CONV_reg/792a09e7f5dbc7f60087887639e797b8f8131ee029ca423846dcb8740e0ace17.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 128, "seconds_generated": 10.24, "mp3_sha256": "960360d5c1adb190bcace6f9017bec31c282997df39307cf3ad500ac10fcff7e", "generation_seconds_batch": 18.527643244946375, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.426280975341797, "target_z": -0.08203123394702483, "ce": 6.149696350097656, "genuineness": 4.331617832183838, "identity": 0.5630402565002441, "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_CONV_reg/794d3b9296e3c4be7c1e4182a2fa3cf095f7d593990763ce0f19c1f5e499a989.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "79eb591f73ac3c526c497ac19919350a5f12b0f3499c945cc076e86d6894345c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1e0b4b5a2dbdc38fce4dc97b1e46f558ba9001a4871af6d20426c3ae31592f1f", "generation_seconds_batch": 18.65636538900435, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8138067722320557, "target_z": 0.2148436818640387, "ce": 5.831989288330078, "genuineness": 4.394378662109375, "identity": 0.4908710718154907, "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_CONV_reg/79eb591f73ac3c526c497ac19919350a5f12b0f3499c945cc076e86d6894345c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "7a6e21d61994d31df1debac3a632b72b6af3771d330179c1bbfc42654eb572ce", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4d1e8a4abfa9ebbdb66e8e13a2a09c7e9af231d6cfe8b2248f21a85daf7ad288", "generation_seconds_batch": 23.631091909017414, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6965293884277344, "target_z": 0.12500007990814307, "ce": 5.816103935241699, "genuineness": 3.948779582977295, "identity": 0.5783135294914246, "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_CONV_reg/7a6e21d61994d31df1debac3a632b72b6af3771d330179c1bbfc42654eb572ce.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": "7b237eff7f7d66640de96dc8ebb474ca2c9acda37b47cfd725499f22e821bb80", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "77ef2f6c211b10f74ac2196c10c8bc9bae680c19c6bd9b111503db009d4231dc", "generation_seconds_batch": 21.832873945008032, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7041778564453125, "target_z": 0.13085940157659223, "ce": 5.7247633934021, "genuineness": 4.507347106933594, "identity": 0.5231444835662842, "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_CONV_reg/7b237eff7f7d66640de96dc8ebb474ca2c9acda37b47cfd725499f22e821bb80.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "7b981d0003305445f8d4fc3c8f7807f82ba822fe103923b02ee576b9b3029027", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "edd748d5dcbe062d22c0114c91271bc7fef5c79f74849ebee187f9d55b6d3562", "generation_seconds_batch": 24.446531435940415, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9406914710998535, "target_z": 1.0781250499425894, "ce": 6.284721851348877, "genuineness": 4.494794845581055, "identity": 0.3385546803474426, "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_CONV_reg/7b981d0003305445f8d4fc3c8f7807f82ba822fe103923b02ee576b9b3029027.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "7ba256e51e43ad47a7bd3a94b29acfcb1ffc10c2fd08a0025beb92959526300c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "98a1e8ac56b5000c2bfbcbf276ab11be6c7fbdbb7c38b9b9987242b47e26c688", "generation_seconds_batch": 18.0278738590423, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.154850482940674, "target_z": 1.242187517836639, "ce": 5.911416053771973, "genuineness": 4.545003414154053, "identity": 0.45581796765327454, "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_CONV_reg/7ba256e51e43ad47a7bd3a94b29acfcb1ffc10c2fd08a0025beb92959526300c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "7be9918c043f32719d43ba2a3f6c35dd65c7e7ff2adbeb4a5ba38bc794b381b9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4a84975d6da1a496f2dc94b46edf4ebe4a4ba339f51367d730329559f37c7fd2", "generation_seconds_batch": 23.665498134098016, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6959381103515625, "target_z": 0.8906249300803748, "ce": 6.117925643920898, "genuineness": 4.356721878051758, "identity": 0.3142283260822296, "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_CONV_reg/7be9918c043f32719d43ba2a3f6c35dd65c7e7ff2adbeb4a5ba38bc794b381b9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "7cd9c872734ba335c722956c76d7cae778cd660c1b2a28befdc94b6a374bc766", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "90c229df8fd20effaeeb87b0484bd7c60b7ffb1490e4f49345996b9061941c7f", "generation_seconds_batch": 18.20648849301506, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8074331283569336, "target_z": 0.20996097468939248, "ce": 5.704906940460205, "genuineness": 4.457139015197754, "identity": 0.6408780217170715, "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_CONV_reg/7cd9c872734ba335c722956c76d7cae778cd660c1b2a28befdc94b6a374bc766.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": "7d674a7d0288a022059d88bc64da7e62eccd916b87222913edebaa5d732b69cb", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4d5cec1be676c05ddf81f3ca045f8cacbdbd7700b0fe0f3011f6103fa0a67070", "generation_seconds_batch": 17.65268068900332, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.277818202972412, "target_z": 0.5703125449483305, "ce": 6.117925643920898, "genuineness": 4.507347106933594, "identity": 0.14722970128059387, "hypothesis": "De sleutels liggen op de keekentafel. 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_CONV_reg/7d674a7d0288a022059d88bc64da7e62eccd916b87222913edebaa5d732b69cb.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "7dd7023cf64fd6278ec5512d7430feef82cc71e841be9e5cef296d8878576b20", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "b5b774d8ee69c2faaf7757da38a77cfb1c4c7a0896769da49e1e5c2a1312c60e", "generation_seconds_batch": 24.757390280952677, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.517472267150879, "target_z": 0.7539062068353334, "ce": 6.181467056274414, "genuineness": 4.344170093536377, "identity": 0.233122318983078, "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_CONV_reg/7dd7023cf64fd6278ec5512d7430feef82cc71e841be9e5cef296d8878576b20.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "7eb840ed72a44d53485269044de5c1e668bc66fe9addfa6e390f604f4caf7a29", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7fd5a3934a939117c622df699f1e46b80fc1734a466b95752348c6d5f32ba61e", "generation_seconds_batch": 22.70547784899827, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1452434062957764, "target_z": 0.4687499343611682, "ce": 5.80021858215332, "genuineness": 4.356721878051758, "identity": 0.6568372845649719, "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_CONV_reg/7eb840ed72a44d53485269044de5c1e668bc66fe9addfa6e390f604f4caf7a29.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "7f328381d57a5887e9605623acc0288f1035bd2f3f4f09e07e828eb227ac2fa9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "63dc643e5f81605731eb04c4194c63a4523d21c65b8cc7e6a5cc5e708ce81609", "generation_seconds_batch": 22.7073774819728, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.06306791305542, "target_z": 1.1718749272265125, "ce": 5.911416053771973, "genuineness": 4.093128204345703, "identity": 0.31840091943740845, "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_CONV_reg/7f328381d57a5887e9605623acc0288f1035bd2f3f4f09e07e828eb227ac2fa9.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": "7f6925bd6b339f33821425583dca51abf57d8e89edf841e931df74da5e03f6db", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Escribí el número en el sobre.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "108a063cc077815c4f7b64cc220f52bed85f9abec48584499b6768eb52a6f6aa", "generation_seconds_batch": 15.024231143994257, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6990787982940674, "target_z": 0.12695312624856472, "ce": 6.308549880981445, "genuineness": 4.419482707977295, "identity": 0.5956829786300659, "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_CONV_reg/7f6925bd6b339f33821425583dca51abf57d8e89edf841e931df74da5e03f6db.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "2bd52ddd836b2a97ba35f61aca2c412863cb4bfc269a9b5d63eeff8c16ed4170", "generation_seconds_batch": 19.46237588895019, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2064318656921387, "target_z": 0.5156250556503139, "ce": 6.395919322967529, "genuineness": 4.482243061065674, "identity": 0.3031221032142639, "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_CONV_reg/7f6fce1a6f733719f45a363f28fee4e24b2f07a3396aa07b42392913d7c24239.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "7fbf75245af66d4db4bafded1fbc73a724f18ab181fbd7bbeaae99132b8ba319", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ca31e26b0622778267f6938339cde07ceac93fce3f3b833efc6cef1d3a7d8e01", "generation_seconds_batch": 23.631091909017414, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.78448748588562, "target_z": 0.19238282703686085, "ce": 5.879645347595215, "genuineness": 4.099404335021973, "identity": 0.6222563982009888, "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_CONV_reg/7fbf75245af66d4db4bafded1fbc73a724f18ab181fbd7bbeaae99132b8ba319.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "7fc651cb0e4f278c74b88b4840a29aad5f86b20e60747f254e3f5ff7883da59d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7c4ca75c81d38be631bb08a5a9feb68db2671b3a288c6220f2b13142280fb36f", "generation_seconds_batch": 21.994029377005063, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.931084394454956, "target_z": 0.30468746646711853, "ce": 5.649308204650879, "genuineness": 4.118232727050781, "identity": 0.4427604079246521, "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_CONV_reg/7fc651cb0e4f278c74b88b4840a29aad5f86b20e60747f254e3f5ff7883da59d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "8017ed16baecbbad301ef9796db4295f2b6132a754a338edf01e2ca5e565c71c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "14e082e4c54bcb10bb015bf18b22223e330cfcaefa3ed3998ce9869ab7da83a3", "generation_seconds_batch": 24.425943535985425, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.318610429763794, "target_z": 0.6015625649253663, "ce": 6.181467056274414, "genuineness": 4.457139015197754, "identity": 0.6429761648178101, "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_CONV_reg/8017ed16baecbbad301ef9796db4295f2b6132a754a338edf01e2ca5e565c71c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "802ac65e663bd0b8d5964df63d37f050de59b94c633de6a5927ff3a83792a034", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "4298472a64eb32dddc0b719bfa89131a69dc437671496b3721d473e183aa9949", "generation_seconds_batch": 26.02799546904862, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3026294708251953, "target_z": -0.17675790837193506, "ce": 5.895530700683594, "genuineness": 4.3127899169921875, "identity": 0.6135126948356628, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "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_CONV_reg/802ac65e663bd0b8d5964df63d37f050de59b94c633de6a5927ff3a83792a034.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "806a93251f93f5217a166b1cd65599ce240739b1470d9c91fef5b00a84e27b93", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "c5078c7337e0dec31def940e2ae63fcea5d59c178f45c6a324f069cc224ecc25", "generation_seconds_batch": 25.497869760962203, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.313511371612549, "target_z": 0.5976562895973387, "ce": 6.205295085906982, "genuineness": 4.055472373962402, "identity": 0.32219088077545166, "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_CONV_reg/806a93251f93f5217a166b1cd65599ce240739b1470d9c91fef5b00a84e27b93.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "80a79a31bc528b43ce2a58ee1a097d813a5ac5f8b812702d2eaeca5a9447b788", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "de6415708b298c12b255f9b9a18c97237b47893b96fe1ee4351333d032cb06e8", "generation_seconds_batch": 24.975468205986544, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.441577911376953, "target_z": -0.07031259061012653, "ce": 6.1894097328186035, "genuineness": 4.2751336097717285, "identity": 0.7677356600761414, "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_CONV_reg/80a79a31bc528b43ce2a58ee1a097d813a5ac5f8b812702d2eaeca5a9447b788.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "a5f4619d9000be04208e2c40ee776998eaeca4d9ab79f7450c030b5aeca2df1a", "generation_seconds_batch": 17.610762433963828, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.272719144821167, "target_z": 0.5664062696203029, "ce": 6.165581703186035, "genuineness": 4.2751336097717285, "identity": 0.7417128682136536, "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_CONV_reg/80c2597a6b8f43b5338a6b37d654868d679f46ded48514b42837ab18fba327ad.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": "80c66c115d7d113d8136e3fdb5a9cadda7901b945760d4cd26a9310b0deca986", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2d45c509427e4a2faced46bda72d388d3ed71f29d556af20b2bc636bd9e608d9", "generation_seconds_batch": 22.032627566950396, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.022275924682617, "target_z": 1.140625089896661, "ce": 6.141753673553467, "genuineness": 4.607763767242432, "identity": 0.6547631621360779, "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_CONV_reg/80c66c115d7d113d8136e3fdb5a9cadda7901b945760d4cd26a9310b0deca986.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "80cf5d6e245905b2da17b9eddc748fffc5d13429baf93404694c91393e7bc17e", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "9a646bf02ea70f0a6f6c16b012b05195a9b0f4e3713cd0b155d83cf2be4c8894", "generation_seconds_batch": 17.123980875941925, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.431380033493042, "target_z": -0.07812495861899733, "ce": 6.1020402908325195, "genuineness": 3.911123275756836, "identity": 0.3890264928340912, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "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_CONV_reg/80cf5d6e245905b2da17b9eddc748fffc5d13429baf93404694c91393e7bc17e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "813220e92d5b78af3eb6f2617ff322bed22fc0a967daf7b0fa41179622677358", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "35401df84bd8e6f5bb4b68fa6de19c6ff0b9eb6d5d60244868f23486566cabbc", "generation_seconds_batch": 23.963112149038352, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.563363552093506, "target_z": 0.7890625021403966, "ce": 6.006728172302246, "genuineness": 4.545003414154053, "identity": 0.5553321838378906, "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_CONV_reg/813220e92d5b78af3eb6f2617ff322bed22fc0a967daf7b0fa41179622677358.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "817e5d6de784a16943f099660116a135d0eaa63d44c245efdba3fc5074ecf96e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "3ff11bcb763b37ceccfc7fd85f88160f8a9d3de22af77875d0384e0b8c79452f", "generation_seconds_batch": 24.221836103941314, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.767324447631836, "target_z": 0.9453124193783914, "ce": 5.998785495758057, "genuineness": 4.457139015197754, "identity": 0.19799400866031647, "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_CONV_reg/817e5d6de784a16943f099660116a135d0eaa63d44c245efdba3fc5074ecf96e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "8193cda1138e91e0ba6df3c0bbee750b998e4de16ae1a543951d083ee3a5ceeb", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b8d4fb47f301d01f1a3389ee58034b529e0db46deecc78bba83d5914a2f4b4e4", "generation_seconds_batch": 17.146196016925387, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8112573623657227, "target_z": 0.21289063552361706, "ce": 6.24500846862793, "genuineness": 4.068024158477783, "identity": 0.3222147822380066, "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_CONV_reg/8193cda1138e91e0ba6df3c0bbee750b998e4de16ae1a543951d083ee3a5ceeb.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "81a06a549def61958d4079d8c19ac0c26a9fc7abb8b4edbcc93cf66b5774d177", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "f0d8b028a49205658b67fec9ceacc21c05d7f957fd9b34fb5b10633f9d03270b", "generation_seconds_batch": 21.84299379598815, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.216629981994629, "target_z": 0.5234376063063689, "ce": 5.974957466125488, "genuineness": 4.344170093536377, "identity": 0.7586100101470947, "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": 0.03999999999999915, "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_CONV_reg/81a06a549def61958d4079d8c19ac0c26a9fc7abb8b4edbcc93cf66b5774d177.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "8240c3c53e6e7189dcd08fbd47be3937836fb908d3d763b08397929ad6c4ba8f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "3a6cb1a320df3f718cb348a7afad250a4dce9c3a8bca76a82fc92bf785cbdc4c", "generation_seconds_batch": 23.43021843000315, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8584232330322266, "target_z": 0.24902336267529906, "ce": 6.149696350097656, "genuineness": 4.118232727050781, "identity": 0.38419973850250244, "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_CONV_reg/8240c3c53e6e7189dcd08fbd47be3937836fb908d3d763b08397929ad6c4ba8f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "83807afda8e498c4d60799cec2187056d5db6b0e1c63c248a118efbb9b7c96ab", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0b09d3ab84ba24ed67085d12cc72b1ec78e49fd3b05514e46f0e142992255038", "generation_seconds_batch": 23.15912739909254, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2880163192749023, "target_z": 0.5781250956043855, "ce": 6.252951145172119, "genuineness": 4.595211505889893, "identity": 0.5146334171295166, "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_CONV_reg/83807afda8e498c4d60799cec2187056d5db6b0e1c63c248a118efbb9b7c96ab.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "8391c2a398e11671c913988dfe03ec3c0ccb78ce946caf08d8553d08f15789ab", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9dd443243dc7ab9f025fe18b748dafe215b23af466711a1c8e8f60385211144d", "generation_seconds_batch": 21.994029377005063, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.982074737548828, "target_z": 0.3437500371002093, "ce": 5.831989288330078, "genuineness": 4.331617832183838, "identity": 0.5645390152931213, "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_CONV_reg/8391c2a398e11671c913988dfe03ec3c0ccb78ce946caf08d8553d08f15789ab.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "83b2325e3767ed12842ab2b1f06587688a56536e34bec1726da3e0a414ecf67c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b53c403a81f77703ca71b863d364913f57172881ba19f4becb31acad881880fd", "generation_seconds_batch": 24.86580517294351, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.297623157501221, "target_z": 1.3515624964326722, "ce": 6.022613525390625, "genuineness": 4.243753433227539, "identity": 0.6236236095428467, "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_CONV_reg/83b2325e3767ed12842ab2b1f06587688a56536e34bec1726da3e0a414ecf67c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "83d02f0e14124c3a797e5277c5b1ffb9623ef7741c67d452eb24bd289fe68d81", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4a2b80a4ea03b03007adb82b2745255ee51fe037cb2104b38d6ac2a372acac5f", "generation_seconds_batch": 23.58173772203736, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3217508792877197, "target_z": -0.162109421553628, "ce": 5.8875885009765625, "genuineness": 4.055472373962402, "identity": 0.0063108280301094055, "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_CONV_reg/83d02f0e14124c3a797e5277c5b1ffb9623ef7741c67d452eb24bd289fe68d81.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": "83e7ad33987792fd734cef3897ea0c00152f5104d880b227362779e3a412eaab", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "4afd21882747a10b509c286767fe2b8a4c905e7832443bb2aace60d71c79539b", "generation_seconds_batch": 21.38254901196342, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2982141971588135, "target_z": 0.5859374636132563, "ce": 6.197352409362793, "genuineness": 4.369274139404297, "identity": 0.5997676849365234, "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_CONV_reg/83e7ad33987792fd734cef3897ea0c00152f5104d880b227362779e3a412eaab.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": "8491504198aa924da78a9940955e2d346f9010a3bdb71f9ef9ef09f7a52fb549", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "57fa45254c38c6e3701606b30b99aa6cb52b1da0745b6c762f94a8a895af49b8", "generation_seconds_batch": 15.237024097004905, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.030515432357788, "target_z": 0.3808593787456942, "ce": 6.070269584655762, "genuineness": 4.557555198669434, "identity": 0.7699447274208069, "hypothesis": "J'ai écrit le numéro sur l'enveloppe. Lis-nous 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_CONV_reg/8491504198aa924da78a9940955e2d346f9010a3bdb71f9ef9ef09f7a52fb549.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "8494e36cca2b2878100a5e247795bd3e9c48cd4a91267bda0d6c2064f9e1a082", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b0d6d26ca7da91fa2121260243e2aee9f83d4659bcd9be33688ca2554d805080", "generation_seconds_batch": 17.130912492983043, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1962337493896484, "target_z": 0.5078125049942589, "ce": 6.221180438995361, "genuineness": 4.494794845581055, "identity": 0.6313878893852234, "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_CONV_reg/8494e36cca2b2878100a5e247795bd3e9c48cd4a91267bda0d6c2064f9e1a082.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "850f4dacc72eb9e9d074c92530936f59bd97414ccc9985edbe4dc4968989c18a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "575fbf9684e2ab4eef8b6822db0d246ec6197ad55d766d2db5611e6258e011af", "generation_seconds_batch": 22.055877617909573, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1152403354644775, "target_z": -0.3203125677792285, "ce": 6.38003396987915, "genuineness": 4.281409740447998, "identity": 0.7325781583786011, "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_CONV_reg/850f4dacc72eb9e9d074c92530936f59bd97414ccc9985edbe4dc4968989c18a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "853b836282ff252ce33d6471c27fa236ecc5fc81ccc1aa267cd1298049c991aa", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b4c7a3bd2139121435a232c4bb279068c6c4403d8900cea616c9aff5ff35db10", "generation_seconds_batch": 16.64608475205023, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6876060962677, "target_z": 0.11816414374589101, "ce": 6.205295085906982, "genuineness": 4.482243061065674, "identity": 0.7062646746635437, "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_CONV_reg/853b836282ff252ce33d6471c27fa236ecc5fc81ccc1aa267cd1298049c991aa.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "869bcb637b4af60a3f36efa9ee974332a63809e154f0f14d65da3b4897665179", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ae827111a9e2882400b59789beb38f50d291b4a48098a5bf37fc397f91cc2788", "generation_seconds_batch": 18.65636538900435, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5665042400360107, "target_z": 0.025390698308586618, "ce": 5.935244083404541, "genuineness": 4.15588903427124, "identity": 0.5654175877571106, "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_CONV_reg/869bcb637b4af60a3f36efa9ee974332a63809e154f0f14d65da3b4897665179.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "86c20e84cfac53f4a0f324ff3590467303a5976e1fc0eb52f7a7739ff861717d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ae032192c850b19b8a7d5cb3ec1bab1d2149b7db00080b2a62e96b888886f713", "generation_seconds_batch": 27.278512349002995, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2013328075408936, "target_z": 0.5117187803222865, "ce": 6.022613525390625, "genuineness": 4.356721878051758, "identity": 0.23528452217578888, "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_CONV_reg/86c20e84cfac53f4a0f324ff3590467303a5976e1fc0eb52f7a7739ff861717d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "870706f7897804473bdbd57f2a3458c1991cc71fc2c89be7c225cecb3591c765", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b1c8c6aa9048b729510a7ff835b20311b33f4c18f4b1377d2f7e994e00593d8e", "generation_seconds_batch": 27.001158267026767, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3390064239501953, "target_z": 0.6171874835902921, "ce": 5.990842819213867, "genuineness": 4.532451152801514, "identity": 0.6648142337799072, "hypothesis": "The trouver la petite boîte. It is pretty la vieille fenêtre. We've regarded ensemble what it contained.", "wer": 0.631578947368421, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/870706f7897804473bdbd57f2a3458c1991cc71fc2c89be7c225cecb3591c765.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "87c6d5daa004b48b086c291b4c6bc83a065182a9dc93b6a2b6947964760942ec", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6b0207c075511b49fbd6defada3d18f0825ad6f7cd627a41bc7e14eb3b552bcf", "generation_seconds_batch": 21.583940683980472, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6398489475250244, "target_z": 0.8476562667664407, "ce": 6.006728172302246, "genuineness": 4.337893962860107, "identity": 0.6550307273864746, "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_CONV_reg/87c6d5daa004b48b086c291b4c6bc83a065182a9dc93b6a2b6947964760942ec.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "87e0826a6c4473a9a4ade7aabd7d4f4f0b96ec5ab532e0e692434e7d447da685", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "8c944746ae9b47b7f94c8c74fd19a803e275c03eb92ceb32c4a18bb4fe5d7235", "generation_seconds_batch": 23.758423281949945, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.030515432357788, "target_z": 0.3808593787456942, "ce": 6.205295085906982, "genuineness": 4.356721878051758, "identity": 0.17666754126548767, "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_CONV_reg/87e0826a6c4473a9a4ade7aabd7d4f4f0b96ec5ab532e0e692434e7d447da685.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "88451f16d2ade8d2a78a9e43c23751958007178a8dbe1c0c739c0e463212e415", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ccb861a5c8a9e80cd37785ef336cb0d17b8135da51be379c42e01f2527c9db9c", "generation_seconds_batch": 18.479565732064657, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.195642471313477, "target_z": 1.2734373551664906, "ce": 6.213237762451172, "genuineness": 4.6956281661987305, "identity": 0.6280186176300049, "hypothesis": "Zelou! 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_CONV_reg/88451f16d2ade8d2a78a9e43c23751958007178a8dbe1c0c739c0e463212e415.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "885b8e38c02624590d354630f0eb5a57ebfc01dc2e15bf90cd99c7fdbbc9e0c8", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7e68aec23be3c13a0755632398d8de88819922e887aa93ceed198c1975c1c26a", "generation_seconds_batch": 23.41782316903118, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6347498893737793, "target_z": 0.8437499914384132, "ce": 5.685050010681152, "genuineness": 4.337893962860107, "identity": 0.5031724572181702, "hypothesis": "Con toda la cajita. Estaba junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.11764705882352941, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/885b8e38c02624590d354630f0eb5a57ebfc01dc2e15bf90cd99c7fdbbc9e0c8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "88c4c6181fa5310fb75054bb9408ee5c73c476d714de1c74eec91abcc721f89d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0c5054e90dec930ed05e7ba914c585229b0b49ca8157747d2c02152cc38d06ea", "generation_seconds_batch": 24.187549999915063, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.732222557067871, "target_z": 0.15234382455715134, "ce": 6.054384231567383, "genuineness": 4.143336772918701, "identity": 0.5771042108535767, "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_CONV_reg/88c4c6181fa5310fb75054bb9408ee5c73c476d714de1c74eec91abcc721f89d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "89034ed6670488d972b454fc388d3f038f119228229dbb8138959450089fee60", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "e2e5362eb3cb8514e6fe69a4274806f8ba47064c67b3af13265398953194ca5e", "generation_seconds_batch": 18.87660593900364, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7411458492279053, "target_z": 0.15917976071940343, "ce": 6.1894097328186035, "genuineness": 4.15588903427124, "identity": 0.28843826055526733, "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_CONV_reg/89034ed6670488d972b454fc388d3f038f119228229dbb8138959450089fee60.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "8a109f8ea7fa7d92b728f6ded235d8556e482c4a25feacba9f7cdafaf1845efe", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b9098bb4356c8912991746b4bf099e06176afcfe37dc1ed116627fc29b5a306e", "generation_seconds_batch": 21.0754584800452, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.441577911376953, "target_z": -0.07031259061012653, "ce": 6.4912309646606445, "genuineness": 4.331617832183838, "identity": 0.5238128304481506, "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_CONV_reg/8a109f8ea7fa7d92b728f6ded235d8556e482c4a25feacba9f7cdafaf1845efe.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "8a71c52c9b8e8d62071fa83fc41e2aeaeef3110a339decea8c3f14715b3dc783", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "49d5edafe4b77d091f531fb0d741299770d436ac882157213d4c48e415bf2f15", "generation_seconds_batch": 23.15912739909254, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.358811378479004, "target_z": 1.3984374350746338, "ce": 6.133810997009277, "genuineness": 4.482243061065674, "identity": 0.4896658658981323, "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_CONV_reg/8a71c52c9b8e8d62071fa83fc41e2aeaeef3110a339decea8c3f14715b3dc783.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "8a765137126890e2b991776fbe061f5d0255c1ea2cda173b707a2526e4f49adf", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a575e85b0054a26658a216021c3bfb8d7be2986e7d6a33b5ccf8967664fa5d27", "generation_seconds_batch": 23.63493035105057, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3390064239501953, "target_z": 0.6171874835902921, "ce": 6.125868320465088, "genuineness": 4.406930446624756, "identity": 0.6910041570663452, "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_CONV_reg/8a765137126890e2b991776fbe061f5d0255c1ea2cda173b707a2526e4f49adf.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "8b7cadd7749021bb4d0dbf156eebd211e7908cc235a3f674ffb7825a2fb6e769", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "2b0033b1afd1d1a0ca2de8189253959b3840800d4f20eaaeea7e74c9bcf226f4", "generation_seconds_batch": 21.996297345962375, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8387107849121094, "target_z": 0.9999999086764079, "ce": 6.157639026641846, "genuineness": 4.331617832183838, "identity": 0.2592441439628601, "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_CONV_reg/8b7cadd7749021bb4d0dbf156eebd211e7908cc235a3f674ffb7825a2fb6e769.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "8bc7463aa1b6b3bbe36e4c4d120b111d4321e0791ed3db3fc43cadb785763099", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b9e1ec523d8aeeaabc0380f709d7f65f2b875ffe7cb2b4cd6ef77120db57a685", "generation_seconds_batch": 25.685742989066057, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.450593948364258, "target_z": 1.4687500256847603, "ce": 5.903473377227783, "genuineness": 4.356721878051758, "identity": 0.48437467217445374, "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_CONV_reg/8bc7463aa1b6b3bbe36e4c4d120b111d4321e0791ed3db3fc43cadb785763099.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "8bed214f9578428e80edb6e0507987eeb32c5b44f80219458fdbbc74fb9f6369", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "db018ad3267478dd2c913fd80ed81ffda1727bffe057d50d82b0b81bfe8b2b8c", "generation_seconds_batch": 23.19873085594736, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6869685649871826, "target_z": 0.11767574517539746, "ce": 6.38003396987915, "genuineness": 4.319066047668457, "identity": 0.6019691228866577, "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_CONV_reg/8bed214f9578428e80edb6e0507987eeb32c5b44f80219458fdbbc74fb9f6369.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "8c3a1b67b7cab237f8c78881a8cbadd99ecb14970d82e336c64b3f3e4de6efbc", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "41e488bb7f576f96a2a9ea4b2e5807a42f84d5d1f1f6eb8e6c220b213506883c", "generation_seconds_batch": 26.323338488931768, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.716334342956543, "target_z": 0.9062500313924848, "ce": 6.332377910614014, "genuineness": 4.632867813110352, "identity": 0.33305680751800537, "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_CONV_reg/8c3a1b67b7cab237f8c78881a8cbadd99ecb14970d82e336c64b3f3e4de6efbc.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "8d037f4bfc9ffb4ba7998e29514b7126a7603173e3c3a98bace76e34f5568000", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "0e072272c6c01b881625b39c19d272a0f1d151d34ea688e7f6f0aba3845c43cf", "generation_seconds_batch": 24.669284758041613, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8795032501220703, "target_z": 1.0312501113006278, "ce": 5.831989288330078, "genuineness": 4.645420074462891, "identity": 0.6395477056503296, "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_CONV_reg/8d037f4bfc9ffb4ba7998e29514b7126a7603173e3c3a98bace76e34f5568000.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "8d13fef09930ae4294c304698a6ad141935105d1d63eb6ac30f497b09be879d4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "76c6b91a279f9ca0fc4bbed3218357e6c8b182056fac04bd50165e2df3618a37", "generation_seconds_batch": 19.196984502952546, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.379798650741577, "target_z": 0.6484375035673278, "ce": 5.855817794799805, "genuineness": 4.20609712600708, "identity": 0.6828893423080444, "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_CONV_reg/8d13fef09930ae4294c304698a6ad141935105d1d63eb6ac30f497b09be879d4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "8d93267a3daa7ab8d824dd9f05b45c3f44ebeb934963661ae174761a85bc7971", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e817e81b106c5f85f0257546e339f37661fdd34d4c83a265d224c30846af4c62", "generation_seconds_batch": 23.972920401021838, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.430788993835449, "target_z": 0.6875000742004186, "ce": 6.133810997009277, "genuineness": 4.758388519287109, "identity": 0.7519353628158569, "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_CONV_reg/8d93267a3daa7ab8d824dd9f05b45c3f44ebeb934963661ae174761a85bc7971.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "8d9ca1f22b390ef317a0704b86df88104bc5d5ffb679903b3e85ebfaeb4ba97d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d9bc35bf95e638da45810ad1896d8a69b18ebcbdd9314a0037d414a63231e9ec", "generation_seconds_batch": 24.24352619005367, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.140735626220703, "target_z": -0.300781191139091, "ce": 6.38797664642334, "genuineness": 4.042920112609863, "identity": 0.5606143474578857, "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_CONV_reg/8d9ca1f22b390ef317a0704b86df88104bc5d5ffb679903b3e85ebfaeb4ba97d.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": "8e2073262655edb65b27bab7de968c14af0291228ca1df4fe662f5502fc133ff", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "ba2b4f89448025fc6d96d4b812afac9ffdb80dadf75c3d8797f9938e37c38626", "generation_seconds_batch": 18.13911257009022, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6347498893737793, "target_z": 0.8437499914384132, "ce": 5.9431867599487305, "genuineness": 4.582659721374512, "identity": 0.11258165538311005, "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": 0.03999999999999915, "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_CONV_reg/8e2073262655edb65b27bab7de968c14af0291228ca1df4fe662f5502fc133ff.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "8e67f47cf2b0b07caf51d40e93b756114d394aaaf56a3f35f5f6fe8907a31628", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "b5d034db8ec817c1f280ca44e0ce8957827d2a1ad8d2703dd89456b76fba3e1d", "generation_seconds_batch": 22.93993466498796, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.103860378265381, "target_z": 1.2031251298507326, "ce": 5.808161735534668, "genuineness": 4.6705241203308105, "identity": 0.6189934611320496, "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": 0.03999999999999915, "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_CONV_reg/8e67f47cf2b0b07caf51d40e93b756114d394aaaf56a3f35f5f6fe8907a31628.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "8e7605fc13d64f49b1039a8c4abacbcad7d814f19c398bf309f632514b5d2858", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7ecad022949d771b164a45720924228aa1f675bf3d03210eddeeb48c409cafd8", "generation_seconds_batch": 24.164906044956297, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.338415145874023, "target_z": 1.3828123337625238, "ce": 5.561938762664795, "genuineness": 4.607763767242432, "identity": 0.5812373161315918, "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_CONV_reg/8e7605fc13d64f49b1039a8c4abacbcad7d814f19c398bf309f632514b5d2858.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "8e87b0fc226ed8f9d7cc06d3cf5839f54240662ed058b8eaf2ff8ca871c826a9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cfe2a0ebabd506b3639a03b1dc61d65b73bc6a598ee0a9558670215fcc04649f", "generation_seconds_batch": 26.117432302911766, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4109838008880615, "target_z": -0.09375005993110731, "ce": 5.831989288330078, "genuineness": 4.005263805389404, "identity": 0.600787878036499, "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_CONV_reg/8e87b0fc226ed8f9d7cc06d3cf5839f54240662ed058b8eaf2ff8ca871c826a9.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": "8eb8c0ecb07af54ebe4f1e19c87eff6fff30c5dfbd1ffcf6c6a74e97eeebf172", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I wrote the number on the envelope.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c350847cb25572f269bdaa74a94d4ffc45c34a41fe757d8a9413eed730d449b0", "generation_seconds_batch": 23.703053252073005, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3390064239501953, "target_z": 0.6171874835902921, "ce": 6.356205940246582, "genuineness": 4.105680465698242, "identity": 0.23835942149162292, "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_CONV_reg/8eb8c0ecb07af54ebe4f1e19c87eff6fff30c5dfbd1ffcf6c6a74e97eeebf172.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "8f0d66e4b3d1d42810d662a0d0b2c73362fc30d12bb1bf6a15b461606f885dd5", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "fc76f9ab3af61d95737452ada99b9184f5e0ea96e37a2e3473911e18b057c4cd", "generation_seconds_batch": 22.218144059064798, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6831443309783936, "target_z": 0.11474608434117288, "ce": 6.38797664642334, "genuineness": 4.143336772918701, "identity": 0.5137148499488831, "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_CONV_reg/8f0d66e4b3d1d42810d662a0d0b2c73362fc30d12bb1bf6a15b461606f885dd5.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "904874e482925d812fc68a2b953cc5257cac40992a494c32db96c89f756e80e4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "b4b99dd7d8223f915cd84c0c510ed59de89cf5182a48d30519c045a9d1e0d33a", "generation_seconds_batch": 18.491748607950285, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.195642471313477, "target_z": 1.2734373551664906, "ce": 6.23706579208374, "genuineness": 4.482243061065674, "identity": 0.25115281343460083, "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_CONV_reg/904874e482925d812fc68a2b953cc5257cac40992a494c32db96c89f756e80e4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "9081c4b202fdae5ecf6e67e85fbd327d142a5aa7c18f9e9a7bf67eb7896a6b4c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "dd43f201ac3fc4479c8fb6240b119ae8810cf39e29a69e15ed134074ce0090e1", "generation_seconds_batch": 17.37831445701886, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8622474670410156, "target_z": 0.25195302350952364, "ce": 6.157639026641846, "genuineness": 4.325342178344727, "identity": 0.1439409852027893, "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_CONV_reg/9081c4b202fdae5ecf6e67e85fbd327d142a5aa7c18f9e9a7bf67eb7896a6b4c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "90ab7091846bc3541e035c311f08e20a1d8db0dede4749cc6780de7c9954df78", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "440897ca01c6c203be65db90d23356d969886f4e8e6aa356b7b04b00199f236d", "generation_seconds_batch": 24.980389069998637, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9814834594726562, "target_z": 1.109374887272441, "ce": 6.435632705688477, "genuineness": 4.369274139404297, "identity": 0.27742236852645874, "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_CONV_reg/90ab7091846bc3541e035c311f08e20a1d8db0dede4749cc6780de7c9954df78.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "920de89ddee3741b563ffd7e2391280fff054dd7e1f8dee871924bcb8a47200a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "285632b285541fe5b9c31fff7a663083e62e6c94f426b1387c29a31d417265bc", "generation_seconds_batch": 24.684078398975544, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.817631244659424, "target_z": 0.21777352534544747, "ce": 5.919358730316162, "genuineness": 4.017816066741943, "identity": 0.3419020175933838, "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_CONV_reg/920de89ddee3741b563ffd7e2391280fff054dd7e1f8dee871924bcb8a47200a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "92748445f4ba740be4afacd01449805d4f81c412a3b3206f3f5ed97873ef30d5", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "169fcd8f8854781381ebe437b7401b9b59d309f2e93d86660f5e02614ef45f88", "generation_seconds_batch": 17.76990672107786, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.563363552093506, "target_z": 0.7890625021403966, "ce": 6.23706579208374, "genuineness": 4.268857479095459, "identity": 0.27105897665023804, "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_CONV_reg/92748445f4ba740be4afacd01449805d4f81c412a3b3206f3f5ed97873ef30d5.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "92803bc90520257005f377a669f4bb18300aa87ec793f7beb1d92c10b2e76701", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "a337674c3bce8319c3cc1639be9c0165c48fd6495b76d2553b48ce630ba94eac", "generation_seconds_batch": 24.049307239009067, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3899967670440674, "target_z": 0.6562500542233828, "ce": 5.974957466125488, "genuineness": 4.287685871124268, "identity": 0.32419729232788086, "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_CONV_reg/92803bc90520257005f377a669f4bb18300aa87ec793f7beb1d92c10b2e76701.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "92adc297127dc6fa73dc62e71dbddba838227aa87f45e096fc2532e9e45863fb", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "94019847ce830f5d490402dad1f6d96062966896d8d399c2ec1a15db5ca1591e", "generation_seconds_batch": 26.127858088933863, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.386763572692871, "target_z": -0.11230463943025767, "ce": 6.332377910614014, "genuineness": 4.137060642242432, "identity": 0.4835299253463745, "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_CONV_reg/92adc297127dc6fa73dc62e71dbddba838227aa87f45e096fc2532e9e45863fb.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "92ff12f0f7e89a30b1aacb1c916383fd697a3bb2f8c57f9bd4c7f78a16f7606d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "904c4730d65044ea935530bf09a754589a31de8fbb30030fd6ca4a5072064a31", "generation_seconds_batch": 24.572221812908538, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6978039741516113, "target_z": 0.12597651175476182, "ce": 6.316492557525635, "genuineness": 4.306513786315918, "identity": 0.459033727645874, "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_CONV_reg/92ff12f0f7e89a30b1aacb1c916383fd697a3bb2f8c57f9bd4c7f78a16f7606d.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": "931ce4982efb6ce7cd98cf6b5b4f574ea10461e74456133dd20b0770ad945056", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "452a3336776b319c48ad87a89a3fe1e42cfff2bbcda1d9817ddfdafe9b13f628", "generation_seconds_batch": 23.513895066920668, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3899967670440674, "target_z": 0.6562500542233828, "ce": 6.38003396987915, "genuineness": 4.7081804275512695, "identity": 0.6416665315628052, "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_CONV_reg/931ce4982efb6ce7cd98cf6b5b4f574ea10461e74456133dd20b0770ad945056.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "933284ae8ca1ad58995d8cdef37d5219c3e48118e3c4d90d93f6fdb0a374490f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4b2e1925f38f311ad79c7ec8b48507332ffb7057ca7fa6b5e3d3a8d4f43b8dd6", "generation_seconds_batch": 26.524054056964815, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4052937030792236, "target_z": 0.6679686975602811, "ce": 5.982900142669678, "genuineness": 4.557555198669434, "identity": 0.6318655610084534, "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_CONV_reg/933284ae8ca1ad58995d8cdef37d5219c3e48118e3c4d90d93f6fdb0a374490f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "940784f84803648013ffdd7b0b947f769e7ad1e858b8ced3a33ed272428d701f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0eaac77802eda3413ccb2508406a6fd7fec4c4e76672a7c2f3ec3306776e1da0", "generation_seconds_batch": 25.823580799042247, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.032473564147949, "target_z": 1.1484372752583476, "ce": 6.173524379730225, "genuineness": 4.457139015197754, "identity": 0.37646764516830444, "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_CONV_reg/940784f84803648013ffdd7b0b947f769e7ad1e858b8ced3a33ed272428d701f.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": "954043a31408cc422247783949b626c90fbd053a1995f348bed11a84be12864b", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "77dddea4b39d0b2e514442bd1de9556e8348d1dfa085948331e466ebc022183b", "generation_seconds_batch": 21.350276642944664, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7469284534454346, "target_z": 0.9296875007134655, "ce": 6.09409761428833, "genuineness": 4.545003414154053, "identity": 0.5276733636856079, "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_CONV_reg/954043a31408cc422247783949b626c90fbd053a1995f348bed11a84be12864b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "958a0ec6c0f7d40a7733746de7b8a22371a8de47df97a06695ac87b79bca08a6", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5254a7ad094bb7274113ae973d4e8dd98d86929344f5aaa5bb0774fb34a917ae", "generation_seconds_batch": 18.730553666013293, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.078956127166748, "target_z": 0.4179687203911791, "ce": 5.919358730316162, "genuineness": 4.105680465698242, "identity": 0.1235392838716507, "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_CONV_reg/958a0ec6c0f7d40a7733746de7b8a22371a8de47df97a06695ac87b79bca08a6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "95f8c6d3dc237c31e60810bdb9a4fe487f4d18e65ce9d4b7ac8202789ea9b778", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b07de47932953e6c3ca63bc92053341430286257cef0770d16dd84f67f784c16", "generation_seconds_batch": 27.043465691967867, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.520612955093384, "target_z": -0.009765596996476644, "ce": 5.998785495758057, "genuineness": 4.268857479095459, "identity": 0.5116872787475586, "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_CONV_reg/95f8c6d3dc237c31e60810bdb9a4fe487f4d18e65ce9d4b7ac8202789ea9b778.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "961f93a8abd8822eceef1d1ac9bc49eb4f4f47066e8913b4c435971e9cdece8b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d7d6e159facfb0506dc044e1e6254be677dd6b4463b928da0633b122afd08a56", "generation_seconds_batch": 16.76785416295752, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.364501476287842, "target_z": 0.6367186775832453, "ce": 6.022613525390625, "genuineness": 4.557555198669434, "identity": 0.36682069301605225, "hypothesis": "Die Tür ist noch offen. Ich habe dein 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_CONV_reg/961f93a8abd8822eceef1d1ac9bc49eb4f4f47066e8913b4c435971e9cdece8b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "96259d749ba7a7e1d56725786d3e3525d6e4d5b9e8a9782c9c7b2df94b812f80", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "756399504a1de0679652cc5d1e7d7f9852ecaeaa2e4c34dd6af4b05b7793646d", "generation_seconds_batch": 22.876277455943637, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6347498893737793, "target_z": 0.8437499914384132, "ce": 5.998785495758057, "genuineness": 4.036643981933594, "identity": 0.5323405861854553, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo, sobre la silla. Dime cuándo estás listo.", "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_CONV_reg/96259d749ba7a7e1d56725786d3e3525d6e4d5b9e8a9782c9c7b2df94b812f80.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": "975941d162634c01232fe74bc3f2f983b50da8691c5e9008d0b39c49e5eea3c4", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "bfd18ad20d9776f82b1e6f77a838b8cb781d1e21eafd6f96da6172f7a5831cfe", "generation_seconds_batch": 19.099420040030964, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7367305755615234, "target_z": 0.9218751327045948, "ce": 5.879645347595215, "genuineness": 4.645420074462891, "identity": 0.766601026058197, "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": 0.03999999999999915, "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_CONV_reg/975941d162634c01232fe74bc3f2f983b50da8691c5e9008d0b39c49e5eea3c4.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": "9831996c37243e34df096a38deb21cb6b81026781b6d1d2a07dadb2fa3aeceff", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The road turns left behind the bridge.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Keep going... until you reach the small wooden gate.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "f121e17f5e46f840d95d5fa51ec32169a784cd0768f81d0742ec27b7e7eb9fc3", "generation_seconds_batch": 15.30347288097255, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4001946449279785, "target_z": 0.6640624222322536, "ce": 6.403861999511719, "genuineness": 4.394378662109375, "identity": 0.19998612999916077, "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_CONV_reg/9831996c37243e34df096a38deb21cb6b81026781b6d1d2a07dadb2fa3aeceff.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "98dbe7a86727646ef5b25faf1e195113bde95ebf5b9e962a73ebcb4fec4bd4dc", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "0e7f10b4873fd967dac1aa0357793a16401da5fef1107660fa178db6cd51be68", "generation_seconds_batch": 24.00714112597052, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.512373447418213, "target_z": 0.7500001141544901, "ce": 6.1894097328186035, "genuineness": 4.582659721374512, "identity": 0.3027832508087158, "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_CONV_reg/98dbe7a86727646ef5b25faf1e195113bde95ebf5b9e962a73ebcb4fec4bd4dc.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": "98f251afd4c0dc8b5c480b69f8092bbe579602969924e752d4c30b654b37e903", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "441587d11dc74eee03b58c40ffafd589b844bf63d1ed1cf8bcad9bf3f9d6cd49", "generation_seconds_batch": 15.29140407196246, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.246633052825928, "target_z": 1.3125001084467656, "ce": 6.24500846862793, "genuineness": 4.570107460021973, "identity": 0.3935644328594208, "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_CONV_reg/98f251afd4c0dc8b5c480b69f8092bbe579602969924e752d4c30b654b37e903.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "993695f6ec502287c5cf5f518cdb8f2fb40352757fdf7cf5f1456698461de717", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8cb84f671d0fc98d4407db1cc399d14a26d844c663d4c328c12ceae06be08f0c", "generation_seconds_batch": 25.22967670892831, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.808116912841797, "target_z": 0.9765626220026113, "ce": 6.284721851348877, "genuineness": 4.6203155517578125, "identity": 0.5516724586486816, "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_CONV_reg/993695f6ec502287c5cf5f518cdb8f2fb40352757fdf7cf5f1456698461de717.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "600545388a4521d881841eaba3582c4f00e75be50c8fc6fff1fb4d0114496cf3", "generation_seconds_batch": 17.796819278039038, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.369600534439087, "target_z": 0.6406249529112729, "ce": 6.348263263702393, "genuineness": 4.519899368286133, "identity": 0.6205669641494751, "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_CONV_reg/99bf1b7d3eed7b885fdae888ab00b4960791c3971d259b46f771306d2801e39b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "9a26fd0d9b16f48d5be6c392ba6750e7cea02fe18d417e4709c6202544c9eb8a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9887cd3acd1fd3c15130140d74379b733adf1b6ebd6fdc7b9ca9b53d01fbd6ec", "generation_seconds_batch": 22.31962956895586, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8979408740997314, "target_z": 0.2792969508057161, "ce": 6.324435234069824, "genuineness": 4.457139015197754, "identity": 0.5461746454238892, "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_CONV_reg/9a26fd0d9b16f48d5be6c392ba6750e7cea02fe18d417e4709c6202544c9eb8a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "9ab7395c8b517a60f27c4575f62cab4e9e8d99900addb971feba9ab8bff0c0b0", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ba262b259fb8ab1293a9cafda83b777cb24a0f0754021003782ae5d98f0edfd2", "generation_seconds_batch": 24.312398375943303, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.552574157714844, "target_z": 1.5468748016565734, "ce": 6.062326908111572, "genuineness": 4.65797233581543, "identity": 0.6347247362136841, "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_CONV_reg/9ab7395c8b517a60f27c4575f62cab4e9e8d99900addb971feba9ab8bff0c0b0.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "9ab87c199f864b62077a6006fcbc444a3ec48b8fadc832014c8c4a5e6ab7b335", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "16be187f00dd1813c11e766aa3d326bd6b5307f110d215b99d2f2d4dae56af4f", "generation_seconds_batch": 22.70547784899827, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2880163192749023, "target_z": 0.5781250956043855, "ce": 5.879645347595215, "genuineness": 4.482243061065674, "identity": 0.4841224253177643, "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_CONV_reg/9ab87c199f864b62077a6006fcbc444a3ec48b8fadc832014c8c4a5e6ab7b335.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "9ae55bd476aa1c27bc353df1ae4b2f4d9f1afd7747da92e57527e9f8ea3f512f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1e7450c6724219ba77a8919a4fec516e824b2f8530a29bec1e4b8118efbb15c5", "generation_seconds_batch": 25.17384561302606, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7571263313293457, "target_z": 0.9374998687223364, "ce": 6.197352409362793, "genuineness": 4.7458367347717285, "identity": 0.6388729810714722, "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_CONV_reg/9ae55bd476aa1c27bc353df1ae4b2f4d9f1afd7747da92e57527e9f8ea3f512f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "9af4800b36d8cfc40806ce7320d90a24ad912ecb5366aed4539ea457d618a60c", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "380e22d7eaa4c6630abb7b76462542a01ae8d6dcb5b408ae9fdfb99cd7f6a3cb", "generation_seconds_batch": 17.123980875941925, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6085710525512695, "target_z": 0.05761715013224113, "ce": 6.181467056274414, "genuineness": 4.137060642242432, "identity": 0.42856183648109436, "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_CONV_reg/9af4800b36d8cfc40806ce7320d90a24ad912ecb5366aed4539ea457d618a60c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "9b1680f43a5dabcdab0186f0c55f983ca00b36cec585068052d4b10a034746e9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "29406f9d1435355cfceddc3897c1c40ff2ceb8a819be2860318e15cf0352109d", "generation_seconds_batch": 24.669284758041613, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5429673194885254, "target_z": 0.7734374008282867, "ce": 5.982900142669678, "genuineness": 4.582659721374512, "identity": 0.5537424087524414, "hypothesis": "J'ai trouvé la petite boîte. Elle était près d'un 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_CONV_reg/9b1680f43a5dabcdab0186f0c55f983ca00b36cec585068052d4b10a034746e9.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": "9b24341b5cd95d3a2f32238c86b72f40444bfcab800766495671b46733ba83db", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "ee8e9f5fbdffb71c09de374c821f3ad6e0911cfa3103cfe71ebef2ddd00caa8f", "generation_seconds_batch": 21.619546915986575, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7010371685028076, "target_z": 0.8945312054084023, "ce": 6.1894097328186035, "genuineness": 4.457139015197754, "identity": 0.728324294090271, "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_CONV_reg/9b24341b5cd95d3a2f32238c86b72f40444bfcab800766495671b46733ba83db.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "9b68444bb5d209bd278d1dca95cdd641d041b90557b12882f353ee81de3f6f6a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9e8f48b12f6a7ae86bac4747230605544a6109921588bbe092d8cef788b8d710", "generation_seconds_batch": 23.185613811016083, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.609254837036133, "target_z": 0.82421879744546, "ce": 6.348263263702393, "genuineness": 4.871357440948486, "identity": 0.6638977527618408, "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_CONV_reg/9b68444bb5d209bd278d1dca95cdd641d041b90557b12882f353ee81de3f6f6a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "9b8df37a2e2a2794d2d0cd087b28a9fd3b3d46b57ad2188493ec92d14743dcf5", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7057bc680a3e2e3bcf9f5fdb3c25002d0771d03b23e520691966b3289d177c18", "generation_seconds_batch": 25.68976305401884, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9412825107574463, "target_z": 0.3125000171231735, "ce": 6.09409761428833, "genuineness": 4.406930446624756, "identity": 0.6329303979873657, "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_CONV_reg/9b8df37a2e2a2794d2d0cd087b28a9fd3b3d46b57ad2188493ec92d14743dcf5.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": "9bbe8bbea5064c1f8dbfbfeee107c88c80688d40a7b316d06b047ebd6ba579fc", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "63363c4b1fcb008aaf98d579bd3c796294cfeaa61721a1a6841d399e65471b5a", "generation_seconds_batch": 19.30197783804033, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.82273006439209, "target_z": 0.22167961802629077, "ce": 6.038498878479004, "genuineness": 4.469690799713135, "identity": 0.5528398752212524, "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_CONV_reg/9bbe8bbea5064c1f8dbfbfeee107c88c80688d40a7b316d06b047ebd6ba579fc.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "2fc29306f6c21ace4bfa51e871e37e86174d976f9cdd44c7562756568b9b3ed7", "generation_seconds_batch": 18.134511395008303, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.821455478668213, "target_z": 0.22070318617967205, "ce": 6.252951145172119, "genuineness": 4.218649387359619, "identity": 0.6941460371017456, "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_CONV_reg/9c09e940c5dab53287ffae2d1a8fa5296f41b2df71714656aaa2bb8830bf5162.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "9c5fb728d6b5d938038cee64f49f16bfceb2a071734b2542b8df77afb2dad0bc", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "173ebb5584f131ab9ba8f27de868a28a6a4ce20f45f86898363c4b0d9d1c3380", "generation_seconds_batch": 22.7073774819728, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9642281532287598, "target_z": 0.3300781647757052, "ce": 6.252951145172119, "genuineness": 4.2374773025512695, "identity": 0.49903780221939087, "hypothesis": "The door is till open. I put your coat on the chair. Tell me when you are ready.", "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_CONV_reg/9c5fb728d6b5d938038cee64f49f16bfceb2a071734b2542b8df77afb2dad0bc.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 128, "seconds_generated": 10.24, "mp3_sha256": "c7637028242f2025b2baeae32f58f742f7fb2c73c670180585d6a60072fe920d", "generation_seconds_batch": 17.481514143990353, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5716030597686768, "target_z": 0.029296790989429933, "ce": 6.117925643920898, "genuineness": 4.400654315948486, "identity": 0.605656087398529, "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_CONV_reg/9c969db20a6db380e08592f6aed9c9189b3db634d677ebe8f573f593ebdfd2d3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "9cfd18760e2dd51f1fc6d572a6eddede2a8fec35058fba4439a96c901f8241a9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d4a2f97de21739a7749dc38b38ad5fe538039964ef7072d94bcb7c9fe2dbdd8f", "generation_seconds_batch": 24.52233663003426, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7979187965393066, "target_z": 0.9687500713465563, "ce": 6.2767791748046875, "genuineness": 4.325342178344727, "identity": 0.46488747000694275, "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_CONV_reg/9cfd18760e2dd51f1fc6d572a6eddede2a8fec35058fba4439a96c901f8241a9.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": "9d1b8399858fd24fea43ceea0951f9ec9bd3c3b225d9a1a31e99496fb894850f", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De weg gaat linksaf na de brug.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Loop verder... tot aan het kleine houten hek.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2a37ec6815f15c26f0f7acc4661aba2fb686542b1cbe24123da862e4a6849e94", "generation_seconds_batch": 21.55607965704985, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6959381103515625, "target_z": 0.8906249300803748, "ce": 6.006728172302246, "genuineness": 4.331617832183838, "identity": 0.46314141154289246, "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_CONV_reg/9d1b8399858fd24fea43ceea0951f9ec9bd3c3b225d9a1a31e99496fb894850f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "9d2775b482572a938ea5dd48c3dd7274d73af7be7be4384d213c48140beedadf", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "82dca7145a988abef4280bdd5bdcae610bad6ba58f1897c54a53bd07a6044bfe", "generation_seconds_batch": 18.20648849301506, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.62896728515625, "target_z": 0.0732422514443511, "ce": 5.895530700683594, "genuineness": 3.998987913131714, "identity": 0.36617791652679443, "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_CONV_reg/9d2775b482572a938ea5dd48c3dd7274d73af7be7be4384d213c48140beedadf.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "9d644afb8f95ec442aabdc4bf44551564030cb2f6614c00dc84c917e2486637d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "80eeb7bb5e763643f35b588bd57c1503c8897f882d26e6eb581e9920ce4fc085", "generation_seconds_batch": 25.20427962206304, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.359402656555176, "target_z": 0.632812584902402, "ce": 6.316492557525635, "genuineness": 4.218649387359619, "identity": 0.2346087247133255, "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_CONV_reg/9d644afb8f95ec442aabdc4bf44551564030cb2f6614c00dc84c917e2486637d.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": "9db634a2910db22a007f3a72988d2c8ddb4e9a177d60864877f9dd7f239dd2a7", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "df118d7ed76adaa85dd8919ed0ccc6427deb80317a2406b2c001ec572ee5d0d4", "generation_seconds_batch": 15.538347241934389, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5531654357910156, "target_z": 0.7812499514843417, "ce": 5.9431867599487305, "genuineness": 4.406930446624756, "identity": 0.6803251504898071, "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_CONV_reg/9db634a2910db22a007f3a72988d2c8ddb4e9a177d60864877f9dd7f239dd2a7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "9e10123b6c0fbcdf293c73b40991598d5f105648b768bb3829a3d7cbad83ea33", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "1e206736455b4f1405d3a5c7500a98893f4c4821012b447759edd05aec9146bc", "generation_seconds_batch": 21.68326976604294, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.419999599456787, "target_z": 1.4453123737165954, "ce": 5.982900142669678, "genuineness": 4.413206577301025, "identity": 0.1431630402803421, "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_CONV_reg/9e10123b6c0fbcdf293c73b40991598d5f105648b768bb3829a3d7cbad83ea33.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "9e22f6529f1bafee5aacb48a0d95e4b9ed853997a2c41cab84ab65043cd075a5", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a1084f0ad781e9b449ccedd1b191998aa7c3f3f06a95c21d0cc05ac5eba544bf", "generation_seconds_batch": 24.83776940801181, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.103860378265381, "target_z": 1.2031251298507326, "ce": 6.078212261199951, "genuineness": 4.632867813110352, "identity": 0.6342796087265015, "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_CONV_reg/9e22f6529f1bafee5aacb48a0d95e4b9ed853997a2c41cab84ab65043cd075a5.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "9e2d3847b91d882f5ab5a2e2d3be777261b5512217d308eda415149b9bf71fc6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "26872c5ed9faa74c62482092e7c55fa4e85c398582be8929bdbfe7a8a73b8447", "generation_seconds_batch": 19.53781936597079, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.175837516784668, "target_z": 0.49218740368214897, "ce": 6.133810997009277, "genuineness": 4.507347106933594, "identity": 0.6191751956939697, "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_CONV_reg/9e2d3847b91d882f5ab5a2e2d3be777261b5512217d308eda415149b9bf71fc6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "9e7465ce61d985ac9a9c07508e26ddb04300c30a6dfb0a03ce4421113afce614", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e7c1a3e18d3ce00317100fc45072988e8040548e3c7f17da93d2742a09e03bf8", "generation_seconds_batch": 24.001647231983952, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7928194999694824, "target_z": 0.9648436133713446, "ce": 5.959072113037109, "genuineness": 4.256305694580078, "identity": 0.44804105162620544, "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_CONV_reg/9e7465ce61d985ac9a9c07508e26ddb04300c30a6dfb0a03ce4421113afce614.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "9ec8f6b234c2f7ee9f5d281eb09b5b8cc9869d453744f4aa7614698046f8571e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f53d11b496308887ed5625457d7d48341baa1c5419f290558bd697f6e477736f", "generation_seconds_batch": 21.425960997934453, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.971285581588745, "target_z": 1.1015625192635703, "ce": 6.078212261199951, "genuineness": 4.557555198669434, "identity": 0.5748255252838135, "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_CONV_reg/9ec8f6b234c2f7ee9f5d281eb09b5b8cc9869d453744f4aa7614698046f8571e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "9f54cbaa4132fb42c9bd563f4273386b64442a174497cefe826fc76ce34d2ef0", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "291f2240e761cb7377ba9f479d27185fd26dd09dbd498a93e98cc0d2e2c1a858", "generation_seconds_batch": 18.853981095016934, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8953912258148193, "target_z": 0.27734372181811023, "ce": 6.09409761428833, "genuineness": 4.432034492492676, "identity": 0.4505721628665924, "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_CONV_reg/9f54cbaa4132fb42c9bd563f4273386b64442a174497cefe826fc76ce34d2ef0.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "9f5e2e50c55a8d525729e5836cbca4e7e65b23fe7b402b1434a8b9ef54ebc61f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "0933efa605696b898573ddf07ff19bc225e302272434cb82d660b0d133262c2f", "generation_seconds_batch": 23.8044083340792, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.205840587615967, "target_z": 1.2812499058225457, "ce": 5.982900142669678, "genuineness": 4.645420074462891, "identity": 0.7120739221572876, "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": 0.03999999999999915, "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_CONV_reg/9f5e2e50c55a8d525729e5836cbca4e7e65b23fe7b402b1434a8b9ef54ebc61f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "9f9dbaae7a79b07a4091269d52f58a861f9510a49aeabdab33728be77c431a22", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "eadbd9afa3c90507eff2b2fbfe6f008d30bf1d18af3ca2d45396ec4229692f81", "generation_seconds_batch": 24.150943858898245, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6041557788848877, "target_z": 0.8203125221174324, "ce": 5.9431867599487305, "genuineness": 4.519899368286133, "identity": 0.4367588758468628, "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_CONV_reg/9f9dbaae7a79b07a4091269d52f58a861f9510a49aeabdab33728be77c431a22.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "9fb23b5bbaff8e2c45951be478e4d1774cd1606663ecbcd3ba8af8196ef97b38", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "270cc379235fbedbd2207f00dd7fa6ff195a0d41a9bbbeeb30db0249ee6c1f09", "generation_seconds_batch": 23.337745286058635, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7571263313293457, "target_z": 0.9374998687223364, "ce": 5.7128496170043945, "genuineness": 4.7081804275512695, "identity": 0.5080958604812622, "hypothesis": "En concrado, la cajita estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro.", "wer": 0.11764705882352941, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/9fb23b5bbaff8e2c45951be478e4d1774cd1606663ecbcd3ba8af8196ef97b38.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "9ff6a2d85bc1f9ddb694c55a167161f16a56c99a7e10bbe31df4d56045e4296f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0ecbdba7d12824c313180418b458133667985815f81320b7efacce849aedcef3", "generation_seconds_batch": 25.491545805009082, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9565796852111816, "target_z": 0.324218843107256, "ce": 5.824047088623047, "genuineness": 3.961331605911255, "identity": 0.456814169883728, "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_CONV_reg/9ff6a2d85bc1f9ddb694c55a167161f16a56c99a7e10bbe31df4d56045e4296f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "a0edfde5d24a8cfcfdc10d2c0d5a18747bee3729f476f2501666087902ae079d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0d92a686b4356d663a6c7ef77db97fe1233b3283669f4ca37980417336f41790", "generation_seconds_batch": 23.673998876009136, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1579909324645996, "target_z": 0.47851553135764485, "ce": 6.300607204437256, "genuineness": 4.394378662109375, "identity": 0.5631228685379028, "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_CONV_reg/a0edfde5d24a8cfcfdc10d2c0d5a18747bee3729f476f2501666087902ae079d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "a124971ca0fba53a301fe344a64dbbd783a0b5d497ffd3ab2bf452c7b9e0f448", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f1ae9a8006e7946c09f8339ebe6f0b77730a62900e65469f97f23099cd8f04db", "generation_seconds_batch": 23.2534371529473, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.369600534439087, "target_z": 0.6406249529112729, "ce": 6.0146708488464355, "genuineness": 4.394378662109375, "identity": 0.5975278615951538, "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_CONV_reg/a124971ca0fba53a301fe344a64dbbd783a0b5d497ffd3ab2bf452c7b9e0f448.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "a15ba435d5abbf76bc53ff68c5189fe747e1e9c9a7e25bae7c4ccace994fd575", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e65964a6331ce487849907476555d996ce0e8a666b052a91e5700cac8ea8e19d", "generation_seconds_batch": 23.86763852695003, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.925985336303711, "target_z": 0.300781191139091, "ce": 5.95112943649292, "genuineness": 4.482243061065674, "identity": 0.7590378522872925, "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_CONV_reg/a15ba435d5abbf76bc53ff68c5189fe747e1e9c9a7e25bae7c4ccace994fd575.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "a1e5a0f2fff29af5659970aea25bde8e1d9083873220c4e819df61239685beed", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "ea9a2b4bd631b38fc1567c1b0dc1ad32c836ed55c9e00368f861fc099bf46f0e", "generation_seconds_batch": 23.30763170600403, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.022275924682617, "target_z": 1.140625089896661, "ce": 6.022613525390625, "genuineness": 4.105680465698242, "identity": 0.43582969903945923, "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_CONV_reg/a1e5a0f2fff29af5659970aea25bde8e1d9083873220c4e819df61239685beed.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "a276b7ffde69cc207d89817e1fec1f1c0e288757943951e8cf9c49e7a212d96c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a888ca848bd302ff9959423e40993d225dad20df71ced69adb7c2e2e85e9b2c2", "generation_seconds_batch": 23.26624842500314, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.430788993835449, "target_z": 0.6875000742004186, "ce": 5.982900142669678, "genuineness": 4.281409740447998, "identity": 0.6064466834068298, "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_CONV_reg/a276b7ffde69cc207d89817e1fec1f1c0e288757943951e8cf9c49e7a212d96c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "a28b21f1695a9b24912d0fbced333b0a151f1af79659f778bb5c7f6fd2fcfc60", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "00617952f874c330c030d8b6e2831ee4cb5d6af5945fa3bd5feca9706fbe12cf", "generation_seconds_batch": 23.978681421955116, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.731631278991699, "target_z": 0.917968674729383, "ce": 6.141753673553467, "genuineness": 4.6956281661987305, "identity": 0.695645809173584, "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_CONV_reg/a28b21f1695a9b24912d0fbced333b0a151f1af79659f778bb5c7f6fd2fcfc60.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "a2917210b4c60224b50eea666d9513eff1f537c3ec42f37e553d5f3498a6baab", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2eeddd5c2417a70ac942893434edd8f9214ac0cb858cd1a419720cc28d25a799", "generation_seconds_batch": 18.034853058983572, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9406914710998535, "target_z": 1.0781250499425894, "ce": 6.09409761428833, "genuineness": 4.7081804275512695, "identity": 0.7252556085586548, "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_CONV_reg/a2917210b4c60224b50eea666d9513eff1f537c3ec42f37e553d5f3498a6baab.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "a2b5d5b514dd1a1401777b26259395514c0a90dde8994cbdb8ca3f8c716e89c9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8b5d29da8e7abeab82e86065755bc33be89ad70e08e713cf24542fa5aea77186", "generation_seconds_batch": 26.17489609203767, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9565796852111816, "target_z": 0.324218843107256, "ce": 5.998785495758057, "genuineness": 4.231201171875, "identity": 0.6955283880233765, "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_CONV_reg/a2b5d5b514dd1a1401777b26259395514c0a90dde8994cbdb8ca3f8c716e89c9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "a32bc3d3513239d4fcdf8444170c371e85152ded89dc5c1a817085deab45f24d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "23bab47680d6bd1a54b32bea3bbe7d3c65fa10f1fc2e91d6d0cd0d0729b4ca74", "generation_seconds_batch": 18.92875500395894, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4970762729644775, "target_z": 0.7382812881704076, "ce": 6.006728172302246, "genuineness": 4.645420074462891, "identity": 0.5773823261260986, "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_CONV_reg/a32bc3d3513239d4fcdf8444170c371e85152ded89dc5c1a817085deab45f24d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "a3433fe7d404526935c2f899c74c09ef841f0afa4d662eae16ef6a69bd5ec711", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "9d5a6b8571e492a08863958e1c50ba1d231aa1c988e96f41eff17aa85e3c5a96", "generation_seconds_batch": 23.671924520051107, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5378684997558594, "target_z": 0.7695313081474434, "ce": 6.062326908111572, "genuineness": 4.331617832183838, "identity": 0.5948168635368347, "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": 0.03999999999999915, "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_CONV_reg/a3433fe7d404526935c2f899c74c09ef841f0afa4d662eae16ef6a69bd5ec711.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "a45e643914e78d85fe29c228a42c24d5550d04a6524bc6b2e5ad019e56e08bc4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9736d0aebf7934338c4a3a1e87fd5f24acf9b8608d0918fb1b4b2dccc44a52eb", "generation_seconds_batch": 25.22967670892831, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4358878135681152, "target_z": 0.6914061668812619, "ce": 6.2767791748046875, "genuineness": 4.532451152801514, "identity": 0.43511003255844116, "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_CONV_reg/a45e643914e78d85fe29c228a42c24d5550d04a6524bc6b2e5ad019e56e08bc4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "a47b2974ffaf8e3ade70eff8d57127493304bed26c0f5df2a4552f6541e49b9b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2e82fb7016b9545d9d376566f6bf3e05d7b2b54a94e6eb5176edf740c7b8b2fc", "generation_seconds_batch": 25.786177610978484, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.609254837036133, "target_z": 0.82421879744546, "ce": 6.213237762451172, "genuineness": 4.319066047668457, "identity": 0.37923452258110046, "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_CONV_reg/a47b2974ffaf8e3ade70eff8d57127493304bed26c0f5df2a4552f6541e49b9b.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": "a57ef837299d0c17d7545fb62c05e49e13be56c4825ccf2a8c2f55c41b32d5a0", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "7227e111fd3a7a02b4e7ab5584123daf0bb773beaa9bf0d569165db3844b1acd", "generation_seconds_batch": 15.12899301899597, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.359402656555176, "target_z": 0.632812584902402, "ce": 6.324435234069824, "genuineness": 4.595211505889893, "identity": 0.5172214508056641, "hypothesis": "Die Schlüsse liegen auf dem Küchentisch. Ich suchte 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_CONV_reg/a57ef837299d0c17d7545fb62c05e49e13be56c4825ccf2a8c2f55c41b32d5a0.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "a59792c4c8f85e5f4bcd6aa00f791f69927b40658023983392b3db8434edaed4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6493c01d31cd845eeb755186beff262bc3be2859d70dd24c4fc540c9d9403c65", "generation_seconds_batch": 23.903037442010827, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1636810302734375, "target_z": -0.2832032261337436, "ce": 5.756534099578857, "genuineness": 4.406930446624756, "identity": 0.5457358360290527, "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_CONV_reg/a59792c4c8f85e5f4bcd6aa00f791f69927b40658023983392b3db8434edaed4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "a643dcd625703bdcdcca358db9ef3ceeea7b48095e8767ef3cd4554eba75da54", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b3918076f76181ba0954bc29356bf4dbf91e634a09064b95d6343b892db60ff2", "generation_seconds_batch": 23.544018284068443, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.124847412109375, "target_z": 0.4531250156962424, "ce": 6.340320587158203, "genuineness": 4.432034492492676, "identity": 0.5329673886299133, "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_CONV_reg/a643dcd625703bdcdcca358db9ef3ceeea7b48095e8767ef3cd4554eba75da54.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": "a64c29b07123a7a31b4d5cd173a38f12c3fa4d11e1d6250ac6941f65f94093eb", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Someone moved the little blue chair.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4b134f1000e8acfe3932888cd7f10bd20a863a64a24f4bc783e846f8af0ea8b0", "generation_seconds_batch": 18.21244785003364, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8285129070281982, "target_z": 0.9921875406675371, "ce": 6.2767791748046875, "genuineness": 4.444586753845215, "identity": 0.24225610494613647, "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_CONV_reg/a64c29b07123a7a31b4d5cd173a38f12c3fa4d11e1d6250ac6941f65f94093eb.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "a6fc0c61afa74939192299c5e94b89fd0eabf418398b39247064bd6d30137afc", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "144c41a21e17be4c6c376ba7cad0d1b00596f5dc35f490e289c318c93700d4d9", "generation_seconds_batch": 23.2534371529473, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2064318656921387, "target_z": 0.5156250556503139, "ce": 5.895530700683594, "genuineness": 4.557555198669434, "identity": 0.7362817525863647, "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": 0.03999999999999915, "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_CONV_reg/a6fc0c61afa74939192299c5e94b89fd0eabf418398b39247064bd6d30137afc.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "a70ca19d6ec06fc1e9060d76aa1c62b9db1581811b037b1899ed83f9f5418077", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "28688966980d9ca2ae0f6e8aee0495b693e5cac4eaa5796df0b7bd0d4b9ac927", "generation_seconds_batch": 22.819248728919774, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6857399940490723, "target_z": 0.8828123794243198, "ce": 6.141753673553467, "genuineness": 4.770940780639648, "identity": 0.5967299938201904, "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_CONV_reg/a70ca19d6ec06fc1e9060d76aa1c62b9db1581811b037b1899ed83f9f5418077.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "a712e2c0cacf6bdce6954d382375a0696af0a26322802a50e141ea11df566150", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "eef026fab093c52090de1b30ab33b1109e183602bfab29c9d27921886ea5f2e8", "generation_seconds_batch": 25.10923226608429, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.961087465286255, "target_z": 1.0937499686075152, "ce": 5.784333229064941, "genuineness": 4.287685871124268, "identity": 0.3968638479709625, "hypothesis": "That door is still open. I put your coat on the chair. Tell me when you are ready.", "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_CONV_reg/a712e2c0cacf6bdce6954d382375a0696af0a26322802a50e141ea11df566150.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "a73e462745f760635007e580ef0db7143500ac0bd2190118c58dff720826a2e4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b350ff7b066118aa95dfd483639c2cc99da83f8761ad0533c2f780c8d3cd7d75", "generation_seconds_batch": 22.2809402680723, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0942533016204834, "target_z": 0.42968754637526163, "ce": 6.173524379730225, "genuineness": 4.519899368286133, "identity": 0.7792725563049316, "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_CONV_reg/a73e462745f760635007e580ef0db7143500ac0bd2190118c58dff720826a2e4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "a749cb8ba1f98f05cc00490417794fefd34e7d4f3bf6277ede6809e504d9b1b8", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "de2f8f9e9c95b435d55e6dc7237cf4e6c600e0042262ae712bcaccb9bb35abb9", "generation_seconds_batch": 20.214222824084572, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7061362266540527, "target_z": 0.8984374807364298, "ce": 6.117925643920898, "genuineness": 4.632867813110352, "identity": 0.27653586864471436, "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_CONV_reg/a749cb8ba1f98f05cc00490417794fefd34e7d4f3bf6277ede6809e504d9b1b8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "a83e48a9ef47a7e2d93f50610792684ecbc2c422e4a8f17771a53355bbc74738", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cc5eb88ae837600eb1b4ae780000cf2f0ad7b5cf73955ce4a807b5bd8a60809a", "generation_seconds_batch": 25.511373008950613, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9055893421173096, "target_z": 0.28515627247416525, "ce": 6.006728172302246, "genuineness": 4.494794845581055, "identity": 0.6937850117683411, "hypothesis": "J'ai trouvé la petite boîte , elle était près de la 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_CONV_reg/a83e48a9ef47a7e2d93f50610792684ecbc2c422e4a8f17771a53355bbc74738.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "a8761aeb84b657c7c5cd7c149d490b429fcbb2c82b342ce8cfe1c9a1ff3caa5b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cf3b3802bcbc5ee1719fc2695ae76a37a96a62cf1286bef4b373949b8e659185", "generation_seconds_batch": 24.118739309953526, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7469284534454346, "target_z": 0.9296875007134655, "ce": 6.181467056274414, "genuineness": 4.482243061065674, "identity": 0.6827648282051086, "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_CONV_reg/a8761aeb84b657c7c5cd7c149d490b429fcbb2c82b342ce8cfe1c9a1ff3caa5b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "a89fa60a32132c6077de7aa7d489c194bff7941208ae7086da77ad64bb79b462", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "99f1db98a784d462d079c476299b4d12be7ec8c83e8108cd3eb6ef7f32ad942c", "generation_seconds_batch": 22.1952163890237, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8387107849121094, "target_z": 0.9999999086764079, "ce": 6.2767791748046875, "genuineness": 4.519899368286133, "identity": 0.3000372648239136, "hypothesis": "I f 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_CONV_reg/a89fa60a32132c6077de7aa7d489c194bff7941208ae7086da77ad64bb79b462.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "d33bbc8e21e45153e51225380b8db9f5b85f6f55555a65398d9b00739ded83cd", "generation_seconds_batch": 16.69407768000383, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.2554636001586914, "target_z": -0.21289063552361706, "ce": 6.2767791748046875, "genuineness": 4.268857479095459, "identity": 0.6964462399482727, "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_CONV_reg/a8bd001b1fcafd2f49b65d73e6ed12567e4c9664909a4d08700c9febfd977540.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "a8e96d5fda170e116557e2944bcdcef55cbc5723c8709c3ce1174976fb74b259", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e1b356248cf8f8b6fd084f455d11898048f1773fe1a31ee5342ce5685e2e45a6", "generation_seconds_batch": 18.06474249495659, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9814834594726562, "target_z": 1.109374887272441, "ce": 5.9431867599487305, "genuineness": 4.482243061065674, "identity": 0.5598524808883667, "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_CONV_reg/a8e96d5fda170e116557e2944bcdcef55cbc5723c8709c3ce1174976fb74b259.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "a8f6c0f22005de2aab4add46b14384db286d54bc03956c13f9eb7a59590e67fd", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8293de429033b8e2645159d167f0e9a730ef32fef32212d68410f72514b35e27", "generation_seconds_batch": 23.51673040597234, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4511849880218506, "target_z": 0.7031249928653444, "ce": 6.038498878479004, "genuineness": 4.595211505889893, "identity": 0.5177104473114014, "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_CONV_reg/a8f6c0f22005de2aab4add46b14384db286d54bc03956c13f9eb7a59590e67fd.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "a92909684eb7da50f02f266373b0aa175acd4f4e2d48b2b8f4a29b5508230355", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "bce1fbfc89a0abd15e9411462979ddfdfa406a958e76ebc63473722d7a027140", "generation_seconds_batch": 25.043361529009417, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.042672157287598, "target_z": 1.156250191208771, "ce": 6.09409761428833, "genuineness": 4.105680465698242, "identity": 0.16154679656028748, "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_CONV_reg/a92909684eb7da50f02f266373b0aa175acd4f4e2d48b2b8f4a29b5508230355.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": "a93f2e32ab70cbf5fb75454017dc1378a6d3fcb55f9eb01b0dcb8a2461cbd350", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Your message arrived after I had left.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1a1db0d19d2ff1278edfac6608bc5b7edf63cb3d248cdfa6df26ed24ae6c7dca", "generation_seconds_batch": 23.61702426301781, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5888586044311523, "target_z": 0.80859369613335, "ce": 6.348263263702393, "genuineness": 4.306513786315918, "identity": 0.2829604744911194, "hypothesis": "Your message arrived after I had left. I r read it twice, before putting the phone away. Uh let me know when we can talk.", "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_CONV_reg/a93f2e32ab70cbf5fb75454017dc1378a6d3fcb55f9eb01b0dcb8a2461cbd350.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "e6e0c44bec0fa87aff3f1342a2f65620ed01046cbf805b248d915c73e6f6d8b9", "generation_seconds_batch": 17.330233343993314, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.084055185317993, "target_z": 0.4218749957192066, "ce": 6.229123115539551, "genuineness": 4.432034492492676, "identity": 0.7525140047073364, "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_CONV_reg/a94b7d5da78ff217f8e5a4f7ae4520900bef0ad764d0e4ee898e69de10c28aea.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "a9e701dec4a66952a5361077150baf4e09bc74ae0877348c818c5936a2f1ffa6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "9bc4d3a9aa5388277eb9e89c53253e1f7f1f46947885d786657978e085e66572", "generation_seconds_batch": 25.305237293010578, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.727123498916626, "target_z": 0.14843754922912386, "ce": 6.197352409362793, "genuineness": 3.860914945602417, "identity": 0.1002834141254425, "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_CONV_reg/a9e701dec4a66952a5361077150baf4e09bc74ae0877348c818c5936a2f1ffa6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "a9effc945ffaae389e398f0061f4b8e4d26b138b5fcb95a8d27a572991143981", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5cf266d486eca24b6a12298abd4d84bc3b61fdda7d679c426cd2c2d16120f9e3", "generation_seconds_batch": 18.92875500395894, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.920295238494873, "target_z": 1.0624999486304794, "ce": 6.046441555023193, "genuineness": 4.570107460021973, "identity": 0.503072202205658, "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_CONV_reg/a9effc945ffaae389e398f0061f4b8e4d26b138b5fcb95a8d27a572991143981.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "aa26b569351808536624519212f6d3838cdf6ec7630d120d1eb31f6e5a561e40", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "de1d045a29c1c12a38c7e029d01f64a02ed966932c620ed4025479ed1f8ab483", "generation_seconds_batch": 16.963250310043804, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.323709487915039, "target_z": 0.6054688402533938, "ce": 6.125868320465088, "genuineness": 4.7834930419921875, "identity": 0.5428202748298645, "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": 0.03999999999999915, "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_CONV_reg/aa26b569351808536624519212f6d3838cdf6ec7630d120d1eb31f6e5a561e40.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "aa8bcaafd4c2f5eebfc504dec731faf5fc1d9053a15c3c272da6c0a47994faa3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cada09dc2032e0314570d002b9a6b3a4c00eda945e5981797650025e58f6c513", "generation_seconds_batch": 23.4627615690697, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.170738697052002, "target_z": 0.4882813110013057, "ce": 6.133810997009277, "genuineness": 4.720732688903809, "identity": 0.7610838413238525, "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_CONV_reg/aa8bcaafd4c2f5eebfc504dec731faf5fc1d9053a15c3c272da6c0a47994faa3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "aaa0d7319e1df7f3394ad2e81c53e3ca39691e9f995c7d56cf6d90679bc7d79d", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "27a014335950ab2f1e3ed3f885fd8d80544ac6c7c82f982906f8e0a9a40bd20f", "generation_seconds_batch": 15.964217353030108, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5639545917510986, "target_z": 0.023437469320980783, "ce": 6.38003396987915, "genuineness": 4.281409740447998, "identity": 0.6896547079086304, "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_CONV_reg/aaa0d7319e1df7f3394ad2e81c53e3ca39691e9f995c7d56cf6d90679bc7d79d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "aaa2279f225d9aa33e04aafb0d5ef7394f9302705453258c683c4accf267a6cf", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "a0d70564e3b74d4689bbddef44d6614cf9793c57e4103a6a33905dabe7c47938", "generation_seconds_batch": 24.757390280952677, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5939576625823975, "target_z": 0.8124999714613774, "ce": 6.213237762451172, "genuineness": 4.1998209953308105, "identity": 0.30485770106315613, "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_CONV_reg/aaa2279f225d9aa33e04aafb0d5ef7394f9302705453258c683c4accf267a6cf.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": "aabb17e7fe0c684a78da4acd686c0cdb055a711585c91ccb613018c6dd7bafa3", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b22cc474e081089a896dab8e421835c9759195490baf94ffaab8559768a7ad12", "generation_seconds_batch": 17.65268068900332, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.471580982208252, "target_z": 0.7187499115302701, "ce": 5.796247482299805, "genuineness": 4.507347106933594, "identity": 0.1382560431957245, "hypothesis": "De sleutels liggen op de keukentafel. Ik kijk 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_CONV_reg/aabb17e7fe0c684a78da4acd686c0cdb055a711585c91ccb613018c6dd7bafa3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "ab4f4f14e78ab3f97f74424b9b0f57d0e45c9a3550024e3c7295a5e4f1c2e7a4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f941b06350236bfa75ceef2deb3ec3e756e6d234603c341fbae4100e88e828a9", "generation_seconds_batch": 24.741782058030367, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.073857069015503, "target_z": 0.41406244506315165, "ce": 6.125868320465088, "genuineness": 4.557555198669434, "identity": 0.633944034576416, "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_CONV_reg/ab4f4f14e78ab3f97f74424b9b0f57d0e45c9a3550024e3c7295a5e4f1c2e7a4.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "9734f2407f5141ac418677cb92db6443cff10d88a312be97fe6d694889cb78ba", "generation_seconds_batch": 18.466284418012947, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1401443481445312, "target_z": 0.4648436590331407, "ce": 6.467403411865234, "genuineness": 4.13078498840332, "identity": 0.3345438241958618, "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_CONV_reg/ab7a94299f968552a7b966fa3391724b652d8e20483888a7767196f29a7f4a69.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "ac44af7cfe239a14a030aac41ea9ea7c07b47c7229e92d0d057b0518e2caac3a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "304d8b19e05bb6a1a3329ad6ff7543c6e17e0a3895f682e14e3232d313dc19fb", "generation_seconds_batch": 25.774161521927454, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.042672157287598, "target_z": 1.156250191208771, "ce": 6.268836498260498, "genuineness": 4.375550270080566, "identity": 0.4756796956062317, "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_CONV_reg/ac44af7cfe239a14a030aac41ea9ea7c07b47c7229e92d0d057b0518e2caac3a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "ac62e2fe8448f9bbaea32b00374d41d44496a43fdc203ea7d85b0398edfdc1fb", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0744e58db68ee9445dd32c435f7fb7640ec4c6b23413a7504979a337188734e8", "generation_seconds_batch": 18.059405258041807, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.216629981994629, "target_z": 0.5234376063063689, "ce": 6.141753673553467, "genuineness": 4.632867813110352, "identity": 0.6967863440513611, "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_CONV_reg/ac62e2fe8448f9bbaea32b00374d41d44496a43fdc203ea7d85b0398edfdc1fb.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "ad473faa4b2752bbeaa732363601b34d26aca914446f29681b0bb1793063b26b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cbf1e7e4259c666e1bfba8c9efac9c236d637164e6b02fd24c12b2b39666d3e5", "generation_seconds_batch": 17.68834655894898, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6806411743164062, "target_z": 0.8789062867434765, "ce": 6.086154937744141, "genuineness": 4.770940780639648, "identity": 0.8181434869766235, "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_CONV_reg/ad473faa4b2752bbeaa732363601b34d26aca914446f29681b0bb1793063b26b.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "dac2bd3d993528fa693d4ff880cc5e4b536e0aae904835cfc9fca9949125193f", "generation_seconds_batch": 18.964713198016398, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8010592460632324, "target_z": 0.20507808486756207, "ce": 6.133810997009277, "genuineness": 4.137060642242432, "identity": 0.5974324345588684, "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_CONV_reg/adcc0121d6c11de6953f8243f84ad890832f75283c96342c1db246fddc443438.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "adda704cfcf60e6d63e28373f7227e898e1aa830f082aa592f1e06540d120a29", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "995bdbb1babb9357092f74169fa1f46301a0e87b9597f5b23ece9a0685208f02", "generation_seconds_batch": 24.187549999915063, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0152182579040527, "target_z": 0.36914055276161173, "ce": 6.062326908111572, "genuineness": 4.394378662109375, "identity": 0.5221874713897705, "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_CONV_reg/adda704cfcf60e6d63e28373f7227e898e1aa830f082aa592f1e06540d120a29.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "adf120b28e73a000ccbbb60c4535c0cce82e46f1a031e507ddbe6bf9fe84a63c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8aca804b143569ca220d7fe6bc6b5e49f6175dbcbe93c50803706b6189708848", "generation_seconds_batch": 23.8299275849713, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.379798650741577, "target_z": 0.6484375035673278, "ce": 6.141753673553467, "genuineness": 4.8588056564331055, "identity": 0.7605702877044678, "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_CONV_reg/adf120b28e73a000ccbbb60c4535c0cce82e46f1a031e507ddbe6bf9fe84a63c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "ae011c82785706fa13387b600645b0102a7c83b00f395dce5a58fdfdca10c912", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f72e0d4e66c6d44892564425d47cbc8043a64b06f51f91074d6075b1d6bfb661", "generation_seconds_batch": 24.19183660298586, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.624551773071289, "target_z": 0.8359374407823582, "ce": 6.149696350097656, "genuineness": 4.607763767242432, "identity": 0.798443078994751, "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_CONV_reg/ae011c82785706fa13387b600645b0102a7c83b00f395dce5a58fdfdca10c912.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "ae1a8745da628eaa8ad16d8f71efd7002f923be0b425b25606ce87ca45f4349f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "3b38b410f9b7c5ec5928d6cd1539c258ec4183b9d4627313a21b8e6e33eaf932", "generation_seconds_batch": 26.484952063998207, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.419999599456787, "target_z": 1.4453123737165954, "ce": 5.967014789581299, "genuineness": 4.432034492492676, "identity": 0.2388264238834381, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. Wanneer je klaar bent.", "wer": 0.10526315789473684, "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_CONV_reg/ae1a8745da628eaa8ad16d8f71efd7002f923be0b425b25606ce87ca45f4349f.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "7e5fb63189dd024392e9ddb2c3251e8a7acacdc954074b2f72fd329e2c3c81cb", "generation_seconds_batch": 16.748898196965456, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6691219806671143, "target_z": 0.10400387285089333, "ce": 6.356205940246582, "genuineness": 4.105680465698242, "identity": 0.1581418514251709, "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_CONV_reg/ae3130a5db36c2790a31e23135fde257111c56fa9c5b8061c1f3e51a8cb61258.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "aee7fc715c173c691c123a6f074523e1d52434e85f664c3057c8913e6042d303", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "d7e8a3e9fbd4fff51f806ab1afc3ffb05b6760dd056a495cb887b70f7e64ea91", "generation_seconds_batch": 17.037154870922677, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3995110988616943, "target_z": -0.10253904243378104, "ce": 5.895530700683594, "genuineness": 4.193545341491699, "identity": 0.633689284324646, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "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_CONV_reg/aee7fc715c173c691c123a6f074523e1d52434e85f664c3057c8913e6042d303.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "aefc2cb332e09a5966f45fcd850bd339409e66e81d3c4bf9bff3db9e920ac1be", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "40a15e287e5891e0d9e18d033b24f160a4616430e37533d996fb201381115300", "generation_seconds_batch": 25.613215090008453, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.379798650741577, "target_z": 0.6484375035673278, "ce": 5.871703147888184, "genuineness": 4.507347106933594, "identity": 0.6112661361694336, "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_CONV_reg/aefc2cb332e09a5966f45fcd850bd339409e66e81d3c4bf9bff3db9e920ac1be.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": "aefd9bce135017f8a220a88f5de8b3bf1dd226d095e4d9ceeaae9123b3989570", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e31c239d8e7d719de89bfd6d49d3da711422c95a32faf0013c79e3348f773748", "generation_seconds_batch": 19.30197783804033, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.790861129760742, "target_z": 0.19726553421150708, "ce": 5.959072113037109, "genuineness": 4.494794845581055, "identity": 0.6006283760070801, "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_CONV_reg/aefd9bce135017f8a220a88f5de8b3bf1dd226d095e4d9ceeaae9123b3989570.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": "af5818cc1d3adf35bf071be48481c893306de87b301c1dfccc273c93fd455580", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "20c4b2c9de85c9279ff6b274a65980ae559598cf7df6f55f8557cc16eb7d6ace", "generation_seconds_batch": 15.34686818299815, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.369600534439087, "target_z": 0.6406249529112729, "ce": 6.300607204437256, "genuineness": 4.6705241203308105, "identity": 0.5004322528839111, "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": 0.120000000000001, "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_CONV_reg/af5818cc1d3adf35bf071be48481c893306de87b301c1dfccc273c93fd455580.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "afec7ae0a76682f9053a0293c74537b3cba0264b5cb36a5fd5f18cb77401ce0d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ffd6d34066cb0016f649de506266d85c8a0a049caba11261113da7cfc7092ed1", "generation_seconds_batch": 22.084413029951975, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.2874250411987305, "target_z": 1.343749945776617, "ce": 6.062326908111572, "genuineness": 4.582659721374512, "identity": 0.6511977910995483, "hypothesis": "J'ai trouvé sa 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_CONV_reg/afec7ae0a76682f9053a0293c74537b3cba0264b5cb36a5fd5f18cb77401ce0d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "b04bb2c531d9edb505da17cfdd0521590e94f824656f7019fa9a7578b0f7f359", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "011390fcce8e8f2eda2d396f569a5c6a241ba831af7ef3766df4c387547d238e", "generation_seconds_batch": 22.084413029951975, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.124847412109375, "target_z": 0.4531250156962424, "ce": 6.054384231567383, "genuineness": 4.469690799713135, "identity": 0.6873885989189148, "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_CONV_reg/b04bb2c531d9edb505da17cfdd0521590e94f824656f7019fa9a7578b0f7f359.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "b06e1e0a1159d71843fe1cde7790faf2375b35c499a3b0ff62b5826e2dbfd3d8", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "10da635b7b81b7a84ca9bf9b1b12e96eeee19d0c007be85da2e929e7f991e986", "generation_seconds_batch": 23.51210604305379, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7571263313293457, "target_z": 0.9374998687223364, "ce": 5.967014789581299, "genuineness": 4.595211505889893, "identity": 0.6808804869651794, "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_CONV_reg/b06e1e0a1159d71843fe1cde7790faf2375b35c499a3b0ff62b5826e2dbfd3d8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "b07de0284ca7b22c57a171c16d8c206d0c2777dd233f6e7529f2a9f1b25587f7", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "329e2d319c4ab58166127d63c41de13635e598e285722d83dd5227ea01dad3a4", "generation_seconds_batch": 27.001158267026767, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9922728538513184, "target_z": 0.35156258775626426, "ce": 6.23706579208374, "genuineness": 4.7081804275512695, "identity": 0.7357496619224548, "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_CONV_reg/b07de0284ca7b22c57a171c16d8c206d0c2777dd233f6e7529f2a9f1b25587f7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "b16f1c350ac769c9febf0ee8498d8c4b453f2c73bed85415a9557088ee16f206", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5fec82705b277099f6a3f4fa62f8dc28b9b2d74f9cf96dd9691cec63931e3b53", "generation_seconds_batch": 18.987081220024265, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4868781566619873, "target_z": 0.7304687375143526, "ce": 6.324435234069824, "genuineness": 4.337893962860107, "identity": 0.3747766613960266, "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_CONV_reg/b16f1c350ac769c9febf0ee8498d8c4b453f2c73bed85415a9557088ee16f206.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "b1dfee2d3c7adadc8931386fa99eca183b7dcd8b438c86c19b84b63c7c4b680b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "190e5a898c45e2493c506ea0cfa324fe30e8f687098636f71ea058dd42e03617", "generation_seconds_batch": 17.945745402015746, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7979187965393066, "target_z": 0.9687500713465563, "ce": 6.078212261199951, "genuineness": 4.381826400756836, "identity": 0.7214810848236084, "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_CONV_reg/b1dfee2d3c7adadc8931386fa99eca183b7dcd8b438c86c19b84b63c7c4b680b.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": "b2da0aef333028946fc2b9d8f495ebca305e83a2eb6635e2ab9b65f3b27e3da6", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Je bericht kwam nadat ik vertrokken was.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik las het twee keer... en legde mijn telefoon weg.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "0bf08b64ee607668b284723f8d8b02ac0c0f81b5ba5aa58fb05f239b6239ed1f", "generation_seconds_batch": 21.619546915986575, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.348613262176514, "target_z": 1.3906248844185787, "ce": 6.165581703186035, "genuineness": 4.3127899169921875, "identity": 0.3416903614997864, "hypothesis": "Je bericht kwam nadat ik vertrokken was. Ik las het twee keer en legde mijn telefoon weg. Wanneer we kunnen praten.", "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_CONV_reg/b2da0aef333028946fc2b9d8f495ebca305e83a2eb6635e2ab9b65f3b27e3da6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "b2e6df894fe9c59c4a4fbc17c76ed338cd04d4efb4039a7d31c441499f594d82", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1054e2b611fe5313f339c84ed88deb623573d97e4db07fc26cbdd69764fa9be5", "generation_seconds_batch": 22.441387692000717, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.42883038520813, "target_z": -0.08007818760660317, "ce": 6.252951145172119, "genuineness": 4.243753433227539, "identity": 0.662498950958252, "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_CONV_reg/b2e6df894fe9c59c4a4fbc17c76ed338cd04d4efb4039a7d31c441499f594d82.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "b2f04b14342c86e28602f5f2ec91e916b2fff069f92057b3c5f53bfc62c961a8", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "fc2cbb4beda51b89f13d9c7d0f7c50ddb0ddff368c50acbde9040582f32016d8", "generation_seconds_batch": 26.88887528993655, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.872445583343506, "target_z": 0.2597655741655786, "ce": 6.109982967376709, "genuineness": 4.68307638168335, "identity": 0.8025946021080017, "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_CONV_reg/b2f04b14342c86e28602f5f2ec91e916b2fff069f92057b3c5f53bfc62c961a8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "b32be224305373b3c15e0a3e18f93f8dd0d4b67226a73b99f987844615760633", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0328d58b1c7ea538c1ba2391809af6c3eb2f855813e62ab1f32acc28438f26f8", "generation_seconds_batch": 24.07771485997364, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.07326602935791, "target_z": 1.1796874778825674, "ce": 5.919358730316162, "genuineness": 4.557555198669434, "identity": 0.6019666194915771, "hypothesis": "La porte est encore ouverte. Donne-je poser ton manteau sur la chaise ? Dis-moi quand tu es prêt.", "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_CONV_reg/b32be224305373b3c15e0a3e18f93f8dd0d4b67226a73b99f987844615760633.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": "b33c23987bea506a8c2b96109ac1403116c3b12bae5bf5c3944994a622fd1052", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "9a3e9cae4d3f4a2d49ac3ace45b1c0d8a747422d669148deec9b13398f2dc92c", "generation_seconds_batch": 21.230708232033066, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9814834594726562, "target_z": 1.109374887272441, "ce": 6.165581703186035, "genuineness": 4.595211505889893, "identity": 0.6426723003387451, "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": 0.03999999999999915, "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_CONV_reg/b33c23987bea506a8c2b96109ac1403116c3b12bae5bf5c3944994a622fd1052.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "b39d1e3455f232ea5d498f951d1c05be8e63e4a723454f7f7c43c61a8cb7ce2c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "042b11e757e3221607a59f57332cd1b0c54a16e5c0ed7a3e6f9b5d76be73da1f", "generation_seconds_batch": 18.47838441701606, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8558738231658936, "target_z": 0.24707031633487742, "ce": 6.022613525390625, "genuineness": 4.174716949462891, "identity": 0.5304450392723083, "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_CONV_reg/b39d1e3455f232ea5d498f951d1c05be8e63e4a723454f7f7c43c61a8cb7ce2c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "b3f4f43d68d8f8ac41152fd3c7ecfcb4df9d0e34fe65082214ea344e9848abd6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "68f92afd5e205c24617d89716bba5dfbaa6977dd8df0058cc08b4645de13c890", "generation_seconds_batch": 18.47838441701606, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.541008949279785, "target_z": 0.005859321668449151, "ce": 5.617537498474121, "genuineness": 3.9425034523010254, "identity": 0.4207693636417389, "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_CONV_reg/b3f4f43d68d8f8ac41152fd3c7ecfcb4df9d0e34fe65082214ea344e9848abd6.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "0f668728e4f5d4b2c6a43b79fee9d1ae0624b47dfd2196a2f3eef6c12d68f8d0", "generation_seconds_batch": 17.8303987220861, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.720749616622925, "target_z": 0.14355465940729345, "ce": 6.268836498260498, "genuineness": 4.306513786315918, "identity": 0.713584840297699, "hypothesis": "Die Schlüsse liegen auf dem Küchentisch. Ich suchte in der Schublade, dann neben 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_CONV_reg/b429ad869b4f073af736e4e6510807542fc4f5324b0eb7f096c719d9002f4f71.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "b4dad9dd7a75d3ea85a623e5b5aee8b3e260c4d403405d414b183e1fb6909f20", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "386544a2254fa1ec2924e278e5ba657dcfe6c56fee0dcafd0ceb917fd920b7e8", "generation_seconds_batch": 24.122404152993113, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7462446689605713, "target_z": 0.16308585340024673, "ce": 6.022613525390625, "genuineness": 4.532451152801514, "identity": 0.658025324344635, "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_CONV_reg/b4dad9dd7a75d3ea85a623e5b5aee8b3e260c4d403405d414b183e1fb6909f20.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "b51e22981488a8ebc7d6d540b92e0f790f51dcc22caa3b26c82f61fe59fe81e9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c141817c0edd5bcfa9e1886c589eb6c2aa3367bedd70ef46aff97f5b25d3752b", "generation_seconds_batch": 23.8299275849713, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4409868717193604, "target_z": 0.6953124422092893, "ce": 6.109982967376709, "genuineness": 4.532451152801514, "identity": 0.6331911087036133, "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_CONV_reg/b51e22981488a8ebc7d6d540b92e0f790f51dcc22caa3b26c82f61fe59fe81e9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "b5d446d135b6366da55f535731be7ae617cf494f1933136daec0bd75564efd3f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "02964b48d9d76575cbc1e6c3d97df2611bde2cc7b5bf1b1fc9d5681fda2012cd", "generation_seconds_batch": 24.673354918020777, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.726532459259033, "target_z": 0.9140625820485397, "ce": 6.181467056274414, "genuineness": 4.7834930419921875, "identity": 0.7747833132743835, "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_CONV_reg/b5d446d135b6366da55f535731be7ae617cf494f1933136daec0bd75564efd3f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "b6395537a5ee22c6eac3b07a366ca387c16682e7fb3ebb0f3b67f890711c3cd2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2e531e6dfae51e5dadc3f54fe77f60549c9a7c03b62cf55bd4f5343a8828883c", "generation_seconds_batch": 23.387461010948755, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3899967670440674, "target_z": 0.6562500542233828, "ce": 6.181467056274414, "genuineness": 4.432034492492676, "identity": 0.5836588144302368, "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_CONV_reg/b6395537a5ee22c6eac3b07a366ca387c16682e7fb3ebb0f3b67f890711c3cd2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "b66f2be9fc7fa0c8bfd45b1cd341beb26155be6c9fd304c8e5a2b4094a1fbfe9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4a14ede520c62d84722bc10787bf6b5ad4831f2b20ad6239d76fdb7a4149839e", "generation_seconds_batch": 25.914695473969914, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.512373447418213, "target_z": 0.7500001141544901, "ce": 6.173524379730225, "genuineness": 4.319066047668457, "identity": 0.49863702058792114, "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_CONV_reg/b66f2be9fc7fa0c8bfd45b1cd341beb26155be6c9fd304c8e5a2b4094a1fbfe9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "b7909cf849cb1e07b92471dbf45a569b49d65e684bc76589efdc2ab759933b99", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "11e712b8d78116ecc6e4b097916b00b49d230615e4d815f369025f7e7cb37736", "generation_seconds_batch": 23.112058863043785, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8928418159484863, "target_z": 0.2753906754776886, "ce": 5.895530700683594, "genuineness": 4.570107460021973, "identity": 0.7691309452056885, "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_CONV_reg/b7909cf849cb1e07b92471dbf45a569b49d65e684bc76589efdc2ab759933b99.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "b7aa2437d4d35bcf6c2c94a1e90a774e7dd79f98e16cb6be45ca6016acef2e03", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "46a46c91ca254f4d18505784954006b39501f75a8d139f051645dc8a468543fc", "generation_seconds_batch": 24.792980027967133, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9616785049438477, "target_z": 0.3281249357880993, "ce": 5.990842819213867, "genuineness": 4.507347106933594, "identity": 0.5088733434677124, "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": 0.03999999999999915, "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_CONV_reg/b7aa2437d4d35bcf6c2c94a1e90a774e7dd79f98e16cb6be45ca6016acef2e03.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "b7ce8d72967893c8ae0fc1079d118f78f7bd295ee457de5d5c448639212bcf86", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "337adfd839144b018171e6723f9b08bff9f7a0a339b4e4755bc42ab91d9bf798", "generation_seconds_batch": 21.998323668027297, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.923435926437378, "target_z": 0.29882814479866937, "ce": 6.157639026641846, "genuineness": 4.432034492492676, "identity": 0.47358953952789307, "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_CONV_reg/b7ce8d72967893c8ae0fc1079d118f78f7bd295ee457de5d5c448639212bcf86.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "b8108139fa6c96f1e8a92401ab85fd081f58bfafeb20b052a3f24a613eeda97e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1edbd4a3df88114bef126b2d7d853bc26c074851f4ca67849517e4aae2b23464", "generation_seconds_batch": 16.788822289090604, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.1228890419006348, "target_z": -0.3144530634635952, "ce": 6.4912309646606445, "genuineness": 4.419482707977295, "identity": 0.4947611689567566, "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_CONV_reg/b8108139fa6c96f1e8a92401ab85fd081f58bfafeb20b052a3f24a613eeda97e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "b81af99222eb64eec9427ea72c425c2070cfb3045ae0fba94e73c2c857e89c65", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "25f85ae107947b8335f468de3dbfe19db883eb78367f944de2e9a3e6d0f43548", "generation_seconds_batch": 23.419881331035867, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.318610429763794, "target_z": 0.6015625649253663, "ce": 6.2767791748046875, "genuineness": 4.582659721374512, "identity": 0.6401081681251526, "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_CONV_reg/b81af99222eb64eec9427ea72c425c2070cfb3045ae0fba94e73c2c857e89c65.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "b84c62bd2658d83c9610db62a5ef51eb49f0428d4a290269a1ad5903779ad23e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "38cf1268b682dd8cf0600ea18506bdc4660a760ae7226ad1a56421ee72368c2b", "generation_seconds_batch": 21.92747243109625, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.619452953338623, "target_z": 0.8320313481015149, "ce": 6.157639026641846, "genuineness": 4.519899368286133, "identity": 0.7125065326690674, "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_CONV_reg/b84c62bd2658d83c9610db62a5ef51eb49f0428d4a290269a1ad5903779ad23e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "b86a613b9d658dbcb94e1edd30da5d008d9a8789a3a16a76b9d0dcc0c08cba6c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "13760e173f91c6a5894429ab346ecd608445db07264b4385196b646afce726e4", "generation_seconds_batch": 23.45556684001349, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5531654357910156, "target_z": 0.7812499514843417, "ce": 6.062326908111572, "genuineness": 4.7081804275512695, "identity": 0.6623932719230652, "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_CONV_reg/b86a613b9d658dbcb94e1edd30da5d008d9a8789a3a16a76b9d0dcc0c08cba6c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "b957165bef077f91873a9738e334af8e69392156ab5072673e1b0aec2382251a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b64292502dfa69cc1656c5d60ba1578ea015cb0edf87a467e473316956f32b30", "generation_seconds_batch": 25.627598464954644, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0432629585266113, "target_z": 0.39062497574217087, "ce": 6.070269584655762, "genuineness": 3.6663577556610107, "identity": 0.42400914430618286, "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_CONV_reg/b957165bef077f91873a9738e334af8e69392156ab5072673e1b0aec2382251a.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": "b9f0be53ce1bf82876a0739e80ef7bd35bbe226cf1804c33faca90c6bd056e6d", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "a6cd1b35e6b158729464fd831e6b58a580a2c76dc9a46ecf5169e63bbc6bf0e5", "generation_seconds_batch": 21.580728361965157, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5021753311157227, "target_z": 0.7421875634984352, "ce": 6.0146708488464355, "genuineness": 4.758388519287109, "identity": 0.8027815818786621, "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_CONV_reg/b9f0be53ce1bf82876a0739e80ef7bd35bbe226cf1804c33faca90c6bd056e6d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "b9f2450a2674c9703e705eb3a12c0a72eedf97591771d73830dd883d3095d21c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0a20316596c634cb3d63b489a9adc44b1eebc020f3e264f39c19da85cccb1830", "generation_seconds_batch": 22.302212193026207, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.625143051147461, "target_z": 0.07031259061012653, "ce": 6.1894097328186035, "genuineness": 4.3504462242126465, "identity": 0.580998420715332, "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_CONV_reg/b9f2450a2674c9703e705eb3a12c0a72eedf97591771d73830dd883d3095d21c.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "be16c5957a9223ebf1722c10b92528af90567e4f4ada4ef6508a456f1a80dff3", "generation_seconds_batch": 17.261695928988047, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0534610748291016, "target_z": 0.3984375263982258, "ce": 6.141753673553467, "genuineness": 4.595211505889893, "identity": 0.7781893014907837, "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_CONV_reg/ba70e7e208a6b15a9236ccd5a69e0e5fd403df2a937202d690effa2cdc45d9f2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "baa10adcc45fe4c49c29400466d0397b7648df81b836f0ee697a294b7d7c3ef8", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "da875c1f1cfc1fa75026278e8a415c4286f7783319c112a79c9df862cb964e8e", "generation_seconds_batch": 22.16825778305065, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.042672157287598, "target_z": 1.156250191208771, "ce": 6.046441555023193, "genuineness": 4.65797233581543, "identity": 0.6592568755149841, "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_CONV_reg/baa10adcc45fe4c49c29400466d0397b7648df81b836f0ee697a294b7d7c3ef8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "baa66f7e96a425d15d5f1aa88b807935b04332450d6a0e19cd224a1c4738f3ed", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ccbdfc9e4f136136bd7926767147f4df98ff2dcc7e94e8d884dd01b07e588056", "generation_seconds_batch": 24.061862407950684, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1070008277893066, "target_z": 0.43945314337173824, "ce": 6.181467056274414, "genuineness": 4.507347106933594, "identity": 0.5950064063072205, "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_CONV_reg/baa66f7e96a425d15d5f1aa88b807935b04332450d6a0e19cd224a1c4738f3ed.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "baf4f7fcd9a162af56912155050db61728238101042796758983299cff8dde6d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "8619892eb7edbd78c86a15decb9e9eb7ab4a6b467fb4ded19bce55f96dcce5a8", "generation_seconds_batch": 21.782437655027024, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.038163900375366, "target_z": 0.38671870041414336, "ce": 5.831989288330078, "genuineness": 4.482243061065674, "identity": 0.7358170747756958, "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": 0.03999999999999915, "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_CONV_reg/baf4f7fcd9a162af56912155050db61728238101042796758983299cff8dde6d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "bb354fb19eaedd0838782e021d7628a02acfa43609e4dc675d5f643401b3d717", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "17d3336240c3f3b6b32978b4113abf7bc73ad672b466b7940f79c0ccbca45b12", "generation_seconds_batch": 25.20427962206304, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.910097122192383, "target_z": 1.0546873979744245, "ce": 6.173524379730225, "genuineness": 4.2374773025512695, "identity": 0.42520302534103394, "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_CONV_reg/bb354fb19eaedd0838782e021d7628a02acfa43609e4dc675d5f643401b3d717.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "bb952f67a1efe118f86532f37ae6ac10c903b2fd33fcd78ab97ec54a257c232c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "97beab00fd4376b4240bf6f53160f1ffcfcdea27981423338d9bd74674d5a409", "generation_seconds_batch": 17.557103842962533, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4001946449279785, "target_z": 0.6640624222322536, "ce": 6.268836498260498, "genuineness": 4.300237655639648, "identity": 0.28777045011520386, "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_CONV_reg/bb952f67a1efe118f86532f37ae6ac10c903b2fd33fcd78ab97ec54a257c232c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "bbcc7cc4365670c91c62d71733548fa30f25ca21891ba06555a85294a7fbd4a3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6737d7fb27c45398c34fa23d6dc3cd7c0d888cd02235f226efec76ec12e70d59", "generation_seconds_batch": 23.378675304935314, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.246633052825928, "target_z": 1.3125001084467656, "ce": 5.967014789581299, "genuineness": 4.394378662109375, "identity": 0.6067076325416565, "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_CONV_reg/bbcc7cc4365670c91c62d71733548fa30f25ca21891ba06555a85294a7fbd4a3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "bbe72a833a1df31a558a4011c73ebdcf64aabe475b5253cb7c54b1661437ac3b", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "071165c7ff8968c6edb9858b364588281fb45840108ed908895a523b777a9c64", "generation_seconds_batch": 19.0542571679689, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.767324447631836, "target_z": 0.9453124193783914, "ce": 6.181467056274414, "genuineness": 4.362998008728027, "identity": 0.778105616569519, "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_CONV_reg/bbe72a833a1df31a558a4011c73ebdcf64aabe475b5253cb7c54b1661437ac3b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "bc9f8ef95ac076417ebfc13f3ef50c7b9fa767a02b5d50bf26b591161ad6f507", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 129, "seconds_generated": 10.32, "mp3_sha256": "4d44c4d28ecb02a33a1f0e3c0f5969195b846d39690e0f3edf932a95db65ad27", "generation_seconds_batch": 15.854248579009436, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.777522563934326, "target_z": 0.9531249700344463, "ce": 6.260893821716309, "genuineness": 4.7081804275512695, "identity": 0.5149523019790649, "hypothesis": "I found the little box. It was beside the old window. We can look inside together.", "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_CONV_reg/bc9f8ef95ac076417ebfc13f3ef50c7b9fa767a02b5d50bf26b591161ad6f507.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": "bd7a4d59997ceb5f566158e162c6deaa82f087c1902aa9b0936155b8c15b94f4", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ab2cf502c59ea822778b742a2df60ed8bb7fc6acaa4989876218e31a4b5e3327", "generation_seconds_batch": 15.12899301899597, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.86930513381958, "target_z": 1.023437560644573, "ce": 6.181467056274414, "genuineness": 4.570107460021973, "identity": 0.8410601615905762, "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_CONV_reg/bd7a4d59997ceb5f566158e162c6deaa82f087c1902aa9b0936155b8c15b94f4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "bde374e7fa39aea7c09a609daf9413c1e67991402edf928d57a9ab59be896b38", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "bfc9446c234aa279159935751059b93a396d8833b3454e1f84cb4693b668b539", "generation_seconds_batch": 23.800228476990014, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6111207008361816, "target_z": 0.059570379119846965, "ce": 6.1020402908325195, "genuineness": 4.406930446624756, "identity": 0.3990287184715271, "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_CONV_reg/bde374e7fa39aea7c09a609daf9413c1e67991402edf928d57a9ab59be896b38.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": "bdfdf53cbbe79df13e99e44488ccec9b711479f8ccb0521d033676f92cbf5c67", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Someone moved the little blue chair.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "1f3d52e35f16e8efa2946cebca9104d62ac0f8072c79f695ab0aa70131bc8654", "generation_seconds_batch": 28.00463885697536, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3950958251953125, "target_z": 0.6601563295514103, "ce": 6.332377910614014, "genuineness": 4.482243061065674, "identity": 0.48395994305610657, "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_CONV_reg/bdfdf53cbbe79df13e99e44488ccec9b711479f8ccb0521d033676f92cbf5c67.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "be71afe868ece2b72afacbad9c50a6f9b243d530dd147abc920e14e80b872618", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8cc72c9a0156b975550ef7bc4870ba9a46824183fbb5500bbec2f71ab61c7085", "generation_seconds_batch": 26.17489609203767, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5582644939422607, "target_z": 0.7851562268123692, "ce": 6.086154937744141, "genuineness": 4.331617832183838, "identity": 0.678800642490387, "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_CONV_reg/be71afe868ece2b72afacbad9c50a6f9b243d530dd147abc920e14e80b872618.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "beb3e44e41fe02cb90bb3de00eec918127bafc9e3ef9016639488e146daf5592", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "f7db2a9c3bd72dc0016fcab9268bd743c66d86303d9685499cd185bfcb201893", "generation_seconds_batch": 24.521114674047567, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6417148113250732, "target_z": 0.08300784844082774, "ce": 6.443574905395508, "genuineness": 4.231201171875, "identity": 0.5192002058029175, "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_CONV_reg/beb3e44e41fe02cb90bb3de00eec918127bafc9e3ef9016639488e146daf5592.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "bebdf60c9d4440d032677a63b2a2d2e384da369c2522013525ae7fa9a01ba8a8", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "01f32b46d06fdc9f9b569ce2bf9c7ce52f1470d64230dd9341153ce67a91c314", "generation_seconds_batch": 23.337745286058635, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9387331008911133, "target_z": 0.31054697078275184, "ce": 5.704906940460205, "genuineness": 4.519899368286133, "identity": 0.5160769820213318, "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_CONV_reg/bebdf60c9d4440d032677a63b2a2d2e384da369c2522013525ae7fa9a01ba8a8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "bec600b4fd3d2d056c791ac6be38d09c244e64869049c9bb14c55a2cf0509ce3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "1647cf48858dfee70bc5b75e83e0458e68223c46fcb7d90b1291b8ec290d470f", "generation_seconds_batch": 22.351519957068376, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4511849880218506, "target_z": 0.7031249928653444, "ce": 6.062326908111572, "genuineness": 4.325342178344727, "identity": 0.5320979356765747, "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_CONV_reg/bec600b4fd3d2d056c791ac6be38d09c244e64869049c9bb14c55a2cf0509ce3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "bf050899a28a97fa54f4e8134977695b6c296c0fc3b26f27bed6713081a3e835", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "c70d2813b01ee2e05434b9031bb51c85c0471d05ac7e8356938e76758b223a20", "generation_seconds_batch": 24.494189434917644, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.818314790725708, "target_z": 0.9843749900114821, "ce": 6.23706579208374, "genuineness": 4.20609712600708, "identity": 0.17485563457012177, "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_CONV_reg/bf050899a28a97fa54f4e8134977695b6c296c0fc3b26f27bed6713081a3e835.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "bf0533e1e44d89d5da1d27047eafb09619348abc535f3661b5563bf98d7ea097", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "06ee3a1f0f263d82447b010e61eb64ca46ee98e1b0483759d48b58c685d343a2", "generation_seconds_batch": 26.55085543403402, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2574219703674316, "target_z": 0.5546874436362205, "ce": 6.1894097328186035, "genuineness": 4.570107460021973, "identity": 0.775860071182251, "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_CONV_reg/bf0533e1e44d89d5da1d27047eafb09619348abc535f3661b5563bf98d7ea097.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": "bf8076b0c08662519b3620cd933151a721cd6e581eab811c20c1fe772c776296", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The keys are on the kitchen table.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "5d93a455b96cc2956621b83e080e794f1d3a51095c029ad407b84b0ae6fa1169", "generation_seconds_batch": 15.288394376984797, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.461383104324341, "target_z": 0.7109375435213994, "ce": 6.308549880981445, "genuineness": 4.293961524963379, "identity": 0.40952572226524353, "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_CONV_reg/bf8076b0c08662519b3620cd933151a721cd6e581eab811c20c1fe772c776296.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "bfb682a7958944e17e4197573ced5698e454e447f905ac96a6170aba507b36da", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7e0d1b9dbdd5caa75c7a88fe3b5a6fa5e0837f4ff64187d2592dbb917ec01bd8", "generation_seconds_batch": 17.945745402015746, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5939576625823975, "target_z": 0.8124999714613774, "ce": 5.7128496170043945, "genuineness": 4.306513786315918, "identity": 0.40841951966285706, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. 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_CONV_reg/bfb682a7958944e17e4197573ced5698e454e447f905ac96a6170aba507b36da.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "c04144a660519706b9d0c631af9056af4e81f6b233009e9d710eebb6f137ee10", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "57e45954cd8ec0218557cbc6b4db05b832d67a11c86a6b9c4ce2b5ffe67b0602", "generation_seconds_batch": 24.00714112597052, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6347498893737793, "target_z": 0.8437499914384132, "ce": 6.316492557525635, "genuineness": 4.519899368286133, "identity": 0.15772415697574615, "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_CONV_reg/c04144a660519706b9d0c631af9056af4e81f6b233009e9d710eebb6f137ee10.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "c0a55a17b2e52111882819aa7606d7921ca4fda790133a693bab6253c8ac0a1c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "456aca6bfec5f8e9110c4a2f4fae363766a5b7f4bb310c5069063aaded2323ed", "generation_seconds_batch": 22.59630184993148, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.646176338195801, "target_z": 0.0864257251983617, "ce": 6.284721851348877, "genuineness": 4.632867813110352, "identity": 0.6388956308364868, "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_CONV_reg/c0a55a17b2e52111882819aa7606d7921ca4fda790133a693bab6253c8ac0a1c.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": "c0b91c2f931d5d7be7cc200d71bb99f55d74900312c88eee883fd0615dc7fe9d", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "6bbd414081ac7cde708b4be2f9e098a880779b2aa4b5f42b3106feaa1fafd58c", "generation_seconds_batch": 15.538347241934389, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.522571325302124, "target_z": 0.757812482163361, "ce": 6.038498878479004, "genuineness": 4.519899368286133, "identity": 0.6328812837600708, "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_CONV_reg/c0b91c2f931d5d7be7cc200d71bb99f55d74900312c88eee883fd0615dc7fe9d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "c0befe4779940bf77f9a60592b65fec14eb6ccef4ba8f9b58c6cd6d5e7d21cd3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7c8be08b4a67467bcfaa899b59549344d3d9fe47c134a881fa7529de8abe470d", "generation_seconds_batch": 25.17384561302606, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.920295238494873, "target_z": 1.0624999486304794, "ce": 5.990842819213867, "genuineness": 4.570107460021973, "identity": 0.613946795463562, "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_CONV_reg/c0befe4779940bf77f9a60592b65fec14eb6ccef4ba8f9b58c6cd6d5e7d21cd3.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": "c0eff14bc4f57f545647f40f5a70bd0e3b1d535e5a4a19cee9d87aab19c313e4", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La carretera gira a la izquierda tras el puente.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sigue... hasta la pequeña puerta de madera.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "12600575e6893640fdcfa9b9a758b9614f62f28fa4d4f911bd603bb0a068a81b", "generation_seconds_batch": 15.058864225982688, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.976975679397583, "target_z": 0.3398437617721818, "ce": 6.125868320465088, "genuineness": 4.595211505889893, "identity": 0.6568862199783325, "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_CONV_reg/c0eff14bc4f57f545647f40f5a70bd0e3b1d535e5a4a19cee9d87aab19c313e4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "c16171347f5586a4ec487079290a384606f00e8dbbbcfbb05d65baf58dc9ccc9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "87c8862bd39871d123980ec71981751b101e11bf8c32ec0c48dea6c9a13bdc5d", "generation_seconds_batch": 25.613215090008453, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6857399940490723, "target_z": 0.8828123794243198, "ce": 6.0146708488464355, "genuineness": 4.482243061065674, "identity": 0.5220911502838135, "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_CONV_reg/c16171347f5586a4ec487079290a384606f00e8dbbbcfbb05d65baf58dc9ccc9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "c17442e0eff1e4f3652eb249deb23a23cfeb64f68acd334e09bee7b9691f298a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "12da762b791a81691a2f2d865e0922c4682c3dabff4339315c4027c700ea73a4", "generation_seconds_batch": 24.986891867010854, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4951179027557373, "target_z": -0.029296790989429933, "ce": 6.324435234069824, "genuineness": 4.3504462242126465, "identity": 0.5812474489212036, "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_CONV_reg/c17442e0eff1e4f3652eb249deb23a23cfeb64f68acd334e09bee7b9691f298a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "c188242f78d0d66560aabca69fcd78c83b4e1954da5b530d95c8d8ffc3c9a4bc", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7bd50ef60bcf8c3e01aba90bfd56f776f4421be40e88a0ed2acb057e6688314b", "generation_seconds_batch": 16.74724169599358, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5888586044311523, "target_z": 0.80859369613335, "ce": 6.006728172302246, "genuineness": 4.469690799713135, "identity": 0.6651358604431152, "hypothesis": "De deur staat nog open. Ik heb je jas op de stoel gelegd. 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_CONV_reg/c188242f78d0d66560aabca69fcd78c83b4e1954da5b530d95c8d8ffc3c9a4bc.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "c19b95ac90825236be28f3cf5b7309363f1b573661f60dd9b49041af3e8afc10", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "acb0d95f72ef2ec660e0cac03243e8a41fdd2c2a084094a67a615494faddb1d0", "generation_seconds_batch": 17.33382919197902, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7367305755615234, "target_z": 0.9218751327045948, "ce": 6.038498878479004, "genuineness": 4.846253395080566, "identity": 0.7649016976356506, "hypothesis": "I've trouved the petite boîte. Elle était près de la vie fenêtre.", "wer": 0.5789473684210527, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/c19b95ac90825236be28f3cf5b7309363f1b573661f60dd9b49041af3e8afc10.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "c1a61426caac23ab4f9ed862e3db6c8f52b88dff98024b1b99582aab7faaa78c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "fcf1124afb23d1d0eb38930278d343cf6dc35cf6728388751712ee7d3d8b8838", "generation_seconds_batch": 22.93993466498796, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1911346912384033, "target_z": 0.5039062296662314, "ce": 5.855817794799805, "genuineness": 4.570107460021973, "identity": 0.5845733284950256, "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": 0.03999999999999915, "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_CONV_reg/c1a61426caac23ab4f9ed862e3db6c8f52b88dff98024b1b99582aab7faaa78c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "c1f45d2d8a2099dd44fca86b36f8ed92bff7dc1a285a7f838cf067818d8162d4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7c1e480b71978771ab9ce5895d79f3904bed73941fa66a4607cd1badb05cd98e", "generation_seconds_batch": 22.158697345061228, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9463815689086914, "target_z": 0.316406292451201, "ce": 6.260893821716309, "genuineness": 4.582659721374512, "identity": 0.45498037338256836, "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_CONV_reg/c1f45d2d8a2099dd44fca86b36f8ed92bff7dc1a285a7f838cf067818d8162d4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "c231cc60894ec06c9c2fa64dc9af38626583cc05109742e9fa29eb8bc4d5d7cf", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e960d5c32a7ad4a1e86ed000ff4897ea683bb366ff566c760f3c5c6136fe2784", "generation_seconds_batch": 15.585673489025794, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5187008380889893, "target_z": -0.011230427413588932, "ce": 6.141753673553467, "genuineness": 3.7416701316833496, "identity": 0.28003644943237305, "hypothesis": "The door is still open. I'd put your coat on the chair. Tell me when you are ready.", "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_CONV_reg/c231cc60894ec06c9c2fa64dc9af38626583cc05109742e9fa29eb8bc4d5d7cf.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "c2ed659992cd332924fbc58f2aa8223141903bb69d395fa52571184738616ba4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4fb080c36f5944e469e8757540e0f69abeed8d372b7f8483194380d59a533efa", "generation_seconds_batch": 17.976073677069508, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4562840461730957, "target_z": 0.7070312681933718, "ce": 6.09409761428833, "genuineness": 4.7458367347717285, "identity": 0.7468891739845276, "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_CONV_reg/c2ed659992cd332924fbc58f2aa8223141903bb69d395fa52571184738616ba4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "c2f026c76e0576e85e5cbf5b0716b3b0005fc2d12b4d2102879f037d385472fa", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "02763a01a3160e6ad7f67e64789637f30f51ce6e1a148a4c3d6b9209e5ec07aa", "generation_seconds_batch": 23.862091740011238, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6857399940490723, "target_z": 0.8828123794243198, "ce": 6.221180438995361, "genuineness": 4.356721878051758, "identity": 0.1511324942111969, "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_CONV_reg/c2f026c76e0576e85e5cbf5b0716b3b0005fc2d12b4d2102879f037d385472fa.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "c30d77eb0513114350571e656918430d3c9916c2f74f65ea882238e327ac7fce", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b6cef9141c23afe5bf4a92d876cac3001313ccd579e1ae042f38f358a3b44070", "generation_seconds_batch": 23.66273452702444, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.813215732574463, "target_z": 0.9804687146834546, "ce": 6.332377910614014, "genuineness": 4.6956281661987305, "identity": 0.6261481046676636, "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_CONV_reg/c30d77eb0513114350571e656918430d3c9916c2f74f65ea882238e327ac7fce.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "c37ed0958719acb1291ad594033bc240f4571d9910dc5ee0e30dcbe6b81c9436", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "dc36a40b640d648654a2c07e374884f46738c65bbf4953dd85f909f9a27886d0", "generation_seconds_batch": 23.477816154947504, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5327694416046143, "target_z": 0.7656250328194159, "ce": 5.935244083404541, "genuineness": 4.400654315948486, "identity": 0.5738030076026917, "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_CONV_reg/c37ed0958719acb1291ad594033bc240f4571d9910dc5ee0e30dcbe6b81c9436.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "220080b40e9fe07bf1c94f01d40023abf9346aa30bda0e491659e35927b4022a", "generation_seconds_batch": 17.481514143990353, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7959601879119873, "target_z": 0.20117180953953456, "ce": 6.062326908111572, "genuineness": 4.419482707977295, "identity": 0.6551313996315002, "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_CONV_reg/c3977db0634952a7f9e4083e6135fbbc87199d34c7cb8c3338515d9ae38c18f4.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": "c3e69fdddabd4d6410fed3a31de3f554d685540eee98b42cf37cf234bea3d49d", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 129, "seconds_generated": 10.32, "mp3_sha256": "a68e98a3ac93e672b944a367832ca50b28c69297c141e7c3acfaf1ee703ac276", "generation_seconds_batch": 23.261017984012142, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4868781566619873, "target_z": 0.7304687375143526, "ce": 6.260893821716309, "genuineness": 4.444586753845215, "identity": 0.5610723495483398, "hypothesis": "Die Schlüssel liegen auf dem Küchentisch. Ich suche in der Schublade. Dann neben dem Fenster. Nimm sie mit, wenn du gehst.", "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_CONV_reg/c3e69fdddabd4d6410fed3a31de3f554d685540eee98b42cf37cf234bea3d49d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "c413ac7ce9323ad6a4bf895bc83d88737776f98c7cef9777b8eccb46341ba702", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "fd057b5176b6b43ec4369d388f08c7fc8829be81f783c1a9969614c6ae6dc861", "generation_seconds_batch": 23.111041725962423, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.665343999862671, "target_z": 0.867187460759394, "ce": 6.316492557525635, "genuineness": 4.507347106933594, "identity": 0.3297908902168274, "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_CONV_reg/c413ac7ce9323ad6a4bf895bc83d88737776f98c7cef9777b8eccb46341ba702.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "c41827434a6c38f0b93f1f95531234ce157021a2e78f3569390d8ecab8cef880", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "81e6ff623f47bc86d343d2cff9a080a57ba393bd114cd341d67d86717d3d7267", "generation_seconds_batch": 24.221836103941314, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.231926918029785, "target_z": 0.5351562496432672, "ce": 6.125868320465088, "genuineness": 4.250029563903809, "identity": 0.31222617626190186, "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_CONV_reg/c41827434a6c38f0b93f1f95531234ce157021a2e78f3569390d8ecab8cef880.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "f3d5a99833cf59026af813aee1853e0a726dbc7420f98169e146bc0eacfd633b", "generation_seconds_batch": 18.134511395008303, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3899967670440674, "target_z": 0.6562500542233828, "ce": 6.229123115539551, "genuineness": 4.432034492492676, "identity": 0.7065010666847229, "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_CONV_reg/c493619abd8899c9b3ce96f2a00a6bcdd900e874ef21436707b986365a8437ee.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "c49cef3d2f30ed571f4b43d7882705879f277a949824f5edd9f7902ab16852b4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7d842d499b101f671f26c160bdd1f83f37977e6c92db4d5d668bb9213e85b409", "generation_seconds_batch": 18.337746488046832, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.520612955093384, "target_z": -0.009765596996476644, "ce": 6.125868320465088, "genuineness": 4.65797233581543, "identity": 0.6636327505111694, "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_CONV_reg/c49cef3d2f30ed571f4b43d7882705879f277a949824f5edd9f7902ab16852b4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "c4ddcfa24f69a686083d40a1d3bb193470c1aa1e93ef81098582060b659d5d6a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5a67dd346867342c9b8b7faaa1305914d1d8b37fb898c52f786600997161bc9a", "generation_seconds_batch": 18.853981095016934, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5665042400360107, "target_z": 0.025390698308586618, "ce": 6.062326908111572, "genuineness": 4.444586753845215, "identity": 0.3064245581626892, "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_CONV_reg/c4ddcfa24f69a686083d40a1d3bb193470c1aa1e93ef81098582060b659d5d6a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "c53716272904962d49cb6e91b19bd46d830de072fcc34e17131007a9eb336649", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5572351971a9794b4786f0f37b900af9e8bf3d13d5f4a2aacac4506090e77999", "generation_seconds_batch": 23.128393740043975, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.359402656555176, "target_z": 0.632812584902402, "ce": 5.959072113037109, "genuineness": 4.394378662109375, "identity": 0.44013115763664246, "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_CONV_reg/c53716272904962d49cb6e91b19bd46d830de072fcc34e17131007a9eb336649.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "c55d0dd4ce1079a203a11df62d119a6a393f169cdd71d75528d5d9b3f6b8e038", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c1f0babdc999fc5f1d5df88a9e6610e948e1a1c9c1bd064d561c0d292793329a", "generation_seconds_batch": 23.963112149038352, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7061362266540527, "target_z": 0.8984374807364298, "ce": 5.911416053771973, "genuineness": 4.545003414154053, "identity": 0.7639771699905396, "hypothesis": "La porte étant 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_CONV_reg/c55d0dd4ce1079a203a11df62d119a6a393f169cdd71d75528d5d9b3f6b8e038.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": "c57a2835394330f16ac8c6bf448872a8b83bb74025fddd012a266f1a3309924f", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "8eb5bfe53fad09d5d4d9a9289ecb691c3d60025ce9a49cca08430ad1f3bbc2e8", "generation_seconds_batch": 23.419356156955473, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3492045402526855, "target_z": 0.625000034246347, "ce": 6.046441555023193, "genuineness": 4.432034492492676, "identity": 0.631778359413147, "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_CONV_reg/c57a2835394330f16ac8c6bf448872a8b83bb74025fddd012a266f1a3309924f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "c58c4c1fbb2d738fb12e461ef1d57654d37f60c004c98fe8c8f681b7dfce4cda", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "96961e0de35ee473c7ab15b4e2f4d8d4891fb011f82f2bc1bbe35164126fba95", "generation_seconds_batch": 24.61782585992478, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4919772148132324, "target_z": 0.7343750128423802, "ce": 6.022613525390625, "genuineness": 4.331617832183838, "identity": 0.455795556306839, "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_CONV_reg/c58c4c1fbb2d738fb12e461ef1d57654d37f60c004c98fe8c8f681b7dfce4cda.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "c631f20c3c558064e8e4a872d3b5bdc8aab35f87717f52726bab13e28709de80", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "961f92b29e8dc20af021a8b163f0520685d4f33b2ce33a34f9869221cd5ce958", "generation_seconds_batch": 18.22077953000553, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.466482162475586, "target_z": 0.7148438188494268, "ce": 5.95112943649292, "genuineness": 4.482243061065674, "identity": 0.6728489398956299, "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": 0.03999999999999915, "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_CONV_reg/c631f20c3c558064e8e4a872d3b5bdc8aab35f87717f52726bab13e28709de80.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "c64a05f66d2d9e1cb531c23e391cb56f42869c83d2b41f8a3b856664df075060", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "1a2e1bbec6ecbaf384dd204c8303ab28cc769015365a7ba160ff7a2bee41e52b", "generation_seconds_batch": 24.836885498953052, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.103860378265381, "target_z": 1.2031251298507326, "ce": 6.125868320465088, "genuineness": 4.331617832183838, "identity": 0.39792001247406006, "hypothesis": "The door um is lopen. I put your coat on the chair. Tell me when you are ready.", "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_CONV_reg/c64a05f66d2d9e1cb531c23e391cb56f42869c83d2b41f8a3b856664df075060.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "c74a33f2fee0eb44ab4d9a12802347bf6992fc3a9ee4321988115a2f25d3674a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c6250583331e52b67fd8fbca2e0ff118b0627921079a25ea1e0bcb353c30ccb0", "generation_seconds_batch": 18.337746488046832, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.06306791305542, "target_z": 1.1718749272265125, "ce": 5.839932441711426, "genuineness": 4.582659721374512, "identity": 0.6436237096786499, "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_CONV_reg/c74a33f2fee0eb44ab4d9a12802347bf6992fc3a9ee4321988115a2f25d3674a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "c74b2b37046a6107a3bbd41194484443af13628f002a243f1a31e68baab02d8e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8a36658d39c29cd44b17c42b324fab3d1df1b9531e204a9e7bcbfac7162afaaa", "generation_seconds_batch": 23.43021843000315, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.180936813354492, "target_z": 0.49609386165736064, "ce": 6.23706579208374, "genuineness": 3.986435890197754, "identity": 0.26151856780052185, "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_CONV_reg/c74b2b37046a6107a3bbd41194484443af13628f002a243f1a31e68baab02d8e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "c765d85f49e495a789bb4bbe347d5d5d346926537180ca9684c5cef073f69cd3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "269e29ae80b2645695d4977d11dfb863b44aaa1cc8a065ba73a3a277af5e4076", "generation_seconds_batch": 24.446531435940415, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6041557788848877, "target_z": 0.8203125221174324, "ce": 6.292664527893066, "genuineness": 4.306513786315918, "identity": 0.4253416061401367, "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_CONV_reg/c765d85f49e495a789bb4bbe347d5d5d346926537180ca9684c5cef073f69cd3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "c77b91a2138d8af4ad98c4c327f46a2af2c221fd7fe8ee9c92faef306762b4b6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b22ce046d2c3fe98bed3b36a5c2b985adc4b334a4caa61c155dd710db44c6c8d", "generation_seconds_batch": 23.66536195005756, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0560104846954346, "target_z": 0.4003905727386475, "ce": 6.173524379730225, "genuineness": 4.099404335021973, "identity": 0.33751583099365234, "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_CONV_reg/c77b91a2138d8af4ad98c4c327f46a2af2c221fd7fe8ee9c92faef306762b4b6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "c7d5f8fbd184ebe680516e75aa44d338385fe02fed88cae958ece9b9588d7043", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "14c7506f6e82f9cfe0717fd734ec7be310d38a2c4c7e3f64fa7224f9f2c11cfc", "generation_seconds_batch": 24.825870117987506, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7080020904541016, "target_z": 0.1337890624108168, "ce": 5.462655067443848, "genuineness": 4.024092197418213, "identity": 0.6651618480682373, "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_CONV_reg/c7d5f8fbd184ebe680516e75aa44d338385fe02fed88cae958ece9b9588d7043.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "c7e86b657f3e8712263cdbb99215e2174fbb90331e43c4e7de70e978e69d9b7f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "20528719015a3095eeca13f00d36609ea4b4628fd8c799e1010925d75a98a1e2", "generation_seconds_batch": 17.146196016925387, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.12994647026062, "target_z": 0.45703129102426987, "ce": 6.292664527893066, "genuineness": 4.15588903427124, "identity": 0.3505643606185913, "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_CONV_reg/c7e86b657f3e8712263cdbb99215e2174fbb90331e43c4e7de70e978e69d9b7f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "c80b4a5f677046fb5c48f6de634b1c50641517ebd0054be9ede9ae3712b5ffd6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "69b9584533314e00f145c3e1f157cfc0279d1e212fb9b6c0b5e310d2fd4c9014", "generation_seconds_batch": 25.0208580349572, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.240166664123535, "target_z": -0.22460927886051535, "ce": 6.395919322967529, "genuineness": 4.080576419830322, "identity": 0.6153900623321533, "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_CONV_reg/c80b4a5f677046fb5c48f6de634b1c50641517ebd0054be9ede9ae3712b5ffd6.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "ee8a20014e3eb88f35efbb7431fbdfdab9c5708adc4c221576be38f01d0b56bf", "generation_seconds_batch": 18.964713198016398, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.614353656768799, "target_z": 0.8281248901263033, "ce": 6.284721851348877, "genuineness": 4.381826400756836, "identity": 0.6267172694206238, "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_CONV_reg/c82131bfaf89850d394bf3a21117deb94436b65306054c3e1cbac0c65e2f913d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "c86d6577b4a040f12669fa1c0dff84051fca603e53203f568a70665e6ef97d1c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3b3db1eb2e2f742c4145bd4e3be04c09c4dbc6f0e45720633a84c8bf7dd491a9", "generation_seconds_batch": 23.673998876009136, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1401443481445312, "target_z": 0.4648436590331407, "ce": 6.197352409362793, "genuineness": 4.532451152801514, "identity": 0.5888797044754028, "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_CONV_reg/c86d6577b4a040f12669fa1c0dff84051fca603e53203f568a70665e6ef97d1c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "c870067a24623799ff3544d42727d2441fdc6dc5228d85564e243b1ea0c71334", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8ca7254b2064e8c0bdfdce7f7200f503af29dce5e7b2388440bf758f3af7b9b6", "generation_seconds_batch": 26.560341875068843, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.721433162689209, "target_z": 0.9101561240733281, "ce": 5.990842819213867, "genuineness": 4.494794845581055, "identity": 0.3158443570137024, "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_CONV_reg/c870067a24623799ff3544d42727d2441fdc6dc5228d85564e243b1ea0c71334.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "c8721b1c4e164c4578b9a746f937b0d8d221cf6bf197e237ac077155708cc5f2", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1eb0825d2bd76ad7294bf0ffee8032854e057a12cd766ecec6432c0bac443696", "generation_seconds_batch": 17.40324608690571, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6805949211120605, "target_z": 0.11279303800075123, "ce": 6.133810997009277, "genuineness": 4.331617832183838, "identity": 0.6729445457458496, "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_CONV_reg/c8721b1c4e164c4578b9a746f937b0d8d221cf6bf197e237ac077155708cc5f2.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "3cc01c027a4abcdb91017f1c383b5f9bb967c5b3ef8adb679d11bab9f45fbeff", "generation_seconds_batch": 18.20118688198272, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1477930545806885, "target_z": 0.47070316334877405, "ce": 6.41180419921875, "genuineness": 4.519899368286133, "identity": 0.4822850823402405, "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_CONV_reg/c8907269f81c0a33f09ba3b9d9f76b6acb2a8250e6881dbc11786cea01719e9d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "c8c47ebfe1569ce09c5b2f7a6b54560d52e988e3e40c7c2edc18004549333898", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "88f96e54abd16cc25299ddc5194c3ea1513bae94967a7660777d107d7e050bef", "generation_seconds_batch": 17.900141792953946, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9814834594726562, "target_z": 1.109374887272441, "ce": 5.974957466125488, "genuineness": 4.7458367347717285, "identity": 0.8063446879386902, "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_CONV_reg/c8c47ebfe1569ce09c5b2f7a6b54560d52e988e3e40c7c2edc18004549333898.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": "c8f62f46e578639bb03c7b7a63ea2d9a7d41e44f005a44def6fdbd34f4315bfc", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d378a5d7b2c2fb80dfd6bb6c8174ef036d9ada929d4c61b4c7d36efd51a901c2", "generation_seconds_batch": 15.111216426012106, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.103860378265381, "target_z": 1.2031251298507326, "ce": 6.23706579208374, "genuineness": 4.720732688903809, "identity": 0.7754678726196289, "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_CONV_reg/c8f62f46e578639bb03c7b7a63ea2d9a7d41e44f005a44def6fdbd34f4315bfc.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": "c918153abcb1f05246a950f26e04310d725c96b84e9db9a7e8f053c48cf7bc9a", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I wrote the number on the envelope.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "ea08f375c87a565ea401cc8351cad11c8a1518a2e5dc46656bf095baa3343e46", "generation_seconds_batch": 17.457468157052062, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.721433162689209, "target_z": 0.9101561240733281, "ce": 6.340320587158203, "genuineness": 4.030367851257324, "identity": 0.312576562166214, "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_CONV_reg/c918153abcb1f05246a950f26e04310d725c96b84e9db9a7e8f053c48cf7bc9a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "c95688e4832c54f1f50dd106114d84fecbcceed951b2ff09f502fffdb8e5250b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "689cdbda86c86876b26debdb23da1034ffab0c185423f5cae424954fa94fadb7", "generation_seconds_batch": 22.165463162004016, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2574219703674316, "target_z": 0.5546874436362205, "ce": 6.173524379730225, "genuineness": 4.519899368286133, "identity": 0.5882818698883057, "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_CONV_reg/c95688e4832c54f1f50dd106114d84fecbcceed951b2ff09f502fffdb8e5250b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "c9754db56c64de4019a5298e10cb0ffcc3f885a9d3ad36466568c2963a3e1c29", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c93259fc25084b8ccfa9fc2125d6583716204f3bb4ee44f83422b6ec97d2492f", "generation_seconds_batch": 26.196921192924492, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2982141971588135, "target_z": 0.5859374636132563, "ce": 6.324435234069824, "genuineness": 4.632867813110352, "identity": 0.5890129804611206, "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_CONV_reg/c9754db56c64de4019a5298e10cb0ffcc3f885a9d3ad36466568c2963a3e1c29.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "c9ada12c091625bfdaf20e3d0ed12832b5f71e24c11a5161616bb9dd4db477f8", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "fdfdb5597a1d29519c36248c27dc3881d049f3afa10a6b6ce677619681c36c38", "generation_seconds_batch": 23.234237976022996, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4970762729644775, "target_z": 0.7382812881704076, "ce": 5.982900142669678, "genuineness": 4.055472373962402, "identity": 0.4702845811843872, "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_CONV_reg/c9ada12c091625bfdaf20e3d0ed12832b5f71e24c11a5161616bb9dd4db477f8.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": "ca77676f64351565b58894eb9f117d0f9a36051fe5716ad4a59bf08281eb842c", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "e25c6cab741010a05e46cd6c0a16811c0198ed74e70399e728ac05337b25ba0d", "generation_seconds_batch": 15.112993807997555, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.490018844604492, "target_z": -0.033203066317457425, "ce": 5.903473377227783, "genuineness": 4.369274139404297, "identity": 0.7287078499794006, "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_CONV_reg/ca77676f64351565b58894eb9f117d0f9a36051fe5716ad4a59bf08281eb842c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "ca9386ea2c0fc1f6d2caa96a32c89891a6bffd891f3e8e7eeead10cc77b49cab", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a4ffd4dc87cf8cf413c058db02a71fb743b421eb0f87978684a1b03cf19440a4", "generation_seconds_batch": 19.196984502952546, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1528921127319336, "target_z": 0.4746094386768015, "ce": 5.7247633934021, "genuineness": 4.457139015197754, "identity": 0.6543139219284058, "hypothesis": "La puerta sí 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_CONV_reg/ca9386ea2c0fc1f6d2caa96a32c89891a6bffd891f3e8e7eeead10cc77b49cab.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "caf3d5e0873c3f3ec6d1d7cb5fbfe54bed551b9216f8c82710c10c7f495897a2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "addf5226f4ea036f4e29961cb6541c5d6f30c11555721f85b721567b5ffd28fc", "generation_seconds_batch": 23.254128345055506, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1095502376556396, "target_z": 0.4414061897121599, "ce": 6.181467056274414, "genuineness": 4.607763767242432, "identity": 0.4111393094062805, "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_CONV_reg/caf3d5e0873c3f3ec6d1d7cb5fbfe54bed551b9216f8c82710c10c7f495897a2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "cb17fd8078db01dddb37f637b308a46fff28b0712518b538e1974426154bc304", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d9b9a8dd8d0f3463973d2d40d3f46a608980c140118abea920ac00fe07db7858", "generation_seconds_batch": 23.66273452702444, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.124256134033203, "target_z": 1.2187498658684741, "ce": 5.895530700683594, "genuineness": 4.532451152801514, "identity": 0.6410071849822998, "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_CONV_reg/cb17fd8078db01dddb37f637b308a46fff28b0712518b538e1974426154bc304.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "cb27fc1cc46f3b4b846336a13bfd8e2ef27da1ddff9d8be89f389b5aa0e1ab71", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b89b754a7c53caea49babbf835eaf1536b15b203a3c5f18adb41a068183987e1", "generation_seconds_batch": 17.18454066000413, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6857399940490723, "target_z": 0.8828123794243198, "ce": 6.2767791748046875, "genuineness": 4.519899368286133, "identity": 0.4086970388889313, "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_CONV_reg/cb27fc1cc46f3b4b846336a13bfd8e2ef27da1ddff9d8be89f389b5aa0e1ab71.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "cc1b8cb19c5341e2659e3d1a0952026b6362338b83cd2a7f0dedf5ec6009f32c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "fbcc7cdba6a8bc49d3664635a52ea37ff512214326fed8b7ca357ce52147d519", "generation_seconds_batch": 22.576245508971624, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5837597846984863, "target_z": 0.8046876034525067, "ce": 6.109982967376709, "genuineness": 4.281409740447998, "identity": 0.20722077786922455, "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_CONV_reg/cc1b8cb19c5341e2659e3d1a0952026b6362338b83cd2a7f0dedf5ec6009f32c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "cc1e6a04cb1524c09f21d74450a3658c1d27b9c4369767e6ef5d0260ea01dbb9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "20e2f08ed67bdc27243a3a1f98ebc5ece3a7737d08282b405be94afd348b30b9", "generation_seconds_batch": 24.707960744970478, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0050203800201416, "target_z": 0.3613281847527409, "ce": 6.213237762451172, "genuineness": 4.570107460021973, "identity": 0.7043202519416809, "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_CONV_reg/cc1e6a04cb1524c09f21d74450a3658c1d27b9c4369767e6ef5d0260ea01dbb9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "cc69228a261c58cb5f7e8b4b8ae9bf4ea0ef54ebe9bdce2a5a8207744d41a659", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f30c04fd31c53828a6fd695dac9e6da6ad19cb7ec5f75815f33cfca29c82a1ed", "generation_seconds_batch": 17.18454066000413, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.369600534439087, "target_z": 0.6406249529112729, "ce": 6.2767791748046875, "genuineness": 4.293961524963379, "identity": 0.4258041977882385, "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_CONV_reg/cc69228a261c58cb5f7e8b4b8ae9bf4ea0ef54ebe9bdce2a5a8207744d41a659.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": "ccb055fba2b5daaa12940f2e5750323a61654b369bc4c319763fe7080fec549f", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "d127c9a39fad8b1b8e26629598b6727167d01e2a446a4dfea53285d4e10fbd3a", "generation_seconds_batch": 15.060472958022729, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4568750858306885, "target_z": -0.05859376462604404, "ce": 6.173524379730225, "genuineness": 4.18099308013916, "identity": 0.44228827953338623, "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_CONV_reg/ccb055fba2b5daaa12940f2e5750323a61654b369bc4c319763fe7080fec549f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "cd7eca063b99e6ad15826fbd1db57ec7412d637b7a5e09891534b421464f5b31", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b513368605206f65142e4385955967e314a4b01576905a4ab4b6ec9ccd6ab918", "generation_seconds_batch": 24.07227897492703, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3899967670440674, "target_z": 0.6562500542233828, "ce": 6.2767791748046875, "genuineness": 4.595211505889893, "identity": 0.610771656036377, "hypothesis": "La porte et ma importance est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt.", "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_CONV_reg/cd7eca063b99e6ad15826fbd1db57ec7412d637b7a5e09891534b421464f5b31.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "cd9c9d231a716506564d84d91395bb8d3bdeb1651f4ffb985df475fa51c5c0f1", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5c23383f7a69b9a5207d324f12857a60a191fe67e3a7fc715242c076a5a164cf", "generation_seconds_batch": 25.953460013959557, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.318019390106201, "target_z": 1.3671875977447823, "ce": 6.181467056274414, "genuineness": 4.482243061065674, "identity": 0.7230380177497864, "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_CONV_reg/cd9c9d231a716506564d84d91395bb8d3bdeb1651f4ffb985df475fa51c5c0f1.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "cdc89e3a4c5335cc8c97c18177b9886721bbe012fbbeaf1f24383b9f34c13056", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3f9768cdbffab66c222990f7422247b2e4ccdbd3e562af0282658bad7d63b966", "generation_seconds_batch": 17.78694170701783, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.737321376800537, "target_z": 0.15624991723799467, "ce": 5.593709468841553, "genuineness": 4.425758361816406, "identity": 0.745107114315033, "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_CONV_reg/cdc89e3a4c5335cc8c97c18177b9886721bbe012fbbeaf1f24383b9f34c13056.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "cdcb475515828d08309eedca5f44d16317007924beff9315e1b81a58a8148abd", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "21a715dc60da7c921085276fcf4143dd96cc02946e28d8b1b60fea5ad27fb59e", "generation_seconds_batch": 23.86763852695003, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.931084394454956, "target_z": 0.30468746646711853, "ce": 6.078212261199951, "genuineness": 4.457139015197754, "identity": 0.5823176503181458, "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_CONV_reg/cdcb475515828d08309eedca5f44d16317007924beff9315e1b81a58a8148abd.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "ce0334fb171fac1f860fc45529e3f17e744c7f4bf51289f78a869398f5049883", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "871e5404b93fb2394fc4524ad8d2b1fba63ee83f1986fceac2163fb3ba911a19", "generation_seconds_batch": 24.474205303005874, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.078956127166748, "target_z": 0.4179687203911791, "ce": 6.1894097328186035, "genuineness": 4.545003414154053, "identity": 0.7357943654060364, "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_CONV_reg/ce0334fb171fac1f860fc45529e3f17e744c7f4bf51289f78a869398f5049883.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "ce0dda70e87e22756959fb40f4380c1c9248db77a0f272e517f4f4f816a9f4ff", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "421cca9274163d83ba1ff09349c0911529cd9401577b948665c18b1d5396f06d", "generation_seconds_batch": 23.793451146106236, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.578660488128662, "target_z": 0.800781145477295, "ce": 5.974957466125488, "genuineness": 4.883909225463867, "identity": 0.7499593496322632, "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_CONV_reg/ce0dda70e87e22756959fb40f4380c1c9248db77a0f272e517f4f4f816a9f4ff.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "ce59441c76f14ea2a50be1b40b9e2ed1191d74bb0a0572fe1899da34d412faf6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "dc38c0ba77db5c6af3c0d68a9506fa5b13d5ce1fdfc2038d212529ec92350946", "generation_seconds_batch": 16.634745346964337, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.971285581588745, "target_z": 1.1015625192635703, "ce": 6.046441555023193, "genuineness": 4.720732688903809, "identity": 0.7893891334533691, "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_CONV_reg/ce59441c76f14ea2a50be1b40b9e2ed1191d74bb0a0572fe1899da34d412faf6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "ce73ebe5fccfc95162e5bdd05ed8833476eaa5ee6ce21372501615aef3b01fd0", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "80658670c76de2ccab0f01c3414d0fcf1cff087180f79a7956f410caf4f851ce", "generation_seconds_batch": 16.11871103802696, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.450593948364258, "target_z": 1.4687500256847603, "ce": 6.157639026641846, "genuineness": 4.293961524963379, "identity": 0.44734591245651245, "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_CONV_reg/ce73ebe5fccfc95162e5bdd05ed8833476eaa5ee6ce21372501615aef3b01fd0.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "cf20c5ffb17c6a8c5521f12c34de98039b2a51817282fab0ecff068b3a70fe34", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d71587fd0f431abcdb6084c9692cd353f8c88cd79509848c5f487d7671cc6c2d", "generation_seconds_batch": 27.043465691967867, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9897232055664062, "target_z": 0.34960935876865845, "ce": 5.95112943649292, "genuineness": 4.394378662109375, "identity": 0.5222043991088867, "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_CONV_reg/cf20c5ffb17c6a8c5521f12c34de98039b2a51817282fab0ecff068b3a70fe34.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": "cf3f587568fc0ff67b572a8b258c17912fec1e1d1141bd59167ff504c694f41d", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "a167edc8b75ce15c5823c1056cda3c241f9041aa2a234d55b886067316aac4a9", "generation_seconds_batch": 17.386320078046992, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.001879692077637, "target_z": 1.124999988584551, "ce": 6.149696350097656, "genuineness": 4.337893962860107, "identity": 0.5277951955795288, "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": 0.03999999999999915, "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_CONV_reg/cf3f587568fc0ff67b572a8b258c17912fec1e1d1141bd59167ff504c694f41d.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "c8fd6ef9406638c2569e5e40f58c13530488337378de72edcc93de27907f7440", "generation_seconds_batch": 17.409522486967035, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0942533016204834, "target_z": 0.42968754637526163, "ce": 6.316492557525635, "genuineness": 4.7332844734191895, "identity": 0.6235452890396118, "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_CONV_reg/cf68e6506caaaad7e7371e72de9d5417fef89211300487ff83acd2484eaab510.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "d047457b63ae200824b194145cb9eed4c098e2d1a418ef69a85374be6ce85e3d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "996fb00863f34c7ece0aeb11eb89e594f746bd52e937afc7eb5500395ce12f6f", "generation_seconds_batch": 17.25354571407661, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1962337493896484, "target_z": 0.5078125049942589, "ce": 6.221180438995361, "genuineness": 4.632867813110352, "identity": 0.5779173970222473, "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_CONV_reg/d047457b63ae200824b194145cb9eed4c098e2d1a418ef69a85374be6ce85e3d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "d06f95c4f7e26dfb8095e7cdce5fc1da28397916dab34bd8b18f4dfe35abb2a2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a24ba2c48d4f6a82132f8bc33921b01c5ea6a7ebf34e302e494dcbafe4e2c3cc", "generation_seconds_batch": 24.474205303005874, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.2874250411987305, "target_z": 1.343749945776617, "ce": 6.0146708488464355, "genuineness": 4.645420074462891, "identity": 0.7155128717422485, "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_CONV_reg/d06f95c4f7e26dfb8095e7cdce5fc1da28397916dab34bd8b18f4dfe35abb2a2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "d08a8d834999bcb2bacef2fb83e458c36c9b849fa79d46966c96fa78ede5dd89", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "051f52ca0b7ec5411c6a6bf5ee9e6c7b36dea9a89a3b944c293feb0aef4af992", "generation_seconds_batch": 24.314591440954246, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.267620086669922, "target_z": 0.5624999942922755, "ce": 6.0305562019348145, "genuineness": 4.582659721374512, "identity": 0.6217189431190491, "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_CONV_reg/d08a8d834999bcb2bacef2fb83e458c36c9b849fa79d46966c96fa78ede5dd89.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": "d09b754788eef520e6bf60b0b21920c9cfcda614b81a65ee00f376f4afbe6801", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I wrote the number on the envelope.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Read it carefully... and check the last two digits.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "71fb8c3aca903f2e84a7b42307f9fe367fbf2ad7d3dbbc2718ca0cb8e19a8aa7", "generation_seconds_batch": 15.262272540014237, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8336119651794434, "target_z": 0.9960938159955646, "ce": 6.348263263702393, "genuineness": 4.268857479095459, "identity": 0.34802332520484924, "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_CONV_reg/d09b754788eef520e6bf60b0b21920c9cfcda614b81a65ee00f376f4afbe6801.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "d0a381cc66b54b4232bf7bb100af3ef156cd4aecb5cf3fba33b1f96a51574c55", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "240110282086f7eb28886338cbce01ec8aef24745e6bc204c39b1930de8f2068", "generation_seconds_batch": 18.09734173596371, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.67740797996521, "target_z": 0.11035159308983603, "ce": 6.316492557525635, "genuineness": 4.162164688110352, "identity": 0.608706533908844, "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_CONV_reg/d0a381cc66b54b4232bf7bb100af3ef156cd4aecb5cf3fba33b1f96a51574c55.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "d100ab89b1ea49a74451d22442ea9032d8e7dbe95470a1f073485bfbbefa7fcb", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d8446a76084c08e124e651198d9bc0539018ffb49885d99e62ec672298c6ac65", "generation_seconds_batch": 24.049307239009067, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7877206802368164, "target_z": 0.9609375206905013, "ce": 6.157639026641846, "genuineness": 4.632867813110352, "identity": 0.3292837142944336, "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_CONV_reg/d100ab89b1ea49a74451d22442ea9032d8e7dbe95470a1f073485bfbbefa7fcb.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "e0b3d71a27f97ea0b021f2b2bca9d7c8fd22ddc46bab055ec7f66d7591c986a6", "generation_seconds_batch": 18.48193127696868, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9514806270599365, "target_z": 0.3203125677792285, "ce": 6.356205940246582, "genuineness": 3.8922953605651855, "identity": 0.3399515151977539, "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_CONV_reg/d11caac31fa3732e2517803823f1b039213fb31e9438619fd83d8b232c894b7d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "d219b89efd1063b0cbc6d2363bb3e1be026a0c442097d03626fafd58f627f618", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "9d85a4bd6889c6b6db75c62d95760cd724c87d14aab197c5d9cc03cdf7c8e0e8", "generation_seconds_batch": 24.298144386033528, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8647971153259277, "target_z": 0.25390625249712945, "ce": 6.205295085906982, "genuineness": 4.068024158477783, "identity": 0.2846851646900177, "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_CONV_reg/d219b89efd1063b0cbc6d2363bb3e1be026a0c442097d03626fafd58f627f618.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "de806317ba0a57d388174947624857ba9f902f2131ea8034ce25ebab0073e4dd", "generation_seconds_batch": 17.610762433963828, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6570117473602295, "target_z": 0.09472649177772605, "ce": 6.427689552307129, "genuineness": 4.256305694580078, "identity": 0.5330102443695068, "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_CONV_reg/d2319f0ba4d4a13f099c61fd2524d2cd47bb09bca092815ed8a29e7f5b0c527c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "d2b2bedcb1278ae396baa0be7bff5e886208acf1dba2c21afbab4c6501ed8588", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "33f9dbbbd892535af00ec2bdbe82591f762d997b6df208b5a73afcbe96f430e9", "generation_seconds_batch": 24.164906044956297, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8795032501220703, "target_z": 1.0312501113006278, "ce": 5.824047088623047, "genuineness": 4.444586753845215, "identity": 0.6435421705245972, "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_CONV_reg/d2b2bedcb1278ae396baa0be7bff5e886208acf1dba2c21afbab4c6501ed8588.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "d305db6bfd1f10451d1fb91573a09046dfdece563497a8ad92b3f1abd4255ef4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0b78282abac5aeb34aa490b9684226922bcc6a195e7af32b34af090cf52fae36", "generation_seconds_batch": 23.58173772203736, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4358878135681152, "target_z": 0.6914061668812619, "ce": 5.998785495758057, "genuineness": 4.362998008728027, "identity": 0.7816145420074463, "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_CONV_reg/d305db6bfd1f10451d1fb91573a09046dfdece563497a8ad92b3f1abd4255ef4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "d322a850bb47e2061e9cc0b7764408b733fbcec160594d56b45155121ccc0b1a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4a27254ab2f86980f108e4fcbe406d155ba9f2863a44267cffb1c5d2c3e0d0ef", "generation_seconds_batch": 17.294604912051, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7367305755615234, "target_z": 0.9218751327045948, "ce": 6.348263263702393, "genuineness": 4.65797233581543, "identity": 0.6046695709228516, "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_CONV_reg/d322a850bb47e2061e9cc0b7764408b733fbcec160594d56b45155121ccc0b1a.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": "d352eabeb89bea7eae0746197c1ea6ced227ead6989c3cc4d4a3e73068cdb318", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7aa0b2983df0798d919666ab0222d4a8f68d0df36fbc4fa8cb74e62e46cebffb", "generation_seconds_batch": 17.52500246197451, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.2931151390075684, "target_z": 0.5820311882852288, "ce": 5.744620323181152, "genuineness": 4.6705241203308105, "identity": 0.78352290391922, "hypothesis": "J'ai écrit le 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_CONV_reg/d352eabeb89bea7eae0746197c1ea6ced227ead6989c3cc4d4a3e73068cdb318.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": "d37e892a36ec336402af948c68b0a4b7ec4a2bf70ed25c74c737cd91632144bc", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Iemand heeft het kleine blauwe stoeltje verplaatst.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond hier... naast de hoge boekenkast.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "644e256a9fc7d4b05c9f69b026ea4274481b0241c3bff9a932a20ced6fa54673", "generation_seconds_batch": 15.068975197034888, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.318610429763794, "target_z": 0.6015625649253663, "ce": 6.125868320465088, "genuineness": 4.419482707977295, "identity": 0.7505067586898804, "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_CONV_reg/d37e892a36ec336402af948c68b0a4b7ec4a2bf70ed25c74c737cd91632144bc.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "d391de9a59afe58abf621a5cfa438e749ad5168b09699545552260d9953d9b0c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "b6ffbda91cf792f70c8f82492e4521478c0e6bfbeb9cbf92e0ec125649cf4af5", "generation_seconds_batch": 16.963250310043804, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.26702880859375, "target_z": 1.3281248444645073, "ce": 5.855817794799805, "genuineness": 4.507347106933594, "identity": 0.6764553785324097, "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": 0.03999999999999915, "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_CONV_reg/d391de9a59afe58abf621a5cfa438e749ad5168b09699545552260d9953d9b0c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "d3a28943c154fa18c917d9f5b3fd552baf2d2b2996f0763d6d1840c4e01ee21c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "bf634f5c43784880bbcc22d4ac87e02f44ca36317b514b6513b515d2c362f8af", "generation_seconds_batch": 21.782437655027024, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1095502376556396, "target_z": 0.4414061897121599, "ce": 6.0305562019348145, "genuineness": 4.557555198669434, "identity": 0.689103364944458, "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_CONV_reg/d3a28943c154fa18c917d9f5b3fd552baf2d2b2996f0763d6d1840c4e01ee21c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "d3c89d8bfcf306841db45bfe428b0e346a2d0e7656d2c4e56fd947f55d27df1f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b33a3f98aa58ffade6272998eae5955060f7a9a8e00b93bdd69972d69b2588ef", "generation_seconds_batch": 22.1952163890237, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8998992443084717, "target_z": 1.0468750299655536, "ce": 6.173524379730225, "genuineness": 4.545003414154053, "identity": 0.3454431891441345, "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_CONV_reg/d3c89d8bfcf306841db45bfe428b0e346a2d0e7656d2c4e56fd947f55d27df1f.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "4ca1d8aad760235cd39d9f650c8b5929e8fd813803c574f4cb83d7eced774046", "generation_seconds_batch": 17.683924596989527, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3523449897766113, "target_z": -0.13867195223264722, "ce": 6.054384231567383, "genuineness": 4.432034492492676, "identity": 0.7915458679199219, "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_CONV_reg/d3fab0ead1fc9dcf899f59bddb0c9d4b54373ba1829b6dc49f77fc4027ce1740.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "d41b0a2252d1a563c2f307356243b3714aba13fce2efa81179c41c35f8261064", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9f74d8fc8e2438fc46504119a6e0ce42b1b2c90f883fdc88bd20dae465e97b53", "generation_seconds_batch": 25.583739039953798, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4919772148132324, "target_z": 0.7343750128423802, "ce": 6.229123115539551, "genuineness": 4.570107460021973, "identity": 0.6505135297775269, "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_CONV_reg/d41b0a2252d1a563c2f307356243b3714aba13fce2efa81179c41c35f8261064.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "d472814c1440ee2bd201dd8739a2b7bac3b672b42cdfb650f30a0c8dfcfa9265", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a5d14e537b4875ed4341962a0c035df80be928c156aa71ec75b373cd2cb64040", "generation_seconds_batch": 22.7725300360471, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3084123134613037, "target_z": 0.5937500142693113, "ce": 6.332377910614014, "genuineness": 4.519899368286133, "identity": 0.39984798431396484, "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_CONV_reg/d472814c1440ee2bd201dd8739a2b7bac3b672b42cdfb650f30a0c8dfcfa9265.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "d4c7053d55c9c811b94ff13551d5dea0722058dc67d51f7ebb3b7ece4c20a13e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "841717693ba8ea298a7efa6e584c3f181c118bccd553001d855dfbaa64c16c1b", "generation_seconds_batch": 26.34194855997339, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9463815689086914, "target_z": 0.316406292451201, "ce": 6.181467056274414, "genuineness": 4.406930446624756, "identity": 0.48098477721214294, "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_CONV_reg/d4c7053d55c9c811b94ff13551d5dea0722058dc67d51f7ebb3b7ece4c20a13e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "d53b8ee7e7ee4f52b2866e2bae776efdff29a02315724604692e4733ece0abd8", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "c0343ca5f3ba4b96ed94c1b73ddb8a39b3974dd71a4060b6e57e1c4aaff04cee", "generation_seconds_batch": 17.76990672107786, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5429673194885254, "target_z": 0.7734374008282867, "ce": 6.24500846862793, "genuineness": 4.262581825256348, "identity": 0.2782478928565979, "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_CONV_reg/d53b8ee7e7ee4f52b2866e2bae776efdff29a02315724604692e4733ece0abd8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "d5f9b38b7ae4e4d5f7bc3252ccc4d12740c75e0430c29d34ce03b79aca371d87", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f4b2ba80791ca128a327f409e5091cb415c7934546b5704c43e668934def7658", "generation_seconds_batch": 17.557103842962533, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.726532459259033, "target_z": 0.9140625820485397, "ce": 6.205295085906982, "genuineness": 4.6203155517578125, "identity": 0.48047858476638794, "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_CONV_reg/d5f9b38b7ae4e4d5f7bc3252ccc4d12740c75e0430c29d34ce03b79aca371d87.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "d6a1963d25cf9141622d0d273d6f51604e9bc95d8f7ae29f57bcd2d8eb00069a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9a6f1db67e82479660fca4b22b8981944ed0664d974bd86c376c730c8cbbe191", "generation_seconds_batch": 17.377999630058184, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.563363552093506, "target_z": 0.7890625021403966, "ce": 6.117925643920898, "genuineness": 4.833701133728027, "identity": 0.6019917726516724, "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_CONV_reg/d6a1963d25cf9141622d0d273d6f51604e9bc95d8f7ae29f57bcd2d8eb00069a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "d703d70f6cc8c9f67cb2b005d675548220c297f0c7e4e0f1c4bc12da06f670e2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "226356335072158c2fc591ccf4ed2e237198849c773bf1fdacc9c05c700284c1", "generation_seconds_batch": 24.481433523003943, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.622593402862549, "target_z": 0.0683593616225207, "ce": 6.141753673553467, "genuineness": 4.268857479095459, "identity": 0.47258272767066956, "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_CONV_reg/d703d70f6cc8c9f67cb2b005d675548220c297f0c7e4e0f1c4bc12da06f670e2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "d7122defd8dcd1c69782514e4544d1095585786fe544e16068cd1e4b595c4030", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4c5cf2c2fc833c0d135c9228d8cb9d7da6a46fc599d2d34dd7bc008451519816", "generation_seconds_batch": 23.79511560301762, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.001879692077637, "target_z": 1.124999988584551, "ce": 6.308549880981445, "genuineness": 4.557555198669434, "identity": 0.38458698987960815, "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_CONV_reg/d7122defd8dcd1c69782514e4544d1095585786fe544e16068cd1e4b595c4030.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "d717af90c2eeb43b18d294be52ed1e44f3862ac67359897f0a4c7b8dea34b19f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a88931d12e728fab63325c6eb6050bd4fb535fc88d76136432f299a5ce0c3788", "generation_seconds_batch": 23.97112848702818, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9508895874023438, "target_z": 1.0859376005986443, "ce": 6.3641486167907715, "genuineness": 4.607763767242432, "identity": 0.751822829246521, "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_CONV_reg/d717af90c2eeb43b18d294be52ed1e44f3862ac67359897f0a4c7b8dea34b19f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "d72d05987666b2c923219f8fb685a04018c48b5b620d9c8e34f05843a0d02546", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d5889d7d7ef7eb89459db09458ade184dfc5d29dec3f500070c232150dce862b", "generation_seconds_batch": 24.718449481064454, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6857399940490723, "target_z": 0.8828123794243198, "ce": 6.1020402908325195, "genuineness": 4.331617832183838, "identity": 0.3418312072753906, "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_CONV_reg/d72d05987666b2c923219f8fb685a04018c48b5b620d9c8e34f05843a0d02546.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "d768c68709eb4e2339ec3ec74176ad6c5883932db31bccaa9717caec4a4127da", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "81b4971d26d8e3883ca778bf7e22d079e9359adcd9654fc61ad0043f8fd0484d", "generation_seconds_batch": 24.312398375943303, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.229968547821045, "target_z": -0.23242182951657034, "ce": 5.855817794799805, "genuineness": 4.111956596374512, "identity": 0.653599739074707, "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_CONV_reg/d768c68709eb4e2339ec3ec74176ad6c5883932db31bccaa9717caec4a4127da.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "d77c7343a4139444a0bf8530a67a26b3d27ed99dbdf07b7fb19857351def61b3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5bcebb41ce0f4ea528babce1fdd3df2514411a1ad56f56176d00ec50d3e44af8", "generation_seconds_batch": 20.255277213989757, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1095502376556396, "target_z": 0.4414061897121599, "ce": 6.078212261199951, "genuineness": 4.419482707977295, "identity": 0.515214204788208, "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_CONV_reg/d77c7343a4139444a0bf8530a67a26b3d27ed99dbdf07b7fb19857351def61b3.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": "d83589a217c5f84faa129251dc628be71e7d7b44d8af0cc74b33d7d893b29041", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Someone moved the little blue chair.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "323f800f1214abc017160b7dd43d38b9a2e9bb208d0b7836aefd36038d63cb89", "generation_seconds_batch": 28.00463885697536, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7367305755615234, "target_z": 0.9218751327045948, "ce": 6.38797664642334, "genuineness": 4.6203155517578125, "identity": 0.39482903480529785, "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_CONV_reg/d83589a217c5f84faa129251dc628be71e7d7b44d8af0cc74b33d7d893b29041.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "d842a7fee218fb76593a5a3fac5962128e5e51cd8001385e613f057fec0b3386", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5cb28ea17eea799dd61cc0d3e87396fd7eb223f2996a86380a3d670a9056e177", "generation_seconds_batch": 22.165463162004016, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.379798650741577, "target_z": 0.6484375035673278, "ce": 6.268836498260498, "genuineness": 4.607763767242432, "identity": 0.406146764755249, "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_CONV_reg/d842a7fee218fb76593a5a3fac5962128e5e51cd8001385e613f057fec0b3386.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "d863b814b904c69e529ddc55787fc8abbd249aa42fc9c4aa94d4c1b2bf7937a7", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ee708f07961644889200eb1adc06e2f00862d817daf7c3b03ff13e266483adf4", "generation_seconds_batch": 16.11406943597831, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.716334342956543, "target_z": 0.9062500313924848, "ce": 6.149696350097656, "genuineness": 4.632867813110352, "identity": 0.5730412006378174, "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_CONV_reg/d863b814b904c69e529ddc55787fc8abbd249aa42fc9c4aa94d4c1b2bf7937a7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "d87bd2cef8aae50e5f4366393157f4207999e50fc09224d405281b2eb10e7840", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "bd1328ca89044a1ec76210cb5d933a1ffbc9e849f58683289cfc3be92eca9e23", "generation_seconds_batch": 25.786177610978484, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.042672157287598, "target_z": 1.156250191208771, "ce": 6.419747352600098, "genuineness": 4.7834930419921875, "identity": 0.24997074902057648, "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_CONV_reg/d87bd2cef8aae50e5f4366393157f4207999e50fc09224d405281b2eb10e7840.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "d8f0d063ea2be6aaf5ac397c558e5ddceffb73dc4ea91db70297582e9b094b5a", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9bb93aa3e9e8d2d60092d19859c97ea4c0f2c5b846aefdbce524c43547b612b6", "generation_seconds_batch": 17.493057081010193, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.923435926437378, "target_z": 0.29882814479866937, "ce": 6.1020402908325195, "genuineness": 4.720732688903809, "identity": 0.5626739859580994, "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_CONV_reg/d8f0d063ea2be6aaf5ac397c558e5ddceffb73dc4ea91db70297582e9b094b5a.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": "d91d5e2bcc750a78308a4ac4ef09a39620d2b83fcad91f86bd7b49b3c785a336", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ton message est arrivé après mon départ.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Je l’ai relu... puis j’ai rangé mon téléphone.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "ea67470a2a0d844e70583c57be9ae860aef4208063c1965f0c02215dabe37196", "generation_seconds_batch": 25.05367086699698, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.16818904876709, "target_z": 0.4863280820136998, "ce": 6.109982967376709, "genuineness": 4.413206577301025, "identity": 0.6446818709373474, "hypothesis": "Ton message is arrivé après mon départ. Je l'ai relu, puis j'ai rangé mon téléphone. Dis-moi quand 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_CONV_reg/d91d5e2bcc750a78308a4ac4ef09a39620d2b83fcad91f86bd7b49b3c785a336.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "d9677b1e0a7f9e9ab75ab90397a806c787a3d52552c06fb30aef80b0997ca20d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "eeae36a5913bf1d4b129aed9f88f7f3602a2e74233b145c858717770cb394c10", "generation_seconds_batch": 21.92747243109625, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.318019390106201, "target_z": 1.3671875977447823, "ce": 5.772419452667236, "genuineness": 4.425758361816406, "identity": 0.16070128977298737, "hypothesis": "Ik heb het 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_CONV_reg/d9677b1e0a7f9e9ab75ab90397a806c787a3d52552c06fb30aef80b0997ca20d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "da03e70195e59242a7e60273f1b95b587bc06f567181d691178ec0cc4e282ee1", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d568d6af2f9a35ccf71c75bb5a3f54c98ef8c4d1e5c377d56f65ea0c0b0b68ed", "generation_seconds_batch": 26.46138009498827, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5142390727996826, "target_z": -0.014648486818307056, "ce": 5.871703147888184, "genuineness": 4.231201171875, "identity": 0.5941187143325806, "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_CONV_reg/da03e70195e59242a7e60273f1b95b587bc06f567181d691178ec0cc4e282ee1.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "da151b003c9e2caada9dbd6c8bef0d6aa33c5824258514e46ffeb8f3d92d7828", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d1597abb5da54b07a0e90481db84ef2fda579e4a8cd18684d7360f3993be4f24", "generation_seconds_batch": 19.307599302963354, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.2529141902923584, "target_z": -0.2148436818640387, "ce": 5.808161735534668, "genuineness": 4.011539936065674, "identity": 0.3508533239364624, "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_CONV_reg/da151b003c9e2caada9dbd6c8bef0d6aa33c5824258514e46ffeb8f3d92d7828.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": "da3a1fde2957a19e02be5b7844a089b0e0c500f834fdf023f1e317394cbb6d80", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Las llaves están en la mesa de la cocina.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Miré en el cajón... y luego junto a la ventana.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "58c49fc449d8aea83743e02cb45b42ff36c6be50ce05eef232d9c8e1d8e94468", "generation_seconds_batch": 17.574974497081712, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3390064239501953, "target_z": 0.6171874835902921, "ce": 5.879645347595215, "genuineness": 4.532451152801514, "identity": 0.7304157018661499, "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_CONV_reg/da3a1fde2957a19e02be5b7844a089b0e0c500f834fdf023f1e317394cbb6d80.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "da54db7a5b9ce8285529dd6a68ad2ce577138df1702bdcb4a9c5629d53770866", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f15964944412d025efae9372768499cf04d617729553620645b29961564c5a73", "generation_seconds_batch": 23.402635614969768, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.910097122192383, "target_z": 1.0546873979744245, "ce": 6.054384231567383, "genuineness": 4.419482707977295, "identity": 0.6187429428100586, "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_CONV_reg/da54db7a5b9ce8285529dd6a68ad2ce577138df1702bdcb4a9c5629d53770866.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "dad834fe43e222f10ca22f7cb33605628456614d390e7a9c0fa8a9dbd163ad76", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "bea46912615f298b04ab4218c2602bd790a65768f3a50fac0ce499756ccb9c2d", "generation_seconds_batch": 19.630737451952882, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3956868648529053, "target_z": -0.10546870326800561, "ce": 6.348263263702393, "genuineness": 4.268857479095459, "identity": 0.4448665678501129, "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_CONV_reg/dad834fe43e222f10ca22f7cb33605628456614d390e7a9c0fa8a9dbd163ad76.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "3801dec2c19c5e996184c6de96871ccd2d35aa0feca7cd88227e4605ef5238ad", "generation_seconds_batch": 18.466284418012947, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.3956868648529053, "target_z": -0.10546870326800561, "ce": 6.300607204437256, "genuineness": 3.8922953605651855, "identity": 0.45374101400375366, "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_CONV_reg/db192910e9ae67de251fb2a990940eab881f66f627c78b50b638ce61f5a63679.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": "db550ca3d34caa3f9e393fb568742e3ab0075fcffaf242b86b624393eab84e9e", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e2d18a27b04475b5d7c027ae756335f8d9b7a180b69ad2ea5b28ea7124c1e1e3", "generation_seconds_batch": 16.04912235995289, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.584350824356079, "target_z": 0.03906257063309076, "ce": 6.133810997009277, "genuineness": 4.432034492492676, "identity": 0.5139949321746826, "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_CONV_reg/db550ca3d34caa3f9e393fb568742e3ab0075fcffaf242b86b624393eab84e9e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "dbafc0dc825dac14345779719000d55630d4c61a931221dfcb6c58c23c24f678", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "3bab856b814c869b075b8c523884f221c62f83ecbcb8749458f6817e954ad79c", "generation_seconds_batch": 24.986368063022383, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0942533016204834, "target_z": 0.42968754637526163, "ce": 6.260893821716309, "genuineness": 4.080576419830322, "identity": 0.36932939291000366, "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_CONV_reg/dbafc0dc825dac14345779719000d55630d4c61a931221dfcb6c58c23c24f678.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "dc4039d412b92b4ce89307bcec765768b8c064d7d51f0f05dec4a375e9a285b9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5207d8d8cd47c8e011f4e45e2588ef333a631c97760ff2d524a735c46d3f2212", "generation_seconds_batch": 22.814648712985218, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.665343999862671, "target_z": 0.867187460759394, "ce": 5.903473377227783, "genuineness": 4.645420074462891, "identity": 0.537577748298645, "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_CONV_reg/dc4039d412b92b4ce89307bcec765768b8c064d7d51f0f05dec4a375e9a285b9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "dcbf5864b1688106aa808f31663db7381c5a3f2486e720a6b3d7080ea60c0609", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2c7b1c76a185a7ef342f06b9ca8e4c2cf6670d2ad48f400376270e8cb57636b0", "generation_seconds_batch": 23.862091740011238, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9922728538513184, "target_z": 0.35156258775626426, "ce": 6.205295085906982, "genuineness": 4.143336772918701, "identity": 0.3057774305343628, "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_CONV_reg/dcbf5864b1688106aa808f31663db7381c5a3f2486e720a6b3d7080ea60c0609.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": "dcc52b729bfd162a39e90e0f2b591d5b33929c5576b6a7902649809ab0be5258", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "93bf16cd8f1b9934b2d9ed9988a49fe11b26478dd51ef36613a059f372aa1a8c", "generation_seconds_batch": 15.254141589975916, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8387107849121094, "target_z": 0.9999999086764079, "ce": 6.078212261199951, "genuineness": 4.570107460021973, "identity": 0.7095952033996582, "hypothesis": "Deine Nachricht kam, als ich schon weg war. Ich las sie zweimal, dann legte ich das Handy weg, sagte 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_CONV_reg/dcc52b729bfd162a39e90e0f2b591d5b33929c5576b6a7902649809ab0be5258.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "ddc4d0e02d424e17475af00ce3e1cd6bb60aed73cf7fbd064ef09cbc32a6caad", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "53bee2394acd7da22c77fac761f7013021ad19cc5b7ddc205be4bd4a38cb559f", "generation_seconds_batch": 24.79566278902348, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.77428936958313, "target_z": 0.18457027638080586, "ce": 6.221180438995361, "genuineness": 4.507347106933594, "identity": 0.7750636339187622, "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_CONV_reg/ddc4d0e02d424e17475af00ce3e1cd6bb60aed73cf7fbd064ef09cbc32a6caad.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "deb679b61612e2ca6d07a7d57bf4dda6c5c45e30e17aca881701fa355f592ad7", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "a5506e0cc48ebb0ea1cf3d985169df9179c1b2bf319fd148ecf0354cb44b2e6e", "generation_seconds_batch": 18.06474249495659, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.885193109512329, "target_z": 0.26953117116205527, "ce": 6.213237762451172, "genuineness": 4.337893962860107, "identity": 0.6629170179367065, "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_CONV_reg/deb679b61612e2ca6d07a7d57bf4dda6c5c45e30e17aca881701fa355f592ad7.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": "defa397ba15ef79bf4b15144c6403e5a8da6d94be382dcd7bd526e1455925f80", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Escribí el número en el sobre.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Léelo bien... y comprueba las dos últimas cifras.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "9751fb2405d4371c54dc1b7fa979ae8fed7a0f4a7fcb7cdbbd464a77b227cb4f", "generation_seconds_batch": 15.024231143994257, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7436952590942383, "target_z": 0.16113280705982508, "ce": 6.070269584655762, "genuineness": 4.331617832183838, "identity": 0.5971671342849731, "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_CONV_reg/defa397ba15ef79bf4b15144c6403e5a8da6d94be382dcd7bd526e1455925f80.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": "df4bebc13c3bde4da85fca5319eb1bde72da5929d2863e6ee8b21ed6e85b063c", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Straße führt hinter der Brücke nach links.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Geh weiter... bis zu dem kleinen Holztor.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "14e96535dcec59feaacb858e15c27cce381a17ae79f2f32f4f2987d17f1afa50", "generation_seconds_batch": 25.643560266005807, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.920886516571045, "target_z": 0.2968750984582477, "ce": 6.133810997009277, "genuineness": 4.720732688903809, "identity": 0.6853266954421997, "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_CONV_reg/df4bebc13c3bde4da85fca5319eb1bde72da5929d2863e6ee8b21ed6e85b063c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "dfab2e9d236c6a4251306a72e5877673f3e803d1fff452efb6a467504ed9e1d9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "091bdac8bc0cc7b9e36db7d47aec37c1ae38d81b939b2b123aa54f0ca41359ba", "generation_seconds_batch": 22.40016520395875, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.132495880126953, "target_z": 0.4589843373646915, "ce": 5.839932441711426, "genuineness": 4.105680465698242, "identity": 0.587009608745575, "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_CONV_reg/dfab2e9d236c6a4251306a72e5877673f3e803d1fff452efb6a467504ed9e1d9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "dfc9ab28bced457b7fce438e6509a41d55a00349d951aacb459ccea4797ae418", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "df92b72c56ba3bf9655a097d4ac36093d620098e639e5d84ed593bc815f7246a", "generation_seconds_batch": 24.33969635493122, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.1854448318481445, "target_z": 1.265625169804804, "ce": 6.2767791748046875, "genuineness": 4.482243061065674, "identity": 0.5352380275726318, "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_CONV_reg/dfc9ab28bced457b7fce438e6509a41d55a00349d951aacb459ccea4797ae418.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "e09e520bf5631fb06c27a73d9d4412eea0b6ed3e7792a3d0b82a22b9d4a3854d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "28383a2a3b2bc88c00b5bf33a887d1b0dc615db2e2f88aa4f9498ad289b89cc6", "generation_seconds_batch": 24.061862407950684, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5837597846984863, "target_z": 0.8046876034525067, "ce": 6.2767791748046875, "genuineness": 4.507347106933594, "identity": 0.4703141152858734, "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_CONV_reg/e09e520bf5631fb06c27a73d9d4412eea0b6ed3e7792a3d0b82a22b9d4a3854d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "e09ef0dd7fffa331e81daa3af3adf5208eae229e162973929f0743827df55743", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "68966d9d5b58a130aa9507e38a6b6a18c242b252886abf437bf8d66c3f3a3304", "generation_seconds_batch": 24.12591344199609, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.481779098510742, "target_z": 0.7265624621863251, "ce": 6.260893821716309, "genuineness": 4.482243061065674, "identity": 0.6555665731430054, "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_CONV_reg/e09ef0dd7fffa331e81daa3af3adf5208eae229e162973929f0743827df55743.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": "e0bea5b9f42e04f50137b39b35319c9261fce92bdb8afde2ab94474dcaa17b68", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "0f8456a21e4292f8a1897c8915a680606d6041aecfb592ebcb9808fdafeffe42", "generation_seconds_batch": 21.434167910018004, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3899967670440674, "target_z": 0.6562500542233828, "ce": 6.086154937744141, "genuineness": 4.7332844734191895, "identity": 0.7347150444984436, "hypothesis": "Les clés sont sur la table de la cuisine, j'ai regardé dans le tiroir, puis t'es près de la fenêtre. Prends-les quand tu partiras.", "wer": 0.041666666666666664, "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_CONV_reg/e0bea5b9f42e04f50137b39b35319c9261fce92bdb8afde2ab94474dcaa17b68.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "e0c0d2e5edb9d97a0bae7d9f6bb0386da754faff97ac0a6e117c097658fe30ef", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2a478e9a4e6bceabd93735cf8a784c254f550733744b7728655d31f060c6ce8f", "generation_seconds_batch": 24.19183660298586, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6041557788848877, "target_z": 0.8203125221174324, "ce": 6.0305562019348145, "genuineness": 4.808597087860107, "identity": 0.671998143196106, "hypothesis": "J'ai trouvé la 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_CONV_reg/e0c0d2e5edb9d97a0bae7d9f6bb0386da754faff97ac0a6e117c097658fe30ef.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "e1467dfaff01ceeac011517eddb948d775209bdc40da6de09158ab2b7b4abfe7", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f569f8c6564b7633861c990440b3f8a6ed3018b811efb0f90f89b4c9cd5f7b56", "generation_seconds_batch": 24.423196024028584, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.205840587615967, "target_z": 1.2812499058225457, "ce": 5.784333229064941, "genuineness": 4.444586753845215, "identity": 0.2527870535850525, "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_CONV_reg/e1467dfaff01ceeac011517eddb948d775209bdc40da6de09158ab2b7b4abfe7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "e14e9f77a265ed5b3de113f5c24e1c197e42eec3faeb5274321ae3fe50d3b338", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "d850493bafafe11aeee98ec7747be03c7fba2165a0f3ad2a32bf6d0933cd6f79", "generation_seconds_batch": 26.484952063998207, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.818314790725708, "target_z": 0.9843749900114821, "ce": 6.006728172302246, "genuineness": 4.413206577301025, "identity": 0.0499175563454628, "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_CONV_reg/e14e9f77a265ed5b3de113f5c24e1c197e42eec3faeb5274321ae3fe50d3b338.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "e1607b229b3878e1157d5570fb45cc1f16b740c0d67f706f933d70e5ff6e8248", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f6910b74b114547aa9397f60f107be24d2742323fd0790ddebcb9958ae176b5b", "generation_seconds_batch": 23.542454715934582, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.726532459259033, "target_z": 0.9140625820485397, "ce": 6.300607204437256, "genuineness": 4.532451152801514, "identity": 0.6251886487007141, "hypothesis": "Ja, 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_CONV_reg/e1607b229b3878e1157d5570fb45cc1f16b740c0d67f706f933d70e5ff6e8248.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "e22f7cba0a604ccb281d2cc919f717e5c33a534dc2abbe463e7b3b79b725cb70", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "fe301c660f5a7bfa074cf5e878e77f0bdea471f01bfe173216387ed794e11f1b", "generation_seconds_batch": 23.378675304935314, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.103860378265381, "target_z": 1.2031251298507326, "ce": 5.959072113037109, "genuineness": 4.406930446624756, "identity": 0.4556308388710022, "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_CONV_reg/e22f7cba0a604ccb281d2cc919f717e5c33a534dc2abbe463e7b3b79b725cb70.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": "e2476c63dc0ee901a06782f7233418dd82d281e437da74ec4a9331a64058e19a", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "54d806dd2e11c9977101c6550210968399e2980ad58edb83874b11f80461a10b", "generation_seconds_batch": 15.060472958022729, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.6633856296539307, "target_z": 0.09960938159955646, "ce": 6.2767791748046875, "genuineness": 4.124508857727051, "identity": 0.6091997623443604, "hypothesis": "Ich schrieb die Nummer auf den Umschlag, lie 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_CONV_reg/e2476c63dc0ee901a06782f7233418dd82d281e437da74ec4a9331a64058e19a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "e26792f2a77623dcbf91836bba55b3c93b9a6d811db5571388ff48bff24f4e9a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d0ada26a587a6790349fee34e2355b60e336df5f0160da63d969abd2fa106627", "generation_seconds_batch": 23.671924520051107, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1503424644470215, "target_z": 0.4726562096891957, "ce": 6.109982967376709, "genuineness": 4.319066047668457, "identity": 0.6711439490318298, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Vélépi la pi. Dis-moi quand tu es prêt.", "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_CONV_reg/e26792f2a77623dcbf91836bba55b3c93b9a6d811db5571388ff48bff24f4e9a.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": "e26b061ef25b01470c40e37c8aa964e20e4e4ace453e19b454346d56720c61c0", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai écrit le numéro sur l’enveloppe.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lis-le bien... et vérifie les deux derniers chiffres.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "6ff40c9c508e52aa50795f5f320ccefe405cae7b1544155bda670ae297041f81", "generation_seconds_batch": 15.237024097004905, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.369600534439087, "target_z": 0.6406249529112729, "ce": 5.959072113037109, "genuineness": 4.607763767242432, "identity": 0.6251000165939331, "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_CONV_reg/e26b061ef25b01470c40e37c8aa964e20e4e4ace453e19b454346d56720c61c0.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "e2801a15efe23098996f0eb83e5b155bdc21b6fd6d66290cf8e271aa64b2bca2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "eb266eaa206871cee12b06ceff2323d31ea4b7b333b337950cb7c2197dce9e8b", "generation_seconds_batch": 23.935674431966618, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0432629585266113, "target_z": 0.39062497574217087, "ce": 6.348263263702393, "genuineness": 4.293961524963379, "identity": 0.42538896203041077, "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_CONV_reg/e2801a15efe23098996f0eb83e5b155bdc21b6fd6d66290cf8e271aa64b2bca2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "e2c2277abac3050f28b2f422c45993eb89258c741ded89b9b3c8dd393b60bc6d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "37b789788f19d595f9e8bef2a0756c4027bfae66a2884768e7f599e696bf1406", "generation_seconds_batch": 23.97112848702818, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.971876621246338, "target_z": 0.3359374864441543, "ce": 6.046441555023193, "genuineness": 4.557555198669434, "identity": 0.7880516052246094, "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_CONV_reg/e2c2277abac3050f28b2f422c45993eb89258c741ded89b9b3c8dd393b60bc6d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "e3a0f5ad8110e4e0d773d7a05ca6ecd5c38896989f9f86a33c3c8bda0169aef6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9bdb94b645f031c430434b10cfb534eb3b3ebe833404f5f04ee692d673a6d31e", "generation_seconds_batch": 16.08214338903781, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7979187965393066, "target_z": 0.9687500713465563, "ce": 5.903473377227783, "genuineness": 4.325342178344727, "identity": 0.4256941080093384, "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_CONV_reg/e3a0f5ad8110e4e0d773d7a05ca6ecd5c38896989f9f86a33c3c8bda0169aef6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "e414d5b9f1658db73caf9ef5e9401ee82a5909a621b9a1e5cfbc8475721d2e30", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5fde19be9f057ea9c34eb1aa20d28d0440e0ce5a5c99e903b11304c95d9629e3", "generation_seconds_batch": 23.338713567936793, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.0834641456604, "target_z": 1.1875000285386226, "ce": 5.8875885009765625, "genuineness": 4.482243061065674, "identity": 0.3670188784599304, "hypothesis": "De deur staat nog open. Ik heb je 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_CONV_reg/e414d5b9f1658db73caf9ef5e9401ee82a5909a621b9a1e5cfbc8475721d2e30.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "e4696a8f1d07d46fd407bf2303a5512033627d1207f9a2a258170541929dec80", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ecc452e47f314688920e7fbe355a1f5de92e5f21dd8a8196ce944e385e3b7f70", "generation_seconds_batch": 22.95933572098147, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8482253551483154, "target_z": 0.24121099466642826, "ce": 6.292664527893066, "genuineness": 4.362998008728027, "identity": 0.653168797492981, "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_CONV_reg/e4696a8f1d07d46fd407bf2303a5512033627d1207f9a2a258170541929dec80.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "e4d4d6bbbe556a4984df1a58babbda8cd387f3c3c841e0c3cd86de6b3d8285e9", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b26a1a571459ccc660e8ed30853ce3a8c622183e29d92127ed2654d02a15a38c", "generation_seconds_batch": 24.61782585992478, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.476680040359497, "target_z": 0.7226561868582977, "ce": 6.046441555023193, "genuineness": 4.281409740447998, "identity": 0.5871902704238892, "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_CONV_reg/e4d4d6bbbe556a4984df1a58babbda8cd387f3c3c841e0c3cd86de6b3d8285e9.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "e5068aeeff497efd94d0b2e47478f7401ddbc37ef70143f103930266a9ecbdad", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d8d9f002146da82fb4a13820122bc505e548cc6bb5111ba21a0fd38a9577a536", "generation_seconds_batch": 23.419881331035867, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.522571325302124, "target_z": 0.757812482163361, "ce": 6.078212261199951, "genuineness": 4.356721878051758, "identity": 0.6942455768585205, "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_CONV_reg/e5068aeeff497efd94d0b2e47478f7401ddbc37ef70143f103930266a9ecbdad.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "e5359c5ebe82cfbe26096b96160669bfa1d8e04be5413951d831c1cc91fc3348", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "bce0c2599387d4d3a57ae93ddd420f9a70a791b198abbf24df8c7896e85860fc", "generation_seconds_batch": 26.42996378301177, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0534610748291016, "target_z": 0.3984375263982258, "ce": 6.252951145172119, "genuineness": 4.6956281661987305, "identity": 0.6492407917976379, "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": 0.03999999999999915, "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_CONV_reg/e5359c5ebe82cfbe26096b96160669bfa1d8e04be5413951d831c1cc91fc3348.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "e539cd6785e8c24cb5306cf29cce5c194d74266e5e6eab00a4ed9f7cc3c9a1b6", "bucket": "vn_S_CONV_reg", "phase": "wording", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d9d0ab6b13104df4afbb1d2e909024b84dde412b29c1da7a9de4377ffb8e1c6c", "generation_seconds_batch": 16.095190137042664, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0101194381713867, "target_z": 0.36523446008076843, "ce": 6.165581703186035, "genuineness": 4.406930446624756, "identity": 0.44330018758773804, "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_CONV_reg/e539cd6785e8c24cb5306cf29cce5c194d74266e5e6eab00a4ed9f7cc3c9a1b6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "e56def814455a2980689bcd75c214f915aa2bc27991168399b5032876df1bf36", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "bf16c580d3ffd6d9a194ac02127d5332cc2b66abb03e262d619d280738f9ad4d", "generation_seconds_batch": 18.913523713010363, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.205840587615967, "target_z": 1.2812499058225457, "ce": 5.863759994506836, "genuineness": 4.821148872375488, "identity": 0.6854843497276306, "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": 0.03999999999999915, "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_CONV_reg/e56def814455a2980689bcd75c214f915aa2bc27991168399b5032876df1bf36.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": "e5d60fd338ac6f30260b11294f7c19448451cbe23d5b4c4b2de8ddd9d1ca6448", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Quelqu’un a déplacé la petite chaise bleue.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était ici... près de la grande bibliothèque.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f1e10c960a5c27ecee0d9a48b529d66f604345e1ee5697d979b98f2f2dad7eae", "generation_seconds_batch": 15.343215593020432, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6041557788848877, "target_z": 0.8203125221174324, "ce": 5.7406487464904785, "genuineness": 4.6705241203308105, "identity": 0.6483298540115356, "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_CONV_reg/e5d60fd338ac6f30260b11294f7c19448451cbe23d5b4c4b2de8ddd9d1ca6448.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "e685bc0ccc1c7ee00995ea4b1643542bf2a7377086e2b71e9be8c98da8a3cd2f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9bcd54806bbdd1df116cdf208f4b2d79de7339f2767ef108aefe20774ace2e74", "generation_seconds_batch": 23.283609776059166, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.599010467529297, "target_z": 0.050292998046679686, "ce": 5.990842819213867, "genuineness": 4.482243061065674, "identity": 0.5197368860244751, "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_CONV_reg/e685bc0ccc1c7ee00995ea4b1643542bf2a7377086e2b71e9be8c98da8a3cd2f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "e6d0edb5e0f97fff6726b44c852f337cf8148fcece23607c386836fb5505bceb", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "e8d31471e9b5a9e2e1de04a8933c461e421a49a83b4b2e7b60ea73bd7fb2c0c6", "generation_seconds_batch": 23.19873085594736, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.094844341278076, "target_z": -0.3359374864441543, "ce": 6.316492557525635, "genuineness": 4.231201171875, "identity": 0.576748251914978, "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_CONV_reg/e6d0edb5e0f97fff6726b44c852f337cf8148fcece23607c386836fb5505bceb.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "eb962d2b2d303948697927010a9c68981d1b84087de01cdc768d2f6d43ff8526", "generation_seconds_batch": 18.20118688198272, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7061362266540527, "target_z": 0.8984374807364298, "ce": 6.459460258483887, "genuineness": 4.406930446624756, "identity": 0.34047335386276245, "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_CONV_reg/e6f1e80cc6606ef00fdc322d5829b94fd65ad80a74ab015f88321e5dc194804c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "e7e6ba7f54ff42b750a6d82dbf3c9751f23813750992d867645ecb3c3f91c28a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "3d6651bad55e8b0c0525c75cc8f8034c7e5b5e08a59dbc02a81106d51c0341f4", "generation_seconds_batch": 23.338713567936793, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.675542116165161, "target_z": 0.875000011415449, "ce": 5.855817794799805, "genuineness": 4.13078498840332, "identity": 0.2927376627922058, "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_CONV_reg/e7e6ba7f54ff42b750a6d82dbf3c9751f23813750992d867645ecb3c3f91c28a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "e8af413b9387ffe445b2b19d84940b7fc1cd8f40a65e6e932950d4a03d8ed71c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "31b031925005b05517512cf3685368876be6bf517e5a942ce032e0cd5ffb5fc9", "generation_seconds_batch": 21.574564099078998, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.297623157501221, "target_z": 1.3515624964326722, "ce": 6.078212261199951, "genuineness": 4.293961524963379, "identity": 0.704830527305603, "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_CONV_reg/e8af413b9387ffe445b2b19d84940b7fc1cd8f40a65e6e932950d4a03d8ed71c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "e8cef0021e2b878f481fa2e6bd237a8315788b3cecf8995887778d584e541537", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "54b5c7ac82799e87e62d395ea10f45631b56f912bb282e4f7d78c2d356668f35", "generation_seconds_batch": 19.299460766022094, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.3492045402526855, "target_z": 0.625000034246347, "ce": 5.998785495758057, "genuineness": 4.080576419830322, "identity": 0.36489129066467285, "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_CONV_reg/e8cef0021e2b878f481fa2e6bd237a8315788b3cecf8995887778d584e541537.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "e8f296d5402bd6feb58074e26805194b60688a1dd1ea7c0cba62616560535655", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3768fb20de1b3cc1a828571a70244ce86abe35cdbb34636916136e3345f0df7c", "generation_seconds_batch": 18.126038942020386, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0993523597717285, "target_z": 0.4335938217032891, "ce": 6.213237762451172, "genuineness": 4.507347106933594, "identity": 0.5811742544174194, "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_CONV_reg/e8f296d5402bd6feb58074e26805194b60688a1dd1ea7c0cba62616560535655.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "63e5190d8e70b711491ef5e46a915f9e2b9a3034e07f071b82042c3d89050ae7", "generation_seconds_batch": 17.409522486967035, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.221728801727295, "target_z": 0.5273436989872122, "ce": 6.3641486167907715, "genuineness": 4.7332844734191895, "identity": 0.3925948441028595, "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_CONV_reg/e9d266c3907611a53d0c5f7688a7e231bba29ead6f87b8ae8d29cb37ab2fae41.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": "e9e36737134f95399831784ce78274c6bc446b0597d88f3a0fa8c29e3d711efa", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b4aee63d060ac70ef93c0fb95fcd333a7d4707b3d3fbde4fbefc2776080688dc", "generation_seconds_batch": 15.111979463952594, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4409868717193604, "target_z": 0.6953124422092893, "ce": 6.157639026641846, "genuineness": 4.7834930419921875, "identity": 0.5309155583381653, "hypothesis": "De sleutels liggen op de keukentafel, kijken de laar, en daarna, naast het raam, neem ze mee als je weggaat.", "wer": 0.18181818181818182, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/e9e36737134f95399831784ce78274c6bc446b0597d88f3a0fa8c29e3d711efa.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "ea4c626a83e601d4aee0af6c56dcd8e436c51b48088d71d0a3ccef865e537592", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4ac653d7bcf87b2cf62e6bb58d3e4fcb0928128388d9ab9ef41108db72fa9446", "generation_seconds_batch": 18.987081220024265, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6500468254089355, "target_z": 0.8554686347753115, "ce": 6.340320587158203, "genuineness": 4.507347106933594, "identity": 0.5288912653923035, "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_CONV_reg/ea4c626a83e601d4aee0af6c56dcd8e436c51b48088d71d0a3ccef865e537592.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": "eabc1fcd286913fe7bcdb33002206fffccc11c06e2b6ac4a61c1165246f273fe", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Les clés sont sur la table de la cuisine.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai regardé dans le tiroir... puis près de la fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "3f08322631e2e06f428de6838b2dc5f6d98fe62eed927e9819e28f0646438307", "generation_seconds_batch": 15.112993807997555, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0968027114868164, "target_z": 0.4316405927156833, "ce": 5.871703147888184, "genuineness": 4.65797233581543, "identity": 0.6361778974533081, "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": 0.03999999999999915, "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_CONV_reg/eabc1fcd286913fe7bcdb33002206fffccc11c06e2b6ac4a61c1165246f273fe.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "eadba68081721c55791190abb0d26af31cdcf724e2965fa29670f9c376dcf01f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ccf823041157327208a3469c8411c704e7c4d2f102d687786134d6706a5e3e04", "generation_seconds_batch": 22.351519957068376, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.267620086669922, "target_z": 0.5624999942922755, "ce": 6.09409761428833, "genuineness": 4.406930446624756, "identity": 0.6718223094940186, "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_CONV_reg/eadba68081721c55791190abb0d26af31cdcf724e2965fa29670f9c376dcf01f.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": "eaf9c182c76567b98a1f0d30dd2c0dbfa77ff76d9b730ef6be54f1efa1ffba26", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Deine Nachricht kam, als ich schon weg war.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich las sie zweimal... dann legte ich das Handy weg.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "58f8ba57954713eb646973a9985295ea586dd16412f3e087f8751855962ca9c3", "generation_seconds_batch": 17.235907964990474, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.920295238494873, "target_z": 1.0624999486304794, "ce": 6.141753673553467, "genuineness": 4.758388519287109, "identity": 0.6731240749359131, "hypothesis": "Deine Nachricht kam, als ich schon weg war. Ich lass sie zweimal, dann legte 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_CONV_reg/eaf9c182c76567b98a1f0d30dd2c0dbfa77ff76d9b730ef6be54f1efa1ffba26.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": "eb1c219a5b5bae0e39fabbd9b787b460c489f9088c09938037adc692bb4c6bb2", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0b2592b5f2c4c6674b7a8571772a41f740b64694b960c216bdbd7bb4c60f07ee", "generation_seconds_batch": 17.555278534069657, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.4460859298706055, "target_z": 0.6992187175373169, "ce": 5.959072113037109, "genuineness": 4.7081804275512695, "identity": 0.6463804244995117, "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_CONV_reg/eb1c219a5b5bae0e39fabbd9b787b460c489f9088c09938037adc692bb4c6bb2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "eb318e17dc856cb51b3936960e12c18066bc9e57dfd46233bae8afd35ca540e3", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "e0166cea0b50d759d9e11e33846419d44c66f305a2b780a8e58beff7bd0f6a0f", "generation_seconds_batch": 20.626477609970607, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.001879692077637, "target_z": 1.124999988584551, "ce": 5.967014789581299, "genuineness": 4.3127899169921875, "identity": 0.7753456830978394, "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_CONV_reg/eb318e17dc856cb51b3936960e12c18066bc9e57dfd46233bae8afd35ca540e3.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "eb7d68a9eb2c82e0d30b76f9ab90bec9ca3167773dd5d585696d659bf84e5e1f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "b285f146644f115a6511118d2c6314352344f1720eae96e0b83eef7a4c745d99", "generation_seconds_batch": 24.481433523003943, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9463815689086914, "target_z": 0.316406292451201, "ce": 6.221180438995361, "genuineness": 4.507347106933594, "identity": 0.5855408906936646, "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_CONV_reg/eb7d68a9eb2c82e0d30b76f9ab90bec9ca3167773dd5d585696d659bf84e5e1f.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "36e71de2b86901ca9a90382aae7fd88a0cdc4d3afb16cf669e3615d0a1af126e", "generation_seconds_batch": 16.938916768995114, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.66019868850708, "target_z": 0.09716793668864125, "ce": 6.038498878479004, "genuineness": 4.256305694580078, "identity": 0.7739014625549316, "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_CONV_reg/ebe89981a89e9506d02bacdfa583f9bfe1f04faa699193749ab6419de85c588f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "ec84e6480ea5602e3dceba0e2e6e6d7fda3bd15fc1b72833f55b0a84ddd939d5", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6e97b86406d994600c61cdee6a92237920479c4a3bcd7813f66755c312cfc60a", "generation_seconds_batch": 16.634745346964337, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.736139297485352, "target_z": 1.6874999828768265, "ce": 5.998785495758057, "genuineness": 4.595211505889893, "identity": 0.7742823362350464, "hypothesis": "He encontrado la cajita. Estaba junto a la. junto a la ventana vieja. ¿Podemos mirar juntos lo que hay dentro?", "wer": 0.17647058823529413, "duration_error_s": 0.040000000000000924, "model_predictions_not_human_ratings": true }, "audio_url": "https://huggingface.co/datasets/laion/Humaneness-Voice-Large-1.5-Intense-Bucket-LoRAs-Demo-Audio/resolve/main/manual-comparison-v1/audio/vn_S_CONV_reg/ec84e6480ea5602e3dceba0e2e6e6d7fda3bd15fc1b72833f55b0a84ddd939d5.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": "eca63b10b1f3ed6026d82a14bf9e298b692f5ac4be2e8a4c613830ccf1b2ac6d", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The keys are on the kitchen table.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I checked the drawer... then looked beside the window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c0b5422a962d660fc41ca4fccbacef74714ceaf7e00078e9885c57e5a2d5e19d", "generation_seconds_batch": 16.802226728061214, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7061362266540527, "target_z": 0.8984374807364298, "ce": 6.268836498260498, "genuineness": 4.262581825256348, "identity": 0.5539467930793762, "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_CONV_reg/eca63b10b1f3ed6026d82a14bf9e298b692f5ac4be2e8a4c613830ccf1b2ac6d.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": "ed520060e24e613d76eeac39cab887b061df19df3828a4101fb999f64cc06b33", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "cbfa61492195cea80790351573a701296d64687cef4b4baa75b0257209317ab9", "generation_seconds_batch": 15.111216426012106, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6857399940490723, "target_z": 0.8828123794243198, "ce": 6.1894097328186035, "genuineness": 4.65797233581543, "identity": 0.6799033880233765, "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_CONV_reg/ed520060e24e613d76eeac39cab887b061df19df3828a4101fb999f64cc06b33.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "ed5f46d27f8ee6c19f9c2f7da6d7eca8838921dbeb9c4f8ff0a8c91117673cae", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "c9da4e8646d6a28c37c306f72a2c5fce604c78ba04614d5b7ebe100f29b7a565", "generation_seconds_batch": 18.22077953000553, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.328808307647705, "target_z": 0.6093749329342371, "ce": 6.022613525390625, "genuineness": 4.808597087860107, "identity": 0.47700315713882446, "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_CONV_reg/ed5f46d27f8ee6c19f9c2f7da6d7eca8838921dbeb9c4f8ff0a8c91117673cae.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": "ed6ef13f222faf69eb22b4f62986e0d78b8c65277c64fdd7a634c2cbfadb0d6d", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "c261fcdf34aa389d76d381d8ec9e4392556479fff226ecaee6c974b4a587d9af", "generation_seconds_batch": 22.54301081900485, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.525712013244629, "target_z": -0.005859321668449151, "ce": 6.086154937744141, "genuineness": 4.306513786315918, "identity": 0.5470579862594604, "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_CONV_reg/ed6ef13f222faf69eb22b4f62986e0d78b8c65277c64fdd7a634c2cbfadb0d6d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "ee10549bc57d98fa616b79532c829347e78a677345d51eeed630dceb9459dc12", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "869ad0f90f5a61074ff7653c693052765eabf5b9dd085aad5dcb6810c4ed1953", "generation_seconds_batch": 18.01774843293242, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.2197704315185547, "target_z": -0.24023438017262533, "ce": 5.911416053771973, "genuineness": 4.18726921081543, "identity": 0.5458105802536011, "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_CONV_reg/ee10549bc57d98fa616b79532c829347e78a677345d51eeed630dceb9459dc12.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": "ee1f4ace4bf18e52e002c8cf98a9be94ca91e11026dcc1f292c105bf814c725e", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La route tourne à gauche après le pont.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Continue... jusqu’à la petite barrière en bois.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "d469d7fc675e38967101f7b9b2fc54d08e9d6855d057f4cef1f9111bce82b263", "generation_seconds_batch": 23.419356156955473, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.77242374420166, "target_z": 0.949218877353603, "ce": 6.09409761428833, "genuineness": 4.532451152801514, "identity": 0.6952086687088013, "hypothesis": "La route tourne 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": 0.03999999999999915, "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_CONV_reg/ee1f4ace4bf18e52e002c8cf98a9be94ca91e11026dcc1f292c105bf814c725e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "ef86b935684cbd3c50a66bdad68d98d6dbfcad82fb8ef7467891b90582f4b0e6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4039530beac96c1369e3f6fba7131509817fe248c56b850b46e6cf0f05ab0d97", "generation_seconds_batch": 25.511373008950613, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.1605405807495117, "target_z": 0.48046876034525066, "ce": 6.157639026641846, "genuineness": 4.607763767242432, "identity": 0.7601807117462158, "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_CONV_reg/ef86b935684cbd3c50a66bdad68d98d6dbfcad82fb8ef7467891b90582f4b0e6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "efb93b74a5fae6635232f0e39ac5b6dab1c5b789f9daee7aa57a823112a3c17b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5c237ca2fbdd4f9dadafd72001e45fa0c388dc7bf60cc5f9f220eef00dfd4530", "generation_seconds_batch": 23.32950896001421, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5575809478759766, "target_z": 0.01855476214633455, "ce": 6.3641486167907715, "genuineness": 4.570107460021973, "identity": 0.6383442878723145, "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_CONV_reg/efb93b74a5fae6635232f0e39ac5b6dab1c5b789f9daee7aa57a823112a3c17b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "eff16ced0ad0fd8bb890db5599f389b3f81696b97228a738a3248fc79252df95", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "93b426886f6299e24846193c8ae039becb5eee363b3dbfa342b40173fb42bc55", "generation_seconds_batch": 22.13093923998531, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8998992443084717, "target_z": 1.0468750299655536, "ce": 6.133810997009277, "genuineness": 4.7834930419921875, "identity": 0.8352092504501343, "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": 0.03999999999999915, "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_CONV_reg/eff16ced0ad0fd8bb890db5599f389b3f81696b97228a738a3248fc79252df95.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "f0267f11d2bb221c819b9f8c70409e8ee812cbb9fe06f17a43d922d19b25be5c", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "adf80c482a746bdb6a243741e35646e2336ff0d1cb11e859b320161c4b7dbd1d", "generation_seconds_batch": 22.59630184993148, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.5027663707733154, "target_z": -0.023437469320980783, "ce": 6.24500846862793, "genuineness": 4.319066047668457, "identity": 0.5716720819473267, "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_CONV_reg/f0267f11d2bb221c819b9f8c70409e8ee812cbb9fe06f17a43d922d19b25be5c.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": "f0c769de0dc7049448d67942b4723e45adc6bde6f983cb29970a719ef1ae2aea", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Schlüssel liegen auf dem Küchentisch.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich suchte in der Schublade... dann neben dem Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b2998efc845c0093f478c81a7cae9df720c558479bf34740bb2b62c7fdd30802", "generation_seconds_batch": 23.261017984012142, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6041557788848877, "target_z": 0.8203125221174324, "ce": 6.213237762451172, "genuineness": 4.494794845581055, "identity": 0.5979930758476257, "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_CONV_reg/f0c769de0dc7049448d67942b4723e45adc6bde6f983cb29970a719ef1ae2aea.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "f14cf165ba8d27f4dc24f5e0bc162abd705c436117654442e835243e86fea136", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "259bab34b229b984f58f87742bb9d331f283a420b4a2664604200e672ece04d1", "generation_seconds_batch": 23.823927422985435, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8998992443084717, "target_z": 1.0468750299655536, "ce": 6.268836498260498, "genuineness": 4.268857479095459, "identity": 0.5909603238105774, "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_CONV_reg/f14cf165ba8d27f4dc24f5e0bc162abd705c436117654442e835243e86fea136.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "f1763963bd64cd3d32b92ba83ba996963663659a453d996753098161382ff780", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "9944fd69c9ac286627bb3dfb95a1ba6b1c1a6284fc41843107993d2a3849e1dd", "generation_seconds_batch": 23.111041725962423, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.660244941711426, "target_z": 0.8632811854313666, "ce": 6.197352409362793, "genuineness": 4.632867813110352, "identity": 0.5021569728851318, "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_CONV_reg/f1763963bd64cd3d32b92ba83ba996963663659a453d996753098161382ff780.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "f1bc0fa7b6ae816d98709b77472390e37b9cb102abd8aa92d9744e3ac789cc03", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6db722e02699dcfa49c32e5854427acae111dcf3b88b15b88e1288c9ff1d6e4f", "generation_seconds_batch": 26.127858088933863, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.589449644088745, "target_z": 0.04296866331393407, "ce": 6.435632705688477, "genuineness": 4.532451152801514, "identity": 0.6487295627593994, "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_CONV_reg/f1bc0fa7b6ae816d98709b77472390e37b9cb102abd8aa92d9744e3ac789cc03.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": "f1cdc091b38f5b904b545ee24caad9316bd9e5c43ce405f15e8afce9a49f5092", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich schrieb die Nummer auf den Umschlag.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lies sie genau... und prüfe die letzten beiden Ziffern.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "02d309259c9ed42d2ee6239f737580174746b424f33df6041ed2536d48d394c3", "generation_seconds_batch": 22.54301081900485, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.728398084640503, "target_z": 0.1494139810757426, "ce": 6.23706579208374, "genuineness": 4.381826400756836, "identity": 0.5074714422225952, "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_CONV_reg/f1cdc091b38f5b904b545ee24caad9316bd9e5c43ce405f15e8afce9a49f5092.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "04d6bb360a184e1951457519870ab43277a76996847acb2c182667079249b752", "generation_seconds_batch": 17.330233343993314, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0407135486602783, "target_z": 0.3886719294017492, "ce": 6.260893821716309, "genuineness": 4.381826400756836, "identity": 0.544090747833252, "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_CONV_reg/f21bc0adf2df6aeda4aa5bc68657c60b57559ffed3b647118e970ebfd0ba6ad8.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "f24b019a7a15bfc5473ce19e8aeff65d43c98dfcae50f1674ffaa7a45a05cc53", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "6924626631ed0afcba9a8dcfe7a450ad50cbcfd1361545f772e124fd1ee13d6d", "generation_seconds_batch": 24.9696722519584, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.726532459259033, "target_z": 0.9140625820485397, "ce": 6.24500846862793, "genuineness": 4.6203155517578125, "identity": 0.6658657789230347, "hypothesis": "La porte est encore 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_CONV_reg/f24b019a7a15bfc5473ce19e8aeff65d43c98dfcae50f1674ffaa7a45a05cc53.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "f313ecbd2c9258ab109355c48bb7504df4cb36299683f9e22be23ae73dc17431", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d91f4fb6ee6210da2c5c730d1c4a58c6faa90301016293db60a679ce9029b318", "generation_seconds_batch": 16.64608475205023, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.382939338684082, "target_z": -0.11523430026448225, "ce": 6.348263263702393, "genuineness": 4.13078498840332, "identity": 0.606544017791748, "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_CONV_reg/f313ecbd2c9258ab109355c48bb7504df4cb36299683f9e22be23ae73dc17431.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "f342155bf5ca141f53ee0d7e9468a2fc28676ba698ccb1f4a4028dd89a20c9b5", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "d09032cf64404bb7d1276865be9a25fba4420b4c28a9b2338c860ff03f759939", "generation_seconds_batch": 24.122404152993113, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4186325073242188, "target_z": -0.08789055561547397, "ce": 5.7406487464904785, "genuineness": 4.042920112609863, "identity": 0.5409652590751648, "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_CONV_reg/f342155bf5ca141f53ee0d7e9468a2fc28676ba698ccb1f4a4028dd89a20c9b5.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "f3b2d1546117effc2f23d162d15f757e7396e3eac06aa9d56abbe9aef5a8c676", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "4d4ffb07623bb27533a31708807382baafbecf0c6919c6be06e305d7ee3dbf95", "generation_seconds_batch": 21.84299379598815, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5480666160583496, "target_z": 0.7773438588034984, "ce": 6.372091293334961, "genuineness": 4.607763767242432, "identity": 0.6878460049629211, "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_CONV_reg/f3b2d1546117effc2f23d162d15f757e7396e3eac06aa9d56abbe9aef5a8c676.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-0", "key": "f3c704cfae3dfa9e4a9e94877179a5f8b49e53c7d6e13544ecd151cee4554452", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La puerta sigue abierta.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] He dejado tu abrigo sobre la silla.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "d8b444d62b7621c9f05fc97a7da8d94ee3d1c37dccb539f9406a521413e67e5b", "generation_seconds_batch": 24.1996202589944, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.8673465251922607, "target_z": 0.25585929883755115, "ce": 5.903473377227783, "genuineness": 4.168440818786621, "identity": 0.5525786876678467, "hypothesis": "La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo.", "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_CONV_reg/f3c704cfae3dfa9e4a9e94877179a5f8b49e53c7d6e13544ecd151cee4554452.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-1", "key": "f4169f5fc150f4129dfb2be95f40cbdfb8e549fe42e76418168483d0b97c94f2", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ik heb het kleine doosje gevonden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Het stond naast het oude raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "e8ab6fa0ccd7977f171578c4e6938c1c5c4b66a61cdbefc7536a03ee7676fb42", "generation_seconds_batch": 23.595628913957626, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.205840587615967, "target_z": 1.2812499058225457, "ce": 6.252951145172119, "genuineness": 4.507347106933594, "identity": 0.4099650979042053, "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": 0.03999999999999915, "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_CONV_reg/f4169f5fc150f4129dfb2be95f40cbdfb8e549fe42e76418168483d0b97c94f2.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "f4641b0a5ad9271b48b23d1615da882c9c4688a3bb893f581cc27f71606b923d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "af3990ba633b99a1263619d5d0b91e8db8a5b9f8130bd91a0a3e7eaa1d2f3685", "generation_seconds_batch": 23.45556684001349, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.5990567207336426, "target_z": 0.816406246789405, "ce": 6.133810997009277, "genuineness": 4.570107460021973, "identity": 0.6181440949440002, "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_CONV_reg/f4641b0a5ad9271b48b23d1615da882c9c4688a3bb893f581cc27f71606b923d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "f4f61555d0e6b2d82e703ab45c559346c27847d01805ddd33bb624d11efdc23a", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5bc6d3c09d1345d8e70936d6d409d4b27049ca36f01bc97567538e9c5cff2d16", "generation_seconds_batch": 26.524054056964815, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.711235284805298, "target_z": 0.9023437560644573, "ce": 6.133810997009277, "genuineness": 4.7081804275512695, "identity": 0.6570810079574585, "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_CONV_reg/f4f61555d0e6b2d82e703ab45c559346c27847d01805ddd33bb624d11efdc23a.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "f66082743e0077a7e241592b2634922a62766875e5a749fab336241db7a7020d", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "5bf0b34950ad5d15bc982dd65511a6e0afc48d5bb73f1bdc427ae98ef89bdb8c", "generation_seconds_batch": 18.65115834691096, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.512373447418213, "target_z": 0.7500001141544901, "ce": 6.1020402908325195, "genuineness": 4.381826400756836, "identity": 0.5930354595184326, "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_CONV_reg/f66082743e0077a7e241592b2634922a62766875e5a749fab336241db7a7020d.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "f6bd2a0ce7477dc7b8b97ce489d00f285cd4d662c7a61e169775f3322cac18f0", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ca7e347e47e97be3dcdfaa0416f7d1553c19e4472022d21b092356e38227212e", "generation_seconds_batch": 23.483065218897536, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.318610429763794, "target_z": 0.6015625649253663, "ce": 5.728734970092773, "genuineness": 4.532451152801514, "identity": 0.662344217300415, "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_CONV_reg/f6bd2a0ce7477dc7b8b97ce489d00f285cd4d662c7a61e169775f3322cac18f0.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": "f6e5bafda3bc2258d8bc0c0ae874b81bf423ff2950575a2c798e4787659568bb", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Jemand hat den kleinen blauen Stuhl umgestellt.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Er stand hier... neben dem hohen Bücherregal.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "213f5ac8faeb70fb4880960bdeaecc232f587c1c26fcf3825588d905824ceea4", "generation_seconds_batch": 15.426218340988271, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.022275924682617, "target_z": 1.140625089896661, "ce": 5.863759994506836, "genuineness": 4.582659721374512, "identity": 0.6004878878593445, "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_CONV_reg/f6e5bafda3bc2258d8bc0c0ae874b81bf423ff2950575a2c798e4787659568bb.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "f728cd0dee69f34dffb1138cc8e1a24e7e534d9772c1f2529214aba5e34f64cf", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0f81ee071539edeb77130d4bd18754307e42217d9cabc4a28ba5384d48d1dbcd", "generation_seconds_batch": 22.441387692000717, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4976673126220703, "target_z": -0.027343744649008277, "ce": 6.340320587158203, "genuineness": 4.306513786315918, "identity": 0.610327422618866, "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_CONV_reg/f728cd0dee69f34dffb1138cc8e1a24e7e534d9772c1f2529214aba5e34f64cf.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "f73ee99be531445fcfcff3ab84b4c1ebb7771015abe02562e2d9d49dc4c60c50", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "392921b05d1c5f8019edd8fb5f5953bd9fdde009c57464f66ebef4b1a5346be7", "generation_seconds_batch": 22.158697345061228, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.9055893421173096, "target_z": 0.28515627247416525, "ce": 6.356205940246582, "genuineness": 4.507347106933594, "identity": 0.5827087163925171, "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_CONV_reg/f73ee99be531445fcfcff3ab84b4c1ebb7771015abe02562e2d9d49dc4c60c50.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "f76fcbbef0f84462aa898892fc83fbe7c76e8f530b77cafbea48a378a737d57b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "ccef3d8173ddf423511d6131e711afc19f4d660a512761beee2275ea82cdae0a", "generation_seconds_batch": 25.494953381014057, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.578660488128662, "target_z": 0.800781145477295, "ce": 6.117925643920898, "genuineness": 4.432034492492676, "identity": 0.7175991535186768, "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_CONV_reg/f76fcbbef0f84462aa898892fc83fbe7c76e8f530b77cafbea48a378a737d57b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "f792fd3e20bc4c7d0265640034fcd735f8d99988b53c1cbb5cdf0069fd0e7a66", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "bc66992db8ca86485d22c4a0a255796c72e371ae367e20acd58386f06c508f70", "generation_seconds_batch": 23.800228476990014, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6041557788848877, "target_z": 0.8203125221174324, "ce": 6.332377910614014, "genuineness": 4.482243061065674, "identity": 0.5343663692474365, "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_CONV_reg/f792fd3e20bc4c7d0265640034fcd735f8d99988b53c1cbb5cdf0069fd0e7a66.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "f0e6434314b0543cc78017afaa0fc4b11a727b3c8f8d9531ce278c2374a0d1f0", "generation_seconds_batch": 17.261695928988047, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.216629981994629, "target_z": 0.5234376063063689, "ce": 6.0305562019348145, "genuineness": 4.400654315948486, "identity": 0.5964339971542358, "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_CONV_reg/f7a65885d0e107603bdce9d92b5fc3b68a68524db7b0b06f79b6538794e8bf5c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "f819301fcf54018156cc055bb8364d8c5bcf36c59004e8a2af9ce1eae606ea42", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "80b913eb145a634fbb4e4fa98df347e8f97418d6f137002640348da53c246613", "generation_seconds_batch": 18.059405258041807, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4046101570129395, "target_z": -0.09863276710575354, "ce": 6.1020402908325195, "genuineness": 4.268857479095459, "identity": 0.7620986104011536, "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_CONV_reg/f819301fcf54018156cc055bb8364d8c5bcf36c59004e8a2af9ce1eae606ea42.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "f821e68f13bd2472fccb7e423c2acfa2869740beaebe42247bd8b57141b71186", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "7839d4f358f976cd7609b0ea8c542318c3ef5b2a52fb327c952183012c6a3fc2", "generation_seconds_batch": 18.87660593900364, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.635340929031372, "target_z": 0.07812495861899733, "ce": 6.221180438995361, "genuineness": 3.8797430992126465, "identity": 0.6157944202423096, "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_CONV_reg/f821e68f13bd2472fccb7e423c2acfa2869740beaebe42247bd8b57141b71186.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_CONV_reg", "phase": "heldout", "candidate": "standard", "prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": "46a12d7ab8b6093270562abc02f28e7d1698684469eb30865b95defe410f7671", "generation_seconds_batch": 17.614973227959126, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.7067272663116455, "target_z": 0.13281244791701388, "ce": 6.292664527893066, "genuineness": 4.400654315948486, "identity": 0.45950978994369507, "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_CONV_reg/f86feddc7af334e5c3d6691e715ece3168f8e7618bd3b63ca7d0171f412fdf9d.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": "f8927a4f04554645a5dde12fae3dc9d0e9432a4495d778e4b72e58042c8be7d4", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De sleutels liggen op de keukentafel.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik keek in de la... en daarna naast het raam.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "2b674ece31755ce73036b317db7d163b6452d880b07c6a7c94611bae5f7b8631", "generation_seconds_batch": 21.594552156049758, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.7979187965393066, "target_z": 0.9687500713465563, "ce": 6.046441555023193, "genuineness": 4.595211505889893, "identity": 0.6998826861381531, "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_CONV_reg/f8927a4f04554645a5dde12fae3dc9d0e9432a4495d778e4b72e58042c8be7d4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "f9288deb631025e8069214b82d174d2067a183e7c49b3946ed8e134f855f60e6", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "3834bcc0a8d7f3e16da3297b0752536f47b3e5a08e6ebffe8ed2a20e53cf474f", "generation_seconds_batch": 17.74644337396603, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6347498893737793, "target_z": 0.8437499914384132, "ce": 5.95112943649292, "genuineness": 4.632867813110352, "identity": 0.612216055393219, "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": 0.03999999999999915, "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_CONV_reg/f9288deb631025e8069214b82d174d2067a183e7c49b3946ed8e134f855f60e6.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "fa2f2053150c0d2aab7f4c96de2c1167c25e414e52b194a76f618c47ca78fc1b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0d5f282da9168574fabd566f68dfddc604d3f11cfa2adbcaa0f8b31d63eb7303", "generation_seconds_batch": 20.626477609970607, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.9814834594726562, "target_z": 1.109374887272441, "ce": 6.078212261199951, "genuineness": 4.419482707977295, "identity": 0.6496373414993286, "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_CONV_reg/fa2f2053150c0d2aab7f4c96de2c1167c25e414e52b194a76f618c47ca78fc1b.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-es-1", "key": "fa3b6c61cb43f41c6d2b61b583d1f87bf033dd7cd6ec7ac1f1dd4acd5e07552e", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] He encontrado la cajita.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba junto a la ventana vieja.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3de6338538cb06a5836cd18f3f202ae3f7f3db8b36ca7c63c783e1108d7a3011", "generation_seconds_batch": 24.86580517294351, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.025416374206543, "target_z": 0.3769531034176667, "ce": 5.71682071685791, "genuineness": 4.406930446624756, "identity": 0.6493912935256958, "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_CONV_reg/fa3b6c61cb43f41c6d2b61b583d1f87bf033dd7cd6ec7ac1f1dd4acd5e07552e.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "fa87b110fc5e1da51be70105194fc6dfd26fcaac67929ee469139dfdd4ec1205", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "fa69ad1dda0f4d2eeb7b16b138ca044bcfd649e3f91b2d8dad4933152b6b662c", "generation_seconds_batch": 23.30763170600403, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.8897011280059814, "target_z": 1.0390624793094987, "ce": 6.268836498260498, "genuineness": 4.532451152801514, "identity": 0.3825696110725403, "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_CONV_reg/fa87b110fc5e1da51be70105194fc6dfd26fcaac67929ee469139dfdd4ec1205.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": "fb12be7a375e2623e5cff6a451253517b2d1fe7d2ee4c644ca322cea44a70ce4", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Your message arrived after I had left.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I read it twice... before putting the phone away.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "b5d9640e58d5b89c93b0243fefa2b460b12278656030d51e77b44d8cedda1289", "generation_seconds_batch": 15.352203065995127, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.0611095428466797, "target_z": 0.404296848066675, "ce": 6.435632705688477, "genuineness": 4.519899368286133, "identity": 0.49858173727989197, "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_CONV_reg/fb12be7a375e2623e5cff6a451253517b2d1fe7d2ee4c644ca322cea44a70ce4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "fb1a5efcd573ac9f6b4aef6f258b426e4b265a554a30ec00214662eb40a12c23", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "1029a41a496f720e7795f66b78119e1f896aa04a94bad6857e56b64354b4cee2", "generation_seconds_batch": 25.258093671989627, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.6041557788848877, "target_z": 0.8203125221174324, "ce": 6.38003396987915, "genuineness": 4.582659721374512, "identity": 0.5848212838172913, "hypothesis": "Die Tür ist davon echt noch offen, Ades. Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist.", "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_CONV_reg/fb1a5efcd573ac9f6b4aef6f258b426e4b265a554a30ec00214662eb40a12c23.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-1", "key": "fb5e9904420c13042ca20eb756193f7885735b6bbae48d3e866bef262e6bed4b", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] I found the little box.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It was beside the old window.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8c0225a3823f07d11b31e580ce9eee30da5fb5c822ed09f5a0edee41fd56f4b3", "generation_seconds_batch": 22.7725300360471, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.920295238494873, "target_z": 1.0624999486304794, "ce": 6.340320587158203, "genuineness": 4.432034492492676, "identity": 0.2720106542110443, "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_CONV_reg/fb5e9904420c13042ca20eb756193f7885735b6bbae48d3e866bef262e6bed4b.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": "fba90fec4eec1fa0b33774640861676f5df5785bfc65c5d34377c3b5adf2591c", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Someone moved the little blue chair.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] It used to stand here... beside the tall bookshelf.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "8eb02caa5b87629a802c68e5884bc93138c811d654be1da4ee9285f4461773c3", "generation_seconds_batch": 18.21244785003364, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.226236820220947, "target_z": 1.2968750071346555, "ce": 6.260893821716309, "genuineness": 4.406930446624756, "identity": 0.321647047996521, "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_CONV_reg/fba90fec4eec1fa0b33774640861676f5df5785bfc65c5d34377c3b5adf2591c.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-1", "key": "fbb7a6e395339e51fedae4f33f5af54bd1b566815289d412e490cdfd31364143", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Sie stand neben dem alten Fenster.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "3c0d46b908bea3a43acf2b2985302a897345a871a7028281cf8584b40fd2b6d9", "generation_seconds_batch": 21.0754584800452, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.4951179027557373, "target_z": -0.029296790989429933, "ce": 6.348263263702393, "genuineness": 4.293961524963379, "identity": 0.36972352862358093, "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_CONV_reg/fbb7a6e395339e51fedae4f33f5af54bd1b566815289d412e490cdfd31364143.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": "fbbb2ac01b11eb67d9b623ea9ddd514bbbc67306216e6b3f8239c82f150010f7", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Tu mensaje llegó después de que me fuera.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Lo leí dos veces... y guardé el teléfono.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 3.0, "ended": true, "frames_generated": 128, "seconds_generated": 10.24, "mp3_sha256": "14bd2b7c56c29a98e1a8987bd8f74def97303b8f83f6c0455677811a91fb9cb1", "generation_seconds_batch": 15.11226050904952, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.522571325302124, "target_z": 0.757812482163361, "ce": 5.8875885009765625, "genuineness": 4.758388519287109, "identity": 0.5426087379455566, "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_CONV_reg/fbbb2ac01b11eb67d9b623ea9ddd514bbbc67306216e6b3f8239c82f150010f7.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-0", "key": "fbf48b0378c8f67288a11f673710379bd0e69776507fe0cf15bf2c34c1569e7f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] La porte est encore ouverte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 127, "seconds_generated": 10.16, "mp3_sha256": "91fc90a0476f3537c5d8d6a7a3bf5ad06dd47114ff1af4d655303b10032fc1bf", "generation_seconds_batch": 17.74644337396603, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.348613262176514, "target_z": 1.3906248844185787, "ce": 6.125868320465088, "genuineness": 4.68307638168335, "identity": 0.7046163082122803, "hypothesis": "La porte est encore ouverte. J'ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt, ni muissomme à cet étonnement", "wer": 0.2777777777777778, "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_CONV_reg/fbf48b0378c8f67288a11f673710379bd0e69776507fe0cf15bf2c34c1569e7f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "fc1f09a4ef0bccb3efd04c9d63ddc4940d7a3488aef6b413c354526cc3913948", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2978c6f2b90eee353e2dadb686a07c77fc1466c42ac44308bdeabeca1127ad19", "generation_seconds_batch": 18.91927820595447, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.022275924682617, "target_z": 1.140625089896661, "ce": 5.990842819213867, "genuineness": 4.557555198669434, "identity": 0.26846304535865784, "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_CONV_reg/fc1f09a4ef0bccb3efd04c9d63ddc4940d7a3488aef6b413c354526cc3913948.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "fcd836f04d915998bc7fd7e7dd530f764a94bb06dc241fb7498cc2b7f809bb77", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "0472bb4d3745fc485030b49e5b49b2f2dca09384355b7aa84dd55fb11fdebfdb", "generation_seconds_batch": 26.68118863797281, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.06306791305542, "target_z": 1.1718749272265125, "ce": 6.24500846862793, "genuineness": 4.519899368286133, "identity": 0.6526095867156982, "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_CONV_reg/fcd836f04d915998bc7fd7e7dd530f764a94bb06dc241fb7498cc2b7f809bb77.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-fr-1", "key": "fd261b532666b2200c53dae289522f43b5eaf6140eb0fa87de2983e2a0680ed4", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] J’ai trouvé la petite boîte.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Elle était près de la vieille fenêtre.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "f37f37a26ff5e5736125bac7b7f8dc93cf7f17d4be81741d4b6139d68bea9a8f", "generation_seconds_batch": 24.375515970983543, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.767324447631836, "target_z": 0.9453124193783914, "ce": 6.0146708488464355, "genuineness": 4.582659721374512, "identity": 0.7013542652130127, "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_CONV_reg/fd261b532666b2200c53dae289522f43b5eaf6140eb0fa87de2983e2a0680ed4.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "fd80e4bbcd29ab4b2962a0170c37d5092dcade13f1baf76b9342d8a2fdb84eaa", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "fe774ef63171c04fdbf9ae56025a4a92e17e273cd0e380f9015a65d4cb06f72f", "generation_seconds_batch": 18.695571545977145, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 4.2874250411987305, "target_z": 1.343749945776617, "ce": 6.284721851348877, "genuineness": 4.645420074462891, "identity": 0.5653969645500183, "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_CONV_reg/fd80e4bbcd29ab4b2962a0170c37d5092dcade13f1baf76b9342d8a2fdb84eaa.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": "fde6e3819384782309055d3cb0931f5909729527a9cd781b9d4f015513103063", "bucket": "vn_S_CONV_reg", "phase": "heldout", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Alguien ha movido la pequeña silla azul.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Estaba aquí... junto a la estantería alta.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 0.5, "cfg": 3.0, "ended": true, "frames_generated": 127, "seconds_generated": 10.16, "mp3_sha256": "a7cd4c4cf5bbbbbb24ef9b58dc73fe866eef193a3c309ee144cc0b275dfe5bc3", "generation_seconds_batch": 21.230708232033066, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.961087465286255, "target_z": 1.0937499686075152, "ce": 6.165581703186035, "genuineness": 4.720732688903809, "identity": 0.6237121820449829, "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": 0.03999999999999915, "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_CONV_reg/fde6e3819384782309055d3cb0931f5909729527a9cd781b9d4f015513103063.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "fdf2b4471d98b0235b5c5a098d10e7d36ced7cb287a39b843c4515b09f0a0a87", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "2bc1aa9f888f3ab039f2b686333de84d8e048b929c294d213a8e61cf01206010", "generation_seconds_batch": 18.980337023036554, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.173288106918335, "target_z": 0.4902343573417273, "ce": 6.125868320465088, "genuineness": 4.218649387359619, "identity": 0.4347759187221527, "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_CONV_reg/fdf2b4471d98b0235b5c5a098d10e7d36ced7cb287a39b843c4515b09f0a0a87.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-en-0", "key": "fe95d78203d1846378c4868b697ac5a11a971656c65f5a03035808b37a32e37f", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] The door is still open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] I put your coat on the chair.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": 128, "seconds_generated": 10.24, "mp3_sha256": "c0284c63fc5acd14cecbccf101c59934e55f6c598d26e9c711c4ff3e2a497f1a", "generation_seconds_batch": 22.72915472392924, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 2.920886516571045, "target_z": 0.2968750984582477, "ce": 6.221180438995361, "genuineness": 4.030367851257324, "identity": 0.4994726777076721, "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_CONV_reg/fe95d78203d1846378c4868b697ac5a11a971656c65f5a03035808b37a32e37f.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-nl-0", "key": "ff2f850f0c506ac59b841c14537a7377d6acd0efe7681b11dff57c5622c37a81", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] De deur staat nog open.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "cbced1ad91a56652fe11440bb6f61d0bf95518e486d7f13086604a6f2c3a2cb5", "generation_seconds_batch": 16.828201264957897, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.716334342956543, "target_z": 0.9062500313924848, "ce": 6.022613525390625, "genuineness": 4.400654315948486, "identity": 0.685088038444519, "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_CONV_reg/ff2f850f0c506ac59b841c14537a7377d6acd0efe7681b11dff57c5622c37a81.mp3" }, { "id": "vn_S_CONV_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_CONV_reg-de-0", "key": "ff43d152190c030d18a6903bd396ab6a489684ab185ea26558be5af1b8436d23", "bucket": "vn_S_CONV_reg", "phase": "grid", "candidate": "manual-ea97be76622a", "prompt": "GENERAL: One adult speaker, clean high-quality recording, no music or sound effects. Craft the delivery to sound like a participant in a deep, meandering thought process shared with a trusted confidant.\nSCRIPT:\n(Employ slight vocal breathiness during transitions to mimic natural thought retrieval.) [3.2 seconds duration] Die Tür ist noch offen.\n(Vary the projection level slightly, moving between confidential softness and moderate conversational volume.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt.\n(Ensure the ending trails off with a sense of unresolved, natural continuation, inviting further response.) [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": "bedee1c2d5ba1177aba0dc85a1925dcd5f251cfcc84021a3af2d85c18f8174e1", "generation_seconds_batch": 24.78121139202267, "original_audio_reused": false, "scoring_contract": "singleton-EARS-ASR-Orange128-preMP3-v1", "metrics": { "valid_waveform": true, "target_raw": 3.364501476287842, "target_z": 0.6367186775832453, "ce": 6.1894097328186035, "genuineness": 4.557555198669434, "identity": 0.4906955659389496, "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_CONV_reg/ff43d152190c030d18a6903bd396ab6a489684ab185ea26558be5af1b8436d23.mp3" }, { "key": "original-vn_S_CONV_reg-de-0-seed1701-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "44d8ce0001ee9e49fe83f4b5494e8a0e6f4ca002e74d0c20c02f2b05b97a2dde", "task_id": "vn_S_CONV_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.578660488128662, "target_z": 0.800781145477295, "ce": 6.356205940246582, "genuineness": 4.607763767242432, "identity": 0.6591275930404663, "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_CONV_reg/de/24f34800d8a470d10b58fab1.mp3" }, { "key": "original-vn_S_CONV_reg-de-0-seed1701-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "5d078e2fe869623d029a61d2bb096c8cd862d29edb211e6748905539ecf1293a", "task_id": "vn_S_CONV_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.882643699645996, "target_z": 0.2675781248216336, "ce": 6.1894097328186035, "genuineness": 4.369274139404297, "identity": 0.5348483920097351, "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_CONV_reg/de/60a53b06ac6c9472b7a758ca.mp3" }, { "key": "original-vn_S_CONV_reg-de-0-seed1701-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "083626828f14dd07dfce274e86f1dced4fedd868e08a60d840ae783f3d12aaf5", "task_id": "vn_S_CONV_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.8647971153259277, "target_z": 0.25390625249712945, "ce": 6.23706579208374, "genuineness": 4.193545341491699, "identity": 0.7152451276779175, "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_CONV_reg/de/eb3ffea50dc1a39fea90dd0c.mp3" }, { "key": "original-vn_S_CONV_reg-de-0-seed1701-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "84b0cb5e048286d8c1608e4381c66607f94824474f3bb1cd00618bd44bf38b7f", "task_id": "vn_S_CONV_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.838027238845825, "target_z": 0.23339844401037327, "ce": 6.125868320465088, "genuineness": 4.331617832183838, "identity": 0.5468755960464478, "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_CONV_reg/de/2e5683e5d4ec2f76fa5dc63a.mp3" }, { "key": "original-vn_S_CONV_reg-de-0-seed1702-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "7a29fd0c5e888be0e1747bdaea47a98a303461bf22ea44cc4b05d061120e33e9", "task_id": "vn_S_CONV_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.7500691413879395, "target_z": 0.1660156968816555, "ce": 6.109982967376709, "genuineness": 4.519899368286133, "identity": 0.6025513410568237, "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_CONV_reg/de/b445aa12eca522e2d80f6aaf.mp3" }, { "key": "original-vn_S_CONV_reg-de-0-seed1702-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "1b65b5250b0d5450ee8cab73eb2b6de01e7fa0aa58f9323a7e23ad3fb4d5c85a", "task_id": "vn_S_CONV_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.4562840461730957, "target_z": 0.7070312681933718, "ce": 6.348263263702393, "genuineness": 4.632867813110352, "identity": 0.6059601306915283, "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_CONV_reg/de/1fa8968f1a2a2cd451578289.mp3" }, { "key": "original-vn_S_CONV_reg-de-0-seed1702-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "cfbd8fb23a4c8f9ea432fe57eef06e52991e93b84471f8e958005fb2a6e160d3", "task_id": "vn_S_CONV_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.6283297538757324, "target_z": 0.07275385287385755, "ce": 6.205295085906982, "genuineness": 4.268857479095459, "identity": 0.5636640787124634, "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_CONV_reg/de/191a16d5ee422d8963c63e12.mp3" }, { "key": "original-vn_S_CONV_reg-de-0-seed1702-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "eb5248df1947b88c4f6b6ef3ed72743a9c079ac70dfa52ccf9ba664cb5497e3a", "task_id": "vn_S_CONV_reg-de-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.1146492958068848, "target_z": 0.4453124650401874, "ce": 6.109982967376709, "genuineness": 4.419482707977295, "identity": 0.4314640164375305, "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_CONV_reg/de/655962111065b48e48d1421f.mp3" }, { "key": "original-vn_S_CONV_reg-de-1-seed1701-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "b58ece670a1077104d3a8f92d5d644b89a636530a384492736252927d9e89211", "task_id": "vn_S_CONV_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.17642879486084, "target_z": -0.2734374464900828, "ce": 6.443574905395508, "genuineness": 4.507347106933594, "identity": 0.6289011240005493, "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_CONV_reg/de/d6695d7475156bbeb0a18991.mp3" }, { "key": "original-vn_S_CONV_reg-de-1-seed1701-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "09f318c89bde2af53943286f0bd13e0da68c61ad31eb1abb8c7f5b29b7683629", "task_id": "vn_S_CONV_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.186626672744751, "target_z": -0.265625078481212, "ce": 6.38797664642334, "genuineness": 4.344170093536377, "identity": 0.3286755084991455, "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_CONV_reg/de/8a3b54f7dd8504bb37db9f74.mp3" }, { "key": "original-vn_S_CONV_reg-de-1-seed1701-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "d5a6a53a8c9e1ad1cb39d182012627182f177c45337bd4544834467470250b9e", "task_id": "vn_S_CONV_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.337048053741455, "target_z": -0.1503905955695455, "ce": 6.41180419921875, "genuineness": 4.042920112609863, "identity": 0.4733392000198364, "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_CONV_reg/de/6a65394ad03c1c7955640389.mp3" }, { "key": "original-vn_S_CONV_reg-de-1-seed1701-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "bb02ae58aab71376107839a626946046c45bf15696b48006be264d02e035453f", "task_id": "vn_S_CONV_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.0540521144866943, "target_z": -0.3671875064211901, "ce": 6.23706579208374, "genuineness": 4.243753433227539, "identity": 0.6643704175949097, "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_CONV_reg/de/83e946a4262d40c0de9db688.mp3" }, { "key": "original-vn_S_CONV_reg-de-1-seed1702-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "244b5a75793169e0eba9fe302b5fdf48a2639089e48b6bf116873e149156e6d8", "task_id": "vn_S_CONV_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.4556002616882324, "target_z": -0.059570379119846965, "ce": 6.403861999511719, "genuineness": 4.381826400756836, "identity": 0.5912182331085205, "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_CONV_reg/de/8b9c40784794c9ae12d00388.mp3" }, { "key": "original-vn_S_CONV_reg-de-1-seed1702-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "6f6798ecdea799828e1b8c643a5be38015722e690e95025ab0722c9c3ebadda8", "task_id": "vn_S_CONV_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.7628166675567627, "target_z": 0.17578129387813213, "ce": 6.3641486167907715, "genuineness": 4.519899368286133, "identity": 0.44632768630981445, "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_CONV_reg/de/a0b04c55afedaf9c621f250d.mp3" }, { "key": "original-vn_S_CONV_reg-de-1-seed1702-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "8d4894091610e91974c8d6b0d315788376c6fbd4d2204917c846113b01ded8f8", "task_id": "vn_S_CONV_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.4938430786132812, "target_z": -0.030273405483232853, "ce": 6.395919322967529, "genuineness": 4.344170093536377, "identity": 0.5881282687187195, "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_CONV_reg/de/ea127e1bd4d5867bf6d9fb3f.mp3" }, { "key": "original-vn_S_CONV_reg-de-1-seed1702-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "3ac08641acb00d3ed0863a48e3de9cb4603374c52391e0a07db8e260eb52f3c7", "task_id": "vn_S_CONV_reg-de-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.639165163040161, "target_z": 0.08105461945322191, "ce": 6.38003396987915, "genuineness": 4.482243061065674, "identity": 0.675316333770752, "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_CONV_reg/de/1674cdb093360654c75d4772.mp3" }, { "key": "original-vn_S_CONV_reg-en-0-seed1701-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "ab4cffc50e3d80946ea00e7a3cc5cc24d69b4c43e15fa36e7c9ce87b83bc51bf", "task_id": "vn_S_CONV_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 1.8704872131347656, "target_z": -0.5078125049942589, "ce": 6.213237762451172, "genuineness": 3.7416701316833496, "identity": 0.41462844610214233, "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_CONV_reg/en/340f788d74c78a7a95dc0450.mp3" }, { "key": "original-vn_S_CONV_reg-en-0-seed1701-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "28b27ae01a9f550eddbf93f76bbb219a350f4a50711cbb2bd33da03025df1564", "task_id": "vn_S_CONV_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.3090033531188965, "target_z": -0.17187501855010465, "ce": 6.181467056274414, "genuineness": 3.898571252822876, "identity": 0.38136231899261475, "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_CONV_reg/en/73464b622f0aed2c9c2ac4d9.mp3" }, { "key": "original-vn_S_CONV_reg-en-0-seed1701-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "b3eccd1e86f913f822dff71a8dbce4e6891a72f2f220bb5fce5c0f0c8795c3b1", "task_id": "vn_S_CONV_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.925985336303711, "target_z": 0.300781191139091, "ce": 6.23706579208374, "genuineness": 4.124508857727051, "identity": 0.2230311930179596, "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_CONV_reg/en/c82edf7db8e2f9cadab200ef.mp3" }, { "key": "original-vn_S_CONV_reg-en-0-seed1701-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "26f02bec22874b102edf6455f80ef0eedba50264710885cd633dd555a7085ac5", "task_id": "vn_S_CONV_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.4071595668792725, "target_z": -0.09667972076533188, "ce": 6.23706579208374, "genuineness": 4.118232727050781, "identity": 0.24127045273780823, "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_CONV_reg/en/08d14621634f20a6c832691c.mp3" }, { "key": "original-vn_S_CONV_reg-en-0-seed1702-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "d375dd1a6a103bef65b3627bcb1b4529e4abf5ffe0f72d1b21b3206099b920bb", "task_id": "vn_S_CONV_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.821455478668213, "target_z": 0.22070318617967205, "ce": 6.23706579208374, "genuineness": 4.118232727050781, "identity": 0.23348623514175415, "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_CONV_reg/en/c590443422be4fd2f1b3d860.mp3" }, { "key": "original-vn_S_CONV_reg-en-0-seed1702-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "71985ed87e430279db8dc84aa1a52fa3f9d6865562d5a4a9697cc13c7d228526", "task_id": "vn_S_CONV_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.5142390727996826, "target_z": -0.014648486818307056, "ce": 6.117925643920898, "genuineness": 3.8044304847717285, "identity": 0.24858053028583527, "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_CONV_reg/en/58894234a5cae19da07645fa.mp3" }, { "key": "original-vn_S_CONV_reg-en-0-seed1702-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "a7add69d9c802ff9e76e49d649ebd9d50c035c31bfaab4f73698416318647a6f", "task_id": "vn_S_CONV_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.508502721786499, "target_z": -0.019042978069643918, "ce": 6.205295085906982, "genuineness": 4.068024158477783, "identity": 0.3637884557247162, "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_CONV_reg/en/d2c674c38a98420768b0d24c.mp3" }, { "key": "original-vn_S_CONV_reg-en-0-seed1702-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "3dfcf6980a4e88a2098802a0cf20a6c2819e95a003fd54a49b3f05868198267a", "task_id": "vn_S_CONV_reg-en-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 1.5696446895599365, "target_z": -0.7382812881704076, "ce": 6.181467056274414, "genuineness": 3.7542223930358887, "identity": 0.26175618171691895, "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_CONV_reg/en/05d0638455e3e01fd3db44ea.mp3" }, { "key": "original-vn_S_CONV_reg-en-1-seed1701-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": 127, "seconds_generated": 10.16, "ended": true, "contract_sha256": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "94b9ac0be157f097efac64dcda51ae64ecafe08098af64123b478e11039f6d20", "task_id": "vn_S_CONV_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.563363552093506, "target_z": 0.7890625021403966, "ce": 6.348263263702393, "genuineness": 4.482243061065674, "identity": 0.489151269197464, "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_CONV_reg/en/29137b58aa427640f1e062a2.mp3" }, { "key": "original-vn_S_CONV_reg-en-1-seed1701-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "8c3321e89427935881375be282ad2ebcb327eb91c09345115c0764578bbd8550", "task_id": "vn_S_CONV_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.2472238540649414, "target_z": 0.5468748929801656, "ce": 6.308549880981445, "genuineness": 4.582659721374512, "identity": 0.22179250419139862, "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_CONV_reg/en/8df1198b14dcc876409214a9.mp3" }, { "key": "original-vn_S_CONV_reg-en-1-seed1701-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": 127, "seconds_generated": 10.16, "ended": true, "contract_sha256": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "56328ea0560700ff05505d391d79a938ce3de46464a1ad5552c690a6b4608dc0", "task_id": "vn_S_CONV_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.02286696434021, "target_z": 0.37500005707724504, "ce": 6.332377910614014, "genuineness": 4.356721878051758, "identity": 0.3552420735359192, "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_CONV_reg/en/d79bccd125b40cf2af3de934.mp3" }, { "key": "original-vn_S_CONV_reg-en-1-seed1701-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "c29a5e9f298f4d6a3b0ae68041687c0d99fe9cc1c108d2fd39860f2f813fc793", "task_id": "vn_S_CONV_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.517472267150879, "target_z": 0.7539062068353334, "ce": 6.23706579208374, "genuineness": 4.482243061065674, "identity": 0.545978307723999, "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_CONV_reg/en/3faab973bfe06473a27914de.mp3" }, { "key": "original-vn_S_CONV_reg-en-1-seed1702-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "720c56023f98ababef58db4dcbed4fac2ef4b4eab592b6a30c4387bbc46d251e", "task_id": "vn_S_CONV_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.318610429763794, "target_z": 0.6015625649253663, "ce": 6.292664527893066, "genuineness": 4.457139015197754, "identity": 0.4584965705871582, "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_CONV_reg/en/1df800f3142ba7ca5cc7cd6c.mp3" }, { "key": "original-vn_S_CONV_reg-en-1-seed1702-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "762a5e535a67485392f9b381640a0896cb1a3a8243b93844856990f4932822bd", "task_id": "vn_S_CONV_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.2574219703674316, "target_z": 0.5546874436362205, "ce": 6.284721851348877, "genuineness": 4.293961524963379, "identity": 0.2910889685153961, "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_CONV_reg/en/a5338188dcc15d20b6f4103c.mp3" }, { "key": "original-vn_S_CONV_reg-en-1-seed1702-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "c43a7539ad9940a28ea7d258db73733394c56c5f503411190c3d38c2fa180885", "task_id": "vn_S_CONV_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.2421250343322754, "target_z": 0.5429688002993222, "ce": 6.221180438995361, "genuineness": 4.3504462242126465, "identity": 0.3042275011539459, "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_CONV_reg/en/3784357d9990f27b2d756e50.mp3" }, { "key": "original-vn_S_CONV_reg-en-1-seed1702-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "ee1492123cd7961d111cf79ccc446354a9163937a3a1e576a16d8991556626e0", "task_id": "vn_S_CONV_reg-en-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.4001946449279785, "target_z": 0.6640624222322536, "ce": 6.197352409362793, "genuineness": 4.319066047668457, "identity": 0.4941454231739044, "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_CONV_reg/en/b4d271f82460f1deb70fee0d.mp3" }, { "key": "original-vn_S_CONV_reg-es-0-seed1701-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "ac76cc54bf862dc207b6ea0249147bfc2c2224cfd583ec49d81c438733edc35c", "task_id": "vn_S_CONV_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.2019238471984863, "target_z": -0.25390625249712945, "ce": 6.062326908111572, "genuineness": 3.9550557136535645, "identity": 0.5462881326675415, "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_CONV_reg/es/5a79b8a1c3cf7c19cab3aa25.mp3" }, { "key": "original-vn_S_CONV_reg-es-0-seed1701-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "eec27e941eb2af34450d5ee15f302abb095ccaccfa37dae4271f535bda50e136", "task_id": "vn_S_CONV_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.17642879486084, "target_z": -0.2734374464900828, "ce": 6.109982967376709, "genuineness": 4.243753433227539, "identity": 0.46990764141082764, "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_CONV_reg/es/e12a7ff5816299bf08443b3d.mp3" }, { "key": "original-vn_S_CONV_reg-es-0-seed1701-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "b95e2b4fe31cf98c88724f618565e4ca0d7df9f762c986cd5b704d8185b53b9a", "task_id": "vn_S_CONV_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 1.6767241954803467, "target_z": -0.6562500542233828, "ce": 6.038498878479004, "genuineness": 4.011539936065674, "identity": 0.6055389642715454, "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_CONV_reg/es/ab0c918c562d304e7321b65c.mp3" }, { "key": "original-vn_S_CONV_reg-es-0-seed1701-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "362ff41247b4ac1dc934216600cb83de0dfeeb993aca5b6a53e191368292a8fa", "task_id": "vn_S_CONV_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 1.6359320878982544, "target_z": -0.6874999828768265, "ce": 6.0305562019348145, "genuineness": 3.911123275756836, "identity": 0.5339090824127197, "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_CONV_reg/es/c2958fa60dd5f5722c43063d.mp3" }, { "key": "original-vn_S_CONV_reg-es-0-seed1702-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "0fd60ccf38cdb35b2df4a0d2bb9d448dc2720844160db3a1e6846b8b3f97a20c", "task_id": "vn_S_CONV_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.474721670150757, "target_z": -0.04492189230153991, "ce": 6.062326908111572, "genuineness": 4.218649387359619, "identity": 0.485903799533844, "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_CONV_reg/es/4c83cff30128d798dffa6e34.mp3" }, { "key": "original-vn_S_CONV_reg-es-0-seed1702-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "63494363ea1434418b65e032cd81057c55918ef098c2014d79b550d372420a68", "task_id": "vn_S_CONV_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 1.9214775562286377, "target_z": -0.4687499343611682, "ce": 5.998785495758057, "genuineness": 3.8797430992126465, "identity": 0.6375206112861633, "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_CONV_reg/es/cd8bd7059edfb08b67613de6.mp3" }, { "key": "original-vn_S_CONV_reg-es-0-seed1702-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "783c8d97b9e2c6272d3d2fa60c4e72f2c5b63c4eb4cfbc2a857356321cc49aba", "task_id": "vn_S_CONV_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.226144313812256, "target_z": -0.23535149035079492, "ce": 6.1894097328186035, "genuineness": 4.20609712600708, "identity": 0.5527219772338867, "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_CONV_reg/es/a6bdf5a7f46980d1fb6aac85.mp3" }, { "key": "original-vn_S_CONV_reg-es-0-seed1702-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "5e78a71f24f2076595a08ee08ac18d7cadc6bf699ee7f320bf092cf8b7b7144d", "task_id": "vn_S_CONV_reg-es-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.233792781829834, "target_z": -0.22949216868234576, "ce": 5.863759994506836, "genuineness": 3.9801597595214844, "identity": 0.554594099521637, "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_CONV_reg/es/2d29fec6eacf26dab46ae74f.mp3" }, { "key": "original-vn_S_CONV_reg-es-1-seed1701-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "d920de35b1c62e4e1449ce64befa1ce447f8e292209c37a72bb82b549365d123", "task_id": "vn_S_CONV_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.0407135486602783, "target_z": 0.3886719294017492, "ce": 6.062326908111572, "genuineness": 4.545003414154053, "identity": 0.6481660008430481, "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_CONV_reg/es/cb5fcc588293ec1442acf2d7.mp3" }, { "key": "original-vn_S_CONV_reg-es-1-seed1701-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "1812c7c196ca47fd03bd8d36ddd0a541870c83d64c416ec85e7b6ee37e45f690", "task_id": "vn_S_CONV_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.4919772148132324, "target_z": 0.7343750128423802, "ce": 6.006728172302246, "genuineness": 4.796044826507568, "identity": 0.6002509593963623, "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_CONV_reg/es/4d86c1cf210c29a1efd891ca.mp3" }, { "key": "original-vn_S_CONV_reg-es-1-seed1701-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "6f3b67d7ed5d87fa9b51e8c9bd3631c2fb3cbe6c84f9a0740eee7793a97d1a2a", "task_id": "vn_S_CONV_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.216629981994629, "target_z": 0.5234376063063689, "ce": 6.038498878479004, "genuineness": 4.65797233581543, "identity": 0.6639608144760132, "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_CONV_reg/es/447ef2e501ea3f9668570114.mp3" }, { "key": "original-vn_S_CONV_reg-es-1-seed1701-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "e96305de9d229653be9bd86acbc66f4714cc66fd2e40cb57496d819a579c72fd", "task_id": "vn_S_CONV_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.28350830078125, "target_z": -0.19140621254305792, "ce": 6.062326908111572, "genuineness": 4.243753433227539, "identity": 0.6264190673828125, "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_CONV_reg/es/36a19806c98f75113bede23f.mp3" }, { "key": "original-vn_S_CONV_reg-es-1-seed1702-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "b71f211b0ff309257b5548467660978bc4148e7ea6767fe8e00a2d918504cca7", "task_id": "vn_S_CONV_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.4708974361419678, "target_z": -0.04785155313576448, "ce": 6.109982967376709, "genuineness": 4.375550270080566, "identity": 0.799119234085083, "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_CONV_reg/es/00d2cae35844ac349c1d265e.mp3" }, { "key": "original-vn_S_CONV_reg-es-1-seed1702-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "287191dc65112a6b4e31d2ea3179ee1622beda7776be7b8f4637f72adc387441", "task_id": "vn_S_CONV_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.522571325302124, "target_z": 0.757812482163361, "ce": 5.95112943649292, "genuineness": 4.6705241203308105, "identity": 0.7611116170883179, "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_CONV_reg/es/900915773aef1bd398593739.mp3" }, { "key": "original-vn_S_CONV_reg-es-1-seed1702-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "403675ff47442e91169ee5a880e4e46d48e8962d8c846ddc24db4509cdb2c806", "task_id": "vn_S_CONV_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.620043992996216, "target_z": 0.06640631528209903, "ce": 6.229123115539551, "genuineness": 4.469690799713135, "identity": 0.6688541173934937, "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_CONV_reg/es/6c30d37ea8f4997609d6a470.mp3" }, { "key": "original-vn_S_CONV_reg-es-1-seed1702-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "23ef672a6a42b92c02ef3e548d3f6597b5d1044e9142c78b98366f6ad61c89d9", "task_id": "vn_S_CONV_reg-es-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.477271318435669, "target_z": -0.04296866331393407, "ce": 6.117925643920898, "genuineness": 4.281409740447998, "identity": 0.5982524156570435, "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_CONV_reg/es/773bca917d324654f82c92bd.mp3" }, { "key": "original-vn_S_CONV_reg-fr-0-seed1701-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "ce0131ed168f1be9145f6c1f6fe62b57ed273bb4459c366e38c7c66e0274744a", "task_id": "vn_S_CONV_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.094844341278076, "target_z": -0.3359374864441543, "ce": 6.197352409362793, "genuineness": 4.055472373962402, "identity": 0.8285737633705139, "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_CONV_reg/fr/ec321f3332121a4ffc9dbf61.mp3" }, { "key": "original-vn_S_CONV_reg-fr-0-seed1701-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "826b1000acd69be2de233679246adbe09a4e11114b71174a17cc49c8e745ddaf", "task_id": "vn_S_CONV_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.3548946380615234, "target_z": -0.1367187232450414, "ce": 6.157639026641846, "genuineness": 4.306513786315918, "identity": 0.7459686398506165, "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_CONV_reg/fr/61f205842a8e4c6bfdfa4992.mp3" }, { "key": "original-vn_S_CONV_reg-fr-0-seed1701-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "9c3bf5a8fcc172eef17b65692a0843e3bad57594331990746695d67390a2483d", "task_id": "vn_S_CONV_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.505953311920166, "target_z": -0.020996024410065577, "ce": 6.038498878479004, "genuineness": 4.20609712600708, "identity": 0.8284317255020142, "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_CONV_reg/fr/14aeed04a67b9eb3854e5290.mp3" }, { "key": "original-vn_S_CONV_reg-fr-0-seed1701-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "d895d689d89053752a57a4eb539dde7de6406b9991cba25c9037420e4953932c", "task_id": "vn_S_CONV_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.437753677368164, "target_z": -0.0732422514443511, "ce": 6.213237762451172, "genuineness": 4.168440818786621, "identity": 0.8481311202049255, "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_CONV_reg/fr/05d9c2043c02fc421f03ee53.mp3" }, { "key": "original-vn_S_CONV_reg-fr-0-seed1702-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "319f5f370485475315dd89db433edfff81990bf1176f76ecbd477a6b565609aa", "task_id": "vn_S_CONV_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.7653660774230957, "target_z": 0.1777343402185538, "ce": 6.197352409362793, "genuineness": 4.432034492492676, "identity": 0.7472943663597107, "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_CONV_reg/fr/355b04e8321395d345f996e6.mp3" }, { "key": "original-vn_S_CONV_reg-fr-0-seed1702-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "c835ad5a29a3d6c67702af50284ae61c19498a5923c6e70c0feaff0e07994f81", "task_id": "vn_S_CONV_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.9004902839660645, "target_z": 0.28124999714613774, "ce": 6.221180438995361, "genuineness": 4.532451152801514, "identity": 0.6511005163192749, "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_CONV_reg/fr/72da61985c68384c62ab7d3b.mp3" }, { "key": "original-vn_S_CONV_reg-fr-0-seed1702-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "7464694dd7492986f7764d6c59e04bf2bbacff013801eabc644c7c5872ccd403", "task_id": "vn_S_CONV_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.375290632247925, "target_z": -0.12109380458011558, "ce": 6.181467056274414, "genuineness": 4.356721878051758, "identity": 0.712196946144104, "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_CONV_reg/fr/07a10fd66590b35468abc5ff.mp3" }, { "key": "original-vn_S_CONV_reg-fr-0-seed1702-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "d5670aac23b2d4a37b5bf2d26c3a204a6bf6742c23f692be2ceb9977b4c1bc40", "task_id": "vn_S_CONV_reg-fr-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.6149449348449707, "target_z": 0.06250003995407154, "ce": 6.022613525390625, "genuineness": 4.344170093536377, "identity": 0.7840471267700195, "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_CONV_reg/fr/71e5411d82b61e78fe0493b8.mp3" }, { "key": "original-vn_S_CONV_reg-fr-1-seed1701-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "f0c69a8ddf49a5fb2b14c0b8b7f95fed4b26292da658da267cd9179a9882209a", "task_id": "vn_S_CONV_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.733497142791748, "target_z": 0.1533202564037701, "ce": 6.165581703186035, "genuineness": 4.457139015197754, "identity": 0.6082763671875, "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_CONV_reg/fr/2c69805eb6a8a007578a419c.mp3" }, { "key": "original-vn_S_CONV_reg-fr-1-seed1701-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "771c7158ec8a69045600ad358977010f161939bec97b53054cdeb7adb7629e6f", "task_id": "vn_S_CONV_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.231926918029785, "target_z": 0.5351562496432672, "ce": 6.149696350097656, "genuineness": 4.557555198669434, "identity": 0.8354436159133911, "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_CONV_reg/fr/72d0e8be5842ac1a00739e7d.mp3" }, { "key": "original-vn_S_CONV_reg-fr-1-seed1701-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "726f172c087df86bcc1c3f9b64239728b7d8e38ad42132c0c3e97684fcb519b6", "task_id": "vn_S_CONV_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.221728801727295, "target_z": 0.5273436989872122, "ce": 6.308549880981445, "genuineness": 4.532451152801514, "identity": 0.7271417379379272, "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_CONV_reg/fr/a15996eabf2e32d615a43b80.mp3" }, { "key": "original-vn_S_CONV_reg-fr-1-seed1701-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "91702a1ebc76f874cd050d1c1cc1ff33bad4f350c34d13c00270893650fecf9f", "task_id": "vn_S_CONV_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.221728801727295, "target_z": 0.5273436989872122, "ce": 6.109982967376709, "genuineness": 4.532451152801514, "identity": 0.7253443598747253, "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_CONV_reg/fr/c1fe0cf3d1569731e35341eb.mp3" }, { "key": "original-vn_S_CONV_reg-fr-1-seed1702-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "1bfd023820ed7516291e475ae40b1ac01f829c2f0d50de9e27a1ca28905ad98b", "task_id": "vn_S_CONV_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.874995231628418, "target_z": 0.26171880315318446, "ce": 6.141753673553467, "genuineness": 4.6203155517578125, "identity": 0.7177805304527283, "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_CONV_reg/fr/f5fee9964bfb2b49a452953c.mp3" }, { "key": "original-vn_S_CONV_reg-fr-1-seed1702-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "59c67b0b730fd8f7b45ce9e712c0f7ac310bdc526e71cf230f92620ea395bab0", "task_id": "vn_S_CONV_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.8795032501220703, "target_z": 1.0312501113006278, "ce": 6.24500846862793, "genuineness": 4.758388519287109, "identity": 0.6571024656295776, "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_CONV_reg/fr/98ec0ccd6fd02a4198c4ae54.mp3" }, { "key": "original-vn_S_CONV_reg-fr-1-seed1702-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "ed0e9cff4f801f7ebbb5dcce1a6592fc59b81cad4aab3c2b55b4deb9684c9ae0", "task_id": "vn_S_CONV_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.2064318656921387, "target_z": 0.5156250556503139, "ce": 6.070269584655762, "genuineness": 4.7081804275512695, "identity": 0.7060475945472717, "hypothesis": "J'ai trouvé la 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/audio/vn_S_CONV_reg/fr/7639b7345a7b5411ee486f8d.mp3" }, { "key": "original-vn_S_CONV_reg-fr-1-seed1702-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "5b2ca21ff66a26f5e879bc3bc43e65ffa116e5d598df430116235b29101b4534", "task_id": "vn_S_CONV_reg-fr-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.517472267150879, "target_z": 0.7539062068353334, "ce": 6.046441555023193, "genuineness": 4.482243061065674, "identity": 0.8281288146972656, "hypothesis": "J'ai trouvé la petite 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/audio/vn_S_CONV_reg/fr/54aa12f703c8cc5428d4a995.mp3" }, { "key": "original-vn_S_CONV_reg-nl-0-seed1701-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "0babfd43c734e2fe3944fe6362949d827ab368d0293b6550c46631a1a6b4cb04", "task_id": "vn_S_CONV_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.0183589458465576, "target_z": -0.3945312510701983, "ce": 6.070269584655762, "genuineness": 4.030367851257324, "identity": 0.6534460783004761, "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_CONV_reg/nl/e74c3d31fd0a5f51c6ee3174.mp3" }, { "key": "original-vn_S_CONV_reg-nl-0-seed1701-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "b034f214fe0a9f03cf3aa89c04e5464cfe68d9b5ac6a57dcb2757137554a50b8", "task_id": "vn_S_CONV_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.7143759727478027, "target_z": 0.13867195223264722, "ce": 6.1894097328186035, "genuineness": 4.300237655639648, "identity": 0.7654026746749878, "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_CONV_reg/nl/9054f43a89c73a26cb83c772.mp3" }, { "key": "original-vn_S_CONV_reg-nl-0-seed1701-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": 127, "seconds_generated": 10.16, "ended": true, "contract_sha256": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "c5eff0ff378ab26714c43983964bd838b59f7670b8823bccbb978f1fe5ef2502", "task_id": "vn_S_CONV_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.3390064239501953, "target_z": 0.6171874835902921, "ce": 6.1020402908325195, "genuineness": 4.20609712600708, "identity": 0.7835365533828735, "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/audio/vn_S_CONV_reg/nl/a90bbef4f9b83606d5c9dce7.mp3" }, { "key": "original-vn_S_CONV_reg-nl-0-seed1701-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "91a34491414193017f20ffdd48e13ae4d04b828d352000e21f24673a623c4ded", "task_id": "vn_S_CONV_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.4511849880218506, "target_z": 0.7031249928653444, "ce": 6.078212261199951, "genuineness": 4.268857479095459, "identity": 0.6795057654380798, "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_CONV_reg/nl/122761ef8a4dadea1d73fa37.mp3" }, { "key": "original-vn_S_CONV_reg-nl-0-seed1702-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "b829e4f6f95c622567d01e20897529535501ecf0f1a48d441850e67ab06a633c", "task_id": "vn_S_CONV_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.279684066772461, "target_z": -0.1943358733772825, "ce": 6.0305562019348145, "genuineness": 4.017816066741943, "identity": 0.6407754421234131, "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_CONV_reg/nl/90ae54e5265eade5bbe2abc7.mp3" }, { "key": "original-vn_S_CONV_reg-nl-0-seed1702-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "45eac3936386bc335f732d267b83b24cb7fd31e79c0eab383bc7d32011351489", "task_id": "vn_S_CONV_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.920295238494873, "target_z": 1.0624999486304794, "ce": 6.109982967376709, "genuineness": 4.218649387359619, "identity": 0.6952003240585327, "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/audio/vn_S_CONV_reg/nl/1314f4148039fa8328618a0e.mp3" }, { "key": "original-vn_S_CONV_reg-nl-0-seed1702-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "de1b4cd0bd34e8a9e23e3c4d0864cdca4c25072ba644b22eb3be2ba7758aef78", "task_id": "vn_S_CONV_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.512373447418213, "target_z": 0.7500001141544901, "ce": 6.09409761428833, "genuineness": 4.419482707977295, "identity": 0.6855734586715698, "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_CONV_reg/nl/9841280bcccbef4d53475dd6.mp3" }, { "key": "original-vn_S_CONV_reg-nl-0-seed1702-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "eee40d3d5910e272011ef2c2414e7f72e24ce61b41a847078e4c9233fe9ea845", "task_id": "vn_S_CONV_reg-nl-0", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.2625210285186768, "target_z": 0.558593718964248, "ce": 6.125868320465088, "genuineness": 4.231201171875, "identity": 0.7169950008392334, "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_CONV_reg/nl/bb36a42fea4c152afc747274.mp3" }, { "key": "original-vn_S_CONV_reg-nl-1-seed1701-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "1e8c46c6462ce52aac12855960610674cc9828493bdac2abc0076611366dbe0c", "task_id": "vn_S_CONV_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.8596980571746826, "target_z": 0.24999997716910197, "ce": 6.1894097328186035, "genuineness": 4.281409740447998, "identity": 0.7704473733901978, "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_CONV_reg/nl/5268a15fe4e534686c4f87a5.mp3" }, { "key": "original-vn_S_CONV_reg-nl-1-seed1701-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "2585d1bff8ab3ae57b96155072f11916ece0af59e037207bd314b6adec492c29", "task_id": "vn_S_CONV_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.5021753311157227, "target_z": 0.7421875634984352, "ce": 5.998785495758057, "genuineness": 4.394378662109375, "identity": 0.7990095615386963, "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_CONV_reg/nl/069d7bb2769756d965f788f7.mp3" }, { "key": "original-vn_S_CONV_reg-nl-1-seed1701-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "ed1f8e08c2f6d3c923d8cec0d53fea8dbaebd96788e7299395e05b8aacbc77a7", "task_id": "vn_S_CONV_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.5027663707733154, "target_z": -0.023437469320980783, "ce": 6.125868320465088, "genuineness": 4.362998008728027, "identity": 0.6259584426879883, "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_CONV_reg/nl/bf18e1503cd34e2c892dc204.mp3" }, { "key": "original-vn_S_CONV_reg-nl-1-seed1701-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "cb636b353200300f9ee696315023969f3829c07eb2d2ab20637ad650a767d0cf", "task_id": "vn_S_CONV_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 1.4676642417907715, "target_z": -0.816406246789405, "ce": 6.197352409362793, "genuineness": 4.143336772918701, "identity": 0.4849599599838257, "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_CONV_reg/nl/69520c6e1bfd4d136d9ce736.mp3" }, { "key": "original-vn_S_CONV_reg-nl-1-seed1702-scale0.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "ee632c206f4b16093631a830385fa6d4d7c8aaac5b56625190a0cf4782c6356b", "task_id": "vn_S_CONV_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.133086919784546, "target_z": -0.3066406954547244, "ce": 6.165581703186035, "genuineness": 4.13078498840332, "identity": 0.6435294151306152, "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_CONV_reg/nl/945b789deccf67876bec916a.mp3" }, { "key": "original-vn_S_CONV_reg-nl-1-seed1702-scale0.5", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "927070facf503a0c43fd452d24a9b6b1ab1985b8734ad4dcccb55d3905db1629", "task_id": "vn_S_CONV_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.695891857147217, "target_z": 0.12451168133764953, "ce": 6.141753673553467, "genuineness": 4.444586753845215, "identity": 0.5304542183876038, "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_CONV_reg/nl/e5d0312de44c039af1bb43b4.mp3" }, { "key": "original-vn_S_CONV_reg-nl-1-seed1702-scale0.75", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "4c4d20909483ef62307faf2e513e803ef24ce973737de2726ef0785292d6e888", "task_id": "vn_S_CONV_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 3.481779098510742, "target_z": 0.7265624621863251, "ce": 6.213237762451172, "genuineness": 4.507347106933594, "identity": 0.7222671508789062, "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_CONV_reg/nl/c450e4ea4ae099c0d47e2572.mp3" }, { "key": "original-vn_S_CONV_reg-nl-1-seed1702-scale1.0", "bucket": "vn_S_CONV_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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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 Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.", "Strong Conversational 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": "1d31e9a54aff9907719f39e868d49a83b5093f78f7671b8e8ad072a4e6f7bb71", "mp3_sha256": "dce8e14fc4d2f295c2f1a16d2b7b6cf83778a09ce3add87fa44865094d82cbab", "task_id": "vn_S_CONV_reg-nl-1", "phase": "standard", "candidate": "standard", "neutral_prompt": "GENERAL: One adult speaker in a clean recording. Strong Conversational Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.\nSCRIPT:\n(Strong Conversational 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 Conversational 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 Conversational 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": 2.2172210216522217, "target_z": -0.24218742651304698, "ce": 6.165581703186035, "genuineness": 4.193545341491699, "identity": 0.7680381536483765, "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_CONV_reg/nl/9f1d29975f0a4a0fe4de24c8.mp3" } ], "source_standard_audio_unchanged": true }