The merged parent used in these specialist comparisons is now released as Humaneness Voice Large 1.6 Alpha: 1.5 + final RealSpeech rank 4, merged once. Apply only the selected new bucket adapter. Do not reload either ancestor.
1.5 Original model · 1.6 merged model · Complete context manuals for humans and agents
Read 1.5 context manual · Read 1.6 + adapter context manual. Exact prompts, CFG, adapter scales and failed gates are retained. New model/documentation contributions: CC BY 4.0; existing upstream and gated specialist terms remain applicable.
The original grid used generic GENERAL/SCRIPT directions at CFG1. It did not systematically use target-specific manual winners. The new study freezes the actual rank-4 manual wording, preserves every literal word, seed, reference, adapter and frame budget, then independently varies CFG and adapter strength. The previous recordings and conclusions remain below.
Ranking:100% requested EARS emotion/style head, standardized using frozen training statistics; low-pole targets reverse the sign. Formal/casual contrasts average the desired head and negative opposite head. CE and genuineness are diagnostics. Higher target z means stronger desired delivery, not a human rating. More CFG is not automatically better.
Selection:20 original-text outcomes per configuration select wording, CFG and scale. Fifty additional unseen-text reference pairs test the frozen choices. A recommended adapter must improve against both the same-CFG no-adapter control and the best manual-only setting. Quality gates: per-take WER≤15%, mean WER≤10%, valid/terminated fraction≥90%, paired identity change≥−0.05 over50pairs. Failed outcomes stay visible. Language-cluster bootstrap intervals use only five fixed reference voices; findings across72targets remain exploratory.
Five languages: EN/DE/FR/NL/ES; original seeds1701/1702; reference187frames/14.96seconds. CFG1/1.25/1.5/2/2.5/3/4/5 × bucket strength0/0.5/0.75/1. CFG4/5 are stress conditions, not presumed improvements. No-reference use of these bucket recommendations has not been validated.
Taxonomy:S_MONO means Monologue, S_TECH means Teacher/Didactic, VOLT means Volatility. The two corrected style recipes and17dimension/contrast recipes are new searches, not historical manual winners.53other targets reuse directly applicable manual recipes.
Best tested agent route: No new configuration passes the held-out gates; retain standard fallback and inspect audio. Bucket scale 0.0, CFG 1.0.
| Held-out condition | n | Target z↑ | Raw head | CE↑ | Genuine↑ | Voice↑ | WER↓ | Valid | QA |
|---|---|---|---|---|---|---|---|---|---|
| standard | 50 | -0.313 | 2.449 | 6.255 | 4.342 | 0.584 | 0.009 | 1.000 | pass |
| manual | 50 | 0.993 | 3.937 | 5.768 | 4.152 | 0.464 | 0.016 | 1.000 | fail |
| adapter | 50 | 1.014 | 3.961 | 5.682 | 4.117 | 0.454 | 0.024 | 0.980 | fail |
| matched_no_adapter | 50 | 0.993 | 3.937 | 5.768 | 4.152 | 0.464 | 0.016 | 1.000 | fail |
Incremental adapter gain versus same-CFG control: 0.021 z;95% cluster interval [-0.09390625801093073, 0.14990232131530667]. Versus best manual-only: 0.021 z. Manual-only versus standard: 1.306 z.
| Scale | CFG | Target z | CE | Genuine | WER | Provisional QA |
|---|---|---|---|---|---|---|
| 0.0 | 1.0 | -0.241 | 6.221 | 4.420 | 0.000 | True |
| 0.5 | 1.0 | -0.013 | 6.183 | 4.342 | 0.003 | True |
| 0.75 | 1.0 | -0.049 | 6.214 | 4.345 | 0.016 | True |
| 1.0 | 1.0 | -0.053 | 6.185 | 4.323 | 0.003 | True |
| 0.0 | 1.25 | 0.022 | 6.246 | 4.413 | 0.000 | True |
| 0.5 | 1.25 | 0.004 | 6.196 | 4.381 | 0.006 | True |
| 0.75 | 1.25 | -0.026 | 6.164 | 4.355 | 0.000 | True |
| 1.0 | 1.25 | 0.104 | 6.187 | 4.350 | 0.047 | True |
| 0.0 | 1.5 | 0.018 | 6.213 | 4.499 | 0.000 | True |
| 0.5 | 1.5 | 0.066 | 6.193 | 4.385 | 0.000 | True |
| 0.75 | 1.5 | 0.312 | 6.166 | 4.399 | 0.000 | True |
| 1.0 | 1.5 | 0.230 | 6.160 | 4.392 | 0.005 | True |
| 0.0 | 2.0 | 0.161 | 6.196 | 4.507 | 0.021 | True |
| 0.5 | 2.0 | 0.394 | 6.104 | 4.448 | 0.008 | True |
| 0.75 | 2.0 | 0.386 | 6.118 | 4.419 | 0.000 | True |
| 1.0 | 2.0 | 0.337 | 6.056 | 4.395 | 0.000 | True |
| 0.0 | 2.5 | 0.408 | 6.129 | 4.517 | 0.022 | True |
| 0.5 | 2.5 | 0.643 | 6.032 | 4.446 | 0.011 | True |
| 0.75 | 2.5 | 0.665 | 5.953 | 4.343 | 0.009 | True |
| 1.0 | 2.5 | 0.806 | 5.860 | 4.255 | 0.009 | True |
| 0.0 | 3.0 | 0.658 | 5.998 | 4.447 | 0.000 | True |
| 0.5 | 3.0 | 0.765 | 5.949 | 4.384 | 0.003 | False |
| 0.75 | 3.0 | 1.038 | 5.844 | 4.353 | 0.003 | True |
| 1.0 | 3.0 | 1.017 | 5.701 | 4.203 | 0.008 | False |
| 0.0 | 4.0 | 0.880 | 5.839 | 4.369 | 0.027 | True |
| 0.5 | 4.0 | 1.043 | 5.638 | 4.234 | 0.045 | True |
| 0.75 | 4.0 | 1.114 | 5.606 | 4.157 | 0.008 | False |
| 1.0 | 4.0 | 1.090 | 5.531 | 4.156 | 0.009 | False |
| 0.0 | 5.0 | 1.162 | 5.671 | 4.196 | 0.000 | False |
| 0.5 | 5.0 | 1.145 | 5.561 | 4.156 | 0.003 | False |
| 0.75 | 5.0 | 1.398 | 5.461 | 4.005 | 0.044 | False |
| 1.0 | 5.0 | 1.178 | 5.384 | 4.057 | 0.005 | False |
direct_manual_recipe
{
"id": "manual-21a285777a88",
"general": "Employ extreme dynamic shifts. Start almost whispered, suggesting secrecy or fear, then transition abruptly to a full, resonant declaration.",
"inline": [
"Initiate at the very edge of audibility, focusing on breath control for intimate delivery.",
"Execute a sudden, powerful surge in volume and projection, making the shift jarring and intentional.",
"Hold the peak energy level briefly, then allow a controlled, slow dissipation of power."
],
"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"
}Exact machine-readable comparison · Agent tool recommendations
Strong Dramatic Style. Base: 1.5 Alpha + merged real-speech rank4; new adapter: rank32. Every row is a matched seed/text/reference test. Scale zero disables the new adapter; all other scales are applied absolutely, never compounded.
Adapter training data: {"fr": 93, "es": 74, "nl": 111, "de": 96, "en": 76}. All available accepted languages were pooled, globally shuffled, for three epochs. Peak LR 2e-05 at the end of the first-epoch warmup, then cosine decay. All original optimizer/RNG checkpoints remain saved.
Taxonomy: S_MONO is monologue, not monotone; VOLT is volatility, not volume. No claim is made that a classifier score alone proves the audible performance.
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 20 | 0.496 ± 0.116 | 6.199 ± 0.110 | 4.259 ± 0.201 | 1.32% ± 4.14 |
| Rank32 × 0.5 | 20 | 0.494 ± 0.104 | 6.181 ± 0.110 | 4.249 ± 0.165 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 20 | 0.429 ± 0.101 | 6.181 ± 0.154 | 4.205 ± 0.270 | 0.26% ± 1.18 |
| Rank32 × 1.0 | 20 | 0.476 ± 0.103 | 6.179 ± 0.162 | 4.258 ± 0.184 | 0.00% ± 0.00 |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.479 ± 0.075 | 6.261 ± 0.086 | 4.096 ± 0.252 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.476 ± 0.106 | 6.229 ± 0.117 | 4.146 ± 0.231 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.424 ± 0.083 | 6.259 ± 0.098 | 4.018 ± 0.283 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.486 ± 0.103 | 6.338 ± 0.156 | 4.183 ± 0.215 | 0.00% ± 0.00 |
| Prompt / seed | Merged base · no new LoRA | Rank32 × 0.5 | Rank32 × 0.75 | Rank32 × 1.0 |
|---|---|---|---|---|
| Seed 1701 The door is still open. I put your coat on the chair. Tell me when you are ready. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.567 · CE 6.24 · Genuine 3.97 · WER 0.0% | Target 0.553 · CE 6.18 · Genuine 3.96 · WER 0.0% | Target 0.531 · CE 6.21 · Genuine 3.74 · WER 0.0% | Target 0.583 · CE 6.13 · Genuine 3.89 · WER 0.0% |
| Seed 1702 The door is still open. I put your coat on the chair. Tell me when you are ready. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.506 · CE 6.15 · Genuine 3.82 · WER 0.0% | Target 0.555 · CE 6.13 · Genuine 4.06 · WER 0.0% | Target 0.448 · CE 6.26 · Genuine 3.98 · WER 0.0% | Target 0.562 · CE 6.39 · Genuine 4.18 · WER 0.0% |
| Seed 1701 I found the little box. It was beside the old window. We can look inside together. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.394 · CE 6.34 · Genuine 4.39 · WER 0.0% | Target 0.329 · CE 6.21 · Genuine 4.09 · WER 0.0% | Target 0.356 · CE 6.40 · Genuine 4.41 · WER 0.0% | Target 0.427 · CE 6.51 · Genuine 4.41 · WER 0.0% |
| Seed 1702 I found the little box. It was beside the old window. We can look inside together. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.448 · CE 6.32 · Genuine 4.21 · WER 0.0% | Target 0.466 · CE 6.40 · Genuine 4.48 · WER 0.0% | Target 0.362 · CE 6.17 · Genuine 3.94 · WER 0.0% | Target 0.371 · CE 6.32 · Genuine 4.25 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.432 ± 0.074 | 6.301 ± 0.038 | 4.319 ± 0.132 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.445 ± 0.087 | 6.279 ± 0.088 | 4.346 ± 0.181 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.331 ± 0.044 | 6.334 ± 0.099 | 4.297 ± 0.080 | 1.32% ± 2.63 |
| Rank32 × 1.0 | 4 | 0.386 ± 0.039 | 6.324 ± 0.097 | 4.476 ± 0.116 | 0.00% ± 0.00 |
| Prompt / seed | Merged base · no new LoRA | Rank32 × 0.5 | Rank32 × 0.75 | Rank32 × 1.0 |
|---|---|---|---|---|
| Seed 1701 Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.535 · CE 6.28 · Genuine 4.44 · WER 0.0% | Target 0.550 · CE 6.40 · Genuine 4.56 · WER 0.0% | Target 0.354 · CE 6.21 · Genuine 4.33 · WER 5.3% | Target 0.363 · CE 6.25 · Genuine 4.38 · WER 0.0% |
| Seed 1702 Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.438 · CE 6.27 · Genuine 4.41 · WER 0.0% | Target 0.479 · CE 6.23 · Genuine 4.43 · WER 0.0% | Target 0.269 · CE 6.31 · Genuine 4.39 · WER 0.0% | Target 0.366 · CE 6.25 · Genuine 4.58 · WER 0.0% |
| Seed 1701 Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.377 · CE 6.29 · Genuine 4.27 · WER 0.0% | Target 0.353 · CE 6.29 · Genuine 4.17 · WER 0.0% | Target 0.369 · CE 6.41 · Genuine 4.21 · WER 0.0% | Target 0.443 · CE 6.45 · Genuine 4.57 · WER 0.0% |
| Seed 1702 Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.378 · CE 6.36 · Genuine 4.16 · WER 0.0% | Target 0.398 · CE 6.20 · Genuine 4.22 · WER 0.0% | Target 0.334 · CE 6.41 · Genuine 4.26 · WER 0.0% | Target 0.370 · CE 6.35 · Genuine 4.37 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.454 ± 0.084 | 6.183 ± 0.050 | 4.445 ± 0.170 | 3.95% ± 7.89 |
| Rank32 × 0.5 | 4 | 0.497 ± 0.055 | 6.166 ± 0.114 | 4.358 ± 0.094 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.459 ± 0.060 | 6.162 ± 0.146 | 4.372 ± 0.307 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.491 ± 0.066 | 6.025 ± 0.086 | 4.242 ± 0.151 | 0.00% ± 0.00 |
| Prompt / seed | Merged base · no new LoRA | Rank32 × 0.5 | Rank32 × 0.75 | Rank32 × 1.0 |
|---|---|---|---|---|
| Seed 1701 La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.342 · CE 6.21 · Genuine 4.63 · WER 0.0% | Target 0.421 · CE 6.10 · Genuine 4.26 · WER 0.0% | Target 0.533 · CE 6.17 · Genuine 3.96 · WER 0.0% | Target 0.515 · CE 5.94 · Genuine 4.08 · WER 0.0% |
| Seed 1702 La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.516 · CE 6.11 · Genuine 4.23 · WER 0.0% | Target 0.516 · CE 6.21 · Genuine 4.44 · WER 0.0% | Target 0.470 · CE 6.17 · Genuine 4.37 · WER 0.0% | Target 0.568 · CE 6.00 · Genuine 4.39 · WER 0.0% |
| Seed 1701 J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.521 · CE 6.20 · Genuine 4.41 · WER 15.8% | Target 0.551 · CE 6.05 · Genuine 4.30 · WER 0.0% | Target 0.388 · CE 5.97 · Genuine 4.48 · WER 0.0% | Target 0.465 · CE 6.02 · Genuine 4.34 · WER 0.0% |
| Seed 1702 J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.437 · CE 6.22 · Genuine 4.51 · WER 0.0% | Target 0.501 · CE 6.30 · Genuine 4.43 · WER 0.0% | Target 0.446 · CE 6.33 · Genuine 4.68 · WER 0.0% | Target 0.415 · CE 6.14 · Genuine 4.15 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.593 ± 0.197 | 6.170 ± 0.128 | 4.234 ± 0.163 | 2.63% ± 5.26 |
| Rank32 × 0.5 | 4 | 0.548 ± 0.156 | 6.148 ± 0.098 | 4.183 ± 0.107 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.536 ± 0.091 | 6.126 ± 0.037 | 4.245 ± 0.096 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.594 ± 0.123 | 6.162 ± 0.060 | 4.205 ± 0.095 | 0.00% ± 0.00 |
| Prompt / seed | Merged base · no new LoRA | Rank32 × 0.5 | Rank32 × 0.75 | Rank32 × 1.0 |
|---|---|---|---|---|
| Seed 1701 De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.852 · CE 6.14 · Genuine 4.09 · WER 0.0% | Target 0.552 · CE 6.05 · Genuine 4.23 · WER 0.0% | Target 0.502 · CE 6.11 · Genuine 4.35 · WER 0.0% | Target 0.624 · CE 6.09 · Genuine 4.24 · WER 0.0% |
| Seed 1702 De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.389 · CE 6.01 · Genuine 4.18 · WER 10.5% | Target 0.607 · CE 6.13 · Genuine 4.12 · WER 0.0% | Target 0.629 · CE 6.18 · Genuine 4.26 · WER 0.0% | Target 0.676 · CE 6.14 · Genuine 4.07 · WER 0.0% |
| Seed 1701 Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.622 · CE 6.31 · Genuine 4.47 · WER 0.0% | Target 0.700 · CE 6.28 · Genuine 4.07 · WER 0.0% | Target 0.426 · CE 6.11 · Genuine 4.26 · WER 0.0% | Target 0.663 · CE 6.21 · Genuine 4.29 · WER 0.0% |
| Seed 1702 Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic 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. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.509 · CE 6.22 · Genuine 4.19 · WER 0.0% | Target 0.332 · CE 6.12 · Genuine 4.31 · WER 0.0% | Target 0.589 · CE 6.10 · Genuine 4.12 · WER 0.0% | Target 0.412 · CE 6.21 · Genuine 4.22 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.521 ± 0.079 | 6.080 ± 0.096 | 4.198 ± 0.170 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.506 ± 0.119 | 6.086 ± 0.056 | 4.214 ± 0.125 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.396 ± 0.113 | 6.025 ± 0.177 | 4.090 ± 0.394 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.423 ± 0.051 | 6.046 ± 0.066 | 4.186 ± 0.200 | 0.00% ± 0.00 |
| Prompt / seed | Merged base · no new LoRA | Rank32 × 0.5 | Rank32 × 0.75 | Rank32 × 1.0 |
|---|---|---|---|---|
| Seed 1701 La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Strong Dramatic 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. (Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.619 · CE 6.02 · Genuine 4.17 · WER 0.0% | Target 0.549 · CE 6.09 · Genuine 4.08 · WER 0.0% | Target 0.465 · CE 5.79 · Genuine 3.71 · WER 0.0% | Target 0.439 · CE 6.13 · Genuine 3.94 · WER 0.0% |
| Seed 1702 La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Strong Dramatic 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. (Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.550 · CE 5.97 · Genuine 4.01 · WER 0.0% | Target 0.648 · CE 6.05 · Genuine 4.19 · WER 0.0% | Target 0.500 · CE 6.05 · Genuine 3.87 · WER 0.0% | Target 0.483 · CE 5.97 · Genuine 4.11 · WER 0.0% |
| Seed 1701 He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.443 · CE 6.15 · Genuine 4.42 · WER 0.0% | Target 0.370 · CE 6.05 · Genuine 4.38 · WER 0.0% | Target 0.248 · CE 6.03 · Genuine 4.61 · WER 0.0% | Target 0.407 · CE 6.03 · Genuine 4.31 · WER 0.0% |
| Seed 1702 He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Dramatic Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Strong Dramatic 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. (Strong Dramatic 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. | Target 0.474 · CE 6.17 · Genuine 4.19 · WER 0.0% | Target 0.459 · CE 6.17 · Genuine 4.21 · WER 0.0% | Target 0.368 · CE 6.22 · Genuine 4.17 · WER 0.0% | Target 0.363 · CE 6.05 · Genuine 4.38 · WER 0.0% |