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: Use manual-only configuration; no clear qualifying incremental adapter gain. Bucket scale 0.0, CFG 1.0.
| Held-out condition | n | Target z↑ | Raw head | CE↑ | Genuine↑ | Voice↑ | WER↓ | Valid | QA |
|---|---|---|---|---|---|---|---|---|---|
| standard | 50 | 0.679 | 2.052 | 6.221 | 4.418 | 0.601 | 0.020 | 0.960 | pass |
| manual | 50 | 0.831 | 1.902 | 6.246 | 4.329 | 0.568 | 0.008 | 1.000 | pass |
| adapter | 50 | 0.872 | 1.861 | 6.245 | 4.320 | 0.583 | 0.007 | 1.000 | pass |
| matched_no_adapter | 50 | 0.831 | 1.902 | 6.246 | 4.329 | 0.568 | 0.008 | 1.000 | pass |
Incremental adapter gain versus same-CFG control: 0.041 z;95% cluster interval [-0.0814257627215332, 0.15260743493542422]. Versus best manual-only: 0.041 z. Manual-only versus standard: 0.151 z.
| Scale | CFG | Target z | CE | Genuine | WER | Provisional QA |
|---|---|---|---|---|---|---|
| 0.0 | 1.0 | 0.796 | 6.190 | 4.303 | 0.011 | True |
| 0.5 | 1.0 | 0.519 | 6.199 | 4.413 | 0.003 | True |
| 0.75 | 1.0 | 0.810 | 6.191 | 4.255 | 0.000 | True |
| 1.0 | 1.0 | 0.766 | 6.173 | 4.296 | 0.000 | True |
| 0.0 | 1.25 | 0.743 | 6.192 | 4.350 | 0.008 | True |
| 0.5 | 1.25 | 0.619 | 6.182 | 4.335 | 0.000 | True |
| 0.75 | 1.25 | 0.679 | 6.199 | 4.311 | 0.008 | True |
| 1.0 | 1.25 | 0.692 | 6.179 | 4.289 | 0.000 | False |
| 0.0 | 1.5 | 0.671 | 6.194 | 4.347 | 0.000 | True |
| 0.5 | 1.5 | 0.624 | 6.177 | 4.341 | 0.003 | True |
| 0.75 | 1.5 | 0.747 | 6.195 | 4.294 | 0.009 | True |
| 1.0 | 1.5 | 0.738 | 6.215 | 4.202 | 0.028 | True |
| 0.0 | 2.0 | 0.638 | 6.220 | 4.334 | 0.014 | True |
| 0.5 | 2.0 | 0.752 | 6.150 | 4.269 | 0.006 | True |
| 0.75 | 2.0 | 0.659 | 6.143 | 4.268 | 0.006 | True |
| 1.0 | 2.0 | 0.649 | 6.156 | 4.260 | 0.000 | True |
| 0.0 | 2.5 | 0.524 | 6.203 | 4.348 | 0.000 | True |
| 0.5 | 2.5 | 0.540 | 6.149 | 4.288 | 0.003 | True |
| 0.75 | 2.5 | 0.669 | 6.166 | 4.298 | 0.000 | True |
| 1.0 | 2.5 | 0.445 | 6.184 | 4.293 | 0.006 | True |
| 0.0 | 3.0 | 0.383 | 6.149 | 4.320 | 0.011 | True |
| 0.5 | 3.0 | 0.450 | 6.132 | 4.300 | 0.031 | True |
| 0.75 | 3.0 | 0.590 | 6.197 | 4.299 | 0.003 | True |
| 1.0 | 3.0 | 0.672 | 6.145 | 4.252 | 0.012 | True |
| 0.0 | 4.0 | 0.385 | 6.141 | 4.291 | 0.003 | True |
| 0.5 | 4.0 | 0.238 | 6.140 | 4.354 | 0.015 | True |
| 0.75 | 4.0 | 0.511 | 6.150 | 4.302 | 0.010 | True |
| 1.0 | 4.0 | 0.356 | 6.137 | 4.351 | 0.019 | True |
| 0.0 | 5.0 | 0.036 | 6.170 | 4.398 | 0.006 | True |
| 0.5 | 5.0 | 0.004 | 6.124 | 4.317 | 0.032 | True |
| 0.75 | 5.0 | 0.210 | 6.131 | 4.332 | 0.021 | True |
| 1.0 | 5.0 | 0.182 | 6.098 | 4.291 | 0.031 | True |
new_canonical_dimension_extension
{
"id": "canonical-extension-P07",
"general": "Emotionally guarded and self-contained, with protected feeling, firm breath and composed articulation. React naturally to a silent partner and keep one consistent adult voice.",
"inline": [
"Emotionally guarded and self-contained, with protected feeling, firm breath and composed articulation. React naturally to a silent partner and keep one consistent adult voice.",
"Emotionally guarded and self-contained, with protected feeling, firm breath and composed articulation. React naturally to a silent partner and keep one consistent adult voice.",
"Emotionally guarded and self-contained, with protected feeling, firm breath and composed articulation. React naturally to a silent partner and keep one consistent adult voice."
],
"recipe": "P07",
"cfg": 1.0,
"timing_syntax": "canonical",
"canonical_labels": true,
"lexical_proxy": false,
"origin": "new_canonical_manual_extension_requires_new_evidence"
}Exact machine-readable comparison · Agent tool recommendations
emotionally detached and unexposed. 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": 58, "es": 45, "de": 46, "nl": 57, "en": 37}. 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 | 0.711 ± 0.153 | 6.163 ± 0.124 | 4.435 ± 0.196 | 0.26% ± 1.18 |
| Rank32 × 0.5 | 20 | 0.632 ± 0.162 | 6.149 ± 0.116 | 4.434 ± 0.170 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 20 | 0.662 ± 0.159 | 6.121 ± 0.160 | 4.367 ± 0.232 | 0.29% ± 1.32 |
| Rank32 × 1.0 | 20 | 0.708 ± 0.143 | 6.169 ± 0.117 | 4.338 ± 0.153 | 1.12% ± 2.88 |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.556 ± 0.187 | 6.283 ± 0.069 | 4.413 ± 0.206 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.446 ± 0.125 | 6.309 ± 0.103 | 4.349 ± 0.152 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.597 ± 0.115 | 6.295 ± 0.089 | 4.212 ± 0.201 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.555 ± 0.136 | 6.203 ± 0.069 | 4.283 ± 0.113 | 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.759 · CE 6.19 · Genuine 4.17 · WER 0.0% | Target 0.533 · CE 6.18 · Genuine 4.14 · WER 0.0% | Target 0.539 · CE 6.37 · Genuine 4.06 · WER 0.0% | Target 0.462 · CE 6.22 · Genuine 4.38 · 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.673 · CE 6.30 · Genuine 4.33 · WER 0.0% | Target 0.572 · CE 6.27 · Genuine 4.33 · WER 0.0% | Target 0.760 · CE 6.22 · Genuine 4.08 · WER 0.0% | Target 0.750 · CE 6.12 · Genuine 4.12 · 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.395 · CE 6.28 · Genuine 4.65 · WER 0.0% | Target 0.318 · CE 6.39 · Genuine 4.44 · WER 0.0% | Target 0.590 · CE 6.21 · Genuine 4.22 · WER 0.0% | Target 0.463 · CE 6.28 · Genuine 4.33 · 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.400 · CE 6.36 · Genuine 4.51 · WER 0.0% | Target 0.363 · CE 6.40 · Genuine 4.48 · WER 0.0% | Target 0.498 · CE 6.37 · Genuine 4.49 · WER 0.0% | Target 0.543 · CE 6.19 · Genuine 4.31 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.816 ± 0.085 | 6.293 ± 0.078 | 4.419 ± 0.154 | 1.32% ± 2.63 |
| Rank32 × 0.5 | 4 | 0.735 ± 0.126 | 6.189 ± 0.103 | 4.462 ± 0.219 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.826 ± 0.023 | 6.205 ± 0.080 | 4.259 ± 0.118 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.815 ± 0.046 | 6.281 ± 0.178 | 4.307 ± 0.061 | 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.840 · CE 6.30 · Genuine 4.48 · WER 5.3% | Target 0.742 · CE 6.08 · Genuine 4.58 · WER 0.0% | Target 0.856 · CE 6.17 · Genuine 4.39 · WER 0.0% | Target 0.797 · CE 6.28 · Genuine 4.26 · 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.894 · CE 6.18 · Genuine 4.30 · WER 0.0% | Target 0.555 · CE 6.13 · Genuine 4.71 · WER 0.0% | Target 0.830 · CE 6.21 · Genuine 4.33 · WER 0.0% | Target 0.813 · CE 6.03 · Genuine 4.38 · 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.836 · CE 6.33 · Genuine 4.29 · WER 0.0% | Target 0.832 · CE 6.28 · Genuine 4.31 · WER 0.0% | Target 0.810 · CE 6.32 · Genuine 4.17 · WER 0.0% | Target 0.878 · CE 6.42 · Genuine 4.26 · 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.696 · CE 6.36 · Genuine 4.61 · WER 0.0% | Target 0.810 · CE 6.27 · Genuine 4.25 · WER 0.0% | Target 0.806 · CE 6.13 · Genuine 4.15 · WER 0.0% | Target 0.770 · CE 6.40 · Genuine 4.33 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.634 ± 0.103 | 6.112 ± 0.023 | 4.548 ± 0.190 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.584 ± 0.043 | 6.110 ± 0.050 | 4.595 ± 0.127 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.469 ± 0.108 | 6.164 ± 0.021 | 4.639 ± 0.142 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.676 ± 0.061 | 6.102 ± 0.056 | 4.504 ± 0.201 | 2.63% ± 5.26 |
| 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.767 · CE 6.14 · Genuine 4.32 · WER 0.0% | Target 0.629 · CE 6.17 · Genuine 4.53 · WER 0.0% | Target 0.622 · CE 6.18 · Genuine 4.52 · WER 0.0% | Target 0.675 · CE 6.09 · Genuine 4.36 · 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.647 · CE 6.11 · Genuine 4.47 · WER 0.0% | Target 0.560 · CE 6.10 · Genuine 4.78 · WER 0.0% | Target 0.466 · CE 6.15 · Genuine 4.53 · WER 0.0% | Target 0.754 · CE 6.11 · Genuine 4.31 · 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.522 · CE 6.11 · Genuine 4.67 · WER 0.0% | Target 0.611 · CE 6.05 · Genuine 4.51 · WER 0.0% | Target 0.388 · CE 6.18 · Genuine 4.68 · WER 0.0% | Target 0.666 · CE 6.04 · Genuine 4.66 · WER 10.5% |
| 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.599 · CE 6.09 · Genuine 4.73 · WER 0.0% | Target 0.537 · CE 6.13 · Genuine 4.56 · WER 0.0% | Target 0.399 · CE 6.14 · Genuine 4.82 · WER 0.0% | Target 0.606 · CE 6.17 · Genuine 4.70 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.843 ± 0.121 | 6.098 ± 0.063 | 4.349 ± 0.216 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.795 ± 0.144 | 6.066 ± 0.056 | 4.332 ± 0.154 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.754 ± 0.155 | 6.007 ± 0.063 | 4.363 ± 0.185 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.847 ± 0.127 | 6.110 ± 0.065 | 4.297 ± 0.109 | 1.32% ± 2.63 |
| 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.673 · CE 6.01 · Genuine 4.20 · WER 0.0% | Target 0.730 · CE 6.09 · Genuine 4.49 · WER 0.0% | Target 0.761 · CE 5.92 · Genuine 4.52 · WER 0.0% | Target 0.665 · CE 6.01 · 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.870 · CE 6.15 · Genuine 4.57 · WER 0.0% | Target 0.634 · CE 6.00 · Genuine 4.39 · WER 0.0% | Target 0.587 · CE 6.07 · Genuine 4.47 · WER 0.0% | Target 0.858 · CE 6.13 · Genuine 4.44 · WER 5.3% |
| 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.961 · CE 6.11 · Genuine 4.13 · WER 0.0% | Target 0.966 · CE 6.13 · Genuine 4.13 · WER 0.0% | Target 0.708 · CE 6.01 · Genuine 4.36 · WER 0.0% | Target 0.918 · CE 6.13 · Genuine 4.31 · 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.866 · CE 6.13 · Genuine 4.49 · WER 0.0% | Target 0.849 · CE 6.05 · Genuine 4.31 · WER 0.0% | Target 0.960 · CE 6.02 · Genuine 4.11 · WER 0.0% | Target 0.946 · CE 6.16 · Genuine 4.19 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.705 ± 0.054 | 6.029 ± 0.089 | 4.443 ± 0.256 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.598 ± 0.109 | 6.068 ± 0.054 | 4.432 ± 0.107 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.667 ± 0.093 | 5.932 ± 0.164 | 4.360 ± 0.294 | 1.47% ± 2.94 |
| Rank32 × 1.0 | 4 | 0.650 ± 0.103 | 6.148 ± 0.118 | 4.300 ± 0.184 | 1.67% ± 3.33 |
| 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.675 · CE 6.08 · Genuine 4.33 · WER 0.0% | Target 0.438 · CE 6.05 · Genuine 4.46 · WER 0.0% | Target 0.581 · CE 5.84 · Genuine 4.26 · WER 0.0% | Target 0.551 · CE 6.01 · Genuine 4.29 · 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.745 · CE 5.90 · Genuine 4.14 · WER 0.0% | Target 0.681 · CE 6.01 · Genuine 4.28 · WER 0.0% | Target 0.796 · CE 5.76 · Genuine 3.99 · WER 0.0% | Target 0.772 · CE 6.09 · Genuine 4.04 · WER 6.7% |
| 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.645 · CE 6.06 · Genuine 4.61 · WER 0.0% | Target 0.651 · CE 6.14 · Genuine 4.53 · WER 0.0% | Target 0.666 · CE 6.01 · Genuine 4.56 · WER 5.9% | Target 0.578 · CE 6.25 · Genuine 4.44 · 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. emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (emotionally detached and unexposed. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (emotionally detached and unexposed. 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. (emotionally detached and unexposed. 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 0.755 · CE 6.08 · Genuine 4.70 · WER 0.0% | Target 0.624 · CE 6.07 · Genuine 4.46 · WER 0.0% | Target 0.626 · CE 6.12 · Genuine 4.63 · WER 0.0% | Target 0.697 · CE 6.24 · Genuine 4.42 · WER 0.0% |