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 5.0.
| Held-out condition | n | Target z↑ | Raw head | CE↑ | Genuine↑ | Voice↑ | WER↓ | Valid | QA |
|---|---|---|---|---|---|---|---|---|---|
| standard | 50 | -0.221 | 0.294 | 6.257 | 4.370 | 0.582 | 0.015 | 1.000 | pass |
| manual | 50 | 1.200 | 1.198 | 5.813 | 4.489 | 0.536 | 0.010 | 1.000 | pass |
| adapter | 50 | 1.138 | 1.159 | 5.729 | 4.452 | 0.531 | 0.022 | 1.000 | fail |
| matched_no_adapter | 50 | 1.200 | 1.198 | 5.813 | 4.489 | 0.536 | 0.010 | 1.000 | pass |
Incremental adapter gain versus same-CFG control: -0.062 z;95% cluster interval [-0.19261720909748642, 0.05173823953074554]. Versus best manual-only: -0.062 z. Manual-only versus standard: 1.421 z.
| Scale | CFG | Target z | CE | Genuine | WER | Provisional QA |
|---|---|---|---|---|---|---|
| 0.0 | 1.0 | 0.305 | 6.202 | 4.498 | 0.000 | True |
| 0.5 | 1.0 | 0.334 | 6.173 | 4.456 | 0.000 | True |
| 0.75 | 1.0 | 0.392 | 6.170 | 4.461 | 0.000 | True |
| 1.0 | 1.0 | 0.360 | 6.135 | 4.466 | 0.003 | True |
| 0.0 | 1.25 | 0.411 | 6.197 | 4.519 | 0.000 | True |
| 0.5 | 1.25 | 0.546 | 6.159 | 4.466 | 0.006 | True |
| 0.75 | 1.25 | 0.522 | 6.145 | 4.415 | 0.003 | True |
| 1.0 | 1.25 | 0.422 | 6.123 | 4.428 | 0.000 | True |
| 0.0 | 1.5 | 0.814 | 6.191 | 4.534 | 0.000 | True |
| 0.5 | 1.5 | 0.754 | 6.092 | 4.540 | 0.000 | True |
| 0.75 | 1.5 | 0.642 | 6.161 | 4.531 | 0.009 | True |
| 1.0 | 1.5 | 0.659 | 6.156 | 4.480 | 0.005 | True |
| 0.0 | 2.0 | 0.778 | 6.074 | 4.478 | 0.003 | True |
| 0.5 | 2.0 | 0.834 | 6.028 | 4.498 | 0.000 | True |
| 0.75 | 2.0 | 0.918 | 6.080 | 4.515 | 0.000 | True |
| 1.0 | 2.0 | 0.895 | 6.103 | 4.559 | 0.006 | True |
| 0.0 | 2.5 | 0.846 | 6.017 | 4.530 | 0.003 | True |
| 0.5 | 2.5 | 1.169 | 6.031 | 4.502 | 0.003 | True |
| 0.75 | 2.5 | 1.263 | 5.980 | 4.516 | 0.000 | True |
| 1.0 | 2.5 | 1.052 | 5.992 | 4.470 | 0.000 | True |
| 0.0 | 3.0 | 1.286 | 5.922 | 4.512 | 0.003 | True |
| 0.5 | 3.0 | 1.204 | 5.911 | 4.481 | 0.000 | True |
| 0.75 | 3.0 | 1.468 | 5.911 | 4.507 | 0.000 | True |
| 1.0 | 3.0 | 1.209 | 5.839 | 4.473 | 0.006 | True |
| 0.0 | 4.0 | 1.414 | 5.759 | 4.483 | 0.011 | True |
| 0.5 | 4.0 | 1.406 | 5.706 | 4.453 | 0.000 | True |
| 0.75 | 4.0 | 1.544 | 5.786 | 4.505 | 0.006 | True |
| 1.0 | 4.0 | 1.479 | 5.854 | 4.517 | 0.000 | True |
| 0.0 | 5.0 | 1.468 | 5.730 | 4.491 | 0.000 | True |
| 0.5 | 5.0 | 1.503 | 5.639 | 4.460 | 0.000 | True |
| 0.75 | 5.0 | 1.547 | 5.661 | 4.470 | 0.006 | True |
| 1.0 | 5.0 | 1.451 | 5.615 | 4.403 | 0.006 | True |
direct_manual_recipe
{
"id": "manual-bd19fff54e5e",
"general": "Focus on vocal placement; project the ache from a high, fragile place in the throat, not from the chest.",
"inline": [
"Begin with a slightly higher register, almost a fragile whisper carrying significant emotional weight.",
"Allow the vowels to stretch and thin, suggesting the yearning is delicate and easily broken.",
"Maintain a shimmering, high-frequency texture that fades gently into the silence following the utterance."
],
"cfg": 2.5,
"recipe": "P06",
"strategy": "local-Gemma-mutation",
"timing_syntax": "canonical",
"canonical_labels": true,
"neutral_event": "keep",
"parent": "cfg-2.5",
"origin": "exact_frozen_manual_recipe"
}Exact machine-readable comparison · Agent tool recommendations
Intense Longing. 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: {"de": 35, "fr": 17, "es": 31, "en": 32, "nl": 2}. 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 | 0.468 ± 0.228 | 6.189 ± 0.120 | 4.387 ± 0.173 | 0.26% ± 1.18 |
| Rank32 × 0.5 | 20 | 0.455 ± 0.220 | 6.189 ± 0.137 | 4.343 ± 0.215 | 0.26% ± 1.18 |
| Rank32 × 0.75 | 20 | 0.506 ± 0.256 | 6.210 ± 0.125 | 4.382 ± 0.183 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 20 | 0.468 ± 0.232 | 6.191 ± 0.149 | 4.301 ± 0.235 | 0.00% ± 0.00 |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.604 ± 0.339 | 6.326 ± 0.081 | 4.206 ± 0.204 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.610 ± 0.317 | 6.332 ± 0.106 | 4.248 ± 0.346 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.760 ± 0.264 | 6.328 ± 0.071 | 4.297 ± 0.209 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.633 ± 0.306 | 6.328 ± 0.142 | 4.222 ± 0.270 | 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open. (Intense Longing. 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. (Intense Longing. 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 0.306 · CE 6.38 · Genuine 3.96 · WER 0.0% | Target 0.370 · CE 6.25 · Genuine 4.23 · WER 0.0% | Target 0.398 · CE 6.27 · Genuine 4.10 · WER 0.0% | Target 0.429 · CE 6.32 · Genuine 3.94 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open. (Intense Longing. 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. (Intense Longing. 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 0.316 · CE 6.25 · Genuine 4.16 · WER 0.0% | Target 0.309 · CE 6.24 · Genuine 3.82 · WER 0.0% | Target 0.730 · CE 6.28 · Genuine 4.14 · WER 0.0% | Target 0.316 · CE 6.13 · Genuine 4.06 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box. (Intense Longing. 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. (Intense Longing. 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 0.920 · CE 6.26 · Genuine 4.26 · WER 0.0% | Target 0.941 · CE 6.44 · Genuine 4.67 · WER 0.0% | Target 0.949 · CE 6.42 · Genuine 4.44 · WER 0.0% | Target 0.927 · CE 6.41 · Genuine 4.51 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box. (Intense Longing. 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. (Intense Longing. 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 0.873 · CE 6.41 · Genuine 4.44 · WER 0.0% | Target 0.821 · CE 6.41 · Genuine 4.27 · WER 0.0% | Target 0.964 · CE 6.35 · Genuine 4.51 · WER 0.0% | Target 0.861 · CE 6.45 · Genuine 4.39 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.424 ± 0.189 | 6.255 ± 0.130 | 4.429 ± 0.178 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.375 ± 0.108 | 6.299 ± 0.118 | 4.432 ± 0.261 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.374 ± 0.144 | 6.273 ± 0.103 | 4.421 ± 0.061 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.427 ± 0.128 | 6.315 ± 0.108 | 4.482 ± 0.062 | 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen. (Intense Longing. 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. (Intense Longing. 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 0.267 · CE 6.11 · Genuine 4.38 · WER 0.0% | Target 0.255 · CE 6.32 · Genuine 4.61 · WER 0.0% | Target 0.241 · CE 6.21 · Genuine 4.42 · WER 0.0% | Target 0.395 · CE 6.28 · Genuine 4.48 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen. (Intense Longing. 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. (Intense Longing. 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 0.274 · CE 6.20 · Genuine 4.68 · WER 0.0% | Target 0.312 · CE 6.13 · Genuine 4.53 · WER 0.0% | Target 0.275 · CE 6.20 · Genuine 4.51 · WER 0.0% | Target 0.266 · CE 6.18 · Genuine 4.44 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. 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. (Intense Longing. 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. (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.494 · CE 6.41 · Genuine 4.27 · WER 0.0% | Target 0.467 · CE 6.36 · Genuine 4.55 · WER 0.0% | Target 0.428 · CE 6.27 · Genuine 4.38 · WER 0.0% | Target 0.483 · CE 6.37 · Genuine 4.43 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. 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. (Intense Longing. 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. (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.660 · CE 6.30 · Genuine 4.38 · WER 0.0% | Target 0.467 · CE 6.38 · Genuine 4.04 · WER 0.0% | Target 0.552 · CE 6.42 · Genuine 4.38 · WER 0.0% | Target 0.564 · CE 6.43 · Genuine 4.57 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.549 ± 0.288 | 6.156 ± 0.093 | 4.435 ± 0.076 | 1.32% ± 2.63 |
| Rank32 × 0.5 | 4 | 0.587 ± 0.256 | 6.132 ± 0.086 | 4.404 ± 0.107 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.705 ± 0.272 | 6.195 ± 0.069 | 4.489 ± 0.143 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.617 ± 0.287 | 6.112 ± 0.059 | 4.438 ± 0.188 | 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte. (Intense Longing. 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. (Intense Longing. 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 0.282 · CE 6.24 · Genuine 4.37 · WER 0.0% | Target 0.395 · CE 6.08 · Genuine 4.32 · WER 0.0% | Target 0.455 · CE 6.17 · Genuine 4.37 · WER 0.0% | Target 0.468 · CE 6.07 · Genuine 4.31 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte. (Intense Longing. 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. (Intense Longing. 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 0.319 · CE 6.17 · Genuine 4.46 · WER 0.0% | Target 0.344 · CE 6.05 · Genuine 4.34 · WER 0.0% | Target 0.483 · CE 6.23 · Genuine 4.44 · WER 0.0% | Target 0.286 · CE 6.05 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. 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. (Intense Longing. 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. (Intense Longing. 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 0.784 · CE 6.20 · Genuine 4.53 · WER 0.0% | Target 0.856 · CE 6.17 · Genuine 4.39 · WER 0.0% | Target 0.929 · CE 6.27 · Genuine 4.44 · WER 0.0% | Target 0.873 · CE 6.17 · Genuine 4.71 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. 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. (Intense Longing. 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. (Intense Longing. 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 0.813 · CE 6.02 · Genuine 4.38 · WER 5.3% | Target 0.755 · CE 6.24 · Genuine 4.56 · WER 0.0% | Target 0.952 · CE 6.11 · Genuine 4.70 · WER 0.0% | Target 0.841 · CE 6.15 · Genuine 4.43 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.275 ± 0.015 | 6.114 ± 0.066 | 4.368 ± 0.114 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.266 ± 0.009 | 6.110 ± 0.094 | 4.270 ± 0.103 | 1.32% ± 2.63 |
| Rank32 × 0.75 | 4 | 0.273 ± 0.012 | 6.166 ± 0.108 | 4.299 ± 0.155 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.263 ± 0.006 | 6.038 ± 0.097 | 4.263 ± 0.053 | 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open. (Intense Longing. 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. (Intense Longing. 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 0.269 · CE 6.11 · Genuine 4.20 · WER 0.0% | Target 0.272 · CE 5.98 · Genuine 4.24 · WER 5.3% | Target 0.269 · CE 6.26 · Genuine 4.14 · WER 0.0% | Target 0.261 · CE 5.95 · Genuine 4.21 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open. (Intense Longing. 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. (Intense Longing. 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 0.263 · CE 6.02 · Genuine 4.39 · WER 0.0% | Target 0.255 · CE 6.11 · Genuine 4.17 · WER 0.0% | Target 0.284 · CE 6.03 · Genuine 4.23 · WER 0.0% | Target 0.271 · CE 5.96 · Genuine 4.33 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. 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. (Intense Longing. 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. (Intense Longing. 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 0.296 · CE 6.17 · Genuine 4.44 · WER 0.0% | Target 0.274 · CE 6.14 · Genuine 4.41 · WER 0.0% | Target 0.258 · CE 6.13 · Genuine 4.31 · WER 0.0% | Target 0.263 · CE 6.13 · Genuine 4.27 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. 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. (Intense Longing. 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. (Intense Longing. 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 0.270 · CE 6.15 · Genuine 4.43 · WER 0.0% | Target 0.262 · CE 6.21 · Genuine 4.26 · WER 0.0% | Target 0.282 · CE 6.25 · Genuine 4.51 · WER 0.0% | Target 0.258 · CE 6.11 · Genuine 4.24 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.487 ± 0.098 | 6.094 ± 0.042 | 4.498 ± 0.180 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.438 ± 0.125 | 6.072 ± 0.053 | 4.361 ± 0.219 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.419 ± 0.065 | 6.090 ± 0.145 | 4.407 ± 0.288 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.401 ± 0.139 | 6.160 ± 0.101 | 4.101 ± 0.321 | 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Intense Longing. 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. (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.495 · CE 6.10 · Genuine 4.36 · WER 0.0% | Target 0.269 · CE 6.05 · Genuine 4.21 · WER 0.0% | Target 0.441 · CE 6.18 · Genuine 4.47 · WER 0.0% | Target 0.564 · CE 6.04 · Genuine 3.93 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Intense Longing. 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. (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.410 · CE 6.07 · Genuine 4.47 · WER 0.0% | Target 0.485 · CE 6.06 · Genuine 4.16 · WER 0.0% | Target 0.324 · CE 5.97 · Genuine 3.99 · WER 0.0% | Target 0.284 · CE 6.12 · Genuine 3.80 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Intense Longing. 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. (Intense Longing. 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 0.623 · CE 6.05 · Genuine 4.41 · WER 0.0% | Target 0.434 · CE 6.15 · Genuine 4.44 · WER 0.0% | Target 0.440 · CE 6.25 · Genuine 4.51 · WER 0.0% | Target 0.287 · CE 6.26 · Genuine 4.14 · 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. Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Longing. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Intense Longing. 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. (Intense Longing. 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 0.422 · CE 6.15 · Genuine 4.76 · WER 0.0% | Target 0.563 · CE 6.03 · Genuine 4.63 · WER 0.0% | Target 0.469 · CE 5.97 · Genuine 4.66 · WER 0.0% | Target 0.470 · CE 6.22 · Genuine 4.53 · WER 0.0% |