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 | 1.574 | 3.666 | 6.231 | 4.350 | 0.599 | 0.013 | 0.980 | pass |
| manual | 50 | 1.862 | 4.080 | 6.261 | 4.125 | 0.530 | 0.010 | 1.000 | fail |
| adapter | 50 | 1.887 | 4.116 | 6.238 | 4.102 | 0.574 | 0.007 | 1.000 | pass |
| matched_no_adapter | 50 | 1.862 | 4.080 | 6.261 | 4.125 | 0.530 | 0.010 | 1.000 | fail |
Incremental adapter gain versus same-CFG control: 0.025 z;95% cluster interval [-0.017812472071414794, 0.06749998754115158]. Versus best manual-only: 0.025 z. Manual-only versus standard: 0.288 z.
| Scale | CFG | Target z | CE | Genuine | WER | Provisional QA |
|---|---|---|---|---|---|---|
| 0.0 | 1.0 | 1.789 | 6.180 | 4.205 | 0.014 | True |
| 0.5 | 1.0 | 1.829 | 6.172 | 4.137 | 0.000 | True |
| 0.75 | 1.0 | 1.811 | 6.232 | 4.216 | 0.003 | True |
| 1.0 | 1.0 | 1.841 | 6.209 | 4.149 | 0.003 | True |
| 0.0 | 1.25 | 1.805 | 6.159 | 4.192 | 0.000 | True |
| 0.5 | 1.25 | 1.713 | 6.217 | 4.179 | 0.006 | True |
| 0.75 | 1.25 | 1.798 | 6.192 | 4.079 | 0.003 | True |
| 1.0 | 1.25 | 1.876 | 6.185 | 4.117 | 0.006 | True |
| 0.0 | 1.5 | 1.731 | 6.150 | 4.097 | 0.003 | False |
| 0.5 | 1.5 | 1.742 | 6.169 | 4.087 | 0.000 | True |
| 0.75 | 1.5 | 1.821 | 6.157 | 4.101 | 0.003 | True |
| 1.0 | 1.5 | 1.835 | 6.187 | 4.145 | 0.000 | True |
| 0.0 | 2.0 | 1.840 | 6.203 | 4.076 | 0.000 | True |
| 0.5 | 2.0 | 1.858 | 6.168 | 4.087 | 0.003 | True |
| 0.75 | 2.0 | 1.781 | 6.132 | 4.049 | 0.000 | False |
| 1.0 | 2.0 | 1.905 | 6.156 | 4.019 | 0.000 | True |
| 0.0 | 2.5 | 1.759 | 6.117 | 4.123 | 0.003 | False |
| 0.5 | 2.5 | 1.817 | 6.135 | 4.072 | 0.003 | False |
| 0.75 | 2.5 | 1.833 | 6.157 | 4.003 | 0.003 | False |
| 1.0 | 2.5 | 1.843 | 6.135 | 4.011 | 0.003 | False |
| 0.0 | 3.0 | 1.685 | 6.099 | 4.012 | 0.010 | False |
| 0.5 | 3.0 | 1.751 | 6.124 | 4.003 | 0.003 | False |
| 0.75 | 3.0 | 1.793 | 6.108 | 3.972 | 0.024 | False |
| 1.0 | 3.0 | 1.759 | 6.084 | 3.998 | 0.003 | False |
| 0.0 | 4.0 | 1.632 | 6.071 | 3.989 | 0.012 | False |
| 0.5 | 4.0 | 1.676 | 6.015 | 3.982 | 0.020 | False |
| 0.75 | 4.0 | 1.764 | 6.090 | 3.924 | 0.020 | False |
| 1.0 | 4.0 | 1.635 | 6.025 | 3.952 | 0.008 | False |
| 0.0 | 5.0 | 1.666 | 5.969 | 3.888 | 0.018 | False |
| 0.5 | 5.0 | 1.654 | 6.051 | 3.935 | 0.006 | False |
| 0.75 | 5.0 | 1.587 | 6.002 | 3.884 | 0.018 | False |
| 1.0 | 5.0 | 1.582 | 6.027 | 3.885 | 0.024 | False |
direct_manual_recipe
{
"id": "manual-9ea232963c9d",
"general": "Focus on precise articulation, ensuring every consonant is cleanly formed while maintaining a relaxed jaw posture.",
"inline": [
"Deliver the opening segment with crisp, clear articulation, emphasizing the initial consonant sounds.",
"Mid-section demands meticulous enunciation of all mid-range vowels and connecting sounds.",
"End with sharp, defined final consonants, giving the narration a clean, authoritative stop."
],
"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
Strong Narrator 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": 40, "es": 65, "nl": 58, "de": 62, "en": 43}. 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 | 0.941 ± 0.039 | 6.209 ± 0.095 | 4.301 ± 0.176 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 20 | 0.940 ± 0.046 | 6.177 ± 0.167 | 4.309 ± 0.158 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 20 | 0.935 ± 0.039 | 6.159 ± 0.128 | 4.259 ± 0.255 | 0.31% ± 1.40 |
| Rank32 × 1.0 | 20 | 0.934 ± 0.061 | 6.193 ± 0.153 | 4.303 ± 0.181 | 0.28% ± 1.24 |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.901 ± 0.039 | 6.235 ± 0.069 | 4.184 ± 0.106 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.932 ± 0.034 | 6.348 ± 0.147 | 4.205 ± 0.269 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.926 ± 0.012 | 6.205 ± 0.115 | 4.187 ± 0.343 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.874 ± 0.097 | 6.324 ± 0.079 | 4.366 ± 0.153 | 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.912 · CE 6.28 · Genuine 4.04 · WER 0.0% | Target 0.931 · CE 6.27 · Genuine 3.90 · WER 0.0% | Target 0.912 · CE 6.13 · Genuine 3.87 · WER 0.0% | Target 0.737 · CE 6.26 · Genuine 4.39 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.845 · CE 6.27 · Genuine 4.30 · WER 0.0% | Target 0.887 · CE 6.20 · Genuine 4.05 · WER 0.0% | Target 0.927 · CE 6.09 · Genuine 3.92 · WER 0.0% | Target 0.875 · CE 6.28 · Genuine 4.22 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.910 · CE 6.25 · Genuine 4.20 · WER 0.0% | Target 0.945 · CE 6.53 · Genuine 4.42 · WER 0.0% | Target 0.924 · CE 6.34 · Genuine 4.52 · WER 0.0% | Target 0.947 · CE 6.44 · Genuine 4.57 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.935 · CE 6.13 · Genuine 4.19 · WER 0.0% | Target 0.967 · CE 6.40 · Genuine 4.44 · WER 0.0% | Target 0.940 · CE 6.26 · Genuine 4.44 · WER 0.0% | Target 0.939 · CE 6.32 · Genuine 4.29 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.955 ± 0.034 | 6.309 ± 0.074 | 4.299 ± 0.137 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.961 ± 0.027 | 6.281 ± 0.152 | 4.390 ± 0.112 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.950 ± 0.043 | 6.293 ± 0.096 | 4.236 ± 0.146 | 1.56% ± 3.12 |
| Rank32 × 1.0 | 4 | 0.952 ± 0.043 | 6.315 ± 0.093 | 4.281 ± 0.186 | 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.943 · CE 6.33 · Genuine 4.34 · WER 0.0% | Target 0.945 · CE 6.17 · Genuine 4.46 · WER 0.0% | Target 0.889 · CE 6.25 · Genuine 4.39 · WER 0.0% | Target 0.958 · CE 6.34 · Genuine 4.35 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.913 · CE 6.21 · Genuine 4.46 · WER 0.0% | Target 0.931 · CE 6.13 · Genuine 4.24 · WER 0.0% | Target 0.957 · CE 6.23 · Genuine 4.31 · WER 0.0% | Target 0.889 · CE 6.19 · Genuine 4.47 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.984 · CE 6.38 · Genuine 4.26 · WER 0.0% | Target 0.981 · CE 6.45 · Genuine 4.49 · WER 0.0% | Target 0.983 · CE 6.44 · Genuine 4.17 · WER 6.2% | Target 0.975 · CE 6.32 · Genuine 4.03 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.980 · CE 6.32 · Genuine 4.13 · WER 0.0% | Target 0.986 · CE 6.36 · Genuine 4.36 · WER 0.0% | Target 0.972 · CE 6.25 · Genuine 4.07 · WER 0.0% | Target 0.985 · CE 6.41 · Genuine 4.28 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.939 ± 0.031 | 6.122 ± 0.105 | 4.463 ± 0.134 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.946 ± 0.011 | 6.096 ± 0.104 | 4.413 ± 0.032 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.935 ± 0.043 | 6.170 ± 0.122 | 4.462 ± 0.227 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.956 ± 0.009 | 6.110 ± 0.070 | 4.267 ± 0.184 | 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.908 · CE 6.05 · Genuine 4.36 · WER 0.0% | Target 0.945 · CE 6.15 · Genuine 4.37 · WER 0.0% | Target 0.954 · CE 6.11 · Genuine 4.19 · WER 0.0% | Target 0.946 · CE 6.05 · Genuine 4.21 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.941 · CE 6.09 · Genuine 4.35 · WER 0.0% | Target 0.941 · CE 5.97 · Genuine 4.41 · WER 0.0% | Target 0.938 · CE 6.08 · Genuine 4.37 · WER 0.0% | Target 0.957 · CE 6.05 · Genuine 4.05 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.982 · CE 6.28 · Genuine 4.51 · WER 0.0% | Target 0.937 · CE 6.21 · Genuine 4.43 · WER 0.0% | Target 0.875 · CE 6.14 · Genuine 4.61 · WER 0.0% | Target 0.969 · CE 6.16 · Genuine 4.33 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.927 · CE 6.07 · Genuine 4.63 · WER 0.0% | Target 0.962 · CE 6.06 · Genuine 4.44 · WER 0.0% | Target 0.973 · CE 6.35 · Genuine 4.68 · WER 0.0% | Target 0.953 · CE 6.18 · Genuine 4.48 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.927 ± 0.033 | 6.183 ± 0.062 | 4.197 ± 0.225 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.892 ± 0.078 | 6.084 ± 0.133 | 4.247 ± 0.070 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.890 ± 0.030 | 6.118 ± 0.031 | 4.162 ± 0.063 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.909 ± 0.053 | 6.132 ± 0.155 | 4.333 ± 0.168 | 1.39% ± 2.78 |
| 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.910 · CE 6.15 · Genuine 3.99 · WER 0.0% | Target 0.946 · CE 6.11 · Genuine 4.14 · WER 0.0% | Target 0.914 · CE 6.16 · Genuine 4.22 · WER 0.0% | Target 0.905 · CE 6.17 · Genuine 4.16 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.941 · CE 6.15 · Genuine 4.27 · WER 0.0% | Target 0.856 · CE 6.04 · Genuine 4.29 · WER 0.0% | Target 0.868 · CE 6.10 · Genuine 4.07 · WER 0.0% | Target 0.915 · CE 5.92 · Genuine 4.22 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.891 · CE 6.28 · Genuine 4.48 · WER 0.0% | Target 0.966 · CE 6.25 · Genuine 4.27 · WER 0.0% | Target 0.917 · CE 6.13 · Genuine 4.18 · WER 0.0% | Target 0.843 · CE 6.15 · Genuine 4.51 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.965 · CE 6.16 · Genuine 4.04 · WER 0.0% | Target 0.800 · CE 5.94 · Genuine 4.28 · WER 0.0% | Target 0.860 · CE 6.09 · Genuine 4.17 · WER 0.0% | Target 0.972 · CE 6.29 · Genuine 4.44 · WER 5.6% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.982 ± 0.004 | 6.195 ± 0.084 | 4.360 ± 0.160 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.971 ± 0.009 | 6.074 ± 0.127 | 4.292 ± 0.154 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.973 ± 0.010 | 6.011 ± 0.080 | 4.250 ± 0.377 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.980 ± 0.009 | 6.082 ± 0.174 | 4.266 ± 0.279 | 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.976 · CE 6.26 · Genuine 4.24 · WER 0.0% | Target 0.981 · CE 5.99 · Genuine 4.10 · WER 0.0% | Target 0.986 · CE 6.01 · Genuine 4.13 · WER 0.0% | Target 0.986 · CE 5.97 · Genuine 4.04 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.984 · CE 6.09 · Genuine 4.21 · WER 0.0% | Target 0.964 · CE 5.94 · Genuine 4.27 · WER 0.0% | Target 0.972 · CE 5.90 · Genuine 3.80 · WER 0.0% | Target 0.971 · CE 5.92 · Genuine 4.14 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.984 · CE 6.26 · Genuine 4.46 · WER 0.0% | Target 0.976 · CE 6.16 · Genuine 4.47 · WER 0.0% | Target 0.961 · CE 6.04 · Genuine 4.70 · WER 0.0% | Target 0.973 · CE 6.30 · Genuine 4.67 · 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 Narrator Style. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Strong Narrator 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 Narrator 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 Narrator 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 0.982 · CE 6.17 · Genuine 4.53 · WER 0.0% | Target 0.964 · CE 6.21 · Genuine 4.33 · WER 0.0% | Target 0.973 · CE 6.09 · Genuine 4.37 · WER 0.0% | Target 0.990 · CE 6.14 · Genuine 4.21 · WER 0.0% |