Very informal/casual, not formal

Standard vs manual prompts: matched LoRA / CFG comparison

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	-0.986	2.704	6.233	4.392	0.602	0.008	1.000	pass
manual	50	-0.375	3.495	6.146	4.373	0.575	0.023	1.000	pass
adapter	50	-0.532	3.307	6.179	4.369	0.590	0.016	0.980	pass
matched_no_adapter	50	-0.442	3.399	6.126	4.391	0.556	0.039	0.960	pass

Incremental adapter gain versus same-CFG control: -0.090 z;95% cluster interval [-0.193347154719618, 0.055709246331492254]. Versus best manual-only: -0.157 z. Manual-only versus standard: 0.611 z.

All32development CFG × strength cells

Scale	CFG	Target z	CE	Genuine	WER	Provisional QA
0.0	1.0	-0.808	6.200	4.399	0.000	True
0.5	1.0	-0.840	6.187	4.339	0.005	False
0.75	1.0	-0.746	6.228	4.390	0.003	False
1.0	1.0	-0.876	6.181	4.323	0.006	True
0.0	1.25	-0.861	6.150	4.358	0.003	True
0.5	1.25	-0.695	6.139	4.337	0.000	True
0.75	1.25	-0.811	6.171	4.345	0.059	True
1.0	1.25	-0.830	6.177	4.383	0.003	True
0.0	1.5	-0.867	6.155	4.317	0.000	True
0.5	1.5	-0.661	6.176	4.403	0.012	False
0.75	1.5	-0.799	6.152	4.338	0.006	True
1.0	1.5	-0.812	6.155	4.327	0.006	False
0.0	2.0	-0.686	6.122	4.338	0.018	False
0.5	2.0	-0.722	6.196	4.317	0.000	True
0.75	2.0	-0.606	6.166	4.389	0.000	False
1.0	2.0	-0.620	6.127	4.339	0.026	False
0.0	2.5	-0.635	6.133	4.325	0.003	True
0.5	2.5	-0.426	6.092	4.347	0.003	True
0.75	2.5	-0.442	6.156	4.382	0.003	True
1.0	2.5	-0.368	6.070	4.403	0.011	True
0.0	3.0	-0.387	6.109	4.416	0.019	True
0.5	3.0	-0.380	6.080	4.326	0.003	True
0.75	3.0	-0.461	6.131	4.343	0.003	True
1.0	3.0	-0.398	6.128	4.365	0.025	True
0.0	4.0	-0.419	6.077	4.331	0.024	True
0.5	4.0	-0.325	6.098	4.323	0.017	True
0.75	4.0	-0.288	6.054	4.402	0.006	True
1.0	4.0	-0.384	6.073	4.287	0.029	True
0.0	5.0	-0.218	6.009	4.348	0.022	True
0.5	5.0	-0.168	6.040	4.259	0.037	False
0.75	5.0	-0.167	6.008	4.280	0.038	False
1.0	5.0	-0.202	6.019	4.290	0.020	False

Selected English acting directions and their provenance

new_canonical_dimension_extension

{
  "id": "canonical-extension-P09",
  "general": "Very informal, loose conversational delivery to a familiar listener, with casual rhythm and spontaneous emphasis. Follow the local performance direction beside each sentence.",
  "inline": [
    "Very informal, loose conversational delivery to a familiar listener, with casual rhythm and spontaneous emphasis. Follow the local performance direction beside each sentence.",
    "Very informal, loose conversational delivery to a familiar listener, with casual rhythm and spontaneous emphasis. Follow the local performance direction beside each sentence.",
    "Very informal, loose conversational delivery to a familiar listener, with casual rhythm and spontaneous emphasis. Follow the local performance direction beside each sentence."
  ],
  "recipe": "P09",
  "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

Condition	n	Target score ↑	Content Enjoyment ↑	Genuineness ↑	WER ↓
Merged base, no bucket adapter	20	0.220 ± 0.129	6.131 ± 0.192	4.398 ± 0.215	0.26% ± 1.18
Rank32 × 0.5	20	0.213 ± 0.166	6.155 ± 0.152	4.402 ± 0.215	0.00% ± 0.00
Rank32 × 0.75	20	0.196 ± 0.129	6.147 ± 0.150	4.343 ± 0.227	0.28% ± 1.24
Rank32 × 1.0	20	0.217 ± 0.144	6.171 ± 0.132	4.401 ± 0.235	0.00% ± 0.00

English

Fixed speaker reference:

Condition	n	Target score ↑	Content Enjoyment ↑	Genuineness ↑	WER ↓
Merged base, no bucket adapter	4	0.211 ± 0.184	6.271 ± 0.107	4.216 ± 0.345	0.00% ± 0.00
Rank32 × 0.5	4	0.292 ± 0.230	6.227 ± 0.115	4.303 ± 0.226	0.00% ± 0.00
Rank32 × 0.75	4	0.251 ± 0.217	6.213 ± 0.091	4.145 ± 0.381	0.00% ± 0.00
Rank32 × 1.0	4	0.247 ± 0.201	6.259 ± 0.055	4.178 ± 0.338	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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.013 · CE 6.13 · Genuine 3.74 · WER 0.0%	Target 0.109 · CE 6.13 · Genuine 4.21 · WER 0.0%	Target 0.087 · CE 6.22 · Genuine 3.99 · WER 0.0%	Target 0.078 · CE 6.19 · Genuine 3.88 · 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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.106 · CE 6.28 · Genuine 4.18 · WER 0.0%	Target 0.078 · CE 6.13 · Genuine 4.03 · WER 0.0%	Target 0.054 · CE 6.09 · Genuine 3.68 · WER 0.0%	Target 0.073 · CE 6.25 · Genuine 3.90 · WER 0.0%
Seed 1701 I found the little box. It was beside the old window. We can look inside together. Exact GENERAL / SCRIPT prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.410 · CE 6.28 · Genuine 4.43 · WER 0.0%	Target 0.476 · CE 6.36 · Genuine 4.51 · WER 0.0%	Target 0.502 · CE 6.30 · Genuine 4.49 · WER 0.0%	Target 0.374 · CE 6.32 · Genuine 4.53 · WER 0.0%
Seed 1702 I found the little box. It was beside the old window. We can look inside together. Exact GENERAL / SCRIPT prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.317 · CE 6.40 · Genuine 4.51 · WER 0.0%	Target 0.506 · CE 6.29 · Genuine 4.47 · WER 0.0%	Target 0.361 · CE 6.25 · Genuine 4.41 · WER 0.0%	Target 0.462 · CE 6.28 · Genuine 4.40 · WER 0.0%

German

Fixed speaker reference:

Condition	n	Target score ↑	Content Enjoyment ↑	Genuineness ↑	WER ↓
Merged base, no bucket adapter	4	0.073 ± 0.029	6.245 ± 0.128	4.308 ± 0.218	0.00% ± 0.00
Rank32 × 0.5	4	0.071 ± 0.030	6.320 ± 0.116	4.382 ± 0.083	0.00% ± 0.00
Rank32 × 0.75	4	0.052 ± 0.020	6.267 ± 0.222	4.327 ± 0.066	0.00% ± 0.00
Rank32 × 1.0	4	0.100 ± 0.036	6.307 ± 0.132	4.449 ± 0.163	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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.096 · CE 6.15 · Genuine 4.53 · WER 0.0%	Target 0.097 · CE 6.21 · Genuine 4.44 · WER 0.0%	Target 0.069 · CE 6.28 · Genuine 4.39 · WER 0.0%	Target 0.144 · CE 6.21 · Genuine 4.56 · 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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.100 · CE 6.13 · Genuine 4.46 · WER 0.0%	Target 0.029 · CE 6.24 · Genuine 4.44 · WER 0.0%	Target 0.055 · CE 5.95 · Genuine 4.35 · WER 0.0%	Target 0.107 · CE 6.20 · Genuine 4.58 · WER 0.0%
Seed 1701 Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen. Exact GENERAL / SCRIPT prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.	Target 0.043 · CE 6.32 · Genuine 4.11 · WER 0.0%	Target 0.074 · CE 6.46 · Genuine 4.27 · WER 0.0%	Target 0.023 · CE 6.46 · Genuine 4.33 · WER 0.0%	Target 0.057 · CE 6.48 · 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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen.	Target 0.053 · CE 6.39 · Genuine 4.13 · WER 0.0%	Target 0.083 · CE 6.37 · Genuine 4.37 · WER 0.0%	Target 0.062 · CE 6.37 · Genuine 4.24 · WER 0.0%	Target 0.092 · CE 6.33 · Genuine 4.22 · WER 0.0%

French

Fixed speaker reference:

Condition	n	Target score ↑	Content Enjoyment ↑	Genuineness ↑	WER ↓
Merged base, no bucket adapter	4	0.233 ± 0.058	6.106 ± 0.095	4.529 ± 0.053	0.00% ± 0.00
Rank32 × 0.5	4	0.207 ± 0.098	6.116 ± 0.069	4.479 ± 0.283	0.00% ± 0.00
Rank32 × 0.75	4	0.238 ± 0.043	6.185 ± 0.072	4.482 ± 0.142	1.39% ± 2.78
Rank32 × 1.0	4	0.266 ± 0.184	6.120 ± 0.110	4.481 ± 0.285	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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.151 · CE 6.01 · Genuine 4.46 · WER 0.0%	Target 0.220 · CE 6.12 · Genuine 4.48 · WER 0.0%	Target 0.197 · CE 6.27 · Genuine 4.60 · WER 5.6%	Target 0.144 · CE 6.05 · Genuine 4.51 · 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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.231 · CE 6.23 · Genuine 4.58 · WER 0.0%	Target 0.099 · CE 6.06 · Genuine 4.12 · WER 0.0%	Target 0.263 · CE 6.16 · Genuine 4.41 · WER 0.0%	Target 0.090 · CE 6.17 · Genuine 4.07 · 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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.266 · CE 6.06 · Genuine 4.53 · WER 0.0%	Target 0.177 · CE 6.21 · Genuine 4.51 · WER 0.0%	Target 0.206 · CE 6.21 · Genuine 4.32 · WER 0.0%	Target 0.487 · CE 6.01 · Genuine 4.72 · 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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.282 · CE 6.13 · Genuine 4.55 · WER 0.0%	Target 0.334 · CE 6.07 · Genuine 4.81 · WER 0.0%	Target 0.284 · CE 6.10 · Genuine 4.61 · WER 0.0%	Target 0.345 · CE 6.25 · Genuine 4.62 · WER 0.0%

Dutch

Fixed speaker reference:

Condition	n	Target score ↑	Content Enjoyment ↑	Genuineness ↑	WER ↓
Merged base, no bucket adapter	4	0.350 ± 0.092	6.146 ± 0.047	4.426 ± 0.105	1.32% ± 2.63
Rank32 × 0.5	4	0.283 ± 0.230	6.110 ± 0.101	4.463 ± 0.130	0.00% ± 0.00
Rank32 × 0.75	4	0.271 ± 0.090	6.062 ± 0.028	4.355 ± 0.118	0.00% ± 0.00
Rank32 × 1.0	4	0.214 ± 0.112	6.132 ± 0.062	4.372 ± 0.034	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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.344 · CE 6.10 · Genuine 4.48 · WER 5.3%	Target 0.251 · CE 6.01 · Genuine 4.31 · WER 0.0%	Target 0.318 · CE 6.05 · Genuine 4.34 · WER 0.0%	Target 0.231 · CE 6.13 · Genuine 4.37 · 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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.234 · CE 6.13 · Genuine 4.33 · WER 0.0%	Target 0.127 · CE 6.19 · Genuine 4.58 · WER 0.0%	Target 0.142 · CE 6.04 · Genuine 4.19 · WER 0.0%	Target 0.366 · CE 6.09 · Genuine 4.42 · 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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.366 · CE 6.21 · Genuine 4.34 · WER 0.0%	Target 0.135 · CE 6.21 · Genuine 4.41 · WER 0.0%	Target 0.347 · CE 6.10 · Genuine 4.43 · WER 0.0%	Target 0.129 · CE 6.09 · Genuine 4.36 · 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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.458 · CE 6.13 · Genuine 4.55 · WER 0.0%	Target 0.617 · CE 6.03 · Genuine 4.56 · WER 0.0%	Target 0.276 · CE 6.05 · Genuine 4.46 · WER 0.0%	Target 0.130 · CE 6.22 · Genuine 4.34 · WER 0.0%

Spanish

Fixed speaker reference:

Condition	n	Target score ↑	Content Enjoyment ↑	Genuineness ↑	WER ↓
Merged base, no bucket adapter	4	0.233 ± 0.086	5.889 ± 0.267	4.509 ± 0.127	0.00% ± 0.00
Rank32 × 0.5	4	0.212 ± 0.124	6.003 ± 0.159	4.383 ± 0.333	0.00% ± 0.00
Rank32 × 0.75	4	0.170 ± 0.077	6.009 ± 0.128	4.407 ± 0.240	0.00% ± 0.00
Rank32 × 1.0	4	0.257 ± 0.139	6.038 ± 0.102	4.525 ± 0.165	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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.	Target 0.332 · CE 6.09 · Genuine 4.63 · WER 0.0%	Target 0.275 · CE 5.86 · Genuine 4.57 · WER 0.0%	Target 0.178 · CE 6.05 · Genuine 4.33 · WER 0.0%	Target 0.239 · CE 6.05 · Genuine 4.43 · WER 0.0%
Seed 1702 La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo. Exact GENERAL / SCRIPT prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo.	Target 0.266 · CE 5.53 · Genuine 4.49 · WER 0.0%	Target 0.047 · CE 6.10 · Genuine 3.89 · WER 0.0%	Target 0.067 · CE 6.16 · Genuine 4.12 · WER 0.0%	Target 0.075 · CE 5.90 · Genuine 4.34 · 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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.203 · CE 5.83 · Genuine 4.57 · WER 0.0%	Target 0.195 · CE 6.17 · Genuine 4.57 · WER 0.0%	Target 0.253 · CE 5.97 · Genuine 4.68 · WER 0.0%	Target 0.306 · CE 6.05 · Genuine 4.68 · 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 prompt GENERAL: One adult speaker in a clean recording. Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Very informal/casual, not formal. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Very informal/casual, not formal. 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. (Very informal/casual, not formal. 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 0.132 · CE 6.09 · Genuine 4.34 · WER 0.0%	Target 0.332 · CE 5.87 · Genuine 4.51 · WER 0.0%	Target 0.182 · CE 5.86 · Genuine 4.49 · WER 0.0%	Target 0.407 · CE 6.15 · Genuine 4.65 · WER 0.0%

Very informal/casual, not formal · rank32 comparison

Humaneness Voice Large 1.5 and 1.6 Alpha — model and manual navigation

Standard vs manual prompts: matched LoRA / CFG comparison

What is compared?

English

German

French

Dutch

Spanish