← all dimensions · markdown (agent skill) · manual home
Perceived Gender VoiceNet dimension guide
Speaker Identitycode vn_GENDslot 51
This dimension measures the perceived gender of the voice (pitch × resonance). Below: the merge-scaling dose-response for the __high and __low LoRA (laion/moss-voicenet-dimension-loras) at 25/50/75/100/125%, the measured target value at each dose, the top-3 ± correlated EmoNet emotions and VoiceNet dimensions, three audio samples per dose, and a grounded prompting tip. Costs are shifts vs. the 24-sample neutral baseline in genuineness, vocal-burst blend and speech quality.
Perceived Gender controls the perceived gender of the voice (pitch × resonance). The __high LoRA shifts it +0.14 at 125% (worst cost: genuineness -0.07); the __low LoRA shifts it -0.51 at 125% (worst cost: burst-blend -0.30). The __low direction is the stronger, more reliable knob here.
▲ HIGH — Perceived Gender __high
Dose-response
| dose | target | Δtarget | Δgenu | Δblend | Δqual |
|---|
| baseline (0%) | 0.616 | — | — | — | — |
| 25% | 0.652 | +0.04 | -0.05 | -0.03 | +0.01 |
| 50% | 0.741 | +0.13 | -0.04 | -0.09 | -0.01 |
| 75% | 0.748 | +0.13 | -0.06 | -0.13 | +0.02 |
| 100% | 0.703 | +0.09 | -0.06 | -0.21 | -0.08 |
| 125% ★ | 0.755 | +0.14 | -0.07 | -0.07 | +0.13 |
EmoNet emotion correlation
Sexual Lust+0.29
Longing+0.28
Awe+0.20
Contentment-0.37
Embarrassment-0.22
Hope/Enthusiasm/Optimism-0.19
VoiceNet dimension correlation
Chest Resonance+0.74
Roughness+0.53
Cartoonish Style+0.44
Register-0.86
Mask Resonance-0.53
Nasal Resonance-0.43
Listen — 3 samples per dose
25%measured target: 0.62 · 0.79 · 0.28
50%measured target: 0.70 · 0.89 · 0.72
75%measured target: 0.81 · 0.80 · 0.72
100%measured target: 0.57 · 0.78 · 0.44
125% ★measured target: 0.63 · 0.79 · 0.73
Merging Perceived Gender __high pushes the dimension: at the recommended 125% dose the measured Perceived Gender score moves from a 0.62 baseline to 0.76 (a +0.14 shift). The dose-response is clean and monotonic — each step pushes the dimension further. That reach is a modest cost — the biggest hit is to genuineness (-0.07); across the board genuineness -0.07, burst-blend -0.07, speech quality +0.13. Emotionally it reads as Sexual Lust (r 0.29), Longing (0.28) and Awe (0.20), and it moves away from Contentment, Embarrassment and Hope/Enthusiasm/Optimism. On other voice dimensions it co-moves most with Chest Resonance (0.74), Roughness (0.53) and Cartoonish Style (0.44), while pulling down Register, Mask Resonance and Nasal Resonance. Practical call: merge __high at 100–125%; expect Sexual Lust and Chest Resonance to rise and genuineness to fall. If you don’t want Sexual Lust bleeding in, offset it with a prompt cue toward Contentment to keep the perceived gender shift without the side-effect.
▼ LOW — Perceived Gender __low
Dose-response
| dose | target | Δtarget | Δgenu | Δblend | Δqual |
|---|
| baseline (0%) | 0.616 | — | — | — | — |
| 25% | 0.340 | -0.28 | -0.03 | -0.08 | -0.03 |
| 50% | 0.323 | -0.29 | -0.06 | -0.13 | +0.09 |
| 75% | 0.132 | -0.48 | -0.10 | -0.32 | +0.11 |
| 100% | 0.127 | -0.49 | -0.11 | -0.34 | +0.08 |
| 125% ★ | 0.103 | -0.51 | -0.11 | -0.30 | +0.09 |
EmoNet emotion correlation
Longing+0.57
Infatuation+0.49
Relief+0.43
Elation-0.27
Impatience and Irritability-0.26
Hope/Enthusiasm/Optimism-0.23
VoiceNet dimension correlation
Cognitive Load+0.83
Chest Resonance+0.78
Fullness+0.66
Register-0.94
Mask Resonance-0.78
Brightness-0.77
Listen — 3 samples per dose
25%measured target: 0.37 · 0.55 · 0.25
50%measured target: 0.79 · 0.53 · 0.23
75%measured target: 0.10 · 0.08 · 0.20
100%measured target: 0.23 · 0.20 · 0.10
125% ★measured target: 0.14 · 0.10 · 0.20
Merging Perceived Gender __low damps the dimension: at the recommended 125% dose the measured Perceived Gender score moves from a 0.62 baseline to 0.10 (a -0.51 shift). The dose-response is clean and monotonic — each step pushes the dimension further. That reach is a steep cost — the biggest hit is to burst-blend (-0.30); across the board genuineness -0.11, burst-blend -0.30, speech quality +0.09. Emotionally it reads as Longing (r 0.57), Infatuation (0.49) and Relief (0.43), and it moves away from Elation, Impatience and Irritability and Hope/Enthusiasm/Optimism. On other voice dimensions it co-moves most with Cognitive Load (0.83), Chest Resonance (0.78) and Fullness (0.66), while pulling down Register, Mask Resonance and Brightness. Practical call: merge __low at 100–125%; expect Longing and Cognitive Load to rise and burst-blend to fall. If you don’t want Longing bleeding in, offset it with a prompt cue toward Elation to keep the perceived gender shift without the side-effect.