← all dimensions · markdown (agent skill) · manual home
Teacher/Didactic Style VoiceNet dimension guide
Speaking Stylecode vn_S_TECHslot 88
This dimension measures how much the voice sounds like a teacher explaining. 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.
Teacher/Didactic Style controls how much the voice sounds like a teacher explaining. The __high LoRA shifts it +0.25 at 125% (worst cost: burst-blend -0.30); the __low LoRA shifts it -0.08 at 25% (worst cost: speech quality -0.07). The __high direction is the stronger, more reliable knob here.
▲ HIGH — Teacher/Didactic Style __high
Dose-response
| dose | target | Δtarget | Δgenu | Δblend | Δqual |
|---|
| baseline (0%) | 0.214 | — | — | — | — |
| 25% | 0.231 | +0.02 | +0.00 | -0.11 | -0.01 |
| 50% | 0.359 | +0.15 | -0.06 | -0.09 | +0.09 |
| 75% | 0.367 | +0.15 | -0.09 | -0.23 | +0.11 |
| 100% | 0.382 | +0.17 | -0.05 | -0.25 | +0.08 |
| 125% ★ | 0.465 | +0.25 | +0.03 | -0.30 | +0.05 |
EmoNet emotion correlation
Concentration+0.62
Emotional Numbness+0.58
Contemplation+0.51
Fear-0.54
Elation-0.39
Distress-0.35
VoiceNet dimension correlation
Valence+0.81
ASMR Style+0.81
Velocity Flux+0.79
Dramatic Style-0.70
Perceived Gender-0.66
Cognitive Load-0.58
Listen — 3 samples per dose
25%measured target: 0.00 · 0.20 · 0.33
50%measured target: 0.33 · 0.25 · 0.33
75%measured target: 0.31 · 0.25 · 0.42
100%measured target: 0.14 · 0.32 · 0.02
125% ★measured target: 0.58 · 0.38 · 0.57
Merging Teacher/Didactic Style __high ramps the dimension: at the recommended 125% dose the measured Teacher/Didactic Style score moves from a 0.21 baseline to 0.47 (a +0.25 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 +0.03, burst-blend -0.30, speech quality +0.05. Emotionally it reads as Concentration (r 0.62), Emotional Numbness (0.58) and Contemplation (0.51), and it moves away from Fear, Elation and Distress. On other voice dimensions it co-moves most with Valence (0.81), ASMR Style (0.81) and Velocity Flux (0.79), while pulling down Dramatic Style, Perceived Gender and Cognitive Load. Practical call: merge __high at 100–125%; expect Concentration and Valence to rise and burst-blend to fall. If you don’t want Concentration bleeding in, offset it with a prompt cue toward Fear to keep the teacher/didactic style shift without the side-effect.
▼ LOW — Teacher/Didactic Style __low
Dose-response
| dose | target | Δtarget | Δgenu | Δblend | Δqual |
|---|
| baseline (0%) | 0.214 | — | — | — | — |
| 25% ★ | 0.138 | -0.08 | +0.00 | +0.03 | -0.07 |
| 50% | 0.283 | +0.07 | -0.02 | +0.03 | +0.06 |
| 75% | 0.251 | +0.04 | -0.04 | -0.21 | +0.06 |
| 100% | 0.174 | -0.04 | -0.06 | -0.21 | -0.02 |
| 125% | 0.222 | +0.01 | -0.05 | -0.18 | +0.03 |
EmoNet emotion correlation
Contemplation+0.52
Concentration+0.50
Emotional Numbness+0.47
Fear-0.39
Elation-0.31
Fatigue/Exhaustion-0.30
VoiceNet dimension correlation
ASMR Style+0.77
Recording Quality+0.77
Esthetics+0.68
Volatility-0.55
Disfluency-0.46
Tension-0.42
Listen — 3 samples per dose
25% ★measured target: 0.10 · 0.03 · 0.33
50%measured target: 0.43 · 0.27 · 0.34
75%measured target: 0.16 · 0.14 · 0.23
100%measured target: 0.37 · 0.27 · 0.06
125%measured target: 0.33 · 0.24 · 0.36
Merging Teacher/Didactic Style __low pulls down the dimension: at the recommended 25% dose the measured Teacher/Didactic Style score moves from a 0.21 baseline to 0.14 (a -0.08 shift). The dose-response is mostly ordered but noisy between doses. That reach is a modest cost — the biggest hit is to speech quality (-0.07); across the board genuineness +0.00, burst-blend +0.03, speech quality -0.07. Emotionally it reads as Contemplation (r 0.52), Concentration (0.50) and Emotional Numbness (0.47), and it moves away from Fear, Elation and Fatigue/Exhaustion. On other voice dimensions it co-moves most with ASMR Style (0.77), Recording Quality (0.77) and Esthetics (0.68), while pulling down Volatility, Disfluency and Tension. Practical call: merge __low at a light 25–50%; expect Contemplation and ASMR Style to rise and speech quality to fall. If you don’t want Contemplation bleeding in, offset it with a prompt cue toward Fear to keep the teacher/didactic style shift without the side-effect.