← all dimensions · markdown (agent skill) · manual home
Storytelling Style VoiceNet dimension guide
Speaking Stylecode vn_S_STRYslot 95
This dimension measures how much the voice sounds like storytelling. 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.
Storytelling Style controls how much the voice sounds like storytelling. The __high LoRA shifts it +0.26 at 100% (worst cost: burst-blend -0.27); the __low LoRA shifts it -0.26 at 125% (worst cost: speech quality -0.29). The __high direction is the stronger, more reliable knob here.
▲ HIGH — Storytelling Style __high
Dose-response
| dose | target | Δtarget | Δgenu | Δblend | Δqual |
|---|
| baseline (0%) | 0.647 | — | — | — | — |
| 25% | 0.664 | +0.02 | -0.03 | +0.01 | +0.10 |
| 50% | 0.694 | +0.05 | -0.06 | +0.03 | +0.10 |
| 75% | 0.817 | +0.17 | -0.07 | -0.11 | +0.06 |
| 100% ★ | 0.911 | +0.26 | -0.09 | -0.27 | +0.09 |
| 125% | 0.851 | +0.20 | -0.07 | -0.31 | +0.03 |
EmoNet emotion correlation
Triumph+0.45
Elation+0.42
Amusement+0.34
Relief-0.60
Fatigue/Exhaustion-0.47
Helplessness-0.42
VoiceNet dimension correlation
Focus+0.86
Articulation Clarity+0.84
Metallic Character+0.81
Vulnerability-0.62
Disfluency-0.61
Genuineness-0.53
Listen — 3 samples per dose
25%measured target: 0.61 · 0.56 · 0.74
50%measured target: 0.64 · 0.61 · 0.68
75%measured target: 0.75 · 0.86 · 0.78
100% ★measured target: 0.85 · 1.00 · 1.00
125%measured target: 0.47 · 0.94 · 0.86
Merging Storytelling Style __high ramps the dimension: at the recommended 100% dose the measured Storytelling Style score moves from a 0.65 baseline to 0.91 (a +0.26 shift). The dose-response is clean and monotonic — each step pushes the dimension further. That reach is a real cost — the biggest hit is to burst-blend (-0.27); across the board genuineness -0.09, burst-blend -0.27, speech quality +0.09. Emotionally it reads as Triumph (r 0.45), Elation (0.42) and Amusement (0.34), and it moves away from Relief, Fatigue/Exhaustion and Helplessness. On other voice dimensions it co-moves most with Focus (0.86), Articulation Clarity (0.84) and Metallic Character (0.81), while pulling down Vulnerability, Disfluency and Genuineness. Practical call: merge __high at 75–100%; expect Triumph and Focus to rise and burst-blend to fall. If you don’t want Triumph bleeding in, offset it with a prompt cue toward Relief to keep the storytelling style shift without the side-effect.
▼ LOW — Storytelling Style __low
Dose-response
| dose | target | Δtarget | Δgenu | Δblend | Δqual |
|---|
| baseline (0%) | 0.647 | — | — | — | — |
| 25% | 0.461 | -0.19 | +0.00 | -0.16 | -0.23 |
| 50% | 0.648 | +0.00 | +0.01 | -0.26 | -0.02 |
| 75% | 0.536 | -0.11 | -0.01 | -0.07 | -0.06 |
| 100% | 0.658 | +0.01 | -0.02 | -0.29 | -0.09 |
| 125% ★ | 0.387 | -0.26 | +0.09 | -0.10 | -0.29 |
EmoNet emotion correlation
Interest+0.75
Affection+0.56
Hope/Enthusiasm/Optimism+0.54
Impatience and Irritability-0.58
Pain-0.57
Fatigue/Exhaustion-0.52
VoiceNet dimension correlation
Narrator Style+0.83
Smoothness+0.82
Voice Age+0.73
Disfluency-0.58
Genuineness-0.46
Playful Style-0.41
Listen — 3 samples per dose
25%measured target: 0.42 · 0.67 · 0.74
50%measured target: 0.63 · 0.61 · 0.27
75%measured target: 0.23 · 0.72 · 0.69
100%measured target: 0.72 · 0.32 · 0.70
125% ★measured target: 0.57 · 0.37 · 0.74
Merging Storytelling Style __low dials back the dimension: at the recommended 125% dose the measured Storytelling Style score moves from a 0.65 baseline to 0.39 (a -0.26 shift). The dose-response is mostly ordered but noisy between doses. That reach is a steep cost — the biggest hit is to speech quality (-0.29); across the board genuineness +0.09, burst-blend -0.10, speech quality -0.29. Emotionally it reads as Interest (r 0.75), Affection (0.56) and Hope/Enthusiasm/Optimism (0.54), and it moves away from Impatience and Irritability, Pain and Fatigue/Exhaustion. On other voice dimensions it co-moves most with Narrator Style (0.83), Smoothness (0.82) and Voice Age (0.73), while pulling down Disfluency, Genuineness and Playful Style. Practical call: merge __low at 100–125%; expect Interest and Narrator Style to rise and speech quality to fall. If you don’t want Interest bleeding in, offset it with a prompt cue toward Impatience and Irritability to keep the storytelling style shift without the side-effect.