← all dimensions · markdown (agent skill) · manual home
Articulation Clarity VoiceNet dimension guide
Rhythm & Timingcode vn_CLRTslot 43
This dimension measures the precision with which consonants and vowels are produced. 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.
Articulation Clarity controls the precision with which consonants and vowels are produced. The __high LoRA shifts it +0.34 at 125% (worst cost: burst-blend -0.25); the __low LoRA shifts it -0.19 at 75% (worst cost: speech quality -0.37). The __high direction is the stronger, more reliable knob here.
▲ HIGH — Articulation Clarity __high
Dose-response
| dose | target | Δtarget | Δgenu | Δblend | Δqual |
|---|
| baseline (0%) | 0.207 | — | — | — | — |
| 25% | 0.249 | +0.04 | -0.03 | +0.02 | +0.08 |
| 50% | 0.355 | +0.15 | -0.04 | -0.06 | +0.06 |
| 75% | 0.454 | +0.25 | -0.11 | -0.22 | +0.12 |
| 100% | 0.488 | +0.28 | -0.10 | -0.26 | +0.11 |
| 125% ★ | 0.552 | +0.34 | -0.11 | -0.25 | +0.13 |
EmoNet emotion correlation
Triumph+0.30
Pride+0.25
Malevolence/Malice+0.21
Relief-0.57
Fatigue/Exhaustion-0.56
Helplessness-0.49
VoiceNet dimension correlation
Attack+0.83
Harmonicity+0.82
Metallic Character+0.80
Vulnerability-0.83
Disfluency-0.79
Casual Style-0.75
Listen — 3 samples per dose
25%measured target: 0.03 · 0.16 · 0.24
50%measured target: 0.55 · 0.16 · 0.51
75%measured target: 0.47 · 0.42 · 0.49
100%measured target: 0.41 · 0.57 · 0.24
125% ★measured target: 0.49 · 0.47 · 0.61
Merging Articulation Clarity __high pushes the dimension: at the recommended 125% dose the measured Articulation Clarity score moves from a 0.21 baseline to 0.55 (a +0.34 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 (-0.25); across the board genuineness -0.11, burst-blend -0.25, speech quality +0.13. Emotionally it reads as Triumph (r 0.30), Pride (0.25) and Malevolence/Malice (0.21), and it moves away from Relief, Fatigue/Exhaustion and Helplessness. On other voice dimensions it co-moves most with Attack (0.83), Harmonicity (0.82) and Metallic Character (0.80), while pulling down Vulnerability, Disfluency and Casual Style. Practical call: merge __high at 100–125%; expect Triumph and Attack 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 articulation clarity shift without the side-effect.
▼ LOW — Articulation Clarity __low
Dose-response
| dose | target | Δtarget | Δgenu | Δblend | Δqual |
|---|
| baseline (0%) | 0.207 | — | — | — | — |
| 25% | 0.173 | -0.03 | +0.02 | +0.02 | -0.04 |
| 50% | 0.033 | -0.17 | +0.06 | -0.03 | -0.22 |
| 75% ★ | 0.017 | -0.19 | +0.10 | -0.12 | -0.37 |
| 100% | 0.117 | -0.09 | +0.08 | -0.11 | -0.34 |
| 125% | 0.134 | -0.07 | +0.05 | -0.19 | -0.33 |
EmoNet emotion correlation
Hope/Enthusiasm/Optimism+0.72
Interest+0.61
Affection+0.49
Fatigue/Exhaustion-0.60
Sadness-0.49
Pain-0.46
VoiceNet dimension correlation
Harmonicity+0.74
Emphasis+0.69
Storytelling Style+0.68
Genuineness-0.67
Disfluency-0.60
Vulnerability-0.50
Listen — 3 samples per dose
25%measured target: 0.18 · 0.10 · 0.28
50%measured target: 0.04 · 0.00 · 0.00
75% ★measured target: 0.04 · 0.00 · 0.02
100%measured target: 0.33 · 0.05 · 0.29
125%measured target: 0.35 · 0.12 · 0.00
Merging Articulation Clarity __low damps the dimension: at the recommended 75% dose the measured Articulation Clarity score moves from a 0.21 baseline to 0.02 (a -0.19 shift). The dose-response is mostly ordered but noisy between doses. That reach is a steep cost — the biggest hit is to speech quality (-0.37); across the board genuineness +0.10, burst-blend -0.12, speech quality -0.37. Emotionally it reads as Hope/Enthusiasm/Optimism (r 0.72), Interest (0.61) and Affection (0.49), and it moves away from Fatigue/Exhaustion, Sadness and Pain. On other voice dimensions it co-moves most with Harmonicity (0.74), Emphasis (0.69) and Storytelling Style (0.68), while pulling down Genuineness, Disfluency and Vulnerability. Practical call: merge __low at 50–100%; expect Hope/Enthusiasm/Optimism and Harmonicity to rise and speech quality to fall. If you don’t want Hope/Enthusiasm/Optimism bleeding in, offset it with a prompt cue toward Fatigue/Exhaustion to keep the articulation clarity shift without the side-effect.