← all emotions
Astonishment Surprise
Best adapter: rank 16 , checkpoint seen 5000 · EmoNet-z -0.44 (baseline -0.57) · blend 0.419 · genu 0.183 · speaker-sim 0.187.
Instruction prompt: GENERAL: A voice that is adult, with extremely carrying startlement and genuine in delivery.
SCRIPT:
Five neutral sentences, each in English / Deutsch / Français, no reference vs with emotion reference (a same-emotion training clip supplies the voice; the prompt applies the emotion).
Each cell = 5 seeds. We show mean±std of the emotion score and play the highest and lowest seed (top = strongest emotion, bottom = weakest), since a fixed seed pins the sampling and single clips vary.
Sentence 1
no reference
z mean +0.01 ±1.14 · bl 0.1 · gn 0.2 ▲ high +2.28
▼ low -0.57
z mean -0.14 ±0.77 · bl 0.3 · gn 0.1 ▲ high +1.40
▼ low -0.56
z mean -0.29 ±0.38 · bl 0.5 · gn 0.1 ▲ high +0.43
▼ low -0.56
with emo reference
z mean -0.56 ±0.00 · bl 0.1 · gn 0.3 ▲ high -0.56
▼ low -0.57
z mean -0.26 ±0.26 · bl 0.1 · gn 0.2 ▲ high +0.06
▼ low -0.57
z mean -0.57 ±0.00 · bl 0.1 · gn 0.5 ▲ high -0.57
▼ low -0.57
Sentence 2
no reference
z mean -0.57 ±0.00 · bl 0.6 · gn 0.1 ▲ high -0.57
▼ low -0.57
z mean -0.18 ±0.47 · bl 0.3 · gn 0.1 ▲ high +0.42
▼ low -0.57
z mean -0.02 ±0.41 · bl 0.6 · gn 0.2 ▲ high +0.56
▼ low -0.56
with emo reference
z mean -0.54 ±0.06 · bl 0.5 · gn 0.3 ▲ high -0.41
▼ low -0.57
z mean -0.56 ±0.01 · bl 0.0 · gn 0.3 ▲ high -0.54
▼ low -0.57
z mean -0.29 ±0.49 · bl 0.1 · gn 0.4 ▲ high +0.68
▼ low -0.57
Sentence 3
no reference
z mean +0.06 ±0.80 · bl 0.2 · gn 0.0 ▲ high +1.39
▼ low -0.57
z mean +1.73 ±0.75 · bl 0.4 · gn 0.1 ▲ high +2.76
▼ low +0.73
z mean -0.57 ±0.00 · bl 0.7 · gn 0.2 ▲ high -0.57
▼ low -0.57
with emo reference
z mean -0.57 ±0.00 · bl 0.1 · gn 0.2 ▲ high -0.57
▼ low -0.57
z mean -0.33 ±0.23 · bl 0.0 · gn 0.2 ▲ high -0.02
▼ low -0.57
z mean -0.57 ±0.00 · bl 0.2 · gn 0.4 ▲ high -0.57
▼ low -0.57
Sentence 4
no reference
z mean -0.57 ±0.00 · bl 0.1 · gn 0.1 ▲ high -0.57
▼ low -0.58
z mean -0.55 ±0.04 · bl 0.3 · gn 0.1 ▲ high -0.47
▼ low -0.58
z mean +0.08 ±0.84 · bl 0.3 · gn 0.2 ▲ high +1.56
▼ low -0.57
with emo reference
z mean -0.57 ±0.00 · bl 0.1 · gn 0.2 ▲ high -0.57
▼ low -0.57
z mean -0.57 ±0.00 · bl 0.0 · gn 0.3 ▲ high -0.57
▼ low -0.57
z mean +0.03 ±1.19 · bl 0.1 · gn 0.6 ▲ high +2.42
▼ low -0.57
Sentence 5
no reference
z mean -0.57 ±0.00 · bl 0.2 · gn 0.1 ▲ high -0.57
▼ low -0.57
z mean +0.57 ±0.97 · bl 0.4 · gn 0.1 ▲ high +1.80
▼ low -0.57
z mean -0.43 ±0.26 · bl 0.5 · gn 0.1 ▲ high +0.09
▼ low -0.57
with emo reference
z mean -0.54 ±0.06 · bl 0.1 · gn 0.1 ▲ high -0.41
▼ low -0.57
z mean -0.57 ±0.00 · bl 0.0 · gn 0.3 ▲ high -0.57
▼ low -0.57
z mean -0.57 ±0.00 · bl 0.1 · gn 0.5 ▲ high -0.57
▼ low -0.57