List your company

Emotional Text-to-Speech on IEMOCAP Low-mismatch

0.908E-SIMCoCoEmo
0.874720.883360.8920.90064Feb 3, 2026

Evaluation results

MethodLinks
CoCoEmoModel=CosyVoice2, Steering strength α=62026.020.9080.6380.8797
CoCoEmoModel=CosyVoice2, Steering strength α=32026.020.9010.5150.8826.08
InstructionModel=CosyVoice22026.020.8970.6490.8722.47
CoCoEmoModel=IndexTTS2, Steering strength α=62026.020.8920.6280.876.85
CoCoEmoModel=IndexTTS2, Steering strength α=32026.020.8890.5140.8726.46
No-steerModel=CosyVoice22026.020.8830.3940.8787.34
Emo vectorModel=IndexTTS22026.020.8810.6710.8465.6
No-steerModel=IndexTTS22026.020.8760.3770.8755.8