Emotional Speech Synthesis on 27 fine-grained emotion tasks (Female)
4.12MOSEmoInstruct-TTS (Dual-Path)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| EmoInstruct-TTS (Dual-Path)mode=zero-shot2026.06 | 4.12 | 3.92 | 4.31 | |
| CosyVoice2mode=zero-shot2026.06 | 3.98 | 3.55 | 4.3 | |
| CosyVoice3mode=zero-shot2026.06 | 3.92 | 3.5 | 4.36 | |
| EmoInstruct-TTS (Emo2embed-Only)mode=zero-shot, ablation=w/o Text Instruct2026.06 | 3.88 | 3.48 | 4.24 | |
| EmoInstruct-TTS (Text-Only)mode=zero-shot, ablation=w/o Emo2emb2026.06 | 3.85 | 3.32 | 4.22 |