Emotional Speech Synthesis Evaluation on Fine-Grained Non-Verbal Expression Data
3.67nMOS (Happy)Only Verbal
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Only VerbalNon-Verbal Cues Granularity=Only Verbal2026.05 | 3.67 | 3.78 | 3.69 | 3.83 | 3.61 | 3.74 | 3.19 | 3.89 | |
| Fine-Grained NV Emotional TTSNon-Verbal Cues Granularity=Fine-Grained NV2026.05 | 3.43 | 4.21 | 3.67 | 4.25 | 3.34 | 4.06 | 3.18 | 4.28 | |
| Coarse-grained NVNon-Verbal Cues Granularity=Coarse-grained NV2026.05 | 3.19 | 3.85 | 3.73 | 4.15 | — | — | — | — |