Speech Naturalness Assessment on CVSS-C
3.74UTMOS ScoreVanilla
Evaluation Results
| Method | Links | |
|---|---|---|
| VanillaSystem=End-to-end, Optimization Stage=After DPO, Back-translation Evaluation Metric=METEOR2026.01 | 3.74 | |
| VanillaSystem=End-to-end, Optimization Stage=After DPO, Back-translation Evaluation Metric=BLEU2026.01 | 3.72 | |
| VanillaSystem=End-to-end, Optimization Stage=Before DPO2026.01 | 3.71 | |
| Chain of ModalitySystem=End-to-end, Optimization Stage=After DPO, Back-translation Evaluation Metric=BLEU2026.01 | 3.7 | |
| Chain of ModalitySystem=End-to-end, Optimization Stage=Before DPO2026.01 | 3.68 | |
| Tri-Task LearningSystem=End-to-end, Optimization Stage=After DPO, Back-translation Evaluation Metric=METEOR2026.01 | 3.68 | |
| Chain of ModalitySystem=End-to-end, Optimization Stage=After DPO, Back-translation Evaluation Metric=METEOR2026.01 | 3.66 | |
| Tri-Task LearningSystem=End-to-end, Optimization Stage=After DPO, Back-translation Evaluation Metric=BLEU2026.01 | 3.66 | |
| Tri-Task LearningSystem=End-to-end, Optimization Stage=Before DPO2026.01 | 3.65 | |
| S2T+TTSSystem=Cascaded2026.01 | 3.04 |