Text-to-Speech on Mizo TTS (evaluation)
3.9DNSMOSVITS
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| VITSFramework=ESPnet2, Text Encoder=Conformer-based, Decoder=HiFi-GAN-based2026.01 | 3.9 | 2.32 | 40.6 | 0.25 | 3.21 | 3.46 | |
| Natural SpeechType=Ground Truth / Reference2026.01 | 3.84 | — | — | — | 4.13 | 4.18 | |
| Tacotron2Framework=ESPnet22026.01 | 3.81 | 2.53 | 41.34 | 0.24 | 2.68 | 2.63 |