Audio Tokenization on Seed-TTS-Eval ZH
3.3PESQ NBMiMo-Audio-Tokenizer
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| MiMo-Audio-TokenizerkBPS=1.552025.12 | 3.3 | 2.71 | 0.89 | 0.93 | |
| XY-TokenizerkBPS=1.02025.12 | 2.88 | 2.24 | 0.87 | 0.9 | |
| BigCodeckBPS=1.042025.12 | 2.88 | 2.26 | 0.8 | 0.91 | |
| XCodec2.0kBPS=0.82025.12 | 2.69 | 2.1 | 0.81 | 0.89 | |
| MimikBPS=1.12025.12 | 2.57 | 2.05 | 0.73 | 0.88 | |
| Baichuan-Audio-TokenizerkBPS=1.02025.12 | 2.37 | 1.84 | 0.78 | 0.86 | |
| GLM-4-Voice-TokenizerkBPS=0.1752025.12 | 1.11 | 1.06 | 0.33 | 0.61 |