Speech Reconstruction on SEED-EN
2.77PESQH-Codec-2.0 (Large)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| H-Codec-2.0 (Large)FS=48k, FPS=6.25, Paras=1.2B2025.12 | 2.77 | 0.88 | 0.93 | |
| MiMo-Audio-TokenizerFS=24k, FPS=25, Paras=1.2B2025.12 | 2.43 | 0.85 | 0.92 | |
| X-Codec2FS=16k, FPS=50, Paras=242M2025.12 | 2.37 | 0.82 | 0.93 | |
| H-Codec-1.5FS=16k, FPS=18.80/20.18, Paras=678M2025.12 | 2.3 | 0.82 | 0.9 | |
| H-Codec-2.0 (Small)FS=48k, FPS=6.25, Paras=236M2025.12 | 2.25 | 0.83 | 0.9 | |
| BigCodecFS=16k, FPS=80, Paras=159M2025.12 | 2.22 | 0.8 | 0.91 | |
| H-Codec-1.0FS=16k, FPS=25, Paras=146M2025.12 | 2.22 | 0.73 | 0.9 | |
| XY-TokenizerFS=16k, FPS=12.5, Paras=-2025.12 | 2.14 | 0.82 | 0.9 | |
| FlexiCodecFS=16k, FPS=7.16/8.49, Paras=216M2025.12 | 2.11 | 0.75 | 0.9 | |
| MimiFS=16k, FPS=12.5, Paras=79M2025.12 | 2.01 | 0.77 | 0.89 | |
| Baichuan-Audio-TokenizerFS=16k, FPS=12.5, Paras=-2025.12 | 1.62 | 0.69 | 0.85 | |
| GLM4-Voice-TokenizerFS=16k, FPS=12.5, Paras=-2025.12 | 1.05 | 0.12 | 0.6 |