Speech Enhancement on LibriSpeech clean (test)
1STOIClean
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| CleanCodebook Size=–, Semantic Sup.=–, Frame Rate=–, Token Rate=–, BPS=–2026.05 | 1 | 4.64 | 1 | 2.19 | |
| BigCodecCodebook Size=8192, Semantic Sup.=No, Frame Rate=80, Token Rate=80, BPS=10402026.05 | 0.94 | 2.68 | 0.84 | 2.91 | |
| WavCodecCodebook Size=1024, Semantic Sup.=No, Frame Rate=12.5, Token Rate=100, BPS=10002026.05 | 0.94 | 2.69 | 0.84 | 2.52 | |
| SimCodecCodebook Size=8192, Semantic Sup.=No, Frame Rate=100, Token Rate=100, BPS=13002026.05 | 0.93 | 2.7 | 0.82 | 2.84 | |
| XY-TokenizerCodebook Size=1024, Semantic Sup.=Yes, Frame Rate=12.5, Token Rate=100, BPS=10002026.05 | 0.92 | 2.49 | 0.84 | 2.46 | |
| X-codec2Codebook Size=65536, Semantic Sup.=Yes, Frame Rate=50, Token Rate=50, BPS=8002026.05 | 0.92 | 2.43 | 0.82 | 2.61 | |
| BiCodecCodebook Size=8192, Semantic Sup.=Yes, Frame Rate=50, Token Rate=50, BPS=6502026.05 | 0.92 | 2.51 | 0.8 | 2.72 | |
| MimiCodebook Size=2048, Semantic Sup.=Yes, Frame Rate=12.5, Token Rate=100, BPS=11002026.05 | 0.91 | 2.26 | 0.74 | 3.24 | |
| WavTokenizerCodebook Size=4096, Semantic Sup.=No, Frame Rate=75, Token Rate=75, BPS=9002026.05 | 0.9 | 2.13 | 0.65 | 4.15 |