Speech Reconstruction on Librispeech (test)
0.9STOIMSR-Codec-612
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MSR-Codec-612Codebook Size=500/512/2048, Nq=3, Token Rate (TPS)=62.5, Bitrates=6122025.09 | 0.9 | 2.73 | 2.09 | 4.13 | 0.83 | |
| WavTokenizerCodebook Size=4096, Nq=1, Token Rate (TPS)=75, Bitrates=9002025.09 | 0.89 | 2.64 | 2.14 | 3.94 | 0.67 | |
| SemantiCodecCodebook Size=16384, Nq=1, Token Rate (TPS)=100, Bitrates=14002025.09 | 0.88 | 2.63 | 2.07 | 2.92 | 0.74 | |
| X-codecCodebook Size=1024, Nq=2, Token Rate (TPS)=100, Bitrates=10002025.09 | 0.86 | 2.88 | 2.33 | 4.21 | 0.72 | |
| Single-CodecCodebook Size=8192, Nq=1, Token Rate (TPS)=23.4, Bitrates=3042025.09 | 0.86 | 2.42 | 1.88 | 3.72 | 0.6 | |
| MSR-Codec-524Codebook Size=500/256/256, Nq=3, Token Rate (TPS)=62.5, Bitrates=5242025.09 | 0.85 | 2.59 | 1.98 | 4.14 | 0.81 | |
| MSR-Codec-424Codebook Size=500/64/32, Nq=3, Token Rate (TPS)=62.5, Bitrates=4242025.09 | 0.84 | 2.37 | 1.82 | 4.15 | 0.8 | |
| SpeechTokenizerCodebook Size=1024, Nq=2, Token Rate (TPS)=100, Bitrates=10002025.09 | 0.77 | 1.59 | 1.25 | 2.28 | 0.36 |