Zero-shot voice conversion on LibriSpeech clean 16 kHz (test)
4.0557UTMOSEZ-VC
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| EZ-VCVC Capable=true, Codec=false, SR=16k, Bitrate=0.45, Train Set (hours)=En(3.1k)+IN5(9.8k), Test Set=LS2026.06 | 4.0557 | 0.6871 | 14.18 | 0.2041 | — | |
| SDP-Codec-16-LVC Capable=true, Codec=true, SR=16k, Bitrate=0.52, Train Set (hours)=MLS+LS (45.5k), Test Set=LS2026.06 | 3.9832 | 0.8405 | 4.11 | 0.6088 | — | |
| BiCodecVC Capable=true, Codec=true, SR=16k, Bitrate=0.65, Train Set (hours)=Em+LS (3k), Test Set=LS2026.06 | 3.9103 | 0.7577 | 3.43 | 0.5517 | — | |
| MSRCodecVC Capable=true, Codec=true, SR=16k, Bitrate=0.52, Train Set (hours)=MLS+TS+Vox (53.6k), Test Set=LS2026.06 | 3.8966 | 0.7396 | 2.18 | 0.5806 | — |