Automatic Speech Recognition on ATCO2 (test)
24.18WER (Synth-only)VC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| VCCategory=Voice Conversion, AAS=true2026.06 | 24.18 | 22.54 | |
| VCCategory=Voice Conversion, AAS=false2026.06 | 26.46 | 22.45 | |
| ACL1→L2 + VC + TTSCategory=Synthetic Mix, AAS=true, Target Accent=L2-ARCTIC, VC (L1→L2)=Yes2026.06 | 28.05 | 22.91 | |
| ACL1→L2 + TTSCategory=Synthetic Mix, AAS=true, Target Accent=L2-ARCTIC, VC (L1→L2)=Yes2026.06 | 28.9 | 21.69 | |
| ACL1→L2Category=Accent Conversion, AAS=true, Target Accent=ATCO2, VC (L1→L2)=Yes2026.06 | 31.4 | 22 | |
| ACL1→L2Category=Accent Conversion, AAS=true, Target Accent=L2-ARCTIC, VC (L1→L2)=Yes2026.06 | 31.48 | 21.95 | |
| ACL1→L2 + VCCategory=Synthetic Mix, AAS=true, Target Accent=L2-ARCTIC, VC (L1→L2)=Yes2026.06 | 32.2 | 22.07 | |
| ACL1→L2Category=Accent Conversion, AAS=true, Target Accent=L2-ARCTIC, VC (L1→L2)=No2026.06 | 32.96 | 22.16 | |
| TTSCategory=Speech Synthesis, AAS=true2026.06 | 33.77 | 21.69 | |
| ACL1→L2Category=Accent Conversion, AAS=true, Target Accent=ATCO2, VC (L1→L2)=No2026.06 | 33.84 | 22.6 | |
| VC + TTSCategory=Synthetic Mix, AAS=true2026.06 | 34.39 | 22 | |
| ACL1→L2Category=Accent Conversion, AAS=false, Target Accent=L2-ARCTIC, VC (L1→L2)=Yes2026.06 | 38.22 | 21.64 | |
| ACL2→L1Category=Accent Conversion, AAS=true2026.06 | 41.6 | 25.92 | |
| TTSCategory=Speech Synthesis, AAS=false2026.06 | 53.88 | 22.9 | |
| Out-of-the-boxCategory=Baseline2026.06 | 63.32 | — | |
| Fine-tuning with real data onlyCategory=Baseline2026.06 | — | 22.69 |