Multi-channel speech separation on WSJ0 + WHAM! (in-domain)
1.98PESQSpatialNet + Fast-GeCo
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| SpatialNet + Fast-GeCoNFE=1+1, RTF=0.0078+0.0068, Type=D+G2026.06 | 1.98 | 79 | 9.5 | 3.07 | 3.52 | 3.97 | |
| DeFTAN2 + Fast-GeCoNFE=1+1, RTF=0.0155+0.0068, Type=D+G2026.06 | 1.96 | 79 | 9.81 | 3.11 | 3.51 | 4.11 | |
| SpatialNet + MeCoNFE=1+1, RTF=0.0078+0.0068, Type=D+G2026.06 | 1.95 | 80 | 9.88 | 3.18 | 3.65 | 4.41 | |
| DeFTAN2 + MeCoNFE=1+1, RTF=0.0155+0.0068, Type=D+G2026.06 | 1.93 | 80 | 10.08 | 3.19 | 3.7 | 4.5 | |
| CrossNet + Fast-GeCoNFE=1+1, RTF=0.0065+0.0068, Type=D+G2026.06 | 1.89 | 77 | 8.92 | 3.05 | 3.34 | 3.95 | |
| DeFTAN2NFE=1, RTF=0.0155, Type=D2026.06 | 1.88 | 75 | 9.31 | 2.94 | 3.12 | 3.92 | |
| SpatialNetNFE=1, RTF=0.0078, Type=D2026.06 | 1.87 | 73 | 8.77 | 2.82 | 2.91 | 3.58 | |
| CrossNet + MeCoNFE=1+1, RTF=0.0065+0.0068, Type=D+G2026.06 | 1.86 | 78 | 9.22 | 3.15 | 3.56 | 4.41 | |
| CrossNetNFE=1, RTF=0.0065, Type=D2026.06 | 1.81 | 72 | 8.29 | 2.79 | 2.81 | 3.52 | |
| DeFTAN2 + MeanFlowNFE=1+1, RTF=0.0155+0.0068, Type=D+G2026.06 | 1.78 | 77 | 10.01 | 3.04 | 3.63 | 4.43 | |
| SpatialNet + MeanFlowNFE=1+1, RTF=0.0078+0.0068, Type=D+G2026.06 | 1.77 | 77 | 9.78 | 3.05 | 3.66 | 4.37 | |
| CrossNet + MeanFlowNFE=1+1, RTF=0.0065+0.0068, Type=D+G2026.06 | 1.71 | 75 | 9.14 | 2.98 | 3.51 | 4.35 |