Demucs v4

Reported on 6 benchmarks across 1 task · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Audio6 results

Speech EnhancementonEARS-WHAM
DNSMOS· 2022-11-15
3.66
best: 3.88 (SGMSE+)
Hybrid Transformers for Music Source Separation arXiv:2211.08553
Speech EnhancementonEARS-WHAM
ESTOI· 2022-11-15
0.71
best: 0.73 (Schrödinger Bridge (PESQ loss))
Hybrid Transformers for Music Source Separation arXiv:2211.08553
Speech EnhancementonEARS-WHAM
PESQ-WB· 2022-11-15
2.37
best: 3.09 (Schrödinger Bridge (PESQ loss))
Hybrid Transformers for Music Source Separation arXiv:2211.08553
Speech EnhancementonEARS-WHAM
POLQA· 2022-11-15
2.97
best: 3.71 (Schrödinger Bridge (PESQ loss))
Hybrid Transformers for Music Source Separation arXiv:2211.08553
Speech EnhancementonEARS-WHAM
SI-SDR· 2022-11-15
16.92
best: 17.85 (Schrödinger Bridge)
Hybrid Transformers for Music Source Separation arXiv:2211.08553
Speech EnhancementonEARS-WHAM
SIGMOS· 2022-11-15
2.87
best: 3.44 (Schrödinger Bridge)
Hybrid Transformers for Music Source Separation arXiv:2211.08553