Speech Enhancement and Automatic Gain Control on MMCSG
-22.68LUFSSE-AGCNet†
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| SE-AGCNet†retrained=LibriAGC dataset2026.06 | -22.68 | -23.04 | 4.8 | 13.86 | 30.36 | |
| MP-SENet (SE)† + pyagcretrained=LibriAGC dataset, post-processing=pyagc2026.06 | -23.74 | -24.08 | 4.65 | 14.06 | 34.91 | |
| MP-SENet (Orig) + pyagctraining_data=DNS Challenge 2020, post-processing=pyagc2026.06 | -29.91 | -30.42 | 5.72 | 44.25 | 50.72 | |
| MP-SENet (Orig)training_data=DNS Challenge 2020, status=Original checkpoint2026.06 | -38.86 | -45.23 | 23.15 | 46.74 | 52.8 | |
| MP-SENet (SE)†retrained=LibriAGC dataset2026.06 | -39.07 | -45.16 | 20.83 | 14.41 | 50.85 | |
| Noisystatus=baseline input2026.06 | -40.24 | -45.92 | 20.21 | 15.12 | 51.36 | |
| MP-SENet (AGC)†retrained=LibriAGC dataset2026.06 | -47.57 | -53.08 | 19.72 | 15.19 | 51.92 |