Table 6 Experimental results of different loss function
From: Supervised Attention Multi-Scale Temporal Convolutional Network for monaural speech enhancement
| Â | PESQ | STOI (%) | SDR |
|---|---|---|---|
Unprocessed | 2.24 | 88.5 | 14.81 |
SA-MSTCN1 (MSE) | 3.35 | 93.8 | 20.02 |
SA-MSTCN1 (\(\mathcal {L}_{1}\)) | 3.44 | 94.3 | 20.61 |
SA-MSTCN2 (MSE) | 3.39 | 93.9 | 20.24 |
SA-MSTCN2 (\(\mathcal {L}_{2}\)) | 3.47 | 94.3 | 20.83 |