Table 3 Comparison of the SDR metric with other models on Musdb18
From: Stripe-Transformer: deep stripe feature learning for music source separation
| Â | Vocals | Drums | Bass | Other | Avg. |
|---|---|---|---|---|---|
Wave-U-Net [15] | 3.25 | 4.22 | 3.21 | 2.25 | 3.23 |
Meta-TasNet [18] | 6.40 | 5.91 | 5.58 | 4.19 | 5.52 |
Open-Unmix [63] | 6.32 | 5.73 | 5.23 | 4.02 | 5.33 |
MMDenseLSTM [27] | 6.60 | 6.41 | 5.16 | 4.15 | 5.58 |
Sams-Net [30] | 6.61 | 6.63 | 5.25 | 4.09 | 5.65 |
D3Net [31] | 7.24 | 7.01 | 5.25 | 4.53 | 6.01 |
Deep ResUNet [29] | 8.98 | 6.62 | 6.04 | 5.29 | 6.73 |
Stripe-Transformer | 7.83 | 7.63 | 5.50 | 5.89 | 6.71 |