From: Single-channel speech enhancement based on joint constrained dictionary learning
SNR (dB) | Frame length = 128 | Frame length = 256 | Frame length = 512 |
---|---|---|---|
− 10 | 6.7389 | 7.8283 | 7.0193 |
− 5 | 4.6526 | 5.8115 | 5.2284 |
0 | 2.8977 | 4.0693 | 3.5235 |
5 | 1.9062 | 2.3522 | 2.2963 |
10 | 0.2983 | 1.0279 | 0.5137 |