Skip to main content

Table 4 The average STOIs (%) with and without postfiltering for the typical DNN-based speech enhancement methods

From: Low-complexity artificial noise suppression methods for deep learning-based speech enhancement algorithms

SNR/dB − 5 0 5 10 Aver. SNR/dB − 5 0 5 10 Aver.
Noisy 60.12 73.29 85.89 91.30 77.65 Noisy 60.12 73.29 85.89 91.30 77.65
CRN 77.46 86.82 93.93 95.70 88.48 DCN 79.46 86.10 93.78 95.73 88.77
SPP-MMSE 76.95 86.55 93.67 95.71 88.22 SPP-MMSE 78.92 85.61 93.59 95.79 88.48
SPP-proposed-1 75.28 84.70 92.76 95.46 87.05 SPP-proposed-1 76.89 84.06 92.72 95.56 87.31
SPP-proposed-2 75.42 84.84 92.41 95.03 86.93 SPP-proposed-2 77.56 84.28 92.17 95.37 87.34
SPP-proposed-3 75.51 85.65 93.62 95.83 87.65 SPP-proposed-3 77.27 85.21 93.46 95.91 87.96
GRN 79.52 87.93 94.23 95.71 89.35 DARCN 79.44 88.99 94.02 95.64 89.52
SPP-MMSE 79.36 87.79 94.03 95.66 89.21 SPP-MMSE 79.26 88.92 93.89 95.63 89.43
SPP-proposed-1 77.34 85.94 93.15 95.60 88.01 SPP-proposed-1 76.95 86.73 92.83 95.37 87.97
SPP-proposed-2 77.61 86.28 92.67 94.97 87.88 SPP-proposed-2 77.54 86.41 91.51 93.75 87.30
SPP-proposed-3 77.94 87.06 94.08 95.86 88.74 SPP-proposed-3 77.65 87.37 93.74 95.58 88.59