Skip to main content

Advertisement

Table 4 The P EL and P NR results when threshold selection is applied

From: Improved monaural speech segregation based on computational auditory scene analysis

Intrusion Threshold selection Hu–Wang tandem Mixture
   system  
  PEL(%) PNR(%) PEL(%) PNR(%) PNR(%)
N1 3.28 2.21 3.34 2.46 27.11
N2 4.89 5.17 4.76 6.08 29.50
N3 6.11 5.50 6.24 5.62 40.56
N4 3.01 5.47 3.03 5.82 22.01
N5 2.58 1.23 2.39 1.64 9.77
N6 4.77 2.08 4.65 2.17 46.31
N7 6.04 6.15 6.19 6.63 36.51
N8 6.95 1.14 7.07 3.23 18.43
N9 7.20 4.98 7.49 5.21 29.46
N10 3.80 0.56 3.84 0.58 11.77
N11 5.85 2.28 6.38 3.02 20.73
N12 4.49 3.23 4.67 3.71 76.79
N13 4.35 2.93 5.09 3.12 13.76
N14 5.95 4.27 6.03 4.80 23.16
N15 4.16 4.60 4.18 5.78 26.84
N16 2.24 2.72 2.03 2.89 6.64
N17 6.08 6.20 6.37 6.94 43.07
Average 4.81 3.57 4.93 4.10 28.38