Skip to main content

Table 4 The mean SNR, PESQ, and IS values obtained from enhancing 24 noisy male and female speech samples at our experiments for the proposed method compared to the other methods for various SNRs.

From: Improved Noise Minimum Statistics Estimation Algorithm for Using in a Speech-Passing Noise-Rejecting Headset

   non engaged engaged
Input SNR 0 5 10 0 5 10
Seg_SNR In −1 2.4 6.2 −1.2 1.56 4.49
  The initial 3.7 6 8.1 1.7 3.94 5.94
  The improved 5.5 6.3 6.9 6.2 7.22 8
SNR_fw In −9 −6 −1 −13 −8.5 −4
  The initial −1 1.3 4.3 −6.2 −2 1.45
  The improved 3 3.8 4.8 2.2 3.8 5.1
PESQ_mos In 1.4 1.6 1.8 1.52 1.68 1.92
  The initial 1.5 1.9 2.2 1.29 1.62 1.96
  The improved 2 2.3 2.4 1.93 2.19 2.4
IS In 2.1 1.3 0.7 3.65 2.91 2.22
  The initial 1.7 1.2 0.9 2.77 2.43 1.91
  The improved 0.6 0.5 0.4 1.63 1.42 1.22