Table 3: Recognition rate for the proposed perceptual generalized spectral subtraction method based on the gammatone and the gammachirp filter banks decomposition (MMSE_GT and MMSE_GC) compared with the spectral modification rule (MMSE).

Babble noise
Methods/SNR051015

MMSE64.81%79.41%89.36%90.13%
MMSE_GT
 MCRA67.69%84.79%93.18%94.97%
 MCRA265.90%83.74%92.24%93.93%
 IMCRA68.76%84.41%93.80%94.62%
 MS66.24%82.15%92.07%93.07%
MMSE_GC
 MCRA57.36%74.40%87.66%93.00%
 MCRA261.29%74.73%87.32%92.89%
 IMCRA67.91%84.08%91.83%93.24%
 MS65.78%75.33%90.33%91.10%

Car noise
Methods/SNR051015

MMSE75.88%85.25%91.40%92.96%
MMSE_GT
 MCRA80.68%89.56%92.64%95.13%
 MCRA279.03%86.83%91.51%93.84%
 IMCRA82.40%90.51%92.06%94.99%
 MS79.78%90.15%91.88%93.35%
MMSE_GC
 MCRA74.77%86.71%91.95%93.69%
 MCRA272.64%82.49%90.33%93.08%
 IMCRA80.18%90.20%92.29%94.89%
 MS79.13%89.75%92.57%93.80%