Towards a unified optimal spectral amplitude estimator for speech enhancement in various low-SNR environments

Speech enhancement using a minimum mean-square error short-time spectral amplitude estimator (MMSE-STSA) has been shown to be very effective against stationary uncorrelated additive wide-band noise. In this paper, we show that this approach can also be used to combat narrow-band noise in low-SNR environments. This was accomplished by the integration of an adaptive time varying noise shaping filter (NSF) with the MMSE-STSA algorithm in order to improve the speech enhancement performance by "whitening" the noisy speech signals. Experiments were conducted using a noisy version of speech signals extracted from the TIMIT database. Such experiments demonstrate that the proposed algorithm yields superior performance in comparison with the basic MMSE-STSA algorithm in severe interfering car noise environments for a wide range of SNRs down to -12 dB.