This paper describes an online algorithm for enhancing monaural noisy speech. First, a novel phase-corrected low-delay gammatone filterbank is derived for signal subband decomposition and resynthesis; the subband signals are then analyzed frame by frame. Second, a novel feature named periodicity degree (PD) is proposed to be used for detecting and estimating the fundamental period ( P0) in each frame and for estimating the signal-to-noise ratio (SNR) in each frame-subband signal unit. The PD is calculated in each unit as the multiplication of the normalized autocorrelation and the comb filter ratio, and shown to be robust in various low-SNR conditions. Third, the noise energy level in each signal unit is estimated recursively based on the estimated SNR for units with high PD and based on the noisy signal energy level for units with low PD. Then the a priori SNR is estimated using a decision-directed approach with the estimated noise level. Finally, a revised Wiener gain is calculated, smoothed, and applied to each unit; the processed units are summed across subbands and frames to form the enhanced signal. The P 0 detection accuracy of the algorithm was evaluated on two corpora and showed comparable performance on one corpus and better performance on the other corpus when compared to a recently published pitch detection algorithm. The speech enhancement effect of the algorithm was evaluated on one corpus with two objective criteria and showed better performance in one highly non-stationary noise and comparable performance in two other noises when compared to a state-of-the-art statistical-model based algorithm.

Previously we described a method of calculating auditory excitation patterns and loudness for steady sounds, based on a nonlinear filterbank. Here the method is extended to deal with time-varying sounds. Firstly, the input waveform is transformed to short-term spectrum by a structure with six FFTs, which using longer signal segments for low frequencies and shorter segments for higher frequencies. Secondly, the excitation patterns are calculated from the short-term spectrum, and the summation of the excitation gives a value for the instantaneous loudness. Thirdly, the short-term loudness is calculated from the instantaneous loudness using an averaging mechanism similar to an automatic gain control system, with attack and release times. Finally the long-term loudness is calculated from the short-term loudness using a similar averaging mechanism, but with longer attack and release time. The method gives good predictions of loudness for both steady sounds and time-varying sounds.