A time-varying smoothing factor for the decision-directed approach in speech enhancement

Frequency domain speech enhancement method is often defined in terms of the a priori SNR (Signal to Noise Ratio). A widely used method to determine the a priori SNR from noisy speech is the decision-directed approach. This paper presents a time-varying smoothing factor in the calculation of the decision-directed approach to improve the speech enhancement. Objective and subjective evaluation confirm superiority in noise suppression and quality of the enhanced speech.