Reliable a posteriori signal-to-noise ratio features selection

This paper addresses the problem of single microphone speech enhancement in noisy environments. State of the art short-time noise reduction techniques are most often expressed as a spectral gain depending on the signal-to-noise ratio (SNR). The well-known decision-directed approach drastically limits the level of musical noise but the estimated a priori SNR is biased since it depends on the speech spectrum estimated in the previous frame. The consequence of this bias is an annoying reverberation effect. We propose a new method, called reliable features selection noise reduction (RFSNR) technique, that is able to classify the a posteriori SNR estimates into two categories: the reliable features leading to speech components and the unreliable ones corresponding to musical noise only. Then it is possible to directly enhance speech using these reliable components leading to an unbiased estimator.