A study of an improved minimum mean-square error speech enhancement algorithm

To achieve good tracking capability without overestimation of various non-stationary noise sources,an improved MMSE-STSA speech enhancement algorithm is proposed in this paper.The algorithm employs weighted noise estimation:the noisy speech is weighted by a weighting factor,which is calculated in accordance with the estimated SNR through a nonlinear function,and the estimated noise is obtained as an average of the weighted noisy speech.The spectrum gain modification in this algorithm can further suppress the residual noise for low SNRs and avoid the excessive suppression.The results indicate that,under the non-stationary noisy environment,the proposed algorithm can not only get a good perfor-mance in enhancement,but also reduce the speech distortion.