Speech Enhancement Using Laplacian Mixture Model under Signal PresenceUncertainty

In this paper, an estimator for speech enhancement based on Laplacian Mixture Model (LMM)has beenproposed. The proposed method, estimates the complex Discerete Fourier Transform(DFT) coefficientsof clean speech from noisy speech using the Minimum Mean Square Error(MMSE) estimator, whenthe clean speech DFT coefficients are supposed mixture of Laplacians and the DFT coefficients ofnoise are assumed zero-mean Gaussian distribution. Furthermore, the MMSE estimator under speechpresence uncertainty and the Laplacian mixture model were derived. It is shown that the proposedestimator has better performance than three estimators based on single Gaussian and single Laplacianmodels. Also under speech presence uncertainty the results become better.