In this paper, we present new speech denosing methods based on some types of maximum a posteriori (MAP) filters applied in the wavelet domain. These methods take into account the statistical properties of the discrete ...