论文部分内容阅读
In this paper,two speech enhancement systems with supergaussian speech modeling are presented. The clean speech components are estimated by Minimum-Mean-Square-Error (MMSE) es-timator under the assumption that the DCT coefficients of clean speech are mode