基于谱减法和经验模式分解的语音增强

Speech Enhancement Baesd on Spectral Subtraction and EMD

本文提出了一种基于谱减法和经验模式分解的语音增强算法。在低信噪比的情况下用谱减法可以去除语音信号中的大部分背景噪声,再对已处理过的信号进行经验模式分解,对前几个IMF进行阈值处理可以进一步增强语音。实验表明:本算法去噪效果优于传统方法。

A speech enhancement algorithm based on spectral subtraction and EMD are proposed in this paper. After the most of noise removed by using spectral subtraction method in low SNR, decomposed the signal by EMD, the speech could be got further enhanced with the first IMFs thresholded. Experimental results show that the proposed algorithm outperforms general algorithms with respect to denosing performance.