一种基于小波变换的语音信号去噪净化方法

语音信号在传输之前快速有效的去噪净化，对于3G通信具有重要的现实意义。由于语音信号的非平稳性，传统的小波阈值去噪算法虽然能够去除一部分语音信号中的噪声，但造成有用语音信号尤其是清音部分的损失，导致去噪后的语音听觉质量下降，达不到很高的信噪比。针对这一问题，本文在分析小波变换原理和传统软硬阈值法的基础上，建立了一种改进的阈值语音信号净化方法。仿真实验结果表明，本方法能有效去除信号中的噪声和较好保留语音细节，达到更佳的语音净化效果。

Method of the speech signal denoising and purification based on wavelet transform

In the field of 3G communication, it is significant that speech signal has been denoised and purified rapidly and effectively before its transmission. For the reason of the non-stationary property of speech, conventional wavelet threshold algorithms are capable of removing some noisy components from the corrupted speech, but inevitably harm the useful speech signal especially the unvoice, and then which causes the degradation of the perceptual quality and non-high signabto-noise ratio. Therefore, through analyzing wavelet transform and conventional methods of soft and hard threshold, a new improved threshold method based on wavelet transform of speech signal denoising and purification is proposed. The simulation results show that the new method is not only able to remove the strong noise more effectively, but also can hold the main details better comparing with the conventional ones.