基于小波阈值法的语音增强算法研究

随着小波变换理论的不断发展，小波变换去噪法不断丰富，小波阈值去噪法以算法简单、计算量小，吸引了众多研究者。通过对小波阈值去噪法传统阈值函数的研究，发现软硬阈值函数都存在着缺点；改进的半软阈值法综合了两者的优点，改正了两者的缺陷，明显改善了语音增强效果。并提出一种基于小波阈值去噪法与谱减法结合的改进算法，在较低信噪比下进行仿真，很好地抑制背景噪声和音乐噪声，减少语音信号的失真。

Algorithm For Speech Enhancement Based on Wavelet thresholding method

With the development of wavelet theory, wavelet transform denoising enrich, the threshold denoising method as a simple algorithm and low computation has attracted many researchers. The research on the threshold denoising, we found that the soft and hard threshold functions both have defects, the improved semi-soft threshold method combines both of the advantages, and correct the defects, the result of speech enhancement has improved significantly.Puts forward a improved algorithm of speech enhancement, the threshold denoising method combined with spectral subtraction, at the situation of low SNR , it can suppression of background noise and residual musical noise, and reduce the voice of the distortion.