基于自适应仿生小波变换的语音增强方法

Speech Enhancement Method Based on Adaptive Bionic Wavelet Transform

分析自适应滤波和小波滤波的原理与方法,提出一种基于自适应仿生小波变换的语音增强方法。该方法首先用仿生小波变换对含噪语音信号进行小波分解,这样可以保证对信号频率和幅值的听觉特性,然后将经仿生小波变换所分离出来的噪声成分作为自适应滤波器的输入。通过选用自适应滤波器的最小二乘算法(RLS)从而实现信噪分离的最佳滤波,以保证去除信号中的相关噪声。实验结果表明,该方法对语音信号有显著的增强效果,能实现语音信号在同频段对噪声成分和有用信号的最佳估计。

The principle and the method of the adaptive filter and the filtering with wavelet transform are analyzed,and a speech enhancement method is presented based on adaptive bionic wavelet transform.Firstly,the noised speech signal is investigated using the bionic wavelet transform.The auditory perception of the frequency and the amplitude signal is maintained.Then,the separated signal of noise by the wavelet transform is the input signal of the adaptive filter. The optimal filtering method for the signal-noise decomposition is realized by using adaptive filtering algorithm—recursive least-squares（RLSs） and the correlated noise of removaled signal is established.Experimental results show that the method has a prominent effect on the speech signal,and realizes the optimal estimation to the valuable signal and the transient signal of noise in the same frequency segment.