本文提出了一种新的基于小波变换和参数滤波的音素分段算法 ,称为小波参数滤波算法 (WPF) .小波变换具有和人耳相似的特性 ,即随着频率的升高 ,分辨率下降 ,我们利用此特性先对信号进行小波域滤波 ,然后利用参数滤波对信号进行分段。实验表明本文提出的算法在性能上优于常规的参数滤波方法 In this paper, a new phoneme segmentation algorithm based on wavelet transform and parametric filtering is proposed, which is called Wavelet Parameter Filtering (WPF) .Wavelet transform has similar characteristics to the human ear, that is, as the frequency increases, the resolution decreases , We use this characteristic to filter the signal in the wavelet domain first, and then use the parameter filtering to segment the signal. Experiments show that the proposed algorithm outperforms the conventional parametric filtering in performance