论文部分内容阅读
针对语音处理问题对离散小波变换及其性质进行了研究，给出了作为语音处理的小波函数及时间因子（ａ，ｂ）必须满足的条件．利用离散小波变换对语音信号的基音提取、频谱分析和清／浊音分割进行了分析．结果表明，小波变换能准确地提取基音和清／浊音分割，并具有较好的稳健性．
In order to solve the problem of speech processing, the discrete wavelet transform and its properties are studied, and the conditions that must be satisfied for the wavelet function and the time factor (a, b) as the speech processing are given. Pitch extraction, spectrum analysis and clear / voiced segmentation of speech signals are analyzed by discrete wavelet transform. The results show that the wavelet transform can accurately extract pitch and clear / voiced segmentation, and has good robustness.