一种快速离散小波变换算法及其在语音信号中的应用  

Fast discrete wavelet transform and its application to speech signal processing

作　　者：徐伟业[1] 

机构地区：[1]南京工程学院通信工程学院,南京211016

出　　处：《计算机工程与应用》2011年第35期143-146,149,共5页Computer Engineering and Applications

基　　金：南京工程学院科研基金项目(No.QKJB2009019)

摘　　要：随着小波分析的理论研究水平不断提高,其应用领域也在不断扩展。特别是其多分辨率分析和Mallat算法在数字信号处理和数字通信中得到了广泛的应用。但是如果直接按照上述算法计算信号的小波分解和重构,其计算量将是很大的。通过对实序列的快速傅里叶变换(FFT)算法的推导及Mallat算法原理的分析,根据离散小波变换算法结构特征,提出了一种基于FFT的快速离散小波变换算法,并从数学理论上进行了论证。同时把该算法应用到实际的语音信号处理中,得到了很好的快速分解和重构效果。With the research standard of wavelet analysis improving,the application field of the wavelet transform is keeping spreading.Especially it is widely used in digital signal processing and digital communication because of its multi-resolu- tion analysis and Mallat algorithm.However, if the signal decomposition and reconstruction is calculated based on the above-mentioned algorithm, the computational complexity will be very large, and the timely processing of signal will be affected.On the basis of analyzing the principle of Mallat algorithm principle,by deriving the real signal Fast Fourier Transform （FFT） algorithm,a fast wavelet transform algorithm based on FFT is proposed in terms of discrete wavelet transform structure in this paper.This algorithm is testified well from mathematical theory.Meanwhile, the proposed fast algorithm is applied to the speech signal processing,and fast decomposition and reconstruction result is obtained well.

关 键 词：快速傅里叶变换 快速离散小波变换 多分辨率分析 信号分解和重构 语音信号处理 

分 类 号：TN911[电子电信—通信与信息系统]