Application Of Wavelet Analysis On Speech De-Noising

The stdudy on signal denoising method is an constant issue in signal processing fields. Typical signal processing mechanism, such as time domain, frequency domain, Short-time Fourier Transform, wigner-ville distribution and so on have their limitions. As the result, their application extent are restricted. Wavelet transform, which developed in late 1980's, is a new kind of time frequency analytic method. It has good localization characteristic in both time and frequency domain; and has been applied for signal de-noising fields.In this thesis, focusing on wavelet transform theory, we analyze its characteristic and stress on takinge its advantage to implement a new remove noise mechanism for non-stationary signal. In addition, some de-noising methods based on wavelet transform is compared, the wavelet thresholding method is studied in the article. The two difficulty, one is which mother wavelet and decomposition scale can get best de-nosing effect in practical application, another important problem is threshold selection, which may be occurred in wavelet thresholding, is completely and deeply discussed. Finally, taking an typically actral use as an example, we analyze the real speech signals by using Matlab's Wavelet Toolbox, and present a modified method of speech de-noising in wavelet domain for the noise pattern. The experiment results show that the method is suitable for the case of non-stationary signal, it is simpler, easier to be complished relatively, and completely meet the demand of the subject. The academic analysis and the research result is consistent, which indicate that the method in the paper has some reference value.