ISBN-13: 9783659231704 / Angielski / Miękka / 2012 / 64 str.
Denoising of any type of signal is a vital part of communication and signal processing system. A signal in the communication system is the information containing part which needs to be processed, but during the process some noise is added in the signal and signal become noisy. The source of noise like noisy engine, pump etc introduces noise over telephone channel or in radio communication device. This is now necessary to denoise that signal or to remove that noise from that signal. Denoising of a signal can be done by using a low pass Butterworth filter, statistically matched wavelet filter and wavelet thresholding method. Wavelet transform is a very helpful method of speech signal analysis and it can be used in many applications for e.g. image processing and signal de-noising. Wavelet transform breaks a speech signal into multi-scale representation. It is also called wavelet thresholding. This technique replaces the coefficients by zero below and above a threshold value. This technique is very useful to minimize the mean square error. This report is based on wavelet denoising algorithm. Number of wavelets is applied on different speech signal and performance is evaluated.