Abstract: By means of the fractionated multi resolution analysis ofthe signals with wavelet transform, many problems can be solved,which can not be solved byFourier transform. From the differences between singular points of noise and signal, and the important features of wavelet transformin highlighting the local profiles of signals both in time-frequencydomains, i.e.,the fact that the maximum of the noise wavelet transform reduces dramaticallywith the increase of the scale, the pressure fluctuation isanalyzed. The results show that the wavelet transform is a good way toeliminate noise; and by the power profiles of de-noised signals, thepressure fluctuation shows a fractal construction, as revealed inthe previous study.