Joint time-frequency analysis of electrochemical noise

Results have been presented for regular Fourier transformation of results of electrochemical noise measurements. The variability of the frequency composition and the averaging character of the classical Fourier analysis have been shown for the example electrochemical noise records. Loss of information concerning the time evolution of the signal has been indicated. These inconveniences render difficult or even make impossible the drawing of conclusions on the basis of a noise signal spectrum of non-stationary character. Two types of transformation have been presented belonging to the class of joint time-frequency analysis methods. The first one, short-time Fourier transformation (STFT) is based on application of a time window, determining the localisation of the spectrum in the time domain. The second, (wavelet transformation) is based on analysis of a signal on the basis of a family of functions formed through translation and scaling of an appropriately chosen localised mother function. Spectrograms obtained by the above methods have been presented, for electrochemical noise recordings. (C) 2001 Elsevier Science B.V. All rights reserved.