Comparison of periodograms of superimposed epochs and Fourier spectra of experimental series

Abstract

Periodograms and Fourier spectra of different experimental time series are calculated. The efficiency of the methods applied to study different signals is considered. The periodogram method often makes it possible to identify periodicities which are hardly identifiable by the Fourier analysis. First of all, it concerns the quasiperiodic processes with a considerable variability in amplitude and form of the rhythmical component, but almost constant period or very weak stable rhythm with a nonsinusoidal form. One more advantage of the periodograms is their low sensitivity to pauses and omissions in the time series. However, data preprocessing is often needed, for instance, filtration of high frequency noise, rarer data interrogation, and analysis of the rhythmical component form. The rhythms revealed by the periodogram method are commonly studied additionally to confirm their significance and to describe their properties. The low computational effectiveness, even relative to the classical Fourier transformation, is one more disadvantage of the periodogram method.