ABSTRACT
Spectral analysis and correlation techniques are an aid to the interpretation of signals and to the systems that generate them. These methods are now widely used for the analysis and interpretation of measurements performed in medicine, geophysics, vibration analysis, communications, and several other areas. Practical applications of spectral and correlation analysis are performed on discrete-time signals. These are obtained either from a sequence of discrete measurements or from the transformation of a continuous signal to digital format using an analog-to-digital converter. Nonstationary signals present a particular problem. In cases where the signal statistical properties change relatively slowly with time, it is possible to select short segments of quasi-stationary data and to use the discrete Fourier transforms or parametric methods to estimate the spectra as mentioned previously.
