Perceptually motivated audio time-frequency analysis

Abstract

Fourier analysis has long been an indispensable tool for the investigation of signals whose frequency content does not change with time. It has likewise been of fundamental importance in engineering for the study of linear time-invariant systems. Clearly, however, most natural and man-made signals do have a time-varying frequency content. Fourier analysis is unfortunately inadequate for the analysis of such signals, as it cannot tell us how the signal’s frequency content changes with time-it can determine which frequency components are present in the signal, but not when they occur. In order to overcome the limitations of Fourier analysis, therefore, a join t time-frequency signal description is required.