Frequency and time—frequency domain analysis tools in measurement

Abstract

A traditional approach to signal processing has been, for a long time, the frequency domain analysis, in which time or space periodicities can be identified. It is a relatively simple approach which usually carries significant information, suitable both as a first-step approach for further signal processing and for feature extraction. Because this approach carries no time information, frequency and time-domain analysis based on wavelets has become increasingly important. This shares a similar analytical approach making use of time-limited functions as the basis of the transform, allowing for space or time localization of short-lived repeating patterns. These are signal-processing tools which require some good understanding of the underlying theory to avoid common pitfalls and circumvent some limitations. Examples are given to show applicability.