ABSTRACT
This chapter explores the usefulness of reassignment by stressing how it permits to overcome the localization and interference trade-off that is usually observed in classical time–frequency (TF) analysis. The reassignment principle is to be detailed in the simplest case of the spectrogram, and some examples illustrate different facets of the technique. The reassignment principle can be applied to time-scale representations of the affine class. The chapter focuses on reassignment in action: starting from the spectrogram case, efficient algorithmic issues, as well as extensions to more general situations, including time-scale distributions such as the scalogram. It provides a number of real-world situations where reassignment may be of effective usefulness, both in exploratory data analysis and in signal manipulation. Beyond TF analysis, useful information on the signal structure can be extracted from the reassignment operators, and used in a signal processing application.
