Abstract

A number of optical processors have been proposed for the generation of time–frequency representations of signals, such as the ambiguity function, spectrogram, and the Wigner distribution. Generalized time-frequency representations (GTFR's) have been proposed by Cohen.¹ Spectrograms, Wigner distributions, the Choi–Williams distribution, and the Zhao-Atlas-Marks (ZAM) GTFR² are subsumed as special cases. We show that Cohen's class of GTFRs can be realized optically by the straightforward augmentation of some established optical architectures by means of the simple operation of apodization. A new architecture is also proposed. In our proposal three coherent optical processors take advantage of the rapid input rate and high time-bandwidth product of one-dimensional acousto-optic transducers. Their use permits real-time processing and greatly increases the flexibility of the systems. A simple digital–optical system capable of performing manipulations on the output of the proposed architectures (such as the separation of the positive and negative portions of the GTFR) is also introduced. Finally, results of computer simulations of systems that are designed to produce the ZAM G1TR are presented and compared with spectrogram performance.

© 1990 Optical Society of America