Phase Perception of Transient Signals Having Identical Power Spectra

Abstract

A model often proposed to account for the detection of signals in noise is a simple energy detector. Such a model consists of a narrow filter followed by a nonlinear device (such as a square-law device) with no memory and some type of integrator. Such devices are, of course, limited in their ability to follow rapid changes in the time waveform and for transient waveforms tend to react to the total energy content regardless of how it is packaged. Can the ear distinguish between brief (duration of 20 msec or less) signals if these signals have identical power spectra? Digital signals were generated via a method suggested by Huffman. The signals have identical power spectra but different phase spectra. Most of the signals we have tested sound very similar but can be discriminated almost perfectly. Many are easily discriminated. Thus, even for very brief stimuli, the ear cannot be treated as a simple energy detector. Some recordings of these signals are presented. [This research was supported by the National Institutes of Health, Public, Health Service, U. S. Department of Health, Education, and Welfare.]