A Functional Model of Signal Processing in the Peripheral Auditory System

This paper attempts to outline a functional model of the first stages of auditory analysis including spectral analysis and the principal transformations of signal intensity.

A model has been implemented by a cascade arrangement of single resonant circuits and buffer amplifiers with output taps at each stage. This analyzer has been designed so as to preserve place-frequency, place-delay, and amplitude frequency characteristics at any tap. Furthermore, the amplitude-coordinate (place) characteristics at a constant frequency conform with Zwicker's data of critical bandwidths and masking. The amplitude processing involves half-wave rectification, instantaneous amplitude compression, and peripheral adaptation. The best agreement with experiments on vowel perception with relative delays between spectral components is provided by a peripheral adaptation controlled by a weighted amplitude selection and 40 ms time constant.