Functional Model of Perception and Its Application to Speech Processing
M. Ross · The Journal of the Acoustical Society of America · 1969
During recent decades, problems in bandwidth reduction of speech signals have been studied from the point of view of the speech-producing mechanism. However, when the emphasis is placed on perception, parameters that correspond to the stimulus-receiving mechanism assume relatively greater importance. In this paper, a functional model of perception is described in which the magnitude of the frequency distribution of voiced sounds at the basilar membrane are expressed in terms of a linear combination of Hermite base functions. In terms of this model, an attempt is made to study and simulate the weighting performed by the ear on the original speech signal. The method of measuring the pertinent coefficients is given; the nature of the analyzer-synthesizer scheme is described; and experimental results are presented, with implications for speech bandwidth reduction. [This research was supported by a NASA Fellowship and the Electrical Engineering Department at Northeastern University.]