Linear prediction analysis of speech based on a pole-zero representation

Publisher Website
Google Scholar
Abstract
Speech analysis and synthesis by linear prediction is based on the assumption that the short-time spectral envelope of speech can be represented by a number of poles. An all-pole representation does not provide an accurate description of speech spectra, particularly for nasals and nasalized sounds. A method for characterizing speech in terms of the parameters of a pole-zero model was presented. In this method, an impulse response representing the composite filtering action of the glottal wave, the vocal tract, the radiation and the speech recording system was first constructed from the speech signal. This impulse response was obtained by performing several stages of all-pole LPC [linear predictive coding] analysis. The pole-zero parameters were determined from the impulse response by solving a set of simultaneous linear equations. The method being noniterative, was very suitable for automatic analysis of speech. The method was applied to real speech data, and the speech spectra derived from the pole-zero model agreed very closely with the actual spectra derived by direct Fourier analysis.

This publication has 0 references indexed in Scilit: