Speech Analysis and Recognition by Computer Using Zero-Crossing Information

Author: De Mori R.  

Publisher: S. Hirzel Verlag

ISSN: 1610-1928

Source: Acta Acustica united with Acustica, Vol.25, Iss.5, 1971-11, pp. : 269-279

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

Some theoretical results on the properties of zero crossings of continuous waveforms are reviewed and applied to speech analysis. A system consisting on an analog preprocessing unit interfaced with a minicomputer is described and some results are presented. They are concerning the automatic formant detection by the estimations of the zero crossing densities of filtered speech signals, an investigation of the statistics of their zero crossing intervals to group them into a small number of classes suitable to be used as input parameters for a speech recognition process and a new method for representing a spoken word by a parametric graph on a plane using time as parameter.