Speech Recognition - Application to the Sinhala Language

Abstract

A simple speech recognizer was implemented for the Sinhala language by using FFT combined with Mel scaled filter banks. The accuracy of the speech recognizer depends on what words are in the vocabulary, the number of sections in the filter bank, and the sampling frequency. In general, short words are difficult and long words are easy to recognize. With a limited set of sounds the technique discussed in this paper can be used to produce a speaker independent speech recognizer. A direct industry application to such a system would be the implementation of a voice command process control system.