An automatic spontaneous speech recognition system for Punjabi language

Abstract

Punjabi is a very tonal language, making employ of a range of tones to distinguish words that would otherwise be alike. Three main tones can be recognized: high-rising-falling, mid-rising-falling, and low rising. Some work has been done in the field of isolated word, connected word, and continuous speech recognition system for Punjabi language. Spontaneous speech recognition is one area where no work has been done so far for Punjabi language. Spontaneous speech and speech from written language are exceptionally dissimilar both acoustically and linguistically. Spontaneous speech contains crammed silence, preservation, faltering, duplications, incomplete vocabulary, and stuttering. In this paper, an effort has been made to build an automatic spontaneous speech recognizer to recognize Punjabi live speech by using speech recognition model using sphinx toolkit.