ON LINE ACQUISITION AND ANALYSIS OF SPEECH SIGNALS FROM THE DEAF

Abstract

A software based system has been developed for on-line acquisition and analysis of speech signals. This system can be used for the hearing impaired persons to provide them visual feedback through computer screen to enable them to improve their speech quality. It is well known fact that deafness causes a major obstacle for the deaf in acquisition of language and speech in the absence of feedback between ear and the apparatus used for sound production. So,. if there is something wrong in the hearing mechanism, it is difficult for a person to hear and produce normal speech . The speech deficiency of the deaf can be improved by giving them speech therapy from highly skilled speech therapist. But there is shortage of such highly skilled therapists. Therefore, the computer based feedback can be used as one of the effective alternative for speech therapist. This disertation deals with a system which has been developed for acquisition and analysis of speech signal which can be used by the deaf for their speech training. The present signal is compared with the standard signal i.e. with the signal of normal speech. The standard signal is derived from the average speech signal obtained for eight different people. The quantitative features like intensity (amplitude), formant frequency, energy in the time and frequency domains are extracted and are individualy and simultaneously displayed on the screen of the computer for the training. The previous signal is also displayed so as to provide better feedback to the deaf to improve their spe-ec!i,, capability. The details are provided in terms of flowcharts and algorithms. The listing of software is also given in the appendix III.