Speech technology and its applications

Abstract

The first stage of capturing a speech signal is to change it from an acoustic signal into an electrical signal, using microphone and amplifier. This signal must be digitized before the computer can handle it, using sampling techniques.

Speech is produced by capturing human speech patterns and analysing them for later reproduction by mechanical means. Text-to-speech synthesis is used in driving voice synthesizers.

Speech recognition is based on the matching of frames of speech. Dynamic programming is a significant step forward in speech recognition.

An application is currently being developed using speech recognition on a PABX.