The speech recognition system described below is designed to control application programs run under MS-DOS on IBM PC AT by separately pronounced words. The system is realized as a TSR-program interfacing the application program via the keyboard buffer. The system uses the Sound Blaster (or compatible) to digitize speech signal. A brief description of the system helping blind PC users to work on PC is given, thereby PC is controlled by voice commands and the keyboard is voiced.
Keywords: speech signal, recognition, personal computer, blind users, learning, system performance