Speech recognition for portable multimedia terminals
Andrew Burstein, R.W. Brodersen · 1996
Portable multimedia terminals will require speech recognition for users to perform command and control operations. The speech recognition may either be implemented in software in computers on a high speed network, or in low-power hardware in the terminal itself. To facilitate the writing of applications using speech recognition, a new extension to the tcl language, along with a toolkit of classes, widgets, and applications, provide an easy to program scAPI and a consistent user interface. This thesis also presents the architecture for low-power hardware to perform speech recognition and describes a custom scCMOS IC that implements an important part of the speech recognition algorithm.