A Brief Overview of Speech Synthesis and Recognition Technologies

Abstract

An overview of several aspects of speech synthesis and recognition technologies is provided as background for subsequent speakers in this session. Specifically, we discuss speech synthesis by rule using automatic text-to-speech conversion and speaker-dependent isolated word recognition. Both of these speech I/O technologies have been developed sufficiently to the point where commercial products are now available for a number of applications. Some of the limitations of these devices are described and suggestions for future research in both synthesis and recognition are outlined.