The MISS speech synthesis system

To provide speech output for a major demonstration project of sophisticated computer based instruction, the Microprogrammed Intoned Speech Synthesizer (MISS) utilizing linear predictive coding techniques was developed. In addition to simple resynthesis of preanalyzed recorded speech, the MISS system can apply procedures to modify the pitch, duration and amplitude of individual words so that they can be concatenated into natural sounding utterances. Text analysis programs provide the linguistic parameters that direct MISS to apply the appropriate manipulations. We evaluate the quality of reproduction and storage requirements for this system in comparison with other systems.