Real-time formant speech synthesis using the TMS320C25

Abstract

Applications of speech synthesis systems depend on high synthetic speech quality, real-time operation and system cost. This paper focuses on the real-time implementation of an improved quality formant synthesis system by using the Texas Instruments TMS320C25 digital signal processor. Owing to using the phoneme as the basic synthesis unit, the required speech database capacity is fairly low, namely, 15 kbytes. Thus, the system cost is decreased rapidly.<>