Abstract
Digital speech synthesis will get more and more important in the new world of information. This paper aims to discuss the techniques of the digital synthesis of Mandarin speech by manipulating its characteristics to decrease the necessary bit rate (or size of memory), or to simplify the process of synthesis, and ultimately to make the synthesis of Mandarin more efficient. The current work focuses on the utilization of the four‐tone behavior of Mandarin. The results have demonstrated that the four‐tone characteristics can be properly used to significantly improve the techniques of Mandarin speech synthesis.