The Key Issues In Uyghur Emotional Speech Synthesis System

As a human voice to communicate the most natural, the most direct and mostconvenient form. Human-computer interaction is mankind for a long time, but also theultimate goal of speech synthesis technology. Now speech synthesis technology hasreached the written language into spoken English, but how can the machinesynthesized voice express human emotion is a problem of speech synthesis research.In this paper, a systematic and deep study has been done about emotional expressionof Uyghur language.this paper has established Uyghur emotional model and Uyghuremotional speech database according to characteristics of Uyghur language. Thus, Ithas set a foundation for emotional speech information processing research in the fieldof Uyghur and other related processing research.Firstly, the emotional speech synthesis technology research status at home andabroad and the research background and significance, followed by several speechsynthesis technology at present are analyzed and compared, describes the process ofestablishing emotional Uyghur corpus. This paper firstly selected in the Uyghurlanguage happy, angry, sad, surprised, calm and other5kind of emotions as thesynthesis of the target emotion, then collect the large of text containing emotion,from the selected representative including all emotion categories (this paper screeneda total of1200sentences)to collect sentences were recorded and manual annotation.Secondly, summarizes the characteristics of each type of emotional prosody: Byusing the STRAIGHT parameter analysis and synthesis algorithm for the extractionand analysis of acoustic parameters of emotional speech calm voice and each goal,Mean while, through the classication and regression tree(CART)algorithm is used toconstruct the prosody prediction model, the prediction model to forecast the varioustarget emotional voice acoustic parameters, and the predicted values were analyzed.Finally, this paper implements a complete Uyghur emotional speech synthesisplat form. all kinds of emotional speech and the synthesis of artificial auditory test, ofthe system to do the emotional speech synthesis experiments verify the feasibility andeffectiveness of the system.