Realization And Improvement Of Chinese Emotional Speech Synthesis Based On HMM

With the development of information technology,the application of intelligent mobile devices has become more and more popular.People have put forward more diverse requirements for human-machine speech interaction.This article aims to develop an application system to study emotional speech synthesis based on Hidden Markov Models and explore ways to improve the naturalness and emotional expression of speech synthesis.The main research work and research results are as follows:1.In order to achieve the experimental content,we construct emotional voice database.For Chinese phonetic sound,we selected,collected,and entered four different emotional voices and corresponding textual data.Based on these things,through the emotional voice database,we can do further study on the characteristics and relationship among the various emotional voice.2.The researcher constructs an HMM-based emotional speech synthesis system and describe the algorithm of a decision-tree model which enable computers to perform emotional speech synthesis automatically.Besides,update the corpus for further training and explain the process of it.3.In order to improve the emotional color of emotional speech synthesis,the researcher has improved the original emotional speech synthesis considering the features of Chinese emotional speech expression and realized the adjustment of emotional level in synthetic speech.The experimental results show that the improved Chinese emotional speech synthesis system has significantly improved the effect of emotional expression compared with the baseline system,achieving the expected research goals.