In a text-to-speech system, emotional expression is one of the most important factors for speech synthesis processing. There are many research projects that have attempted to develop a system which can read text, such as a novel, with emotion. Emotional expression makes a story come alive and catches the listeners' attention. Research on emotional expression usage in a Thai language text-to-speech synthesis has never been done before. This paper studies the classification of Thai emotional words by using interjection words defined in the Thai Royal Institute Dictionary (1999). This study makes a comparison between the words defined in the dictionary and a survey of annotators' opinion. The results show that the annotators' opinion of the emotional expression of the interjection words is significantly different from that defined in dictionary. Furthermore, these results indicate that there are many factors that have an effect on emotional expression such as sex, and age. Therefore, these factors should be utilized to identify emotional expressions in text for improving the efficiency of Thai text-to-speech development in the future.
展开▼
