A combined approach to the polysemy problems in a Chinese to Taiwanese TTS system

This paper proposes a combined approach to the polysemy problems in a Chinese to Taiwanese text-to-speech (TTS) system. Polysemy means there are words with more than one meaning or pronunciation. For example, there are two kinds of pronunciation for the word (he) in Taiwanese. They are /yi7/ and /yin7/. The first pronunciation, /yi7/, can mean `he´ or `him´; and the other one, /yin7/, means `his´. The correct pronunciation of a word affects the comprehensibility (or clarity) and fluency of Taiwanese speech. We had shown that using language models to solve the polysemy problems can have very good results. We propose a combined approach to the polysemy problems in this paper. There are six words with polysemy problems to be solved in this paper. They are (you), (I), (he), (no), (up), and (down). The numbers of pronunciation of these six words are 2, 2, 2, 6, 3, and 4, respectively. Results show that the proposed combined approach can achieve higher accuracy than the existing methods, WU and DLC.