A dictionary-based approach for Thai Text to Speech (TTTS)

Abstract

The idea of text to speech by a computer is an enhancement of the human learning ability. Due to the fact that each person has individual ability of visualization, the receiving of information in the form of voice helps make everything become easier. The objective of this research is to develop computer software that can translate Thai Text to Speech (TTTS). The TTTS consists of four modules, which are 1) Input Thai text file, 2) Text processing, 3) Dictionary Lookup, and 4) Speech synthesis. The TTTS dictionary contains more than 5,300 Thai primitive words and their sound waves. The experiment is conducted on 30 Thai text files, which contain more than 3200 Thai words. The reading precision of the TTTS is around 74.05 percent, with the reading speech of 0.88 second per word.