Design of a syllable based Bengali Text-to-Speech System

This paper describes the design of a syllable based Text-to-Speech for Bengali language. We divide our system into four different phases-Texts Normalization, Syllable Detection, Syllable Selection and Sound File Collection. In Text Normalization phase, we will split words and mark their type. Different rules of pronunciation to detect syllable will be applied in syllable detection phase. In Syllable Selection phase, we will split each syllable from normalize text according to a remark and search them into database. Finally, in the Sound File Collection phase, we will collect sound file corresponding with syllable and concatenate them to produce speech. We hope that our system will perform better as we need few concatenation points to produce a word than any other Bengali Text-to-Speech synthesis system. KeywordsText Normalization, Syllable Detection, Syllable Selection, Sound File Collection