An efficient speech generation method based on character and modifier of Bangla PDF Document

Abstract

In this paper we present a method for Bangla speech generation from Bangla PDF document. Our main goal is to generate almost natural speech from Bangla PDF document. For this we have proposed a method which performs three major tasks. One is PDF to text conversion, then text to ASCII conversion, and then follows the character and modifier rules while reading text and finally speech generation from the concatenate sound files of text. We have analyzed the proposed method and the existing methods with respect to error rate of reading all the Bangla words and also the time complexity.