Integration of Fuzzy If-Then Rule with Waveform Concatenation Technique for Text-to-Speech Synthesis in Odia

Abstract

This paper describes a novel approach to Text-to-Speech synthesis in Odia language which may be used as an application in small hand held devices, benefiting people to listen to different literatures or mythological documents available in Odia. The model uses a fuzzy if-then rule text classification technique to identify the type of units involved and is being integrated with a waveform concatenative technique to produce speech out of the input text in Odia. To analyze the performance of the proposed technique, we have considered different types of words in Odia with respect to the speech quality evaluation parameters and the results of our experiments shows the effectiveness of the technique in producing intelligible speech in Odia.