Embedded Unit Selection with Limited Data Base for Tamil Text-To-Speech Synthesis

This paper describes a modular, related to the development of speech synthesis system for Tamil language. We propose a new design and development of the functional architecture of unit selection Tamil TTS engine using unit selection techniques with limited resources such as mobile devices, robotics, and household devices with no considerable speech quality degradation. The proposed model implemented in FPGAs to optimizes the parameter extraction to perform efficient speech synthesis.