Study of Speech Synthesis Systems Towards Implementation of a Multilingual Text-To-Speech System

Location

CSU

Student's Major

Computer Information Science

Student's College

Science, Engineering and Technology

Mentor's Name

Mahbubur Rahman Syed

Mentor's Department

Computer Information Science

Mentor's College

Science, Engineering and Technology

Description

This work researches the state-of-the-art Text-To-Speech (TTS) systems through a comprehensive study of the technological evolution of different TTS systems. We also studied several speech synthesis approaches, available tools for TTS systems with an objective to identify their applicability towards development of a Bangla TTS system. To our knowledge very little or no work has been done in this area. This work is organized in two phases. First, a comparative survey of different TTS systems was conducted to evaluate their potential to be used for a Bangla TTS system and therefore approaching a multi-lingual ITS system. It is hypothesized that the use of existing TTS systems in other languages cannot be used for implementation of a Bangla System. Several distinguishing characteristics of Bangla sound were identified that make it necessary to build a Bangla TTS framework. In the second phase, the development of a preliminary framework is underway. Initial experiments are being conducted to test the output quality by simple concatenation of elementary speech units defined by us.

This document is currently not available here.

Share

COinS
 

Study of Speech Synthesis Systems Towards Implementation of a Multilingual Text-To-Speech System

CSU

This work researches the state-of-the-art Text-To-Speech (TTS) systems through a comprehensive study of the technological evolution of different TTS systems. We also studied several speech synthesis approaches, available tools for TTS systems with an objective to identify their applicability towards development of a Bangla TTS system. To our knowledge very little or no work has been done in this area. This work is organized in two phases. First, a comparative survey of different TTS systems was conducted to evaluate their potential to be used for a Bangla TTS system and therefore approaching a multi-lingual ITS system. It is hypothesized that the use of existing TTS systems in other languages cannot be used for implementation of a Bangla System. Several distinguishing characteristics of Bangla sound were identified that make it necessary to build a Bangla TTS framework. In the second phase, the development of a preliminary framework is underway. Initial experiments are being conducted to test the output quality by simple concatenation of elementary speech units defined by us.