Please use this identifier to cite or link to this item:
Title: Text-to-speech synthesis
Authors: Tun Naung Lin.
Keywords: DRNTU::Engineering::Electrical and electronic engineering
Issue Date: 2006
Abstract: This project is a comprehensive introduction to Text-To-Speech (TTS) synthesis. Two areas of speech synthesis are explored. The first part concerns natural language processing and the inherent problems that are encountered in speech synthesis. The overall organization of Natural Language Processing (NLP) module in a TTS synthesizer is examined. Each component of the morpho-syntactic analyzer is functionally described and available solutions are browsed. A decision tree, called Classification and Regression Tree (CART) which automatically induce lexical-to-surface pronunciations mappings from a labeled corpus, is introduced for phonetization. The second part focuses on the digital signal processing techniques that can deliver the expected high-quality synthetic speech, with a special emphasis on concatenative approach. Time-Domain Pitch Synchronous Overlap Add (TDPSOLA) algorithm is extensively discussed for prosodic modification. Finally, a simple limited vocabulary Text-To-Speech synthesis system incorporation NLP component and DSP component is emulated. The synthesized speech demonstrates a significant level of intelligibility and a high level of naturalness.
Description: 69 p.
Schools: School of Electrical and Electronic Engineering 
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Theses

Files in This Item:
File Description SizeFormat 
  Restricted Access
5.86 MBAdobe PDFView/Open

Page view(s) 50

Updated on Jun 16, 2024


Updated on Jun 16, 2024

Google ScholarTM


Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.