Constructing concatenating speech synthesis system using realistic data from audiobook

• Main Author: Tham, Yew Loong
• Other Authors: Chng Eng Siong
• Format: Final Year Project
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Computer science and engineering
• Online Access: http://hdl.handle.net/10356/62588
• Institution: Nanyang Technological University

In the past few decades, technology has advanced and made human lifestyles more convenient. One example is the Speech Synthesis System. It not only brings convenience to users, it also plays a substantially important role for user to learn foreign languages as well as provides assistance to speak on behalf of speech impaired people. These systems are generally designed for languages such as English, French, German, Mandarin, Japanese, etc. However, there are limitations to how this system can pronounce the synthesized speech. Many researchers around the world are trying to overcome these limitations. Therefore, it leads to many different forms of developments in the process of sentence constructions. For English sentences, it is made up of combinations of words. In turn these words are made up of phonemes. In terms of speech, each word is pronounced from different syllables. This project will develop the basic Speech Synthesis System. The experimental results will demonstrate the speech constructing and concatenating using the data from an Obama audiobook.