Theory, Methodology and Implementation of The Malay Text-To-Speech System

摘要

This paper deals with the improvements and modifications done on the first edition of the Malay Language Text-to-Speech System, SUM (acronym for Sintesis Ucapan Melayu). A simple review on human speech production system and synthetic speech production system is discussed. Modifications on the voicing source include an additional KLGLOTT88 model added into the software. Theory and methodology on the voicing source are discussed in this paper. Characteristics of natural speech such as breathiness and flutter are studied. In addition, more constructive rules are added to improve the capability of the previous Malay TTS SUM system. The synthesiser model is based on the Klatt’s formant synthesiser, KLSYN88.