AMHARIC SPEECH SYNTHESIS SYSTEM AND ITS APPLICATIONS TO MULTIMEDIA AND TELECOMMUNICATIONS

摘要

This paper presents the design of a syllabic based Text-to-Speech (TTS) synthesizer for Amharic language and discusses the possible application areas of the synthesizer, mainly in multimedia and telecommunications. Amharic is the official language of Ethiopia. It is one of the widely spoken language and Semitic languages having its own script. The designed Amharic Text-to-Speech (AmhTTS) system is parametric and rule-based system that uses a source filter model for speech production and a Log Magnitude Approximation (LMA) filter as the vocal tract filter. As far as we have the information, our system is the first syllabic based TTS system designed for Amharic language. The system was evaluated by syllable and word listening test and a reasonable result of 89% syllable and 98% word correct-rates were achieved. The paper also summarizes the important issues to be considered to improve the intelligibility and naturalness of the synthesizer for our next version.