Design and Development of an Arabic Text-To-Speech Synthesizer

Abstract

Computers that can interact with humans via speech had become a dream for scientists since the early stages of the computer age. Scientists from different areas have been interested in producing human speech artificially for more than two decades and they made a lot of research to reach this dream. Many advances in Text To Speech (TTS) synthesis systems have been achieved in languages like English and French. Languages like Arabic have been largely ignored. This paper presents a TTS system for Arabic that uses allophone/diphone concatenation method. It takes a text as input and produces corresponding speech in Arabic. In this system, the output is available in one male voice only. Since Arabic is a verbal language, the TTS engine developed can be used also for other verbal languages with some minor modifications involving the specific language.