Toward an Arabic text-to-speech system

الملخص EN

This paper des€r؛bes an experimental system for algorithmic based segmental Arabic text-t()-speech system.

The system utilizes a comprehensive set of Arabic allophones for speech synthesis in accordance with b©th Modern Standard Arabic and the classical Arabic language.

The proposed all©ph©ne set contains about 150 Arabic allophones and v©wel/c©ns©nant combinations to simplify computer voice production.

The paper also describes an extensive set of Arabic letter-t©-s©und rules from voweled text (with full diacritics) based on this all©ph©ne set.

This set of rules governs €©n$©nant/v©wel combinations, usage of short and long vowels, the coarticulation effects of emphatics and pharyngeals, pronunciation of the particu-larization, rules for word final aspiration, Tartween and Ghonna rules, rules for conversion, suppression, and assimilation, and rules for combining words.