An Automatic Tool for High Quality Arabic Speech Synthesis

Speech synthesis TTS (text-to-speech) is the process of converting the written text into machine generated synthetic speech. Concatenative speech synthesis systems render speech by concatenating pre-recorded speech units. This work describes the Arabic TTS synthesis system. This system uses an automatic tool based on concatenation of the Arabic diphone with MBROLA synthesizer. The quality of a synthesized speech is improved by analyzing the spectrum features of voice source in various F0 ranges and timbres in detail. It generates speech synthesis based on estimation and optimization of the Arabic prosody by classifying the voice source into different types. The developed model enhances the quality of the naturalness, and the intelligibility of speech synthesis in various speaking environment.