ASCII based transcription systems for languages with the Arabic script: the case of Persian

We discuss transcription systems needed for automated spoken language processing applications in languages such as Persian that use the Arabic script for writing. The work is described in the context of a speech-to-speech translation system development for English and Persian. This system can easily be modified for Arabic, Dari, Urdu and any other language that uses the Arabic script. The proposed system has two components. One is a phonemic based transcription of sounds for acoustic modeling in automatic speech recognizers and for text-to-speech synthesizers, using ASCII based symbols, rather than International Phonetic Alphabet symbols. The other is a hybrid system, that provides a minimally-ambiguous lexical representation that explicitly includes vocalic information; such a representation is needed for language modeling and machine translation.