Natural language processing for Urdu TTS system
H. Kabir · 2002
Natural Language Processing (NLP) provides the foundation for every text-to-speech system. NLP caters all the issues related to the language. NLP for Urdu Textto-Speech System is a part of research project being done at CRULP, NUCES, Lahore. Its objective is to generate phonetic stream from plain Urdu text for the production of natural and intelligible speech. To carry out this transformation from text to speech, the NLP performs a series of steps. These steps are tokenization of input text, generation of phonemic stream, syllabification of words present in text, primary stress assignment on words, application of phonological rules and resyllabification of words. This paper focuses on all these processes from the perspective of Urdu.