The project involved several studies involving synthesis of intonation in text-to-speech systems. A number of algorithms were developed for preprocessing texts in order to extract contextual information (e.g. coreference) as well as lexico-syntactic information (e.g. word class) that trigger the assignment of different kinds of pitch accents and determine their relative levels of prominence.
|Effective start/end date||1992/01/01 → 1996/12/31|