The Jess Blizzard Challenge 2006 Entry
Peter J. Cahill, Julie Carson-Berndsen · 2006
This paper describes the version of the Jess system that participated in the Blizzard Challenge 2006. The Jess system consists of a suite of software tools for processing text and speech. The largest component of the system is a multi-platform unit selection speech synthesiser that uses Unicode and the International Phonetic Alphabet (IPA). The system has been designed to be modular so that different synthesiser algorithms can be implemented in a single instance of the system, allowing for alternative techniques to be compared with the same input and target data. In this paper we discuss the algorithms used in the Jess synthesiser to produce speech, insights gained from participation in the Blizzard Challenge 2006, and our intended areas of future work.