EUROSPEECH 2003 - INTERSPEECH 2003 8th European Conference on Speech Communication and Technology Geneva, Switzerland September 1-4, 2003

Evaluating and Correcting Phoneme Segmentation for Unit Selection Synthesis

John Kominek, Christina L. Bennett, Alan W. Black

Carnegie Mellon University, USA

As part of improved support for building unit selection voices, the Festival speech synthesis system now includes two algorithms for automatic labeling of wavefile data. The two methods are based on dynamic time warping and HMM-based acoustic modeling. Our experiments show that DTW is more accurate 70% of the time, but is also more prone to gross labeling errors. HMM modeling exhibits a systematic bias of 15 ms. Combining both methods directs human labelers towards data most likely to be problematic.

Full Paper

Bibliographic reference.  Kominek, John / Bennett, Christina L. / Black, Alan W. (2003): "Evaluating and correcting phoneme segmentation for unit selection synthesis", In EUROSPEECH-2003, 313-316.