A Comparison of Unit Selection Techniques in Limited Domain Speech Synthesis

Warning

This publication doesn't include Faculty of Sports Studies. It includes Faculty of Informatics. Official publication website can be found on muni.cz.
Authors

GAURA Pavel BATŮŠEK Robert

Year of publication 2003
Type Article in Proceedings
Conference Proceedings of TSD 2003
MU Faculty or unit

Faculty of Informatics

Citation
Field Informatics
Keywords speech synthesis; corpus; unit selection; segmentation; processing
Description Traditional concatenative speech synthesizers equipped with a small amount of speech segments suffer from the lack of naturalness. On the other hand, corpus-based speech synthesizers are able to produce much more natural speech. This paper presents a comparison of two new unit-selection methods in the corpus-based speech synthesis. An experimental comparison of comprehensibility and naturalness of all three approaches is provided here. The results are compared with one widely-used unit-selection method.
Related projects:

You are running an old browser version. We recommend updating your browser to its latest version.

More info