Second European Conference on Speech Communication and Technology

Genova, Italy
September 24-26, 1991


Automatic Diphone Segmentation

P. A. Taylor, Stephen D. Isard

Centre for Speech Technology Research, University of Edinburgh, UK

A method is described for automatically segmenting a database of diphones for speech synthesis purposes. This involves using hidden Markov models to find phoneme boundaries and then a spectral mismatch minimisation algorithm to choose the precise diphone boundaries.

Full Paper

Bibliographic reference.  Taylor, P. A. / Isard, Stephen D. (1991): "Automatic diphone segmentation", In EUROSPEECH-1991, 709-711.