14th International Congress of Phonetic Sciences (ICPhS-14)

San Francisco, CA, USA
August 1-7, 1999

Use of Articulatory and Spectral Information for Speech Training

Marie-Christine Haton, Jean-Paul Haton

LORIA, Université Henri-Poincaré, Nancy 1, France

The improvement of speech recognition and speech analysis techniques favor the development of tools for various applications related to speech learning: pathological voice assessment, aid to diagnosis, aid to oral communication, aid to speech perception. We arc dealing in this paper with speech education (and reeducation), and with foreign language teaching.
   Based on our experience gained in the development of the SIRENE 2 system, we present the different aspects of the ongoing work concerning speech analysis and feature extraction, automatic guidance of a trainee, and performance assessment. Emphasis is given to the importance of acoustic-phonetic analysis at different abstraction levels. These points arc illustrated with practical examples.

Full Paper

Bibliographic reference.  Haton, Marie-Christine / Haton, Jean-Paul (1999): "Use of articulatory and spectral information for speech training", In ICPhS-14, 793-796.