The ESCA Workshop on Speech Synthesis

September 25-28, 1990
Autrans, France

Acoustic-Articulatory Inversion Based on a Neural Controller of a Vocal Tract Model

Alain Soquet, Marco Saerens, Paul Jospa

Institut de Phonétique and IRIDIA, University Libre de Bruxelles, Belgium

We propose to use the qualitative knowledge provided by the Distinctive Region and Mode theory (Mrayati, Carré & Guérin, 1988), coupled with a neural controller (Saerens & Soquet, 1989a), to produce an acoustic-articulatory inversion of an eight-region vocal tract. This paper presents preliminary results for vowels.


Full Paper

Bibliographic reference.  Soquet, Alain / Saerens, Marco / Jospa, Paul (1990): "Acoustic-articulatory inversion based on a neural controller of a vocal tract model", In SSW1-1990, 71-74.