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

Mixed Physical Modeling Techniques Applied to Speech Production

Matti Karjalainen

Helsinki University of Technology, Finland

The Kelly-Lochbaum transmission-line model of the vocal tract started the discrete-time modeling of speech production. More recently similar techniques have been developed in computer music towards a more generalized methodology. In this paper we will study the application of mixed physical modeling to speech production and speech synthesis. These approaches are Digital Waveguides (DWG), Finite Difference Time-Domain schemes (FDTD), and Wave Digital Filters (WDF). The equivalence and interconnectivity of these schemes is shown and flexible real-time synthesizers for articulatory type of speech production are demonstrated.

Full Paper

Bibliographic reference.  Karjalainen, Matti (2003): "Mixed physical modeling techniques applied to speech production", In EUROSPEECH-2003, 2445-2448.