8th International Conference on Spoken Language Processing

Jeju Island, Korea
October 4-8, 2004

Speech Production Based on Lossy Tube Models: Unit Concatenation and Sound Transitions

Karl Schnell, Arild Lacroix

Goethe-University Frankfurt am Main, Germany

Discrete time tube models describe the propagation of plane sound waves through the vocal tract. Therefore they are important for speech analysis and production. The standard lossless tube model is extended by introducing distributed frequency dependent losses. In comparison to a previous investigation it is shown how the lossy vocal tract model can be used for speech production. The vocal tract areas of the lossy model are estimated from speech signals by an optimization algorithm. With the aid of the obtained parameters resynthesis can be successfully performed. For a synthesis of new utterances analyzed diphones are concatenated by linear transitions of the vocal tract areas. Furthermore for a generation of sound transitions also a nonlinear area transition is discussed considering horizontal area movements and a component of a neutral vocal tract configuration.

Full Paper

Bibliographic reference.  Schnell, Karl / Lacroix, Arild (2004): "Speech production based on lossy tube models: unit concatenation and sound transitions", In INTERSPEECH-2004, 505-508.