7th International Workshop on Models and Analysis of Vocal Emissions for Biomedical Applications (MAVEBA 2011)

Florence, Italy
August 25-27, 2011

Parameter Optimization for a Time-Dependent Multi-Mass Model for the Pharyngo-Esophageal Segment

B. Hüttner, Michael Döllinger, Georg Luegmair, Ulrich Eysholdt, A. Ziethe, E. Gürlek

Department of Phoniatrics & Pediatric Audiology, University Hospital Erlangen, Erlangen, Germany

Laryngeal cancer may necessitate a complete removal of the larynx. Consequently, the required sound source for voiced communication is lost. Alternatively, a substitute sound signal can be generated by tissue vibrations in the pharyngoesophageal (PE) segment. The quality of the substitute voice significantly depends on the vibration characteristics of the PE segment. For investigation purpose, the tissue vibrations are detected by endoscopic interventions with a high-speed camera and are quantified by a biomechanical multi-mass model. As pseudoglottal vibrations present variations in frequency and amplitude we suggest an expansion of the biomechanical PE model for a time-dependent multi-mass model. Additionally, we propose a block based optimization procedure to fit the model dynamics to real PE vibrations. First results demonstrate the performance of the time-dependent model and the optimization procedure being comparable to those of the non-stationary PE model and the timedependent vocal fold model.

Index Terms. PE segment, model optimization

Full Paper (reprinted with permission from Firenze University Press)

Bibliographic reference.  Hüttner, B. / Döllinger, Michael / Luegmair, Georg / Eysholdt, Ulrich / Ziethe, A. / Gürlek, E. (2011): "Parameter optimization for a time-dependent multi-mass model for the pharyngo-esophageal segment", In MAVEBA-2011, 49-52.