Third International Workshop on Models and Analysis of Vocal Emissions for Biomedical Applications (MAVEBA 2003)

Florence, Italy
December 10-12, 2003

Poincaré Sections for Pitch Mark Determination in Dysphonic Speech

Martin Hagmüller, Gernot Kubin

Institute of Communications and Wave Propagation, Graz University of Technology, Graz, Austria

In this paper a Poincaré approach to pitch mark determination is presented. While speech has been interpreted in terms of nonlinear systems theory for quite some time, not much effort has been made to exploit this knowledge in the problem of pitch mark detection. This algorithm uses nonlinear state space embedding and calculates the Poincaré section at a chosen point in state space, pitch-marks are then found at the crossing of the trajectories with the Poincaré plane. The procedure is performed framewise to account for the changing dynamics of the speech production system. First results show promising performance, comparable to the pitch marking algorithm used in 'Praat', and outperforming it in case of irregular voices.

Index Terms. Dysphonic speech, state-space-embedding, Poincaré section, pitch-marks

Full Paper (reprinted with permission from Firenze University Press)

Bibliographic reference.  Hagmüller, Martin / Kubin, Gernot (2003): "Poincaré sections for pitch mark determination in dysphonic speech", In MAVEBA-2003, 281-284.