Modelling the voice production process using a nonparametric approach - Archive ouverte HAL Access content directly
Conference Papers Year : 2009

Modelling the voice production process using a nonparametric approach

E. Cataldo
  • Function : Author
  • PersonId : 885597
R. Sampaio
  • Function : Author
  • PersonId : 864067
J. Lucero
  • Function : Author


The aim of this work is to discuss model uncertainties, in the case of the biomechanics of phonation. A number of mechanical models of voice production have been proposed in past years; but, in general, they have a deterministic nature. In previous works, data uncertainties were incorporated to a two-mass model of the vocal folds to perform a probabilistic analysis of the fundamental frequency of the oscillation. The present work is intended as a follow-up, to explore the effect of model uncertainties, using a non-parametric probabilistic approach based on the use of a probabilistic model for symmetric positive-definite real random matrices applying the Maximum Entropy Principle. The theory is discussed and numerical examples are presented to show that the predictability of the model may be improved when model uncertainties are taken into account. It is shown that some realizations of the output radiated pressure obtained are similar with samples of voice signals obtained from people with some pathologies, as nodulus and papilomas, in their vocal folds.
Fichier principal
Vignette du fichier
conference-2009-COMPDYN-Rhodes-cataldo-sampaio-soize-lucero-preprint.pdf (384.28 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-00692800 , version 1 (01-05-2012)


  • HAL Id : hal-00692800 , version 1


E. Cataldo, R. Sampaio, Christian Soize, J. Lucero. Modelling the voice production process using a nonparametric approach. COMPDYN 2009, ECCOMAS Thematic Conference on Computational Methods in Structural Dynamics and Earthquake Engineering, Jun 2009, Rhodes, Greece, Greece. pp.1-12. ⟨hal-00692800⟩
237 View
66 Download


Gmail Facebook Twitter LinkedIn More