Publication detail

Three-dimensional Numerical Analysis of Czech Vowel Production

HÁJEK, P. ŠVANCARA, P. HORÁČEK, J. ŠVEC, J.

Original Title

Three-dimensional Numerical Analysis of Czech Vowel Production

English Title

Three-dimensional Numerical Analysis of Czech Vowel Production

Language

en

Original Abstract

Spatial air pressures generated in human vocal tract by vibrating vocal folds present sound sources of vowel production. This paper simulates phonation phenomena by using fluid-structure-acoustic scheme in a three-dimensional (3D) finite element model of a Czech vowel [o:]. The computational model was composed of four-layered M5-shaped vocal folds together with an idealized trachea and vocal tract. Spatial fluid flow in the trachea and in the vocal tract was obtained by unsteady viscous compressible Navier-Stokes equations. The oscillating vocal folds were modelled by a momentum equation. Large deformations were allowed. Transient analysis was performed based on separate structure and fluid solvers, which were exchanging loads acting on the vocal folds boundaries in each time iteration. The deformation of the fluid mesh during the vocal fold oscillation was realized by the arbitrary Lagrangian-Eulerian approach and by interpolation of fluid results on the deformed fluid mesh. Preliminary results show vibration characteristics of the vocal folds, which correspond to those obtained from human phonation at higher pitch. The vocal folds were self-oscillating at a reasonable frequency of 180 Hz. The vocal tract eigenfrequencies were in the ranges of the formant frequencies of Czech vowel [o:] measured on humans; during self-oscillations the formants shifted to lower frequencies.

English abstract

Spatial air pressures generated in human vocal tract by vibrating vocal folds present sound sources of vowel production. This paper simulates phonation phenomena by using fluid-structure-acoustic scheme in a three-dimensional (3D) finite element model of a Czech vowel [o:]. The computational model was composed of four-layered M5-shaped vocal folds together with an idealized trachea and vocal tract. Spatial fluid flow in the trachea and in the vocal tract was obtained by unsteady viscous compressible Navier-Stokes equations. The oscillating vocal folds were modelled by a momentum equation. Large deformations were allowed. Transient analysis was performed based on separate structure and fluid solvers, which were exchanging loads acting on the vocal folds boundaries in each time iteration. The deformation of the fluid mesh during the vocal fold oscillation was realized by the arbitrary Lagrangian-Eulerian approach and by interpolation of fluid results on the deformed fluid mesh. Preliminary results show vibration characteristics of the vocal folds, which correspond to those obtained from human phonation at higher pitch. The vocal folds were self-oscillating at a reasonable frequency of 180 Hz. The vocal tract eigenfrequencies were in the ranges of the formant frequencies of Czech vowel [o:] measured on humans; during self-oscillations the formants shifted to lower frequencies.

Keywords

Biomechanics of voice; Compressible flow; Finite element method; Fluid-structure-acoustic interaction; Simulation of phonation

Released

13.01.2021

Publisher

Brno University of Technology, Institute of Solid Mechanics, Mechatronics and Biomechanics

Location

Brno

ISBN

978-80-214-5896-3

Book

Engineering mechanics 2020

Edition

26

Edition number

1

Pages from

182

Pages to

185

Pages count

4

URL

Documents

BibTex


@inproceedings{BUT167725,
  author="Petr {Hájek} and Pavel {Švancara} and Jaromír {Horáček} and Jan G. {Švec}",
  title="Three-dimensional Numerical Analysis of Czech Vowel Production",
  annote="Spatial air pressures generated in human vocal tract by vibrating vocal folds present sound sources of vowel production. This paper simulates phonation phenomena by using fluid-structure-acoustic scheme in a three-dimensional (3D) finite element model of a Czech vowel [o:]. The computational model was composed of four-layered M5-shaped vocal folds together with an idealized trachea and vocal tract. Spatial fluid flow in the trachea and in the vocal tract was obtained by unsteady viscous compressible Navier-Stokes equations. The oscillating vocal folds were modelled by a momentum equation. Large deformations were allowed. Transient analysis was performed based on separate structure and fluid solvers, which were exchanging loads acting on the vocal folds boundaries in each time iteration. The deformation of the fluid mesh during the vocal fold oscillation was realized by the arbitrary Lagrangian-Eulerian approach and by interpolation of fluid results on the deformed fluid mesh. Preliminary results show vibration characteristics of the vocal folds, which correspond to those obtained from human phonation at higher pitch. The vocal folds were self-oscillating at a reasonable frequency of 180 Hz. The vocal tract eigenfrequencies were in the ranges of the formant frequencies of Czech vowel [o:] measured on humans; during self-oscillations the formants shifted to lower frequencies.",
  address="Brno University of Technology, Institute of Solid Mechanics, Mechatronics and Biomechanics",
  booktitle="Engineering mechanics 2020",
  chapter="167725",
  doi="10.21495/5896-3-182",
  edition="26",
  howpublished="online",
  institution="Brno University of Technology, Institute of Solid Mechanics, Mechatronics and Biomechanics",
  number="1",
  year="2021",
  month="january",
  pages="182--185",
  publisher="Brno University of Technology, Institute of Solid Mechanics, Mechatronics and Biomechanics"
}