
Influence of tissue changes in superficial lamina propria on production of Czech vowels
Hájek, P. ; Švancara, P. ; Horáček, Jaromír ; Švec, J.
Superficial lamina propria (SLP) is a waterlike vocal fold (VF) layer located directly under overlying epithelium. Its material properties affect VF motion and thus resulting spectrum of produced sound. Influence of stiffness and damping of the SLP on sound spectrum of Czech vowels is examined using a twodimensional (2D) finite element (FE) model of a human phonation system. The model consists of the VF (structure model) connected with an idealized trachea and vocal tract (VT) (fluid models). Five VTs for all Czech vowels [a:], [e:], [i:], [o:] and [u:] were used and their geometry were based on MRI data. Fluid flow in the trachea and VT was modelled by unsteady viscous compressible NavierStokes equations. Such a formulation enabled numerical simulation of a fluidstructureacoustic interaction (FSAI). Selfsustained oscillations of the VF were described by a momentum equation including large deformations and a homogeneous linear elastic model of material was used. Fluid and structure solvers exchange displacements and boundary forces in each iteration. During closed phase VFs are in contact and fluid flow is separated. We can observe that both the damping and the stiffness of the SLP substantially influence the amplitude and frequency of VFs vibration as well as the open time of the glottis.\n


Experimental and computer modelling study of glottal closing velocity during phonation
Horáček, Jaromír ; Radolf, Vojtěch ; Bula, Vítězslav ; Šidlof, P. ; Geneid, A. ; Laukkanen, A. M.
This preliminary study shows that the impact stress between the colliding vocal folds during phonation should not be evaluated from the maximum velocity of the glottal closing because the velocity of the closing diminishes just before the glottal closure. This phenomenon, which can be caused by a pressure cushion effect in the fast narrowing glottal gap, is demonstrated with measurements from high speed camera images recorded from human and on a physical laboratory model for vowel [u:] phonation and on a threemass computer model employing a Hertz model of impact force. For a more detailed future study of this phenomenon a faster camera has to be used. \n


Aerodynamic transfer of energy to vibrating vocal folds for different driving mechanisms
Valášek, J. ; Sváček, P. ; Horáček, Jaromír
This paper studies the mutual energy transfer between the fluid flow, described by incompressible NavierStokes equations, and the elastic body represented by vocal folds. The aerodynamic energy transfer function describes the amount and more importantly the sign of the energy exchange. It determines if the vocal fold vibrations are selfexcited or prescribed.The energy transfer function is studied for three different driving mechanisms introduced by different inlet boundary conditions (BC). The most frequently used inlet BCs for incompressible model of fluid flow approximated by the finite element method are either Dirichlet BC giving the inlet velocity or donothing type of BC prescribing the pressure difference between the inlet and the outlet. Since the numerical simulations with both aforementioned BCs do not provide results observed experimentally the newly introduced BC based on the penalization approach seems as remedy. The numerical model consists of strongly coupled partitioned scheme based on the stabilized finite element method.\n\n

 
 

Numerical investigation of acoustic characteristics of 3D human vocal tract model with nasal cavities
Vampola, T. ; Štorkán, J. ; Horáček, Jaromír ; Radolf, Vojtěch
Acoustic resonance characteristics of 3D human vocal tract model without and with nasal and\nparanasal cavities were computed. Nasal cavities (NC) form the side branches of the human vocal tract and exhibit antiresonance and resonance properties which influence the produced voice quality. Developed FE models of acoustic spaces of nasal and vocal tract for vowel /a:/ are used to study the influence of (NC) on phonation. Acoustics frequencymodal characteristics are studied by modal analysis and numerical simulation of acoustic signals in time domain is performed by transient analysis of the FE models.


The Influence of Different Geometries of Human Vocal Tract Model on Resonant Frequencies
Valášek, Jan ; Sváček, Petr ; Horáček, Jaromír
This paper presents the transfer function approach in order to determine the acoustic resonant frequencies of a human vocal tract model. The transfer function is introduced here as an acoustic pressure ratio between input amplitude at glottis position and output amplitude at mouth opening given by the solution of Helmholtz equation. This equation is numerically approximated by finite element method. The influence of different boundary conditions are studied and also different locations of excitation and microphone. Four different vocal tract geometries motivated by vocal tract geometry for vowel [u:] are investigated. Its acoustic resonance frequencies in range 100  2500 Hz are computed and compared with published results. Further, the transient acoustic computation with different acoustic analogies are performed. The frequency spectra of Lighthill analogy, acoustic wave equation and perturbed convective wave equation are compared, where the vocal tract model with best frequency agreement with published results was chosen. The dominant frequencies correspond with predicted frequencies of transfer function approach.\n


Computationally Efficient Model of the Human Vocal Fold
Štorkán, J. ; Vampola, T. ; Horáček, Jaromír
One mass model of the vocal folds with three degrees of freedom in 2D space was created and used to simulate the movement of the vocal folds. Vocal folds are modeled as a solid mass stored flexibly in 2D. The model is excited by aerodynamic forces. The flow is solved by analytical model incompressible and nonviscous fluid with constant flow. In case of close of the glottis are aerodynamic forces replaced by Hertz model of the contact forces. Movement equations are solved by numerical method. The model allows to solve the movement of the vocal folds in the time domain, pressure field acting on the vocal folds or contact pressures.


Effect of turbulence in FE model of human vocal folds selfoscillation
Hájek, P. ; Švancara, Pavel ; Horáček, Jaromír ; Švec, J.G.
The purpose of the study is to determine whether a turbulence model in fluid flow calculation affects the vocal folds (VF) vibration and the acoustics of human vocal tract (VT). The objective is examined using a twodimensional (2D) finite element (FE) model of the fluidstructureacoustic interaction for selfsustained oscillations of the VF. The FE model consists of the models of the VF, the trachea and a simplified model of the human VT. The developed FE model includes large deformations of the VF tissue and VF contact interrupting the airflow during glottis closure. The airflow is modelled by the unsteady viscous compressible NavierStokes equations, without and with the Shear Stress Transport (SST) turbulence model. Fluidstructure interaction (FSI) and morphing of the fluid mesh are realized using Arbitrary LagrangianEulerian (ALE) approach. The method is applied on the FE model of the VT shaped for the Czech vowel [a:]. Also effect of varying stiffness of the superficial lamina propria (SLP) is analyzed. The numerical simulations showed that considering of the turbulence affects mainly higher frequencies apparent in a frequency spectrum of the VT acoustics.


Influence of the nasal cavities to human voice quality
Vampola, T. ; Horáček, Jaromír
Nasal cavities (NC) form the side branches of the human vocal tract and exhibit antiresonance and resonance properties which influence the produced voice quality. This study investigates the possibility of these resonances to contribute to the speaker's or singer's formant cluster around 3  5 kHz. A reduced finite element (FE) model was created which allows numerical simulation of the effects of changing the volumes of NC on the acoustic resonance and antiresonance characteristics of the vocal tract. This model, created from an accurate threedimensional (3D) FE model of the human vocal tract for vowel [a:] and [i:] is computationallyeffective and allows parametric changes of the volume connecting the nasal tract with the human vocal tract. Developed FE models of acoustic spaces of nasal and vocal tract for vowels /a:/ and /i:/ are used to study the influence of (NC) on phonation of these vowels. Acoustics frequencymodal characteristics are studied by modal analysis and numerical simulation of acoustic signals in time domain is performed by transient analysis of the FE models.
