[2]. In the two-mass model [17],the lower mass is made thicker (vertical dimension in the coronal plane) and more massive than the upper mass in an effort to simulate the effects of the body layer. However,since this arrangement does not allow for a coupled oscillation of both the cover and body layers,the two-mass model is essentially a ‘‘cover’’ model rather than a ‘‘cover-body’’ model. It was this limitation that motivated the introduction of a three-mass model [18] that was intended to include the effect of the cover-body vocal fold structure but also maintain the simplicity of a low-dimensional system. As shown in Fig. 8,this model essentially adds a ‘‘body’’ mass that is positioned lateral to the two cover masses. The cover masses are both connected to the body mass via spring and damper systems that represent stiffness of the cover tissue as well as the effective coupling stiffness between the body and cover. The body mass,in turn,is connected to a rigid lateral boundary with another spring and damper system that account for the effective stiffness of the body which will depend on the level of contraction of the muscle tissue. Finally,to account for shear forces in the cover,the two cover masses are coupled to each other with another spring-damper element. In addition to coupled oscillation of the cover and body layers,the advantage of the threemass model is that that physiologically realistic control parameters characterizing the cover and body tissue are more easily determined (than with the two-mass approach) when the discretization imposed by the model more closely follows anatomical boundaries. For example,a contraction of the thyroartenoid muscle (muscle in the ‘‘body’’) will increase the stiffness of the body but may not necessarily stiffen the cover. Other more complex models have been developed to simulate the layered vocal fold structure and also account for the vibrational patterns in the transverse plane. For Fig. 7 A diagram of a two-mass vocal fold model [17]. Mucosal wave as well as lateral motion is represented. Fig. 8 The three-mass model [18] of the vocal folds. In addition to the mucosal wave and lateral motion,this model accounts for the coupling between the cover and body. B. H. STORY: SOUND SOURCE FOR VOWELS 199 example,the two-mass approach was modified such that a single vocal fold was represented by eight coupled transverse sections,each with two masses in the coronal plane that had both lateral and vertical degrees of freedom [19,20]. This model allows for simulation of the vibrational pattern of the vocal folds in both the coronal and traverse planes. A further step up in complexity came with a continuum mechanics model [21] and a more recent finiteelement model [22] both of which provide a precise physiological and mechanical representation of the tissue layers of the vocal folds without lumping large anatomical regions into a few mass elements. Their large number of degrees of freedom allows for detailed study of the complex vibratory pattern observed in human vocal folds. Interestingly,however,these complex models capable of producing many modal vibration patterns,have shown that vocal fold vibration is largely dominated by only 2–3 modes of vibrations [10],similar to movement patterns shown in Fig. 5. Thus,the lumped-element models seem to capture enough of the vibratory characteristics to still serve as a useful research tool if fine detail is unnecessary.
4. INTERACTION OF THE VOICE SOURCE AND THE VOCAL TRACT In this section a three-mass model [18] will be used to demonstrate self-oscillation of the vocal folds under the influence of several vocal tract configurations. Mathematical details concerning the mechanical aspects of the model can be found in the original publication [18] but the aerodynamic equations have been updated [23]. Table 1 gives the relevant parameter values that will be used throughout the following discussion unless otherwise stated. Note that important characteristics of phonation, such as fundamental frequency,voice quality,amplitude of vibration,may be altered by changing the parameters in Table 1. However,in keeping with the theme of this special issue concerning vowels,the following sections focus only on changing the vocal tract configuration.
[2]. In the two-mass model [17],the lower mass is made thicker (vertical dimension in the coronal plane) and more massive than the upper mass in an effort to simulate the effects of the body layer. However,since this arrangement does not allow for a coupled oscillation of both the cover and body layers,the two-mass model is essentially a ‘‘cover’’ model rather than a ‘‘cover-body’’ model. It was this limitation that motivated the introduction of a three-mass model [18] that was intended to include the effect of the cover-body vocal fold structure but also maintain the simplicity of a low-dimensional system. As shown in Fig. 8,this model essentially adds a ‘‘body’’ mass that is positioned lateral to the two cover masses. The cover masses are both connected to the body mass via spring and damper systems that represent stiffness of the cover tissue as well as the effective coupling stiffness between the body and cover. The body mass,in turn,is connected to a rigid lateral boundary with another spring and damper system that account for the effective stiffness of the body which will depend on the level of contraction of the muscle tissue. Finally,to account for shear forces in the cover,the two cover masses are coupled to each other with another spring-damper element. In addition to coupled oscillation of the cover and body layers,the advantage of the threemass model is that that physiologically realistic control parameters characterizing the cover and body tissue are more easily determined (than with the two-mass approach) when the discretization imposed by the model more closely follows anatomical boundaries. For example,a contraction of the thyroartenoid muscle (muscle in the ‘‘body’’) will increase the stiffness of the body but may not necessarily stiffen the cover. Other more complex models have been developed to simulate the layered vocal fold structure and also account for the vibrational patterns in the transverse plane. For Fig. 7 A diagram of a two-mass vocal fold model [17]. Mucosal wave as well as lateral motion is represented. Fig. 8 The three-mass model [18] of the vocal folds. In addition to the mucosal wave and lateral motion,this model accounts for the coupling between the cover and body. B. H. STORY: SOUND SOURCE FOR VOWELS 199 example,the two-mass approach was modified such that a single vocal fold was represented by eight coupled transverse sections,each with two masses in the coronal plane that had both lateral and vertical degrees of freedom [19,20]. This model allows for simulation of the vibrational pattern of the vocal folds in both the coronal and traverse planes. A further step up in complexity came with a continuum mechanics model [21] and a more recent finiteelement model [22] both of which provide a precise physiological and mechanical representation of the tissue layers of the vocal folds without lumping large anatomical regions into a few mass elements. Their large number of degrees of freedom allows for detailed study of the complex vibratory pattern observed in human vocal folds. Interestingly,however,these complex models capable of producing many modal vibration patterns,have shown that vocal fold vibration is largely dominated by only 2–3 modes of vibrations [10],similar to movement patterns shown in Fig. 5. Thus,the lumped-element models seem to capture enough of the vibratory characteristics to still serve as a useful research tool if fine detail is unnecessary. 4. INTERACTION OF THE VOICE SOURCE AND THE VOCAL TRACT In this section a three-mass model [18] will be used to demonstrate self-oscillation of the vocal folds under the influence of several vocal tract configurations. Mathematical details concerning the mechanical aspects of the model can be found in the original publication [18] but the aerodynamic equations have been updated [23]. Table 1 gives the relevant parameter values that will be used throughout the following discussion unless otherwise stated. Note that important characteristics of phonation, such as fundamental frequency,voice quality,amplitude of vibration,may be altered by changing the parameters in Table 1. However,in keeping with the theme of this special issue concerning vowels,the following sections focus only on changing the vocal tract configuration.
การแปล กรุณารอสักครู่..