Skip to main content
PMC home page
The Journal of the Acoustical Society of America logoLink to The Journal of the Acoustical Society of America
. 2010 Dec 6;128(6):EL378–EL383. doi: 10.1121/1.3515838

On the acoustical relevance of supraglottal flow structures to low-frequency voice production

Zhaoyan Zhang 1,✉, Juergen Neubauer 1
PMCID: PMC3013154  PMID: 21218861

Abstract

The supraglottal flow exhibits many complex phenomena such as recirculation, jet instabilities, jet attachment to one vocal fold wall, jet flapping, and transition to turbulence. The acoustical relevance of these flow structures to low-frequency voice production was evaluated by disturbing the supraglottal flow field using a cylinder and observing the consequence on the resulting sound pressure field. Despite a significantly altered supraglottal flow field due to the presence of the cylinder, only small changes in sound pressure amplitude and spectral shape were observed. The implications of the results on our understanding of phonation physics and modeling of phonation are discussed.

Introduction

There have been many recent studies on characterization of the supraglottal flow field (for a detailed list of such studies see the literature review in, e.g., Neubauer et al., 2007; Luo et al., 2009; Triep and Brucker, 2010). These studies show that the downstream glottal flow is highly three-dimensional and exhibits many complex phenomena such as recirculation, jet attachment to one vocal fold wall, jet instabilities (vortex shedding and roll-up), jet reattachment to vocal tract walls, jet flapping, and transition to turbulence. While accounting for such phenomena significantly increases the computational load of phonation models, the acoustical relevance of such flow features, i.e., how they affect the fluid–structure interaction and eventually the acoustics of the produced sound and its perception, is still unclear. Considering that the ultimate goal of phonation models is to accurately predict the characteristics of the radiated sound, the supraglottal flow structures, if can be shown to be acoustically irrelevant, do not have to be included in such phonation models.

From the sound production point of view, on the one hand, these complex flow structures in the downstream glottal flow field are sound sources of quadrupole type (dipole type when obstacles present in the pathway of airflow) and radiate sound mostly at high frequencies (generally above 2 kHz) due to the small length scales associated with the flow structures (Zhang et al., 2002; Howe and McGowan, 2007). Therefore, these flow features have to be accurately modeled if the high-frequency component of voice is to be reproduced.

On the other hand, it has been postulated that the unsteady supraglottal flow structures may affect the intraglottal flow and pressure distribution, which is the driving force of vocal fold vibration, and therefore affect the low-frequency sound production. Once separated from the vocal fold walls, the glottal jet starts to develop jet instabilities and is therefore susceptible to downstream disturbances, especially when the glottis takes on a divergent shape. In this way, the unsteady supraglottal flow structures may strongly interact with the boundary layer at the glottal exit and affect flow separation point within the glottal channel (Hirschberg et al., 1996). Previous works suggested that asymmetric vortical structures downstream may cause jet deflection or oscillation near the glottal exit (Thomas and Chu, 1989; Shinwari et al., 2003; Neubauer et al., 2007; Luo et al., 2009). For confined jet flows, a strong recirculation zone exists surrounding the jet region. The unsteady flow in the recirculation zone may also strongly influence the jet orientation, particularly during glottal opening and closing (Alleborn et al., 1997; Neubauer et al., 2007). However, it is unclear how and to what extent such upstream influence, if it exists indeed, can affect the intraglottal pressure distribution and the resulting voice.

This study attempted to quantify the acoustical relevance of supraglottal flow structures in the low-frequency or the harmonic component of phonation in a self-oscillating model of the vocal folds. The focus on the low-frequency harmonic component was justified by the fact that, in normal phonation, the low-frequency harmonic component dominates the inharmonic component (the so-called noise component due to flow turbulence) and therefore is perceptually more important (Klatt and Klatt, 1990).

Our approach was to disturb the supraglottal flow field and observe how the low-frequency component of the sound (amplitude, fundamental frequency, and spectral shape) produced by the disturbed supraglottal flow field would be different from that produced by the undisturbed supraglottal flow field. The hypothesis is that if the supraglottal flow structures are essential to the production of the low-frequency harmonic component of voice, significant changes in the low-frequency sound should be observed if these supraglottal flow structures are significantly altered. Otherwise, if these supraglottal flow structures are acoustically irrelevant, alterations of the supraglottal flow field would not produce noticeable difference in the low-frequency sound field.

Methods

The same experimental setup as in Zhang et al. (2006a) was used. It consisted of an expansion chamber, a 12.4-cm-long uniform circular tracheal tube (inner diameter of 2.54 cm), a two-layer self-oscillating vocal fold model, and a 2.8-cm-long uniform rectangular vocal tract tube. Airflow was supplied through a pressurized flow supply. The short vocal tract was used to provide confinement to the jet flow exiting the vocal fold model. The short tube length also helped to reduce possible supraglottal acoustic feedback (Zhang et al., 2006a), which allowed us to focus on aerodynamic feedbacks from the supraglottal flow field. The two-layer vocal fold model had a Young’s modulus of approximately 9.3 and 3.0 kPa for the body (inner) and cover (outer) layers of the vocal fold model, respectively. Phonation onset occurred at around 3.09 kPa at a phonation frequency of 147 Hz, which was accompanied by a sudden increase in the sound pressure amplitude. Beyond onset, the sound pressure amplitude increased significantly with increasing upstream pressure. For example, for a mean upstream pressure of 3.27 kPa (at which the results below were obtained), a ±10% change in upstream sound pressure amplitude roughly corresponds to a ±0.05 kPa change in the mean upstream pressure.

Previous studies showed that the physical model exhibited qualitatively similar vibratory pattern (Zhang et al., 2006b, 2009, 19) and supraglottal flow field pattern (Neubauer et al., 2007; Drechsel and Thomson, 2008) to that observed in excised-larynx (Khosla et al., 2007) and in-vivo canine models (Dollinger et al., 2005). In general, the vocal fold models exhibited vibratory patterns (e.g., onset frequency and pressure, pressure–flow relationship, vibration amplitude, and flow modulation, etc.) similar to that typical of humans. The glottal jet exhibited cyclic variations in both the jet width and jet length during one oscillating cycle, with the jet disappearing at the end of the closing phase and reappearing at the opening phase of the next cycle. Readers are referred to these previous studies for further details of the vibratory and flow field pattern of this physical model.

To disturb the supraglottal flow, a hexagonal cylinder of 1.27 mm diameter was aligned in the anterior–posterior direction downstream of the vocal fold model [Fig. 1a] and traversed in the flow direction at different left–right locations. The cylinder was long enough to cover the entire anterior–posterior span of the vocal fold model but slightly smaller than the cross-sectional dimension of the vocal tract tube. Figure 1b defines the coordinate system which was used later in Sec. 3. The axial location (x) of the cylinder was measured with reference to the point of maximum downstream excursion of the vocal fold model [Fig. 1b]. During experiments, this reference location (x = 0) was first identified using high-speed recordings of the vocal fold model and then used for determination of cylinder locations in later experiments. The left–right direction (y-axis) was defined with regard to the nominal glottal centerline.

Figure 1.

Figure 1

Open in a new tab

(Color online) (a) Superior view of the vocal fold model and the cylinder used to disturb the supraglottal flow. (b) Side (frontal) view of the vocal fold model during vibration and the definition of coordinate system.

The effectiveness of the cylinder in disturbing the supraglottal flow was verified by visualizing the supraglottal flow field using a high-speed laser and a high-speed camera. Comparison between the flow field with and without cylinder shows that the supraglottal flow field was significantly disturbed by the presence of the cylinder except for extreme off-center cylinder locations in the left–right direction. The jet was either deflected to one side or split into two separate jets at the location of the cylinder, and the distribution and evolution of vortical structures were also significantly altered (Fig. 2).

Figure 2.

Figure 2

Open in a new tab

(Color online) Typical images of the supraglottal flow field in the left–right and axial plane during one oscillation cycle: (a) Without cylinder; (b) cylinder was about at the glottal center (x = 4.5 mm, y = 0 mm); and (c) cylinder was on the right of the glottal center (x = 4.5 mm, y = 1.5 mm). The cylinder was highlighted by a hexagon. The black horizontal strip to the left of the hexagon is the shadow of the cylinder as the laser light sheet came from the left side.

For each cylinder location, the mean subglottal pressure, the mean flow rate, and the acoustic pressure inside the tracheal tube (2 cm upstream from the entrance of the glottis) and outside were measured for a 1-s period.

Note that high-frequency tonal sounds were generated due to jet impingement onto the cylinder when the cylinder was placed directly inside the jet flow. To focus on the low-frequency voiced sound production and facilitate comparison to cases when the cylinder was away from the jet axis and no jet tones were produced, the outside sound pressure was low-pass filtered with a cut-off frequency of 2.5 kHz. No filtering was applied to the upstream sound pressure.

Results

Figure 3a shows the normalized amplitude of the outside sound pressure as the cylinder was traversed in the flow direction, for five left–right locations. The five left–right locations include two locations on the left (locations 1 and 2), one approximately at the glottal center (location 3), and two locations on the right (locations 4 and 5). The data were normalized by their corresponding values when the cylinder was traversed to the farthest point downstream from the vocal fold model. Therefore, by definition, a value significantly different from one indicates a significant influence of the presence of the cylinder on the sound pressure amplitude. During the course of each experiment, the mean upstream pressure was manually kept at 3.27 kPa.

Figure 3.

Figure 3

Open in a new tab

(Color online) The normalized outside acoustic pressure amplitude (top) and fundamental frequency (bottom) as the cylinder was traversed in the flow direction for five cylinder locations in the left–right direction, ○: cylinder at y = −2 mm (location 1); ◇: cylinder at y = −1.4 mm (location 2); □: cylinder at y = 0 mm (glottal center, location 3); ▵: cylinder at y = 1.2 mm (location 4); and +: cylinder at y = 2.2 mm (location 5). The two straight lines in Fig. 3a indicate the borderlines of ±10% change in sound pressure amplitude.

For all five left–right locations of the cylinder, the influence on the sound pressure amplitude decreased as the cylinder moved away from the vocal fold model and was less than 10% for cylinder locations x > 3 mm [Fig. 3a], which compares with a maximum glottal opening width of about 3 mm in the left–right direction. For location 3 at the glottal center and location 5 on the right, this region of negligible influence (sound pressure amplitude change less than 10%) extended even further upstream to x > 0.75 mm.

Figure 3b shows the corresponding fundamental frequency as a function of the cylinder location in the flow direction, for the five cylinder locations in the left–right direction. Similar to that of the sound pressure amplitude, changes in the fundamental frequency decreased as the cylinder was traversed away from the vocal fold model. For all cylinder locations x > 3 mm, changes in fundamental frequency were within ±2 Hz.

Although not shown in this paper, both the subglottal and outside sound spectra remained qualitatively the same as the cylinder was traversed (except for the tonal sound above 3 kHz in the outside sound spectra, due to jet impingement onto the cylinder).

Discussion

The results of this study show that the presence of the cylinder had little influence on the sound pressure amplitude (less than 10% change) and fundamental frequency (less than 2 Hz change) when the cylinder was located further than 3 mm downstream of the glottal exit (x > 3 mm or one maximum glottal width). For cylinder locations in the jet center or outside of the jet, this region of negligible influence extended further upstream to 0.75 mm downstream of the glottal exit (x > 0.75 mm or 0.25 times the maximum glottal width).

A rough estimation of the upstream influence of a circular non-rotating cylinder of diameter 1.47 mm (the width of the hexagonal cylinder of this study in the left–right direction) in a uniform potential flow suggests that a change in upstream velocity greater than 10% is induced at locations within 1.59 mm upstream of the cylinder. Therefore, the results of this study indicate that disturbances in the part of the supraglottal flow field that was approximately 1.4 mm downstream of the glottal exit had little influence on the low-frequency sound production. In this study, the jet flow during most part of the oscillating cycle persisted for at least 6 mm downstream of the glottal exit before it broke down into turbulence. Therefore, this region of negligible influence included a significant part of the jet region as well as the turbulence region further downstream.

The fact that significant alterations in supraglottal flow structures did not lead to significant change in sound production indicates that jet instabilities, the detailed vortical structures, and transition to turbulence in the supraglottal flow field may only have little influence on the low-frequency vocal fold vibration and sound production.

Previous studies have suggested possible upstream aerodynamic feedback mechanisms by which asymmetric downstream vortical structures can induce deflection or oscillation in upstream shear layer or flow separation patterns (Thomas and Chu, 1989; Neubauer et al., 2007). However, the observed small influence of the disturbed supraglottal flow field in this study seems to indicate that such aerodynamic feedback mechanisms either did not induce significant aerodynamic changes in the upstream flow field or, if they did, such aerodynamic changes were not able to induce significant acoustic changes.

Such small influence is consistent with our experience with excised larynx experiments, in which no noticeable changes in phonation were observed when the flow was disturbed in the supraglottal region. Also, recent studies have shown that predictions from phonation models using low-dimensional flow models (both one-dimensional and two-dimensional) compared well with experiments, despite the many complex flow phenomena being neglected in these models (Pelorson et al., 1994; Zhang et al., 2002; Ruty et al., 2007; Kaburagi and Tanabe, 2009), indicating possibly a minor role of the supraglottal flow structures in phonation.

We would like to emphasize the importance of more systematic study to further clarify the acoustic relevance of supraglottal flow structures in more realistic phonatory conditions. The vibratory behavior of the physical models used in this study was slightly different from that observed in human. It is unclear how the observations of this study would be affected under vibratory patterns typical of humans or different pathological conditions. Also, when close to bifurcation boundaries, it is possible that the vocal fold vibration and sound production may be more susceptible to disturbances in supraglottal flow structures. Further studies are needed to quantify the influence of the supraglottal flow on phonation in more realistic phonatory conditions. If the observation of this study can be generalized to more realistic phonatory conditions, it will then suggest that the supraglottal flow structures do not need to be accurately resolved in phonation models for predictions of reasonable accuracy (e.g., 10% difference in amplitude and a few hertz in phonation frequency).

ACKNOWLEDGMENT

This study was supported by Grant No. R01 DC009229 from NIDCD, NIH.

References and links

  1. Alleborn, N., Nandakumar, K., Raszillier, H., and Durst, F. (1997). “Further contributions on the two-dimensional flow in a sudden expansion,” J. Fluid Mech. 330, 169–188. [Google Scholar]
  2. Dollinger, M., Berry, D., and Berke, G. (2005). “Medial surface dynamics of an in vivo canine vocal fold during phoantion,” J. Acoust. Soc. Am. 117, 3174–3183. [DOI] [PubMed] [Google Scholar]
  3. Drechsel, J. S., and Thomson, S. L. (2008). “Influence of supraglottal structures on the glottal jet exiting a two-layer synthetic, self-oscillating vocal fold model,” J. Acoust. Soc. Am. 123, 4434–4445. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Hirschberg, A., Pelorson, X., Hofmans, G., van Hassel, R. R., and Wijnands, A. P. J. (1996). “Starting transient of the flow through an in-vitro model of the vocal folds,” in Vocal Fold Physiology: Controlling Complexity and Chaos, edited by Davies P. J. and Fletcher P. J. (Singular, San Diego: ), pp. 31–46. [Google Scholar]
  5. Howe, M. S., and McGowan, R. S. (2007). “Sound generated by aerodynamic sources near a deformable body, with application to voice speech,” J. Fluid Mech. 592, 367–392. [Google Scholar]
  6. Kaburagi, T., and Tanabe, Y. (2009). “Low-dimensional models of the glottal flow incorporating viscous-inviscid interaction,” J. Acoust. Soc. Am. 125, 391–404. [DOI] [PubMed] [Google Scholar]
  7. Khosla, S. M., Murugappan, S., Gutmark, E. J., and Scherer, R. C. (2007). “Vortical flow field during phonation in an excised canine larynx model,” Ann. Otol. Rhinol. Laryngol. 116, 217–228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Klatt, D. H., and Klatt, L. C. (1990). “Analysis, synthesis, and perception of voice quality variations among female and male talkers,” J. Acoust. Soc. Am. 87, 820–857. [DOI] [PubMed] [Google Scholar]
  9. Luo, H. X., Mittal, R., and Bielamowicz, S. A. (2009). “Analysis of flow-structure interaction in the larynx during phonation using an immersed-boundary method,” J. Acoust. Soc. Am. 126, 816–824. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Neubauer, J., Zhang, Z., Miraghaie, R., and Berry, D. A. (2007). “Coherent structures of the near field flow in a self-oscillating physical model of the vocal folds,” J. Acoust. Soc. Am. 121, 1102–1118. [DOI] [PubMed] [Google Scholar]
  11. Pelorson, X., Hirschberg, A., van Hassel, R. R., and Wijnands, A. P. J. (1994). “Theoretical and experimental study of quasisteady-flow separation within the glottis during phonation. Application to a modified two-mass model,” J. Acoust. Soc. Am. 96, 3416–3431. [Google Scholar]
  12. Ruty, N., Pelorson, X., Van Hirtum, A., Lopez-Arteaga, I., and Hirschberg, A. (2007). “An in vitro setup to test the relevance and the accuracy of low-order vocal folds models,” J. Acoust. Soc. Am. 121, 479–490. [DOI] [PubMed] [Google Scholar]
  13. Shinwari, D., Scherer, R. C., DeWitt, K. J., and Afjeh, A. A. (2003). “Flow visualization and pressure distributions in a model of the glottis with a symmetric and oblique divergent angle of 10 degrees,” J. Acoust. Soc. Am. 113, 487–497. [DOI] [PubMed] [Google Scholar]
  14. Thomas, F. O., and Chu, H. C. (1989). “An experimental investigation of the transition of a planar jet: Subharmonic suppression and upstream feedback,” Phys. Fluids A 1, 1566–1587. [Google Scholar]
  15. Triep, M., and Brucker, C. (2010). “Three-dimensional nature of the glottal jet,” J. Acoust. Soc. Am. 127, 1537–1547. [DOI] [PubMed] [Google Scholar]
  16. Zhang, Z., Mongeau, L., and Frankel, S. H. (2002), “Experimental verification of the quasi-steady approximation for aerodynamic sound generation by pulsating jets in tubes,” J. Acoust. Soc. Am. 112(4), 1652–1663. [DOI] [PubMed] [Google Scholar]
  17. Zhang, Z., Neubauer, J., and Berry, D. A. (2006a). “The influence of subglottal acoustics on laboratory models of phonation,” J. Acoust. Soc. Am. 120, 1558–1569. [DOI] [PubMed] [Google Scholar]
  18. Zhang, Z., Neubauer, J., and Berry, D. A. (2006b). “Aerodynamically and acoustically-driven modes of vibration in a physical model of the vocal folds,” J. Acoust. Soc. Am. 120, 2841–2849. [DOI] [PubMed] [Google Scholar]
  19. Zhang, Z., Neubauer, J., and Berry, D.A. (2009). “Influence of vocal fold stiffness and acoustic loading on flow-induced vibration of a single-layer vocal fold model,” J. Sound Vib. 322, 299–313. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from The Journal of the Acoustical Society of America are provided here courtesy of Acoustical Society of America

RESOURCES