69 7 glottic behavior in vocal frymar 02, 2020 · all visual hsdp data were recorded using color...

7 Glottic behavior in vocal fry69

Chapter 7Glottic behavior in vocal fry. A preliminary study using LVS & HSDP with analysis by kymography, P-FFT & Nyquist plotsKrzysztof Izdebski, Enrico Di Lorenzo, Ronald R. Ward, Yuling Yan & Matthew Blanco

Abstract

Glottal fry voice falls within the lowest range of the human vocalization and is produced voluntarily or as a part of dysphonia. Although acoustics of fry are common, physiologic data are scarce, specifically with regard to glottis behavior. LVS and nowadays HSDP pro-vide a whole new way to visually investigate laryngeal behavior and glottis posturing during any phonatory modes, providing detailed real-time information about laryngeal biomechanics (e.g., mucosal wave, wave motion directionality, glottic area wave form, asymmetry of vibrations within and across vocal folds (VF), and contact area of the glot-tis. These observations are fundamental to our understanding and modeling of both normal and disordered phonation. In this preliminary report, we focus on direct HSDP in vivo observations of not only the glottic region, but also on the entire supraglottic laryngeal posturing during production of vocal fry. In subsequent reports we examine hiss and over-pressured phonation modes produced in a non-pathological setting. We also contrast these findings to normative phonation and to each other. Analysis includes spatio-temporal vibration patterns of VF, multi-line kymograms, spectral P-FFT analysis, and Nyquist spatio-temporal plots. The presented examples reveal that supraglottic con-traction assists in prolonged closed phase of the vibratory cycle and that the prolonged closed phase is long in vocal fry. These findings need to be compared to pathologic pho-nation containing vocal fry voice in order to arrive at a better differential diagnosis and if/when a better treatment protocol is desired.

Keywords: HSDP, LVS, female voice, vocal fry, mucosal wave, glottic cycle, GAW, PFFT, color analysis, Nyquist plots

Introduction

Vocal fry—also referred to as fry or pulse register, laryngealized voice, creak or creaky voice, glottal fry, rattle, scrape, or strohbass—is the lowest of the human voice qualities sounding like a popping sound [1]. Vocal fry occurs in normal and pathological conditions [2] and can be used to express various emotions or as a form of affected vocal genre [3]. In fact vocal fry is becoming very popular among young women. Recent Youtube postings [4] and popular press articles are focusing on the phenomenon of vocal fry in the female speech and both types of media talk about a female (teenage) vocal fry epidemic sweep-ing our country [5-10].

Normal & Abnormal Vocal Folds Kinematics: HSDP, OCT & NBI®, Volume II: Applications70

Vocal fry has been studied acoustically and physiologically in the past [1, 6, 11] (also see Chapter 6 in this volume). This research established fry to be compatible with a short opened glottis phase (pulse). Recent advances in biomedical optical imaging contribute greatly to our understanding of voice physiology and encourage us to study vocal fry using HSDP. This technology towers over non-optical measurements (e.g., inverse filter-ing, EGG, EMG), which cannot directly record VF vibrations or provide visual information about the laryngeal vibratory laryngeal systems. From the very beginning of laryngeal imaging with a mirror, acquisition of VF images in action has played a crucial role in estab-lishing our fundamental knowledge on how the human voice (VF vibrations) is produced [11]. Significant advances came with the advent of capturing VF vibratory images by film or video and even further gains have been made with the introduction of LVS [12]. LVS provides a visual image of the glottal movement representing a composite image of vi-bratory cycles rather than a sequential representation of consecutive glottis cycles as in HSDP [2, 11]. LVS performs best when observing periodic VF vibrations but suffers when aperiodic VF vibratory patterns are analyzed. Moreover, VF oscillations can range from approximately 40-4000 Hz (or above 8000 Hz at the extreme), hence faster than standard video capture rate imaging systems are a must.

Interestingly, the need for high speed filming to comprehend the biomechanics of the VF was already recognized more than seventy years ago [13], but because of techni-cal complexities, high speed motion pictures were extremely rare occurrences even in the voice physiology laboratories [11].

High-speed capturing rates are more readily available courtesy of modern optics and electronics and are commercially packaged into HSDP systems [14-15]. Despite many cumbersome features of these systems, HSDP is now beginning to emerge as an impor-tant tool for observing the laryngeal biomechanics in vivo for both research and clinical applications [16-18].

These HSDP systems are now capable of recording glottal images at the rates above 16,000 f/s, though rates between 2000 and 4000 frames are more customary and can be obtained in full color. Higher rates require black and white imaging. No matter the actual speed or frame acquisition, analysis of such massive data (i.e., 8000 or 16,000 frames for 4 seconds of phonation at 2000 and 4000 f/s, respectively) requires substantial pro-cessing time and specialized signal analysis algorithms that are part of the commercial systems or are customized [19-20]. Our group [19-23] applies these custom designed automatic tracing systems (e.g., Vocalizer®) to describe laryngeal motions of the glottis from both normal subjects and from patients [19]. HSDP observations also reveal for the first time, except when the ultra-high speed filming was produced in the past [11], that glottis activity is not produced in isolation, but that it is accompanied by posturing and movements of the entire supraglottic area. This oscillatory behavior of the supraglottic larynx had been rarely described or discussed previously.

In this chapter we focus on glottic biomechanics during production of vocal fry by a female subject using HSDP. We feel that these data will be instrumental in modifying phonatory models, specifically with respect to acoustic and aerodynamics of phonation [24-26].

Materials and methods

All visual HSDP data were recorded using color HSV system (KayPENTAX Model 9710, NJ, USA). Standard phonoscopic signal acquisition without topical anesthesia was used [2]. The rate of video frames was set to 2000 with maximum available image resolution of 512 x 512 pixels. Acquired signals were processed by KIPS (KayPENTAX) when appropri-ate and using custom programs such as Vocalizer® elsewhere [17-18].


HSDP images were compared to LVS images obtained from the same subject on the same date with the use of standard LVS equipment (Kay Elemetrics RLS 9100, NJ, USA).

Fry phonation was executed on exhalation by a female subject during production of a sustained /i/ sound. This mode was chosen to contrast these findings to pathological conditions in which such modes constitute primary symptoms. For example fry phona-tion presents in dystonia, traumatic brain injury, closed head injury, or various emotional voice qualities [3, 27].

Results

All modes of phonation are determined by the action of the intrinsic and the extrinsic la-ryngeal muscles that determine specific pitch ranges and the transitions to these ranges from the modal range [2, 28-30]. These muscular adjustments are expressed in VF posi-tioning shown in detail of HSDP, which are superior to those obtained from LVS.

Fry phonation: Glottic wave

Figure 1 (A & B) shows glottic view during fry phonation form LVS (A) and HSDP (B) re-cordings respectively with a transoral rigid endoscope. Note that HSDP displays more vividly the moment of glottis separation.

Figure 1. Glottic separation display from LVS (A) and from HSDP (B).

Findings with KIPS Kymography showed extremely prolonged closed phase of the glottic cycle. This is illustrated in Figure 2.

Using Vocalizer ® technology, we showed that release of phonatory wave was very brief and consisted of a double pulse creating a single sound. This phase release pre-dominated.

This double pulse was generated by two separate locations along the glottis (Figure 3) and represented glottis separation of very short duration and a very long closed phase that was longitudinally limited to mid-glottic portion. These pulses are represented by the two separate illuminated areas of the glottis (outlined in green). Also as shown in this illustration, a significant supraglottic constriction of the FVF is present during vocal fry production.


FFT

Point FFTs for fry voice are shown in Figure 4. Note that left versus right vocal cord equal distant PFFTs are similar and that low F0 is mixed with noise.

Nyquist plots

To provide more information on the acoustics of fry voice mode we also generated Ny-quist plots form voice signals alone. These findings (left plot) are contrasted to normal phonation (right plot) from the same subject shown in Figure 5.

Acoustic data corresponds well to the physiologic data, showing a double loop plot that in our opinion corresponds to the double pulse observed in the mucosal wave pat-tern as shown in Figure 3.

Figure 2. Note extreme long closed phase of the glottis cycle and compres-

sion of the ventricular folds.

Figure 3. Demonstration of double glottis pulse in fry mode. Analysis by Vocalizer®.


Figure 5. Nyquist plot from acoustics for vocal fry contrasted against modal normal phonation for same female speaker.

Note concentration of F0 in the center for fry.

Summary and conclusions

LVS and HSDP are informative about supraglottic topography, with LVS missing fine move-ments of these structures. HSDP is superior in demonstrating details of the mucosal wave. Based on these findings we conclude that mucosal wave is suppressed in fry vocal mode showing a very long closed portion of the glottal cycle and a very short pulse. Gen-erated fry sound may be composed from double pulses generated at two locations within the glottis but proceeding very rapidly as to form a double pulse repetition pattern. Pulse is mixed with noise and medial compression of the entire glottis causes momentary ces-sation of the vibratory cycle.

Fry mode also showed supraglottic contraction. This medial motion of the supraglot-tic structures is non-existing in normal (modal) phonation produced by the same sub-ject. In the fry mode, vocal cord oscillations are suppressed despite VF and FVF midline approximation. Acoustic Nyquist plots revealed fine structure of F0 of fry while PFFT showed only minor left versus right FFT differences.

Figure 4. PFFT of fry derived from L and R locations.


References

1. Large, J., 1972. Towards an integrated physiologic-acoustic theory of vocal registers. NATS Bull. 28, 18-36.

2. Izdebski, K., 2011. Clinical voice assesment: The role & value of phonatory func-tion studies (Chapter 29), in: Lalwani, A. (Ed.), Current Diagnosis & Treatment Otolaryngology–Head and Neck Surgery, third ed. Lange, New York.

3. Izdebski, K., 2008. Emotions in the Human Voice, Vol. I-III. Plural Publishing, San Diego.

4. Vocal fry epidemic: http://www.youtube.com/watch?v=UsE5mysfZsY5. Yuasa, I., 2010. Creaky voice: A new feminine voice quality for young urban-

oriented upwardly mobile american women. Am. Speech 85, 315-337.6. Wolk, L., Abdelli-Beruh, N., Slavin, D., 2012. Habitual use of vocal fry in young

adult female speakers. J. Voice 26, 111-116.7. Fessende, M., 2011. “Vocal fry” creeping into U.S. speech. Science Now.8. Steinmetz, K., 2011. Get your creak on: Is “vocal fry” a female fad? Time:

Health & Family.9. Greenfield, R., 2011. Vocal fry isn’t just for college girls. The Atlantic Wire.10. Greenwood, V., 2011. The linguistic phenomenon du jour: Vocal fry. Discover.11. Moore, P., von Leden, H., 1958. Dynamic variations of the vibratory pattern

in the normal larynx. Folia Phoniatr. (Basel) 10, 205-238.12. Izdebski, K., Ross, J., Klein, J., 1990. Rigid transoral laryngovideostroboscopy

(phonoscopy). Seminars in Speech and Hearing, Thieme.13. Farnsworth, D., 1940. High speed motion pictures of the human vocal cords

(and film). Bell Lab. Rec. 18, 203-208.14. http://www.kayelemetrics.com15. http://www.richard-wolf.com/16. Hertegård, S., 2005. What have we learned about laryngeal physiology

from high-speed digital videoendoscopy? Curr. Opin. Otolaryngol. Head Neck Surg. 13, 152-156.

17. Sonninen, A., Laukkanen, A., 2003. Hypothesis of whiplike motion as a possi-ble traumatizing mechanism in vocal fold vibration. Folia Phoniatr. Logop. 55, 189-198.

18. Deliyski, D., 2005. Endoscope motion compensation for laryngeal high-speed videoendoscopy. J. Voice 19, 485-496.

19. Yan, Y., Ahmad, K., Kunduk, M., Bless, D., 2005. Analysis of vocal-fold vibrations from high-speed laryngeal images using Hilbert transform based methodology. J. Voice 19, 161-175.

20. Jiang, T., Luo, S., Yan, Y., 2013. Software for automatic analysis of image and sound data simultaneously acquired from high-speed videoendoscopy. Proc. SPIE BiOS. International Society for Optics and Photonics, 856520.

21. Yan, Y., Izdebski, K., 2010. Integrated spatio-temporal analysis of high-speed laryngeal imaging and acoustic signals: Their role and applications in the study of normal and abnormal vocal functions, in: Demenko, G., Wagner, E. (Eds.), Speech and Language Technology, vol. 12/13. Polish Phonetic Association, Poznan, pp. 15-38.

22. Yan, Y., Kunduk, E., Izdebski, K., 2009. Acoustics contribute to the differential diagnosis of movement disorders of vocalization. Paper presented at the XVII Annual Pacific Voice Conference PVSF/UCLA.


23. Yan, Y., Demenko, G., Izdebski, K., Bhandari, A., 2010. Using Nyquist plots to describe real and mimicking traumatic phonation. Paper presented at XVIII Annual Pacific Voice Conference PVSF/UCLA.

24. Fant, G., 1970. Acoustic Theory of Speech Production. Mouton De Gruyter, Netherlands.

25. Granqvist, S., Hertegård, S., Larsson, H., Sundberg, J., 2003. Simultaneous analysis of vocal fold vibration and transglottal airflow: Exploring a new experi-mental setup. J. Voice 17, 319-330.

26. Titze, I., 1994. Principles of Voice Production. Prentice Hall, Englewood Cliffs.27. Izdebski, K., 1984. Overpressure and breathiness in spastic dysphonia:

An acoustic (LTAS) and perceptual study. Acta Otolaryngol. 97, 122-129.28. Izdebski, K., 2011. Phonetotopic organization of phonation: Evidence

from electrophysiology, aerodynamics, acoustics and kinestetics. Paper presented at 2011 CoMeT Meeting, Frankfurt.

29. Faaborg-Andersen, K., 1957. Electromyographic investigation of intrinsic laryn-geal muscles in humans. Acta Physiol. Scand. 41, Suppl. 140.

30. Shipp, T., 1968. A technique for examination of laryngeal muscles during phona-tion. Proc. First Int. Cong. Electromyographic Kinesiology.

69 7 glottic behavior in vocal frymar 02, 2020 · all visual hsdp data were recorded using color...

Documents