Adjustment of Glottal Confgurations in

Christian T. Herbst and Jan G. Švec

INTRODUCTION

ocal is the central quality in singing. On a long-term level basis, singers and their teachers are concerned with “building the instrument,” that is, establishing habitual motor control and behavioral patterns for voice production Vwithin the limits of their chosen singing style (this process is well refected by the German term Stimmbildung). On the other hand, singers must also be able to vary voice timbre on an ad hoc level, allowing for artistic Christian T. Herbst expression. Here, agility and muscular fne control are the main features of a “good” voice. Te human voice production mechanism consists of three basic physi- ological layers: the respiratory system (power source), the larynx (sound source), and the vocal tract (sound modifers).1 Te most obvious method for voice timbre control is via the sound modifers, that is, articulation by introducing changes into the vocal tract shape.2 Fine tuning the vowel color (in relation to pitch) is a well established concept for varying the vocal timbre and optimizing sound output in singing.3 At the laryngeal (sound source) level, vocal timbre can be modifed by two Jan G. Švec basic physiological adjustments: (1) adduction of the posterior (cartilaginous) glottis, controlled by the singer when changing the quality of voice between “breathy” and “pressed”;4 and (2) via the singing voice register.5 Whereas each of these physiological parameters (adduction of the posterior glottis and vocal registers) is well researched and understood individually, little is known about their relation to each other. Tis is particularly true in the realm of voice pedagogy, where the proper terminology is not well estab- lished and sometimes not well comprehended. It is therefore desirable to fll this gap in our understanding of the singing voice by attempting to relate the two concepts to each other.

TWO TYPES OF GLOTTAL ADDUCTION

Te glottis can be conceptually divided into two parts: the membranous glottis (i.e., the portion made up of the vocal folds from the anterior commissure Journal of Singing, January/February 2014 to the tip of the vocal processes); and the cartilaginous glottis (i.e., the poste- Volume 70, No. 3, pp. 301–308 Copyright © 2014 rior part of the glottis, consisting of the arytenoid cartilages and their vocal National Association of Teachers of Singing processes; see Figure 1b).

January/February 2014 301 Christian T. Herbst and Jan G. Švec

Figure 1. Schematic illustration of two types of glottal adduction: (a) membranous medialization; middle and bottom graph: coronal section through larynx, showing the efect of TA muscle contraction; (b) conceptual separation of glottis into membranous and cartilaginous portion; (c) cartilaginous adduction.

A closer examination of the two glottal adjustments the vocal folds; in particular, it bulges out the inferior revealed a causal relationship between (1) the singer’s portion of the vocal fold, and also reduces the width of intended sound quality; (2) the muscular adjustments the membranous glottis.7 Tis maneuver can be called made in the larynx; (3) their impact on glottal adduc- membranous medialization through vocal fold bulging tion (membranous and/or cartilaginous glottis); (4) the (short: membranous medialization).8 A contraction of vibratory properties of the vocal folds and their efect the TA increases the vertical phase delay in vocal fold on glottal airfow; and (5) the actual sound that is being vibration and results in a longer closed phase (and thus a produced. larger closed quotient).9 From an acoustic point of view, Te chest vs. register is mainly controlled this enhances the output of high-frequency energy,10 via the thyroarytenoid (TA, vocalis) muscle (see Figure thus creating a “brighter” or more “resonant” timbre. 1a). Whereas in falsetto the TA is more or less relaxed, This is perceived as a “heavier” by both it is contracted in chest (particularly at higher pitches), singers and listeners. thus thickening, shortening, and medially bulging the Te voice quality along the dimension of “breathy” body of the vocal fold, while slackening the vocal fold vs. “pressed” is maintained through the adductors, that cover.6 Te medial bulging of the membranous portion is, the lateral cricoarytenoid (LCA) and interarytenoid of the vocal fold adducts the membranous portion of (IA) muscles, and their antagonist, the posterior cri-

302 Journal of Singing Adjustment of Glottal Confgurations in Singing

Figure 2. Schematic illustration of the efect of cartilaginous adduction and membranous medialization through vocal fold bulging in singing. For each adduction type, two schematic graphs are shown: top view of vocal folds, arytenoids, and thyroid cartilage (lef); sagittal view of larynx, with schematic drawings of thyroid cartilage, cricoid cartilage, and thyroarytenoid muscle (right). Te arrows indicate the primary changes in the vocal fold position for each case. Te expected theoretical contributions of the internal laryngeal muscles (TA, LCA, and IA) on the three dimensional airspace of the glottis are also indicated. Te main action of the TA is to bulge out the inferior portion of the vocal fold; the LCA adducts the vocal processes and thus also the superior membranous portion of the vocal fold (indicated by the small gray arrows in aDducted falsetto); and the IA seals the posterior cartilaginous glottis (Christian T. Herbst, Qingjun Qiu, Harm K. Schutte, and Jan G. Švec, “Membranous and Cartilaginous Vocal Fold Adduction in Singing,” Journal of the Acoustical Society of America 129, no. 4 [April 2011]: 2253–2262; used with permission). coarytenoid (PCA) muscle, which is an abductor (see tis (caused by the efect of LCA activity on the superior Figure 1c).11 Te main efect of activity in the LCA and part of the membranous vocal fold portion). Tis is due IA muscles is the adduction of the cartilaginous glottis to the fact that the vocal folds terminate posteriorly at via the positioning of the arytenoid cartilages. Whereas the vocal processes of the arytenoid cartilages, hence any the LCA adducts the vocal processes (and hence also the manipulation of the cartilaginous glottis is propagated superior membranous portion of the vocal folds), the to a some extent to the membranous glottis. A higher IA muscles seal the posterior part of the cartilaginous degree of cartilaginous adduction results in a longer glottis. Te maneuver induced by the combined action closed phase (and thus a larger closed quotient),15 creat- of LCA and IA is thus termed cartilaginous adduction.12 ing stronger high frequency components in the radiated Te cartilaginous glottis can either be fully adducted, or sound.16 there can be a posterior glottal chink (PGC) of variable Schematically, these two types of glottal adduction size.13 In the latter case, vocal fold contact during a glottal (cartilaginous adduction and membranous medializa- cycle is reduced or nonexistent, which results in a steady tion) and the resulting glottal confgurations, as well as airflow during the “pseudo” closed phase, therefore the expected roles of the internal laryngeal muscles, are introducing noise components into the voice source.14 displayed in Figure 2. It should be noted that no single As a side efect, cartilaginous adduction also has an muscle adducts one and only one compartment of the infuence on the confguration of the membranous glot- vocal folds, particularly when considering the vertical

January/February 2014 303 Christian T. Herbst and Jan G. Švec

TABLE 1. Four distinct phonation types, as created by dif- muscle tension dysphonia.18 On the other hand, falsetto ferent degrees of cartilaginous adduction and membranous register (where the thyroarytenoid muscle is relaxed) medialization. can be produced with full closure of the posterior glottis, Cartilaginous Membranous such as in the mid and upper range of trained female Voice quality adduction medialization classical singing, and in singing. Te timbral efect of variations of the closed phase can aBducted falsetto — — aDducted falsetto ++ — be observed in the spectra shown in the bottom of Figure aBducted chest — ++ 3. When going from aBducted falsetto to aDducted chest, aDducted chest ++ ++ the ever increasing duration of the closed phase correlates with a decreasing spectral slope, resulting in stronger high frequency components in the acoustic signal for dimension of the membranous portion of the vocal folds. phonations with more adduction of the cartilaginous However, these fner aspects of vocal fold geometry are or membranous glottis (i.e., a “heavier registration”).19 not relevant for the creation of a basic pedagogic model, which should have the potential to be applied easily in TWO TYPES OF ADDUCTION— both singing pedagogy and speech therapy. A PEDAGOGIC MODEL Recently it has been shown that both trained and untrained singers can separately infuence the degree of Since cartilaginous adduction and membranous medial- cartilaginous adduction and membranous medialization, ization can be controlled separately by both trained and thus being able to create four distinct voice qualities by untrained singers, these two physiological parameters four distinct glottal confgurations: aBducted falsetto can be displayed in a two dimensional plane in order (FaB), aDducted falsetto (FaD), aBducted chest (CaB) to create a pedagogic model for sound quality adjust- and aDducted chest (CaD; Table 1 and Figure 3).17 ments made on the laryngeal level (Figure 4). Tis model As expected, the main difference between the consists of four quadrants, representing the four sound aBducted phonations and their aDducted counterparts qualities described above. Diferent singing styles can be was observed in the confguration of the posterior glot- mapped onto the model, showing their main mechanism tis. All the subjects had a greater posterior glottal chink of production, such as:20 and thus a less adducted cartilaginous glottis in the two t aBducted falsetto: e.g., untrained female singing (high aBducted (breathy) phonation types (“aBducted fal- tessitura), “naïve” male falsetto, “lighter” registration setto” and “aBducted chest”) than in the two aDducted (both male and female) in pop and jazz; (nonbreathy) phonation types (“aDducted falsetto” and t aDducted falsetto: e.g., trained female classical (high “aDducted chest”). tessitura), countertenor; Te main diference between the phonations in the t aBducted chest: e.g., untrained naïve singing (lower two different registers (falsetto and chest) is seen in range), “lighter” registration (both male and female) larger vertical phase differences of vocal fold vibra- in pop and jazz (low tessitura); tion, a prevalence of mucosal waves, and the duration t aDducted chest: e.g., trained male classical, belting, of the closed phase. Te observed values difered from rock, soul. individual to individual. In some cases, there was an Te model is in no way restricted to the production of overlap of closed quotient values for aBducted chest four distinct phonation types. On the contrary, gradual and aDducted falsetto. In aBducted chest, the effect adjustments along the two major axes (cartilaginous of the vocal fold bulging caused by contraction of the adduction and membranous medialization) are pos- thyroarytenoid muscle in chest register can be counter- sible. Tese fne controlled muscular adjustments allow acted by a decreasing adduction in the posterior glottis, experienced singers to create a large variety of sound allowing for a “breathy chest” phonation. Te resulting at the glottal level, increasing their expressional glottal confguration is, in its extreme case, comparable freedom within the aesthetic boundary conditions of to the one observed in hyperfunctional breathy voice or their chosen singing style. Also, the model can provide

304 Journal of Singing Adjustment of Glottal Confgurations in Singing

Figure 3. Images and signals documenting the vocal fold vibration in the four phonation types: aBducted falsetto (FaB), aDducted falsetto (FaD), aBducted chest (CaB), aDducted chest (CaD). From top to bottom: [Row 1] pairs of videostroboscopic images at the phases of maximum vocal fold contact and maximum glottal opening. Te arrows point at the position of the vocal processes of the arytenoid cartilages, which are slightly set apart in FaB and CaB and close together in FaD and CaD; [Row 2] typical videokymographic images at the place of maximal vibration amplitude of the vocal folds; [Row 3] typical EGG signals (dark gray) and their frst derivative, dEGG (light gray), for all phonation types, normalized both in amplitude and time. Te x-axis represents normalized glottal cycle duration, and the y-axis shows EGG and dEGG signal amplitude, locally normalized per glottal cycle; [Row 4] spectrogram of sustained notes (taken from the audio channel of the videokymographic recordings) produced with phonation types A, B, C, and D consecutively. (Reprint from Christian T. Herbst, Sten Ternström, and Jan G. Švec, “Investigation of Four Distinct Glottal Confgurations in Classical Singing—A Pilot Study,” Journal of the Acoustical Society of America 125, no. 3 [March 2009]: EL104–EL109, used with permission.) Acoustic and video fles documenting the four phonation types can be freely obtained from http://dx.doi.org/10.1121/1.3057860. pedagogic strategies, suggesting physiological adjust- sical singers. We postulate that in classical singing (as ments to successfully create desired singing styles. opposed to other singing styles, such as belting) the desired quality above the primo (around pitch A MODEL CASE FROM THE 21 D4, ca. 295 Hz) would have to be aDducted falsetto, SINGING STUDIO since (a) chest phonation is usually restricted to the A typical example for the application of the model lower range; and (b) breathy phonation, created by described here would be the mid-range of female clas- aBducted falsetto, is outside the aesthetic boundaries.

January/February 2014 305 Christian T. Herbst and Jan G. Švec

A valid pedagogic strategy would try to establish the aDducted falsetto at higher pitches (around B to F ), ♭4 5 via calling (but not shouting) at vowel /u/, or using “pri- mal sounds.”21 Tis sound quality could then be applied to other vowels, and to pitches all the way down to the primo passaggio. When attempting to achieve a successful transition from (aDducted) chest to aDducted falsetto on ascending scales, there is the danger that the singer “gets stuck” in the chest register. In order to avoid this, it is advisable that ascending scales are sung only when: (1) the student reliably succeeded in producing aDducted falsetto in the primo passaggio on descending scales; and (2) she managed the falsetto-chest register transition at the primo passaggio without abrupt timbral changes.

CONCLUSION

Figure 4. Relationship between the two types of adduction Te singing voice timbre can be controlled at the laryn- and the four phonation types described here. Tis geal sound source by two types of glottal adduction: pedagogic model is not limited to distinct and extreme cartilaginous adduction (via the voice quality along the phonation types; rather, it encourages gradual, fne controlled adjustments along the two major dimensions, dimension of “breathy” vs. “pressed”) and membranous in order to increase the singer’s timbral variability for medialization (via the choice of register, that is, chest vs. enhanced artistic expression. falsetto). Both trained singers and nonsingers of both sexes can separately control these two physiological parameters. Cartilaginous adduction and membranous Untrained female singers have a tendency to produce medialization can be mapped onto a two-dimensional sounds above the primo passaggio in either aDducted plane in order to create a pedagogic model for sound chest register, or in aBducted falsetto. In other words, quality adjustments made on the laryngeal level. Tis they either “push” the chest register up (usually at the model may help to better understand diferent singing risk of an abrupt register transition at the upper end of styles, and may be applied in the voice studio. Te sug- the chest register range), or they produce sof breathy gested model allows both singing teachers and voice sounds, in both cases severely limiting the dynamic therapists to better understand and diagnose their cli- range and timbral variability of the voice at those par- ents’ voice production habits on a physiological level. ticular pitches. Tis could eventually lead to novel pedagogic methods Te difculty of switching from aDducted chest to (or to better understanding and applying existing ones), aDducted falsetto can be explained physiologically. thus increasing the efciency and efectiveness of teach- When changing from chest to falsetto register, the TA ing and therapy. muscle relaxes and thus the membranous medialization is reduced. Te resulting decrease of overall vocal fold ACKNOWLEDGEMENTS adduction could be counteracted by a slight increase of cartilaginous adduction via the LCA and IA muscles, in Te research has been supported in the Czech Repub lic order to keep the timbral change at a minimum. Such a by the European Social Fund Projects OP VK CZ.1.07/ maneuver (decreasing TA while increasing LCA and IA 2.3.00/20.0057, and CZ.1.07/2.3.00/30.0004 “POST-UP,” activity, respectively) is, however, both ambivalent and and in Austria by the ERC Advanced Grant SOMACCA. complex. Since the three involved muscles are all inner- We thank Dr. Ingo R. Titze for his helpful suggestions vated by the recurrent laryngeal nerve, the required level and comments on the manuscript. of fne control might be hard to reach by some singers.

306 Journal of Singing Adjustment of Glottal Confgurations in Singing

NOTES 10. James L.Flanagan, “Some Properties of the Glottal Sound 1. David M. Howard and Damian T. Murphy, Voice Science, Source,” Journal of Speech and Hearing Research 1 (June Acoustics, and Recording (San Diego, CA: Plural Publishing, 1958): 99–116. 2007), 33. 11. During phonation, a minor, supplementary role in arytenoid 2. Peter Ladefoged, A Course in Phonetics, vol. 4 (Orlando, FL: adduction/abduction is played also by the thyroarytenoid Harcourt, 2001), 193–216. (TA) and cricothyroid (CT) muscles; for details see Zemlin, 126–135. 3. Johan Sundberg, “The Acoustics of the Singing Voice,” Scientifc American 236 (March 1977): 82–91; Donald G. 12. Herbst et al. Miller, Resonance in Singing (Princeton, NJ: Inside View 13. Maria Södersten, Stellan Hertegard, and Britta Hammarberg, Press, 2008), 187–206; Berton Coffin, Coffin’s Overtones “Glottal Closure, Transglottal Airfow, and Voice Quality of (Metuchen, NJ: Te Scarecrow Press, 1980); in Healthy Middle-Aged Women,” Journal of Voice 9, no. 2 Matthias Echternach, Johan Sundberg, Tobias Baumann, (June 1995): 182–197. Michael Markl, and Bernhard Richter, “Vocal Tract Area 14. Te phase where air fow reached a minimum during a glottal Functions and Formant Frequencies in ’ cycle is sometimes called the “closed phase,” regardless of Modal and Falsetto Registers,” Journal of the Acoustical the actual occurrence of glottal closure. Society of America 129, no. 6 (June 2011): 3955–3963; Elodie 15. Herbst et al. Joliveau, John Smith, and Joe Wolfe, “Acoustics: Tuning of Vocal Tract Resonance by ,” Nature 427, no. 16. Herbst et al.; Christian T. Herbst, David M. Howard, and Jan 6970 (January 2004): 116; Brad Story, “Vowel Acoustics for G. Švec, “Te Sound Source in Singing—Basic Principles and Speaking and Singing,” Acta Acustica united with Acoustica Muscular Adjustments for Fine-Tuning Vocal Timbre,” in G. 90 (July/August 2004): 629–640; Johan Sundberg and Jörgen Welch, D. M. Howard, and J. Nix, eds., Te Oxford Handbook Skoog, “Dependence of Jaw Opening on Pitch and Vowel of Singing (Oxford, UK: Oxford University Press, in print). in Singers,” Journal of Voice 11, no. 3 (September 1997): 17. In a previously conducted pilot study, these sound qualities 301–306; Johan Sundberg, “Articulatory Interpretation of were labelled “naïve falsetto” (aBducted falsetto), “counter- Te ‘Singing Formant’,” Journal of the Acoustical Society of falsetto” (aDducted falsetto), “lyrical chest” (aBducted America 55, no. 4 (April 1974): 838–844. chest), and “full chest” (aDducted chest); see Note 19. 4. Johan Sundberg, “Te Voice as a Sound Generator,” in J. 18. Murray Morrison, Linda A. Rammage, Gilles M. Belisle, Sundberg et al., Research Aspects on Singing (Stockholm: C. Bruce Pullan, and Hamish Nichol, “Muscular Tension Royal Swedish Academy of , 1981), 6–14. Dysphonia,” Journal of Otolaryngology 12, no. 5 (October 5. Ingo R. Titze, Principles of Voice Production, Second Printing 1983): 302–306. (Salt Lake City: National Center for Voice and Speech, 2000), 19. Christian T. Herbst, Sten Ternström, and Jan G. Švec, 281–305. To keep this manuscript concise, this discussion is “Investigation of Four Distinct Glottal Confgurations in limited to the two main registers: falsetto and chest. While Classical Singing—A Pilot Study,” Journal of the Acoustical vocal fry and are recognized as singing voice Society of America 125, no. 3 (March 2009): EL104-EL09. registers in their own right, they can only be produced at the extreme ends of the human fundamental frequency range 20. Tis is a nonexhaustive list, intended solely to provide some and are less ofen employed during singing performance. basic orientation. Please note that some of the singing styles could also make use of other phonation types. 6. Titze, 17; Minoru Hirano, “Morphological Structure of the Vocal Cord as a Vibrator and Its Variations,” Folia 21. We readily acknowledge that many pedagogues use the term Phoniatrica 26, no. 2 (1974): 89–94. “head” or “mixed head” for the range immediately above the primo passaggio in classical female singing. However, 7. Willard R. Zemlin, Speech and Hearing Science—Anatomy in order to maintain the simplicity of the model proposed and Physiology, 4th ed. (Boston: Allyn and Bacon, 1998), here, the latter is limited to the two registers chest and fal- 104–119. setto, which can easily be distinguished from each other by 8. Christian T. Herbst, Qingjun Qiu, Harm K. Schutte, and measuring the closed or contact quotient. Please note that Jan G. Švec, “Membranous and Cartilaginous Vocal Fold the proposed model allows for gradual changes is muscle Adduction in Singing,” Journal of the Acoustical Society of activity, thus allowing for a “mixed register.” For an in- America 129, no. 4 (April 2011): 2253–2262. depth discussion of the mixed register of “voix mixte,” see 9. Titze, 126. Janice Chapman, Singing and Teaching Singing. A Holistic

January/February 2014 307 Christian T. Herbst and Jan G. Švec

Approach to Classical Voice (San Diego, Oxford, Brisbane: visualize chaotic phenomena in nonlinear systems, and EGG wavegrams, Plural Publishing, 2006), 1–22. a technique for visualizing and analyzing vocal fold vibration properties. At the Laboratory of Bio-Acoustics at the University of Vienna he has created a setup for excised larynx experiments. The gathered data are used to Christian T. Herbst, born 1970, earned a master’s degree in classical perform cross-species comparisons of physical sound production mecha- voice pedagogy from the Mozarteum University, Salzburg, Austria. During nisms in mammals. In this context Christian recently published a paper his career as voice pedagogue (Tölzer Knabenchor, among others), about the low frequency physical sound production mechanisms in African Christian took interest in the physics and physiology of the voice. After a elephants in the prestigious Science journal. www.christian-herbst.org one year stay as visiting researcher at the Center for Computer Research in Music and Acoustics (CCRMA), Stanford University, he decided to put Jan G. Švec, PhD is an internationally renowned Czech physicist per- voice science as his future professional focus. He received his PhD in forming basic research on production of human voice. He has worked as biophysics from the Department of Biophysics, University of Olomouc, a research scientist at the Center for Communication Disorders (Medical Czech Republic, under the supervision of Dr. Jan Švec. From 2009 to Healthcom) in Prague, the Czech Republic, at the National Center for Voice 2013, Christian worked at the Department of Cognitive Biology, University and Speech in Denver, CO, USA, and at the University of Groningen, the of Vienna. He currently holds a position as postdoctoral researcher at the Netherlands. Currently he is at the Palacky University Olomouc, the Czech Department of Biophysics (Jan Švec group), University of Olomouc. He Republic, and serves also as an associate research scientist at the Voice is an editorial board member of Logopedics, Phoniatrics and Vocology, Centre Prague. He designed videokymography, the method for high-speed and resident editor of the Journal of Voice. visualization of vocal fold vibrations, which is used for advanced diagnosis Christian has published several prize-winning papers on laryngeal con- of voice disorders. He collaborates with numerous research teams across fgurations in singing and on electroglottography (EGG), a noninvasive the world and lectures world wide. From 2004 to 2011 he served as technique to gather physiological information of laryngeal dynamics the chairman of the Voice Committee of the International Association of during vocalization. He has invented the phasegram, a new method to Logopedics and Phoniatrics (IALP).

ADVERTISE WITH NATS! Journal of Singing – Email Blasts – Mailing Lists – Website Advertising

Advertising with NATS allows you to reach 7,300+ NATS members, university subscribers, webpage viewers, and music students all over the world.

CUSTOMIZE! Maximize your marketing budget by contacting us to customize an advertising package for your next marketing campaign. Advertising packages include print, web, postal, and email advertising!

Mention this ad and receive up to 15% off pricing on your next advertising campaign!

For information about advertising with NATS contact: Tom Strother Phone: 904-992-9101 x304 Email: [email protected]

308 Journal of Singing