Vocal Tract Shapes in Different Singing Functions Used in Musical Theater Singing—A Pilot Study *Matthias Echternach, †Lisa Popeil, *,‡Louisa Traser, §Sascha Wienhausen, and *Bernhard Richter, *zFreiburg and xOsnabr€uck, Germany, and yLos Angeles, California
Summary: Objective. Singing styles in Musical Theater singing might differ in many ways from Western Classical singing. However, vocal tract adjustments are not understood in detail. Material and Methods. Vocal tract shapes of a single professional Music Theater female subject were analyzed concerning different aspects of singing styles using dynamic real-time magnetic resonance imaging technology with a frame rate of 8 fps. The different tasks include register differences, belting, and vibrato strategies. Results. Articulatory differences were found between head register, modal register, and belting. Also, some vibrato strategies (‘‘jazzy’’ vibrato) do involve vocal tract adjustments, whereas others (classical vibrato) do not. Conclusions. Vocal tract shaping might contribute to the establishment of different singing functions in Musical Theater singing. Key Words: Musical Theater–Belting–Vocal tract–Magnetic resonance imaging–Vibrato–Chest voice–Head voice– Modal register. INTRODUCTION Musical Theater singing differs in numerous ways from Western Classical singing. One reason could be that, in contrast to most classical voice performances, the use of microphones and electronic amplification are now commonly used by Musical Theater singers. This allows for the use of ‘‘uneconomic’’ voice qualities such as breathy or rough voice (as heard in Rock and Pop styles) to be used as valid aesthetic choices. As Musical Theater in America developed in the early 20th Century, the lack of amplification and the aesthetic of music made for and enjoyed by the common man required a loud singing technique, which differed in character from traditional classical vocal ideals. Such technique was often denoted as belting. Since then, the term belting has been often used and defined in many different ways. Today, there is no agreement in voice pedagogy nor in the scientific world how belting should be defined. In addition, there are descriptions of subtypes of belting such as ringy, brassy, nasal, speech-like, and heavy.1 Some authors have postulated that belting might be related to a register function.2,3 In this respect, it is assumed that belting is considered an extension of the modal or chest register to higher fundamental frequencies (F0s). However, belting could also be related to aesthetic or timbral aspects of singing production. Compared with Western Classical singing technique, it has been shown that belting exhibits differences in all aspects of voice production. It is often assumed that belting is associated with a loud voice4–7 and high subglottic pressure.7,8 In a recent Accepted for publication January 20, 2014. This work was supported by the Deutsche Forschungsgesellschaft (DFG), Grant Ri1050/ 4-1. Presented in part at the Annual Meeting of the Voice Foundation; May 30, 2012 to July, 3, 2012; Philadelphia, Pennsylvania. From the *Institute of Musicians’ Medicine, Freiburg University Medical Center, Freiburg, Germany; yVoiceworks, Los Angeles, California; zDepartment of Otorhinolaryngology, Freiburg University Medical Center, Freiburg, Germany; and the xInstitute of Music, Osnabr€ uck University of Applied Science, Osnabr€uck, Germany. Address correspondence and reprint requests to Matthias Echternach, Institute of Musicians’ Medicine, Freiburg University Medical Center, Breisacher Str. 60, 79106 Freiburg, Germany. E-mail: [email protected] Journal of Voice, Vol. -, No. -, pp. 1-7 0892-1997/$36.00 Ó 2014 The Voice Foundation http://dx.doi.org/10.1016/j.jvoice.2014.01.011
study by Sundberg et al,9 however, it was shown that the degree of subglottic pressure varied according to the substyle of belting. The authors observed high subglottic pressure only for heavy belt, whereas the values for ringy and brassy belting substyles were within the range found in the female classical singing style.9 Many studies have focused on the voice source. Using flow glottograms, it was found that the voice source fundamental was stronger for the female classical voice (head register) compared with heavy belt.9 Also, the closed quotient was greater for the belting substyles.9 This last finding agrees with numerous electroglottographic studies in which the contact quotient was greater for belting.3,4,10,11 Lebowitz and Baken12 were not able to confirm such differences between ‘‘belt’’ and ‘‘legit’’ in their study. Legit is a term often used to denote a classically based technique, primarily using head register in females, used in Musical Theater. An increased speed quotient for belting was also found.12 As a consequence, it might be safe to assume that vocal fold oscillatory patterns differ between belting and legit singing styles. It can also be assumed that different registers are being used and that different muscle actions are being used. Indeed, in electromyographic studies by Estill,4 it was found that vocalis muscle activity was greater for belting in comparison with Western Classical singing. Furthermore, muscles crucial for vocal tract shaping showed different activation for belting in contrast to female operatic (classical) singing.4 Later, Kochis-Jennings et al13 found different laryngeal muscle activities for what the authors denoted as chest, chestmix, headmix, and head. It has been often assumed that there are special vocal tract configurations associated with belting. Osborne6 has pointed out that belting is associated with an open vowel configuration. In contrast to operatic singing, some authors observed a higher first formant frequency for belting.3,7,9 Schutte and Miller3 suggested that this would be caused by a higher larynx position for belting. They postulated that this might be necessary to match first formant to the second harmonic for open vowels. A comparable tuning strategy was also reported, recently, by Bourne and Garnier14 for the ‘‘chesty belt,’’ whereas legit failed to show
2 such a tuning strategy. However, Sundberg et al9 were not able to confirm these results using inverse filtering techniques. Other tuning strategies were found for /i/ and /e/ by Bestebreurtje and Schutte.15 Furthermore, tongue position may be an integral element in belting.5,16 Here, it was observed that the tongue was elevated. Titze and Worley17 suggested from their modeling experiments that classical singing is associated with an inverse trumpet shape of the vocal tract, whereas belting is related to a trumpet shape configuration. Because modifications of vocal tract shape are mostly not visible from the outside, differences between Musical Theater singing and Classical singing are not yet understood in detail. Most studies on Musical Theater singing have focused on belting with several authors investigating an alternate speechlike type of phonation. However, in contrast to belting for which sometimes a lack of vibrato is postulated,5,6 there are also different vibrato techniques used in Musical Theater singing when compared with Western Classical singing. Very little investigation of vocal tract shape differences in varying Musical Theater vibrato types has been undertaken so far. Dynamic real-time magnetic resonance imaging (MRI) techniques for some time have offered the opportunity to analyze two-dimensional vocal tract shapes in singing. Such techniques have been successfully applied in studies concerning different singing functions such as high female singing,18,19 registers,20–22 and yodeling.23 The present study was performed to analyze vocal tract shape comparing different vocal functions used in Musical Theater singing in a single professional Musical Theater female subject. It has been hypothesized that vocal tract shape changes when varying singing tasks. Therefore, it was expected that (1) the vocal tract configuration would differ between heavy belt on one hand and head voice on the other, (2) the vocal tract shape differs between modal and head register with respect to different vowel conditions, and (3) there would be vocal tract shape differences when using different vibrato strategies.
MATERIAL AND METHODS All experiments were performed in a single professional female Musical Theater singer and pedagogue subject. The subject was the same subject analyzed by Sundberg et al,9 in a previous study. At the time of the recordings, the subject did not complain of any vocal symptoms. Laryngeal pathologies were excluded by videostroboscopic examinations. Experiments were performed radiologically with the 3.0 T TIM TRIO (Siemens, Erlangen, Germany) MRI device in a supine position.22 Real-time MRI was performed with a temporal resolution of 8 images per second. The detailed MRI setup was identical to previous studies.22 The audio signal was recorded by means of an optic microphone (Fa. MR confon, Magdeburg, Germany, OptiMRI Noise Reduction Software, Fa. Optoacoustics Ltd., Or-Yehuda, Israel). As in our previous studies, the subject was provided with the audio signal over headphones (Fa. MR confon, Magdeburg, Germany) as acoustic feedback.
Journal of Voice, Vol. -, No. -, 2014
In the first experiment, the subject was asked to sing ascending pitches from G3 up to C6 on the vowel /e/ in either head voice or belting, respectively. Because the definition of belting might be considered as heterogeneous in the literature, the authors accepted the subject’s own definition of the heavy belt substyle, described before.9 Because other researchers might differ with this definition, the authors provide representative online Supplementary video material of all experiments including the original audio signal. The vowel /e/ was chosen because some researchers and many Musical Theater singers report that belting is best accomplished using this or the vowel /a/. However, because register and vowel choice might contribute to different Musical Theater singing styles, the same subject was asked in a second experiment to sing on the vowels /a, e, i, o, u/ in a descending triad (C5, G4, E4, C4) either in modal or in head register, respectively. Because it has been widely accepted that varying widths of the pharyngeal cavity impact vocal timbre, the subject was asked to sustain the vowel /a/ on C4 with a (1) neutral pharynx position, (2) constricted pharynx position, and (3) wide pharynx position. Finally, because vibrato type might differ between styles, subject was asked to produce two different kinds of vibrato on G4 on the vowel /ae/. The subject herself denoted the vibrato styles as (1) ‘‘classical’’ and (2) ‘‘jazzy’’ vibrato. During the MRI recordings, four experts were present as follows: (1) a singing pedagogue and full professor of Musical Theater singing, (2) a professional Musical Theater singer and medical student, (3) and (4) both otolaryngologists and Western Classical singers. The experts checked that the subject produced the desired tasks. Directly after each sequence, the subject was asked if she felt that she had produced the tasks acceptably. Only sequences found acceptable by all experts as well as by the singer herself were subsequently analyzed. In each frame of the MRI material, a series of measures were taken, as described previously.22 In addition, the supralaryngeal diameter was defined as the shortest distance between the epiglottis and the arytenoid cartilages. In the audio recording, the F0 was identified using PRAAT software (University of Amsterdam, the Netherlands). The relationship between F0, task, and each of the MRI measures was then analyzed. Due to the fact that this study is a single subject study, comparing statistical analysis for mean values for the distances and angles with respect to the registers was not feasible.
RESULTS In the first experiment, on the vowel /e/, the subject showed clear vocal tract shape differences between the head voice and belting tasks for the different pitches (Figure 1 and online Supplementary Video 1). Results showed greater lip opening, jaw opening, and a smaller pharynx width for the belting task. Furthermore, the larynx position was found to be much higher in belting. Finally, the epilaryngeal tube, as measured by means of the supralaryngeal diameter, was found much narrower for the belting task. The changes with rising pitch were
Matthias Echternach, et al
Vocal Tract in Musical Theater Singing
3
FIGURE 1. Articulatory data for the belting and head voice, respectively, on vowel /e/ in relation to F0. rather homogeneous for the different tasks, belting and head voice, respectively. Because belting might be, in part, related to vocal registers, it was of interest to compare modal register with head voice. In addition, we wanted to determine if vowel shapes alter as vocal register changes. Therefore, an experiment was performed on lower pitches so that nearly all vowels could be established. The online Supplementary Video 2 shows the MRI movie of the experiment and Figure 2 shows the related articulatory data. Here, it is shown that for most vowels, the modal register is associated with a greater lip opening and greater pharynx width. The larynx position was lower and the laryngeal angle was smaller, whereas the supralaryngeal diameter was greater for head voice. As described before, the pharynx width might be relevant for different singing styles. The subject was, therefore, asked to produce, on a sustained pitch and on the vowel /a/, a neutral, narrow, and wide pharynx position (online Supplementary Video 3). Figure 3 shows representative vocal tract shapes concerning these tasks. As shown in Figure 4, the pharynx width was much higher for the neutral position in relation to the narrow position. Interestingly, the wide pharynx position was associated with a lower value for the measure of the distance between back of the tongue and the pharyngeal back wall. However, the larynx position was much lower and the lower pharynx was much widened (Figures 3 and 4). As a final task, the subject was asked to produce different types of vibrato. The subject herself denoted the first as classical and the second as a jazzy vibrato. As shown in online Supplementary Video 4, the classical vibrato exhibited much less vocal tract modification compared with the jazzy vibrato. This was also reflected in some articulatory data, as shown in Figure 5. Here, the larynx position and the supralaryngeal diameter showed dramatically increased changes during the jazzy
vibrato. However, as shown in the spectrum, the vibrato amplitude and frequency was nearly comparable between the two different vibrato styles. During the experiment, the subject felt that, for some of the tasks, the supine position, the noisy environment, and especially the constrained neck bracing somehow negatively affected performance. Even so, it was the experts’ opinion that the sound produced during the experiments was very close to the sound subject produced in real conditions and the subject agreed in every experiment that the task was fulfilled in a valid way. The online Supplementary Videos 1–4 (corresponding to experiments 1–4) include the original audio files and relay the acoustic sound made during the experiment.
DISCUSSION This study presents articulatory data of a single professional singer subject in two singing styles used in Musical Theater singing, specifically belting and head voice. This article shows that there are numerous vocal tract shape differences associated with different singing styles. Without microphones and electronic amplification, singers must create carrying power to fill a theater with sound. To this end, Western Classical singers often use formant clustering24 or formant tuning strategies.25 At the beginning of the 20th Century, a special vocal technique was developed in Musical Theater singing, which is associated with a loud and powerful voice and has been denoted as belting. This technique is not only used for enhancing vocal power but also for aesthetic reasons, as belting might be considered as an extension of natural yelling and projected speech. As a consequence, belting is currently used in Pop and Rock singing as well, although the use of microphones in these styles is standard. However, there is no uniform definition of belting among
4
Journal of Voice, Vol. -, No. -, 2014
FIGURE 2. Articulatory data for the modal and head register for different vowel conditions and pitches. voice pedagogues and scientists. Furthermore, there are descriptions of subtypes of belting.1,9 Therefore, it seems rather difficult to analyze such a phenomenon because people could disagree that the analyzed samples really
reflect belting. This explains why the online Supplementary Video material has been included in this article. Although the subject and the experts agreed that the first belting task reflected the commonly held definition, the authors are
Matthias Echternach, et al
Vocal Tract in Musical Theater Singing
5
FIGURE 3. Representative magnetic resonance images for the neutral (left), narrow (middle), and wide (right) pharynx position. open to readers’ interpretation if the performance indeed demonstrates belting or not. Many studies have been performed to study the aerodynamic and voice source properties of belting. In contrast, there are only few studies focusing on the vocal tract. Lawrence2 and Sundberg et al7 observed a higher larynx position for belting. In contrast, Lovetri et al26 found that belting could also be produced with a relatively low larynx position. However, the presented MRI data seem to support the observation of a higher laryngeal position for the measured single subject. As pointed out in the introduction, Titze and Worley17 suggested from their modeling experiments that classical singing is associated with an inverse trumpet shape of the vocal tract, whereas belting is related to a trumpet shape configuration. Our data also show that the pharynx width and the epilaryngeal tube were much smaller and the lip opening greater for the belting in comparison with head voice, suggesting a more trumpet-like vocal tract shape. Our results are, therefore, in agreement with laryngoscopic observations by Lawrence2 and Sundberg et al7 who observed a constriction of the epilaryngeal tube and the pharynx. However, it should be mentioned here that our material is limited to the two-dimensional mid-
sagittal plane. Therefore, the medial-lateral dimension is not reflected in our material. Some authors have suggested increased nasality for belting5 and others have not.16 In our single subject, the velopharyngeal area was closed all the time during the experiment suggesting that the nasal cavities might not be part of the resonance system. It is sometimes assumed that the belting is a register function in which the modal register is raised over the region where usually the register transition from modal to head register would occur. In our second experiment, we found several differences between the use of modal and head voice in the articulatory data. For example, in head voice, a lower larynx angle was seen which could reflect more laryngeal tilting due to a dominance of the cricothyroid muscle. Similar to the belting condition of the first experiment, the modal register was associated with greater lip opening, higher larynx position, slightly higher tongue dorsum, and smaller supralaryngeal diameter. The pharynx width was greater for the modal register for most vowel conditions. Concerning the /e/ vowel condition, however, the pharynx width showed nearly no difference between registers. This is in contrast to the belting data of the first experiment. Therefore, it seems reasonable to assume
FIGURE 4. Articulatory data for the neutral (gray), narrow (black) and wide (white) pharynx position.
6
Journal of Voice, Vol. -, No. -, 2014
FIGURE 5. Acoustic spectrum of the audio signal for the ‘‘classical’’ vibrato (upper row), and ‘‘jazzy’’ vibrato (second row) with the corresponding data for the larynx position and supralaryngeal diameter.
that for the measured subject, belting is not exactly the same as a modal register. In general, vocal tract shape differences between modal and head registers were visible for most vowel samples. Therefore, it seems as if the differences between the analyzed registers were nearly vowel-independent. As pointed out previously, the pharynx width might be an important contributor to the differences in the results. In our third experiment, there are a lot of articulatory differences associated with varying pharynx widths. Interestingly, when the subject was asked to produce a wide pharynx position, the smallest distance between the tongue and the pharyngeal back wall was even smaller. The subject produced a greater pharynx width by lowering the larynx. As a consequence, the sound produced resembled a covered voice in Western Classical singing technique. In some styles of Musical Theater singing, vibrato technique can differ. Belting, as a style, is sometimes thought to be associated with no or little vibrato.5,6 In Western Classical singing,
it is expected that vibrato is a consequence of laryngeal muscle activities.27 Therefore, it could be assumed that vocal tract changes play a minor role in the classical vibrato. As shown in our material here, the classical vibrato task shows nearly no changes in the vocal tract configuration. In contrast, the jazzy vibrato was associated with changes of the larynx position and the supralaryngeal diameter. As shown in the online Supplementary Video 3, there is also strong, horizontal activity around the hyoid bone, which is not reflected in the articulatory measurements. Despite a comparable vibrato amplitude and frequency, these differences might contribute to the perceptive differences of both vibrato styles, classical and jazzy, respectively. Finally, it should be mentioned that, similar to our previous investigations, these experiments also are limited due to the performance in supine position. In a recent study using a very low Tesla-MRI device, it was shown that for professional tenors, the position (supine vs upright) influence articulatory data for singing conditions by only a small amount.28 In this respect,
Matthias Echternach, et al
Vocal Tract in Musical Theater Singing
it was shown that the larynx position was lower in the upright position.28 Because larynx position seems to be of importance in belting tasks, the subjective sense of hindrance in the neck region might have influenced the data. However, both the subject and the experts agreed after each experiment that the tasks were fulfilled in a valid way. CONCLUSION In singing, there are numerous articulatory differences between head voice and belting voice production. These include differences in lip opening, pharynx width, larynx position, tongue dorsum height, and the supralaryngeal diameter. There are also articulatory differences between modal register and belting. Therefore, it should be pointed out that belting does not equal the use of modal register above the lower passaggio. Some vibrato styles (jazzy) include vocal tract adjustments, whereas others (classical) do not. Acknowledgments The authors would like to thank Martin Schumacher, MD and Hansj€ org Mast for technical assistance and Judith Kroll, for discussions. There are no financial interests. APPENDIX Supplementary material Supplementary videos associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.jvoice.2014. 01.011. REFERENCES 1. Popeil L. The multiplicity of belting. J Singing. 2007;64:77–80. 2. Lawrence V. Laryngological observations on belting. J Res Singing. 1979; 2:26–28. 3. Schutte HK, Miller DG. Belting and pop, nonclassical approaches to the female middle voice: some preliminary considerations. J Voice. 1993;7: 142–150. 4. Estill J. Belting and classic voice quality: some physiological differences. Med Probl Perform Art. 1988;3:37–43. 5. Miles B, Hollien H. Wither belting. J Voice. 1990;4:64–70. 6. Osborne C. The Broadway voice, part I: just singin’ in the pain. Hi-Fidelity. 1979;29:57–65. 7. Sundberg J, Gramming P, Lovetri J. Comparisons of pharynx, source, formant, and pressure characteristics in operatic and musical theatre singing. J Voice. 1993;7:301–310.
7
8. Hein M. Die Gesangtechnik des Beltings – eine Studie u€ber Atemdruck, Lungenvolumen und Atembewegungen. Hamburg, Germany: Dissertation, Hamburg University; 2010. 9. Sundberg J, Thalen M, Popeil L. Substyles of belting: phonatory and resonatory characteristics. J Voice. 2012;26:44–50. 10. Evans M, Howard DM. Larynx CQ in female belt and opera qualities. Voice. 1993;2:7–14. 11. Barlow C, Lovetri J. Closed quotient and spectral measures of female adolescent singers in different singing styles. J Voice. 2010;24:314–318. 12. Lebowitz A, Baken RJ. Correlates of the belt voice: a broader examination. J Voice. 2011;25:159–165. 13. Kochis-Jennings KA, Finnegan EM, Hoffman HT, Jaiswal S. Laryngeal muscle activity and vocal fold adduction during chest, chestmix, headmix, and head registers in females. J Voice. 2012;26:182–193. 14. Bourne T, Garnier M. Physiological and acoustic characteristics of the female Music Theater voice. J Acoust Soc Am. 2012;131:1586–1594. 15. Bestebreurtje ME, Schutte HK. Resonance strategies for the belting style: results of a single female subject study. J Voice. 2000;14: 194–204. 16. Bourne T, Kenny D. Perceptual Descriptions of Legit and Belt Voice Qualities in Musical Theater. Perth, Australia: Proceedings of the ANATS Conference; 2008. 17. Titze IR, Worley AS. Modeling source-filter interaction in belting and highpitched operatic male singing. J Acoust Soc Am. 2009;126:1530. 18. Sundberg J. Articulatory configuration and pitch in a classically trained soprano singer. J Voice. 2009;23:546–551. 19. Echternach M, Sundberg J, Arndt S, Markl M, Schumacher M, Richter B. Vocal tract in female registers—a dynamic real-time MRI study. J Voice. 2010;24:133–139. 20. Echternach M, Sundberg J, Arndt S, et al. Vocal tract and register changes analysed by real-time MRI in male professional singers—a pilot study. Logoped Phoniatr Vocol. 2008;33:67–73. 21. Echternach M, Sundberg J, Markl M, Richter B. Professional opera tenor’s vocal tract configurations in registers. Folia Phoniatr Logop. 2010;62: 278–287. 22. Echternach M, Traser L, Markl M, Richter B. Vocal tract configurations in male alto register functions. J Voice. 2011;25:670–677. 23. Echternach M, Markl M, Richter B. Vocal tract configurations in yodelling—prospective comparison of two Swiss yodeller and two nonyodeller subjects. Logoped Phoniatr Vocol. 2011;36:109–113. 24. Sundberg J. Articulatory interpretation of the ‘‘singing formant’’. J Acoust Soc Am. 1974;55:838–844. 25. Sundberg J. Formant technique in a professional female singer. Acustica. 1975;32:89–96. 26. Lovetri J, Lesh S, Woo P. Preliminary study on the ability of trained singers to control the intrinsic and extrinsic laryngeal musculature. J Voice. 1999; 13:219–226. 27. Sundberg J. The Science of the Singing Voice. Dekalb, IL: Northern Illinois University Press; 1987. 28. Traser L, Burdumy M, Richter B, Vicari M, Echternach M. The effect of supine and upright position on vocal tract configurations during singing—a comparative study in professional tenors. J Voice. 2013;27: 141–148.
Summary: Objective. Singing styles in Musical Theater singing might differ in many ways from Western Classical singing. However, vocal tract adjustments are not understood in detail. Material and Methods. Vocal tract shapes of a single professional Music Theater female subject were analyzed concerning different aspects of singing styles using dynamic real-time magnetic resonance imaging technology with a frame rate of 8 fps. The different tasks include register differences, belting, and vibrato strategies. Results. Articulatory differences were found between head register, modal register, and belting. Also, some vibrato strategies (‘‘jazzy’’ vibrato) do involve vocal tract adjustments, whereas others (classical vibrato) do not. Conclusions. Vocal tract shaping might contribute to the establishment of different singing functions in Musical Theater singing. Key Words: Musical Theater–Belting–Vocal tract–Magnetic resonance imaging–Vibrato–Chest voice–Head voice– Modal register. INTRODUCTION Musical Theater singing differs in numerous ways from Western Classical singing. One reason could be that, in contrast to most classical voice performances, the use of microphones and electronic amplification are now commonly used by Musical Theater singers. This allows for the use of ‘‘uneconomic’’ voice qualities such as breathy or rough voice (as heard in Rock and Pop styles) to be used as valid aesthetic choices. As Musical Theater in America developed in the early 20th Century, the lack of amplification and the aesthetic of music made for and enjoyed by the common man required a loud singing technique, which differed in character from traditional classical vocal ideals. Such technique was often denoted as belting. Since then, the term belting has been often used and defined in many different ways. Today, there is no agreement in voice pedagogy nor in the scientific world how belting should be defined. In addition, there are descriptions of subtypes of belting such as ringy, brassy, nasal, speech-like, and heavy.1 Some authors have postulated that belting might be related to a register function.2,3 In this respect, it is assumed that belting is considered an extension of the modal or chest register to higher fundamental frequencies (F0s). However, belting could also be related to aesthetic or timbral aspects of singing production. Compared with Western Classical singing technique, it has been shown that belting exhibits differences in all aspects of voice production. It is often assumed that belting is associated with a loud voice4–7 and high subglottic pressure.7,8 In a recent Accepted for publication January 20, 2014. This work was supported by the Deutsche Forschungsgesellschaft (DFG), Grant Ri1050/ 4-1. Presented in part at the Annual Meeting of the Voice Foundation; May 30, 2012 to July, 3, 2012; Philadelphia, Pennsylvania. From the *Institute of Musicians’ Medicine, Freiburg University Medical Center, Freiburg, Germany; yVoiceworks, Los Angeles, California; zDepartment of Otorhinolaryngology, Freiburg University Medical Center, Freiburg, Germany; and the xInstitute of Music, Osnabr€ uck University of Applied Science, Osnabr€uck, Germany. Address correspondence and reprint requests to Matthias Echternach, Institute of Musicians’ Medicine, Freiburg University Medical Center, Breisacher Str. 60, 79106 Freiburg, Germany. E-mail: [email protected] Journal of Voice, Vol. -, No. -, pp. 1-7 0892-1997/$36.00 Ó 2014 The Voice Foundation http://dx.doi.org/10.1016/j.jvoice.2014.01.011
study by Sundberg et al,9 however, it was shown that the degree of subglottic pressure varied according to the substyle of belting. The authors observed high subglottic pressure only for heavy belt, whereas the values for ringy and brassy belting substyles were within the range found in the female classical singing style.9 Many studies have focused on the voice source. Using flow glottograms, it was found that the voice source fundamental was stronger for the female classical voice (head register) compared with heavy belt.9 Also, the closed quotient was greater for the belting substyles.9 This last finding agrees with numerous electroglottographic studies in which the contact quotient was greater for belting.3,4,10,11 Lebowitz and Baken12 were not able to confirm such differences between ‘‘belt’’ and ‘‘legit’’ in their study. Legit is a term often used to denote a classically based technique, primarily using head register in females, used in Musical Theater. An increased speed quotient for belting was also found.12 As a consequence, it might be safe to assume that vocal fold oscillatory patterns differ between belting and legit singing styles. It can also be assumed that different registers are being used and that different muscle actions are being used. Indeed, in electromyographic studies by Estill,4 it was found that vocalis muscle activity was greater for belting in comparison with Western Classical singing. Furthermore, muscles crucial for vocal tract shaping showed different activation for belting in contrast to female operatic (classical) singing.4 Later, Kochis-Jennings et al13 found different laryngeal muscle activities for what the authors denoted as chest, chestmix, headmix, and head. It has been often assumed that there are special vocal tract configurations associated with belting. Osborne6 has pointed out that belting is associated with an open vowel configuration. In contrast to operatic singing, some authors observed a higher first formant frequency for belting.3,7,9 Schutte and Miller3 suggested that this would be caused by a higher larynx position for belting. They postulated that this might be necessary to match first formant to the second harmonic for open vowels. A comparable tuning strategy was also reported, recently, by Bourne and Garnier14 for the ‘‘chesty belt,’’ whereas legit failed to show
2 such a tuning strategy. However, Sundberg et al9 were not able to confirm these results using inverse filtering techniques. Other tuning strategies were found for /i/ and /e/ by Bestebreurtje and Schutte.15 Furthermore, tongue position may be an integral element in belting.5,16 Here, it was observed that the tongue was elevated. Titze and Worley17 suggested from their modeling experiments that classical singing is associated with an inverse trumpet shape of the vocal tract, whereas belting is related to a trumpet shape configuration. Because modifications of vocal tract shape are mostly not visible from the outside, differences between Musical Theater singing and Classical singing are not yet understood in detail. Most studies on Musical Theater singing have focused on belting with several authors investigating an alternate speechlike type of phonation. However, in contrast to belting for which sometimes a lack of vibrato is postulated,5,6 there are also different vibrato techniques used in Musical Theater singing when compared with Western Classical singing. Very little investigation of vocal tract shape differences in varying Musical Theater vibrato types has been undertaken so far. Dynamic real-time magnetic resonance imaging (MRI) techniques for some time have offered the opportunity to analyze two-dimensional vocal tract shapes in singing. Such techniques have been successfully applied in studies concerning different singing functions such as high female singing,18,19 registers,20–22 and yodeling.23 The present study was performed to analyze vocal tract shape comparing different vocal functions used in Musical Theater singing in a single professional Musical Theater female subject. It has been hypothesized that vocal tract shape changes when varying singing tasks. Therefore, it was expected that (1) the vocal tract configuration would differ between heavy belt on one hand and head voice on the other, (2) the vocal tract shape differs between modal and head register with respect to different vowel conditions, and (3) there would be vocal tract shape differences when using different vibrato strategies.
MATERIAL AND METHODS All experiments were performed in a single professional female Musical Theater singer and pedagogue subject. The subject was the same subject analyzed by Sundberg et al,9 in a previous study. At the time of the recordings, the subject did not complain of any vocal symptoms. Laryngeal pathologies were excluded by videostroboscopic examinations. Experiments were performed radiologically with the 3.0 T TIM TRIO (Siemens, Erlangen, Germany) MRI device in a supine position.22 Real-time MRI was performed with a temporal resolution of 8 images per second. The detailed MRI setup was identical to previous studies.22 The audio signal was recorded by means of an optic microphone (Fa. MR confon, Magdeburg, Germany, OptiMRI Noise Reduction Software, Fa. Optoacoustics Ltd., Or-Yehuda, Israel). As in our previous studies, the subject was provided with the audio signal over headphones (Fa. MR confon, Magdeburg, Germany) as acoustic feedback.
Journal of Voice, Vol. -, No. -, 2014
In the first experiment, the subject was asked to sing ascending pitches from G3 up to C6 on the vowel /e/ in either head voice or belting, respectively. Because the definition of belting might be considered as heterogeneous in the literature, the authors accepted the subject’s own definition of the heavy belt substyle, described before.9 Because other researchers might differ with this definition, the authors provide representative online Supplementary video material of all experiments including the original audio signal. The vowel /e/ was chosen because some researchers and many Musical Theater singers report that belting is best accomplished using this or the vowel /a/. However, because register and vowel choice might contribute to different Musical Theater singing styles, the same subject was asked in a second experiment to sing on the vowels /a, e, i, o, u/ in a descending triad (C5, G4, E4, C4) either in modal or in head register, respectively. Because it has been widely accepted that varying widths of the pharyngeal cavity impact vocal timbre, the subject was asked to sustain the vowel /a/ on C4 with a (1) neutral pharynx position, (2) constricted pharynx position, and (3) wide pharynx position. Finally, because vibrato type might differ between styles, subject was asked to produce two different kinds of vibrato on G4 on the vowel /ae/. The subject herself denoted the vibrato styles as (1) ‘‘classical’’ and (2) ‘‘jazzy’’ vibrato. During the MRI recordings, four experts were present as follows: (1) a singing pedagogue and full professor of Musical Theater singing, (2) a professional Musical Theater singer and medical student, (3) and (4) both otolaryngologists and Western Classical singers. The experts checked that the subject produced the desired tasks. Directly after each sequence, the subject was asked if she felt that she had produced the tasks acceptably. Only sequences found acceptable by all experts as well as by the singer herself were subsequently analyzed. In each frame of the MRI material, a series of measures were taken, as described previously.22 In addition, the supralaryngeal diameter was defined as the shortest distance between the epiglottis and the arytenoid cartilages. In the audio recording, the F0 was identified using PRAAT software (University of Amsterdam, the Netherlands). The relationship between F0, task, and each of the MRI measures was then analyzed. Due to the fact that this study is a single subject study, comparing statistical analysis for mean values for the distances and angles with respect to the registers was not feasible.
RESULTS In the first experiment, on the vowel /e/, the subject showed clear vocal tract shape differences between the head voice and belting tasks for the different pitches (Figure 1 and online Supplementary Video 1). Results showed greater lip opening, jaw opening, and a smaller pharynx width for the belting task. Furthermore, the larynx position was found to be much higher in belting. Finally, the epilaryngeal tube, as measured by means of the supralaryngeal diameter, was found much narrower for the belting task. The changes with rising pitch were
Matthias Echternach, et al
Vocal Tract in Musical Theater Singing
3
FIGURE 1. Articulatory data for the belting and head voice, respectively, on vowel /e/ in relation to F0. rather homogeneous for the different tasks, belting and head voice, respectively. Because belting might be, in part, related to vocal registers, it was of interest to compare modal register with head voice. In addition, we wanted to determine if vowel shapes alter as vocal register changes. Therefore, an experiment was performed on lower pitches so that nearly all vowels could be established. The online Supplementary Video 2 shows the MRI movie of the experiment and Figure 2 shows the related articulatory data. Here, it is shown that for most vowels, the modal register is associated with a greater lip opening and greater pharynx width. The larynx position was lower and the laryngeal angle was smaller, whereas the supralaryngeal diameter was greater for head voice. As described before, the pharynx width might be relevant for different singing styles. The subject was, therefore, asked to produce, on a sustained pitch and on the vowel /a/, a neutral, narrow, and wide pharynx position (online Supplementary Video 3). Figure 3 shows representative vocal tract shapes concerning these tasks. As shown in Figure 4, the pharynx width was much higher for the neutral position in relation to the narrow position. Interestingly, the wide pharynx position was associated with a lower value for the measure of the distance between back of the tongue and the pharyngeal back wall. However, the larynx position was much lower and the lower pharynx was much widened (Figures 3 and 4). As a final task, the subject was asked to produce different types of vibrato. The subject herself denoted the first as classical and the second as a jazzy vibrato. As shown in online Supplementary Video 4, the classical vibrato exhibited much less vocal tract modification compared with the jazzy vibrato. This was also reflected in some articulatory data, as shown in Figure 5. Here, the larynx position and the supralaryngeal diameter showed dramatically increased changes during the jazzy
vibrato. However, as shown in the spectrum, the vibrato amplitude and frequency was nearly comparable between the two different vibrato styles. During the experiment, the subject felt that, for some of the tasks, the supine position, the noisy environment, and especially the constrained neck bracing somehow negatively affected performance. Even so, it was the experts’ opinion that the sound produced during the experiments was very close to the sound subject produced in real conditions and the subject agreed in every experiment that the task was fulfilled in a valid way. The online Supplementary Videos 1–4 (corresponding to experiments 1–4) include the original audio files and relay the acoustic sound made during the experiment.
DISCUSSION This study presents articulatory data of a single professional singer subject in two singing styles used in Musical Theater singing, specifically belting and head voice. This article shows that there are numerous vocal tract shape differences associated with different singing styles. Without microphones and electronic amplification, singers must create carrying power to fill a theater with sound. To this end, Western Classical singers often use formant clustering24 or formant tuning strategies.25 At the beginning of the 20th Century, a special vocal technique was developed in Musical Theater singing, which is associated with a loud and powerful voice and has been denoted as belting. This technique is not only used for enhancing vocal power but also for aesthetic reasons, as belting might be considered as an extension of natural yelling and projected speech. As a consequence, belting is currently used in Pop and Rock singing as well, although the use of microphones in these styles is standard. However, there is no uniform definition of belting among
4
Journal of Voice, Vol. -, No. -, 2014
FIGURE 2. Articulatory data for the modal and head register for different vowel conditions and pitches. voice pedagogues and scientists. Furthermore, there are descriptions of subtypes of belting.1,9 Therefore, it seems rather difficult to analyze such a phenomenon because people could disagree that the analyzed samples really
reflect belting. This explains why the online Supplementary Video material has been included in this article. Although the subject and the experts agreed that the first belting task reflected the commonly held definition, the authors are
Matthias Echternach, et al
Vocal Tract in Musical Theater Singing
5
FIGURE 3. Representative magnetic resonance images for the neutral (left), narrow (middle), and wide (right) pharynx position. open to readers’ interpretation if the performance indeed demonstrates belting or not. Many studies have been performed to study the aerodynamic and voice source properties of belting. In contrast, there are only few studies focusing on the vocal tract. Lawrence2 and Sundberg et al7 observed a higher larynx position for belting. In contrast, Lovetri et al26 found that belting could also be produced with a relatively low larynx position. However, the presented MRI data seem to support the observation of a higher laryngeal position for the measured single subject. As pointed out in the introduction, Titze and Worley17 suggested from their modeling experiments that classical singing is associated with an inverse trumpet shape of the vocal tract, whereas belting is related to a trumpet shape configuration. Our data also show that the pharynx width and the epilaryngeal tube were much smaller and the lip opening greater for the belting in comparison with head voice, suggesting a more trumpet-like vocal tract shape. Our results are, therefore, in agreement with laryngoscopic observations by Lawrence2 and Sundberg et al7 who observed a constriction of the epilaryngeal tube and the pharynx. However, it should be mentioned here that our material is limited to the two-dimensional mid-
sagittal plane. Therefore, the medial-lateral dimension is not reflected in our material. Some authors have suggested increased nasality for belting5 and others have not.16 In our single subject, the velopharyngeal area was closed all the time during the experiment suggesting that the nasal cavities might not be part of the resonance system. It is sometimes assumed that the belting is a register function in which the modal register is raised over the region where usually the register transition from modal to head register would occur. In our second experiment, we found several differences between the use of modal and head voice in the articulatory data. For example, in head voice, a lower larynx angle was seen which could reflect more laryngeal tilting due to a dominance of the cricothyroid muscle. Similar to the belting condition of the first experiment, the modal register was associated with greater lip opening, higher larynx position, slightly higher tongue dorsum, and smaller supralaryngeal diameter. The pharynx width was greater for the modal register for most vowel conditions. Concerning the /e/ vowel condition, however, the pharynx width showed nearly no difference between registers. This is in contrast to the belting data of the first experiment. Therefore, it seems reasonable to assume
FIGURE 4. Articulatory data for the neutral (gray), narrow (black) and wide (white) pharynx position.
6
Journal of Voice, Vol. -, No. -, 2014
FIGURE 5. Acoustic spectrum of the audio signal for the ‘‘classical’’ vibrato (upper row), and ‘‘jazzy’’ vibrato (second row) with the corresponding data for the larynx position and supralaryngeal diameter.
that for the measured subject, belting is not exactly the same as a modal register. In general, vocal tract shape differences between modal and head registers were visible for most vowel samples. Therefore, it seems as if the differences between the analyzed registers were nearly vowel-independent. As pointed out previously, the pharynx width might be an important contributor to the differences in the results. In our third experiment, there are a lot of articulatory differences associated with varying pharynx widths. Interestingly, when the subject was asked to produce a wide pharynx position, the smallest distance between the tongue and the pharyngeal back wall was even smaller. The subject produced a greater pharynx width by lowering the larynx. As a consequence, the sound produced resembled a covered voice in Western Classical singing technique. In some styles of Musical Theater singing, vibrato technique can differ. Belting, as a style, is sometimes thought to be associated with no or little vibrato.5,6 In Western Classical singing,
it is expected that vibrato is a consequence of laryngeal muscle activities.27 Therefore, it could be assumed that vocal tract changes play a minor role in the classical vibrato. As shown in our material here, the classical vibrato task shows nearly no changes in the vocal tract configuration. In contrast, the jazzy vibrato was associated with changes of the larynx position and the supralaryngeal diameter. As shown in the online Supplementary Video 3, there is also strong, horizontal activity around the hyoid bone, which is not reflected in the articulatory measurements. Despite a comparable vibrato amplitude and frequency, these differences might contribute to the perceptive differences of both vibrato styles, classical and jazzy, respectively. Finally, it should be mentioned that, similar to our previous investigations, these experiments also are limited due to the performance in supine position. In a recent study using a very low Tesla-MRI device, it was shown that for professional tenors, the position (supine vs upright) influence articulatory data for singing conditions by only a small amount.28 In this respect,
Matthias Echternach, et al
Vocal Tract in Musical Theater Singing
it was shown that the larynx position was lower in the upright position.28 Because larynx position seems to be of importance in belting tasks, the subjective sense of hindrance in the neck region might have influenced the data. However, both the subject and the experts agreed after each experiment that the tasks were fulfilled in a valid way. CONCLUSION In singing, there are numerous articulatory differences between head voice and belting voice production. These include differences in lip opening, pharynx width, larynx position, tongue dorsum height, and the supralaryngeal diameter. There are also articulatory differences between modal register and belting. Therefore, it should be pointed out that belting does not equal the use of modal register above the lower passaggio. Some vibrato styles (jazzy) include vocal tract adjustments, whereas others (classical) do not. Acknowledgments The authors would like to thank Martin Schumacher, MD and Hansj€ org Mast for technical assistance and Judith Kroll, for discussions. There are no financial interests. APPENDIX Supplementary material Supplementary videos associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.jvoice.2014. 01.011. REFERENCES 1. Popeil L. The multiplicity of belting. J Singing. 2007;64:77–80. 2. Lawrence V. Laryngological observations on belting. J Res Singing. 1979; 2:26–28. 3. Schutte HK, Miller DG. Belting and pop, nonclassical approaches to the female middle voice: some preliminary considerations. J Voice. 1993;7: 142–150. 4. Estill J. Belting and classic voice quality: some physiological differences. Med Probl Perform Art. 1988;3:37–43. 5. Miles B, Hollien H. Wither belting. J Voice. 1990;4:64–70. 6. Osborne C. The Broadway voice, part I: just singin’ in the pain. Hi-Fidelity. 1979;29:57–65. 7. Sundberg J, Gramming P, Lovetri J. Comparisons of pharynx, source, formant, and pressure characteristics in operatic and musical theatre singing. J Voice. 1993;7:301–310.
7
8. Hein M. Die Gesangtechnik des Beltings – eine Studie u€ber Atemdruck, Lungenvolumen und Atembewegungen. Hamburg, Germany: Dissertation, Hamburg University; 2010. 9. Sundberg J, Thalen M, Popeil L. Substyles of belting: phonatory and resonatory characteristics. J Voice. 2012;26:44–50. 10. Evans M, Howard DM. Larynx CQ in female belt and opera qualities. Voice. 1993;2:7–14. 11. Barlow C, Lovetri J. Closed quotient and spectral measures of female adolescent singers in different singing styles. J Voice. 2010;24:314–318. 12. Lebowitz A, Baken RJ. Correlates of the belt voice: a broader examination. J Voice. 2011;25:159–165. 13. Kochis-Jennings KA, Finnegan EM, Hoffman HT, Jaiswal S. Laryngeal muscle activity and vocal fold adduction during chest, chestmix, headmix, and head registers in females. J Voice. 2012;26:182–193. 14. Bourne T, Garnier M. Physiological and acoustic characteristics of the female Music Theater voice. J Acoust Soc Am. 2012;131:1586–1594. 15. Bestebreurtje ME, Schutte HK. Resonance strategies for the belting style: results of a single female subject study. J Voice. 2000;14: 194–204. 16. Bourne T, Kenny D. Perceptual Descriptions of Legit and Belt Voice Qualities in Musical Theater. Perth, Australia: Proceedings of the ANATS Conference; 2008. 17. Titze IR, Worley AS. Modeling source-filter interaction in belting and highpitched operatic male singing. J Acoust Soc Am. 2009;126:1530. 18. Sundberg J. Articulatory configuration and pitch in a classically trained soprano singer. J Voice. 2009;23:546–551. 19. Echternach M, Sundberg J, Arndt S, Markl M, Schumacher M, Richter B. Vocal tract in female registers—a dynamic real-time MRI study. J Voice. 2010;24:133–139. 20. Echternach M, Sundberg J, Arndt S, et al. Vocal tract and register changes analysed by real-time MRI in male professional singers—a pilot study. Logoped Phoniatr Vocol. 2008;33:67–73. 21. Echternach M, Sundberg J, Markl M, Richter B. Professional opera tenor’s vocal tract configurations in registers. Folia Phoniatr Logop. 2010;62: 278–287. 22. Echternach M, Traser L, Markl M, Richter B. Vocal tract configurations in male alto register functions. J Voice. 2011;25:670–677. 23. Echternach M, Markl M, Richter B. Vocal tract configurations in yodelling—prospective comparison of two Swiss yodeller and two nonyodeller subjects. Logoped Phoniatr Vocol. 2011;36:109–113. 24. Sundberg J. Articulatory interpretation of the ‘‘singing formant’’. J Acoust Soc Am. 1974;55:838–844. 25. Sundberg J. Formant technique in a professional female singer. Acustica. 1975;32:89–96. 26. Lovetri J, Lesh S, Woo P. Preliminary study on the ability of trained singers to control the intrinsic and extrinsic laryngeal musculature. J Voice. 1999; 13:219–226. 27. Sundberg J. The Science of the Singing Voice. Dekalb, IL: Northern Illinois University Press; 1987. 28. Traser L, Burdumy M, Richter B, Vicari M, Echternach M. The effect of supine and upright position on vocal tract configurations during singing—a comparative study in professional tenors. J Voice. 2013;27: 141–148.
