TMUS 8329: Forming Formants 1
Forming Formants: Models of Vocal Tract Configuration for Sung Primary Vowels
Angerhofer, Thomas Erik
University of Colorado Boulder
Author Note
TMUS 8329 Pedagogy Project for DMA in Voice Performance and Pedagogy TMUS 8329: Forming Formants 2
Abstract
The familiar adage from the old Italian school of singing, “si canta come si parla,” or “one sings as one speaks,” is as familiar today to teachers of singing as it was in centuries past. There is little argument, however, that the requirements of acoustic, classical singing differ from speech in at least three important ways: projection, intonation, and duration. In order to meet the musical and expressive demands of acoustic, classical vocal music, it would stand to reason that the manner in which a singer shapes the cavities and utilizes the articulators would differ from speech. Using the primary vowels /i e a o u/ as a point of reference, this document will begin by describing standard vocal tract configuration during speech. It will then take a chronological approach to detailing vocal tract shaping and use of articulators as they are outlined in both seminal and current vocal pedagogy texts. Finally, this document will attempt to categorize these findings to point out similarities and differences in thought.
Keywords: Vowels, Formants, Articulation, Resonator, Vocal Tract
TMUS 8329: Forming Formants 3
Forming Formants: Models of Vocal Tract Configuration for Sung Primary Vowels
There are many well-known adages in the lexicon of vocal pedagogy, but perhaps none so infamous as “si canta come si parla.” Some cling to the idea that one sings as one speaks as though it were an absolute truth. While it may be more commonly used today in the instruction of contemporary music styles, perhaps this is only due to the decades of criticism this expression has received among experts in classical singing. Richard Miller’s March 1987 article in the
NATS Journal drew attention to some of the issues related to this saying, an argument that continues to resonate with authors, as exemplified in the November 2018 Journal of Singing article, “Phonetic (L)imitation.”
I have personally been familiar with the axiom for over 30 years and have heard it used in varying contexts. Nevertheless, I have never thought of it as a rule or being particularly instructive, but more as an aphorism. To me, it is a general observation regarding delivery; because the art of singing harnesses both music and text for communication, it is important that sung sounds come across to the listener in a manner similar to speech. That is to say, there are some qualities of speech that should be maintained in singing, particularly the elements of flow and prosody, including intonation, tone, stress, and rhythm.
Authors critical of the saying and its application for classical singing have noted that singing and speech vary in at least three important ways: projection, intonation, and duration. In conversational speech, the proximity and size of the audience is minimal. Intensity levels, therefore, can remain relatively low while still being intelligible. In classical singing, the distance and size of the audience is far greater. In conversational speech, variation of pitch is rather limited – except in moments of heightened emotion. Intonation is typically limited to less than a perfect 4th up or down from the habitual or mean pitch, resulting in a functional pitch range of TMUS 8329: Forming Formants 4 less than an octave. Professional classical singers are required to have a functional pitch range of over two octaves in order to meet the demands of many compositions. With the exception of auctioneers and patter songs, duration varies greatly between speech and singing. It is frequently the case that the sung line is delivered at a slower tempo than in conversational speech, requiring greater specificity from the singer in defining vowel sounds and timing the change from sound to sound.
Given the different requirements of speech and singing, along with the goal of having them interpreted similarly by the listener, there would seem to be a logical extension: in order for them to sound similar, they must be done differently. Nevertheless, after over 30 years of vocal instruction, it wasn’t until quite recently that the concept of singing articulation as being disparate from speech was approached in any formal or systematic way in my training. Of course, there were casual mentions regarding the formation of different sounds, such as “the edges of the tongue should stay in contact with the top teeth” or “the tip of the tongue should remain touching the bottom teeth.” After realizing some benefits from training a method of articulation that is singing-specific, my curiosity was sparked as to how leading pedagogues have approached the subject of articulation.
Using the primary vowels /i e a o u/ as a point of reference, this document will begin by describing standard vocal tract configuration during speech. It will then take a chronological approach to detailing vocal tract shaping and use of articulators as they are outlined in both seminal and current vocal pedagogy texts. Finally, this document will attempt to categorize these findings to point out similarities and differences in thought.
TMUS 8329: Forming Formants 5
Articulatory Phonetic Model: Speech
Articulatory phonetics is the subfield of phonetics primarily concerned with how humans produce the sounds of speech, rather than the physical properties of spoken sounds, or acoustic phonetics. Singing voice science owes much of its foundational knowledge to these two disciplines; since the invention of the Bell Telephone in 1876 and the Edison Phonograph in
1877, linguists and acousticians have been using increasingly more sophisticated tools for recording, analyzing, and synthesizing the sounds of speech. Many of the acoustic and articulatory models for speech have been transferred and applied to singing, continuing to inform our current understanding.
Fig. 1: Places of Articulation
In figure 1, we see the places of articulation; these are the points in the vocal tract where maximum obstruction may occur due to the moveable structures. An obstruction is formed when two articulators move into proximity. The moving articulator is considered active, while the stationary one is considered passive. These places of articulation, where passive and active articulators meet, are labeled: 1. Exolabial, 2. Endo-labial, 3. Dental 4. Alveolar 5. Post-alveolar, TMUS 8329: Forming Formants 6
6. Pre-palatal, 7. Palatal, 8. Velar, 9. Uvular, 10. Pharyngeal, 11. Glottal, 12. Epiglottal, 13.
Radical, 14. Postero-dorsal, 15. Antero-dorsal, 16. Laminal, 17. Apical, 18. Sub-apical.
In the case of vowels, the structures that contribute the most are the cavities of the throat and mouth, or the laryngopharynx, oropharynx, and the oral cavity, the soft palate, the tongue, and the lips. The laryngo- and oro- pharynges, together with the oral cavity, are responsible for resonating the source harmonics that create the spectral tone colors that we identify as vowels.
The bulge in the tongue divides or couples the pharynges and the oral cavity, raising or lowering the resonance frequency of the two spaces or encouraging them to resonate more as a singular cavity. The velum, or soft palate, can raise or lower, contributing the first formant space and controlling nasality. The lips can potentially shorten or lengthen the vocal tract by retracting at the corners (spreading or smiling) or protruding forward (puckering), raising or lowering all or the resonating frequencies of the vocal tract.
Fig. 2: Describing Vowels
Spoken vowels are described by a combination of factors, with the position of the highest point in the tongue being the primary descriptor. If the bulge in the tongue is close to the roof of the mouth, the vowel is considered “closed.” If the tongue is seated lower in the mouth, with the bulge of the tongue less pronounced, the vowel is considered “open.” When the highest point of the tongue is nearer to the anterior portion of the mouth, it is considered a “front” vowel, but when the highest point is approaching the posterior portion of the mouth, it is considered a TMUS 8329: Forming Formants 7
“back” vowel. There is also the distinction of the “central” vowel, where the bulge in the tongue is more toward the middle of the mouth.
Fig. 3: Rounded vs. Unrounded
Another dimension used to describe spoken vowels is rounded vs. unrounded. This value refers to the movement or shaping of the lips during articulation. This may not be the most accurate descriptor; for the rounded vowels there are varying degrees of “pucker” or moving the corners of the mouth inward and/or slightly forward, but for the unrounded vowels there is sometimes an element of “spreading” or pulling the corners of the mouth back. Loose and moon are fine examples of the rounding, while please and me often go beyond neutral to have an element of spreading.
Chiba and Kajiyama’s research, detailed in The Vowel: Its Nature and Structure, is among the most comprehensive and influential works in the field of phonetics. Focusing on the primary vowels /i e a o ɯ/ (the unrounded /ɯ/ is more idiomatic to Japanese than the rounded /u/), their work is presented here as it is representative of the perspective of articulatory phonetics from the field of speech. In addition, it will provide a baseline for comparison with the varying positions on articulatory phonetics in singing. TMUS 8329: Forming Formants 8
/i/ Vowel Articulation
Fig. 4: /i/ vowel articulation
Articulation for the /i/ vowel is done with the bulge in the tongue toward the roof in the anterior portion of the mouth, or post-alveolar region, thus making it a “close unrounded front” vowel. Chiba and Kajiyama note that, “the front of the tongue is raised in such a manner that both sides of it touch the hard palate and a narrow channel of nearly uniform breadth is made in the middle of the tongue (p. 118).” The resulting division of the vocal tract is a larger space in the back (where F1 resonates) and a smaller space in the front (where F2 resonates). In speech, the lips can be left neutral; however, the corners of the mouth are often pulled back or “spread” for extra brightness. TMUS 8329: Forming Formants 9
/e/ Vowel Articulation
Fig. 5: /e/ vowel articulation
Articulation for the /e/ vowel is done with the bulge in the tongue slightly farther back in the mouth than the /i/ vowel, near the pre-palatal or palatal region. You will notice from figures 4 and 5 that in the /i/ vowel, the point where the tongue is closest to the roof of the mouth is at point 3, while /e/ is closest at point 4. Additionally, the /e/ vowel is twice the distance from the roof of the mouth at its closest point when compared to /i/, and the /e/ has slightly more space in the anterior portion of the mouth while the /i/ has a more generous space in the posterior portion of the mouth. The /e/ vowel is labeled a “close-mid unrounded front” vowel due to the increased distance from the roof of the mouth. The authors note that the vertical and horizontal openings of the mouth in the /e/ vowel are slightly greater than /i/; however, like the /i/ vowel, it is often the case that the corners of the mouth are pulled back or “spread.” TMUS 8329: Forming Formants 10
/a/ Vowel Articulation
Fig. 6: /a/ vowel articulation
The /a/ used by Chiba and Kajiyama is naturally found in Japanese language, and it is said to be akin to the “ah” sound in Italian. This “brighter” /a/ sounds like the “ah” in “mine,” which is colored by the /i/ that follows in the diphthong. In contrast, typical American English also uses the “darker” /ɑ/, as in the word “father.” Figure 3 shows that the difference between the two is minimal, with both vowels being “open / unrounded” and the place of articulation being in the posterior pharyngeal region for both vowel sounds. “In pronouncing /a/, the upper surface of the tongue is held almost flat, while the back of the tongue approaches the back wall of the pharynx and thus forms a constriction which separates the mouth cavity from a smaller pharynx cavity situated above the vocal chords (p. 123).” The measurements of the height and width of the mouth opening are greatest in this vowel, and like the /i/ and /e/, it is often “spread” in speech articulation. TMUS 8329: Forming Formants 11
/o/ Vowel Articulation
Fig. 7: /o/ vowel articulation
The /o/ vowel is described as a “close-mid rounded back” vowel; notice that the point of articulation is pharyngeal in nature, like the /a/ vowel. At the same time, the vertical mouth opening slightly less than in /a/. The authors note that the constriction is slightly higher and broader than in the /a/ vowel. They also state, “the mouth cavity assumes a shape somewhat like a Helmholtz resonator with a round enclosure and a small aperture (p. 126).” The vertical space inside the oral cavity in both the /a/ and /o/ is generally much greater than at any point in the /i/ and /e/ vowels.
/ɯ/ Vowel Articulation
It is important to note that Chiba and Kajiyama use the unrounded /ɯ/ rather than the rounded /u/ that is found in American English or Italian. As stated earlier, the unrounded /ɯ/ is TMUS 8329: Forming Formants 12 naturally found in the Japanese language and has slightly different properties than the /u/ vowel.
It is presented here for the sake of continuity and comparison.
Fig. 8: /ɯ/ vowel articulation
The authors assert, “the Japanese /ɯ/ is clearer; this is due to a difference in the manner of pronunciation – namely, in pronouncing u, the lips are rounded, and the constriction formed by the back of the tongue and soft palate divides the vocal cavity into two parts of nearly the same size… whereas in the case of (ɯ) the mouth cavity is a little smaller, with the jaws less widely opened and the lips held in their normal state (there occurs no lip-rounding) (pp. 126-
127).” Pictured below is a side by side comparison of the unrounded /ɯ/ and the rounded /u/. TMUS 8329: Forming Formants 13
Fig. 9: unrounded /ɯ/ vs rounded /u/
In addition to the lip rounding, and often protruding, in the formation of the /u/ vowel, the place of articulation is more uvular in /ɯ/, while the /u/ is more velar (positions 7-8 in /ɯ/ and 6-
7 in /u/). Further, the bulge in the tongue looks to be slightly more pronounced in /u/, narrowing the cavity even more than /ɯ/ and at a more definite point. Despite their minor differences, the overall vocal tract shape is quite similar. These articulatory models provide a point of reference for investigating the possible similarities and differences between vowel formations in speech and singing, as well as variations between models offered by notable pedagogues.
Articulatory Phonetic Models: Singing
Determining where, when, and with whom to begin this study is a challenge. Certainly, there have been many notable works on the subject of vocal pedagogy dating back hundreds of years. Many of these predate Chiba and Kajiyama’s 1958 publication, and many of them have useful and relevant information. For this project I have chosen to begin the review of articulatory models from singing after the date of their publication, when its implications had already made their way into the field of singing voice science. The contributions from the field of vocal pedagogy will be presented here in chronological order from date of publication. TMUS 8329: Forming Formants 14
Vennard, William. Singing: the Mechanism and the Technic. 1967
Vennard, along with a cadre of other teachers of singing like Ralph Appelman, heralded in a new wave of interest in singing voice science in the middle of the 20th century. His research in the areas of anatomy, physiology, and acoustics simultaneously substantiated and challenged many of the long-held beliefs within the classical singing community regarding the nature of the vocal instrument. He worked closely with Janwillem van den Berg, the Dutch speech scientist and medical physicist largely responsible for developing the modern model of vocal fold function known as the myoelastic-aerodynamic theory of voice production (van den Berg also designed the first implantable, adjustable rate pacemaker). During his tenure at the University of
Southern California, Vennard trained numerous successful singers; among the most notable is mezzo-soprano Marilyn Horne.
Vowel Articulations.
Fig. 10: Vennard’s Vowels
Vennard’s 1967 book only contains films of 3 of the primary vowels, as much of his work centers on the physical nature of sound, or acoustics. The point of articulation for /i/ in Vennard and van den Berg’s film is similar to Chiba and Kajiyama’s, alveolar or post-alveolar, with perhaps a slightly larger channel between the tongue and the roof of the mouth. Both the /u/ and TMUS 8329: Forming Formants 15
/a/ in figure 10 are different versions of the vowels from the ones analyzed in The Vowel…;
Vennard’s /u/ has a tongue that sits lower in the mouth with a point of articulation near Chiba’s reference point 9, which is more toward the posterior of the pharynx. In fact, Vennard’s /u/ tongue articulation more closely resembles Chiba’s /a/, which is pharyngeal, while Vennard’s /ɑ/ tongue articulation is more closely related to Chiba’s uvular /ɯ/ - albeit Chiba’s mouth / jaw opening is smaller, creating a narrower channel in the oral cavity. While we are unable to compare all of the primary vowels graphically, Vennard makes the following statements about /e/ and /o/; “we may say that the vowels Ay and Ee are formed with the tongue… With Ay, the tongue partition is farther back (p. 131)” and “if the singer rounds his lips over the vowel Ah, that is, changes nothing except to purse the lips a little, the pitch of the formant is lowered, and the vowel becomes Oh (p. 134).”
Moriarty, John. Diction. 1975
Born in 1930, John Moriarty has distinguished himself as a vocal coach and collaborative pianist, conductor, stage director, and artistic administrator with leading organizations throughout the United States. His interest in vocal music led to studies with Pierre Bernac, among others, later becoming the chair of the Opera Department at the New England
Conservatory of Music. While not a vocalist himself, he worked regularly with many of the finest established artists of the era, as well as playing a pivotal role in the development of generations of young artists. While it is not intended to be a scientific text, Diction does an excellent job of categorizing vowels by their most active articulators, as well as providing detailed instructions on their formation. TMUS 8329: Forming Formants 16
Fig. 17: lip and tongue vowels
Vowel Articulations.
Moriarty begins his section on vowel formations with a statement on vowel classification, immediately calling our attention to a potential challenge between the phonetics of speech and singing. “Most phoneticists classify vowels in three groups: frontal, central and back. These terms refer to the position of the tongue… The word ‘back’, however, is anathema to many singers and teachers of singing. For this and additional reasons, we find it more useful to classify the vowels in five groups as tongue vowels, lip vowels (according to which factor is most active in their formation), mixed vowels, neutral vowels and nasalized vowels” (p. 7). Since this study focuses on the primary vowels /i e a o u/, the material presented herein will pertain only to the tongue and lip vowels.
Tongue Vowels.
In describing the tongue hump, or lack thereof, Moriarty again departs from the standard classifications of “closed” and “open.” Instead he opts for “high” and “low,” respectively. Rather than including x-ray films or drawings, Moriarty instructs us to speak a few words and notice the difference in the tongue positions. The first of these are “me” and “met,” noting that /i/ has a higher tongue position than /ɛ/. He also adds, “it is possible to pronounce the vowel [i] without TMUS 8329: Forming Formants 17 smiling. Instructions here and below about smiling, puckering lips, etc., are given for the purpose of leading the singer to the easiest way of making accurate vowel sounds, all of which can be pronounced with a variety of facial expressions” (p. 9).
In order to find the /e/ vowel, Moriarty suggests sustaining the /i/ vowels and gradually lowering the tongue and the jaw until one reaches /ɛ/. Halfway between those vowel sounds, one will encounter /I/; half way between /i/ and /I/ (or ¼ way between /i/ and /ɛ/) is where one finds
/e/. “If you continue lowering the tongue and jaw, keeping a smile, you will find the French vowel [a]…” (p. 11)
Lip Vowels.
Again using familiar words as a guide, Moriarty instructs us to say the word boon “with well-rounded, projecting lips and a lightly dropped jaw” (p. 13) and contrast it with the word
Bought. With the lips transforming from circular to vertical oval as the jaw drops, one can find many vowels in between. Moriarty considers /U/ to be the halfway point between /u/ and /ɔ/, with the lips still puckered like /u/, albeit more vertically spaced. /o/, then, is found halfway between /u/ and /U/. Moriarty considers the more fronted /a/ to be characteristic of French and part of the spectrum of tongue vowels, while the more back /ɑ/ “is the long, relaxed vowel heard in Italian in words like sala, lana, pane, etc. It is the most relaxed vowel: the jaw is dropped, the lips are totally relaxed, and the tongue lies flat in the mouth with its tip resting against the back of the lower teeth” (p. 47).
Coffin, Berton. Coffin’s Overtones of Bel Canto & Sounds of Singing. 1980 & 1987
As Professor of Voice and Chair of the Voice Division at the University of Colorado at
Boulder, where he served for nearly 31 years, Berton Coffin’s contributions to the field of voice go beyond singing voice science and pedagogy. His Phonetic Readings of Songs and Arias, TMUS 8329: Forming Formants 18
Word-by-Word Translations of Songs and Arias, and Singer’s Repertoire series have found their way into the reference section of music libraries everywhere. His work with Dr. Pierre Delattre, phonetician and Professor of French, let to many valuable contributions in the field of acoustic phonetics. Coffin dedicates Appendix K of his Sounds of Singing to an article Delattre published in 1968.
Vowel Articulations.
Fig. 11: Coffin-Delattre TMUS 8329: Forming Formants 19
Delattre’s article, written while he was still teaching at the University of California Santa
Barbara, begins with a section from Molière’s Le Bourgeois Gentilhomme:
“PROFESSOR OF PHILOSOPHY: . . . There are five vowels or voices: A, E, I, O, U. MONSIEUR JOURDAIN: I understand all that. P. OF PHIL.: The vowel A is sounded by opening the mouth very wide, -- A. M. J.: A, A. Yes. P. OF PHIL.: The vowel E is sounded by bringing the lower jaw to the upper jaw, -- A, E. M. J.: A, E; A, E. Bless me! How fine that is! P. OF PHIL.: The vowel I is formed by bringing the jaws still closer together, and stretching the corners of the mouth toward the ears, -- A, E, I. M. J.: A, E, I, I, I. That’s true. Hurrah for science! P. OF PHIL.: The vowel O is sounded by opening the jaws and drawing in the lips at the two corners, -- O. M. J.: O, O. Nothing could be more true. A, E, I, O, I, O. It is admirable! I, O; I, O. P. OF PHIL.: The mouth must be opened exactly like a round O. M. J.: O, O, O. You are right. O, -- ah! what a fine thing it is to know something! P. OF PHIL.: The vowel U is sounded by bringing the teeth together without entirely joining them, and protruding the lips outwardly, while bringing them narrowly together without actual contact: O, U. M. J.: O, U, U; the truest thing that ever was, -- U. P. OF PHIL.: Both your lips should be stretched out as if you were making a grimace; so that if you should ever want to make a face at any one and ridicule him you have only to say “U”. M. J.: U, U. True enough. Ah! why didn’t I learn that in my youth?”
While not exactly scientific in nature, this passage does bring to light some lay observations about vowel articulations. First, the /a/ is thought of as having a large mouth opening, with /e/ and /i/ becoming progressively smaller in opening; notice the instruction for
“spreading” the /i/ vowel. The /o/ is described to have a similar jaw height as /a/, only with a narrower width, and the /u/ is narrower and smaller still, with protruded lips. The scrutiny of the professor is limited to what can be seen by the naked eye, specifically the external movements and positions of the face.
In comparing Delattre’s models with Chiba and Kajiyama’s, vowels /i/, /e/ and /a/ are somewhat similar. The difference between Chiba and Kajiyama’s /a/ and /o/ is only slight, while
Delattre’s tongue shape begins to transform in the /o/ vowel as it moves toward the next position TMUS 8329: Forming Formants 20 of /u/. While both Delattre’s and Chiba’s models come from speech, perhaps it is the difference in the Japanese /ɯ/ that lends to a different movement pattern as the tongue moves from /a/, through /o/, on the way to /ɯ/.
Fig. 12: famous Italians
These two films, found in Coffin’s Overtones of Bel Canto, are borrowed from Russell’s
Speech and Voice (1931). The upper film shows the /i/, /ɑ/ and /u/ vowels of Italian baritone
Pasquale Amato (1878-1942), whose career began in the early 1900’s singing in leading Italian opera houses under the baton of Toscanini. The lower film depicts the same vowels as sung by the tenor Enrico Caruso (1873-1921), whose 260 commercial recordings led to international pop stardom. Both of these career singers received their training at the Conservatory of San Pietro a
Majella in Naples under the same teacher, Beniamino Carelli.
At first glance, the height of the tongue hump is noticeable in each of the sung vowels.
While /ɑ/ has the greatest degree of “openness”, it is difficult to ascertain where the place of articulation is exactly – back, toward the posterior portion of the pharynx, or upward, toward the TMUS 8329: Forming Formants 21 velar or uvular regions. From an acoustic standpoint, this would change the dimensions of the F1 and F2 resonating cavities, raising or lowering their resonance frequencies. Both Amato and
Caruso have a more released jaw in each of their vowels when compared to the models of Chiba and Kajiyama; the /u/ vowel is likely to retain its intelligibility due to the use of the lips when compared with /ɯ/. Nevertheless, it is the height of the back of the tongue, which Coffin notices on page 183, that remains the defining characteristic of these two singers. Despite their differing fächer, one must wonder if their common Italian heritage, and therefore language, their comparable instruction, or both, might be contributing factors to this phenomenon.
Miller, Richard. The Structure of Singing… 1986
For those whose studies of vocal pedagogy predate Y2K, Richard Miller (1926-2009) is among the most notable and prolific figures. His training as a singer took place in the United
States, France, and Italy, with his major professional appointment in Switzerland; it is no wonder that he took a keen interest in comparing and contrasting the international schools of singing.
During his over 40 years of teaching at Oberlin Conservatory, his interest in physiology and acoustics led to the establishment of one of the first singing voice laboratories on a college campus. Through NATS, Miller created a vocal pedagogy forum that served as a resource for teachers throughout North America and beyond. TMUS 8329: Forming Formants 22
Vowel Articulations.
Fig. 18: Miller’s vowels
The drawing and films Miller included for his vowel articulations are both taken from speech models, published between 1962 and 1974. For this study, we are principally interested in lines 1, 2, 5, 7, and 8, to the left, as well as the films for heed, hod, and who’d, to the right (there is no representation for either the /e/ or the /o/ vowels). While the overall shapes of the vowels are similar to Chiba and Kajiyama’s, it appears that the front-back dimension of the channel, especially in heed or the /i/ vowel, presented by Miller is longer. More specifically, the place of articulation along the roof of the mouth is over a greater distance in Miller, stretching from the post-palatal to near the velar region. Miller’s /ɑ/ is nearly identical to Chiba’s /a/, with the articulation happening toward the posterior pharyngeal wall. The /u/ in figure 18 has a noticeably higher, upward articulation near the velum when compared to the /ɯ/ in figure 8, but it does seem to be congruent with the rounded /u/ in figure 9. TMUS 8329: Forming Formants 23
Fig. 19: Vowel Modification
Miller includes more specific verbal instruction in his chapter on the well balanced vowel, which begins with a description of “the acoustic at-rest posture”, which he thinks of as a
“home base for the speech mechanism” (p. 69). Miller makes a number of points throughout the book regarding vowel modification and/or adjustment of the shape of the vocal tract for singing in different parts of the range. On the left of figure 19 is a drawing indicating the “need” for the soprano to have a wider (or taller) mouth opening when singing in the upper range, changing the quality of the sung vowel (the image is borrowed from Sundberg’s 1977 article in Scientific
American). On the right is Miller’s chart for moving toward a more open vowel in the upper part of the range, or for moving toward a more closed vowel in the zona di passaggio. His detailed instructions for each of the primary vowels follow.
/i/.
Miller refers to /i/ as “the most frontal of all the vowels (the closest, regarding forward tongue posture)… the buccal orifice is narrowed… and is horizontally elongated as in a pleasant facial expression. The mandible is in a posture that shows limited space between the two rows of teeth… With the exception of the apex of the tongue, which contacts the lower teeth, the front of the tongue is elevated so that the tongue arches almost to the roof of the mouth. The highest degree of tongue elevation occurs in front of the middle point of the hard palate. The tongue contacts portions of the hard palate on either side” (p. 71). TMUS 8329: Forming Formants 24
/e/.
“The vowels [e] is more closely related to [i] than any other sound… most singers are aware of the increased lateral contact of the tongue with the teeth on [e]. A slightly lower jaw posture than in [i], and the wider shape to the oral cavity, result in more frontal spaciousness in
[e] than in [i]” (p. 72).
/ɑ/.
“When singing /ɑ/, the lips part, the mandible lowers, and the tongue lies flat on the floor of the mouth cavity… the attraction of the vowel /ɑ/ for many singers lies in its avoidance of tongue constriction of the vocal tract” (p. 71)
/o/.
“The lips are separated with [o], but they are more rounded than with [ɔ], and they protrude somewhat more… it is easy to exaggerate the physical distance from [ɔ] to [o] in singing… the tongue is depressed in its anterior portion; elevation of the posterior part of the tongue… is also present” (p. 73).
/u/.
“… the back portion of the tongue… elevates, leaving little space between the tongue and the soft palate… there is little space in the posterior buccopharyngeal area because more room exists in the forward part of the mouth” (p. 73).
Sundberg, Johan. The Science of the Singing Voice. 1987
Sundberg begins the preface to his book by stating, “this is an attempt to present and discuss research on the human voice in singing as compared with speech. A book with this aim seemed needed: research reports in this area are scattered in various journals ranging in subject from astronautics to musicology…” He continues, “This book does not treat the singing voice in TMUS 8329: Forming Formants 25 the usual way. Actually, most of the several terms normally used for singing are not at all mentioned in this text, the reason being that different persons tend to mean different things with these terms” (ix). Originally trained as a musicologist, Sundberg’s interest in acoustics began with a dissertation on organ pipes, which then led to his analysis of the singing voice as a musical instrument.
Vowel Articulations.
Lindblom and Sundberg began by examining long vowels, as opposed to glides, and the use of five articulators: the jaw, the tongue body, the tongue tip, the lips, and the larynx. He writes, “different vowels are often associated with different larynx positions. The main rule seems to be that the larynx is raised in vowels pronounced with spread lips, such as /i/, and lowered for vowels pronounced with rounded lips, such as /u:/” (97). Sundberg also notices a correlation between fundamental frequency and the position of the larynx in speech; the higher the spoken pitch, the higher the larynx. Further, he states that the down and back movement of the jaw has a narrowing effect on the pharyngeal cavity. The pictures below are limited to the tongue body and its relationship to the jaw. TMUS 8329: Forming Formants 26
Fig. 13: Lindblom-Sundberg model
One of the fist observations Sundberg made is that many of the vowels can be made with only a slight movement of the tongue. An argument can be made that for the vowels /i/, /e/, and
/u/, modeled in the top two drawings, there is some element of “closeness” as the bulge in the tongue articulates upward – likely approaching the roof of the mouth. The vowels /ɑ/ and /o/ also seem to indicate an articulation of the tongue toward the posterior of the pharynx, with place of articulation for /ɑ/ being slightly lower in the vocal tract, again showing the possibility for a degree of “closeness.” The tongue tip, according to Sundberg, was mostly passive. In the /u/ vowel it slides back a bit along the floor of the mouth, but it mostly seems to be in a state of readiness for the articulation of consonants. Movements upward, downward, anteriorly, and posteriorly seem to have little effect on the vowel. In any case, Lindblom and Sundberg believe that the “articulators move according to certain patterns” (96). TMUS 8329: Forming Formants 27
Fig. 14: speech vs. singing
Sundberg included the above illustration to show how the sung pitch can influence articulation. In the speech level model, the bulge in the tongue is more forward for /i/, but the bulge in the tongue is nearly identical for /ɑ/ and /u/. The frequencies used in the other models seem to indicate a reference pitch of A, with the upper right being A3, the bottom left A4, and the bottom right A5; in actuality, these frequencies lay between A and A#, so the subjects must be using a higher tuning system. The tongue movement is less pronounced with each successive octave. Generally speaking, these articulations look remarkably close the films of Amato and
Caruso; the tongue generally articulates upward at some point, with the exception of /ɑ/ and /u/ articulating toward the posterior of the pharynx with the sung pitch of A3. Sundberg also makes mention of a “pitch-dependent jaw opening” (p. 129), calling for more formal measurements to be made to show a correlation between higher sung pitches and an increased jaw opening.
Doscher, Barbara M. The Functional Unity of the Singing Voice, 2nd Ed. 1994
Barbara Doscher earned all of her degrees in voice at the University of Colorado at
Boulder with Berton Coffin as her major professor. She went on to teach at the same university alongside her mentor following the completion of her doctorate and went on to train many successful professional singers. Doscher was considered an elite teacher, and was featured in the TMUS 8329: Forming Formants 28
2017 book, A Spectrum of Voices: Prominent American Voice Teachers Discuss the Teaching of
Singing.
Fig. 15: velo-pharyngeal axis
Doscher’s book contains information from Chiba and Kajiyama, Berton Coffin, Delattre, and Richard Miller, regarding articulatory phonetics. Figure 15 makes a point that the tongue cannot be in two places at once; the place of articulation cannot be both “front” and “back.” She makes several useful points. One, “if there is any single concept on which most singing teachers appear to agree, it is the idea of an ‘open throat’ (not that they agree about how to get it or what
‘open’ means)” (p. 111). Two, “freedom and mobility” of the tongue are directly related to having those two qualities in the sung tone. Three, “few vowels are pronounced with a flat tongue… it is self-evident that very little adjustment for vowel differentiation is possible” (p.
115). TMUS 8329: Forming Formants 29
Vowel Articulations.
Fig. 16: Doscher’s vowels
In addition to the movement of the tongue, Doscher also indicates movement of the jaw for vowel articulation. Both /i/ and /u/ maintain a “closeness” with the roof of the mouth, with the /i/ being more alveolar the front and the /u/ being more velar; while the /ɑ/ shows the tongue laying lower in the mouth, the jaw is also released down and back. The place of articulation for /
ɑ/ is toward the posterior of the pharynx; however, the back of the tongue is shown to be slightly higher in the mouth than the tongue tip. This could be thought of as resulting from the jaw position, but when shown in contrast to the /æ/ vowel with the same jaw opening, it is noticeably higher.
(Bunch) Dayme, Meribeth. Dynamics of the Singing Voice, 4th Ed. 1997
Meribeth (Bunch) Dayme (1938-2019) contributed in many ways to the teaching of singing, from her collaboration on The Singing Book, a favorite for class voice instruction, to her establishment of a holistic approach she called CoreSinging®. While working on her PhD. at the
University of Southern California with William Vennard, she instructed singing in the School of
Dentistry. This might account for her precise and detailed, medical approach to anatomy and TMUS 8329: Forming Formants 30 physiology in this book, which utilizes cadaver photos to illustrate many of the physical structures of the voice.
Vowel Articulations.
Fig. 20: Bunch’s vowels
In the left half of figure 20, we see an example of this book’s comprehensive approach to anatomy. The X-ray sketches to the right are the only ones dedicated to articulatory phonetics, showing the tongue, jaw, and lip positions for the vowels /i/, /a/ and /u/. The sung pitch for each of these examples is A4, and each is performed by the same singer. While the /i/ looks to be complimentary to both the speech and singing models already presented, the slightly higher back of the tongue in the /a/ is a departure from Chiba and Kajiyama. The other noticeable feature is the movement of the jaw – closest for the /i/, most open for the /a/, and slightly less for the /u/.
Bunch later states, “formation of proper vowel sounds requires little or no movement of the lower jaw and lips (a notable exception is the /u/ and /y/ in German and French), however beginners tend to overlook this and include much too much jaw movement. Ideally, the singer will allow the muscles which elevate the mandible, such as the temporalis, to maintain its TMUS 8329: Forming Formants 31 balance which will in turn free the tongue to move as needed and the soft palate to remain“ (p.
101). Throughout chapters 5 and 6, Bunch reiterates that there is a wealth of language specific information regarding articulatory phonetics and urges singers to undertake ongoing study of the foreign languages in which they sing. While she presents little else dealing with explicit vowel formations, she leaves many relevant gems which I will include below.
• “The basic Italian vowels used in singing are /a/, /e/, /i/, /o/, and /u/… All additional
vowel sounds are modifications of the above” (p. 101).
• “The lips are able to articulate all vowels efficiently when they are free and relaxed. Lips
drawn tightly over the teeth will absorb vibrations and dampen the tone. Lips held in the
“smiling” position tend to become tense and produce too bright a sound. Because singing
will magnify errors made in speech, teachers will find it necessary to spend a
considerable amount of time correcting student’s vowels” (p. 102).
• “Wrinkled brows, looks of worry or a fixed stare are certainly indicative of a tone that
lacks freedom… The singer’s face need never be contorted while singing; rather
approach a natural look with the lips relaxed ready to help in articulation” (pp. 103-104).
• “Singers… may have the rather common speech habit of protruding the lower jaw rather
than lowering it… In singing, this forward displacement, and any other deviation, alters
the shape of the resonator and position of the larynx and creates problems of articulation
and resonance” (p. 104).
• “The most commonly observed fault is a tongue pulled away from the teeth and heaped
up at the back; and such a lump… fills a large volume of the oropharyngeal isthmus…
thus absorbing vibrations and reducing the vocal quality… Glossal tension generally
works against elevation of the palate, however, some singers will attempt to get the TMUS 8329: Forming Formants 32
needed space by consciously depressing the tongue, a habit that is detrimental to good
sound. This habit results in the resonating chamber being shortened and narrowed and the
sound becoming harsh, strident and forced” (p. 104).
Miller, Donald Gray. Registers in Singing. 2000
Donald Miller is best known among voice pedagogues as the developer of VoceVista, a software program for analyzing the physical properties of sung sounds as well as the contact quotient of the glottis. This software has had a tremendous impact in the singing voice science community, providing an accessible tool for interested researchers with a growing set of applications and data for evidence-based practices. Dr. Miller taught at Syracuse University for over twenty years when he moved to Groningen, the Netherlands, in 1987 to devote himself to research on the acoustics and physiology of the singing voice as an associate of the Groningen
Voice Research Lab.
Vowel Articulations.
Most of Miller’s work focuses on singing voice acoustics; among his discoveries are the differing resonance strategies successfully employed by singers in the same general fach.
Pavarotti and Carreras, for example, tend to tune their second formant to the third harmonic in their upper range to achieve brilliance. Domingo and Kraus, on the other hand, rely less on this tuning and more on boosting the singer’s formant cluster, a group of resonances in the upper octave of the piano, for their characteristic ring. In Chapter 10 of Registers in Singing, Miller presents a Comparison of Vocal Tract Formants in Singing and Nonperiodic Phonation.
Comparing sung phonation with both vocal fry and ingressive phonation, both of which are non- periodic, Miller suggests, “the close similarity in acoustic data in combination with the relative dissimilarity in spatial data indicates that the accurate imitations are not primarily the result of TMUS 8329: Forming Formants 33 imitating the singing postures, but have instead an aural basis” (p. 188). For each of the sung phonations, the nominal pitch used is D3 (most likely sung by a low male voice) to ensure a high density of harmonics and a low likelihood or tendency toward formant tuning as a registration strategy.
/i/ Vowel Articulation.
Fig. 21: Miller’s /i/
The sung /i/ vowel in figure 21 most closely resembles Richard Miller’s model. Rather than having a more specific place of articulation toward the alveolar ridge, both Millers present an /i/ with a longer place of articulation, extending further along the roof of the mouth from the alveolar to velar regions with a long, narrow channel extending from front to back. TMUS 8329: Forming Formants 34
/ɑ/ Vowel Articulation.
Fig. 22: Miller’s /ɑ/
Unlike some of the sung models we’ve seen, D. Miller’s /ɑ/ is more like the speech model of Chiba and Kajiyama. Films and sketches of Italianate and Sweedish singers, as well as a few non-European models, show a slight downward slope from the back of the tongue to the tongue tip in /ɑ/. Those models derived from speech tend to show an opposite tongue slope; usually the back of the tongue is lower than the front, with the place of articulation toward the posterior pharyngeal wall. TMUS 8329: Forming Formants 35
/u/ Vowel Articulation.
Fig. 23: Miller’s /u/
Figure 23 shows the greatest deviation from the norm; only Vennard’s /u/ comes close to
D. Miller’s tongue position. Every other model, spoken or sung, shows a place of articulation close to the velum toward the posterior portion of the roof of the mouth. While Vennard’s model is somewhat flat, D. Miller’s /u/ has a clear downward slope at the back of the tongue. Because we cannot hear the sample from which this MRI was taken, it gives cause to wonder about the possible variables, such as skill of the singer, voice type, and sung pitch, and how these might have consequences for the data.
Smith, W. Stephen. The Naked Voice. 2007
Realizing early in his career that he was drawn to teaching, Steve Smith (b. 1950) credits many of his realizations about the teaching of singing to his time studying with Inez Lunsford
Silberg at Oklahoma City University. From her he claims to have learned the qualities of a good sound, but his technical approach came from interpreting and organizing her imagery-based instruction into a more concrete, action-oriented protocol. His success working with developing TMUS 8329: Forming Formants 36 artists at the Houston Grand Opera Studio and the Aspen Music Festival and School led to an appointment at the Julliard School in 1998. Since 2011, he has been on the faculty at
Northwestern University; his students include Christine Brewer, Joyce DiDonato, Rod Gilfry,
Brian Mulligan, and Eric Owens.
Vowel Articulations.
In the spirit of full disclosure, my personal interest in the subject of articulatory phonetics in singing came from my exposure to Steve Smith. I had a teaching colleague who studied with
Smith dating back to his time at Oklahoma Christian College. Not only did I have a fondness for this teacher’s kind manner and generous spirit, but he often expressed to me that he was not blessed with a wonderful instrument – in order to make good sounds, he had to really learn how to use it well. When the opportunity arose to work with Smith at the 2017 NATS Intern program,
I was exposed to a highly organized approach to singing that was unlike anything I had been shown in my previous 30 years of vocal instruction; among the foundations of this approach are aurally motivated, “refined” vowel articulations.
Smith believes that purifying the vowels helps to coordinate the articulatory muscles in a way that prevents them from having an adverse effect on the musculature of the larynx during phonation. “To form pure, disentangled vowels, we must move the articulators (tongue, jaw, lips, and soft palate) in ways that are quite different from normal speech. For clearly defined vowels, the articulators must become more disentangled from the voice (muscles involved with phonation). To accomplish this, it is necessary to coax the articulators to function independently of each other, which is not typical of normal speech” (p. 57). Smith uses a number of “tools” to persuade the articulators, the tongue in particular, to reconfigure themselves for these “refined” vowels as they contribute to efficient singing. TMUS 8329: Forming Formants 37
/i/ Vowel Articulation.
Fig. 24: Smith’s /i/
There are two elements that one will notice in each of Smith’s vowel articulation models.
First, the jaw is always released. That means the jaw is opened down and back, and it plays no role in the articulation of the vowels. Second, the place of articulation remains relatively constant near the velar region of the roof of the mouth. This is the same spot where many teachers of singing advocate articulating the /ŋ/, as in “hungry”; the articulation is upward toward the roof of the mouth rather than backward toward the posterior pharyngeal wall. This is a huge departure from most spoken and sung models of /i/; according to Smith, in the fronted /i/ “the whole tongue is thrust forward, which causes tension in the base of the tongue and pulls the larynx up with it. We cannot get free phonation with the normal, forward-thrusted [i]” (p. 57).
He expounds, “the key to speaking a free [i] is to let the jaw drop as if saying an [ɑ].
While saying [i] with the jaw in this position, the tip of the tongue may pull back from the back of the lower front teeth, the arch of the tongue moves farther back, and it touches the very back TMUS 8329: Forming Formants 38 of the molars. It is common when someone drops the jaw to the [ɑ] position and then tries to say
[i], that the resulting sound is [I] as in ‘bit.’ The singer must insist on the sound of [i]…”(pp. 58-
59). Like D. Miller, Smith indicates that vocal tract formation is motivated aurally. In addition to the place of articulation toward the velar region of the oral cavity, the channel or length of articulation from front to back is longest with /i/ - extending toward the post-alveolar region.
This results in a longer pharyngeal cavity and a smaller oral cavity. The lips remain neutral.
/e/ Vowel Articulation.
Fig. 25: Smith’s /e/
Smith’s /e/ is very similar to his /i/; the jaw is released, the lips are neutral, the tip of the tongue may also lose contact with the bottom teeth, and the place of articulation is identical. The main difference is the length of the channel or articulation of the tongue. The channel for /e/ is slightly shorter than /i/, extending only to the palatal region, resulting in a slightly larger oral cavity when compared to /i/. Like all vowels, it is motivated aurally; maintaining a stable larynx, TMUS 8329: Forming Formants 39 a released jaw, and neutral lips, while audiating the pitch and the vowel will provoke the tongue to find its “refined” position.
/ɑ/ Vowel Articulation.
Fig. 26: Smith’s /ɑ/
The /ɑ/ vowel is intentionally used in Smith’s model as it is most natural to speakers of
American English. The /ɑ/ is, in Smith’s model, the neutral or default position upon which all other vowels are built. The place of articulation is the same in all vowels, upward near the velum, and in the /ɑ/ there is little to no channel, making the oral cavity slightly larger than /e/, while the pharyngeal cavity remains stable or constant. This position is somewhat similar to what the films of the Italianate singers demonstrated, which Coffin examined in figure 12, although the back of the tongue is noticeably higher in Smith’s model. TMUS 8329: Forming Formants 40
/o/ and /u/ Vowel Articulation.
Fig. 27: Smith’s /o/ and /u/
The /o/ and /u/ are considered “lip” vowels; they share the same tongue articulation as the
/ɑ/ but have an additional rounding of the lips – without engaging the jaw. Instead of puckering or protruding the lips, the lips simply come together. This keeps the length of the vocal tract stable throughout all the vowels. The manner in which they move might necessarily be different than how they are accustomed to moving in speech, as the jaw remains released in singing. There is noticeably more movement in the upper lip than in the lower lip to accomplish the successively smaller openings required for /o/ and /u/.
In many ways, the “refined” /ɑ/ acts as a home base for all of the vowels; the tongue vowels, /e/ and /i/, start from /ɑ/ and have a progressively longer channel. The lip vowels, /o/ and
/u/, start from /ɑ/ with the lips creating an increasingly smaller opening – the inner dimensions of the space remain the same. This manner of articulation allows for the first formant space and overall tube length to remain stable across all vowels, while manipulating the second formant TMUS 8329: Forming Formants 41 space as needed. All other sung vowels are found in the spaces between the “refined” primary vowels or by combining elements of the lip and tongue vowels.
McCoy, Scott. Your Voice: An Inside View. 2nd Edition. 2012
As Professor of Voice and Pedagogy at the Ohio State University, Scott McCoy is the director of the Swank Voice Laboratory. He has served as president of NATS and has published regularly in the Journal of Singing. His commitment to teacher education and contributions to singing voice science has influenced a generation of evidence-based pedagogues.
Vowel Articulations.
Much like Donald Miller’s work, McCoy’s text devotes a greater percentage of its content to acoustics. As the formation of the resonator, the position and movement of the articulators, and how they divide the vocal tract determine the potential resonance frequencies of the tract and its cavities, the objective of an evidence-based approach would be to utilize the data and findings to support best practices in application. Interestingly, McCoy presents models from speech, rather than singing, taken from Peterson and Barney’s 1952 study. TMUS 8329: Forming Formants 42
/i/ Vowel Articulation.
Fig. 28: McCoy’s /i/
The shape of the tongue in McCoy’s model is similar to the one presented by Chiba and
Kajiyama. It also appears to have much in common with the drawings and films of Bunch,
Doscher, Sundberg, and Richard Miller. The place of articulation is toward the alveolar ridge.
/e/ Vowel Articulation.
Fig. 29: McCoy’s /e/ TMUS 8329: Forming Formants 43
McCoy’s /e/ is also similar to those presented by Chiba, Sundberg, and Coffin. The place of articulation for /e/ is slightly farther back than the /i/, in the palatal region of the mouth.
/a/ Vowel Articulation.
Fig. 30: McCoy’s /a/
The tongue shape in McCoy’s /a/ features a slightly higher back of the tongue than some of the other models presented. It is difficult to tell from both the drawing and McCoy’s description if the place of articulation is more upward, toward the velar region of the mouth, or backward toward the posterior pharyngeal wall.
/o/ Vowel Articulation.
Fig. 31: McCoy’s /o/ TMUS 8329: Forming Formants 44
McCoy explains that the /o/ is farther “back” than the /a/, with more “closeness” at the point of articulation. Indeed, the tongue-hump is extending upward toward the velar and uvular region in a way that creates an inclined plane with the tongue, from behind the bottom front teeth up toward the soft palate.
/u/ Vowel Articulation.
Fig. 32: McCoy’s /u/
While McCoy instructs that the lips be “rounded into a small, cylindrical opening” for the
/u/ vowel, he does not advocate for puckering or protruding them forward. The /u/ has the greatest degree of “closeness” toward the back of the mouth, with the place of articulation extending even further toward the velum and uvula than the /o/. Hence, the tongue forms a wedge in the /u/ vowel that is even more severe than the /o/.
Jones, David L. A Modern Guide to Old World Singing. 2017
David Jones admits in his biography to having suffered from numerous vocal issues early in his career. His studies with Alan Lindquest helped to correct many of the difficulties he had been experiencing, and inspired him to seek out the instruction and advice of teachers who were connected to the old Italian school of singing (Lindquest’s first two teachers were pupils of TMUS 8329: Forming Formants 45
Manuel Garcia). Over the past 30 years, he has been compiling information and exercises to educate a new generation of teachers on what he calls the “Swedish-Italian” method of singing.
Vowel Articulations.
Fig. 33: Jones’ vowels
Jones’ book is far from scientific; much of the terminology he uses and how he describes certain phenomena can be problematic. Figure 33 is the only diagram referencing vowel articulations in his book; that being said, his simplistic, principle centered approach is described at length. He specifies, “to inspire maximum resonance or ring in each vowel, the Italian and
Swedish-Italian Schools encouraged the use of the ng /ŋ/ as home position for the tongue. Note that this tongue posture is a flexible home base… For maximum results, the ng /ŋ/ tongue position should be encouraged directly after each consonant… the back of the pharynx must remain open. The tongue tip must be free to move to the different positions of the five Italian vowels… While use of the ng /ŋ/ tongue position is a practical way of bringing the tongue more forward and out of the pharynx, it must be carefully produced with the middle of the tongue, not the back of the tongue or tongue root” (p. 56).
The tongue tip positions in figure 33 are meant mostly for the middle range. “Toward the upper passaggio and into the high range, the tongue-tip position must be deep below the gum TMUS 8329: Forming Formants 46 line for all vowels” (p. 57). Jones also asserts, “when working on tongue-tip positions, the jaw should unhinge slightly down and back. If you stabilize the jaw slightly down and back and speak the five Italian vowels without closing the jaw, you will feel the proper position of the tongue-tip for each vowel sound naturally” (p. 57). Jones’ approach to vowel articulation seems to have the most in common with Smiths; the place of articulation is upward toward the velar region, similar to /ŋ/, with the /a/ vowel as a “home base.” All other vowels are formed by adjusting the position of the front of the tongue, the lips, or both, maintaining a stable pharyngeal space and vocal tract length through each of the vowels.
Observations and Recommendations
When I first embarked on this exploration of articulatory phonetics, there were a few things that I was expecting to find. First, as I had previously found in the field of acoustic phonetics, research from the field of linguistics has had a profound impact on singing. Because the two fields utilize the same organs for the ultimate purpose of communication, it would seem plausible to adopt studies and models from an already highly developed body of knowledge.
Second, based on my many years of private study and academic pursuits, I had a hunch that it would be unlikely to uncover many sources with a highly organized approach to acoustic phonetics for singing. As Vennard states, “the phoneticist concerns himself more with whether the vowels are recognizable or not, and less with their desirability for singing. We all know that there are half a dozen recognizable Ah’s, but the problem is to select and cultivate the one that is most pleasing to the ear” (p. 111).
Categories of Articulatory Phonetics
One of the goals of this project is to categorize the findings. In a previous project focusing on acoustic phonetics, I took a chronological approach to my investigation. In doing so, TMUS 8329: Forming Formants 47
I found that much of the research built and expanded upon previous discovery. Such was not the case with articulatory phonetics; some of the most recent authors included models from speech that were published half a decade earlier, while another recent author claimed that his singing- specific model had historical precedent. Nevertheless, I believe it is possible to place each of the authors’ models into one of the three following categories.
Speech-Based Models.
While most of the pedagogy texts were more or less related to speech-based models, like the reference model of Chiba and Kajiyama, none of them had enough similarities to be labeled
“speech-based.”
Hybrid Models.
Many of the authors identified articulations of the jaw, lips and tongue that loosely resemble speech-based models; however, they also indicate important differences in the general release of the jaw, increased buccal and pharyngeal space, and place or manner of articulation.
For these reasons, I would place the following authors in the “hybrid” category: Vennard,
Moriarty, Coffin, R. Miller, Doscher, Bunch, D. Miller, and McCoy.
Singing-Specific Models.
The authors included in this category propose a manner of articulation that differs dramatically from those spoken models of articulatory phonetics. Sundberg is perhaps the most subtle example of this; his observations are almost like those of an outside observer – collecting data from good singers on good singing without bias from personal practice. Smith and Jones, however, depart more radically from the norm with their advocacy for retaining “closeness” near the place of articulation for /ŋ/, or near the velum and top, back molars in articulating each of the vowels. TMUS 8329: Forming Formants 48
Contributions
While there was no noticeable, chronological progression in theory regarding articulatory phonetics in singing, each of the authors presented a wealth of information that might be helpful in determining best practices. Listed below are selected contributions from each of the authors.
• Chiba and Kajiyama provided a strong foundation for examining articulatory
phonetics, with precise mappings of the tongue, lips, jaw, and other articulatory
structures.
• Vennard identified “twang”, a characteristic of sound that is desirable in the
proper proportion, as being related to the position of the tongue (a higher back of
the tongue toward the velum and a lower tongue tip behind the bottom, front
teeth).
• Moriarty’s classification of the vowels, especially tongue, lip, and mixed, are
among the most practical and least confusing, as articulatory and acoustic
terminology often overlap and/or clash.
• Coffin’s models showed a movement toward maintaining “closeness” in many of
the vowels, further differentiating spoken and sung vowel articulations.
• Among other things, Richard Miller made mention of an “acoustic at rest posture”
from which the articulators can move with ease and efficiency.
• Sundberg noticed a potential relationship between vertical jaw opening and range
/ registration (acoustic and laryngeal).
• Doscher states that very little adjustment of the articulators is required in
differentiating vowel sounds, and that freedom and mobility of the tongue is
paramount to producing a good sound. TMUS 8329: Forming Formants 49
• Bunch asserts that effective articulation requires little to no movement of the jaw
and lips.
• Donald Miller sets forth the idea that vowel articulation might have an aural basis;
the shaping of the vocal tract is informed by pre-hearing or audiating the pitch and
vowel.
• Smith presents a highly organized approach to singing-specific articulation.
• McCoy’s evidence-based approach calls for new and ongoing research to support
best practices.
• Jones’ Swedish-Italian approach might possibly indicate a historical precedent for
singing-specific articulation.
Basis of Articulation / Articulatory Settings
Linguistic researchers have contributed greatly to our understanding of the vocal mechanism. In 1968, the distinguished linguist Noam Chomsky hypothesized an articulatory
“baseline” which he called the “basis of articulation;” a neutral position for the articulators and the configuration of the vocal tract, motivated by the mental impulse to speak, defaulted to in preparation for and in-between units of speech. Richard Miller eludes to this position of readiness in singing, motivated by what Smith calls “the urge for utterance,” which allows the articulators to position themselves for enunciation with minimal effort and movement.
Chomsky’s theory has been built upon in what Honikman calls “articulatory settings.”
Each language has characteristic phonemes, thus establishing patterns of movement for the articulators that are unique to that language. One of the challenges in enunciating the sounds of a foreign language is that our articulators are “geared” to move in patterns that pertain to the language with which we are most familiar. The standard settings and native patterns of our home TMUS 8329: Forming Formants 50 language, coupled with inadequate aural familiarity with the new language, lend to numerous physical and cognitive impediments to articulating efficiently and effectively in a foreign language. Enunciating with ease would require the development of both a physical and mental
“gear” for that language.
In classical singing, we not only have the task of articulating in less familiar languages, but we also have the language of music that places additional responsibilities on the vocal apparatus. These requirements include, but are not limited to projection, intonation, and duration.
Sustainable projection requires favorable source-filter interactions, or vocal tract shapes that can lessen the muscular effort while boosting the sounds being produced at the source. The extended range of singing may also require alteration of the vocal tract shape in a way that promotes intelligibility and efficient interactions. The length and timing of phonemes in vocal music may require clearer definition of the vowel sounds, as well as alterations to those sounds in order to maintain the communicative aspects of speech in singing. Music, therefore, becomes yet an additional language that effects the articulatory phonetics of singing.
Developing articulatory settings that are specific to singing would seem to be a logical extension of this theory. Just as the movement patterns from native language can inhibit smooth and effortless movement in a foreign language, so too can the habits of unprojected, minimally intoned, and undefined speech result in labored articulatory movements in singing. Many of the authors warn as to how undue tension in the articulators can negatively affect phonation and resonance. And, while this exploration is dedicated to articulatory phonetics, the resulting shapes of the vocal tract are largely responsible for the clarity of diction and efficiency of vocal production. TMUS 8329: Forming Formants 51
Recommendation
Bozeman declares that close timbre (F1=H2) and “whoop” timbre (F1=H1), both desirable sound qualities with efficient source-filter interactions found in classical singing, require a stable tube length. This means that the larynx should remain settled, the velum stretched, and the lips neither protruded nor retracted. And while the oral cavity promotes clarity and “twang” in the sound, it is the resonance of the pharyngeal space that most directly interacts with the vocal folds. The only models that truly promote stability in this space, behind the place of articulation of the tongue-hump, are those that are singing-specific. All other models show the tongue encroaching on the pharyngeal space in some of the vowel articulations.
Respecting evidence-based practice, more research on singing-specific models of articulation must be done in order to have appropriate supporting data. Furthermore, in order to ensure the integrity of the study, it is important to qualify the subjects. As in the films of Amato and Caruso, we can be sure that these subjects were “making good sounds.” However, Smith cautions that making good sounds does not necessarily correlate with good production. It is entirely possible that a subject, endowed with a fine instrument, can produce compelling, yet inefficient sounds. Hence, finding those subjects to study that first, make excellent sounds, second, do it very well, and third, know exactly what they are doing and how they are doing it might prove to be the most challenging aspects of any serious study of singing-specific articulatory phonetics. Nevertheless, finding skilled subjects from which to gather data on singing-specific articulation and the resulting acoustic properties would be an important step in verifying the efficacy of this approach to forming formants.
TMUS 8329: Forming Formants 52
Bibliography
Bozeman, Kenneth W. Practical Vocal Acoustics: Pedagogic Applications for Teachers and
Singers. Pendragon Press, 2013.
Bunch, Maribeth. Dynamics of the Singing Voice. Fourth Edition, Springer, 1997.
Chiba, Tsutomu, and Kajiyama, Masato. The Vowel: Its Nature and Structure. Phonetic Society
of Japan, 1958.
Chomsky, N. and Halle, M. The Sound Pattern of English. 1968.
Coffin, Berton. Coffin’s Overtones of Bel Canto. Scarecrow Press, 1980.
Coffin, Berton. Coffin’s Sounds of Singing. Second Edition, Scarecrow Press, 1987.
Doscher, Barbara M. The Functional Unity of the Singing Voice. Second Edition, Scarecrow
Press, 1994.
Honikman, B. “Articulatory Settings” In Honour of Daniel Jones. Longmans, Green, 1964.
Jones, David L. An Old World Guide to New World Singing. Self-Published, 2017.
McCoy, Scott. Your Voice: An inside View. Second Edition, Inside View Press, 2012.
Miller, Donald Gray. Registers in Singing: Empirical and Systematic Studies in the Theory of the
Singing Voice. Thesis University of Groningen, the Netherlands, 2000.
Miller, Richard. The Structure of Singing: System and Art in Vocal Technique. Schirmer, 1986.
Moriarty, John. Diction: Italian, Latin, French, German… the sounds and 81 exercises for
singing them. E. C. Schirmer Music Company, 1975.
Smith, W. Stephen. The Naked Voice: A Wholistic Approach to Singing. Oxford University Press,
2007. TMUS 8329: Forming Formants 53
Sundberg, Johan. The Science of the singing Voice. Northern Illinois University Press, 1987.
Vennard, William. Singing: the Mechanism and the Technic. Revised Edition, Greatly Enlarged,
Carl Fischer, Inc., 1967.