Voice Science and Early Experiments

In 1964, following four years of post-graduate study and minor professional gigs in Milano, Berlin, and Vienna, and with a growing family, I needed a “steady job.” It was my good fortune to land a position as voice instructor (later Professor of Voice) at Syracuse University (SU), where I happily remained for nearly 25 years. Among the advantages I found in upstate New York was proximity to Tri-Cities Opera, a respected company in nearby Binghamton, where I sang leading roles from the standard repertory. An even greater stroke of luck was that SU had started an initiative modeled on the recently created Voice Foundation, using up-to-date technology for research in laryngology, speech language therapy, and at least potentially, singing instruction. During my earlier years in Europe, I had followed my interest in languages and began to delve more deeply into phonetics (production of speech sounds), which then led to interest in acoustics and voice science. At the SU lab, we singing teachers were invited to observe our vocal folds during singing by videostroboscopy, an unusual opportunity in those days.

In connection with the Voice Foundation initiative at SU, singing teacher Jo Estill had arrived as a graduate student. It was her task to pursue the relationship between voice science and singing, an interest she and I shared. Jo and speech scientist Ray Colton offered a one-week summer workshop that offered an introduction to those interested in the topic, with participants largely from outside of the university. Jo and I had stimulating discussions, and I gradually found my way to the early Voice Foundation symposia in New York City.

In those same years of the late 1970s, I was engaged as the bass soloist for an oratorio performance at Calvin College in Grand Rapids and met pedagogue Richard Miller of Oberlin Conservatory, who was the tenor soloist. I invited Miller to give a workshop at SU and as we became friends, I learned about his extended work with Harm Schutte, an otolaryngologist and voice physiology researcher in Groningen, The Netherlands. Coincidentally, Groningen was also where esteemed voice pedagogue William Vennard had worked with voice scientist Janwillem van den Berg, author of the myoelastic-aerodynamic theory of voice production. It was from his time with van den Berg that Vennard found the basis for his highly regarded book Singing: The Mechanism and the Technic. Thus in 1982, I resolved to visit Prof. van den Berg, who then introduced me to Schutte. Both men then arranged for me to spend the spring semester of 1984 in Groningen. And that is how I came to spend my first sabbatical with Dr. Schutte, an otolaryngologist. I saw the sabbatical as a unique opportunity to learn as much as possible about voice science with the essential help of a laryngologist who was licensed to perform invasive procedures.

One such procedure that Harm wanted to try entailed passing a rather rigid catheter (originally intended for the urinary tract!) through the posterior commissure of the glottis, allowing direct measurement of the rapidly varying sub- and supra-glottic pressures. A few other Dutch colleagues had used that experimental protocol for speech research, but ours would be the first systematic application to the singing voice. (It also turned out to be the last to date, as far as I know.)

To be honest, I was not looking forward to swallowing the catheter and spent as much time as possible in avoidance, preferring to instead learn about other aspects of invasive procedures (i.e., video laryngoscopy) as well as about Schutte’s own dissertation work on subglottic pressures. Schutte had successfully measured average subglottic pressure with an esophageal balloon introduced nasally and swallowed into the esophagus, a relatively uncomplicated affair that sounds far worse than it is.  When we did finally get around to pressure measurements with the hard catheter through the posterior commissure of the glottis, the initial result was a lot of coughing by me, with only short stretches of sustained pitch. I feared that the experiment had failed, but then saw the ‘beautiful’ display of oscillating pressures above and below the glottis in the few sustained moments of successful phonation. When examined together with the electroglottographic signal (EGG), which indicates the instants of glottal closing and opening, the results were quite satisfying. To avoid coughing spasms in future experiments, we discovered that 15 minutes of rest after catheter insertion was a terrific help. I then used a sort of [v] sound before the vowel, which helped further to initiate a smooth vibration of the vocal folds that would then continue into the sustained vowel. The procedure never became easy for me, but we returned to it for four more extended sessions, and were even able to collect data from other available professional and amateur singers.

The results of these early experiments were quite revealing. The pressure data, together with the EGG data, showed that at glottal closing supraglottic pressure was minimal while subglottic pressure was maximal; thus, the basic acoustic impulse was a negative one, producing a wave of rarefaction, not compression. This result contradicted common scientific thought at the time. At that point in voice science history, the intuitive understanding of voice production assumed that subglottic pressure increased in the closed phase of the glottis causing the vocal folds to burst open (the open phase), resulting in the production of a train of pressure waves at glottal opening that generated the quasi-periodic fundamental frequency. It was assumed that the acoustic pressure wave began with a compression, or positive impulse that supplied the acoustic energy of voicing. Sadly, van den Berg was not supportive of our conclusions and died two years later without resolving the question.

But it turns out that we were not quite finished with the rigid laryngeal catheter. Schutte had been well acquainted with an excellent professional soprano who was willing to try our experiment, reasoning that if I could manage it, so could she. It seemed logical that glottal adduction would be inherently reduced in the ‘falsetto’ register typical of classical female singing compared to the operatic male voice in which the high range is produced in a type of modified ‘chest’ register. Because of this, the catheter in the posterior commissure should be less of an obstacle during phonation. This proved to be true and the catheter presented no great difficulty for her throughout the entire singing voice range. The particularly surprising finding from her data was this: the peak-to-peak variation measured in supraglottic pressure included a brief phase where it momentarily exceeded the subglottic pressure, suggesting that airflow through the glottis was momentarily reversed. This totally unexpected finding gave me an idea.

I had made my debut as a presenter a year earlier at the Voice Foundation Symposium in June of 1984, reporting on suggestive findings from our experiments with van den Berg. The following year, Gunnar Fant, a giant in the field of voice science research, announced a conference to be held on the island of Gotland in the Baltic Sea. Invited to participate were top international voice scientists and clinicians, including Ingo Titze and Johan Sundberg, whom I admired from the Voice Foundation Symposia as well as from their published research. Additionally, noted speech scientist Martin Rothenberg, whom I had come to know as a colleague at SU, and Harm Schutte were included. Those invited were expected to give presentations. I had the idea that our sub- and supra-glottic pressure measurements might be worthy of this group, and I proposed to write a paper that Harm would deliver. Though the organizers were somewhat reluctant to accept the proposal, the paper was well received, with special mention by Fant, and subsequently published. Already in my fifties, I had the feeling of having belatedly crashed a party of the “big boys.”

Harm Schutte and I continued productive investigations of the singing voice for the next 25 years, contributing annually at the Voice Foundation Symposia, as well as publishing a series of articles in refereed journals. In 1987 I resigned my professorship at SU and moved to Groningen to devote myself full-time to scientific research. In particular, I was deeply interested in investigating how modern technology might be employed to inform practical studio instruction of the singing voice.

The Origins of VoceVista

In the 1990s the computer had entered into the considerations of the Groningen Voice Research Lab, but I gave it little attention. In fact, born as I was in 1933, I was reasonably certain that I was beyond the age where I would ever get involved with computers. But thanks to Harm Schutte, who patiently guided me through the learning process, I soon changed my mind. Personal computers had evolved about then to the point where they could do spectrum analysis in ‘real time,’ giving immediate evidence of the effects of formant tuning on the relative strength of the harmonics of the voice source. (Formants are the variable dominant resonance frequencies of the vocal tract that produce the several vowel sounds.) The Groningen University Hospital provided Harm with the resources to have a computer program created — in MS-DOS — that would show and record the microphone signal, spectrum analysis, and electroglottograph (EGG) signal simultaneously.

Anxious to present this new technological tool to singing teachers, I was initially surprised that our presentations at the Voice Foundation symposia appeared to arouse little interest among the participants. For example, we identified “formant tuning” as a key practical measure by which skilled singers, through adjustments of the resonances of the vocal tract, successfully negotiated the challenges of “passaggio” in producing climactic operatic high notes. The basic problem comes from the fact that as the spacing between the harmonics rises with increasing fundamental frequency, the technology for automatically identifying the formant frequencies gradually fails. When we showed, however, that the formant frequencies can nonetheless be reliably, if not precisely, identified by the skillful use of vocal fry, that news was received with indifference. This response was perhaps to be expected, since the expertise of the singing teacher is manifest in hearing the desired result, rather than seeing a visual display of harmonics of the voice source. Nonetheless, I resolved to continue efforts in this line of research, which showed great promise in my own practical teaching. And though there was no enthusiastic stampede from a large group of singing teachers, there was enthusiastic, deep interest from a few.

Jim Doing, a young American tenor living in The Netherlands in 1996 to whom I gave occasional singing lessons, landed a position teaching singing at the University of Missouri. Jim was the first professional singer to make deliberate use of visual formant tuning feedback in our lessons together. When asked about his “research interests”, he mentioned our work on formant tuning and the university responded by offering to fund such research. Part of the funds were used to send me, along with some bulky apparatus, to work with Jim and his students for what turned out to be a successful few weeks. I recall that Jim had one tenor student in particular with a remarkable voice who beautifully demonstrated the key “passaggio” maneuver enabled by informed formant tuning.

It also happened that John Stuart, a tenor friend with a notable operatic career who had settled at Washington University in St. Louis, was on the planning committee for the upcoming biennial national conference of the national Association of Teachers (NATS). John was familiar with my work and he proposed to give it a place on the program. So, Jim Doing, his star student, and I drove from Kansas City, Missouri to St. Louis, all the while rehearsing our afternoon presentation in the car. The response was positive. Well-known baritone Ron Hedlund declared that he wanted VoceVista for the University of Illinois (UIUC), and Dan Ihasz, a young singer teaching at SUNY Fredonia, expressed his enthusiasm and began a years-long process of mastering its application to studio teaching and to his own singing technique. Dan continues to use VoceVista with his Fredonia voice and pedagogy students, some of whom are now building successful careers of their own.

There was also Garyth Nair, a choral conductor, singing teacher, and pedagogical researcher at Drew University in Madison, NJ, who was an early user of the internet and a promotor of new technology. I had met Garyth in 1997 at the Voice Foundation Symposium, where I presented the original VoceVista. After that original meeting, we decided to work together to make the software more user-friendly for singing teachers. A first step was to make it compatible with Microsoft Windows, which was then a new operating system. Garyth had earlier been in contact with the American engineer Richard Horne, who had put a free spectrum analysis program on the internet. A simple email inquiry was enough to get Horne to agree to meet to discuss VoceVista. But there was to be an additional twist. In addition to spectral analysis, I was hoping to add electroglottographic capability (from Martin Rothenberg’s Glottal Enterprises) to the software program. And so it was Richard Horne who would subsequently integrate both the acoustic and EGG signals into the original Windows version of VoceVista.

But still a serious impediment to the wider use of the VoceVista system remained in the high price of the EGG, which had been designed and available at the time for medical use only. After some inquiries, Schutte was able to find a resourceful Dutch engineer, Gerrie Goeree, who recreated and personally manufactured a unique VoceVista EGG that resulted in excellent waveform signals, all for a price that singers and singing teachers would be able to afford.

The following decade brought improvements in our software, principally Richard Horne’s integration of his own spectrogram software into VoceVista. This new technology made it possible to view a singer’s resonance adjustments in real-time, thus allowing the signals to be projected for an audience to see the changes as the singer made them. It was then in June of 2007 that we invited pedagogue Scott McCoy, author of Your Voice, an Inside View, to conduct the first-ever “wired master class” for some students of Dan Ihasz in Fredonia. The class was a great success and Scott began to use the program on his own later that summer.

Knowing that Scott had previously self-published Your Voice, which provides a basic acoustic and physiological background for vocal pedagogy, I began to envision VoceVista as an extension and practical realization of the voice science contained in his book. Recently chosen as the incoming president of NATS, Scott would have influence over the program content for the 2008 biennial conference in Nashville, so he made the bold decision to begin the program with a wired master class using VoceVista. This development allowed me to see an important opportunity, if not an obligation, to write a book that could serve as an extended manual for the use of VoceVista and the EGG in time to be available at the NATS conference. I asked Scott if he would edit and publish it, and without hesitation he agreed. Thus, I began work on a draft of Resonance in Singing.

The actual writing of Resonance in Singing proceeded smoothly as I had already covered most of the content in a series of published articles that resulted in my dissertation for the Ph.D. in Medical Sciences (University of Groningen in 2000) entitled Registers in Singing: Empirical and Systematic Studies in the Theory of the Singing Voice. Scott turned out to be a helpful and prompt editor, even while fulfilling his full-time faculty position at Westminster Choir College and managing the presidency of NATS. Though my writing was momentarily interrupted by an emergency appendectomy in the spring of 2008, the book was ready in time for the conference in Nashville, where the wired master class with VoceVista would be the opening plenary item on the agenda.

I think it safe to say that few in the audience fully understood what was presented in that abbreviated master class, but the connection of pedagogy with voice science was obvious. A small group of interested singing teachers — we nicknamed ourselves “Fryers” from the use of vocal fry to locate the formants — continued to exchange information and share insights, promoting our biennial workshops on the Physiology and Acoustics of Singing (PAS) in San Antonio, Stockholm, and Las Vegas (after previous workshops in Groningen and Denver).

The basic acoustic analysis software with the addition of EGG became known as VoceVista Pro, and together with the book Resonance in Singing, continues to be an important tool for singing teachers throughout the world. It is the basis for a doctoral dissertation, soon to be published, by Stephen Robertson, head of Voice Studies at the Royal Conservatoire of Scotland. It covers his use of the program in instructing students, as well as analyzed data from successful professionals on the international opera circuit.

Though gaining in popularity in the voice pedagogy world, VoceVista was about to undergo yet another period of change. Bodo Maass, who had created the software program Overtone Analyzer, offered to update and improve the VoceVista software, adding a video signal to the basic audio and EGG signals, along with other technological advances. We are at present (January 2020) in the homestretch towards the finished product, and are additionally in the process of a major edit for the second edition of Resonance in Singing to match the updates to VoceVista Video Pro.