Change in Singing Voice Production,

Objectively Measured
*Harm K. Schutte, James A. Stark, and Donald G. Miller
Sackville, New Brunswick, Canada and Groningen, The Netherlands

Summary: Although subglottal pressures in conversational speech are

relatively easily measured and thus known, the higher values that sometimes
occur in singing (especially in tenors) have received little attention in the
literature. Still more unusual is the opportunity to measure a large-scale
change over decades in the application of pressure in singing production. This
study compares measurements of subglottal pressure in a tenor/singing teacher
(JS) at two points in his career: in his early thirties, when he was a subject in
HSs dissertation study on the efficiency of voice production; and recently, in his
fifties, in connection with JSs forthcoming book on the history of the pedagogy
of Bel Canto. Although a single case study, its points of special interest include
the high values initially measured (up to100 cm H2O) and the reduction of this
figure by more than 50% in the maximal values of the recent measurements. The
study compares these values with those of other singers in the same laboratory
(both with esophageal balloon and directly, with a catheter passed through the
glottis) and in the literature, as well as discusses in detail the problems pertaining
to the measurement (repeatability, correcting for lung volume, etc.).
As a sophisticated subject, JS makes some pertinent observations about the
changes in his use of subglottal pressure.
Key Words: Subglottal pressureAir flow rateSinging voiceTenor
Measuring procedure.

INTRODUCTION
Although subglottal pressures in conversational
speech are relatively easily measured indirectly,13
and thus known, the higher values that sometimes
occur in singing (especially in tenors) have received
little attention in the literature. In general, researchers have reported subglottal pressures for conversational speech ranging from about 7 to 10 cm H2O,
reaching around 10 to 12 cm H2O for loud speech,
and about 40 cm H2O for shouting. Some authors
have reported subglottal pressures in singing to reach
40 to 70 cm H2O, with the upper limit rarely above
60 cm H2O.4,59 Furthermore, Schutte has measured
even higher levels, especially for tenors, who have
reached 100 cm H2O or more on high, loud notes.10,11

Accepted for publication January 30, 2002.

Paper presented at the 26th Annual Symposium: Care of the
Professional Voice, The Voice Foundation, Philadelphia, PA,
June 1997.
From the *Department of Music, Mount Allison University,
Sackville, New Brunswick, Canada; Groningen Voice Research Lab, Department of BioMedical Engineering, Faculty
of Medical Sciences, Groningen, The Netherlands.
Address correspondence and reprint requests to Harm K.
Schutte, Groningen Voice Research Lab., Dept of BioMedical
Engineering, Faculty of Medical Sciences, Ant. Deusingloan 1,
9713 AV, Groningen, The Netherlands. E-mail: h.k.schutte@med.
rug.nl
Journal of Voice, Vol. 17, No. 4, pp. 495501
2003 The Voice Foundation
0892-1997/2003 $30.000
doi:10.1067/S0892-1997(03)00009-2

495

496

HARM K. SCHUTTE ET AL

The Groningen Voice Research Lab has now had

the unusual opportunity to measure a marked change
in the levels of subglottal pressures in the singing of
one such tenor-subject (one of the authors) spanning
a period of more than 20 years. This study compares
measurements of the subject, a professional tenor
and singing teacher, at two points in his career, first
when he was in his mid-thirties and was a subject in
Schuttes dissertation, The Efficiency of Voice Production (1980)11, and second, when he was in his
mid-fifties and writing a book on the history of
vocal pedagogy.12
Although this is a single case study, its points
of special interest include the high values initially
measured, and the marked reduction of this figure
in the maximal values of the recent measurements.
The subject associates this difference with a deliberate change in his vocal techniquea change that
was based on his interpretation and application of
singing methods described in certain historical voice
treatises. In this study, we will also discuss in detail
the problems pertaining to the measurements of subglottal pressures.
METHOD
In measuring subglottal pressure in professional
singers, one might encounter a few typical problems,
depending, among other things, on the measurement
technique employed. At present, one can choose
between five techniques:
I. Direct Measurements
1) Direct measurement by using a needle inserted
through the cricothyroid membrane has obvious psychological drawbacks for the singer.
One would be very reluctant to do this with
singers in the midst of their careers, although some important data have been obtained in the past by using this technique.13
2) Direct measurement by means of a catheter
with wide-band pressure transducer, which is
passed through the glottis, recording the
pressure changes above and below the vocal
folds.14
II. Indirect Measurements
1) The technique of using intra-oral pressure
measurement, interpolating a value for the
Journal of Voice, Vol. 17, No. 4, 2003

subglottal pressure from a following and

preceding occlusive (/bp/method after Rothenberg15), is routinely accepted for pressures
used in speech, but is questionable for the
measuring of the high pressures singers tend
to use15.
2) Using an esophageal balloon: a) while keeping
track of the changing lung volume during phonation; b) The Van den Berg maneuver (explained below). In this study, results are given
of both methods employing the esophageal
balloon.

Van den Berg introduced the indirect method,

named after him, in 195616. This method is based on
an informed application of the principles of lung
mechanics. Intrapleural pressure, that is the pressure
between the pleura blades, is directly related to
the intrathoracic pressure and can be measured by
means of an esophageal balloon, brought into the
lower part of the esophagus. Changes in the intrathoracic pressure and thus in the pressure below the
vocal folds during phonation are indirectly measured
in this way. However, changes in lung volume during
breathing and singing influence the value of the
intrapleural pressure. Esophageal pressure during
phonation thus has two components: (1) a negative
component, which is a function of lung volume,
(the greater the lung volume the larger the negative component), and (2) a further component,
tracheal pressure, amounting to the pressure difference between atmospheric pressure and pressure at
the trachea. This second component, also called subglottal pressure, is of course positive in the act of
expiration against any barrier presented by the narrowed glottis, or other articulatory structures of the
vocal tract. There is a well known relationship
between lung volume and esophageal pressure under
the condition of an open airway, and this can be
used in distinguishing the two different components
of the esophageal pressure. The easiest way to monitor the momentary value of the lung volume is
by measuring the amount of air that is displaced by
breathing. This requires a pneumotachograph, connected to either a mouthpiece or a mask, either of
which produces an effective elongation of the vocal

SINGING VOICE MEASURED

tract and a distortion of its normal acoustic properties. The mask influences at the very least the auditory feedback of the singer-subject, who must then
cope with acoustic conditions of the vocal tract that
differ from the normal conditions of singing.
As the value of the subglottal pressure used in
singing tones is simply added to the first (lungvolume dependent) component of the esophageal
pressure, a phonation stop at the end of a tone
will reduce the value of the esophageal pressure by
the value of that pressure at the end of the phonation.
This is the basis of the Van den Berg maneuver16;
however, a proper and well-executed maneuver must
fulfill three conditions: (1) abrupt cessation of the
act of phonation, because even a small decrescendo
will result in an unwanted reduction of subglottal
pressure; (2) maintenance of an open glottis for a
short period immediately following the cessation of
phonation (otherwise intrathoracic pressure reflected
in the esophageal pressure can remain high); and (3)
avoidance of inhalation or exhalation directly following the phonation stop, because any change in
lung volume or transient pressure difference between trachea and outside air will change the value
of the esophageal pressure. Both approaches to
measuring subglottal pressure using an esophageal
balloon (II2a and b) are depicted in Figure 1. Detailed information on the reliability and accuracy of
the method can be obtained from Schutte11.
In 1974, the airflow rate was measured to register
the changes of lung volume. In this way, highly
reliable values were obtained for subglottal pressure.
These values were also occasionally checked using
the Van den Berg maneuver. In 1996, a higher level of
sophistication in experimental procedure, as well as
of singing skill, made it possible to use the Van den
Berg maneuver exclusively. This avoids the disadvantage of the elongated vocal tract and its disturbing
effect on the singers strategies for optimizing resonances. There were in principle no experimental procedural differences in obtaining subglottal pressure.
The only difference existed in the choice of the
pitches used in the measurements. In 1974, the measurements were aimed at obtaining aerodynamic
data (and calculated efficiency) over the full dynamic range, divided into steps of five decibels, at
preselected fundamental frequencies (A2, E3, A3,

497

E4, G4, and A4)11. In 1996, data of subglottal pressure were obtained over the whole fundamental
frequency range in semitone steps for three subjective levels of loudness. The whole range was first
produced at a comfortable level, then fortissimo, and
finally pianissimo.
To avoid the inherent tendency to adjust subglottal
pressure incrementally in adjacent semitones, each
phonation was removed from the previous one by
a perfect fifth plus or minus a semitone (see Figure
2). The phonation series starts with the G above
middle C, and then the middle C is sung. The third
tone is then one semitone lower than the first sung
tone, G-flat above middle C, and then follows B
below middle C. This is continued until the lowest
attainable pitch is reached. Upward, the series starts
with A-flat above middle C, followed by E-flat, a
perfect fifth above that tone, then A-natural, and so
on, until the highest possible pitch is reached. In
this way, the singer could concentrate fully on the
production of a tone of optimal quality, judged by
his experienced and trained ear.
All signals on both occasions were registered
on a Mingograph MT800, eight-channel ink writer,
which was also used for monitoring the pressure
changes in the esophagus. In case of an imperfect
Van den Berg maneuver or an involuntary increase
of the esophageal pressure due to swallowing, the
phonation was repeated immediately.
RESULTS AND DISCUSSION
The results of the measurement series have been
depicted in Figure 3. The 1974 measurements consist
of two measuring series, 2 days apart. The data have
been lumped together, except for the data on G4
(392Hz) and A4 (440Hz). Data on G4 (five instances
of phonations) were obtained only in the first session,
and the data obtained on A4 (440Hz) show differences due to an indisposition caused by travel fatigue
on the first measuring day and will thus be shown
separately.
The 1996 data are depicted by open circles for the
pianissimo tones and closed circles for the fortissimo
sung tones. The measuring points are connected with
lines for the whole frequency range. It is interesting
to note that the differences in sound pressure level
between artistically usable fortissimo and pianissimo
Journal of Voice, Vol. 17, No. 4, 2003

498

HARM K. SCHUTTE ET AL

FIGURE 1. A verification of the indirect measuring method registering the changes

in the esophageal pressure by means of the esophageal balloon, upon the abrupt
cessation of phonation, the so-called Van den Berg maneuver, is depicted. For a brief
moment after phonation, the lung volume is kept constant with an open glottis. The
pressure in the esophagus drops to the value pc corresponding with the lung volume at
that moment. The pressure drop delta-p equals ps pr, which is the sum of the subglottal
pressure (ps) during phonation plus a pressure (pr) that equals the product of the viscous
resistance and the air flow rate during phonation. The pressure pr is negligible during
singing, because the flow is relatively low.

tones amounts to no more than about 15 dB, which

is much lower than the usual decibel differences
between loudest and softest (threshold) sung tones in
a voice range profile (phonetogram)11. It is clear that
pianissimo in artistic singing is not the same as
the softest phonation level. For this reason, we use the
musical term pianissimo for this loudness level.
At the other extreme of loudness, however, this distinction does not apply, and the upper contour of
the voice range profile is essentially identical to
fortissimo.
The 1974 data do not contain information on all the
semitones over the whole dynamic fundamental
frequency range, but is restricted to five fundamental frequencies. Together with the 1996 data, these
are shown in Figure 3. Especially for the midrange fundamental frequencies E3 (165Hz) and A3
(220Hz), the softest tone in 1974 is considerably
low in SPL, which also explains the low subglottal
Journal of Voice, Vol. 17, No. 4, 2003

pressure on E3 and A3, approaching the lowest pressure for a sustainable tone.
The most salient difference between the data from
1974 and 1996 measurements is in the values for
subglottal pressure in the upper part of the range.
For the E4 and A4, the 1996 values for subglottal
pressure are lower by about 30 cm H2O, making
the pressures for fortissimo roughly equivalent to
those for the softest possible tones at the earlier date.
Even on A3, subglottal pressure is higher by one third
in 1974. With the exception of A4, the highest pitch
measured in 1974, however, fortissimo SPL is only
marginally lower in 1996, in spite of the reduction
in subglottal pressure. (The 1974 values for G4, as
well as the 1974a values for A4, are uncharacteristically low. They were included in the figure for the
sake of completeness, but were deemed aberrant,
due to travel fatigue of the subject on the day of
their registration.)

SINGING VOICE MEASURED

The attainment of the 1974 sound pressure levels
at the reduced subglottal pressures of 1996 is evidence of improved efficiency, the exact nature of
which is a matter of conjecture. An undesirable consequence of the change in technique is the apparent
loss of the ability to produce the highest pitches
(those above G4) with an SPL appropriate to high
subglottal pressure. The high SPL and relatively
low subglottal pressure of the 1974a A4, however,
are an indication that the subglottal pressure may
not be the decisive factor in the SPL. In any case, it
is evident from the full set of data that the relationship
between subglottal pressure and SPL is not a simple
one, and that other factors, such as proximity of
harmonics to resonances of the vocal tract, can play
an important role.

PEDAGOGICAL CONSIDERATIONS
The marked differences in subglottal pressures in
the case of our tenor-subject was due to his deliberate
change in vocal technique over a 22-year period,
guided largely by his interpretation and application
of vocal methods described in certain historical treatises on singing. At the time of the 1974 measurements, the singer was applying the advice of the
celebrated 19th-century Italian voice teacher Francesco Lamperti and his equally famous son, Giovanni Battista Lamperti, both of whom seemed
to advocate singing with high subglottal pressures.
Francesco advocated holding back the breath, and
not permitting more air [to escape] than is absolutely
necessary.17 Giovanni amplified this comment by
saying that the breath should be held back by strong
glottal resistance, resulting in compressed breath.
This compressed breath was the foundation of his
method18,19. In Germany, at about the same time,
Georg Armin (1909) and Rudolph Schilling (1925)
developed a theory called Stauprinzip, or breath
damming, which pointed to the use of high subglottal pressure and strong glottal resistance20,21. Richard
Miller has described breath damming as a technique
of breath retention through marked sub-glottal muscular pressures. The flow of breath is stemmed by
the glottis as a result of muscular tension similar
to that experienced in a painful groan or grunt.
According to Miller, this groaning utterance, called

499

FIGURE 2. Pitch protocol used in this study. See text for

explanation.

Stonlaut, is the primitive power of the vocal instrument, and that a long list of successful German
singers in this century have given allegiance to it,
including a number of Wagnerian Heldentenore22.
Our tenor-subject, whose early vocal training had
been based on relaxation techniques with low breath
pressures and high rates of airflow, modified his
technique by developing strong glottal closure and
high levels of subglottal pressure, consistent with the
concepts of compressed breath and Stauprinzip. It
was these higher pressures that were measured in
Groningen in 1974. They were consistent with
similarly high pressures in the other tenor subject
in Schuttes measurement11, noted that the technique
of Stauprinzip has been rejected generally, because it may lead to damage of the voice. However,
Schilling, in 1922, did not categorically dismiss
Stauprinzip, and Schutte concurred that without
further investigation, this (the high subglottal pressure) is insufficient reason to discard this singing
method completely11.
Nevertheless, after learning to sing with these
high pressures, our tenor-subject became concerned
that he might be generating higher pressures than
were optimal for his type of lyrical singing. He found
it difficult to execute mezza voce, diminuendo, messa
di voce, or other forms of low-intensity singing. As
well, he characterized the voice quality as heavy
or labored, and lacking in buoyancy. He therefore turned to the treatises of another famous 19thcentury voice teacher, Manuel Garcia II, who also
advocated strong glottal closure, but with steady,
moderate, and prolongued pressure. Garcia
maintained that glottal closure should be achieved
with firm adduction of the interarytenoid muscles
by pinching the glottis. He further maintained that
Journal of Voice, Vol. 17, No. 4, 2003

500

HARM K. SCHUTTE ET AL

FIGURE 3. The results of the measurements in 1974 and 1996. In the upper part, the
values are given for the Sound Pressure Level, measured at a microphone-mouth distance of
30 cm. In the lower part, the values for subglottal pressure are given. See text for discussion.

the firm closure of the arytenoid cartilages had the

effect of shortening the vibrating portion of the glottis and reducing airflow rates, and gave a bright edge
(eclat) to the sound source. Garcias vocal method
was based on the coordination of glottal settings
and vocal tract adjustments intended to put the
singer in possession of all the tints of the voice23,24.
Garcias pupil, Hermann Klein, renewed the call for
compressed breath, but noted that the singer must
be guided by ease and economy of breath pressure25. He said that Garcias first rule was to repress the breathing power and bring it into proper
proportion with the resisting force of the throat and
larynx26. In interpreting Garcias advice, our tenorsubject again modified his technique, this time by
reducing the subglottal pressures while maintaining
Journal of Voice, Vol. 17, No. 4, 2003

strong glottal closure. The Groningen measurements

of 1996 confirmed the reduction of both subglottal
pressures and airflow rates compared to the 1974
measurements. In the singers opinion, this was accomplished by consciously reducing the degree of
contraction of the expiratory muscles during singing.
The chief vocal exercise used to achieve this was
the messa di voce, that is, the crescendo-decrescendo
of a long note.
CONCLUSION
What seems to us remarkable in this study is the
degree to which the subject was able, through conscious application of a principle, to change substantially the levels of subglottal pressure used

SINGING VOICE MEASURED

habitually in singing. The high levels of the 1974
measurements resulted from a singing technique acquired deliberately, and the reduction reflected in
the 1996 measurements was equally the result of a
deliberate change. The fact that SPL was only affected marginally by the reduction is also remarkable, considering that subglottal pressure is generally
recognized as the controlling factor in vocal loudness. Finally, the fact that this subject was able,
without damage to his vocal folds, habitually to
use pressures in singing that many authorities might
consider excessive (eg, Proctor, who regards 25 cm
H2O as unusually high pressure in singing) is an
indication that singing methods advocating high subglottal pressures (eg Stauprinzip) are not necessarily
physiologically detrimental.

REFERENCES
1. Rothenberg M. The Breath-Stream Dynamics of SimpleReleased-Plosive Production.Basel, Switzerland: Karger, 1968.
2. Rothenberg M. Interpolating subglottal pressure from oral
pressure. J Speech Hear Dis. 1982;47:219220.
3. Lofqvist AG, Carlborg B, Kitzing P. Initial validation of
an indirect measure of subglottal pressure during vowels.
J Acoust Soc Am. 1982;72:633635.
4. Netsell R. Subglottal and intraoral air pressures during the
intervocalic contrast of /t/ and /d/. Phonetica. 1969;20:6873.
5. Proctor DF. Breathing, Speech and Song. Vienna, Austria:
Springer-Verlag; 1980.
6. Khambata AS. Anatomy and physiology of voice production: the phenomenal voice. In: Critchley M, Henson RA,
eds. Music and the Brain. London: William Heinemann
Medical Books; 1977:5977.
7. Sears TA. Some neural and mechanical aspects of singing. In:
Critchley M, Henson RA, eds. Music and the Brain. London:
William Heinemann Medical Books; 1977:7894.
8. Wyke BD. Ventilatory and Phonatory Control Systems.
London: Oxford University Press; 1977.
9. Baken RJ, Orlikoff RF. Clinical Measurement of Speech and
Voice. San Diego, CA: Singular Publishing Group; 2000.

501

10. Catford JC. Fundamental Problems in Phonetics.

Bloomington, Indiana: Indiana University Press; 1977.
11. Schutte HK. The Efficiency of Voice Production. 1980.
Thesis University of Groningen.
12. Stark J. Bel Canto: A History of Vocal Pedagogy. Toronto,
Canada: University of Toronto Press; 1999.
13. Rubin HJ, LeCover M, Vennard WD. Vocal intensity: subglottic pressure and air flow relationships in singers. Folia
Phoniatrica. 1967;19:393413.
14. Miller DG, Schutte HK. Characteristic patterns of sub- and
supraglottal pressure variations within the glottal cycle. In:
Lawrence VL, eds. Transcr XIIIth Symp Care Prof Voice.
Part I: Scientific Sessions. New York: The Voice Foundation,
1985:7075
15. Rothenberg M. A new inverse-filtering technique for deriving the glottal air flow waveform during voicing. J Acoust
Soc Am. 1973;53:16321645.
16. Van den Berg Jw. Direct and indirect determination of the
mean subglottic pressure. Folia Phoniatrica. 1956;8:124.
17. Lamperti F. A Treatise on the Art of Singing. New York:
E. Schuberth; 1871.
18. Lamperti F. The Technics of Bel Canto. New York: G.
Schirmer; 1905.
19. Brown WE. Vocal Wisdom: Maxims of Giovanni Battista
Lamperti. New York: Taplinger; 1957.
20. Armin G. Das Stauprinzip oder die Lehre von dem Dualismus der menschlichen Stimme. Strassburg: Carl Bongard;
1909.
21. Schilling R. Untersuchungen uber die Atembewegungen
beim Sprechen und Singen. Monatsschr Ohrenheilkd Laryngorhinol. 1925;59:5180,134153,313343,454467,643668.
22. Miller R. English, French, German and Italian Techniques
of Singing: A Study in National Tonal Preferences and How
They Relate to Functional Efficiency. Metuchen, NJ: The
Scarecrow Press Inc; 1977.
23. Garcia M. Traite complet de lart du chant. In Two
Parts. Paris: Chez lAuteur; 1847. Reprint ed. Geneve:
Minkoff; 1985:Part I:24; Part II:1821.
24. Garcia M. Hints on Singing. New York: The Joseph Patelson
Music House, 1982. Translated from the French by Beata
Garcia. New and Revised Edition. Hermann Klein, editor.
London: E. Ascherberg. 1984:13, 22.
25. Klein H. The Bel Canto, with Particular Reference to the
Singing of Mozart. London: Humphrey Milford, Oxford
University Press; 1923:2124.
26. Klein H. Thirty Years of Musical Life in London, 18701900. London: Heinemann; 1903:36.

Journal of Voice, Vol. 17, No. 4, 2003