This article was downloaded by: [University of West Florida]

On: 31 December 2014, At: 06:43

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer
House, 37-41 Mortimer Street, London W1T 3JH, UK

Voice and Speech Review

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/rvsr20

Peer Reviewed Article Subglottal Pressure

Variation in Actors' Stage Speech
a b
Malin Abrahamsson & Johan Sundberg
a
Malin Abrahamsson. Born in Vänersborg, Sweden in 1974. Received a master's
degree in Speech Therapy from the University of Gothenburg in 2002, and is
currently working as a speech pathologist and therapist at the Department of
Logopedics and Phoniatrics, at Sahlgrenska University Hospital, in Gothenburg,
Sweden. Within her own company Vocalization, Ms. Abrahamsson also works with
voice development aimed at speech and singing voices. She has extensive practical
experience in the art of music; both as choir director, background singer and as a
soloist.
b
Johan Sundberg. Born in 1936. Received his PhD musicology, doctor HC University
of York, UK, 1996. Professor of music acoustics at KTH (Royal Inst. Technology)
Stockholm, founded and headed its music acoustics research group, retired 2001.
Research particularly on singing voice and music performance. Written The Science
of the Singing Voice (1987) and The Science of Musical Sounds (1991), edited or co-
edited many proceedings of music acoustic meetings. He has practical experience
in performing music (choir and solo singing). Member of Royal Swedish Academy of
Music, Swedish Acoustical Society (President 1976–81) and fellow of the Acoustical
Society of America, its Silver Medal in Musical Acoustics 2003.
Published online: 25 Nov 2013.

To cite this article: Malin Abrahamsson & Johan Sundberg (2007) Peer Reviewed Article Subglottal Pressure Variation
in Actors' Stage Speech, Voice and Speech Review, 5:1, 343-346, DOI: 10.1080/23268263.2007.10769783

To link to this article: http://dx.doi.org/10.1080/23268263.2007.10769783

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”)
contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors
make no representations or warranties whatsoever as to the accuracy, completeness, or suitability
for any purpose of the Content. Any opinions and views expressed in this publication are the opinions
and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of
the Content should not be relied upon and should be independently verified with primary sources of
information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands,
costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or
indirectly in connection with, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or
systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in
any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Peer Reviewed Article by Malin Abrahamsson & Johan Sundberg

Subglottal Pressure Variation in Actors' Stage Speech

*This investigation was co-author MAs master thesis carried out during her speech and
language pathologist and therapist education at the Sahlgrenska University Hospital,
Gothenburg, Sweden. Co-author JS served as the supervisor. A more complete account of
the research is under preparation.

Introduction
In training actors, many teachers pay attention to the breathing habits. Most
voice teachers agree that the breathing habits are important to the voice, and
that inappropriate breathing habits can damage the voice.

In conversational speech the typical breathing pattern is quite straightforward.

We inhale a small quantity of air, perhaps half a liter. With that quantity of
air in the lungs, the elasticity of the respiratory apparatus provides an over-
Downloaded by [University of West Florida] at 06:43 31 December 2014

pressure of air that is sufficient for this type of speech. But the expiratory
elasticity forces decrease with decreasing lung volume. Therefore, as the
person releases the air and the lung volume decreases while the utterance is
Malin Abrahamsson. Born in Vanersborg, Sweden in
produced, expiratory muscular effort is needed to maintain the necessary air
1974. Received a master's degree in Speech Therapy pressure in the lung. Moreover, the lung pressure, the primary controlling
from the University of Gothenburg in 2002, and is factor for vocal loudness, is generally quite constant in neutral speech. Stress
currently working as a speech pathologist and thera-
pist at the Department of Logopedics and Phoniatrics
and phrasing are typically reflected in terms of variation of syllable duration
at Sahlgrenska University Hospital, in Gothenburg, and voice pitch. Thus, the breathing apparatus mostly has an easy job when
Sweden. Within her own company Vocalization, Ms. we talk.
Abrahamsson also works with voice development
aimed at speech and singing voices. She has exten-
sive practical experience in the art of music; both as Things are more complex in acting. Acting is often performed in front of
choir director, background singer and as a soloist. large audiences requiring a loud voice, and a loud voice requires high lung
pressure and consumes a relatively large amount of air. This implies that large
lung volumes are needed.

Some years ago, co-author JS and collaborators measured the lung pressure in
a professional actor who recited a poem as if he were addressing a large audi-
ence (Sundberg et al. 1993). Lung pressure was about 10 cm H20, which is
about twice as high as needed for conversational speech. But, more surprising,
the pressure was far from constant. Instead, it showed a number of sudden
increases and decreases. As this investigation was a single subject study, the
question remained if this actor's pressure behavior was representative of actors
in general.

Experiment
Three male and three female adult actors, all professionals, volunteered as
subjects. They recited the same Swedish poem as if addressing a large audi-
Johan Sundberg. Born in 1936. Received his PhD
ence. They also gave some samples of conversational speech. The speech was
musicology, doctor HC University of York, UK, 1996. picked up by a head mounted microphone. Air pressure data were obtained
Professor of music acoustics at KTH (Royal Inst. by asking them to swallow a thin catheter inserted through the nose. At the
Technology) Stockholm, founded and headed its
music acoustics research group, retired 2001.
tip of the catheter there was a pressure transducer, and the catheter was
Research particularly on singing voice and music per- adjusted such that this transducer was located in the esophagus. Figure 1 (next
formance. Written The Science of the Singing Voice page) shows a picture of one of the subjects during the experiment. This pro-
(1987) and The Science of Musical Sounds (1991 ),
edited or co-edited many proceedings of music
cedure was straightforward and was not much more dramatic than swallowing
acoustic meetings. He has practical experience in spaghetti. The esophageal pressure is lower than the subglottic pressure, since
performing music (choir and solo singing). Member the lung elasticity contributes only to the latter. Nevertheless, pressure
of Royal Swedish Academy of Music, Swedish
Acoustical Society (President 1976-81) and fellow of
changes that take place in the trachea, e.g., changes of subglottic pressure
the Acoustical Society of America, its Silver Medal in during speech, are reflected as changes of the esophageal pressure.
Musical Acoustics 2003.

343
 Voice and Speech Science, Vocal Health
Subglottal Pressure Variation in Actors' Stage Speech by Malin Abrahamsson & Johan Sundberg (continued)

not unexpected in view of the degree of vocal loudness needed

in stage speech when addressing a great audience.

More interesting is that the pressure curves for the perform-

ance recitation contains a number of more or less sharp peaks,
while, as expected, the curves for neutral speech are quite
smooth. The pressure peaks occurring in the performance
recitation were mostly at least twice as high as the pressure
peaks in the conversational speech.

The Swedish poem contained 52 syllables, so the entire

recitation material consisted of 6 x 52 = 312 syllables. In this
Downloaded by [University of West Florida] at 06:43 31 December 2014

Figure 1. Photograph of one of the subjects during the recording session. material we observed a total of 77 clear pressure pulses. Of
The head mounted microphone was fastened to the end of a copper wire these 43 appeared on stressed vowels, 17 on unstressed vowels,
covered by white plastic tubing. The pressure transducer was inserted
and 17 on consonants, nine of which were continuants such as
through the nose and fastened by means of a band aid. The subject gave
his permission for use of the picture of himself for publication purposes. /l, m, n/. These pulses were obviously produced by respiratory
muscles. Pressure pulses of comparable amplitudes never
occurred or occurred rarely in the recordings of the actors'
The pressure and speech signals were recorded on separate conversational speech.
tracks on a multichannel TEAC PCM DAT recorder. The audio
signal was calibrated by recording a sustained vowel, the sound A careful analysis of the timing of these pulses showed quite
pressure level (sPL) of which was measured by means of a interesting results. About half of the pulses were synchronous
sound level meter that was held next to the recording micro- with stressed vowels, while the remainder appeared on certain
phone. The SPL value thus observed was announced on the categories of consonants, such as /m/, In/, and Isl, i.e., on
recording. The pressure transducer was calibrated by recording consonants which, unlike stop consonants, can be sustained,
its output signal in free air and while it was immersed into also called continuant consonants. The consonants produced
water. The distance between the transducer and the water with pressure pulses did not seem to be strategically chosen.
surface was announced on the recording as the water
column pressure. Figure 3 shows two examples of consonants produced with
pressure pulses, one produced by a male actor (arrow in left
Results panel) and the other by a female actor (arrow in right panel).
Figure 2 compares pressure curves recorded from two actors' The male subject increased his pressure by more than 50%
conversational speech and from their reciting of the poem. during the vowel lo:/ such that the pressure peak occurred at
The differences are striking. The pressures were much higher the onset of the consonant /m/. During this pressure increase
for the stage speech than for the conversational speech. One the sound level and Fo also increased. The Fo peak occurred
actor actually used pressures as high as 60 cm H10. This is slightly earlier than the pressure peak. Pressure, Fo and sound
level decreased after the pressure peak. The
Conversational Speech Stage Speech
female subject synchronized a pressure peak
with the nasal /n/ in the word /svenska/
~
:z:
(Swedish). The pressure pulse was accompanied
by a marginal increase of the sound level and
.! an Fo decrease.
!

!Q. The relative amplitude of each pressure pulse

= observed in the entire material was measured as
1 illustrated in Figure 4, by computing the ratio
"'
..c
between the amplitude of the peak (Ppeak) and
! the average of the pressure amplitudes that
appeared just before (Pbefore) the pressure
Time[s] increase ending with the peak, and at the end
Figure 2. Examples of two actors' esophageal pressure during conversational speech and of the pressure decrease that followed the peak
stage speech (left and right panels, respectively). Note the numerous pressure peaks in the (Pafter). This relative pulse amplitude, henceforth
stage speech.
 Voice and Gender
The Voice & Speech Review

RPA, thus reflected the maximum relative

0
'O pressure increase during the pulse.
"
<(

The results revealed that the actors produced

carefully timed pressure pulses when they recited
the poem, and thus controlled their breathing
0
"'I apparatus quite accurately. To find an answer to
E
u the question whether or not these pressure pulses
iii
if were performed to produce stress, we asked a
panel of listeners to rate the degree of stress of
each syllable in each of the six actors' renderings
of the poem. The listeners gave their ratings by
Downloaded by [University of West Florida] at 06:43 31 December 2014

markings on visual analogue scales, one for each

syllable.

We then examined the correlation between the

median of the ratings of perceived syllable stress
and all RPA values exceeding 1.25. Also obtained
was the relation between the median of the
ratings and (1) the sound level, (2) the mean F0
during the syllable, and (3) the duration of
the syllable.

Time[s) Time[s)
Figure 3. Examples of timing of the pressure pulses, marked by arrows, in Actor RPA N SPL FO Dur
two of the actors. In the upper part of the figure the curves show, from top:
audio, sound level (l), pressure (P), and voice fundamental frequency (FO). 1F 0.066 7 0.477 0.177 0.704
The lower part of the figure shows the associated spectrogram. In the left 2F 0.265 12 0.466 0.254 0.658
panel the pressure pulse occurs at the onset of the consonant Im/, and in the
right panel the pressure peak is synchronized with the consonant /n/. 3F 0.368 9 0.512 0.389 0.721

1M 0.673 9 0.540 0.331 0.688

2M 0.570 16 0.403 0.299 0.732
3M 0.221 24 0.517 0.308 0.662

Table 1. Correlation coefficients (linear regression) between the

median of the quantized rated syllable stress and the relative
pulse amplitude (RPA), mean SPL during the syllable (SPL),
mean Fo during the syllable (Fo). and syllable duration (Dur). N
represents the number of RPAs exceeding 1.25. The correlations
0
'O with SPL, Fo and Dur were based upon all 52 syllables in the
"
<( poem. Correlation coefficients above 0.600 are bolded.

A quite unexpected and interesting result was obtained,

as can be seen in Table 1. There was no clear relation-
2
I ship between the respiratory pressure pulses and the
E
u
iii
ratings of syllable stress. Rather, these stress ratings
if seemed more related to syllable duration. For the different
actors, the correlation between the median of the
ratings of stress and the RPA values varied between
Time 0.066 and 0.673. The correlation between mean rated
Figure 4. Illustration of the method used for measuring the relative pressure stress and sound level, mean Fo and syllable duration
amplitudes PRA. The amplitude of the pressure peak (Ppeak) is divided by was clearly higher, duration showing the highest
average of the pressures occurring before (Pbefore) and after (Patter) the peak.

345
 Voice and Speech Science, Vocal Health
Subg!.ottal Pressure Variation in Actors' Stage Speech by Malin Abrahamsson & Johan Sundberg (continued)

correlations, between 0.658 and 0.732. This shows that our six Our results show that in stage speech subglottal pressure is
actors used syllable duration as a main tool for signaling stress. changed quickly and accurately dependent upon the text
spoken. A need to change subglottal pressure quickly and
Discussion accurately occurs also in singing and wind instrument playing,
The pressures measured in some of our actor subjects were where this pressure must be adapted not only to loudness but
quite extreme. While in conversational speech subglottal pres- also to pitch (Fuks and Sundberg 1999; Leanderson et al.
sures normally are lower than 10 cm H20, pressures in the 1987). It is thought-provoking that singers, wind
vicinity of 40 cm H20 were not uncommon in the actors' instrumentalists, and actors tend to use the term "support" for
stage speech. One actor even reached pressures near 80 cm an important aspect of their breathing technique, thus sug-
H20, as was shown in Figure 2. There are, however, strong gesting that "support" is related to the need to constantly
reasons for regarding these pressure readings as reliable. First, change subglottal pressure.
for each subject the pressure transducer was calibrated directly
Downloaded by [University of West Florida] at 06:43 31 December 2014

after the experiment. Second, the actors used normal pressures In summary, this investigation has shown that actors
during conversational speech. Third, the stage speech was of performing recitation as if in front of a large audience:
course quite loud. It should also be mentioned that the actor 1. change subglottal pressure quickly and accurately,
who used the highest pressures had a harsh voice during 2. producing pressure pulses synchronized with not only
the day of the recording, possibly because of some vocal stressed vowels bur also consonants.
fold disorder. 3. The pressure pulses did not increase the degree of
perceived stress for the syllable where they appeared;
The results pose the question why the actors produced these 4. rather, stress was perceived when syllable duration was
carefully timed pressure pulses. Enhancing text intelligibility longer.
seems a likely answer, at least in some cases. In the total 5. It seems likely that the pressure pulses increase text
material consisting of 312 syllables, 77 relative pulse amplitude intelligibility.
values exceeding 1.25 were observed. Of these, slightly less
than half occurred on unstressed vowels or on consonants, and
nine of the 17 pulses observed during consonants occurred
during continuants such as /l, m, n/. This means that, on
average, a pressure pulse was produced on every fourth
syllable. Pulses of similar amplitudes were very rarely observed
in the actors' conversational speech, so it seems quite likely References
Sundberg, J., N. Elliot, P. Gramming & L. Nord "Short-term variation of
that the production of short pressure pulses belong to the charac- subglottal pressure for expressive purposes in singing and stage
teristics ofactors' stage speech. speech. A preliminary investigation~ Journal of Voice 7, (1993), 227-234.
Fuks, L. & J. Sundberg "Blowing pressures in bassoon, clarinet, oboe
and saxophone'; Acustica/Acta acustica 85, (1999), 267-277.
The pressure pulses appearing on consonants are crucial to Leanderson, R., J. Sundberg & C. von Euler. "Role of the diaphragmatic
activity during singing: a study of transdiaphragmatic pressures':
text intelligibility. An increase of subglottal pressure enhances J Appl Physiol 62, (1987), 259-270.
the higher overtones of the source spectrum more than the
lower overtones, such that the higher overtones become more
prominent in the spectrum. As consonants are associated with
characteristic patterns in the second and third formants, the
amplitudes ofpartials in the higher part ofthe spectrum are rele-
vant to the identification ofconsonants. At the same time conso-
nants are not very loud, but their higher partials in particular
will become louder if they are produced with a high lung pressure.
In this sense, the consonantal pressure pulses would contribute to
text intelligibility.

We would also like to speculate that the key to the respiratory

behavior that produces these pressure pulses is not so much
awareness of breathing as the desire to communicate and the
urge to reach distant listeners. Nevertheless, pedagogy tends to
profit from the teachers' awareness of the physical realities
underlying the skills being taught.