Aerobic Exercise as a Warm-up for Singing:

Acoustic Impacts
*Monica A. McHenry and †Joseph Evans, *Valhalla, New York, and †Houston, Texas

Summary: Objectives. In a previous work, it was found that a 30-minute aerobic workout significantly increased
singers’ sound pressure level and airflow during voicing, suggesting a shift to flow phonation. This companion study
was designed to assess the impact of the same workout on pitch accuracy, vibrato rate, extent and regularity, and the
singing power ratio.
Study Design. This study is a cohort experimental study.
Methods. Twenty-two students in an academic vocal performance program participated. They performed an aerobic
workout for 30 minutes. Before and after the workout, they sang the first seven notes of the “Star-Spangled Banner”
on /pa/, producing seven /pa/s on the last note. The students then sang an ascending and descending scale to the ninth
on “ah.” The following measures were obtained from the “Star-Spangled Banner”: pitch accuracy calculated on the
seventh note (“by”); and vibrato rate, regularity, and extent, calculated on the most sustained sixth note (“see”). For
the scale, the following measures were calculated from each note: pitch accuracy; vibrato rate, regularity, and extent;
and the singing power ratio.
Results. There were no significant differences from pre- to postworkout across any measures.
Conclusions. It appears that an aerobic workout positively impacts the respiratory driving force for voice produc-
tion but does little for phonation. Critical for performance is the fine tuning and balancing across the respiratory, laryngeal,
and resonance systems. It appears that this can only be achieved with vocalization exercises, facilitating coordination
within and across the physiological systems involved in the complex art of bel canto.
Key Words: Vibrato–Singing power ratio–Pitch accuracy–Workout–Bel canto.

INTRODUCTION
Warming up before any physical activity is typically recom-
mended. Based on exercise physiology, a physical warm-up
facilitates muscle contraction and relaxation speed, movement
efficiency as a result of reduced viscous resistance in active
muscles, oxygen delivery and use, nerve transmission, and blood
flow.1 Applying these principles to vocal warm-up, Elliot et al2
speculate that increased blood flow in the laryngeal muscles would
result in decreased vocal fold viscosity, enhancing effortless
production.
There are many detailed investigations of the ideal warm-up
to optimize specific types of physical performance. One issue
addressed is the specificity of the warm-up.3 Investigators stressed
the importance of discipline-specific warm-ups, with a goal of
preparing the muscles that could potentially limit perfor-
mance. Another work focuses on the importance of warm-up to
reduce muscle damage that may contribute to delayed-onset
muscle soreness.4 These investigators also highlighted the im-
portance of cool-down to increase circulation, facilitating the
removal of cellular waste products resulting from exercise.
Another aspect to be considered is the timing of the physical
warm-up. As expected, longer elapsed times have a detrimen-
tal effect on performance, although this has only been studied
Accepted for publication October 27, 2016.
Presented at the 45th Annual Symposium: Care of the Professional Voice, June 5, 2016,
Philadelphia, PA.
From the *New York Medical College, Valhalla, New York; and the †School of Music,
University of Houston, Houston, Texas.
Address correspondence and reprint requests to Monica A. McHenry, Department of
Speech-Language Pathology, New York Medical College, 30 Plaza West, Ste. 214, Valhalla,
NY 10595. E-mail: monica_mchenry@nymc.edu
Journal of Voice, Vol. 31, No. 4, pp. 438–441
0892-1997
© 2017 The Voice Foundation. Published by Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jvoice.2016.10.023
under limited conditions.5 All of the variables impacting the ef-
fectiveness of warm-up for a physical activity are relevant to vocal
warm-ups as well. DeFatta and Sataloff6 emphasized the im-
portance of individualization based on the singers’ Fächer,
anticipated performance demands, individual response to warm-
up exercises, and general fitness level. These recommendations
are, for the most part, based on knowledge of muscle physiol-
ogy and anecdotal evidence, rather than on empirical data. One
of the challenges in translating the exercise physiology litera-
ture to vocal hygiene is the complexity of the vocal fold tissues.
Knowledge of muscle tissue is relevant to the body of the vocal
folds, as well as the supporting musculature, but does not address
the viscoelastic properties of the superficial layer, critical for ef-
fortless vocal fold vibration.7 Subtle details of the investigations
drastically impact the interpretation of results. For example,
Sandage et al7 measured phonatory threshold pressure immedi-
ately after completion of submaximal exercise, whereas in a
previous work, we obtained aerodynamic measures only after
rest, hydration, and a return to resting pulse rate.8 Finally, a much
neglected consideration in optimizing performance and mini-
mizing delayed negative effects of performance is cooling-
down vs vocal rest. Gottliebson9 found that a systematic cool-
down routine resulted in reported improved vocal function
compared to vocal rest or conversation. The cool-down con-
sisted of glides, arpeggios, and humming, designed to return to
an optimal pitch for speech. It was difficult to determine the source
of the reported benefit. The author speculated that cooling down
could have increased awareness of optimal resonance, or may
simply have been a psychological benefit. In a previous work
on the delayed effect of performance, we found a benefit of vocal
rest.10 Specifically, individuals who did not sing in a choir on
the morning after performing demonstrated better vocal function

Downloaded for Anonymous User (n/a) at Pontifical Catholic University of Chile from ClinicalKey.com by Elsevier on October
07, 2021. For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
Monica A. McHenry and Joseph Evans Impact of Aerobic Exercise as a Warm-Up for Singing
than those who did. It is clear that both warm-up and cool-
down exercises warrant further investigation, specifically regarding
their impact on vocal performance.
Warming up before vocal performance is considered stan-
dard practice. Students learn warm-up strategies in voice lessons,
and customize and adapt the most beneficial for regular prac-
tice. Typically, singers complete vocal exercises to prepare for
performance. The majority of singers report some sort of warm-
up, although its duration varies markedly.11 Anecdotally, the
greatest warm-up challenge appears to be for individuals with
higher Fächer who were required to perform early in the morning,
such as at a church service. Preparation for such a perfor-
mance requires more extensive vocalization because of the
typically increased mass of the vocal folds in the morning sec-
ondary to fluid accumulation in the superficial layer. A number
of years ago, Titze12 detailed the specific physiological ben-
efits of what he considered to be the five best vocal warm-up
exercises. These exercises are ideally designed to prepare for bel
canto. Bel canto is a term used to describe a style of classical
singing characterized by a well-rounded tone with very smooth
and effortless productions throughout the vocal range, with no
evidence of register shifts.13,14 In the first recommended exer-
cise, semi-occluded vocal tract exercises optimize vocal fold
configuration and lowered phonation threshold pressure. Two-
octave pitch glides stretch the vocal folds and optimized the
interaction between the thyroarytenoid and cricothyroid muscles,
facilitating the negotiation of passaggio, a key component of bel
canto. Vowel sequence /a/-/i/ scales loosen the tongue and jaw.
Messa di voce serves to coordinate laryngeal musculature with
changing lung pressure. Finally, staccato on arpeggios estab-
lishes an ideal vocal onset. Vocalization exercises are focused
on preparing and coordinating the respiratory and laryngeal
systems for the exacting requirements of performances in the
genre of bel canto. They do not, however, engage the entire body
in the warm-up.
In previous work, it was speculated that a physical, aerobic
warm-up may prove as beneficial as a purely vocal warm-up.8
The reader is referred to the previous study8 for more details on
the benefits of both vocal and aerobic warm-up strategies. The
earlier findings focused on the aerodynamic contributions to voice
production. Sixteen graduate and six undergraduate students com-
pleted a 30-minute treadmill workout in their aerobic heart rate
range. Aerodynamic data were collected before and after the
warm-up. It was found that the workout significantly increased
mean sound pressure level and mean airflow during voicing, sug-
gesting a shift to flow phonation, or an optimization of airflow.
It is recognized that simply increasing airflow in isolation can
lead to a weak, breathy production. The interpretation of the pres-
ence of flow phonation was based on the concomitant increase
in mean sound pressure level, which may have reflected not only
increased airflow but also an optimization of vocal tract
configuration.15 It was concluded that an aerobic workout had
beneficial effects on the aerodynamics of voice production.
The present work represents an acoustic analysis of the samples
obtained pre- and postworkout. It was hypothesized that singers
would improve in the following: pitch accuracy; vibrato rate,
extent and regularity; and the singing power ratio (SPR).
Downloaded for Anonymous User (n/a) at Pontifical Catholic University of Chile from ClinicalKey.com by Elsevier on October
07, 2021. For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
439
METHODS
The study was approved by the Institutional Review Board, and
all participants completed a consent form. Singers were as-
sessed after having been awake for at least two hours, with
minimal voice use. Before aerobic exercise, their resting heart
rate was established. Maximum heart rate was calculated as
220 − age. Finally, the target aerobic heart rate was calculated
as 60%–80% of their maximum heart rate.
Participants
Twenty-two students in an academic vocal performance program
participated. There were 16 graduate and six undergraduate stu-
dents. Voice types were 11 sopranos, 1 mezzo soprano, 5 tenors,
2 bass-baritones, and 3 baritones. Ages ranged from 18 to 38
years, with a mean of 24 years. Fitness levels were obtained using
the NASA Physical Activity Scale, shown to be moderately well
correlated with objectively measured physical activity values.16
Reported fitness levels were highly varied, ranging from 0 (avoid
walking or exertion) to 8 (running 16–20 mi/wk or walking
21–26 mi/wk or spending 6–8 h/wk in comparable physical ac-
tivity). The mean activity level across all participants was 6.1
(standard deviation [SD] = 1.8). A rating of 6 represents running
6–10 mi/wk, walking 7–13 mi/wk, or spending 1–3 h/wk in com-
parable physical activity, suggesting that the majority of singers
were relatively active.
Tasks
Before the aerobic workout, the participants wore a mask at-
tached to a pneumotachograph to capture airflow, with a pressure
sensing catheter placed behind the lips to capture intraoral air
pressure (Phonatory Aerodynamic System, PENTAX Medical,
Montvale, New Jersey, USA). The participants then sang the first
seven notes of the “Star-Spangled Banner” (SSB) on /pa/, in a
key appropriate for their Fächer. The singers produced a syl-
lable train of seven /pa/s on the seventh note of the SSB. They
repeated the abbreviated SSB three times. They then sang an as-
cending and descending scale to the ninth on “ah,” again in an
appropriate key, recorded in a quiet room on an iPad (Apple,
Cupertino, California, USA). For the scale, the singers were given
only the starting pitch and were not accompanied further.
After initial data acquisition, the singers’ pulse rates were moni-
tored as they walked or ran on a treadmill. The treadmill speed
was adjusted as needed to maintain the target heart rate range
for 30 minutes. After completion of the workout, the singers drank
½ L of water. When the singers’ heart rates had returned to their
preworkout level, data acquisition was repeated.
Data analysis
The audio recording of the SSB /pa/ productions were ana-
lyzed using Praat.17 Pitch accuracy was obtained from the second
of three trials, and was averaged across the second through sixth
repetitions of the seventh highest note. Pitch accuracy was de-
termined by comparing the target fundamental frequency with
the singer’s actual fundamental frequency, and converting the
hertz difference to semitones. All measures of vibrato were ob-
tained from the sixth note (“see”) in the second trial, because
the singers sustained that note longer than they did the repeated
440
/pa/ productions of the seventh note. Vibrato rate was deter-
mined by counting the number of vibrato cycles in the stable
portion of the vowel and dividing by time in seconds. Vibrato
regularity was determined by obtaining the SD of each period
of vibrato in the stable portion of the vowel. Vibrato extent was
calculated by determining the lowest and highest frequencies in
a vibrato cycle and converting those values to a semitone extent.
Pre- and postvibrato data were not obtained if the singer pro-
duced a straight tone in either condition. Two singers were
eliminated because of straight tone production.
The ascending and descending scale to the ninth was ob-
tained for only 10 singers because of technical problems. Pitch
accuracy and vibrato rate, extent, and regularity were obtained
for each note in the scale as described earlier. Additionally, the
SPR18 was determined by subjecting each note to a spectral anal-
ysis. The difference in the highest amplitude from 0 to 2000 Hz
and the highest amplitude from 2000 to 4000 Hz was calcu-
lated. The values were averaged across all notes in the scale. All
data were subjected to a paired t test to compare pre- and
postworkout values. Because there were nine comparisons of in-
terest, the Bonferroni correction was applied to an alpha level
of .05. To be significant, a p value of .006 was required.
Intrajudge reliability
Reliability was determined by remeasuring all data for a male
and a female participant. Reliability data were averaged across
participants and pre- and postconditions. Differences between
original and repeated measures were SSB pitch accuracy = .17
semitones, SSB vibrato rate = .09 cycles/second, SSB vibrato vari-
ability = .001 second, SSB vibrato extent = .07 semitones, scale
pitch accuracy = .008 semitones, scale vibrato rate = .001 cycles/
second, scale vibrato variability = .002 second, scale vibrato
extent = .02 semitones, and scale SPR = .5.
RESULTS
Pitch accuracy for the SSB was calculated based on the differ-
ence (above or below) from the target note to the actual note
on the seventh note (ie, “by”). The values in hertz were con-
verted to semitones. The mean pitch accuracies were −.13
(SD = .4) preworkout and −.23 (SD = .4) post workout. There
was no significant difference (t = 2.1, df = 20; p = 0.29). Ten par-
ticipants were less accurate post workout by more than .1
semitone; six were within .1 semitone pre- to post workout; and
five were better post workout by more than .1 semitone.
Pitch accuracy for the scale was determined for 10 singers
and was calculated for each note in the ascending and descend-
ing scale to the ninth. The differences in semitones between the
actual and target pitches were calculated. The individual differ-
ences of all notes across singers were input for statistical analysis.
The mean pitch differences were −.32 (SD = .47) preworkout and
−.29 (SD = .57) post workout. There was no significant differ-
ence (t = −.5, df = 162, p = 0.57). Inspection of the means for
each singer revealed the following patterns pre- to post workout:
five became more flat in pitch, one became less flat, three moved
from flat to slightly sharp, and one moved from sharp to flat. It
should be noted that target accuracies were not uniform across
notes, although there was a tendency for singers who were flat
Downloaded for Anonymous User (n/a) at Pontifical Catholic University of Chile from ClinicalKey.com by Elsevier on October
07, 2021. For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
Journal of Voice, Vol. 31, No. 4, 2017
to be so across notes. Although one might expect the greatest
difference in the highest note of the scale, often the greatest dif-
ferences occurred earlier in the ascending scale, with the highest
note typically the most accurate.
The vibrato rate for the SSB was calculated based on the most
sustained sixth note (ie, “see”). The mean vibrato rate both pre-
and post workout was 5.7 (preworkout SD = .8, post workout
SD = .5). There was no significant difference (t = 2.1, df = 18,
p = 0.51). Three singers, who demonstrated what would be con-
sidered a slow vibrato rate, increased it to a more appropriate
rate. These values in vibrato cycles per second were 4.6–5.5,
4.1–5.3, and 4.9–6.3, moving to within the optimal vibrato range.19
Vibrato regularity for the sustained sixth note of the SSB was
calculated by determining the SD of the period of each vibrato
cycle. The mean vibrato regularities were .015 (SD = .008)
preworkout and .012 (SD = .009) post workout. There was no
significant difference (t = 2.1, df = 18, p = 0.41). For seven in-
dividuals, the vibrato rate became more regular, four by .01, and
three by .02. Vibrato regularity for seven individuals was un-
changed, whereas four singers demonstrated increased irregularity
by .01 and one increased by .02.
The vibrato extent for the SSB was calculated by determin-
ing the lowest and highest frequencies in a cycle of vibrato, and
converting this value to semitones. The mean vibrato extent
preworkout was 2.3 (SD = .9) and post workout was 2.4
(SD = 1.0). The vibrato extent is greater than ideal (roughly one
semitone around the target pitch18) and may reflect overdriving
in response to singing while wearing a mask over the nose and
mouth. There was no significant difference in the pre- to
postworkout values (t = .69, df = 12, p = 0.5).
The vibrato rate for the scale, averaged across notes, was 5.6
for both pre- and post workout. There was no significant dif-
ference (t = .2, df = 9, p = 0.8). Vibrato regularities, averaged across
notes, were .04 preworkout (SD = .1) and .001 post workout
(SD = .004). There was no significant difference in the pre- to
postworkout values (t = .99, df = 9, p = 0.4).
Vibrato extents for the scale, averaged across notes, were 1.3
preworkout (SD = .6) and 1.6 post workout (SD = .4). This dif-
ference approached significance (t = 2.5, df = 9, p = 0.04). It should
be noted that the difference was due to some extent to three
singers producing straight tones in the preworkout condition. Two
singers decreased vibrato extent in the post condition, but did
so only slightly.
The SPR was calculated only for the scale because of the pres-
ence of the mask during the SSB condition. The pre-SPR was
19.4 (SD = 6.9), whereas the post-SPR was 17.0 (SD = 7.6). This
difference was not significant (t = 2.2; df = 9; p = 0.06).
Because there were virtually no significant acoustic differ-
ences before and after the workout, and the majority of singers
exercised frequently, there was no attempt to relate physical fitness
to vocal performance.
DISCUSSION
The present study was designed to expand previous work3 that
demonstrated positive vocal and respiratory changes with an
aerobic workout. In the earlier work, singers demonstrated a sig-
nificantly increased sound pressure level, and as an important
Monica A. McHenry and Joseph Evans Impact of Aerobic Exercise as a Warm-Up for Singing
concomitant, increased airflow during voicing. It appears, however,
that improvements with exercise are limited to aerodynamic
changes. In the present work, singers overall showed virtually
no improvement in target pitch accuracy, vibrato rate, extent, and
regularity, and the SPR.
Some individual changes are of interest nonetheless. Our find-
ings for three individuals demonstrated a normalization of the
vibrato rate from slow to within normal limits, comparable to
changes noted by Moorecroft and Kenny.20 It is possible that the
aerobic workout increased activation across their physiological
systems, an effect that has also been achieved by vocalization
alone.5
In an earlier work,21 the effects of a vocal warm-up in isola-
tion were compared with a vocal warm-up combined with a
5-minute workout in an aerobic heart rate range. Participants were
nonsingers, however, and the vocal warm-up was one designed
to prepare actors for the stage. Thus, measures critical for singers
such as pitch accuracy and features of vibrato were not ob-
tained. Nonetheless, it was found that a combined approach
reduced phonation threshold pressure and decreased noise in the
voice (ie, jitter and noise-to-harmonics ratio). Thus, a warm-
up strategy that combines vocalization with an aerobic workout
remains appealing.
CONCLUSIONS
It appears that an aerobic workout positively impacts the respi-
ratory driving force for voice production. The singer has an easier
time generating increased airflow during voicing, resulting in a
healthy increase in sound pressure level. This is especially crucial
for classical opera and concert singers, who are required to
perform without amplification or sound enhancement in large
theaters with full orchestras.
An aerobic workout has been shown to reduce phonation
threshold pressure, but this is a relatively isolated physiologi-
cal change in the vocal fold tissues. Although there are clear
aerodynamic benefits, data from the current work demonstrate
that a purely aerobic workout is not adequate to prepare the voice
for classical performance. The vocal loads sustained during suc-
cessful projection with orchestral accompaniment necessitate
careful preparation. Exercises designed for vocal fold flexibil-
ity at very full amplitudes and sustained pitches at the extremes
of the range are essential both for immediate bel canto vocal-
ism and for rapid recovery and continued vocal health after such
performances. Critical for performance is the fine tuning and bal-
ancing across the respiratory, laryngeal, and resonance systems.
It appears that these can only be achieved with vocalization
Downloaded for Anonymous User (n/a) at Pontifical Catholic University of Chile from ClinicalKey.com by Elsevier on October
07, 2021. For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
441
exercises, facilitating coordination within and across the phys-
iological systems12 involved in the complex art of bel canto.
REFERENCES
1. McArdle WD, Katch FL, Katch VL. Exercise Physiology. 5th ed.
Philadelphia, PA: Lippincott Williams & Wilkins; 2001:575.
2. Elliot N, Sundberg J, Gramming P. What happens during vocal warm-up?
J Voice. 1995;9:37–44.
3. Sanders A, Keiner M, Schlumberger A, et al. Effects of functional exercises
in the warm-up on sprint performances. J Strength Cond Res. 2013;27:995–
1001.
4. Olsen O, Sjøhaug M, van Beekvelt M, et al. The effect of warm-up and
cool-down exercise on delayed onset muscle soreness in the quadriceps
muscle: a randomized controlled trial. J Hum Kinet. 2012;35:59–68.
5. Spitz MG, Wenefick RW, Mitchell JB. The effects of elapsed time after
warm-up on subsequent exercise performance in a cold environment.
J Strength Cond Res. 2014;28:1351–1357.
6. DeFatta RA, Sataloff RT. The valuye of vocal warm-up and cool-down
exercises. Questions and controversies. J Singing. 2012;69:173–175.
7. Sandage MJ, Connor NP, Pascoe D. Voice function differences following
resting breathing versus submaximal exercise. J Voice. 2013;27:572–
578.
8. McHenry M, Evans J. Aerobic exercise as a warm-up for singing:
aerodynamic changes. J Voice. 2015;doi:10.1016/j.jvoice.2015.08.011. [Epub
ahead of print]. Retrieved 4/1/16.
9. Gottliebson RO. Efficacy of cool-down exercises in the practice regimen
of elite singers [doctoral dissertation]. Cincinnati, OH: University of
Cincinnati; 2011.
10. McHenry MA, Evans J, Powitzky E. Vocal assessment before, after, and
the day after opera performance. J Voice. 2015;30:186–191.
11. Gish A, Kuncuk M, Sims L, et al. Vocal warm-up practices and perceptions
in vocalists: a pilot survey. J Voice. 2012;26:e1–e10. doi:10.1016/
j.voice.2010.10.005. Retrieved 7/1/2014.
12. Titze IR. Voice research: the five best vocal warm-up exercises. J Singing.
2001;57:51–52.
13. Tommasini A. Bel canto: audiences love it, but what is it? New York Times,
Nov 28, 2008:AR22.
14. Lani J. Bel canto and the art of singing. Opera J. 2003;36:3–36.
15. Titze I. On flow phonation and airflow management. J Singing. 2015;72:57–
58.
16. Sieminski DJ, Cowell LL, Montgomery PS, et al. Physical activity monitoring
in patients with peripheral arterial occlusive disease. J Cardiopulm Rehabil.
1997;17:43–47.
17. Boersma P, Weenink D. Praat: Doing phonetics by computer (Version 5.3.35).
Amsterdam, ND: U Amsterdam. Available at: http://www.fon.hum.uva.nl/
praat. Accessed December 8, 2012.
18. Omori K, Ashutosh K, Carroll L, et al. Singing power ratio: quantitative
evaluation of singing voice quality. J Voice. 1996;10:228–235.
19. Isherwood N. Vocal vibrato: new directions. J Singing. 2009;65:71–283.
20. Moorecroft L, Kenny DT. Vocal warm-up produced acoustic change in
singers’ vibrato rate. J Voice. 2012;26:667, e13-8. doi:10.1016/j.voice
.2011.10.007. Retrieved 7/5/2014.
21. McHenry M, Johnson J, Foshea B. The effect of specific versus combined
warm-up strategies on the voice. J Voice. 2009;23:572–576.