Vocal Performance of Group Fitness Instructors

Before and After Instruction: Changes in Acoustic

Measures and Self-Ratings
Katherine Dallaston and Anna F. Rumbach, Brisbane, Australia

Summary: Objectives. (1) To quantify acute changes in acoustic parameters of the voices of group fitness instruc-
tors (GFIs) before and after exercise instruction. (2) To determine whether these changes are discernible perceptually by
the instructor.
Study Design. This is a pilot prospective cohort study.
Methods. Participants were six female GFIs, based in Brisbane, Australia. Participants performed a series of vocal
tasks before and after instruction of a 60-minute exercise class. Data were obtained pertaining to fundamental frequency
(pitch), intensity (volume), jitter, shimmer, harmonic-to-noise ratio (HNR), maximum duration of sustained phonation
(MDSP), and pitch range. Additionally, self-ratings of voice quality were obtained before and after instruction. Data
were analyzed using the Wilcoxon signed rank test.
Results. Significant increases (P
0.05) were found in fundamental frequency and intensity after instruction. No sig-
nificant changes in jitter, shimmer, HNR, or MDSP were found before and after instruction. For the group, no significant
change in self-ratings of voice quality occurred before and after instruction.
Conclusions. Statistically significant changes in pitch and volume were found on acoustic analysis. However, these
subtle changes remained within the limits of what is considered normal and representative of the participants age and
gender. Further research into the effects of exercise instruction on the voice is needed.
Key Words: Professional voice useAerobics instructorGroup fitness instructorVocal loadingVoice disorder
Acoustics.

INTRODUCTION and an equal number with complaints of intermittent hoarse-

Group fitness instructors (GFIs) use the voice to engage, moti-
vate, and cue participants safely through an exercise routine.
However, voice problems such as hoarseness, pitch changes,
and vocal tract dryness are commonly reported by GFIs.14
One recent study reported over 78% of GFIs experience acute
voice symptoms during or immediately after instructing.4
This may not be surprising considering the GFIs need to
perform vigorous exercise and speak simultaneously. Addition-
ally, the need to compete vocally with loud background music
and environmental noise presents a combination of environ-
mental and physiological stressors that may contribute to the
development of voice problems. Despite a number of studies
exploring GFIs self-perception of voice difficulties,14 there
has been limited objective exploration of occupational voice
changes in this population.
To date, only one study has reported objective data relating to
acute voice changes during and immediately after instruction.
The study, conducted by Wolfe et al,5 assessed the vocal func-
tion of six female GFIs using measures of acoustic perturbation
and electroglottography (EGG) before and after a 30-minute
period of simulated exercise instruction. The subjects were
randomly selected from the results of a previous study2 to
include three GFIs with few or no reported voice problems
Accepted for publication February 12, 2015.
From the School of Health and Rehabilitation Sciences, The University of Queensland,
Brisbane, Australia.
Address correspondence and reprint requests to Anna F. Rumbach, School of Health and
Rehabilitation Sciences, The University of Queensland, St Lucia, Brisbane 4072, Australia.
E-mail: a.rumbach@uq.edu.au
Journal of Voice, Vol. 30, No. 1, pp. 127.e1-127.e8
0892-1997/$36.00
2016 The Voice Foundation
http://dx.doi.org/10.1016/j.jvoice.2015.02.007
ness, complete or partial voice loss, vocal fatigue, and strain.
The study failed to elucidate any significant difference in acous-
tic parameters or EGG before and after instruction. However,
significant differences were found in vocal behavior and func-
tion between the two groups. During the simulation, instructors
who reported voice problems used a louder voice and phonated
for a higher percentage of time, compared with instructors
without voice problems. This is consistent with the hypothesis
that vocal loading (ie, prolonged phonation at a higher than
usual effort level) contributes to a GFIs risk of voice problems.
Significantly, more jitter, lower harmonic-to-noise ratio (HNR),
and less periodicity of vocal fold vibration were found in in-
structors with vocal problems. This finding suggests that self-
report of vocal problems in GFIs will likely correlate with
objective acoustic and instrumental measures, and that further
investigation using these objective measures is appropriate.
The generalizability of Wolfe et als5 results is limited by the
study design, which used a simulated 30-minute period of exer-
cise instruction, rather than a real exercise class environment.
Subjects were asked to perform as they would in a typical ses-
sion, although were confined to a 4-ft area, asked to always face
a freestanding microphone, and were without an audience or
authentic background noise. Although one of the authors, a
certified GFI, confirmed that each participant carried out the in-
struction in a realistic fashion, the degree to which the instruc-
tion was typical of each individuals usual performance was not
reported. To obtain results that can be reliably extrapolated to
the profession at large, objective measurements of vocal func-
tion of GFIs are required using a study scenario of an authentic
group fitness class, rather than a simulation of one. Further-
more, developing methodology that can feasibly be used to
127.e2
elucidate changes in vocal performance across a large cohort,
whose gender distribution reflects the profession, is important.
Therefore, the present study acted to pilot methodology that
was able to (1) quantify any acute changes in acoustic parame-
ters in the voices of GFIs before and after a period of authentic
exercise instruction and (2) to determine whether these changes
are discernible perceptually by the GFIs. In view of the high
vocal demands of the profession and the high prevalence of
self-reported acute vocal symptomatology, it was hypothesized
that some degree of measurable vocal change at the acoustic
level would be detected. Furthermore, understanding whether
GFIs could perceptually discriminate acute voice changes
was considered important as it has been hypothesized that
self-awareness of subtle voice change may reflect their capacity
to self-manage ongoing vocal health.
METHOD
Procedure
A total of six female GFIs aged between 21 and 46 years (me-
dian 40 years) participated in this study. All participants were
certified and registered GFIs and were currently earning income
from teaching at least one type of group fitness class per week at
a university-based fitness center located within Brisbane,
Australia. Recruitment of participants was conducted in coop-
eration with the fitness centers manager, who distributed invi-
tations to participate to all staff via e-mail and/or verbal
discussion. Recruitment was conducted over a 1-month period
(August 2014) that aligned with the commencement of the uni-
versity semestera time that typically offers consistency in re-
gards to size and number of classes taught by GFIs each week.
On recruitment, participants provided informed consent and
were assigned a de-identified participant number to ensure
participant confidentiality.
Each participant underwent a voice evaluation before and af-
ter instruction of a 60-minute group exercise class that was part
of their regular teaching schedule at this facility. For this reason,
the study was not able to control for the time of the scheduled
class. All classes were conducted in a purpose-built, air-condi-
tioned, group fitness studio that was able to accommodate up to
50 exercise class participants. All GFIs had access to a surround
sound stereo system and a headset microphone. However,
microphone use was not mandatory in this facility. As the aim
of the study was to examine GFIs voice use in an authentic
setting, the researchers did not control variables surrounding
vocal hygiene and teaching practices that would alter their usual
teaching behaviors (eg, classes taught per day or week, level of
hydration before and during exercise instruction). Vocal use
before the voice assessments was reported to be consistent
with normal daily vocal use for each individual.
Ethical approval to conduct this research was obtained from
the University of Queenslands Behavioural and Social Sci-
ences Ethics Committee.
Questionnaires. Before voice assessment, participants were
asked to complete two questionnaires: (1) a previously pub-
lished questionnaire that gathered information regarding GFI
demographics, lifestyle, teaching practices, and experience of
Journal of Voice, Vol. 30, No. 1, 2016
voice symptomatology4 and (2) the Voice Handicap Index
(VHI)6 to determine the impact of voice difficulties on the par-
ticipants daily functioning and quality of life.
Voice self-rating. Participants were asked to rate their voice
quality before and after instruction using a scale of 1 to 10, with
1 representing worst or most abnormal voice and 10 represent-
ing best or most normal voice. After instruction, participants
were also asked to comment on any voice changes that occurred
during the period of instruction and, if any were experienced, to
elaborate on exactly which sensory and/or perceptual changes
they perceived.
Voice assessment. Each participants vocal function was
assessed twice, once before and once after a 60-minute period
of exercise instruction by a qualified speech-language patholo-
gist. Recordings were obtained within 30 minutes before and af-
ter instruction and not before the participants respiratory rate,
as perceived by the GFI, had returned to resting rate. Assess-
ments were conducted face-to-face in a small room with low
ambient noise (<50 dB), as measured by a sound pressure level
meter,7 and each session took approximately 20 minutes to
complete.
Participants were seated for the duration of the assessment
and asked to perform a series of vocal tasks. Audio recordings
were captured using a head-mounted condenser microphone
(AKG C544L), positioned 5 cm from the lips to the side of
the mouth, to minimize air burst noise.8 A Micro Mic Phantom
Adapter (AKG MPA VL) was used to connect the microphone
to a Korg MR-1000 Professional 1-Bit Digital Audio Recorder,
which digitized the microphone signals at a sampling rate of
48 000 Hz. To obtain accurate measures of vocal intensity, cali-
bration of the microphone was conducted before data collection
using a comparison method.9
Vocal tasks included maximum duration of sustained phona-
tion (MDSP), maximal pitch range, oral passage reading, and
conversational speech. The order in which vocal tasks were per-
formed and the use of standardized verbal instructions (and
demonstrations when applicable) were consistent across both
data collection time points for all participants.
MDSP on the vowel /a/ was performed under two conditions:
comfortable (ie, modal) intensity and minimal intensity (MI).
Participants were required to perform five trials of the task in
each condition to optimize the reliability of the data.10 To deter-
mine maximal pitch range, participants were instructed to glide
or step from one end of their pitch range to the other, following a
model from the assessor. Participants performed the task in both
ascending and descending directions (sustaining the highest and
the lowest pitch for 5 seconds if able). A total of five trials in
each direction were performed to allow participants opportunity
to demonstrate the true limits of their phonation range.11
As prior research has shown vocal behavior to vary depend-
ing on elicitation method,12 more than one task was used to
obtain recordings of connected speech. First, an oral passage
reading task was conducted using a passage of text containing
a representative phonetic sample (the Grandfather passage13).
Participants were instructed to familiarize themselves with
the text before reading the passage aloud using a comfortable,
Katherine Dallaston and Anna F. Rumbach Vocal Performance of Group Fitness Instructors
conversational voice.8 Second, a 3-minute conversational sam-
ple was obtained by asking participants to Tell me a happy
story or Tell me about a happy experience regarding group
fitness classes. Participants were prompted to continue talking
until a full 3-minute sample had been obtained for analysis.8,14
Data analysis
Analysis of voice recordings was conducted using Praat free-
ware.15 MDSP was computed using the average of five sus-
tained vowels at comfortable intensity levels.10 Maximal
pitch range was computed by calculating the average of the
five highest fundamental frequency (F0) values and the average
of the five lowest F0 values produced during any maximal pitch
range trials. Recordings of continuous phonation (MDSP) were
analyzed using the following measures: F0 (Hz); average inten-
sity (dB); frequency perturbationjitter (%); amplitude pertur-
bationshimmer (dB); and average HNR. To optimize the
reliability of perturbation and HNR measures, recordings
were truncated, leaving 1.5 seconds on each side of the tempo-
ral midpoint for analysis.16 Recordings of connected speech
tasks (oral passage reading and conversational speech) were
analyzed using measures of F0 and average intensity.
Statistical analysis was performed with Stata software
(Version 10.0). Relationships between preinstruction and post-
instruction data were analyzed using the Wilcoxon signed rank
test for analysis. A nonparametric test was used because of the
small number of participants and because not all parameters
were normally distributed. The significance level was set at P

0.05, and P values between 0.05 and 0.07 were considered
a trend toward significance.
Participant demographics
Participants (n 6) had been teaching group fitness classes for
127 years (median 5 years) and were teaching on average
between 2 and 20 classes (median 5 classes) per week, 2
6 days (median 4.75 days) per week. Four of the six
FIGURE 1. Voice Handicap Index (VHI) scores for six female GFIs. Items are grouped into three domains that represent physical (P), functional
(F), and emotional (E) aspects of the impact of voice disorders on quality of life. Total VHI scores are the combined total of each subscale. Total
possible score is 120.
127.e3
participants reported sometimes experiencing voice changes
either during or immediately after instructing; however, no par-
ticipants reported a history of diagnosed voice disorders. They
reported no cigarette smoking, no regular medication use, min-
imal alcohol consumption, and few allergies.
The impact of voice difficulties on daily functioning and
quality of life varied considerably for the group (Figure 1).
VHI scores ranged from 8 to 42 (median 14). For the group,
nearly half (48.2%, n 54) of VHI scores related to the phys-
ical subscale, 38.4% (n 43) related to the functional subscale,
and 13.4% (n 15) related to the emotional subscale. The two
highest scoring statements for the groupThe sound of my
voice varies throughout the day and People have difficulty
understanding me in a noisy roomwere also the only two
statements reported to occur more often than never by all par-
ticipants. The impact of voice on emotional aspects varied, with
half of participants reporting no emotional impact of any kind.
RESULTS
The voices of six GFIs were evaluated before and after instruc-
tion of a 60-minute exercise class. Data pertaining to self-
perception and acoustic parameters (F0, intensity, jitter,
shimmer, HNR, MDSP, and pitch range) of vocal quality are
presented. Details of these analyses are as follows.
Self-perception of vocal quality
There was considerable variation in the group in regards to self-
reported acute voice changes immediately after instruction. Half
of the group (n 3; GFI2, GFI5, and GFI6) reported no acute
voice changes. The remaining participants reported acute onset
of auditory-perceptual changes (GFI1), sensory changes (GFI4),
or a combination of both (GFI3). Perceptual self-ratings before
and after instruction are presented in Figure 2. Preinstruction
self-ratings (with 1 representing worst or most abnormal voice
and 10 representing best or most normal voice) were found to
roughly reflect VHI scores; participants with high VHI scores
127.e4 Journal of Voice, Vol. 30, No. 1, 2016

FIGURE 2. Self-ratings of sensory perceptual voice quality of six GFIs before and after instruction, where 1 worst or most abnormal voice and
10 best or most normal voice.

reported low preinstruction self-ratings, and vice versa. For the
group, self-ratings did not change significantly before and after
instruction (P 0.83). Three participants reported an improve-
ment in voice quality after instruction, one reported no change,
and two reported a decline in voice quality. Those who reported
an improvement in voice quality after instruction described a
reduction in croakiness and the sensation of a warmed-up
voice that was easier to control. Those who reported a decline
in voice quality after instruction described the sensations of a
lump in the throat, slight soreness, tightness, or discomfort.
Acoustic parameters
Fundamental frequency. Measures of F0 across all tasks
were generally within the range expected of normal female
adults17 (Table 1), with the exception of one participant
(GFI3). Significantly higher F0 was found after instruction dur-
ing sustained vowels at comfortable (P
0.001) and minimal (P

0.001) intensity and during oral passage reading (P 0.02).
F0 during conversation for the group also increased after instruc-
tion; however, the difference was not significant (P 0.11).
Intensity. For the group, higher values of intensity were found
after instruction across all tasks (Table 2). Results showed sig-
nificant increase in intensity after instruction during sustained
vowels, both at comfortable intensity (P
0.001) and MI (P

0.004). Although higher after instruction, the increases in in-
tensity values during oral passage reading and conversation
were not significant for the group.
Maximum duration sustained phonation. To allow for
appropriate comparison to previous studies, MDSP measures are
presented for sustained vowels at comfortable volume. Results

TABLE 1. TABLE 2.
Fundamental Frequency (Hz) of Six GFIs Before and After Vocal Intensity (dB) of Six GFIs Before and After
Instruction Instruction
Vocal Vocal
Task Median SD Range z P Task Median SD Range z P
Sustained vowels CI Sustained vowels CI
Pre 208.97 26.32 151.97231.69 3.733 <0.001* Pre 81.67 5.07 72.0389.20 4.720 <0.001*
Post 227.52 24.40 165.21250.71 Post 84.72 3.85 78.2392.37
Sustained vowels MI Sustained vowels MI
Pre 195.47 26.29 148.01231.34 4.535 <0.001* Pre 75.95 2.62 70.9679.41 2.952 <0.004*
Post 228.27 29.64 164.12262.23 Post 78.01 4.69 71.2087.92
Oral passage reading Oral passage reading
Pre 207.90 22.06 159.05209.05 2.201 0.02* Pre 80.68 5.44 71.9687.59 0.314 0.75
Post 214.50 27.64 161.61230.30 Post 80.26 4.74 75.6189.97
Conversation Conversation
Pre 181.30 28.09 157.89221.13 1.572 0.11 Pre 81.11 3.89 76.5186.86 0.524 0.60
Post 190.56 27.18 160.80235.15 Post 80.53 3.64 79.0788.55
Abbreviations: SD, standard deviation; CI, comfortable intensity; MI, min- Abbreviations: SD, standard deviation; CI, comfortable intensity; MI, min-
imal intensity. imal intensity.
*P value significant at
0.05. *P value significant at
0.05.
Katherine Dallaston and Anna F. Rumbach Vocal Performance of Group Fitness Instructors 127.e5

FIGURE 3. MDSP for six GFIs before and after instruction. *Estimated expected MDSP for women aged 1365 years (Baken and Orlikoff17).
MDSP, maximum duration of sustained phonation.

varied considerably between participants (Figure 3), with only one fold edema.20 Edema (or swelling) is the natural reaction of tis-
participant (GFI2) achieving an MDSP within the normal range sue to physical trauma or misuse, which is experienced during
for adult women.17 Preinstruction and postinstruction comparison vocal loading through repetitive, forceful glottal collisions.11
revealed no significant differences for the group (P 0.68). After a period of prolonged or excessive vocalization, some
Perturbation. As phonatory stability is more difficult to studies have observed edema,21,22 whereas others have not.23
maintain at MI,18,19 perturbation results are presented for Currently, there are no videostroboscopic studies of GFIs after
sustained vowels at MI only (Table 3). Preinstruction and post- a period of instruction, although it is likely GFIs are at
instruction comparison showed no significant differences in the
jitter values for the group (P 0.42). A trend toward increased TABLE 3.
shimmer after instruction (P 0.07) was observed; however, Vocal Function of Six GFIs Before and After Instruction
this finding was not statistically significant.
Variable Median SD Range z P
Harmonic-to-noise ratio. Comparison of preinstruction Jitter (%)
and postinstruction HNR values revealed no significant differ- Pre 0.43 0.45 0.162.23 0.792 0.42
ence on this measure (Table 3). Post 0.45 0.37 0.221.86
Pitch range. Maximal pitch ranges differed considerably be- Shimmer (dB)
tween participants (Figure 4). Maximum (highest) pitch did not Pre 0.23 0.09 0.140.57 1.790 0.07
significantly change for the group before and after instruction Post 0.26 0.19 0.140.92
HNR (dB)
(P 0.11); however, participants minimum (lowest) pitch
CI
significantly increased (P 0.02) after instruction. Pre 26.14 2.94 21.0432.20 1.532 0.12
Post 28.46 4.35 16.7932.54
MI
DISCUSSION Pre 24.14 3.96 17.5332.17 0.689 0.49
A statistically significant degree of voice change for funda- Post 24.38 3.40 18.9830.31
mental frequency and intensity measures were found in this MDSP (s)
study. These results support previous research findings, which Pre 13.81 4.80 4.4924.19 0.401 0.68
reported 42.679% of GFIs experienced some degree of voice Post 13.13 5.14 5.1428.15
change during or immediately after exercise instruction.1,2,4 Pitch range (Hz)
However, changes after a single, one-hour period of exercise in- Minimum
struction did not cause changes significant enough to alter Pre 163.05 30.63 90.02172.14 2.201 0.02*
Post 171.62 11.40 153.00182.81
acoustic parameters so that they deviate from age and gender
Maximum
norms. Although this study recruited only a small participant Pre 384.29 127.95 265.78686.13 1.572 0.11
cohort, the findings contribute to the limited evidence base of Post 406.73 187.54 319.07839.61
the effects of authentic exercise instruction on the voice.
Notes: Jitter and shimmer values are presented for vowels sustained at
A major finding of the study was significantly higher F0 after minimal intensity only.
instruction during sustained vowels and oral reading. This Abbreviations: SD, standard deviation; CI, comfortable intensity; MI, min-
imal intensity; MDSP, maximum duration of sustained phonation.
result was unexpected in light of one of the physiological
*P value significant at
0.05.
changes known to occur after a period of vocal loading: vocal
127.e6
considerable risk for vocal fold edema; in addition to vocalizing
to large groups over background noise, the force of glottal col-
lisions during group fitness instruction may be amplified by
suboptimal postures,24 reduced breath support,25 and concur-
rent strenuous exercise.26 Edematous vocal folds are thicker
than nonedematous folds.11 The mass of the vocal folds has
an inverse relationship with the F0 of phonation; as the mass
of the vocal folds increases, the F0 of phonation decreases.11,27
In the present study, a degree of vocal fold edema was expected,
and therefore, so too was a decrease in F0 after instruction. One
possible explanation why this was not found is that a 60-minute
exercise instruction was not of sufficient duration to cause
edema to the degree that it can be detected acoustically. Studies
that have observed vocal fold edema after prolonged or exces-
sive vocalization21,22 have done so following vocal loading
tasks that were longer than 60 minutes in duration using
videostroboscopic measures.
It is also possible that the acoustic signs of edema were
masked by a second physiological change that occurs during
vocal loading: excessive laryngeal tension.20 During vocal
loading, speakers may develop hyperfunctional behavior in
the form of excessive tension of the laryngeal muscles.20,28
Excessive tension of the cricothyroid muscles results in
higher than normal longitudinal tension of the vocal folds.29
The relationship between tension of the vocal folds and F0 of
phonation can be likened to the tension of a string and the pitch
at which it vibrates; as tension increases, so too does F0.11
Wolfe et al5 observed a significant increase in F0 during a 30-
minute period of simulated instruction and suggested that this
finding may be an acoustic indication of increased laryngeal
tension. Similarly, in the present study, the increase in F0 after
instruction may be indicative of increased laryngeal tension that
continued after exercise instruction had ceased. This was
further evidenced acoustically by the finding that during postin-
struction pitch glides, no participant was able to match their
preinstruction minimal (lowest) pitch. As participants were
asked to demonstrate their full pitch range, this finding suggests
that participants were unable to voluntarily relax the
FIGURE 4. Minimal (lowest) and maximal (highest) pitch of six GFIs before and after instruction. Min, minimum; Max, maximum.
Journal of Voice, Vol. 30, No. 1, 2016
longitudinal tension of the vocal folds to the degree that was
achieved before instruction.
Another major finding of the study was significantly higher
vocal intensity after instruction for sustained vowels across
both comfortable intensity and MI conditions. During the MI
condition, participants were instructed to produce the vowel
using your quietest voice, just above a whisper, so it can be
assumed that attempts were representative of true ability to
phonate as quietly as possible. The reason for increased inten-
sity of MI phonation after instruction may be related to the
development of increased laryngeal tension secondary to vocal
loading, as discussed previously. The tension of the vocal folds
creates resistance to vibration.11 Consequently, in conditions of
increased vocal fold tension, higher subglottal air pressure
levels are needed to initiate vocal fold vibration and higher in-
tensity levels are produced as a result.30 Importantly, however,
the minimum subglottal air pressure needed to initiate vocal
fold vibration is known to be dependent on frequency.30 The
fact that participants produced postinstruction sustained vowels
at higher F0 than at preinstruction therefore confounds the find-
ings. Future studies would therefore benefit from controlling for
F0 of sustained vowels.
MDSP is conventionally thought to measure laryngeal effi-
ciency and respiratory support11; most participants MDSP fell
below the expected range for adult women.17 This was surprising
as it was expected that GFIs would have relatively good cardio-
vascular and respiratory endurance because of their regular prac-
tice of phonating while simultaneously exercising. However,
research has shown there are a number of additional influencing
variables, including age, sex, number of trials, and elicitation
technique10 that affect performance on this task. The present
study design does not allow inference on which of these factors
contributed toward the findings of below-average MDSP for the
group. The only discernible finding was that no significant
change in MDSP before and after instruction could be found.
Changes in laryngeal efficiency were further explored using
measures of perturbation (jitter/shimmer7). No significant
changes were found in acoustic perturbation measures before
Katherine Dallaston and Anna F. Rumbach Vocal Performance of Group Fitness Instructors
and after instruction. Similarly, Wolfe et al5 found no signifi-
cant changes in perturbation measures of six GFIs before and
after a 30-minute exercise instruction simulation. This finding
is contrary to the expectation of vocal deterioration after vocal
loading.20 Prolonged or excessive vocalization is thought to in-
crease the risk of tissue damage that impairs the efficiency of
phonation.11 Changes to perturbation levels reflect minute
changes in phonatory stability,7 caused by changes in vocal
fold mass, tension, and biomechanics.11 Experimentally, vocal
loading has been shown to cause significant increases in edema,
edge irregularity, and erythema.21 However, identification of
changes in phonatory stability using acoustic measures has
been inconsistent.21,22,31 In the present study, the higher
frequency and increased intensity of postinstruction vowels
may be a confounding factor, as there is evidence that F0 and
intensity of phonation affect measures of phonatory
stability.18,32 Future studies would also therefore benefit from
controlling for intensity of sustained vowels.
Similarly, there were no changes found in HNR before and
after instruction. A possible reason for no significant increase
in HNR in the present study may again be that a single 60-
minute period of exercise instruction is not of sufficient dura-
tion to cause change in the adequacy of glottal closure. After
a period of vocal loading, an increase in glottal air leakage
may be predicted because of suboptimal closure of fatigued
folds.28 There is no universally accepted mathematical defini-
tion of HNR, and so it is unclear whether the current partici-
pants present with HNR within normal limits.33
There was no apparent trend in the groups self-perceptions
of vocal quality before and after instruction; after instruction,
three participants perceived an improvement in voice quality,
one perceived no change, and two perceived a decline. Has-
kell34 described a conceptual model for vocal self-perception
in which a speakers perception of voice production is informed
by both auditory feedback (how the voice sounds) and tactile-
proprioceptive feedback (how the voice feels). Interestingly,
when asked to comment on self-ratings, participants who
perceived an improvement in voice quality after instruction
offered comments relating to both auditory and tactile-
proprioceptive aspects of voice production; however, those
who perceived a decline in voice quality commented only on
tactile-proprioceptive aspects. This finding may indicate that
changes in laryngeal sensation are more salient to GFIs than
changes in sound quality. If this is the case, efforts to increase
vocal health awareness in this population would be wise to
exploit this saliency by targeting education toward the implica-
tions of laryngeal discomfort. Future studies may benefit from
using more than one self-rating scale to investigate GFIs vocal
self-perception to delineate the auditory aspects and tactile-
proprioceptive aspects of self-perception.
By comparing perceptual self-ratings with acoustic mea-
sures, the study aimed to determine whether acoustic changes
were discernible by GFIs. To date, studies that have compared
self-ratings and acoustic measures have not been successful in
demonstrating a strong correlation between these parame-
ters.3537 In the present study, post hoc analysis of the
acoustic findings showed their comparison with the self-
127.e7
rating scale was flawed. The design of the self-rating scale
was such that it asked participants to judge their voice in rela-
tion to quality on a scale from the worst to the best voice. How-
ever, the only two acoustic parameters for which there were
significant changes were F0 and intensity, which correlate
with the perceptual qualities of pitch and volume, respec-
tively.11 Importantly, an increase in pitch or volume does not
necessarily constitute a better or worse voice. The issue
of whether acoustic changes were discernible by the instructors
was therefore not something that this study was able to deter-
mine. A design consideration for future studies is the potential
use of multiple self-ratings scales, such as volume, pitch, effort,
and quality of voice.
Limitations and future directions
This is the first study to objectively quantify changes in the voi-
ces of GFIs before and after a period of authentic instruction.
The results provide pilot data toward the consideration of a
larger, more controlled study. Given the pilot nature of the
study, a number of limitations around the design of the study
are evident. First, the studys small sample size with strong fe-
male bias limits the generalizability of the results. The sample
was also heterogeneous in regards to self-report of voice prob-
lems, length of time in the industry, teaching frequency, type of
exercise class taught, vocal behavior, and vocal hygiene. Con-
trolling for these factors in future studies may provide valuable
information about their influence. The objective of this study
was to gain an authentic view of change in before and after
vocal performance after a 1-hour period of exercise. Although
control of extraneous variables is scientifically desirable, this
would interfere with the authenticity of the participants
behavior, shifting the study to examine simulated teaching ex-
periences. Future studies may seek to incorporate rigorous
monitoring of vocal behavior and vocal hygiene variables to
further inform their relationship with acoustic findings. Addi-
tionally, controlling for level of vocal training may be informa-
tive.22,32 Another consideration is the use of physiological
monitoring before and after exercise instruction to ensure
that all postassessments are carried out only once the
participants heart and respiratory rates had returned to
resting rate.
Future studies that use laryngoscopic examination to com-
plement acoustic measures are required. Comparison of the
appearance of the larynx before and after a period of exercise
instruction may provide a rationale for any changes detected
in acoustic and perceptual parameters. However, exploration
of acute voice change in an authentic exercise setting requires
a quick transition from conclusion of the period of exercise in-
struction to the completion of the postexercise voice assess-
ment. Failure to transport the necessary equipment to the
fitness center may allow a period of voice recovery which could
alter results. Therefore, use of laryngoscopy may be better
suited for experimental designs that incorporate simulated
group fitness exercise tasks.
A better understanding of acute voice changes in GFIs
would be gained by supplementing preinstruction and post-
instruction data with data relating to vocal function during
127.e8
instruction. Wolfe et al5 found significant differences in the
vocal behavior during instruction of GFIs with reported
vocal problems compared with those without. Vocal
behavior during instruction is also likely dependent on per-
sonality, individual teaching style, and type of exercise pro-
gram taught. It could be surmised, for example, that
instructors of high intensity aerobic classes produce
different vocal qualities to instructors of yoga classes. To
best understand the acute voice changes after a period of in-
struction, more information about voice usage during in-
struction is necessary. The use of ambulatory phonation
monitoring may be one way to achieve this.
The issue of compromised vocal health among GFIs is clear;
however, a lack of research in the field prevents the implemen-
tation of effective prevention, intervention, and management
strategies. Further research is needed to validate acute changes
in acoustic parameters in the voices of GFIs before and after in-
struction. Furthermore, longitudinal studies are needed to track
vocal changes in GFIs over time.
Acknowledgments
The authors wish to thank Stacie Park for her assistance in data
collection and also thank all the GFIs who donated their time to
participate in this research.
REFERENCES
1. Heidel SE, Torgerson JK. Vocal problems among aerobic instructors and
aerobic participants. J Commun Disord. 1993;26:179191.
2. Long J, Williford HN, Scharff Olson M, Wolfe V. Voice problems and risk
factors among aerobics instructors. J Voice. 1998;12:197207.
3. Newman C, Kersner M. Voice problems of aerobics instructors: implica-
tions for preventative training. Logoped Phoniatr Vocol. 1998;23:177180.
4. Rumbach AF. Vocal problems of group fitness instructors: prevalence of
self-reported sensory and auditory-perceptual voice symptoms and the
need for preventative education and training. J Voice. 2013;27:524.
e11524.e21.
5. Wolfe V, Long J, Youngblood HC, Williford H, Olson MS. Vocal parameters of
aerobic instructors with and without voice problems. J Voice. 2002;16:5260.
6. Jacobson BH, Johnson A, Grywalski C, Silbergleit A, Jacobson G,
Benninger MS, Newman CW. The voice handicap index (VHI): develop-
ment and validation. Am J Speech Lang Pathol. 1997;6:66.
7. Titze IR. Workshop on Acoustic Voice Analysis: Summary Statement. Salt
Lake City, UT: National Center for Voice and Speech; 1995.
8. Ma E, Yiu EM-L. Handbook of Voice Assessments. San Diego: Plural Pub;
2011.
9. Boersma P, Weenink D. PRAAT Manual: Sound Pressure Calibration 2004.
10. Speyer R, Bogaardt HCA, Passos VL, et al. Maximum phonation time: vari-
ability and reliability. J Voice. 2010;24:281284.
11. Colton RH, Casper JK, Leonard R. Understanding Voice Problems: A Phys-
iological Perspective for Diagnosis and Treatment. Philadelphia: Wolters
Kluwer Health/Lippincott Williams & Wilkins; 2011.
12. Zraick RI, Skaggs SD, Montague JC. The effect of task on determination of
habitual pitch. J Voice. 2000;14:484489.
Journal of Voice, Vol. 30, No. 1, 2016
13. Darly FL, Aronson AE, Brown JR. Motor Speech Disorders. Philadelphia:
W.B. Saunders; 1975.
14. Zraick RI, Birdwell KY, Smith-Olinde L. The effect of speaking sample
duration on determination of habitual pitch. J Voice. 2005;19:197201.
15. Boersma P, Weenink D. Praat: Doing Phonetics by Computer (Version 5.3.
80). 2014.
16. Vogel AP, Maruff P. Comparison of voice acquisition methodologies in
speech research. Behav Res Methods. 2008;40:982987.
17. Baken RJ, Orlikoff RF. Clinical Measurement of Speech and Voice. San
Diego: Singular Thomson Learning; 2000.
18. Gelfer MP. Fundamental frequency, intensity, and vowel selection: effects
on measures of phonatory stability. J Speech Hear Res. 1995;38:1189.
19. Orlikoff RF, Kahane JC. Influence of mean sound pressure level on jitter
and shimmer measures. J Voice. 1991;5:113119.
20. Solomon NP. Vocal fatigue and its relation to vocal hyperfunction. Int J
Speech Lang Pathol. 2008;10:254266.
21. Mann EA, McClean MD, Gurevich-Uvena J, Barkmeier J, McKenzie-
Garner P, Paffrath J, Patow C. The effects of excessive vocalization on
acoustic and videostroboscopic measures of vocal fold condition. J Voice.
1999;13:294302.
22. Scherer RC, Titze IR, Raphael BN, Wood RP, Ramig LA, Blager RF. Vocal
fatigue in a trained and an untrained voice user. In: Baer T, Sasaki CT,
Harris KS, eds. Laryngeal Function in Phonation and Respiration. San
Diego: Singular Publishing; 1991:533555.
23. Stemple JC, Stanley J, Lee L. Objective measures of voice production in
normal subjects following prolonged voice use. J Voice. 1995;9:127133.
24. Rubin JS, Mathieson L, Blake E. Posture and voice. J Singing. 2004;60:
271275.
25. Doust JH, Patrick JM. The limitation of exercise ventilation during speech.
Respir Physiol. 1981;46:137147.
26. Naito A, Niimi S. The larynx during exercise. Laryngoscope. 2000;110:
11471150.
27. Mathieson L. Greene and Mathiesons The Voice and its Disorders. Lon-
don, Philadelphia: Whurr Publishers; 2001.
28. Welham NV, Maclagan MA. Vocal fatigue: current knowledge and future
directions. J Voice. 2003;17:2130.
29. Pretterklieber ML. Functional anatomy of the human intrinsic laryngeal
muscles. Eur Surg. 2003;35:250258.
30. Titze IR. Phonation threshold pressure: a missing link in glottal aerody-
namics. J Acoust Soc Am. 1992;91:29262935.
31. Gelfer MP, Andrews ML, Schmidt CP. Effects of prolonged loud reading on
selected measures of vocal function in trained and untrained singers. J
Voice. 1991;5:158167.
32. Orlikoff RF. Vocal jitter at different fundamental frequencies: a
cardiovascular-neuromuscular explanation. J Voice. 1989;3:104112.
33. Oller LL. Analysis of Voice Signals for the Harmonics-to-Noise Crossover
Frequency. Masters thesis. Stockholm, Sweden: KTH Royal Institute of
Technology; 2008.
34. Haskell JA. Vocal self-perception: the other side of the equation. J Voice.
1987;1:172179.
35. Buekers R. Are voice endurance tests able to assess vocal fatigue. Clin Oto-
laryngol Allied Sci. 1998;23:533538.
36. Lehto L, Laaksonen L, Vilkman E, Alku P. Changes in objective acoustic
measurements and subjective voice complaints in call center customer-
service advisors during one working day. J Voice. 2008;22:164177.
37. Rantala L, Vilkman E. Relationship between subjective voice complaints
and acoustic parameters in female teachers voices. J Voice. 1999;13:
484495.