Laryngoscopic, acoustic, perceptual, and functional assessment of voice in rock singers

Original Paper
Folia Phoniatr Logop 2013;65:248–256
DOI: 10.1159/000357707

Published online: March 18, 2014

Laryngoscopic, Acoustic, Perceptual,
and Functional Assessment of Voice in
Rock Singers
Marco Guzman a Macarena Barros c Fernanda Espinoza c Alejandro Herrera c
Daniela Parra c Daniel Muñoz b Adam Lloyd d
a

School of Communication Sciences and b Faculty of Medicine, University of Chile, and c School of Communication
Sciences, Andres Bello National University, Santiago, Chile; d Voice Care Center, Ear, Nose, Throat, and Plastic Surgery
Associates, Orlando, Fla., USA

Key Words
Rock singing · Growl voice · Falsetto · Singing voice ·
Singing Voice Handicap Index · Laryngoscopy ·
Acoustic analysis · Hyperfunction

Abstract
Objective: The present study aimed to vocally assess a group
of rock singers who use growl voice and reinforced falsetto.
Method: A group of 21 rock singers and a control group of
18 pop singers were included. Singing and speaking voice
was assessed through acoustic, perceptual, functional and
laryngoscopic analysis. Results: No significant differences
were observed between groups in most of the analyses.
Acoustic and perceptual analysis of the experimental group

demonstrated normality of speaking voice. Endoscopic evaluation showed that most rock singers presented during
singing voice a high vertical laryngeal position, pharyngeal
compression and laryngeal supraglottic compression. Supraglottic activity during speaking voice tasks was also observed. However, overall vocal fold integrity was demonstrated in most of the participants. Slightly abnormal observations were demonstrated in few of them. Singing voice
handicap index revealed that the most affected variable was
the physical sphere, followed by the social and emotional
spheres. Conclusions: Although growl voice and reinforced

© 2014 S. Karger AG, Basel
1021–7762/14/0655–0248$39.50/0
E-Mail
www.karger.com/fpl

falsetto represent laryngeal and pharyngeal hyperfunctional
activity, they did not seem to contribute to the presence of
any major vocal fold disorder in our subjects. Nevertheless,
we cannot rule out the possibility that more evident vocal
fold disorders could be found in singers who use these techniques more often and during a longer period of time.
© 2014 S. Karger AG, Basel

Introduction

Voice quality in singing is affected by both laryngeal
and vocal tract configuration. The sound produced by the
larynx is in turn determined by the vibratory patterns of
the vocal folds. These patterns vary depending on vocal
registers (vocal fry, modal, falsetto, whistle), mode of
phonation (pressed, normal, flow, breathy and whisper)
as well as singing style (belting, opera, rock, etc.) [1].
Modifications of vocal tract configuration are important not only to produce phonetic-articulatory features,
but also because it is an important way to shape the vocal

quality. Particularly, the lower vocal tract structures (epilaryngeal tube, pyriform sinuses and hypopharynx) seem
to play a crucial role in the quality of the singing voice.
Earlier studies have reported that the lower part of the vocal tract changes depending on the style of singing. In clasMarco Guzman
University of Chile, School of Communication Sciences
Av. Independencia 1027
Santiago (Chile)
E-Mail guzmanvoz @ gmail.com


sical singing, wide pyriform sinuses and a wide hypopharynx, as well as a narrow epilaryngeal tube have been observed [2, 3]. According to Sundberg [4, 5], the area ratio
between these two last structures would be the main explanation for the singer’s formant cluster. On the other hand,
during belting (singing technique used in pop styles, musical theater and others) a narrow hypopharynx, pyriform
sinuses and epilaryngeal tube have been observed [4, 5].
Moreover, in different singing styles it is common to
find vocal sounds that would probably be categorized as
pathological vocal qualities if they were produced by nonsingers during conversational voice usage. However,
these sounds are typically used as expressive and stylistic
vocal resources by some contemporary commercial music singers. One of these vocal resources is the so-called
growl voice, which is commonly used in jazz, blues, pop,
gospel and rock among others [6, 7]. Moreover, similar
productions have been found in ethnic singing from
Brazil, Japan and South Africa [8]. When this vocal resource is used in rock, it is also called death growl, death
metal vocals, guttural vocals, death grunts, unclean vocals, and harsh vocals [9].
Perceptually, the growl voice is similar to other types
of vocal qualities such as roughness and hoarseness,
which are considered perceptual signs of voice disorders.
Nevertheless, the growl voice is always a vocal effect or
expressive emphasis and is typically not a permanent way
to produce voice [10]. Even though some singers use
growl voice extensively during a song, it is usually not

found in normal speaking voice phonation [11].
According to Sakakibara et al. [12] growl voice is produced through the simultaneous vibration of the vocal
folds and laryngeal supraglottic structures. The vocal
folds vibrate periodically and the aryepiglottic folds generate subharmonics. These subharmonics have also been
found in other studies [13–15]. A radiologic study showed
that during growl voice production the larynx rose to the
fourth vertebra and there was a large anterior-posterior
(A-P) laryngeal constriction [12]. In addition, some subjects demonstrated vibration of both right and left aryepiglottic folds in phase, whereas in other singers the phase
was slightly different and in some instances aryepiglottic
vibration was completely aperiodic and unstable.
An important group of performers that commonly use
growl voice are rock singers, specifically in some rock
subgenres such as screamo, thrash metal, nu metal, black
metal, heavy metal, death metal, and hardcore punk [16].
Rock singers not only use growl voice as part of their
vocal technique, but also many other vocal resources such
as the reinforced falsetto. The naive falsetto (falsetto with-

out training) is characterized by a decreased degree of
glottal adduction, a reduction in amplitude of the electroglottographic signal which may come from a reduction of
the contact surface area between the vocal folds, which
could be related to a reduction in the thickness of the vocal folds compared to modal register [17–19]. The reinforced falsetto usually presents a greater glottal adduction,
and hence sounds louder, brighter and has more harmonic energy in the high spectral region (2–5 kHz) than naive
falsetto [20]. Musical styles where singers may use reinforced falsetto are: blues, glam rock, epic metal, soul, etc.
The reinforced falsetto in contemporary commercial music is usually associated with a shortening of the vocal tract
and a very open mouth configuration. Perceptually, it may
sound pressed and similar to screaming [20].
Since both growl voice and reinforced falsetto are
probably produced with laryngeal and vocal tract constrictions, they could be labeled as vocal resources based
on hyperfunctional vocal activity, which in turn might be

potentially harmful to the phonatory mechanism. Nevertheless, there are no empirical studies supporting this assumption. The present study aimed to perceptually,
acoustically, functionally, and laryngoscopically assess a
group of rock singers who use growl voice and/or reinforced falsetto. In particular, the question was explored of
whether or not frequent use of growl voice and reinforced
falsetto in singing had an effect on the characteristics of
the speaking voice.

Voice Assessment in Rock Singers

Folia Phoniatr Logop 2013;65:248–256
DOI: 10.1159/000357707

Methods
Participants
Twenty-one rock singers (19 men and 2 women) participated
as subjects in the present study (experimental group). Their average age was 26 years, with a range from 19 to 34 years. All were
native Spanish speakers and recruited from amateur rock bands.
Since one of the authors of the present study (A.H.) belongs to the
singing rock field, he contacted and recruited the participants. The
inclusion criteria were: (1) experience producing growl voice and/
or reinforced falsetto for at least 1 year, and (2) practice of the technique at least 4 times a month during performances or rehearsals.
None of the participants reported previous voice therapy/training.
None of the participants had a hearing impairment. All participants were asked to explain and produce the voice techniques they
use in rock singing before undergoing endoscopic examination.
This procedure was perceptually verified by 4 of the authors of the
present study, all of whom were speech-language pathologists with
experience in singing voice.
A control group was also included in the present study. Eighteen pop singers (14 men and 4 women) with more than 2 years of
experience performing different pop styles (except rock) were recruited for the same voice assessment. Their average age was 28
years, with a range from 20 to 32 years. The experiments were con-

249


ducted with the understanding and the written consent of each
participant. Participants from both experimental and control
groups underwent an evaluation session lasting no more than 1 h.
This session included (1) application of the Singing Voice Handicap Index (S-VHI), (2) voice recording, and (3) laryngoscopic evaluation. Individual explanation and demonstrations were provided
about the required tasks before performing the examinations.
Questionnaire Application
All participants were asked to complete the Spanish adaptation
and validation of the S-VHI. This self-administrated questionnaire
is a health status instrument designed to assess the voice handicap
resulting from singing voice problems [21, 22]. The S-VHI has important psychometric properties of reliability and validity. It contains 36 items chosen to address the physical, emotional, social,
and economic impact of singing voice problems. Each item is individually scored on a 5-point Likert scale anchored by ‘never’
(score of 0) and ‘always’ (score of 4).
Recording Procedure
Subjects were required to produce three repetitions of a sustained vowel /a/ using a comfortable fundamental frequency and
loudness level; each vowel was sustained for 6 s. The vowel productions were recorded in an acoustically treated booth, with an ambient noise level below 30 dB. For the recording, a Focusrite Saffire
6 USB interphase (Focusrite, Calif., USA) and an omnidirectional
condenser microphone (Samson MM01, Samson Technologies
Corp., Hauppauge, N.Y., USA) positioned at a distance of 20 cm
from the mouth, with a 45° inclination angle was used. Samples
were digitally recorded using Steinberg Cubase LE4 (Steinberg,
Germany) software in a WAV format at a sampling rate of 44.1
kHz with 16 bits/sample quantization.
Acoustic Analysis
Acoustic analysis of all recorded sustained vowels included harmonic-to-noise ratio (HNR), frequency perturbation measure (jitter), and amplitude perturbation measure (shimmer). They were obtained by PRAAT software, version 5.0.23 (Bergsma & Weenink,
2008). Recorded samples were edited prior to acoustic analysis. The
first and last second of each sample were excluded, leaving the central 4 s for analysis. The editing procedure was carried out using

Goldwave v5.58 software (GoldWave Inc., St. John’s, Nfld., Canada).
Laryngoscopic Assessment
After voice recordings, participants underwent a laryngoscopic examination. They were asked to sit upright in a comfortable
chair. Assessment of laryngeal activity was carried out with a flexible fiberoptic endoscope (Pentax VNL-1170K, KayPENTAX, Lincoln Park, N.J., USA) connected to a video camera and a light
source. Analog images were digitalized with a Pentax EPK-1000
(KayPENTAX) digital processor. All examinations were performed without topical nasal anesthesia. The laryngoscopic procedure was carried out by a laryngologist with more than 20 years of
experience in singing and speaking voice assessment. During singing voice tasks, the flexible endoscope was placed near the tip of
the uvula, allowing a full view of the pharynx and larynx. This
placement was set by securing the fiberscope against the alar cartilage of the nose with the laryngologist’s finger. A steady placement of the fiberscope is crucial since observation of laryngeal
height adjustments and other laryngeal configurations can be af-

250

Folia Phoniatr Logop 2013;65:248–256
DOI: 10.1159/000357707

fected by movement of the endoscope. During regular speaking
endoscopic procedures, the flexible endoscope was placed right
above the epilaryngeal tube. For the purposes of this study, two
main aspects were observed during the laryngoscopic procedure:
(1) regular speaking voice laryngoscopic assessment (for both
groups), and (2) assessment during production of growl voice and
reinforced falsetto (only for the experimental group). Phonatory
tasks during the first stage were: sustained vowels and running
speech tasks. To assess growl voice, subjects from the experimental
group were asked to produce a sustained vowel /a/ using a comfortable fundamental frequency (middle range) and loudness level.
To evaluate reinforced falsetto, participants were required to produce a sustained vowel in the highest possible pitch using the most
comfortable vowel and loudness level. Recall that not all the participants produced both growl voice and reinforced falsetto. This
information was previously obtained during the recruitment process. Possible laryngeal constrictions, vertical laryngeal position
and pharyngeal activity were assessed during both growl voice and

reinforced falsetto after laryngoscopic recording was made.
Visual Evaluation of Laryngoscopic Samples
Two fellowship-trained laryngologists blinded to the purpose
and nature of the study were asked to review the laryngoscopic examinations and rate the degree of A-P laryngeal compression, medial laryngeal compression, pharyngeal constriction, and vertical
laryngeal position on a 5-point scale. For vertical laryngeal position 1 = very low, 5 = very high, for medial laryngeal compression
1 = very opened, 5 = very narrow, for A-P laryngeal compression
1 = very opened, 5 = very narrow, and for pharyngeal width 1 =
very wide, 5 = very narrow. Moreover, raters were asked to evaluate the presence of masses, edema, erythema, glottal chink, and
other possible laryngeal alterations (presence  = 1, absence  = 0).
This visual assessment was performed for videos obtained from
both experimental and control groups. For the experimental group
only, a similar visual evaluation was carried out to evaluate growl
voice and reinforced falsetto. The degree of medial and A-P laryngeal compression, medial laryngeal compression, pharyngeal constriction, and vertical laryngeal position was assessed on the same
5-point scale. All sound was removed. Each laryngoscopic examination could be reviewed as many times as desired.
Perceptual Evaluation
Voice audio samples of sustained vowels from both experimental and control groups were perceptually assessed by 3 external raters. All three repetitions of the vowel (from each participant) were
used for this part of the analysis. This group of blinded judges consisted of speech-language pathologists with at least 8 years of experience working with patients with voice disorders. The order of
recordings was randomized. Perceptual assessment was performed
with the GRBAS scale [23]. The GRBAS scale evaluates five parameters: G = grade of dysphonia, R = roughness, B = breathiness, A =
asthenic quality, and S = strain quality. Perceptual variables were
assessed using a 4-point scale (from 0 to 3), where 0 = normal, 1 =
slight, 2 = moderate, and 3 = severe. Raters could replay each sample as many times as they wanted before making their determination and moving on to the next recording. The evaluation was performed in a sound-treated room using a laptop computer and a
high-quality Audioengine loudspeaker (Audioengine, Kowloon,
Hong Kong). The listeners were located at approximately 2 m from
the loudspeaker. All the listeners reported normal hearing.

Guzman/Barros/Espinoza/Herrera/Parra/
Muñoz/Lloyd

Statistical Analysis
Descriptive statistics were calculated for the variables, including mean and standard deviation. A kappa test was performed to
assess the interrater concordance for each auditory perceptual parameter and cut-off point of >0.60 was used. Intraclass correlation
coefficients were computed to assess intrarater concordance.
Moreover, a Spearman correlation analysis between acoustic parameters, S-VHI and perceptual assessment was performed; the
correlational analysis only used the G (grade) dimension from the
GRBAS scale.
The results and the comparisons between experimental and
control groups were assessed using Wilcoxon’s rank-sum test for
continuous variables, and χ2 test and Fisher’s exact for categorical
variables. A p value <0.05 was considered to be statistically significant. Stata 12.0 (StataCorp. 2011, College Station, Tex., USA:
StataCorp LP) statistical software was used for analysis.

Results

Acoustic Analysis
Results (mean and standard deviation) of acoustic
analysis for experimental and control groups are presented in table 1. Group means showed normal values for jitter, shimmer, and HNR. The normal threshold for jitter
and shimmer was considered to be 1 and 3%, respectively.
HNR greater than 10 dB was considered normal. There
were no significant differences between both groups.
Laryngoscopic Evaluation
Table 2 shows the results from the intra- and interrater
reliability analysis. A good intrarater concordance was
demonstrated for each judge. Moreover, the 2 blinded
judges obtained high agreement (interrater reliability) for
all laryngoscopic features (kappa value >0.60). No significant differences were found between groups for all laryngoscopic features. Therefore, there are no structural or
functional laryngeal differences when comparing rock
and pop singers.
Of the 21 subjects from the experimental group, 7

(33.3%) presented with normal laryngeal anatomy. This
includes normality of all laryngeal structures including
the vocal folds. The remaining participants demonstrated
varying degrees of alteration in some laryngeal structures.
Eight of them presented with slight vocal fold erythema,
and 9 singers showed a slight degree of edema in the posterior laryngeal commissure (possibly due to laryngealpharyngeal reflux). Moreover, 2 subjects presented with
slight asymmetry of the false vocal folds and 5 participants demonstrated excessive laryngeal mucous. No mass
lesions (polyps, nodules, cysts, edemas, etc.) were found
in any participant.
Voice Assessment in Rock Singers

Table 1. Mean and standard deviation of acoustic analysis for ex-

perimental and control groups

Jitter
Shimmer
HNR

Experimental

Control

p value

0.29±0.17
1.55±0.54
23.64±3.84

0.36±0.21

1.68±0.51
22.40±4.62

0.2989
0.4924
0.3874

Table 2. Intra- and interrater reliability analysis for laryngoscopic

assessment

Masses
Edema
Erythema
VLP
MLC
A-P LC
Pharyngeal
constriction
Glottal chink
Others

Experimental group

Control group

interrater
(kappa;
p value)

intrarater
(ICC;
p value)

interrater
(kappa;
p value)

intrarater
(ICC;
p value)

–
0.57; 0.009
0.66; 0.007
0.70; 0.005
0.81; 0.0001
0.88; <0.0001

–
0.73; 0.0006
0.68; 0.0011
0.67; 0.0017
0.68; 0.005
0.67; 0.0017

–
0.55; 0.01
0.51; 0.02
0.71; 0.005

0.83; 0.0001
0.85; <0.0001

–
0.54; 0.012
0.57; 0.009
0.82; 0.0001
0.68; 0.0011
0.66; 0.007

0.88; <0.0001 0.73; 0.0006 0.82; 0.0001
0.55; 0.01
0.54; 0.012 0.55; 0.01
0.51; 0.02
0.49; 0.03
0.66; 0.007

0.88; <0.0001
0.66; 0.007
0.54; 0.012

ICC = Intraclass correlation coefficient; VLP = vertical laryngeal position; MLC = medial laryngeal compression; LC = laryngeal compression.

During phonation, of the total number of rock singers
(n = 21), 15 (71.4%) showed normal laryngeal function,
including that of the vocal folds. The remaining participants presented some degree of abnormality. Four of
them demonstrated a slight glottal chink. Furthermore,
slight arytenoid asymmetry was observed in 3 subjects
during phonation. Despite these slight alterations, none of
them demonstrated a major functional laryngeal problem.

When considering supraglottic tissue function during
conversational voice usage, 16 participants (66.6%) from
the experimental group demonstrated supraglottic compression. Nine of them showed A-P compression, 4 presented medial compression, and 3 presented both A-P
and medial compression.
Growl Voice and Reinforced Falsetto Laryngoscopic
Assessment
Table 3 presents the results from intra- and interrater
reliability analysis for growl voice and reinforced falsetto.
A good intrarater concordance was demonstrated for
each judge. Moreover, the 2 blinded judges obtained high
Folia Phoniatr Logop 2013;65:248–256
DOI: 10.1159/000357707

251


Table 3. Intra- and interrater reliability analysis for growl voice and reinforced falsetto
Growl voice

Vertical laryngeal position
Medial laryngeal compression
A-P laryngeal compression
Pharyngeal constriction

Reinforced falsetto

judge 1

judge 2

kappa (p value) ICC (p value)
(interrater)
(intrarater)

judge 1

judge 2

kappa (p value)
(interrater)

ICC (p value)
(intrarater)

3±1.60
3.8±0.94
4.53±0.74
4.13±1.06

2.93±1.33
4.33±0.97
4.53±0.74
4±1

0.56 (0.0011)
0.37 (0.0017)
0.73 (0.0086)
0.60 (0.0057)

4.66±0.49

4.08±0.79
3.5±0.79
4.41±0.51

4.75±0.45
4±0.95
3.66±0.88
4.66±0.49

0.58 (0.0008)
0.33 (0.002)
0.83 (0.0005)
0.58 (0.0038)

0.73 (0.0086)
0.68 (0.0011)
0.67 (0.0017)
0.84 (0.0001)

0.83 (0.0006)
0.55 (0.001)
0.68 (0.0011)
0.80 (0.0005)

ICC = Intraclass correlation coefficient.

Table 4. Scores from growl voice and reinforced falsetto

techniques

Vertical laryngeal position
Medial laryngeal compression
A-P laryngeal compression
Pharyngeal constriction

Growl
voice

Reinforced p value
falsetto

2.96±1.44
4.06±0.90
4.53±0.69
4.06±0.97

4.70±0.33
4.04±0.72
3.58±0.82
4.54±0.39

0.0004
0.4693
0.0032
0.0639

agreement (interrater reliability) for all laryngoscopic
variables in both growl voice and reinforced falsetto.
Figure 1 and table  4 show the scores from both growl
voice and reinforced falsetto. Significant differences were

found between groups for vertical laryngeal position and
A-P laryngeal compression. No differences were demonstrated for medial laryngeal compression and pharyngeal
constriction.
Twelve of the 21 participants produced reinforced
falsetto and 15 produced growl voice. Therefore, 6 singers performed both techniques. All singers who performed reinforced falsetto chose the vowel /a/ to perform the required high pitch and presented a high vertical laryngeal position and open-mouth configuration
during examination. Moreover, in 5 of the 12 participants (41.6%) a medial laryngeal compression was observed. Four subjects (33.3%) presented both medial and
A-P compression. The remainder (n = 3, 25%) did not
demonstrate laryngeal compression during reinforced
falsetto phonation. Additionally, pharyngeal constriction was demonstrated during endoscopic examination
in 11 of the 12 participants (92%) during reinforced falsetto phonation.
For growl phonation, A-P supraglottic compression was
observed in only 15 participants. One participant showed
only medial compression (7%). In most of the participants
252

Folia Phoniatr Logop 2013;65:248–256
DOI: 10.1159/000357707

Table 5. Mean and standard deviation of S-VHI for experimental
and control groups

Physical
Social
Emotional
Economic
Total

Experimental

Control

p value

19.28±9.01
8.19±7.33
1.57±2.03
0.42±0.87
29.47±16.76

12.33±6.45
0.66±1.23
1.73±2.37
0.66±1.17
15.4±6.99

0.0154
0.0004
0.8277
0.4891
0.0043

Vertical laryngeal Medial laryngeal
position
compression

A-P laryngeal
compression

Pharyngeal
constriction

5
4
3
2
1
0
Growl Falsetto Growl Falsetto Growl Falsetto Growl Falsetto

Fig. 1. Scores from growl voice and reinforced falsetto techniques.

who produced growl voice (n = 11, 73%), both medial and
A-P supraglottic compression was demonstrated. Furthermore, pharyngeal constriction was demonstrated in 10
participants (67%). Thirteen singers (87%) produced a vertical laryngeal position change. Four of them demonstrated
a higher vertical laryngeal position compared to the rest
level. Most of them (9 subjects) presented a lower vertical
laryngeal position during growl voice.
Guzman/Barros/Espinoza/Herrera/Parra/
Muñoz/Lloyd


Table 6. Spearman correlation analysis between acoustic parameters, S-VHI and perceptual assessment for the experimental group

S-VHI
Jitter
Shimmer
HNR
G

S-VHI

Jitter

Shimmer

HNR

G

1
rho = 0.23; p = 0.3143
rho = 0.11; p = 0.6135
rho = –0.31; p = 0.1576
rho = 0.30; p = 0.1855

–
1
rho = 0.53; p = 0.0124
rho = –0.70; p = 0.003
rho = 0.26; p = 0.2422

–
–
1
rho = –0.68; p = 0.007
rho = 0.38; p = 0.086

–
–
–

1
rho = –0.46; p = 0.0328

–
–
–
–
1

From perceptual analysis through GRBAS scale, only G (grade of dysphonia) was used.

Questionnaire Results
Results (mean and standard deviation) of S-VHI for
experimental and control groups are presented in table 5.
Significant differences were found in physical and social
spheres between groups. There were no differences for
emotional and economic spheres. Total scores were also
significantly different between groups.
For the experimental group, results showed that the
highest rated variables were in the physical sphere, rated
19.28 ± 9.01 points (26.78%) from the total of 72 possible
points. However, even the highest rating was less than a
third of the total possible points. The ratings were typically 1 = almost never (18 subjects, 85.7%) or 2 = sometimes (3 subjects, 14.3%).
Subjects from the experimental group responded ‘almost never’ in the physical sphere (52.38%) for the variable ‘It takes a lot of effort to sing’. The rest (47.61%) responded ‘almost never’ for the variables ‘I have trouble
making my voice do what I want it to’ and ‘I have trouble
controlling the raspiness in my voice’. Therefore, more
than half of the subjects reported that almost never they
have the sensation of effort during singing.
The social sphere was the second most affected variable in the experimental group (1.57 ± 2.03 points,
19.64%), followed by the emotional sphere (8.19 ± 7.33

points, 14.47%), from the total of 56 points and 8 points,
respectively. Moreover, it is important to mention that
the most common response in all items of the emotional
sphere was the option 0 = never (9 subjects, 43%). The less
affected S-VHI variable was the economic sphere (0.42 ±
0.87 points, 5.35%).
Perceptual Evaluation
Results of the perceptual evaluation by raters using
the GRBAS scale for experimental and control groups
indicated that all the parameters obtained a kappa value
above 0.60 and p value <0.05, except roughness for the
control group. Therefore, there was a high interrater
Voice Assessment in Rock Singers

concordance for most of perceptual variables. Asthenia
(A) demonstrated the lowest value (0) and grade of dysphonia (G) the highest value (0.34) for the experimental
group. Moreover, the 2 blinded judges obtained high
intrarater reliability. There were no significant differences between both groups for any of the perceptual
parameters.
Correlation Analysis
The results of correlation analysis are presented in table 6. Statistically significant correlation (p value <0.05)
was found between grade of dysphonia (G) and HNR.
Moreover, jitter correlated with shimmer and HNR. No
correlations were found with S-VHI.

Discussion

The term ‘edgy voice’ is defined as all the sounds that
are commonly used by rock singers and sound like a distortion of the ‘clean voice’. Some of these edgy sounds are
growling, guttural voice, raspy voice and pressed high

notes [24]. The present study aimed to perceptually,
acoustically, functionally, and laryngoscopically assess a
group of 21 rock singers who use two of these ‘edgy’
sounds: growl voice and reinforced falsetto.
Results revealed that both the growl voice and the reinforced falsetto are characterized by supraglottic compression, pharyngeal constriction and changes in vertical laryngeal position. The rough and raspy perceptual quality
of growl voice production is likely produced by the vibration of the supraglottic structures involved in the observed
laryngeal compression. Borch et al. [25] assessed the
acoustic and physiology of ‘distorted tones’ in rock singers
and found significant supraglottic activity including the
aryepiglottic folds, anterior part of the mucosa covering
the arytenoids, and ventricular folds (medial compression). An earlier investigation performed in growl voice
Folia Phoniatr Logop 2013;65:248–256
DOI: 10.1159/000357707

253


technique also showed simultaneous vibration of laryngeal supraglottic structures [12]. In the present study,
growl voice demonstrated a significantly higher degree of
aryepiglottic (A-P) compression than reinforced falsetto.
More than 70% of subjects from the experimental
group demonstrated not only aryepiglottic constriction,
but also medial supraglottic compression during growl
voice. Fuks et al. [13] in a high-speed video study reported approximation and self-sustained ventricular fold vibration in different singing modes, one of which was similar to Mongolian bass-type singing (perceptually similar
to growl voice quality). Additionally, Lindestad et al. [26]
analyzed the bass-type throat singing. The ventricular
folds had low amplitude of vibration with incomplete
closure.
The supraglottic laryngeal activity found in the present
study during singing could be considered a hyperfunctional behavior and hence a possible risk factor for developing a voice disorder for the singers who frequently use

it. Several definitions of laryngeal hyperfunction exist,
but a recurrent feature in almost all descriptions includes
excessive laryngeal musculoskeletal activity, force, or tension [27, 28]. The basic paradigm that evaluates laryngeal
hyperfunction is to look for compression of the supraglottic structures during phonation [29]. This supraglottic activity may be separated into two components: (1)
A-P supraglottic activity and (2) medial supraglottic activity [30–32]. In the present study most of the participants from the experimental group not only demonstrated laryngeal supraglottic activity during singing, but 14 of
them (67%) also presented with some slight degree of laryngeal supraglottic activity during speaking voice tasks.
Therefore, it would be possible to speculate that these
subjects (rock singers) could have laryngeal hyperfunction, maybe caused by the use of growl voice and reinforced falsetto during singing. Nevertheless, when comparing supraglottic laryngeal activity between groups
during speaking voice tasks, no significant differences
were found. Therefore, it is unlikely that the slight hyperfunction observed during speaking is caused by growl
voice and reinforced falsetto.
Despite the apparent widespread acceptance of these
postures as endoscopic signs of vocal hyperfunction,
some studies have shown that this supraglottic activity
could be present in subjects with normal voice [20, 30, 33,
34]. Sama et al. [33] assessed the prevalence of laryngeal
hyperfunction in two groups: subjects diagnosed with
functional dysphonia and a group of subjects with normal
voice. No differences were found. Similarly, Stager et al.
[30] found the presence of anterior-posterior compres254

Folia Phoniatr Logop 2013;65:248–256
DOI: 10.1159/000357707

sion in patients with nodules, functional dysphonia, and
normal voice.
The fact that both acoustic and perceptual analysis in
the present study revealed normal and quasi-normal
voices, respectively, might be another reason to question
that the supraglottic activity present in rock singers during speaking voice tasks is really a sign of laryngeal hyperfunction and hence a risk factor for vocal fold injury. In

addition, significant correlation was found between perceptual grade of dysphonia (G) and all the acoustic parameters. Moreover, no significant differences were
found between groups when comparing acoustics and
perceptual data.
The laryngeal and pharyngeal configurations observed
during both growl voice and reinforced falsetto in most
of participants is not surprising since previous studies
have reported these compressions as normal and even
desirable vocal adjustment in various styles of singing.
Studies carried out with belters suggested that this technique is typified by epiglottis tilted over the larynx [35].
Yanagisawa et al. [3] also reported aryepiglottic constriction during twang, belting, and opera qualities. Pershall
et al. [36] demonstrated similar results. Supraglottic
compression has also been found in Middle Eastern singing [37]. In addition, there is empirical evidence that A-P
constriction can contribute to an acoustic advantage
(voice quality) [5, 38].
Even though growl voice possesses a quality that may
sound pressed or constricted, the major work may not be
performed by the vocal folds, but by the laryngeal supraglottic structures and pharynx. This assumption is based
on the fact that some vocal coaches teach their students
to produce a breathy voice quality before adding the vibration of supraglottic structures when producing growl
voice. Moreover, it is also taught that growl voice should
not be produced by tightening the vocal folds, but to relax
them and using more abdominal support (abdominal
muscle contraction instead of laryngeal effort) [24, 39,
40]. This might explain the laryngoscopic normality
found in most of our participants.
The reinforced falsetto demonstrated high vertical laryngeal position, pharyngeal constriction and supraglottic compression. However, they may not be possible
causes of the slight vocal folds alterations found during
laryngoscopy. The use of adequate resonance strategies to
avoid an excessive vocal fold muscle effort during high
pitch phonation might be a suitable explanation. The resonance strategies used during reinforced falsetto in high

pitches were assessed in the same subject group in a previous study [20]. Authors reported that high values of the
Guzman/Barros/Espinoza/Herrera/Parra/
Muñoz/Lloyd


two first vocal tract formant frequencies (F1 and F2) were
observed in all participants. A clear increase of the spectral energy in one or two harmonics due to the coincidence or proximity with one or two formants was demonstrated in all singers. These outcomes would imply that
the total sound pressure level of the radiated sound during reinforced falsetto production may not be supported
by increasing vocal effort (high glottal resistance and high
subglottic pressure), but by an adequate vocal tract configuration strategy. Borch et al. [25] reported the same
formant changes in rock singing. They stated that the values reflected a relatively high larynx position and open
mouth configuration. In this regard, earlier studies have
demonstrated that a similar vocal tract configuration is
classically displayed by sopranos during production of
high pitches to avoid laryngeal muscle effort [41–43].
Additionally, according to Titze and Story [44], Story
et al. [45] and Titze [46], an important clinical issue pertains to the conflict of accepting (and even encouraging)
the use of a narrow epilaryngeal tube for healthy voice
production. The authors state that the source-filter interaction and vocal tract inertance may be increased by narrowing the epilaryngeal tube in an A-P direction. Therefore, this A-P narrowing could constitute a benefit for
vocal fold oscillation and vocal fold abduction [44–46].
Although, there is some evidence that the supraglottic
activity demonstrated in the present study might not be a
real hyperfunctional and detrimental laryngeal behavior,
it is not possible to assure that over a longer period of time
the same results would be found. Most of our participants
only practice growl voice and reinforced falsetto occasionally (at least 4 times a month) and also most of them
have been using these techniques for about 1 year.
Furthermore, the possibility that rock singing might
be harmful to the phonatory mechanism cannot be ruled
out. A previous investigation revealed that rock style is

characterized by higher values of subglottic pressure,

closed quotient, and pitch range than other nonclassical
singing styles. Moreover, regarding mode of phonation,
rock singing was perceptually rated as the most pressed
style among the others. This was objectively corroborated
by the lowest normalized amplitude quotient value (high
degree of phonatory pressedness) [47, 48]. Loud phonation and high subglottic pressures were also found in rock
singing in a study conducted by Borch et al. [25].
Functional assessment through S-VHI showed results
that seem to be in line with the above-mentioned findings. Although the most affected variable was the physical
sphere and it was higher in rock singers than in pop singers, it was affected only in 26.78% of the experimental
group. Furthermore, most subjects responded 1 = almost
never or 2 = sometimes. These outcomes could indicate
that participants do not consider they have a significant
physical impairment with their singing voice.

Conclusions

Although both growl voice and reinforced falsetto
have physiologic characteristics that are usually associated with laryngeal hyperfunction (high vertical laryngeal position, pharyngeal constriction and both medial and
A-P supraglottic compression), they do not seem to contribute to laryngeal disorders in the assessed group of
rock singers. Perceptual, acoustic, functional and laryngoscopic assessment demonstrated no major alterations
in most of the participants. Likely a proper resonance
strategy in reinforced falsetto and a decreased glottal adduction in growl voice could be the factors that contribute
to the avoidance of voice problems in singers that use
these vocal resources that are classically labeled vocal
abuse. We do not overlook the possibility that vocal fold
disorders could be found if singers used these techniques
more frequently and for longer periods of time.

References
1 Imagawa H, Sakakibara K, Tayama N, Niimi
S: The effect of the hypopharyngeal and supraglottic shapes on the singing voice. Proc
Stockholm Music Acoustics Conf, Stockholm, 2003.
2 Sundberg J, Gramming P, Lovetri J: Comparisons of pharynx, source, formant and pressure characteristics in operatic and musical
theatre singing. J Voice 1993;7:301–310.
3 Yanagisawa E, Estill J, Kmucha S, Leder S: The
contribution of aryepiglottic constriction to
‘ringing’ voice quality: a videolaryngoscopic

Voice Assessment in Rock Singers

4
5
6
7

study with acoustic analysis. J Voice 1989; 3:
342–350.
Sundberg J: The Science of the Singing Voice. Illinois, Northern Illinois University Press, 1989.
Sundberg J: Articulatory interpretation of the
singing formants. J Acoust Soc Am 1974; 55:
838–844.
Catford JC: Fundamental Problems in Phonetics.
Edinburgh, Edinburgh University Press, 1977.
Esling JH: Pharyngeal consonants and the
aryepiglottic sphincter. J Int Phonet Assoc
1996;26:65–88.

8 Zemp H: Les Voix du Monde: une anthologie
des expressions vocales. Paris, Musée de
l’Homme, 1996.
9 Bang K: Your Head: The Rise and Fall of Heavy
Metal. New York, Three Rivers Press, 2002.
10 Locos A, Bonada L: Emulating rough and growl
voice in spectral domain. Proc 7th Int Conf on
Digital Audio Effects (DAFX-04), Naples, 2004.
11 Whitehead RL, Metz D-E, Whitehead BH: Vibratory patterns of the vocal folds during
pulse register phonation. J Acoust Soc Am
1984;75:1293–1297.

Folia Phoniatr Logop 2013;65:248–256
DOI: 10.1159/000357707

255


12 Sakakibara K, Fuks L, Imagawa H, Tayama N:
Growl voice in ethnic and pop styles. Proc Int
Symp on Musical Acoustics (ISMA 2004),
Nara, 2004.
13 Fuks L, Hammarberg B, Sundberg J: A selfsustained vocal ventricular phonation mode:
acoustical, aerodynamic and glottographic
evidences. KTH TMH-QPSR 1998;3:49–59.
14 Lindestad PA, Sodersten M, Merker B, Granqvist S: Voice source characteristics in Mongolian ‘throat singing’ studied with highspeed imaging technique, acoustic spectra,
and inverse filtering. J Voice 2001;15:78–85.
15 Sakakibara K, Konishi T, Kondo K, Murano
EZ, Kumada M, Imagawa H, Niimi S: Vocal
fold and false vocal fold vibrations and synthesis of Khoomei. Proc ICMC 2001, Japan,

2001.
16 York W: Voices from the Hell. San Francisco,
Bay Guardian, 2004.
17 Miller DG, Svec JG, Schutte HK: Measurement of characteristic leap interval between
chest and falsetto registers. J Voice 2002; 16:
8–19.
18 Hirano M: Vocal mechanisms in singing: laryngological and phoniatric aspects. J Voice
1988;2:51–69.
19 Henrich N, d’Alessandro C, Castellengo M,
Doval B: Glottal open quotient in singing:
measurements and correlation with laryngeal
mechanisms, vocal intensity, and fundamental frequency. J Acoust Soc Am 2005; 117:
1417–1430.
20 Guzman M, Barros M, Espinoza M, Herrera
A, Parra D, Lloyd A: Resonance strategies in
rock singers. J Singing, in press.
21 Cohen SM, Jacobson BH, Garrett CG,
Noordzij JP, Stewart MG, Attia A, Ossoff RH,
Cleveland TF: Creation and validation of the
singing voice handicap index. Ann Otol Rhinol Laryngol 2007;116:402–406.
22 García-López I, Nez-Batalla F, Gavilán J,
Górriz-Gil C: Validación de la versión en espol del índice de incapacidad vocal (S-VHI)
para el canto. Acta Otorrinolaringol Esp 2010;
61:247–254.

256

23 Hirano M: Clinical Examination of Voice. Vienna, Springer, 1981.
24 Martin B: Pro Secrets of Heavy Rock Singing.
London, Sanctuary Publishing, 2002.

25 Borch DZ, Sundberg J, Lindestad PA, Thalén
M: Vocal fold vibration and voice source aperiodicity in ‘dist’ tones: a study of a timbral
ornament in rock singing. Logoped Phoniatr
Vocol 2004;29:147–153.
26 Lindestad P, Södersten M, Merker B, Granqvist S: Voice source characteristics in Mongolian ‘throat singing’ studied with highspeed imaging technique, acoustic spectra,
and inverse filtering. J Voice 2001;15:78–85.
27 Hillman RE, Holmberg EB, Perkell JS, Walsh
M, Vaughan C: Objective assessment of vocal
hyperfunction: an experimental framework
and initial results. J Speech Hear Res 1989;32:
373–392.
28 Hillman RE, Gress C, Hargrave J, Walsh M,
Bunting G: The efficacy of speech-language
pathology intervention: voice disorders. Semin Speech Lang 1990;11:297–310.
29 Rubin A, Praneetvatakul V, Gherson S, Moyer C, Sataloff R: Laryngeal hyperfunction during whispering: reality or myth? J Voice 2006;
20:221–127.
30 Stager S, Bielamowicz S, Regnell J, Gupta A,
Brakmeier J: Supraglottic activity: evidence of
hyperfunction or laryngeal articulation? J
Speech Hear Res 2000;43:229–238.
31 Behrman A, Dahl L, Abramson A, Schutte H:
Anterior-posterior and medial compression
of the supraglottis: signs of nonorganic dysphonia or normal postures? J Voice 2003; 17:
403–410.
32 Von Deorsten PG, Izdebski K, Ross JC, Cruz
RM: Ventricular dysphonia: a profile of 40
cases. Laryngoscope 1992;102:1296–1301.
33 Sama A, Carding PN, Price S, Kelly P, Wilson
JA: The clinical features of functional dysphonia. Laryngoscope 2001;111:458–463.
34 Stager S, Bielamowicz SA, Gupta A, Marullo

S, Regnell JR, Barkmeier JM: Quantification
of static and dynamic supraglottic activity. J
Speech Lang Hear Res 2001;44:1245–1256.

Folia Phoniatr Logop 2013;65:248–256
DOI: 10.1159/000357707

35 Lawrence V: Laryngological observations on
belting. J Res Singing 1979;2:26–28.
36 Pershall KE, Boone S: Supraglottic contribution to voice quality. J Voice 1987;1:186–190.
37 Hamdan AL, Sibai A, Moukarbel RV, Deeb R:
Laryngeal biomechanics in Middle Eastern
singing. J Voice 2006;20:579–584.
38 Hanayama E, Camargo Z, Tsuji D, Pinho S:
Metallic voice: physiological and acoustic features. J Voice 2009;23:62–70.
39 Baxter M: The Rock-N-Roll Singer’s Survival
Manual. UK, Library of Congress cataloging,
1990.
40 Surnami K, Mitchell K: Rock Singing Techniques. Los Angeles, Alfred Music Publishing,
2001.
41 Sundberg J, Skoog J: Dependence of jaw opening on pitch and vowel in singers. J Voice
1997;11:301–308.
42 Hertegard S, Gauffin J, Sundberg J: Open and
covered fiberoptics, inverse singing as studied
by means of filtering, and spectral analysis. J
Voice 1990;4:220–230.
43 Sundberg J, Gramming P, LoVetri P: Comparisons of pharynx, source, formant, and
pressure characteristics in operatic and musical theatre singing. J Voice 1993;7:301–310.
44 Titze I, Story B: Acoustic interactions of the
voice source with the lower vocal tract. J

Acoust Soc Am 1997;101:2234–2243.
45 Story B, Laukkanen A-M, Titze I: Acoustic
impedance of an artificially lengthened and
constricted vocal tract. J Voice 2000; 14: 455–
469.
46 Titze I: Voice training and therapy with a
semi-occluded vocal tract: rationale and scientific underpinnings. J Speech Lang Hear
Res 2006;49:448–459.
47 Borch DZ, Sundberg J: Some phonatory and
resonatory characteristics of the rock, pop,
soul, and Swedish dance band styles of singing. J Voice 2011;25:532–537.
48 Sundberg J, Thalen M, Alku P, Wilkman E:
Estimating perceived phonatory pressedness
in singing from flow glottograms. J Voice
2004;18:56–62.

Guzman/Barros/Espinoza/Herrera/Parra/
Muñoz/Lloyd


Copyright: S. Karger AG, Basel 2014. Reproduced with the permission of S. Karger AG,
Basel. Further reproduction or distribution (electronic or otherwise) is prohibited without
permission from the copyright holder.