ARTICLE IN PRESS
Glottal Airflow and Glottal Area Waveform Characteristics of
Flow Phonation in Untrained Vocally Healthy Adults
~, *zBloomington, Indiana, yStockholm, Sweden, and xMadrid,
*Rita R. Patel, †Johan Sundberg, ‡Brian Gill, and §Filipa M.B. La
Spain
Summary: Objective. To examine flow phonation characteristics with regard to vocal fold vibration and voice
source properties in vocally healthy adults using multimodality voice measurements across various phonation
types (breathy, neutral, flow, and pressed) and loudness conditions (typical, loud, and soft).
Participants and Methods. Vocal fold vibration, airflow, acoustic, and subglottal pressure was analyzed in 13
untrained voices (six female and seven male). Participants repeated the syllable / pæ:/ using breathy, neutral,
flow, and pressed phonation during typical, loud, and soft loudness conditions. Glottal area (GA) waveforms
were extracted from high-speed videoendoscopy; glottal flow was derived from inverse filtering the airflow or the
audio signal; and subglottal pressure was measured as the intraoral pressure during /p/ occlusion.
Results. Changes in phonation type and loudness conditions resulted in systematic variations across the relative
peak closing velocity derived from the GA waveform for both males and females. Amplitude quotient derived
from the flow glottogram varied across phonation types for males.
Conclusion. Multimodality evaluation using the GA waveform and the inverse filtered waveforms revealed a
complex pattern that varied as a function of phonation types and loudness conditions across males and females.
Emerging findings from this study suggests that future large-scale studies should focus on spatial and temporal
features of closing speed and closing duration for differentiating flow phonation from other phonation types in
untrained adults with and without voice disorders.
Key Words: Flow phonation−High-speed videoendoscopy−Glottal area waveform−Inverse filtering.
INTRODUCTION
Phonation can be varied within wide limits with regard to
both loudness and phonation type, ranging from breathy/
hypofunctional to pressed/hyperfunctional. Traditionally,
phonation types are characterized by variations in glottal
flow (Figure 1) using the non-invasive technique called glottal inverse filtering. In glottal inverse filtering, the effects of
the vocal-tract filter characteristics and the lip radiation are
cancelled from the speech output to provide estimates of the
glottal source characteristics during phonation. The resulting flow glottograms which shows the glottal volume velocity waveform provides information related to the type of
phonation and other glottal source characteristics noninvasively. Several time-based parameters (eg, open quotient,
speed quotient, closed quotient, pulse amplitude) and frequency-based parameters (eg, fundamental frequency, ratio
of the first and second harmonics) can be derived from the
flow glottogram to estimate the glottal excitation characteristics.1 Breathy/hypofunctional phonation has a large
Accepted for publication July 30, 2020.
Sources of financial support or funding: The Voice Foundation Research Grant.
Presentation: Oral presentation at the 48th Annual Symposium: Care of the Professional Voice, The Voice Foundation, Philadelphia, PA, USA, 2019 & The Pan-European Voice Conference, Copenhagen, Denmark, 2019.
From the *Department of Speech, Language and Hearing Sciences, Indiana University, Bloomington, Indiana; yDivision of Speech, Music, and Hearing, KTH Royal
Institute of Technology, Stockholm, Sweden; zVoice Department, Indiana University,
Bloomington, Indiana; and the xDepartment of Didactics, School Organization and
Special Didactics, Faculty of Education, The National Distance Education University
(UNED), Madrid, Spain.
Address correspondence and reprint requests to Rita R. Patel, Department of
Speech, Language and Hearing Sciences, Indiana University, 200 South Jordan Avenue, Bloomington, IN 47405-7002. E-mail: patelrir@indiana.edu
Journal of Voice, Vol. &&, No. &&, pp. &&−&&
0892-1997
© 2020 The Voice Foundation. Published by Elsevier Inc. All rights reserved.
https://doi.org/10.1016/j.jvoice.2020.07.037
amplitude of the flow oscillation and lacks vocal fold closure. Pressed/hyperfunctional phonation, by contrast, is
characterized by a small amplitude of the flow oscillation,
the minimum of which reaches zero. Neutral phonation,
which is typically used in normal conversation, is characterized by larger peak-to-peak amplitude of the flow oscillation
than pressed/hyperfunctional phonation and is produced
with complete or almost complete vocal fold closure. Flow
phonation has been defined as a phonation type produced
with the largest peak-to-peak flow amplitude, where the
minimum still reaches zero. The characteristic airflow for
flow phonation is reduced compared to breathy phonation
but higher compared to typical or neutral phonation.2,3 In
this sense, flow phonation can be regarded as a phonation
type located between breathy and neutral along the
breathy-pressed continuum.2,3
Phonation types are relevant both pedagogically and clinically. Breathy/hypofunctional phonation is inefficient, as
air is expended without generating complete vocal fold closure. The opposite extreme of pressed/hyperfunctional phonation is produced with excessive muscular and physical
effort/force. Habitual use of hyperfunctional phonation
may lead to voice disorders.4,5
Several therapeutic approaches have been proposed for
rehabilitation/habilitation of hyperfunctional phonation.
One such approach used in voice therapy, aimed at reducing
vibration dose, increased vocal fold adduction,6 and excessive muscular effort7,8 in patients with muscle tension dysphonia,9 is called flow phonation. Flow phonation is
associated with a higher ratio between output sound pressure level and input subglottal pressure compared to neutral
and pressed phonations. Thus, flow phonation has been
ARTICLE IN PRESS
2
FIGURE 1. Typical examples of flow glottograms for the indicated phonation types of breathy, flow, neutral, and pressed.
regarded, in some sense, as optimal.2 Voice production in
flow phonation is known to feel effortless and efficient.10
Flow phonation generates a strong voice source fundamental, indicating a touch glottal adduction.2 Since the pulse
amplitude from the flow glottograms (AC amplitude) is
high in flow phonation, it implies good airflow with a clear,
but not so long closed phase and a strong voice source fundamental.
Flow phonation should not be confused with another
voice therapy approach called stretch-and-flow,11,12 where
the goal is to facilitate a balance between the respiratory,
phonatory, and articulatory subsystems of voice production
using a hierarchical approach of unvoiced to voice airflow
stimuli. Furthermore, therapeutically, there is often confusion between breathy phonation and flow phonation. In
studies on healthy male participants, breathy phonation has
been found to be produced with low subglottal pressure and
incomplete glottal closure causing a larger glottal airflow
than flow phonation. The ratio between sound level and
subglottal pressure is much lower in breathy phonation
compared to both flow phonation and neutral phonation.
Airflow in flow phonation is lower than in breathy phonation but higher than neutral phonation.2,3,13,14 Thus, flow
phonation and breathy phonation are associated with substantially different voice source properties.2,3,13,14 It is
important to note the possibility of using breathy phonation
as a means to teach patients/students how to produce flow
phonation, even though the goal is to produce flow phonation without a breathy voice quality.
The goal of voice treatment when using the flow phonation approach typically is to improve phonatory physiology
by guiding the patient to combining high glottal airflow
with complete glottal closure.2,15 Treatment outcomes clinically are measured using laryngeal imaging, acoustic
Journal of Voice, Vol. &&, No. &&, 2020
evaluation, aerodynamic evaluation, or measurement of
patient self-perception of the impact of the voice disorder
on their daily function.16−18 The advantage of direct visualization of the vocal folds is that it can be used to provide
feedback of the vocal fold contact during voice training.
Inverse filtering, though useful for noninvasive evaluations
of the glottal source characteristics, has not gained widespread clinical use, probably because it is time-consuming
and has limitations in accurately estimating glottal source
properties across all voice types. For example, high-pitch
voices of females and children including the presence of
nasalization makes estimating source characteristics unreliable with inverse filtering and thus limiting its clinical applicability across the range of voice conditions routinely
encountered in the clinics.1 Direct visualization of the vocal
folds through methods of laryngeal imaging is the gold standard to evaluate the structure and the function of the source
/ vocal folds.17,19 Stroboscopy is the gold-standard in laryngeal imaging, although limited for tracking cycle-to-cycle
vocal fold vibrations in severely aperiodic voices.20,21 The
increased temporal resolution of high-speed videoendoscopy
over stroboscopy has the advantage of capturing cycle-tocycle variation in glottal area (GA) rather than an apparent
motion.20,22,23
Most studies evaluating flow phonation have used inverse
filtering.2,3,14,15,24,25 Therefore, it is important to evaluate
the relationship between airflow and the GA by simultaneously using inverse-filtering and high-speed videoendoscopy. It is well-established that the airflow pulse is
asymmetrical26,27 and skewed to the right while the GA
waveform is more symmetrical.28,29 Theoretically, this relationship has been investigated for normal and loud phonations.30,31 Direct relationship between simultaneous airflow
and GA has been examined in only a handful of studies.
Hertegard, Gauffin, and Karlsson (1992)32 examined the
relationship between GA and flow in two males and three
females (26−45 years) in chest register at frequencies around
175 Hz, 200 Hz, and 250 Hz for females and 175 Hz,
220 Hz, and 240 Hz for males in normal and breathy phonation using simultaneous inverse filtering, electroglottography, and videostroboscopy. The findings revealed that the
flow glottogram volume velocity waveform, henceforth the
flow glottogram, in males and females during complete closure had an offset of 20−30 mL/second for typical phonation. Furthermore, the authors found a small hump at the
beginning of the closed phase in the flow glottograms at low
pitch, which coincided with a large mucosal wave on videostroboscopy, revealing a clear vertical phase difference
between the upper and lower margins. This vertical component was more pronounced in males compared to females
but it was not observed for both genders during high
pitch phonation. Similar piston movement was later confirmed by Hertegard and Gauffin (1995).33 Piston movement during flow phonation was also confirmed by
Granqvist et al (2003)34 in a trained male and a trained
female participant using simultaneous recordings of airflow,
ARTICLE IN PRESS
Rita R. Patel, et al
Glottal Airflow and Glottal Area Waveform Characteristics
electroglottography, air pressure, acoustic signal, and highspeed videoendoscopy at 1900 frames/second.
Presently there are limited numbers of large group studies
describing in quantitative terms the relationship between
glottal airflow and vibratory parameters as derived from
methods of simultaneous airflow, subglottal pressure, and
laryngeal imaging during flow phonation. The present study
extends the previous efforts by Granqvist et al (2003)34 just
mentioned and by Gauffin & Sundberg (1989)2 and Sundberg (1995),3 in which flow glottogram characteristics were
quantified in male participants. In order to describe flow
phonation in quantitative terms, the relationship between
subglottal pressure, vibratory amplitude, and glottal flow
during breathy, flow, neutral, and pressed phonation was
examined using an ambitious experimental setup of simultaneous high-speed videoendoscopy, inverse filtering, subglottal pressure, electroglottography, and acoustic recordings in
untrained adult male and female vocies. This study seeks to
address the following two main questions:
1. Do the vibratory amplitude, the relative peak closing
velocity (PCV), the flow pulse amplitude, and the maximum flow declination rate decrease systematically as a
function of phonation modes - breathy, flow, neutral,
and pressed - in adult males and females?
2. What is the relationship between total flow (mL/second) and GA (pixels) and its derivatives of (i) flow
amplitude quotient versus area amplitude quotient, (ii)
duration of flow decrease phase (ms) and duration of
the area decrease phase (ms). Additionally the relationship between the relative PCV and the amplitude-tolength ratio (ALR), both derived from the GA waveform, was explored.
METHOD
Participants
Nineteen untrained young adults (18−45 years), ten females
and nine males participated in the experiment. The lower
age limit was established to conform to the NIH definition
of an adult participant, while the upper age limit was established to reduce the confounding effects of voice changes
due to senescence. Children were excluded from this initial
3
study because the determinants for flow phonation may be
different from those for adults and warrant a separate study.
For inclusion in the study, the participants had negative histories of vocal pathology, no identifiable vocal fold pathology on stroboscopic screening, negative history of smoking,
were able to produce different types of phonations (breathy,
flow, neutral, and pressed), and had no more than up to
4 years of training in classical music. All were perceptually
judged to have a normal voice (overall grade = 0) by a certified speech language pathologist (co-author RP) who has
over 20 years of experience in using the Consensus Auditory
Perceptual Evaluation of Voice scale for determining voice
status and voice disorders.35
Participants were recruited from advertisements and fliers
placed around the Indiana University Campus following
approval by the Institutional Review Board. Two participants (one male & one female) were excluded as they were
unable to produce the different types of phonations. Additionally, two participants (one male & one female) participated in the first half of the experiment without endoscopy
and hence were excluded from further analysis. Data verification of expected average flow values across the various
phonation types (Figure 2) revealed that two female participants did not perform as expected for breathy phonation.
Their flow glottogram signal had minimum flow values at
or near zero rather than well above zero flow and was therefore excluded. Data analysis was performed on the remaining seven males (M = 24.5 § 2.07 years) and six females
(M = 28.86 § 5.11 years).
Data collection/instrumentation
Simultaneous high-speed videoendoscopy, airflow, subglottal pressure, electroglottography, and audio recordings were
conducted for each participant during a series of the syllables /pæ:/ in typical (conversational), loud, and soft phonation for each of the phonation types of breathy, flow,
neutral, and pressed. High-speed videoendoscopic recordings were captured at 2000 frames per second with the
PentaxMedical model 9710 (Montvale, New Jersey). Simultaneous audio (omnidirectional electret microphone
TCM110, V-JEFE), flow (Glottal Enterprises MSIF2), subglottal pressure (Glottal Enterprises PG100E), and
FIGURE 2. Example of the flow glottograms displayed during the recording produced by a male subject with each of the indicated phonation types.
ARTICLE IN PRESS
4
Journal of Voice, Vol. &&, No. &&, 2020
electroglottographic (Glottal Enterprises EG2-PCX)
recordings were captured at 44 kHz on separate tracks using
the data acquisition system (National Instruments, Austin
Texas, USB 6221) and recorded directly with the high-speed
videoendoscopy. The data acquisition system (National
Instruments, Austin Texas, USB 6221) has the capability to
capture up to eight channels of simultaneous data with a
combined sampling rate of 250 kHz and is driven by the
clock-out signal from the high-speed video system, thus
ensuring synchronization.36−38 A xenon light source of 300
watts was used for the high-speed videoendoscopic recordings.
An omnidirectional head mounted microphone was
placed at a mouth-to-microphone distance of 10−12 cm at a
fixed angle of 45 degrees to the side of the mouth. The
microphone was calibrated by recording a sustained phonation, the sound pressure level of which was measured at the
recording microphone at 30 cm mouth-to-microphone distance by means of a sound level meter (RadioShack) with
C-weighting and slow response.17,39
The airflow signal was captured from a Rothenberg flow
mask connected to a Glottal Enterprises MSIF2 inverse filter unit. Each participant held the mask such that it fitted
snuggly to the face. Calibration of the flow signal was conducted by means of the calibration unit provided by the
manufacturer which drives 140 mL of air through the mask.
High-speed videoendoscopic recordings were captured
using a standard digital flexible nasal fiberscope (Pentax
FNL-10RP3) introduced through a custom hole in the flow
mask without application of topical anesthetic to the nasal
mucosa. The fiberscope was fitted tightly in the hole by
means of a custom plug, thus preventing airflow leakage
through the mask.
Subglottal pressure was recorded as the oral pressure during /p/ occlusion captured by means of a pressure transducer
attached to a thin plastic tube. The tube was introduced into
a hole in the flow mask and then into the corner of the participant’s mouth. Subglottal pressure was captured using
the Glottal Enterprises new PG100E system, which leaves a
DC signal output proportional to the pressure. The transducer was calibrated by means of the unit provided by the
manufacturer, which produces a pressure of 20 cm H2O.
The pressure signal was monitored on an oscilloscope during the recordings. The sound pressure level from the microphone is strongly influenced by the relationship between the
strongest partial and its distance to the first formant and
hence is not a reliable estimate of vocal loudness level.40,41
In this study, subglottal pressure, which is a physiologically
more relevant measure of vocal loudness compared to vocal
sound pressure level,42−44 was collected to verify the vocal
loudness levels of typical, loud, and soft conditions. Since
subglottal pressure was used to verify that the participants
performed the task of loud, soft, and typical loudness, it
was not included for statistical analyses. The experimental
set-up is depicted in Figure 3.
Prior to recordings, participants were trained (co-authors
BG & FL) to produce pressed, neutral, flow, and breathy
phonation types at self-selected comfortable pitch and three
loudness levels on /pæ:/ syllable sequences. During the
recording the participants were seated in a comfortable
exam chair in a double walled sound treated booth. The participants were provided with real-time visual feedback in
terms of flow glottogram, obtained from the Glottal Enterprises MSIF1 inverse filter. While the participant sustained
a loud /æ:/ vowel its two formant filters were tuned using a
ripple-free closed phase as the criterion. The unit’s output
was displayed on an oscilloscope screen in front of the participant, adjusted to show a time window of about 5 seconds.
During the recordings the participant was asked to view
the oscilloscope screen and to produce a flow glottogram
signal that: (i) for Neutral phonation had a minimum as
FIGURE 3. Schematic illustration of the experimental set-up.
Rita R. Patel, et al
5
Glottal Airflow and Glottal Area Waveform Characteristics
close to zero flow as possible; (ii) for Flow phonation
showed a large amplitude while keeping the minimum close
to zero flow; (iii) for Pressed phonation had as low amplitude as possible; and (iv) for Breathy phonation showed a
signal with a minimum well above zero flow. If the participants were able to produce three consecutive repetitions of
/pæ:/ at the target flow levels, they were considered sufficiently trained for the experiment (Figure 2). The criterion
of successful training based on the ability to produce three
consecutive repetitions of /pæ:/ at the target flow levels was
empirically determined since three consecutive repetitions of
/pæ:/ are required for measurement of the subglottal pressure.17,45 During the experiment, the experimenters (authors
JS, FL, and BG) verified that the participants were maintaining the effect of training by visual inspection of the oscilloscope screen for target flow levels and auditory-perceptual
judgments of the target production including the loudness
level.
In each of the four phonation types, the participants produced in typical, loud, and soft voice a string of five /pæ:/
syllables in one breath at the rate of approximately 1.5−2
syllables per second.46 Thus each syllable was approximately 0.5 seconds in duration. The experimenters confirmed whether the participants were truly producing typical
versus soft versus loud phonation by simultaneous measurement of the subglottal pressure (Figure 3). Thus, each syllable was approximately 0.5 seconds in duration. The syllable
/ pæ:/ was chosen, as subglottal pressure can be captured
from the oral pressure during the /p/ occlusion, and the
vowel /ae/ is optimal for inverse filtering due to its high and
widely separated first and second formants. In addition, it
tends to be pronounced with complete occlusion of the velopharyngeal port.47 In instances where it was impossible to
visualize the free margins of the vocal folds during phonation the vowel /æ:/, the vowel /i/ was used instead. Problems
to visualize the free margins of the vocal folds generally
occurred for pressed phonation. In the current analyses only
one data point with pressed phonation using the vowel /i:/
was included.
The following signals were simultaneously recorded on
separate tracks of the data acquisition system: high-speed
videoendoscopic, audio, electroglottograph, subglottal pressure, and airflow. Calibration signals for sound pressure
level, subglottal pressure, and flow were recorded in a separate file each day.
Overall, each participant was recruited to produce a total
of 12 conditions (four phonation types x three degrees of
vocal loudness). The participants performed the various
types of phonations in the neutral condition first, followed
by loud, and then the soft condition. The participants performed all the conditions first without the endoscope and
then with the endoscope in place. For the study only the
data with the endoscope in place was further analyzed.
Data collection was completed in about 2 hours for each
participant. Only two participants (one male and one
female) completed all phonation types across the three loudness conditions (Table 1). Producing target phonations at
TABLE 1.
Summary table of number of tokens per condition and
phonation type across male (M) female (F) participants
Phonation Type
Condition
Typical
Loud
Soft
Breathy
Flow
Neutral
Pressed
M
F
M
F
M
F
M
F
7
3
1
4
2
0
6
3
1
5
2
1
5
3
1
6
4
1
6
2
0
4
2
0
soft loudness level was challenging for the untrained participants in the study and hence soft loudness was the last condition recorded in the study. All participants first produced
neutral, then flow, then pressed, and finally breathy phonation in typical loudness level, followed by loud and then soft
loudness levels. Due to the level of difficulty of the task and
the fatiguing nature of the experiment with the flow mask
with simultaneous intraoral pressure tube and flexible
nasendoscopy in place, 11 participants completed the target
productions at only typical and loud conditions. All recordings were conducted at the Vocal Physiology and Imaging
Laboratory at the Department of Speech, Language, and
Hearing Sciences, Indiana University.
Analysis
The steady-state of the vowel appearing in the middle three
syllables was selected for analysis of vocal fold vibrations
and inverse filtering. Segmentation and analysis of the highspeed recordings along with the acoustic signal was conducted using an automated endoscopic imaging analysis
tool called the Glottis Analysis Tool.48 GA waveforms were
derived from the high-speed videoendoscopy data. The following parameters were derived for each of five subsequent
cycles and averaged: GA, pixels, ALR, Relative CV, Speed
quotient (SQ), GA index (GAI), Closing quotient (CQ),
Maximum area declination rate (MADR), Amplitude quotient (AQ), Closing Duration (CDarea [ms]) and Stiffness
index (SI) (Table 2).
The flow signal was analyzed using the Sopran software
(Svante Granqvist, available at www.Tolvan.com). As a
first step, the flow signal was calibrated in mL/s using the
integral of the signal produced by 140 mL of air driven
through the Rothenberg mask from a syringe, and multiplying the average of that signal with its duration in seconds.
Average flow (mL/s) values were calculated from the calibrated flow signal.
Inverse filtering of the flow signal was performed after
resampling the recordings to 16 kHz and specifying the
microphone distance. The tuning of the formant filters, on
average one per kHz, was made manually (co-author JS)
using the “Inverse filter” module of the Sopran software.
This module displays the flow glottogram and the voice
source spectrum simultaneously in separate windows. For
6
Journal of Voice, Vol. &&, No. &&, 2020
TABLE 2.
Description of the dependent variables derived from
high-speed videoendoscopy
High-Speed
Videoendoscopy
Parameters
Definition
Glottal area (GA)
Glottal area in pixels
from the visible glottal
contour
Ratio between changing
glottal area and the
glottal length
Maximum speed with
which the area of the
glottis would be changing during opening or
closing of the glottis
movement is approximated as a sine curve
Ratio of the opening and
closing duration of a
glottal cycle
Ratio between the minimum and maximum
glottal opening
The portion of the time
during a glottal cycle
when the glottis is
closing
Maximum closing
velocity
Maximum closing velocity normalized with the
amplitude of the GAW
Time in ms during which
the vocal folds are
closing
Maximum speed with
which the area of the
glottis is changing during opening or closing
normalized to the total
varying glottal area
Amplitude-to-length ratio
(ALR)
Relative peak closing
velocity (PCV)
Speed quotient (SQ)
Glottal area index (GAI)
Closing quotient (CQ)
Maximum area declination rate (MADR)
Amplitude quotient (AQ)
Closing duration (CDarea
[ms])
Stiffness index (SI)
tuning the filters two criteria were used, a ripple-free closed
phase and a source spectrum envelope as free from dips and
peaks near the formants as possible. After completing the
tuning of the inverse filters, the resulting flow glottograms
were saved in a separate channel of the recording and the
frequencies and bandwidths of the inverse filters were saved
into a log file. Subsequently, the flow glottograms were analyzed using the Glottal flow parameter measurement module of the Sopran software. Three flow glottogram
properties were manually marked, (i) the period, (ii) the
closed phase, and (iii) the mean airflow during the quasiclosed phase. The following flow glottogram parameters
were then automatically computed and recorded into the
same log file, which was then exported to an excel file: Flow
pulse amplitude (PAflow mL/s), Maximum Flow Declination
Rate (MFDRflow mL/s2, Amplitude Quotient (AQflow ms),
Closed Quotient (CQflow), Level difference between the first
and second harmonics (H1H2flow), Duration of Closing
Phase (CDflow (ms)), Speed Quotient (SQflow), and Glottal
Resistance (GRflow) (Figure 4). The definitions of parameters extracted from flow glottograms are given in Table 3.
Using the Sopran software, the subglottal pressure was
measured as the plateau of the oral pressure signal during
the /p/ occlusion for the middle of the five syllables.17 A total
of 69 tokens (Table 1) were subjected to statistical analysis
for males and females across condition (typical, loud, and
soft) and phonation types (breathy, flow, neutral, and
pressed) after verification of expected flow values (Figure 2)
from a total of 258 tokens collected from 13 participants.
As mentioned, the participants viewed the flow glottogram of their voice during the experimental recordings. The
participants were instructed to produce a large pulse amplitude in flow and breathy phonation and small pulse amplitude in pressed phonation. For breathy phonation they were
asked to produce a signal well above zero flow (Figure 2).
Qualitative observation from Figure 5 reveals that the pulse
amplitude was greatest in flow and breathy phonation and
smallest in pressed phonation. The two middle panels show
the corresponding mean values for subglottal pressure. On
qualitative observation the subglottal pressure was generally
higher in loud than in typical, and male participants used
higher pressures than the female participants. The right panels show the corresponding averages of glottal resistance. As
expected, they were lowest in breathy and highest in pressed
on qualitative observation.
Statistical analysis
A two-way Analysis of Variance was used to statistically
evaluate the effect of phonation type (breathy, flow, neutral,
pressed) and condition (typical, loud, and soft) on the
dependent variables derived from inverse filtering and highspeed videoendoscopy within males and females respectively. Simple effects tests using the least square method
were conducted to test the interactions. For question 2,
regarding the relationship between total flow and GA, Pearson’s correlation coefficients were computed and evaluated
for statistical significance using Analysis of Variance.
Results were considered significant for p ≤ 0.5. Statistical
analysis was performed using SAS 9.4 (SAS Institute Inc.,
Cary, NC).
RESULTS
I. Changes in GA waveform and airflow parameters as a
function of breathy, flow neutral, and pressed phonation types during typical loudness:
(a) Male: The relative PCV, derived from the GA
waveform, was the only parameter that systematically varied significantly across all combinations of
Rita R. Patel, et al
Glottal Airflow and Glottal Area Waveform Characteristics
7
FIGURE 4. Measures of the flow glottogram, its derivative (upper and lower left graphs) and its spectrum (right panel).
TABLE 3.
Description of the dependent variables derived from the inverse filtered flow glottograms
Inverse Filtered Parameters
Definition
Flow pulse amplitude (PAflow mL/s)
Maximum flow declination rate (MFDRflow (mL/s2)
Amplitude quotient (AQflow)
AC amplitude of the flow pulse
Negative peak amplitude of the derivative of the flow glottogram
Ratio between PAif and MFDRif also referred to as ‘Td’ by Fant
(1995).61
The portion of the time during a glottal cycle when the glottis is
closed or nearly closed according to the inverse filtered waveform
Level difference between first and second partial of the voice source
spectrum
Duration of the decreasing part of the flow pulse
Ratio between the durations of the increasing and decreasing part of
the flow pulse
Ratio of subglottal pressure to glottal airflow
Closed quotient (CQflow)
H1H2flow
Duration of closing phase (CDflow (ms))
Speed quotient (SQflow)
Glottal resistance (GRflow)
phonation types of breathy, flow, neutral, and
pressed during typical loudness (F = 0.001 [3, 27],
p < 0.0001) (Figure 6). The PCV was large for
breathy, followed by neutral, flow, and pressed. GA
(F = 4.23 [3, 27], p = 0.014) was significantly larger
in breathy (M = 473.93 § 286.52) compared to the
neutral (M = 180.41 § 143.40, t [27] = 2.90,
p = 0.007), flow (M = 176.31 § 92.26, t [27] = 3.09,
p = 0.005), and pressed (M = 85.74 § 78.23,
t [27] = 4.03, p < 0.001) phonation types. AQ was
significantly reduced in breathy (M = −3.74 § 0.81)
compared to the flow (M = −2.87 § 0.45,
t [27] = −2.87, p = 0.008), neutral (M = −2.82 §
0.55, t [27] = −2.87, p = 0.008), and pressed
(M = −2.48 § 0.57, t [27] = −4.16, p < 0.001) phonation types (Table 4). As hypothesized the ALR
was not statistically significant across phonation
types (F = 0.10 [3, 27], p = 0.96).
The AQflow was significantly greater in breathy
(M = 1.26 § 0.29) compared to the pressed
(M = 0.89 § 0.25, t [27] = 3.67, p = 0.013), neutral
(M = 0.83 § 0.27, t [27] = 2.76, p = 0.010), and flow
(M = 0.78 § 0.19, t [27] = 3.67, p = 0.001) phonation types. The PAflow (F = 3.00 [3, 27], p = 0.05)
and MFDRflow (F = 2.44 [3, 27], p = 0.86) were not
statistically significant across phonation types as
hypothesized.
Female: The only parameter that was statistically
significant across all combinations of phonation
types during typical loudness was the relative PCV
derived from the GA waveform (F = 0.001 [3, 21],
p < 0.0001). The PCV was large for breathy,
8
Journal of Voice, Vol. &&, No. &&, 2020
FIGURE 5. Mean and standard deviations of pulse amplitude, glottal resistance, and subglottal pressure for breathy, flow, neutral, and
pressed phonation types during typical and loud voice produced by males(upper panels) and females (lower panels).
FIGURE 6. Mean and standard deviations of maximum flow declination rate derived from inverse filtering, and relative peak closing
velocity, amplitude-to-length ratio from high-speed video for breathy, flow, neutral, and pressed phonation types during typical and loud
voice produced by males (upper panels) and females (lower panels).
followed by flow, neutral, and pressed phonations
(Table 5). The MADR was significantly increased
in pressed (M = −37.09 § 16.37) compared to the
flow (M = −124.84 § 45.53, p = 0.024) and breathy
(M = −128.73 § 86.84, p = 0.025) phonation types.
The GA (F = 2.93 [3, 21], p = 0.05) was significantly
increased in breathy (M = 241.76 § 119.87) compared to the neutral (M = 107.09 § 69.86,
Rita R. Patel, et al
9
Glottal Airflow and Glottal Area Waveform Characteristics
TABLE 4.
Mean and standard deviations of the dependent variables for males across typical loudness
Condition
Typical
Phonation
Breathy
Flow
Neutral
Pressed
Variable
N
Mean
Std Dev
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow (mL/s2)
Amplitude quotientflow
Duration closing phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow (mL/s2)
Amplitude quotientflow
Duration closing phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow (mL/s2)
Amplitude quotientflow
Duration closing phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow (mL/s2)
Amplitude quotientflow
Duration closing phaseflow (ms)
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
0.29
1.28
15.36
170.75
146395.57
1.26
2.32
0.04
473.93
309042.13
1.75
0.83
10.31
0.36
−190.37
−3.74
0.32
2.71
0.38
1.85
9.64
200.58
270708.33
0.78
1.57
0.11
176.31
200315.19
1.89
1.31
10.14
0.30
−159.23
−2.87
0.37
2.23
0.43
1.77
9.45
142.81
201367.6
0.83
1.68
0.11
180.41
200450.85
2.26
1.07
10.44
0.32
−162.40
−2.82
0.38
2.60
0.54
1.25
6.67
80.65
99549.83
0.89
1.68
0.12
0.33
3.87
135.71
129746.64
0.29
0.61
0.02
286.52
326452.24
0.62
0.30
6.81
0.07
185.17
0.81
0.06
0.73
0.14
0.42
3.99
88.61
144164.02
0.19
0.24
0.05
92.26
96002.75
0.51
0.83
3.69
0.10
49.81
0.45
0.07
0.83
0.13
0.26
3.53
82.09
172512.9
0.27
0.28
0.07
143.40
150595.29
1.12
0.22
3.20
0.12
82.44
0.55
0.06
1.03
0.12
0.52
2.15
40.27
64915.65
0.25
0.63
Minimum
0.17
0.92
8
80
58833
0.80
1.52
0.02
49.41
60543.62
1.27
0.59
6.81
0.25
−584.00
−4.85
0.25
1.50
0.25
1.37
4.7
113
247710
0.46
1.22
0.07
77.97
140586.27
1.30
0.83
6.86
0.15
−255.00
−3.41
0.31
1.20
0.28
1.52
5.2
93
98668
0.57
1.36
0.05
47.03
49825.25
1.34
0.92
7.25
0.14
−273.80
−3.56
0.31
1.10
0.40
0.64
4.1
38
141100
0.56
0.90
Maximum
0.51
1.76
20.1
464
424250
1.62
3.22
0.06
890.39
1019928.53
3.07
1.32
25.69
0.44
−52.20
−2.33
0.43
3.60
0.59
2.54
15.2
370
558340
0.97
1.87
1
347.64
394956.89
2.67
2.99
16.25
0.44
−108.70
−2.32
0.49
3.40
0.56
2.13
13
288
507240
1.21
2
0.19
379.11
443605.77
4.13
1.46
14.93
0.44
−54.80
−2.23
0.45
3.78
0.70
2.07
9.6
148
83388
1.20
2.72
(Continued)
10
Journal of Voice, Vol. &&, No. &&, 2020
TABLE 4. (Continued )
Condition
Phonation
Variable
N
Mean
Std Dev
Minimum
Maximum
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
5
6
6
5
6
6
6
6
6
6
6
0.35
85.74
73350.57
2.19
2.03
7.11
0.21
−75.07
−2.48
0.43
1.63
0.22
78.23
17243.26
0.46
1.60
2.96
0.11
18.10
0.57
0.08
0.88
0.24
20.18
52227.48
1.53
0.43
2.53
0.08
−105.30
−3.52
0.30
0.50
0.72
232.27
98428.87
2.73
4.80
11.13
0.31
−51.60
−2.01
0.50
2.50
TABLE 5.
Mean and standard deviations of the dependent variables for females across typical loudness
Condition
Phonation
Variable
N
Mean
Std Dev
Minimum
Maximum
Typical
Breathy
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow (mL/s2)
Amplitude quotientflow
Duration closing phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow (mL/s2)
Amplitude quotientflow
Duration closing phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow (mL/s2)
Amplitude quotientflow
Duration closing phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
4
4
4
4
4
4
4
4
4
4
3
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
5
0.28
1.02
19.76
71.0
100850
0.71
1.31
0.11
241.76
256696.42
1.69
0.70
8.89
0.36
−128.73
−2.41
0.45
1.30
0.24
1.70
19.93
94.4
152963
0.63
1.11
0.27
147.78
242709.97
1.75
0.97
8.39
0.38
−124.84
−2.40
0.43
1.41
0.30
1.42
17.43
83.6
1277335.9
0.65
1.18
0.19
107.09
153484.30
1.83
0.06
0.45
3.9
14.2
12593
0.12
0.21
0.08
119.87
141434.33
0.20
0.35
6.38
0.04
86.84
0.28
0.05
0.18
0.04
0.21
5.02
38.0
59636.63
0.06
0.16
0.18
38.17
53924.28
0.38
0.07
2.59
0.06
45.53
0.15
0.03
0.17
0.09
0.37
7.31
56.1
70807.14
0.15
0.12
0.15
69.86
94551.70
0.70
0.22
0.52
17.9
55
91224
0.59
1.11
0.03
128.33
79944.56
1.47
0.28
1.91
0.30
−242.50
−2.66
0.38
1.03
0.22
1.46
13.9
55
78060
0.59
0.90
0.10
94.81
191367.87
1.15
0.84
4.88
0.33
−203.90
−2.51
0.40
1.29
0.22
1.01
11.5
31
44822
0.36
1.05
0.07
40.18
63819.49
1.28
0.37
1.47
24
87
112190
0.84
1.59
0.23
409.65
424723.38
1.83
1.00
17.03
0.41
−31.40
−2.04
0.50
1.39
0.30
1.90
26.4
132
206030
0.71
1.29
0.53
188.93
334509.80
2.07
1.03
11.44
0.47
−86.20
−2.15
0.47
1.60
0.43
1.98
31.2
19.6
249770
0.70
1.37
0.45
199.79
317772.10
3.00
Flow
Neutral
(Continued)
Rita R. Patel, et al
11
Glottal Airflow and Glottal Area Waveform Characteristics
TABLE 5. (Continued )
Condition
Phonation
Pressed
Variable
N
Mean
Std Dev
Minimum
Maximum
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow (mL/s2)
Amplitude quotientflow
Duration closing phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
6
6
6
6
6
6
6
4
4
4
4
4
4
4
4
4
4
3
4
4
4
4
4
4
4
1.01
7.13
0.30
−79.72
−4.72
0.45
1.20
0.51
1.27
6.14
72.92
197717.75
0.43
0.99
0.48
79.70
62812.94
2.03
1.20
5.11
0.31
−37.09
−2.67
0.43
1.45
0.38
2.94
0.07
46.32
5.88
0.04
0.36
0.13
0.64
1.89
52.48
132378.85
0.21
0.22
0.43
60.33
19663.94
0.94
1.16
0.96
0.16
16.37
0.51
0.07
0.71
0.29
2.84
0.20
−160.60
−16.70
0.41
0.69
0.37
0.50
4.8
12
23151
0.21
0.06
0.08
18.94
38376.62
1.08
0.54
4.45
0.10
−54.80
−3.22
0.35
0.40
1.39
11.01
0.41
−22.28
−2.02
0.50
1.62
0.66
2.05
8.9
140
345060
0.67
0.13
1.09
161.78
83473.95
2.96
2.93
6.54
0.49
−19.00
−2.00
0.50
1.88
t [21] = 2.94, p = 0.008) and pressed (M = 79.90 §
60.33, t [21] = 3.24, p = 0.004) phonation types. The
ALR was not statistically significant across phonation types as hypothesized (F = 1.23 [3, 21],
p = 0.32).
The AQflow (F = 5.26 [3, 21], p = 0.007) was significantly greater in breathy (M = 0.71 § 0.12) compared
to the flow (M = 0.63 § 0.06, t [21] = 2.11, p = 0.047)
phonation type. The CQflow (F = 3.20 [3, 21],
p = 0.044) was significantly greater in pressed
(M = 0.51 § 0.13) compared to neutral (M = 0.30 §
0.09, t [21] = −3.38, p = 0.003), breathy (M = 0.28 §
0.06, t [21] = −3.83, p < 0.001), and flow (M = 0.24 §
0.04, t [21] = −3.23, p = 0.004) phonation types. As
hypothesized the PAflow (F = 0.83 [3, 21], p = 0.49)
and MFDRflow (F = 0.08 [3, 21], p = 0.90) were not
statistically significant across phonation types.
II. Changes in GA waveform and airflow parameters as a
function of breathy, flow neutral, and pressed phonation types during loud condition:
(a) Male: The only parameter that was statistically significant across all combinations of phonation types
during loud phonation was the relative PCV
derived from the GA waveform (F = 0.001 [3, 27],
p < 0.0001) . For the loud condition the PCV was
large for breathy, followed by flow, pressed, and
neutral phonations. The CDflow was significantly
greater in breathy (M = 2.01 § 0.43) compared to
the flow (M = 1.93 § 0.27, t [27] = 2.60, p = 0.015),
pressed (M = 1.76 § 0.58, t [27] = 2.32, p = 0.028),
and neutral (M = 1.70 § 0.42, t [27] = 2.60,
p = 0.015), phonation types (Table 6).
(b) Female: The only parameter that was statistically
significant across all combinations of phonation
types during loud phonation was the relative PCV
derived from the GA waveform (F = 0.001 [3, 21],
p < 0.0001). The PCV was large for breathy followed by flow, neutral, and pressed phonations.
The AQflow (F = 5.26 [3, 21], p = 0.007) was significantly greater in pressed (M = 0.64 § 0.15) compared to flow (M = 0.62 § 0.10, t [21] = −3.38,
p = 0.007), and neutral (M = 0.56 § 0.16,
t [21] = −3.71, p = 0.001) and lower compared to
breathy (M = 0.71 § 0.14, t [21] = −2.38,
p = 0.027) phonation types. The CDflow (F = 3.62
[3, 21], p = 0.030) was significantly greater in
pressed (M = 1.17 § 0.43) compared to the flow
(M = 1.14 § 0.03, t [21] = −2.93, p = 0.008),
breathy (M = 1.06 § 0.02, t [21] = −3.15,
p = 0.005), and neutral (M = 1.03 § 0.10,
t [21] = −4.08, p < 0.001) phonation types. The
GRflow (F = 4.53 [3, 21], p = 0.013) was significantly
greater in pressed (M = 0.59 § 0.58) compared to
the neutral (M = 0.24 § 0.14, t [21] = −2.46,
p = 0.023), flow (M = 0.22 § 0.16, t [21] = −2.83,
p = 0.01), and breathy (M = 0.10 § 0.06,
t [21] = −3.11, p = 0.005) phonation types. The
SQflow (F = 2.91 [4, 21], p = 0.046] was significantly
greater in neutral (M = 1.35 § 0.29, t [21] = −3.03,
p = 0.006) compared to pressed (M = 1.20 § 0.42,
t [21] = 3.66, p = 0.002), and breathy (M = 0.80 §
0.34, t [21] = −2.48, p = 0.022), phonation types
and lower compared to flow (M = 1.42 § 0.26,
t [21] = −3.03, p = 0.006) (Table 7).
12
Journal of Voice, Vol. &&, No. &&, 2020
TABLE 6.
Mean and standard deviations of the dependent variables for males across loud condition
Condition
Phonation
Variable
N
Mean
Std Dev
Minimum
Maximum
Loud
Breathy
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow (mL/s2)
Amplitude quotientflow
Duration closing phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow (mL/s2)
Amplitude quotientflow
Duration closing phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow (mL/s2)
Amplitude quotientflow
Duration closing phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow (mL/s2)
Amplitude quotientflow
Duration closing phaseflow (ms)
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
2
2
2
2
2
2
2
031
1.35
14.60
178.19
173529.33
1.12
2.01
0.12
398.50
343514.20
1.82
0.96
11.89
0.34
−214.20
−3.23
0.35
2.29
0.34
1.37
11.17
203.03
290326.67
0.76
1.93
0.08
222.00
265175.95
2.17
1.15
11.01
0.29
−211.53
−2.66
0.38
1.96
0.38
1.63
10.59
143.75
235866.67
0.64
1.70
0.08
145.13
157838.52
2.02
1.02
9.11
0.32
−135.23
−2.68
0.38
2.29
0.38
1.44
10.16
121.58
136808.5
0.91
1.76
0.03
0.26
2.67
108.68
121508.90
0.20
0.43
0.07
267.08
262653.84
0.56
0.47
7.21
0.12
145.84
0.20
0.05
0.84
0.08
0.31
1.24
7.78
104239.13
0.25
0.27
0.02
82.63
98031.62
0.08
0.31
3.25
0.06
47.01
0.20
0.02
0.21
0.11
0.10
1.57
20.17
54415.84
0.19
0.42
0.08
26.88
28402.13
0.15
0.16
1.41
0.03
16.22
0.24
0.03
0.20
0.17
0.10
1.75
47.71
64433.691
0.08
0.58
0.28
1.16
11.7
64
47198
0.97
1.69
0.04
165.97
58949.73
1.33
0.60
4.20
0.20
−303.30
−3.35
0.30
1.40
0.28
1.02
10
194
217860
0.47
1.69
0.07
149.44
152053.08
2.07
0.92
8.92
0.22
−240.70
−2.88
0.36
1.79
0.27
1.56
9.6
123
191410
0.42
1.23
0.07
116.24
128426.46
1.87
0.90
8.13
0.28
−150.20
−2.85
0.35
2.10
0.25
1.37
8.9
88
91247
0.85
1.36
0.34
1.64
16.9
280
289560
1.35
2.5
0.17
690.21
576649.80
2.43
1.49
18.48
0.44
−45.90
−3.01
0.40
3.08
0.43
1.63
12.5
206
409790
0.96
2.23
0.11
311.94
325288.56
2.22
1.50
14.76
0.33
−157.30
−2.48
0.40
2.20
0.50
1.75
12.4
164
296550
0.75
2.03
0.16
169.40
185109.54
2.16
1.20
10.73
0.35
−118.00
−2.41
0.42
2.49
0.50
1.50
11.4
155
182370
0.96
2.17
Flow
Neutral
Pressed
(Continued)
ARTICLE IN PRESS
Rita R. Patel, et al
13
Glottal Airflow and Glottal Area Waveform Characteristics
TABLE 6. (Continued )
Condition
Phonation
Variable
N
Mean
Std Dev
Minimum
Maximum
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude 1uotient
Stiffness index
Closing durationarea (ms)
2
2
2
2
2
2
2
2
2
2
2
0.08
127.04
173207.86
1.39
1.89
16.03
0.25
−145.80
−2.55
0.43
1.70
0.16
55.80
8445.06
0.16
0.81
4.11
0.12
8.49
0.03
0.03
0.85
0.16
87.59
167236.30
1.28
1.31
13.13
0.17
−151.80
−2.57
0.41
1.10
0.39
166.50
179179.41
1.50
2.46
18.94
0.34
−139.80
−2.53
0.45
2.30
TABLE 7.
Mean and standard deviations of the dependent variables for females across loud condition
Condition
Phonation
Variable
N
Mean
Std Dev
Minimum
Maximum
Loud
Breathy
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow
Amplitude quotientflow
Duration closing phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow
Amplitude quotientflow
Duration closing phaseflow
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
Closed quotientflow
Speed quotientflow
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
4
4
0.33
0.80
13.99
73.99
106064.5
0.71
1.06
0.10
111.83
226795.95
4.42
1.16
10.34
0.26
−102.98
−2.30
0.48
1.09
0.32
1.42
17.86
137.97
210545
0.62
1.14
0.22
61.21
134382.37
1.70
1.77
9.13
0.22
−78.55
−2.27
0.47
0.89
0.37
1.35
0.08
0.34
4.97
0.06
20174.46
0.14
0.02
0.06
115.33
264664.82
4.60
0.17
2.41
0.16
86.59
0.37
0.02
0.14
0.12
0.26
10.54
115.61
152289.59
0.10
0.03
0.16
40.93
78423.23
0.57
0.47
5.00
0.02
53.39
0.34
0.04
0.15
0.18
0.29
0.28
0.56
10.5
73.95
91799
0.62
0.87
0.06
30.28
39649.66
1.17
1.03
8.63
0.15
−164.20
−2.56
0.47
0.99
0.23
1.24
10.4
56
102860
0.55
1.14
0.11
32.27
78928.78
1.30
1.44
5.59
0.21
−116.30
−2.51
0.44
0.79
0.19
0.92
0.39
1.04
17.5
74
120330
0.81
1.08
0.15
193.38
413942.23
7.67
1.28
12.04
0.37
−41.75
−2.04
0.49
1.19
0.40
1.61
25.3
220
318230
0.69
1.16
0.33
90.16
189835.97
2.10
2.10
12.66
0.24
−40.80
−2.02
0.50
1.00
0.56
1.52
Flow
Neutral
(Continued)
ARTICLE IN PRESS
14
Journal of Voice, Vol. &&, No. &&, 2020
TABLE 7. (Continued )
Condition
Phonation
Pressed
Variable
N
Mean
Std Dev
Minimum
Maximum
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow
Amplitude quotientflow
Duration closing phaseflow
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow declination rateflow
Amplitude quotientflow
Duration closing phaseflow
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area declination rate
Amplitude quotient
Stiffness index
Closing durationarea (ms)
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
14.99
93.25
182087.5
0.56
1.03
0.24
51.67
122583.53
2.13
1.88
6.63
0.21
−75.10
−2.02
0.50
0.85
0.50
1.20
7.02
87.52
156200
0.64
1.17
0.59
43.70
84421.58
4.22
1.22
6.96
0.19
−62.70
−2.06
0.49
0.84
7.13
40.75
104910.42
0.16
0.10
0.14
41.22
79693.15
0.93
0.73
2.65
0.09
45.97
0.02
0.00
0.31
0.05
0.42
0.71
88.64
173962.41
0.15
0.43
0.58
32.63
43071.28
2.10
0.01
5.90
0.07
40.87
0.07
0.02
0.21
6.6
46
78620
0.39
0.92
0.12
20.67
54458.08
1.50
1.19
3.58
0.14
−140.40
−2.05
0.50
0.60
0.46
0.91
6.5
25
33190
0.54
1.05
0.18
20.62
53965.58
2.73
1.21
2.79
0.14
−91.60
−2.11
0.48
0.69
21.9
144
289460
0.78
1.14
0.44
111.32
230572.95
3.50
2.88
10.03
0.33
−42.80
−2.01
0.50
1.29
0.51
1.50
7.5
150
279210
0.75
1.47
0.99
66.77
114877.57
5.70
1.23
11.13
0.24
−33.80
−2.01
0.50
0.99
III. Changes in GA waveform and airflow parameters as a
function of breathy, flow neutral, and pressed phonation types during soft condition:
(a) Male: The only parameter that was statistically
significant across all combinations of phonation
types during soft loudness was the relative PCV
derived from the GA waveform (F = 0.001 [3, 27],
p < 0.0001). For soft condition the PCV was largest
for flow phonation followed by breathy, and neutral
phonations (Table 8). The PAflow (F = 3.00 [3, 27],
p = 0.048) was significantly greater in flow phonation (212 mL/s) compared to neutral (89 mL/s,
t [27] = 2.15, p = 0.041).
(b) Female: The only parameter that was statistically
significant across all combinations of phonation
types during soft loudness was the relative PCV
derived from the GA waveform between flow and
neutral (F = 0.001 [3, 21], p < 0.0001) (Table 9).
IV. Relationship between GA and flow:
An important aspect of the results and our second study
question is to what extent the High-Speed Videoendoscopy
(HSV) and the flow data agree. There is a significant positive relationship between the amplitude quotient from
inverse filtering and HSV, r (41) = 0.65, p < 0.001 (Figure 7).
There was also a statistically significant positive relationship between the HSV closing phase and the flow glottogram closing phase r (41) = 0.76, p < 0.001 and between
the entire GA and the total glottal flow r (41) = 0.47,
p = 0.001. In most cases the duration of glottal closing on
HSV is longer than the duration of glottal flow decrease
(Figure 8). The Figure 9 shows the relationship between
GA and total glottal flow. A contributing factor to the
low value of the determination coefficient is that the flow
to area relationship must depend on both subglottal pressure and phonation type; a given subglottal pressure will
produce a greater flow through a wide than through a narrow GA. In addition, female adults have shorter vocal
ARTICLE IN PRESS
Rita R. Patel, et al
15
Glottal Airflow and Glottal Area Waveform Characteristics
TABLE 8.
Mean and standard deviations of the dependent variables for males across soft loudness
TABLE 9.
Mean and standard deviations of the dependent variables for females across soft loudness
Condition
Phonation
Variable
N
Mean
Condition
Phonation
Variable
N
Mean
Soft
Breathy
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse
amplitudeflow (mL/s)
Maximum flow
declination rateflow
(mL/s2)
Amplitude quotientflow
Duration closing
phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area
declination rate
Amplitude quotient
Stiffness index
Closing
durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse
amplitudeflow (mL/s)
Maximum flow
declination rateflow
(mL/s2)
Amplitude quotientflow
Duration closing
phaseflow
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area
declination rate
Amplitude quotient
Stiffness index
Closing
durationarea (ms)
Closed quotientflow
Speed quotientflow
H1H2flow
Flow pulse
amplitudeflow (mL/s)
Maximum flow
declination rateflow
(mL/s2)
Amplitude quotientflow
Duration closing
phaseflow (ms)
Glottal resistanceflow
Glottal area
Peak closing velocity
Speed quotient
Glottal area index
Amplitude-length-ratio
Closing quotient
Maximum area
declination rate
Amplitude quotient
Stiffness index
Closing
durationarea (ms)
1
1
1
1
0.21
1.56
19.9
106
Soft
Flow
1
0.42
1
79219
1
1
1
1.71
13.0
214
1
1
1.34
2.04
1
278970
1
1
1
1
1
1
1
1
0.04
591.80
341287.44
1.90
0.74
11.38
0.35
–211.50
1
0.77
1
1.05
1
0.06
1
1
78.86
169269.01
1
1
1
–3.39
0.30
2.30
1
1
1
1
0.22
1.72
19.6
212
1
1
1
1.40
2.10
10.84
1
1
0.20
–115.60
1
148610
1
–2.34
1
1
1.43
1.96
1
1
0.44
1.00
1
1
1
1
1
1
1
1
0.08
462.48
424190.96
1.13
0.90
13.47
0.47
–226.30
1
0.23
1
1
1
1.57
16.4
193
1
196440
1
1
1
–4.06
0.29
3.20
1
0.98
1
1
1
1
0.26
1.72
14.5
89
1
1.51
1
0.05
1
81757
1
1
27.30
72199.51
1
1
1.09
1.87
1
1
1
1
1
1
1
1
0.08
292.31
319310.10
1.29
1.05
13.67
0.42
–193.90
0
1
1
3.05
7.72
1
1
0.04
–58.50
1
–2.00
1
1
1
–3.62
0.28
2.90
Closed
quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow
declination
rateflow (mL/s2)
Amplitude
quotientflow
Duration closing
phaseflow (ms)
Glottal
resistanceflow
Glottal area
Peak closing
velocity
Speed quotient
Glottal area index
Amplitude-lengthratio
Closing quotient
Maximum area
declination rate
Amplitude
quotient
Stiffness index
Closing durationarea (ms)
Closed
quotientflow
Speed quotientflow
H1H2flow
Flow pulse amplitudeflow (mL/s)
Maximum flow
declination
rateflow (mL/s2)
Amplitude
quotientflow
Duration closing
phaseflow (ms)
Glottal
resistanceflow
Glottal area
Peak closing
velocity
Speed quotient
Glottal area index
Amplitude-lengthratio
Closing quotient
Maximum area
declination rate
Amplitude
quotient
Stiffness index
Closing durationarea (ms)
1
1
0.50
0.19
Flow
Neutral
Neutral
ARTICLE IN PRESS
16
Journal of Voice, Vol. &&, No. &&, 2020
FIGURE 7. Amplitude quotient derived from the high-speed video data (x-axis) as a function of the associated values derived from the flow
glottogram (y-axis), produced in the four phonation types for typical vocal loudness. Open and filled symbols refer to female and male participants, respectively respectively. The dashed line and the equation refer to the trendline across all participants and phonation types.
FIGURE 8. Duration of the closing phase (ms) from the high-speed video data (x-axis) as function of the duration of the closing phase
derived (ms) from the flow glottograms (y-axis) produced in the four phonation types for typical vocal loudness. Open and filled symbols
refer to female and male participants respectively. The dashed line and the equation refer to the trendline across all participants and phonation types.
ARTICLE IN PRESS
Rita R. Patel, et al
Glottal Airflow and Glottal Area Waveform Characteristics
17
FIGURE 9. Glottal area in pixels (x-axis) from the high-speed video data as a function of the total flow from the flow glottograms (mL/s)
(y-axis) produced in the four phonation types for typical vocal loudness. Open and filled symbols refer to female and male participants
respectively. The dashed line and the equation refer to the trendline across all participants and phonation types.
folds than male adults. This should be the reason why
the female subjects showed lower flow and area values
than the male subjects. There was a statistically significant
positive relationship between the relative PCV and the
ALR r (41) = 0.84, p < 0.001, both derived from HSV
(Figure 10).
DISCUSSION
The goal of this study was to quantify the most salient
source characteristics of flow phonation in males and
females. A comprehensive set of source measurements were
evaluated from simultaneous high-speed video endoscopy,
inverse filtering, subglottal pressure, acoustic recordings,
and electroglottography. In this study, analysis of electroglottography was not conducted as the main goal was to
evaluate source characteristics from direct examination of
the vocal folds against flow glottogram characteristics of
flow phonation. Young, vocally healthy adult males and
females performed a series of the syllables /pæ:/ in typical
(conversational), loud, and soft phonation in each of the
phonation types breathy, flow, neutral, and pressed.
Source characteristics of flow phonation from highspeed videoendoscopy
Systematic changes in shape of the GA waveform during
various phonation types resulted in changes in the relative
PCV derived from high-speed videoendoscopy for both
male and female participants across typical, loud, and soft
phonations. For females, as expected, the relative PCV systematically decreased from breathy to flow to neutral and
pressed phonation types for typical and loud condtions
(Figure 6). The PCV also decreased from flow to neutral
phonation during the soft condition as expected. Systematic
changes in PCV suggest that the GA is changing quickly
during the closing phase for breathy compared to pressed
phonations.
Contrary to the female participants, the PCV was large
for breathy, followed by neutral, flow, and pressed phonation for typical loudness for male participants. For the loud
condition, the relative PCV was large for breathy followed
by flow, pressed, and neutral. The difference between neutral and flow in the typical condition (Table 4) and between
pressed and neutral in the loud condition was small
(Table 6). Further empirical studies are required to examine
whether or not the trend of the PCV observed in this study
across phonation types holds true with a larger number of
male participants. For the soft condition, the PCV was large
for flow followed by breathy and neutral phonation types.
Prior studies have investigated variations in relative PCV to
analyze changes during modal, low, and high pitch conditions, and soft, medium, and loud vocal loudness conditions
using stroboscopy.49 Using high-speed videoendoscopy
Patel et al (2015)50 investigated normalized mid-membranous closing velocity to evaluate changes in speed of vocal
fold closure during typical phonation across healthy male,
ARTICLE IN PRESS
18
Journal of Voice, Vol. &&, No. &&, 2020
FIGURE 10. Amplitude-to-length ratio (x-axis) as a function of relative peak closing velocity (pixels/s) (y-axis) produced in the phonation
modes represented by symbols at typical degree of vocal loudness. Open and filled symbols refer to female and male participants respectively. The dashed line and the equation refer to the trendline calculated across all participants and phonation types.
females, and children.50 Children had the highest normalized closing velocity followed by males, and females. In the
current study overall the relative PCV was smaller for
females compared to males; however, this was not subjected
to statistical analysis. Normalized closing velocity has also
been used to evaluate the changes in vibratory characteristics due to impact stress in vocal nodules.37,51 Findings of
this study suggest that the relative PCV derived from the
GA waveform was useful for distinguishing the various phonation types of breathy, flow, neutral, and pressed; however,
there appears to be gender differences and a complex relationship in how the relative PCV is varied across vocal loudness conditions during breathy, flow, neutral, and pressed
phonation types for males. According to previous studies,
this could be caused by the vertical phase difference between
the upper and lower margins of the vocal folds,28,32,33 where
the convergence angle is widest for breathy voice and smallest for pressed voice.
Contrary to our expectation, statistical difference was not
observed across changes in vibratory amplitude (ampltiudeto-length ratio) across phonation types in males or females
for various vocal loudness levels. The vibratory amplitude
however, systematically reduced from breathy to flow to
neutral and pressed in females for typical and loud vocal
loudness levels, although it was not statistically significant.
For the soft condition, the amplitude was larger for flow
phonation compared to neutral. For male participants,
changes in phonation type resulted in expected changes
across typical phonation; however, the relative vibratory
amplitude was marginally larger for neutral (10.44) than
flow phonation (10.14). Contrary to our hypothesis, the
loud condition resulted in the largest vibratory amplitude
for pressed, followed by breathy, flow, and neutral phonations in males. For the soft condition, vibratory amplitude was large for neutral (13.67), followed by flow
(13.47), and breathy phonation (11.38) in males. There
appears to be a gender difference in how amplitude is
varied across various loudness conditions for breathy,
flow, neutral, and pressed phonations. Figure 10 shows
an overview of the data in terms of a scatter plot of
PCV as a function of the ALR. As expected, the velocity
increases with increasing ALR. It can also be noted that
for the male participants, pressed and neutral phonation
tended to produce lower velocity values than flow and
breathy phonation compared to females.
Unexpectantly, the amplitude quotient, which reflects
the length of the glottal closing phase was largest for
breathy, followed by flow, neutral, and pressed phonation during typical loudness; however, statistical difference was observed only between breathy and the other
phonation types in males. Statistical difference was not
observed in amplitude quotient across females. Findings
suggest that male participants not only vary closing
speed but also vary the duration of the closing phase to
achieve target changes in the GA waveform for breathy,
flow, neutral, and pressed phonation. Females on the
other hand, predominantly vary closing speed as a function of phonation types.
ARTICLE IN PRESS
Rita R. Patel, et al
Glottal Airflow and Glottal Area Waveform Characteristics
Source characteristics of flow phonation from flow
glottogram
The second question that we wanted to answer with the
present study was what the relationship is between parameters derived from the glottal airflow and the GA waveform.
Contrary to our hypothesis flow pulse amplitude and the
maximum flow declination rate did not achieve statistical
significance, but varied with phonation type as expected.
The flow pulse amplitude was consistently large for flow
phonation followed by neutral, pressed, and breathy phonation types for typical and loud sound level for both females
and males, however, this was not statistically significant
possibly because being underpowered. For the males’ typical and loud conditions, MFDR was large for flow, followed by the neutral, breathy and pressed phonation types,
whereas for the females’ typical sound level condition, the
MFDR was highest in pressed followed by flow, neutral,
and breathy phonations. Flow phonation at soft loudness
levels resulted in larger flow pulse amplitudes and maximum
flow declination rate compared to neutral phonation.
Several studies have investigated the effects of varying
subglottal pressure on the voice source flow waveform in
untrained adult males and adult females,45,52,53 however
there are few studies investigating variations in flow glottograms across phonation types in females. S€
odersten et al
(1995)54 recorded videofiberstroboscopy, glottal airflow and
EGG in healthy middle-aged women during soft, middle
and loud phonation and had expert listeners rate voice samples along visual analogue scales representing breathiness,
hypofunction, and hyperfunction. A correlation of approximately 0.5 between pulse amplitude and subglottal pressure
in voices perceived as breathy and hypofunctional was
reported by S€
odersten et al (1995),55 while for voices perceived as hyperfunctional the correlation was -0.52.
For male participants, the flow pulse amplitude was consistently larger for flow, followed by breathy, neutral, and
pressed phonation types across both typical and loud vocal
loudness levels. This is in accordance with findings from one
male particpant in Sundberg (1995)3 and 6 males participants in Gauffin & Sundberg (1989)2. However, this difference failed to reach statistical significance in the present
study. For the soft vocal loudness condition, the flow pulse
amplitude was large for flow phonation followed by breathy
and neutral phonation types. The maximum flow declination rate was large for flow phonation across typical, loud,
and soft vocal loudness levels and smallest for pressed phonation across typical and loud vocal loudness levels. Overall, it appears that the flow pulse amplitude was a bit small
for flow phonation in females and large for males, whereas
the maximum area declination rate was large for flow phonation in females and males during typical condition.
The statistical analysis revealed very few significant flow
glottogram differences between the phonation modes. For
example, the inverse filtered closed quotient was statistically
larger for pressed compared to breathy, flow, and neutral
phonation during typical loudness only in females, but not
in males. This was surprising, as earlier studies have
19
observed that the closed quotient as well as several other
flow glottogram parameters differed systematically between
phonation types. Prior studies have shown MFDR, QClosed,
H1H2 and NAQ to have significant correlations with expert
listeners’ ratings of degree of phonatory pressedness.2,13,14
The lack of statistical significance in the present investigation could be due to the challenges posed by the extended
duration of the recordings and the complex experimental
setup, which required participants to produce the various
phonation types with a flow mask, intraoral tube, and a flexible endoscopy. In addition, the participants were untrained
and not used to consistently shifting between phonation
types. Other studies investigating the relationship between
airflow and phonation types typically were performed with
trained participants and without the endoscope in
place.2,34,56
The speed quotient from the flow glottogram was statistically larger for flow compared to breathy, neutral, and
pressed phonations during loud sound level in females suggesting a skewing of the glottal volume velocity waveform
to the right. In males, the speed quotient was larger in flow
compared to breathy, neutral, and pressed phonation types
during typical sound level, again suggesting that the duration of the opening phase was longer compared to the closing phase for neutral phonation. This assymmetry between
the rising and falling parts of the flow glottogram can be
explained by the nonlinear source-filter interaction.30,31
Similar to high-speed videoendoscopy, the amplitude
quotient from inverse filtering was significantly greater for
breathy compared to flow, neutral, and pressed phonation
types in male and female participants, suggesting that the
length of the closing phase is longer. This finding is similar
to the findings reported by Alku & Vilkman (1996)57 in four
males and females and Alku et al (2002)58 in five speakers
with inverse filtering.
The glottal resistance was significantly larger for pressed
compared to neutral, breathy, and flow phonation in males
during typical sound level. This is in accordance with results
from previous investigations on individuals with various
vocal pathologies by Yanagihara (1970)59 and from exiosed
canine laryngeal expirements by Alipour et al (1997).60
Relationship between flow and area
The amplitude quotient, which is a measure related to the
degree of glottal abduction,1,61,62 decreased more slowly in
the GA than the flow (Figure 7). The difference between the
area-based and the flow-based versions is that the former
includes the glottal leakage. As expected, the females produced lower values compared to the males. Also, as anticipated, the GA during the closing phase decreased more
slowly than the flow in both males and females (Figure 8).
Thus, the flow glottogram was more skewed to the right
than the area waveform. This is in good agreement with the
theory of source-filter interaction31,63,64 and is similar to the
findings in one male and one trained female in Granqvist
et al (2003)34 and Alku et al (2019)62 on 10 speakers for the
ARTICLE IN PRESS
20
vowel /i/. At the beginning of the glottal opening, a small
airflow is produced for a given GA, however, at the end of
the open phase due to the inertia and the non-linear sourcefilter interaction, a large airflow is produced for a given glottal opening.31,65−67 The data reveal that the changes in phonation type are mainly due to the changes in the closing
phase in the glottal volume velocity waveform and the GA.
During closing, the GA decrease is slower than the flow
decrease in both males and females (Figure 8).
CONCLUSION
This study provides insights into the relationship between
the GA and flow glottograms across various phonation
types of breathy, flow, neutral, and pressed. Emerging findings from this study suggest that the pulse amplitude is not
the determining factor for flow phonation in the untrained
participants in this study. The closing speed and duration of
the closing, conversely, were robust in differentiating the
phonation types. The flow pulse amplitude however, is useful for visual feedback purposes in vocal training. The study
findings futher suggest the use of PCV from the GA waveform and amplitude quotient from both the GA waveform
and flow glottograms has the potential for monitoring
changes in treatment targeting flow phonation. Untrained
male participants in this study not only varied closing speed
but also varied the duration of the closing phase to achieve
target changes in the GA waveform for breathy, flow, neutral, and pressed phonation. Females on the other hand,
predominantly vary closing speed as a function of phonation types. Findings from this study could lead to the development of empirically based quantitative treatment targets
for training flow phonation in adults with voice disorders
and assessment of treatment outcomes by providing the
most salient objective measures of flow phonation in individuals with voice disorders. Having only two participants
with complete data set is a limitation of the study. Future
studies should consider scheduling the experiment for two
days to facilitate complete data acquisition.
REFERENCES
1. Alku P. Glottal inverse filtering analysis of human voice production-a
review of estimation and parameterization methods of the glottal excitation and their applications. Sadhana. 2011;36:632–650.
2. Gauffin J, Sundberg J. Spectral correlates of glottal voice source waveform characteristics. J Speech Hear Res. Sep 1989;32:556–565.
3. Sundberg J. Vocal fold vibration patterns and modes of phonation.
Folia Phoniatr Logop. 1995;47:218–228.
4. Hillman RE, Holmberg EB, Perkell JS, et al. Objective assessment of
vocal hyperfunction: an experimental framework and initial results.
J Speech Hear Res. Jun 1989;32:373–392.
5. Boone DR, McFarlane S, Von Berg S, et al. The voice and voice therapy. 8th Edition ed. Boston: Akkyn & Bacon Communication Sciences
and Disorders; 2010.
6. Patel R, Liu L, Galatsanos NP, et al. Differential vibratory characteristics of adductor spasmodic dysphonia and muscle tension dysphonia
on high-speed digital imaging. Ann Otol Rhinol Laryngol. 2011;120:
21–32.
Journal of Voice, Vol. &&, No. &&, 2020
7. Gartner-Schmidt JL. Flow Phonation. In: Stemple J, Hapner ER, eds.
Voice therapy: Clinical Studies. 4th Edition ed. San Diego, CA: Plural
Publishing; 2013.
8. Gartner-Schmidt JL, Roth DF, Zullo TG, et al. Quantifying component parts of indirect and direct voice therapy related to different voice
disorders. J Voice. 2013;27:210–216.
9. McCullough G, Zraick RI, Balou S, et al. Treatment of laryngeal
hyperfunction with flow phonation: A pilot study. J Laryngol Voice.
2012;2:64–69.
10. Titze. On flow phonation and airflow management. J Singing.
2015;72:57–58.
11. Watts CR, Diviney SS, Hamilton A, et al. The effect of stretch-andflow voice therapy on measures of vocal function and handicap.
J Voice. 2015;29:191–199.
12. Watts CR, Hamilton A, Toles L, et al. A randomized controlled trial
of stretch-and-flow voice therapy for muscle tension dysphonia. Laryngoscope. 2015;125:1420–1425.
13. Millgard M, Fors T, Sundberg J. Flow glottogram characteristics
and perceived degree of phonatory pressedness. J Voice.
2016;30:287–292.
14. Sundberg J, Thalen M, Alku P, et al. Estimating perceived phonatory
pressedness in singing from flow glottograms. J Voice. 2004;18:56–62.
15. Sundberg J. Flow glottogram and subglottal pressure relationship in
singers and untrained voices. J Voice. 2017;32:23–31.
16. Behrman A. Common practices of voice therapists in the evaluation of
patients. J Voice. 2005;19:454–469.
17. Patel R, Awan SN, Barkmeier-Kraemer J, et al. Recommended protocols for instrumental assessment of voice: American speech-languagehearing association expert panel to develop a protocol for instrumental
assessment of vocal function. Am J Speech Lang Pathol. 2018;June:
1–19.
18. Roy N, Barkmeier-Kraemer J, Eadie T, et al. Evidence-based clinical
voice assessment: a systematic review. Am J Speech Lang Pathol.
2013;22:212–226.
19. Stachler RJ, Francis DO, Schwartz SR, et al. Clinical practice guideline: hoarseness (Dysphonia) (Update). Otolaryngol Head Neck Surg.
2018;158(1_suppl):S1–S42.
20. Patel R, Dailey S, Bless D. Comparison of high-speed digital imaging
with stroboscopy for laryngeal imaging of glottal disorders. Ann Otol
Rhinol Laryngol. 2008;117:413–424.
21. Olthoff A, Woywod C, Kruse E. Stroboscopy versus high-speed glottography: a comparative study. Laryngoscope. 2007;117:1123–1126.
22. Poburka BJ, Patel RR, Bless DM. Voice-Vibratory assessment with
laryngeal imaging (VALI) form: reliability of rating stroboscopy and
high-speed videoendoscopy. J Voice. 2017;31:e1-513.e514.
23. Hirano M, Bless D. Videostroboscopic Examination of the Larynx. San
Diego, California: Singular Publishing Group, INC.; 1993.
24. Sundberg J, Fahlstedt E, Morell A. Effects on the glottal voice source
of vocal loudness variation in untrained female and male subjects.
J Acoust Soc Am. 2005;117:879–885.
25. Sundberg J, Titze I, Scherer R. Phonatory control in male singing:
a study of the effects of subglottal pressure, fundamental frequency, and mode of phonation on the voice source. J Voice.
1993;7:15–29.
26. Rothenberg M. A new inverse-filtering technique for deriving the glottal air flow waveform during voicing. J Acoust Soc Am. 1973;53:1632–
1645.
27. Gauffin-Lindquist J. Inverse filtering, instrumentation and techniques.
Speech Transmission Laboratory, Quarterly Progress and Status Report
(Stockholm, Royal Institue of Technology). 4, 19641–4.
28. Titze. The physics of small-amplitude oscillation of the vocal folds.
J Acoust Soc Am. 1988;83:1536–1552.
29. Titze. Regulating glottal airflow in phonation: application of the maximum power transfer theorem to a low dimensional phonation model.
J Acoust Soc Am. 2002;111:367–376.
30. Rothenberg M, Zahorian S. Nonlinear inverse filtering technique for
estimating the glottal-area waveform. J Acoust Soc Am. 1977;61:1063–
1070.
ARTICLE IN PRESS
Rita R. Patel, et al
Glottal Airflow and Glottal Area Waveform Characteristics
31. Titze. Nonlinear source-filter coupling in phonation: theory. J Acoust
Soc Am. 2008;123:2733–2749.
32. Hertegard S, Gauffin J, Karlsson I. Physiological correlates of the
inverse filtered flow waveform. J Voice. 1992;6:224–234.
33. Hertegard S, Gauffin J. Glottal area and vibratory patterns studied
with simultaneous stroboscopy, flow glottography, and electroglottography. J Speech Hear Res. 1995;38:85–100.
34. Granqvist S, Hertegard S, Larsson H, et al. Simultaneous analysis of
vocal fold vibration and transglottal airflow: exploring a new experimental setup. J Voice. 2003;17:319–330.
35. Kempster GB, Gerratt BR, Verdolini Abbott K, et al. Consensus auditory-perceptual evaluation of voice: development of a standardized
clinical protocol. Am J Speech Lang Patho. 2009;18:124–132.
36. McDonnell M, Sundberg J, Westerlund J, et al. Vocal fold vibration
and phonation start in aspirated, unaspirated, and staccato onset.
J Voice. 2011;25:526–531.
37. Patel RR, Unnikrishnan H, Donohue KD. Effects of vocal fold nodules on glottal cycle measurements derived from high-speed videoendoscopy in children. PLoS One. 2016;11: e0154586.
38. Patel R, Walker R, Dollinger M. Oscillatory onset and offset in young
vocally healthy adults across various measurement methods. J Voice.
2017;31: e517-512 e524.
39. Svec,
Granqvist S. Tutorial and guidelines on measurement of sound
pressure level (SPL) in voice and speech. J Speech-Lang Hear Res.
2018;61:441–461.
40. Gramming P, Sundberg J. Spectrum factors relevant to phonetogram
measurement. J Acoust Soc Am. 1988;83:2352–2360.
41. Titze IR. Acoustic interpretation of the voice range profile (phonetogram). J Speech Hear Res. 1992;35:21–34.
42. Plant RL, Younger RM. The interrelationship of subglottic air pressure, fundamental frequency, and vocal intensity during speech.
J Voice. 2000;14:170–177.
43. Schutte H. The efficiency of voice production. Dissertation. 1980.
44. Bjorklund S, Sundberg J. Relationship between subglottal pressure and
sound pressure level in untrained voices. J Voice. 2016;30:15–20.
45. Holmberg EB, Hillman RE, Perkell JS. Glottal airflow and transglottal air pressure measurements for male and female speakers in soft,
normal, and loud voice. J Acoust Soc Am. 1988;84:511–529.
46. Holmberg EB, Hillman RE, Perkell JS, et al. Comparisons among
aerodynamic, electroglottographic, and acoustic spectral measures of
female voice. J Speech Hear Res. 1995;38:1212–1223.
47. Plexico LW, Sandage MJ. Influence of vowel selection on determination of phonation threshold pressure. J Voice. 2012;26:673.e7–673.e12.
48. Lohscheller J, Toy H, Rosanowski F, et al. Clinically evaluated procedure for the reconstruction of vocal fold vibrations from endoscopic
digital high-speed videos. Med Image Anal. 2007;11:400–413.
49. Woo P. Quantification of videostrobolaryngoscopic findings−measurements of the normal glottal cycle. Laryngoscope. 1996;106(3 Pt 2 Suppl
79):1–27.
21
50. Patel R, Donohue KD, Unnikrishnan H, et al. Kinematic measurements of the vocal-fold displacement waveform in typical children and
adult populations: quantification of high-speed endoscopic videos.
J Speech Lang Hear Res. 2015;58:227–240.
51. Dejonckere PH, Kob M. Pathogenesis of vocal fold nodules: new
insights from a modelling approach. Folia Phoniatr Logop. 2009;61:
171–179.
52. Sundberg J, Andersson M, Hultqvist C. Effects of subglottal pressure
variation on professional baritone singers' voice sources. J Acoust Soc
Am. 1999;105:1965–1971.
53. Holmberg EB, Hillman R, Perkell JS. Glottal airflow and transglottal air pressure measurements for male and female speakers
in low, normal and high pitch. J Voice. 1989;3:294–305.
54. Sodersten M, Lindestad PA. Glottal closure and perceived breathiness
during phonation in normally speaking subjects. J Speech Hear Res.
1990;33:601–611.
55. Sodersten M, Hertegard S, Hammarberg B. Glottal closure, transglottal airflow, and voice quality in healthy middle-aged women. J Voice.
1995;9:182–197.
56. Sundberg J. Objective characterization of phonation type using amplitude of flow glottogram pulse and of voice source fundamental.
J Voice. In press.
57. Alku P, Vilkman E. Amplitude domain quotient for characterization
of the glottal volume velocity waveform estimated by inverse filtering.
Speech Commun. 1996;18:131–138.
58. Alku P, Backstrom T, Vilkman E. Normalized amplitude quotient for
parametrization of the glottal flow. J Acoust Soc Am. 2002;112:701–
710.
59. Yanagihara N. Aerodynamic examination of the laryngeal function.
Studia Phonologica. 1970;5:45–51.
60. Alipour F, Scherer RC, Finnegan E. Pressure-flow relationships
during phonation as a function of adduction. J Voice. 1997;11:
187–194.
61. Fant CG. The LF-model revisited. Transformations and frequency
domain analysis. Speech Transm Lab Q Prog Status Rep. 1995;36:119–
156.
62. Alku P, Murtola T, Malinen J, et al. Skewing of the glottal flow with
respect to the glottal area measured in natural production of vowels.
J Acoust Soc Am. 2019;146:2501.
63. Fant G, ed. Acoustic Theory of Speech Production. The Hague, Netherlands: Mouton; 1960.
64. Rothenberg M. Measurement of airflow in speech. J Speech Hear Res.
1977;20:155–176.
65. Titze Riede T, Popolo P. Nonlinear source-filter coupling in phonation: vocal exercises. J Acoust Soc Am. 2008;123:1902–1915.
66. Jiang J, Zhang Y, Ford CN. Nonlinear dynamics of phonations in
excised laryngx experiments. J Acoust Soc Am. 2003;14:2198–2205.
67. Herzel H, Berry DA. Analysis of vocal disorders with methods from
nonlinear dynamics. J Speech Hear Res. 1994;37:1008–1019.