SUBGLOTTAL PRESSURE MEASURES DURING
VOCAL FRY PHONATION
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The author expresses sincere gratitude to his com-
mittee chairman Dr. Donald Dew for his continued support
and encouragement during the preparation and execution of
the experiment and also throughout the author's training
program at the University of Florida. Further gratitude is
expressed to the members of the dissertation committee for
their suggestions and contributions to the ideas expressed
herein. Special thanks are given to the subjects who will-
ingly submitted to the procedures of the experiment.
The author wishes to acknowledge the invaluable support
from the faculty, staff, and students of the Communication
Sciences Laboratory and to the staff of the Bronchopulmonary
Laboratory, J. Hillis Miller Health Center. The author is
especially indebted to Mr. Robert Idzikowski for his assis-
tance in preparing the instrumentation for the study, to
Dr. J. O. Harris and Mrs. Joyce Fedik for their assistance
during the experimental sessions, and to Drs. Harry Hollien
and John F. Brandt for their constructive criticism provided
during'the preparation of preliminary drafts of this report.
This research was supported by the National Institutes
of Health grants NB-05475 and NB-06459.
TABLE OF CONTENTS
ACKNOWLEDGMENTS . . . . . . . .
LIST OF TABLES .......
LIST OF FIGURES . . . . . . . .
1 INTRODUCTION AND STATEMENT OF
II PROCEDURES ....
III RESULTS . .
IV DISCUSSION ....
V SUMMARY AND CONCLUSIONS
APPENDIX A . . .
APPENDIX B . . .
APPENDIX C . . .
APPENDIX D . . .
PROBLEM . .
LIST OF TABLES
1 Summary of three-way analysis of
variance to determine the effects
of vowels and phonation conditions
upon the subglottal pressures. .. ... 26
2 Comparison of subglottal pressures
at each phonation condition. The
marginal values contain the sum of
the subglottal pressures for both
vowels at each phonation condition.
The values in the matrix represent
the differences between each marginal
pair . . . . . . ... . 28
3 Summary of three-way analysis of
variance to determine the effects of
the mouthpiece and phonation conditions
upon subglottal pressure .. . 31
4 Comparison of subglottal pressures at
each phonation condition. The marginal
values contain the sum of the subglottal
pressures for both mouthpiece conditions
at each phonation condition. The values
in the matrix represent the differences
between each marginal pair . .. 33
5 Summary of the analysis of variance for
air flow as a function of phonation
conditions ... . . . . . . 36
6 Comparison of the air flow rates at each
phonation condition. The marginal values
in the matrix represent the difference
between each marginal pair. . . 38
7 Phonational ranges for the five subjects
used in the present investigation. The
measures are reported in Hz for the vocal
fry and modal ranges. . .. .. ... 57
8 Mean fundamental frequency, air pressure,
rate of air flow, and relative intensity
for the vowels /a/ and /i/ at five
phonation regions for Subject 1. In con-
ditions 1-10 rate of air flow was not
recorded. . . . . . . . . 64
9 Mean fundamental frequency, air pressure,
rate of air flow, and relative intensity
for the vowels /a/ and /i/ at five phonation
regions for Subject 2. In conditions 1-10
rate of air flow was not recorded. . . 65
10 Mean fundamental frequency, air pressure,
rate of air flow, and relative intensity
for the vowels /a/ and /i/ at five
phonation regions for Subject 3. In con-
ditions 1-10 rate of air flow was not
recorded ... . . ........ . 866
11 Mean fundamental frequency, air pressure,
rate of air flow, and relative intensity
for the vowels /a/ and /i/ at five
phonation regions for Subject 4. In con-
ditions 1-10 rate of air flow was not
recorded. . . .. . . .. 67
12 Mean fundamental frequency, air pressure,
rate of air flow, and relative intensity
for the vowels /a/ and /i/ at five
phonation regions for Subject 5. In con-
ditions 1-10 rate of air flow was not
recorded. ... .. . . . . .. 68
13 Mean fundamental frequency, subglottal
pressure, and rate of air flow at five
phonation conditions for five subjects
producing /a/ and /i/. . . . .. 69
LIST OF FIGURES
1 Block diagram of equipment used to
provide the reference signals, record
the pressure, flow and voice signals,
and monitor subjects' output. . . . 11
2 Subglottal pressure (cm H20) plotted
for the vowels /a/ and /i/ during
three vocal fry and two modal phonation
conditions. . . . .. .. . 24
3 Mean subglottal air pressure (cm H20)
for phonation for /a/ with and without
the air flow recording apparatus in the
mouth at three vocal fry and two modal
phonation conditions. .. . .. .. . 29
4 Air flow rate as a function of phonation
condition. Data points represent means
for five subjects. Combined range of
subjects at each condition indicated by
dotted lines. . .. .. . . . . 34
5 Mean subglottal pressure as a function of
the relative intensity for five subjects
phonating the vowels /a/ and /i/ in vocal
fry. The values for /i/ are underlined . 40
6 Mean subglottal pressure as a function of
the relative intensity for five subjects
phonating the vowel /a/ in vocal fry with
the mouthpiece in place . . . .. 41
7 Mean air flow rate as a function of the
relative intensity for five subjects
phonating the vowel /a/ in vocal fry. . 43
8 Schematic representation of the output
from the Multitrace Oscillograph used
in measuring subglottal pressure and rate
of air flow ... . .. . ... . . . 62
INTRODUCTION AND STATEMENT OF PROBLEM
Recently, Hollien and Michel (1968) demonstrated that
vocal fry can be appropriately classified as one of the three
frequency ranges used by normal speakers during phonation.
Each of these ranges, falsetto, modal (or mid-range),l and
vocal fry, appears to have distinctive characteristics with
regard to its 1) production, 2) acoustic wave form, and 3)
perceptual attributes. The absolute fundamental frequencies
of the ranges will vary from speaker to speaker, but for any
single speaker, falsetto is produced with the highest funda-
mental frequencies, modal is produced in the middle frequency
The mid-range of the total phonational range is some-
times considered to be the modal register (Morner, Franeson,
and Fant, 1963). As noted by Hollien and Michel (1968), how-
ever, the term "voice register" appears to have no adequate,
accepted definition. In fact, various investigators have
divided the total phonational range into as few as one and as
many as five "voice registers" (Morner, Franeson, and Fant,
1963). It was not the purpose of the research to provide a
description of the term "voice register"; rather, the intent
was to describe certain relationships which occur in vocal
fry, a mode of phonation perceptually described as a quasi
periodic series of relatively distinct pulses.
range, and vocal fry is produced in the lowest range of
Because vocal fry has been recognized only recently
as a normal phonatory range (Hollien, Moore, Wendahl, and
Michel, 1966) which appears to occur routinely during speech,
it is of particular interest to the scientist. There have
been several investigations concerning the description and
perception of fry (Coleman, 1963; Wendahl, Moore, and Hollien,
1963; Hollien and Michel, 1968; Michel and Hollien, 1968;
Hollien and Wendahl, 1968); information concerning the
operation of the laryngeal structures during vocal fry pho-
nation, however,is somewhat lacking. Specifically, pre-
vious investigations of vocal fry have provided information
concerning the approximate range of vocal fold repetition
rates. For example, in a recent investigation by Hollien
and Michel (1968), the phonational ranges of 12 males and
11 females were found to be from two to 78 pulses per second."
This study confirms the results of earlier investigations
reporting on the approximate range of vocal fry repeti-
tion rates (Michel and Hollien, 1968; McGlone, 1967) and
2When the term, "pulse," is used, it refers to a
function which makes one or more excursions from a base-
line and which has a finite baseline time (Hubbs, 1966).
suggests that fry is characterized by fundamental frequen-
cies lower than those in the modal register.
Hollien and Wendahl (1968)have also reported that
subjects can accurately match the repetition rate of vocal
fry pulse patterns to electronically produced pulses occurring
at low frequencies. Thus, in fry there appears to be suffi-
cient periodicity in the signal for listeners to perceive it
as having low fundamental frequency.
Although low fundamental frequencies are characteris-
tic of vocal fry phonation, Coleman (1963) reported that the
damping of the wave rather than its repetition rate was the
important factor in its perception. Using an electronic
laryngeal analog to vary the damping factor, he reported that
when the damping factor reached a critical value, fry was
always perceived; when the damping factor was less, fry was
not perceived regardless of the frequency. Wendahl, Moore,
and Hollien (1963) have also demonstrated that the vocal fry
acoustic wave forms of human productions show high damping.
Having examined a large number of phonellographic tracings,
they found fry to be characterized by a highly damped wave
form in addition to its low frequency. It appears, therefore,
that vocal fry is a distinct mode of phonation with relatively
low fundamental frequencies and a wave form which is highly
In an investigation of glottal wave forms, Timke,
Von Leden, and Moore (1958), also provide data which would
support the contention that the vocal fry wave form is highly
damped. Using high speed motion picture photography, they
found the vocal fry glottal wave form to be the result of a
pulse-like opening and closing of the folds followed by a long
closed phase. Indeed, some examples of fry phonation have
been observed (Fant, 1964; Coleman, 1968) which consist of a
double or triple pulse followed by the long closed phase.
Wendahl, Moore, and Hollien (1963) report that while these
patterns were found in their phonellographic tracings, the
most common vocal fold vibratory pattern in vocal fry is one
of a single glottal pulse followed by the characteristically
long closed phase.
Although the production of modal range phonation has
been extensively investigated,the operation of the laryngeal
mechanism during vocal fry has not been specified. This
lack of research may be attributed in part to the miscon-
ception that vocal fry was considered a pathological con-
dition which occurred only infrequently. Clearly, although
vocal fry may be pathological if used extensively, it does
exist as a product of the normal larynx as pointed out pre-
viously. If regarded as a normal phonatory event, then
the mechanism of fry production may be explained within a
framework of other types of normal phonation, namely the
myoelastic-aerodynamic theory of phonation (Van den Berg, 1958).
Investigations of the voice produced during modal pho-
nation have shown frequency to be highly correlated with vocal
fold length,3 mass, and thickness4 and the subglottal pressures
needed to maintain vocal fold movement (Kunze, 1962). While
considerable data have been presented showing the relationship
between vocal fold length and fundamental frequency of pho-
nation in the modal range, it has been shown that the length
of the folds does not appear to vary as a function of the rep-
etition rate in fry (Hollien, Damste, and Murry, 1969).
Investigations of vocal fold thickness (see footnote
four) have generally shown that, as the fundamental frequency
of phonation decreases, the thickness of the vocal folds
increases. Although the reports of the thickness of the vocal
folds during fry as yet are predominantly qualitative, they
indicate that during vocal fry donationn the folds ap-ar to
he rather thick but not necessarily tense. These observations
were noted by Coleman (1968) from examination of high speed
motion pictures, by Hollien, Damste, and Murry (1969)
from a strictly subjective comparison of lateral x-rays
3Irwin, 1940; Brackett, 1947; Sonninen, 1954, 1956;
Hollien, 1960; Hollien and Moore, 1960; Hollien, 1962;
Wendler, 1964; Damste, Hollien, Moore, and Murry, 1968.
4Hollien and Curtis, 1960; Hollien, 1962a, 1962b;
Hollien, Coleman, and Moore, 1968.
and from laminagraphs (Hollien, 1968) taken during modal and
Most of the data concerning the aerodynamic forces
present during phonation exist for modal range phonation.
Kunze (1962, 1964), Isshiki (1964), and Perkins and Yanagi-
hara (1968) investigated mean rate of air flow over a wide
range of frequencies and found a positive relationship be-
tween air flow rate and intensity. These investigators
found air flow rate to have little or no relationship to
the fundamental frequency of phonation. Additional findings
by Kunze (1962) demonstrated the importance of the air flow
measuring system; he has shown that the absolute values of
flow rates are significantly lower when a respirometer is
used in place of a pneumotachograph.
Air flow rates during sustained phonation of vocal
fry have been obtained by McGlone (1967) using a respirometer.
He found flow rates ranging from 2.0 ml/sec to 71.9 ml/sec
for repetition rates ranging from 10.9 to 52.1 pps. A Pearson
Product-Moment Correlation Coefficient was carried out for
the two sets of measures and found to be .26. These results
for fry phonation, like those obtained for higher frequency
phonation, did not demonstrate any systematic relationship
between air flow rate and repetition rate.
As yet, there are no experimental results of the sub-
glottal pressures produced during vocal fry phonation.
Several investigators have obtained subglottal pressure mea-
sures for low frequency modal phonation of the vowel, /a /
(Van den Berg, 1956; Isshiki, 1959; Ladefoged, 1962; Kunze,
1962); however, their results appear to conflict. For example,
Van den Berg (1956) reported a mean subglottal pressure of
10 cm HI20 for phonation at 145 Hz as compared to 9 cm H20
for phonation at 97 Hz. Isshiki (1959) found an average
pressure of 9 cm H20 for sustained phonation at 123 Hz and
5.5 cm H20 for phonation at 98 Hz. While Ladefoged's (1962)
results show a similar trend, Kunze (1962) noted that as the
subject produced tones below or above the 30 percent point of
his phonational range, subglottal pressure increased. Thus,
the only implication appears to be that frequency increases
as a function of subglottal pressure over the upper portion
of the phonational range. Due to sampling techniques, the use
of a single vowel, and procedural considerations,5 however,
5The measures reported by Ladefoged and his associates
(Draper, Ladefoged, and Whitteridge, 1957, 1959; Ladefoged,
Draper, and Whitteridge, 1958; Ladefoged (1962) are somewhat
higher than those reported by Van den Berg (1956) and Isshiki
(1959). This is, as Kunze (1964) has shown, due to the fact
that Ladefoged's results are based on estimates of subglottal
pressure obtained from recordings of the intraesophageal pres-
sure rather than intratracheal pressure measures. It has been
previously demonstrated (Fry, Stead, Ebert, Lubin, and Wells,
1952; Mead and Whittenberger, 1953) that recordings of intra-
esophageal pressure consist of the intratracheal (airway) pres-
sure which drives the folds plus the pressure required to over-
come the elastic resistance of the lungs. Since this latter
resistance is not constant over the expiratory cycle, it ap-
pears that intraesophageal pressure is not a valid estimate
of subglottal pressure.
this relationship is not clear, especially at the low end of
the modal frequency region.
Kunze's (1962) subglottal pressure measures based
upon recordings of the vowel /a/ by 10 subjects at five points
along their phonational range appear to be the most comprehen-
sive of those reported. He found that from the 90 percent to
10 percent points of the subjects phonational range (excluding
vocal fry), subglottal pressure decreases as the fundamental
frequency decreases except at the subjects' 10 percent point.
In addition, and perhaps of major significance, Kunze demon-
strated the importance of obtaining direct (intratracheal)
measures of subglottal pressures (see footnote five) rather
than indirect estimates of pressure from an esophogeal bal-
loon. He concluded that ". there is little doubt that
intratracheal pressure measures provide the only satisfactory
estimates of subglottal pressure under conditions of sustained
phonation and during connected speech."
Statement of the Problem
Previous studies of the aerodynamic factors contribut-
ing to laryngeal operation have been confined predominantly
to the'middle and upper portions of the phonational range
and have demonstrated that 1) increases in subglottal pres-
sure are usually accompanied by increases in the fundamental
frequency of phonation, 2) subglottal pressure is directly
related to vocal intensity, 3) air flow appears to be re-
lated to vocal intensity, and 4) there is no systematic
relationship between the fundamental frequency of phonation
and air flow rate.
To date, only the variable of air flow rate has been
examined during vocal fry. Specifically, air flow was found
to be significantly lower in vocal fry than in modal phonation
and to have no relationship to fundamental frequency of pho-
nation (McGlone, 1967). In order to understand the operation
of the laryngeal mechanism during vocal fry, it would appear
necessary to examine both air pressure and air flow rate as
they relate to changes in the repetition rate and vocal
intensity of vocal fry phonation.
The major purpose of this study was to determine if
the subglottal pressures and rates of air flow produced dur-
ing vocal fry differ from those in the mid-phonational range.
A second purpose was to investigate the relationship between
the fundamental frequency of phonation of vocal fry and both
subglottal air pressure and rate of air flow. Finally, con-
sideration was given to the observed relationships between
variations in vocal intensity and the subglottal pressure/
air flow data.
The plan of this study was to obtain simultaneous
measures of the intratracheal air pressure, air flow rate,
and sound pressure during vocal fry and mid-range phonation
from five adult males. These measures were obtained during
sustained phonation of the vowels /a/ and /i/. These vowels
which represent two extreme supraglottal configurations were
chosen to determine if varying the supraglottal structures
affects subglottal pressures. The samples were produced at
three vocal fry and two mid-range fundamental frequencies.1
A block diagram of the instrumentation employed in the
present study is shown in Figure 1. While the subject phonat-
ed in a supine position, subglottal pressure was measured di-
rectly through a hypodermic needle inserted between the first
and second tracheal rings, air flow was measured through a
pneumotachograph connected to a mouthpiece held in place by
1In order to avoid considerable confusion in the dis-
cussion of the mode of phonation, the term "fundamental fre-
quency" will be used throughout the remainder of this study
to refer to the quasi periodic laryngeal signal produced
during vocal fry and mid-range phonation. While this term
departs somewhat from standard acoustic terminology (Hubbs,
1966), it appears to be descriptive of the frequency of the
repetitive wave forms found both in vocal fry and mid-range
a flange fitting between the lips and teeth, and sound pres-
sure was measured with a microphone and probe tube inserted
into the connecting tube of the pneumotachograph. Follow-
ing amplification by separate channels of a data amplifier,
each of the three signals was recorded on both an oscillo-
graphic writer and an FM magnetic tape recorder.
For monitoring phonation, the acoustic wave was dis-
played on the face of an oscilloscope in view of the subject
and experimenter and also was presented to each of them by
means of individual earphones. In addition, the subject was
fitted with a second earphone to monitor the frequency of
phonation; the reference signal was generated by a sine-wave
oscillator for mid-range frequencies while a square-wave
generator provided reference signals for vocal fry.
Instrumentation for measuring and calibrating subglottal
After applying a local anesthetic (one percent xylo-
caine) to the laryngeal area, a physician inserted an 18
gauge hypodermic needle (3.5 inches long with an inside
diameter of .033 inches) between the first and second tracheal
rings so that it was perpendicular to the flow of air through
the trachea. The other end of the needle was connected to a
Statham PM 131 TC differential pressure transducer via a 12 cm
plastic tube having an inside diameter of .125 inches. The
resultant electrical signal was amplified by an Electronics-
for-Medicine DR-8 dc pressure amplifier and recorded simul-
taneously on onechannel of an associated Multitrace Oscillo-
graphic Writer and on one channel of a Honeywell 8100 FM
magnetic tape recorder. Kunze (1964) and Perkins and Yana-
gahara (1968) have shown that direct pressure recordings made
in this manner provide consistent measures of the intratracheal
pressure driving the folds.
The intratracheal pressure measuring system was cali-
brated daily by reference to a U-tube manometer scaled in
centimeters of water (cm H20). The manometer was connected
to the pressure transducer and associated oscillographic
writer by means of a Y-valve having one end open to the atmo-
sphere and the other attached to a syringe. By applying a
steady pressure with the syringe, a specific displacement of
water in the manometer could be calibrated with a specific
pressure change shown on the oscillographic writer. By this
method it was found that the transducer and its system were
linear within the levels needed in the present study.
Instrumentation for measuring air flow
Air flow measurements were made by a pneumotachograph
which operates according to the Law of Poiseuille. That is,
along a rigid tube, there is a decrease in pressure which is
proportional to the velocity of flow per unit length when the
flow is laminar (Rossier, Buhlmann, and Wiesinger, 1954).
Therefore, continuous recording of the pressure diff-
ferences between two points yields a function which,
when integrated, indicates the volume of air moved per
unit of time. In the present study, air flow was recorded
by connecting a Fleish No. 0 pneumotachograph to a mouth-
piece 2.12 inches in length and held in place by a flange
fitting between the lips and teeth. The pneumotachograph
was connected through a Statham PM 97 pressure transducer
to the data amplifier and recording systems. This device,
used with a noseclip, appeared to provide reliable measures
of air flow while avoiding the problems of dead air spaces
and improperly fitting face masks experienced by previous
investigators (Isshiki, 1959; Kunze, 1962; Hardy, 1965).
Calibration of the air flow recording system was
accomplished by passing a known flow rate produced by a
variable speed Emerson RJV-192 vacuum cleaner through a
Fisher-Porter rotameter and to the transducer. The ro-
tameter recorded the flow rate in liters per second; thus,
the deflections on the oscillograph were converted directly
to milliliters of air flow per second (ml/sec).
Sound pressure measurements
Relative intensity levels2 during the vocal fry
conditions were measured by a Brueland Kjaer 4134 one-half
inch condenser microphone with a probe tube six centimeters
long and an inside diameter of .2 mm. The frequency response
of the probe tube and microphone was found to be + 3 dB over
a range of 20-1000 Hz. The microphone probe was fitted into
an insert in the wall of the pneumotachograph and was connect-
ed to its associated Bruel and Kjaer 2801 Microphone Power
Supply. The signal in turn was fed to one channel of the
data recorder, the FM tape recorder, and a Tektronix 310A
oscilloscope. The oscilloscope provided the subject and the
experimenter with a visual trace of the phonation which was
used to maintain constant intensity for each phonation con-
In order to maintain a constant frequency during
phonation, the subject was provided with a reference signal
produced by a General Radio 1313A Oscillator and presented
through a Telephonics TDH 39 earphone, as shown schematically
The intensity levels for the vocal fry conditions
were not predetermined since it was observed in a pilot study
that vocal intensity, measured in front of the mouth tended
to increase as the repetition rate increased.
in Figure 1. The accuracy of the generator was verified by
a Hewlett Packard 5214L Preset Counter. The frequency set-
tings were determined from the information obtained from
the subjects' phonational ranges (Appendix A). The vocal fry
reference signals were square waves set at slow, medium, and
fast frequencies of the subject's fry phonational range. The
reference signals for the 10 and 30 percent frequencies of
the modal range were sine waves set at the predetermined fre-
quencies for each subject.
Five adult males were selected for the investigation.
The subjects volunteered to perform the tasks and were chosen
on the basis of the following criteria: 1) they were capable
of phonating in vocal fry over a range of repetition rates;
and 2) they showed no evidence or history of voice disorders
or laryngeal pathology.
Prior to the experimental task, each subject's modal
and vocal fry phonational ranges were determined. The modal
range was obtained in the traditional fashion by having
each subject sing up and down the musical scale until he
produced his highest and then his lowest sustainable notes.
In order to monitor his production, he was instructed to match
a series of pure tones presented at tone interval's. The
subject and the experimenter acted as judges to determine
which tones represented the extremes of his modal range.
The highest and lowest frequencies produced by the subject
and agreed upon by the experimenter were considered as the
modal phonational range boundaries.
The vocal fry range was determined by a procedure
similar to that described by Hollien and Michel (1968).
Specifically, each subject practiced producing vocal fry
at various repetition rates, then varied the rates upward
and downward until reaching the lowest and highest sustainable
repetition rate. When the experimenter and subject agreed
upon these limits, a four-second sample was recorded on
magnetic tape. To measure repetition rate, tape loops of
the samples were played through a General Radio 1900A Wave
Analyzer coupled to a Hewlett Packard 5214L Preset Counter.
The wave analyzer was set to the tracking generator mode
having a three-cycle bandwidth. When the largest deflection
appeared on the voltmeter, the frequency at that point was
recorded from the counter. The mid and fry phonational ranges
of all subjects are shown and discussed in Appendix A.
After administering a local anesthetic to the laryngeal
area, a physician inserted the hypodermic needle for recording
the intratracheal pressure. The needle was connected to the
transducer and the subject was given ample time to adjust
to the experimental apparatus. Each subject was then in-
structed to produce three samples of the vowels /a/ and /i/
at slow, medium, and fast repetition rates within his vocal
fry phonational range as well as at his 10 and 30 percent
points of the modal range. Subsequently, the mouthpiece
leading from the pneumotachograph was fitted and the subject
was asked to produce samples of the vowel /a/ at the three
vocal fry and two modal range frequencies. All recordings
were made while the subject was in the supine position. A
set of earphones presented the reference signal at one ear
while the subject monitored his vocal output at the other ear.
Thus, in the first series of recordings, subglottal pressure
was recorded during phonation of two vowels at five frequencies.
In the second series, subglottal air pressure and air flow were
recorded for the vowel /a/ only at the five frequency con-
ditions since it appeared that recordings of /i/ would be
greatly distorted in phonemic quality or would result in air
leakage around the mouthpiece. Appendix B contains the re-
cording itinerary followed throughout.
For each subject, frequency and pressure recordings
were made for the vowels /a/ and /i/ at three frequency re-
gions in vocal fry and two in the modal range. During a
separate recording period, the frequency, pressure, and air
flow were obtained for the vowel /a/ only at five frequency
regions. It was necessary to convert the outputs of the
multitrace oscillographic writer into numerical values (Appen-
dixC) to estimate the actual subglottal pressure and rate
of air flow. This was accomplished by measuring the pressure
shifts produced on the oscillographic writer at 100 milli-
second intervals for three seconds and converting the values
to centimeters of water for subglottal air pressure and milli-
liters of air per second for air flow using the appropriate
calibration factors previously obtained. For each phonation
a mean and standard deviation was computed. The means of three
samples at a particular phonation condition were computed for
frequency, subglottal air pressure, rate of air flow, and
relative intensity3 and used as the criterion measures.
The fundamental frequency data were obtained from the
Honeywell FM tape recorder. The recorder was slowed by a
factor of eight and played through a Sanborn 150-1300D Dynagraph.
3The relative intensity levels were obtained by record-
ing a two-volt sine wave on the tape recorder and playing it
back through a Bruel and Kjaer 2112 Audio Frequency Spectrometer
which acted as an attenuater. The output or the spectrometer
was coupled to a Bruel and Kjaer 2305 Graphic Level Recorder
which recorded the reference signal at mid-scale with a 50 dB
potentiometer. The measures reported arerelative to this scale
value rather than to the actual 2-volt signal. The use of the
scale values eliminates the large negative dB values.
Wave-to-wave measurements of the output of each sample were
made by measuring the wave from peak-to-peak, dividing this
by eight times the speed of the dynagraph to get the period.
Using the formula F = period measures were converted to
frequency in Hz.
Two analyses of variance with repeated measures were
performed on the subglottal pressure measures to determine
the effects of vowels, the mouthpiece, and the frequency of
phonation on the subglottal pressures (Lindquist, 1953). The
first was a frequency-by-vowel-by-subject anaylsis to determine
the effects of the phonation condition and vowel. The second
tested the effects of phonation condition and addition of the
mouthpiece on subglottal pressure. Since only one vowel was
used during the pressure and flow recordings, a one-way analysis
of variance with repeated measures was performed on the air flow
data to determine the effects of frequency changes on air flow
rate. Thus, it was possible to test differences in pressure
for effects due to frequency, vowel, and insertion of the
mouthpiece and in air flow for effects due to frequency of
To determine the relationship between the actual
repetition rate and subglottal pressure during vocal fry
phonation, a Pearson Product-Moment Correlation Coefficient
was computed for the variables of repetition rate and subglottal
pressure. Similarly, a Pearson Product-Moment Correlation
Coefficient was computed between repetition rate in vocal
fry and rate of air flow.
The major purpose of this study was to determine if
the subglottal pressures and rates of air flow produced
during vocal fry phonation differ from those produced in
modal range phonation. This study also investigated the
relationships of both subglottal air pressure and rate of
air flow with a) fundamental frequency and b) vocal intensity.
In order to accomplish these purposes, simultaneous recordings
of the intratracheal air pressure, air flow, and sound pres-
sure were obtained from five subjects phonating the vowels
/a/ and /i/ at five fundamental frequency regions, three in
the vocal fry and two in the modal range. Three samples at
each experimental condition were obtained; the criterion
measures were the mean frequency, mean subglottal pressure,
mean rate of air flow, and average relative intensity for
three samples in each condition.
It should be noted that the subglottal air pressures
and mean rates of air flow obtained in the present study
varied from subject to subject as well as between successive
samples produced by the same subject. For example, subject
two was found to have mean subglottal air pressures which
were higher than those of most other subjects; however,
the frequencies at which he was phonating were similar to
the other.subjects. Moreover, examination of the raw data
showed that subject two had little sample-to-sample varia-
tion in comparison to subject one. The reader should be
aware of the existence of the variability found in this as
well as other investigations (Kunze, 1962) when interpret-
ing the present data.
Subglottal Pressure During Vocal Fry
Relationship of subglottal pressure to phonation condition
The mean subglottal pressures for the vowels /a/ and
/i/ produced at three vocal fry and two modal frequency
regions of the subjects' ranges are shown in Figure 2.
Figure 2 clearly indicates that the subjects' subglottal
air pressure was greater in vocal fry than in the two modal
range conditions. Moreover, there were no reversals in this
trend at any phonation condition. From Figure 2 it can also
be seen that as subjects increased frequency of phonation
in vocal fry, the subglottal pressure increased. In the
modal range conditions, however, the subglottal pressures
decreased as the subjects' frequency increased from the 10 percent
S/Ia/ O /I/
Figure 2. Subglottal pressure (cm H20) plotted for the
vowels /a/ and /i/ during three vocal fry and two modal
region1 of their modal phonational range. In vocal fry, the
overall mean increase in subglottal pressure from the slow
repetition rates to the fast rates was statistically signif-
icant at the five percent level. In view of the variability
exhibited by the subjects during the production of the modal
range frequencies, however it can only be shown that as the
subjects increased their fundamental frequency, subglottal
pressures decreased significantly as the fundamental frequency
approached the 30 percent region of a subject's modal pho-
Figure 2 also shows that the subglottal pressure trends
which were found for the /a/ were similar to those found for
/i/; although the values for /i/ were higher in all but one
condition, the differences were not significant.
To determine if the changes in subglottal pressure as
a function of the five phonation conditions (i.e., slow, me-
dium, and fast fry, and 10 and 30 percent modal), and vowels
were statistically significant, a treatments by treatments by
subjects analysis of variance (Lindquist, 1953) was performed.
The results of this analysis are presented in Table 1. The
effect of phonation condition was significant at the .05 level;
While all subjects did not produce all samples at the
exact 10 and 30 percent frequencies of their modal phonational
range, the means of the three samples were within three semi-
tones of the desired frequency.
Table 1. Summary of three-way analysis of variance to
determine the effects of vowels and phonation conditions
upon the subglottal pressures.
Source of Variation SS df MS F
A Phonation condition 125.38 4 31.35 4.52*
B Vowels 4.33 1 4.33 1.56
S Subjects 81.47 4 20.37
AB 1.95 4 .49 .53
AS 110.51 16 6.91
BS 11.08 4 2.77
ABS 14.67 16 .92
Total 349.39 49
(F = 3.01, F
the .05 level
the effects due to vowels and the AB interaction were not
statistically significant at the tested level.
To determine which of the phonation conditions
accounted for the overall significance of this effect, the
Newman-Keuls test of treatment differences following an
overall significant F-ratio was applied to the data. This
statistic tests the difference between any number of means
arranged in increasing order of magnitude while maintaining
the level of significance equal to alpha. The results of
this test presented in Table 2 indicate that all phonation
conditions were significantly different from each other.
That is, the pressures in all samples of vocal fry were
significantly greater than the pressures in all samples of
mid-range phonation. Furthermore, the matrix shows that
the pressures at the 10 percent modal frequency were sig-
nificantly greater than those at the 30 percent region and
also that the pressures at each of the three fry frequency
regions were significantly different from each other.
Effects of mouthpiece and phonation condition upon subglottal
As described previously, subglottal pressures for the
vowel /a/ were recorded when the subject had a mouthpiece in-
serted for the purpose of recording air flow rate and when
the mouthpiece was not inserted. Figure 3 presents the
40 > i
O 0 cn
H-F m .
0 o -r
>~ it d i
^ ^ +*
ri 3 r
6-1 S I .
0 0 C 0
o 0 l
0l C (
0 0 (0 0
0 0 C lI
/ / WITHOUT
-Mean subglottal air pressure (cm H20) for
for /a/ with and without the air flow recording
in the mouth at three vocal fry and two modal
results for the mean subglottal pressures for the /a/ under
these two conditions plotted as a function of phonation
condition. Figure 3 indicates that the subglottal pressures
for both treatment conditions, i.e., with and without the
mouthpiece, increased as a function of the phonation con-
dition in vocal fry. The subglotta] pressures in the modal
range were lower than those in vocal fry and tended to de-
crease as the subjects increased frequency of phonation.
When the subglottal pressures for the two recording condi-
tions are compared at any one phonation condition, it can
be seen that the pressures produced while the subject was
fitted with the mouthpeice were lower than when the mouth-
piece was removed. Thus, it appears that the addition of
the mouthpiece for recording air flow rate caused a reduction
in the subglottal pressures.
An analysis of variance was performed to determine if
the changes in pressure as a function of the two variables
was statistically significant; Table 3 summarizes this anal-
ysis. The effects due to insertion of the mouthpiece were
not significant at the .05 level. The effects due to pho-
nation condition were significant at the .05 level; the AB
interaction was not significant at the tested level.
The Newman-Keuls test was applied to the data to de-
termine which of the phonation conditions accounted for the
Table 3. Summary of three-way analysis of variance to
determine the effects of the mouthpiece and phonation
conditions upon subglottal pressure.
Source of Variation SS df MS F
A Phonation condition 99.73 4 24.93 7.16*
B Mouthpiece 8.00 1 8.00 1.04
S Subjects 96.00 4 24.00
AB 3.69 4 .92 .52
AS 55.79 16 3.48
BS 30.63 4 7.66
ABS 28.08 16 1.76
*Significant at the .05 level
(F 05,4 16= 3.01, F 7.71).
overall significance of this effect. Table 4 presents the
results of this test. It can be seen that all fry pressures
were significantly different from those in the modal range.
The pressures within each range, however, were not statis-
tically significant from each other at the .05 level.
Thus, it was found that subglottal pressure during
vocal fry is greater than during modal range phonation. It
was also found that subglottal pressure increases as a func-
tion of frequency in vocal fry over the range of frequencies
produced by the five subjects in this experiment. Subglottal
pressure decreased from the 10 to 30 percent frequencies in
the modal range. Furthermore, the pressures produced during
phonation of /i/ in the vocal fry and modal frequency ranges
are greater than the pressures produced during phonation of
/a/; however, this difference was not statistically signifi-
cant. Finally, the insertion of a mouthpiece for recording
air flow rate has no statistically significant effect upon
the pressures produced while phonating /a/; however, the
pressures produced under this condition are less than when
the mouthpiece is not in place.
Air flow rate and mode of phonation
Mean rate of air flow was obtained during phonation of
/a/ at the five phonation conditions. The results of these
measures are shown in Figure 4. In Figure 4, mean rate of
0 W 4 0
' N 4 1-
n C i- i-
0 0 a)
^ bd Q
C ,- 0
160- I I
20 i I I
I I !
20 I ..
I I I I I
SLOW MEDIUM FAST 10% 30%
VOCAL FRY RANGE MODAL RANGE
Fi gure 4. Air flow rate as a function of phonation condition.
Data points represent means for five subjects. Combined range
of subjects at each condition indicated by dotted lines.
air flow is plotted as a function of the three vocal fry
and two modal range frequencies of phonation. The extended
lines from the data points indicate the range of air flow
rates at that condition for the combined group of subjects.
From Figure 4, it is apparent that the mean air flow rate
during vocal fry phonation is lower than during modal range
phonation. While there appears to be an increaseof flow rate
as frequency increases, the ranges obtained from the subjects
tend to mask this trend. It should be pointed out that al-
though the combined air flow ranges in vocal fry and modal
phonation overlap, there was no overlap within any one sub-
ject's mean flow rates during the two types of phonation.
That is, subjects with vocal fry flow rates relatively great-
er than the mean flow rates also had relatively greater modal
flow rates than the overall means at those two conditions. The
mean flow rates for each subject can be found in Appendix C.
A one-way analysis of variance with repeated measures
across subjects (Winer, 1962) was performed to determine if
the flow rates for the five conditions of phonation were sig-
nificantly different. Table 5 summarizes the analysis indi-
cating that the overall F-ratio was significant at the tested
level. When submitted to a Newman-Keuls test to determine what
conditions accounted for the overall significance, the data
indicate that the flow rates between vocal fry and modal
were significantly different at the .05 level, however, the
Table 5. Summary of the analysis of variance
as a function of phonation conditions.
for air flow
Source of Sum of Mean
Variation Squares dF Square Fobt
Between Subjects 18.63 4 4.66
Within Subjects 76.84 20 3.84
Phonation Conditions 54.69 4 13.67 9.91*
Error 22.15 16 1.38
*Significant at the
.05 level of confidence
flow rates within the vocal fry and modal conditions were
not statistically significant from each other. These re-
sults are summarized in Table 6. It may be concluded that
the subjects in this study produced relatively low pressures
in vocal fry compared to those produced during modal range
Variations in Air Pressure and Flow in Vocal Fry Phonation
Another purpose of this study was to examine the mean
air pressure and air flow rate as they relate to repetition
rate and vocal intensity during vocal fry. As described
previously, subglottal pressures and flow rates were obtained
for three samples at each vocal fry condition. The data are
shown in Appendix D.
It should be noted that during the experiment, sub-
jects failed to produce repetition rates at the extremes of
their fry phonational ranges which were obtained prior to
the experiment. For example, the pre-experimental vocal
fry ranges (shown in Appendix A) varied from 22 to 92 Hz
while the experimental values ranged only from 40 to 90 Hz.
Nonetheless, as may be seen in Appendix D, relatively fast,
medium, and slow repetition rates were obtained for each
Relationship between repetition rate and subglottal pressure
In order to investigate the relationship between the
actual repetition rates and subglottal pressures produced by
1- 11 )-i
>, >,i~ a)tt
o o r
c a a
j ^ Q)
> k *H
-4 > H
the subjects in this experiment, a Pearson Product-Moment
Correlation Coefficient (Hays, 1963) was computed between
subglottal pressure and repetition rate. A correlation of
.27 was found for this relationship, which when tested for
significance using a t-ratio, was found to be not signifi-
cant at the .05 level.
Relationship between repetition rate and air flow rate
A Pearson-r was computed between repetition rate and
mean rate of air flow during sustained phonation of the vowel
/a/. An r equal to -.29 was obtained and found to be non-
significant at the .05 level.
Intensity variations related to subglottal air pressure and
air flow rate
In view of the large variability in the subglottal
pressure measures and the variability observed in the rela-
tive intensity measures, an investigation of the relationships
between intensity changes and both subglottal pressure and air
flow rate was undertaken. Figure 5 is a plot of each subject's
mean subglottal air pressure as a function of the relative
intensity during phonation of the vowels /a/ and /i/. Fig-
ure 6 represents this relationship for the conditions with
the mouthpiece inserted. Figures 5 and 6 show that there is
a positive relationship between subglottal air pressure and
relative intensity. The Pearson Product-Moment Correlation
Coefficient computed between these two variables was .98.
I I I I I
I I 1 I i1
17 18 19 20 21
I I !
22 23 24
Figure 5. Mean subglottal pressure as a function of the
relative intensity for five subjects phonating the vowels
/a/ and /i/ in vocal fry. The values for /i/ are under-
I I I I I I
20 21 22. 23 24 25 26 27
RELATIVE INTENSITY (dB)
Figure 6. .Mean subglottal pressure as a function of the
relative intensity for five subjects phonating the vowel
/a/ in vocal fry with the mouthpiece in place.
In Figure 7 the mean rate of air flow is plotted-as a function
of the relative intensity. It can be seen that air flow rate
generally increased as a function of the relative intensity
and that the rate of increase varied from subject to subject.
For the five subjects used in the present investigation,
it may be concluded that subglottal pressure and air flow tend
to increase as vocal fry repetition rate increases. In addi-
tion, increases in subglottal pressure and to a lesser degree,
air flow,produced during vocal fry are related to increases
in the relative intensity level of vocal fry phonation.
90 X S-2
80- El S-4
S 70 /?
u 40 -
z 30 -
0 II I I I 1 I
20 21 22 23 24 25 26 27 28
RELATIVE INTENSITY (dB)
Figure 7. Mean air flow rate as a function of the
relative intensity for five subjects phonating the
vowel /a/ in vocal fry.
The subjects in the present investigation utilized
essentially normal laryngeal mechanisms (as determined by the
subject-selection criteria) to produce vocal fry and mid-
range phonation which were found to be highly dissimilar in
terms of the aerodynamic measures accompanying them. The
results indicate that both the mean subglottal air pressure
and the mean air flow rate during vocal fry are significant-
ly different from those during low frequency modal phonation.
The subglottal pressures produced during vocal fry were
greater than those recorded in low modal phonation while
the mean rate of air flow during fry was less than that for
mid-range phonation. Subglottal pressure tended to increase
as a function of the repetition rate during vocal fry; it
decreased as the subjects went from approximately their 10
percent to their 30 percent mid-range frequencies. Air flow,
although highly variable during both types of phonation,
tended to decrease slightly as vocal fry repetition rate
increased. During modal phonation the mean air flow rates
were essentially the same for the two conditions. These
results and their relationships to previous studies are
discussed in the following paragraphs.
Subglottal Pressure and Type of Phonation
Prior to this study, no measures of subglottal pres-
sure during vocal fry phonation had been reported. Nonethe-
less the finding of relatively large subglottal pressures
departs somewhat from a previous prediction of low pressures
during vocal fry (Hollien, Moore, Wendahl, and Michel, 1966).
Several factors may be considered to account for the high
pressures. First of all, the subglottal pressure may increase
as the subject deviates from the region at which he usually
phonates. The results of this study and of Kunze's (1962)
show that the smallest subglottal pressures are in the region
usually used during normal conversation. Kunze found that as
a subject deviates either below or above the 30 percent fre-
quency in the modal phonational range, the subglottal pres-
sure tends to increase. In fact he found pressures in the
upper region of the modal range which were of the same gen-
eral magnitude as those found for vocal fry in this study.
The work of Flanagan and Langraf (1967) suggests a
possible explanation of the unexpected high pressures found
during vocal fry phonation. Using the variables of subglottal
pressure, vocal cord tension, and the properties of the con-
tacting surfaces of the folds at closure, Flanagan and Langraf
have shown that subglottal pressure is related to the duty
cycle of the wave form. Although their model is not specif-
ically related to vocal fry phonation, their calculated
values indicate that with a moderate amount of tension and
relatively long closed time, the accompanying subglottal pres-
sure is higher than at short closed times and extremely high
or low levels of tension. This model is in agreement with
earlier work by Van den Berg (1956) who suggested that when
the folds are vibrating with a long closed phase, the sub-
glottal pressures will be relatively large and produce a
rapid opening of the folds. He also suggests that if the
area of the opening is small, air flow rate could be expected
to be low. In fry the area of the opening appears to be tem-
porally small with respect to the total period (Coleman, 1968).
Therefore, pressure may build during the long closed phase and
then force apart the apparently large vocal fold mass.
The high pressures in fry may also be related to the
effects of the volume of air in the lungs at the beginning and
end of phonation. It has been reported by Draper, Ladefoged,
and Whitteridge (1957), by Kunze (1964),and by Perkins and
Yanagihara (1968) that pressure is greatest when lung volume
is greatest. In the present experiment, subjects produced
the fry and mid-range samples directly after inhalation,
that is at relatively large lung volumes. The obtained values
for air flow clearly indicate that the subject must retain
a relatively larger lung volume for a longer period of time
in vocal fry than in modal phonation. Consequently, high
pressures could be expected during sustained vocal fry pho-
nation when lung volume is relatively large. However, since
fry appears to occur normally at the ends of sentences and
phrases at relatively low lung volumes, it is not unlikely
that it could also be produced with low subglottal pressures.
There appears to be a need for additional research concerning
lung volume and subglottal pressure during vocal fry. In
particular, subglottal pressure should be measured at various
lung volumes when the speaker is reported to be phonating in
The subglottal pressures obtained during the modal pho-
nation conditions of this study are in agreement with those
obtained by Kunze (1962) although it should be remembered
that he used the 10 and 30 percent points of the phonational
range including the falsetto frequencies. Therefore, although
the actual frequencies at which his subjects were phonating
were not reported, it may be assumed that they were phonating
at frequencies somewhat higher than the present subjects.
Nevertheless, he found a statistically significant drop in
pressure amounting to approximately one centimenter of water
as the subjects went from their 10 percent frequency to their
30 percent frequency of the modal-falsetto phonational range.
In the present study there was also a significant change in
the subglottal pressure between the two points of each sub-
ject's range; however, due to the variability both within
and between subjects, it can only be said that there was a
drop in pressure as the subjects approached their 30 percent
modal frequency from their 10 percent modal frequency region.
Air Flow and Type of Phonation
The flow rates found during vocal fry phonation tend
to agree with those previously reported by iMcGlone (1967).
Using a respirometer, he obtained flow rates ranging from
2.0 to 71.9 ml/sec. The present results are in general agree-
ment with McGlone's; however, the frequency ranges over which
air flow was measured in the two studies are somewhat different.
His subjects produced repetition rates ranging from 10.9 to
52.1 pps while the subjects in the present study had a com-
bined range from 39.7 to 90.4 pps. Neither study demonstrated
a significant relationship between repetition rate and air
flow rate while both showed that flow rate is highly variable
between subjects and within consecutive samples from the same
When compared to the flow rates-obtained for modal
phonation, the fry air flow values are considerably lower.
The mid-range flow rates in this study extended from 60.0
to 177.1 ml/sec. While this range overlaps the fry air flow
range, there was no overlap for any one subject. In general,
the mean air flow rates obtained during modal phonation for
the present subjects (107.6 and 121.7 ml/sec for the 10 and
30 percent frequencies respectively) agree with those ob-
tained by Kunze (1962), Isshiki (1964) and by Perkins and
The flow rates recorded during vocal fry are generally
lower than those accompanying falsetto phonation. Isshiki's
(1964) lowest reported falsetto air flow rate was 59.4 ml/sec;
most of the fry values reported in this study and in MlcGlone's
(1967) are lower. Thus, it appears that the air flow associated
with the production of vocal fry is less than that for most
other phonational events. In this respect, the present data
support the prediction of Hollien, Moore, Wendahl, and Michel
Effects of Vowels on Modal and Vocal Fry Phonation
The effects of the vowels upon the subglottal pressure
during vocal fry appear similar to results obtained for modal
range phonation (Ladefoged and McKinney, 1963). That is, dur-
ing vocal fry the /i/ was consistently associated with greater
subglottal air pressure than the /a/. When individual samples
of /a/ and /i/ produced by the same subject and having the
same relative intensity are compared, the subglottal pressures
are greater for phonation of /i/. Ladefoged and McKinney (1963)
report that for a given sound pressure level, subglottal pres-
sures produced during sustained phonation of /i/ are greater
than those produced during /a/. Thus, as one changes the
configuration of the vocal tract from /a/ to /i/, the sub-
glottal pressure is increased; a relationship can be noted
for vocal fry phonation also.
Effects of Intensity Changes
Consideration must also be given to the effects of
the relative intensity upon the fundamental frequency, sub-
glottal pressure, and air flow rate. Since the intensity
and the subglottal pressure varied simultaneously during
vocal fry phonation, it would appear that the effect of these
two variables on the repetition rate should be considered to-
gether. Although the subjects,with the help of the experimeter
attempted to control intensity, they appeared to increase in-
tensity as repetition rate increased during the practice and
the experimental sessions. Thus, with the present subjects,
it was not possible to completely separate the increases in
repetition rate and vocal intensity with respect to increases
in the subglottal pressure.
Variability Associated with Larvngeal Operation
Although there are certain consistent patterns in the
subglottal pressures of vocal fry and modal phonation which
have been demonstrated in this and other studies, some indi-
viduals appear to vary greatly in both the air pressure and
the air flow rate which accompanies phonation. This variabil-
ity may be the result of individuals exhibiting various patterns
of pressure/flow/frequency relationships as suggested by
Smith (1954) and Van den Berg (1958). For example, since
the lungs are highly elastic, the pressure created by their
recoil varies directly with their volume. Thus, while the
subjects inhaled prior to phonation, there is no way to
ascertain that they had inhaled maximally or that they all
began with the same lung volume for each phonation sample.
Furthermore, the task was one which demanded sophistication
in control of the intrinsic and extrinsic laryngeal muscula-
ture. Failure to maintain this high level of control may
have resulted in significant changes in phontaion. That
such changes occur has been demonstrated by Hollien (1962b),
Hollien and Mloore (1960) and Damste, Hollien, Moore, and
Murry (1968). These authors have shown that within the
framework of myoelastic-aerodynamic vocal fold operation,
small changes in vocal fold length and thickness may be asso-
ciated with relatively large changes in the fundamental fre-
quency of phonation. While changes in vocal fold length as
a function of the fry repetition rate have not been demon-
strated (Hollien, Damste, and Murry, 1969), additional in-
formation is needed to understand the relationships among
such variables as vocal fold tension, thickness, and repe-
tition rate during vocal fry.
SUMMARY AND CONCLUSIONS
The purpose of this investigation was to determine
the subglottal pressure and air flow rates which accompany
vocal fry phonation. A secondary purpose was to examine
the relationships between the repetition rate in vocal fry
and the variables of subglottal air pressure and mean rate
of air flow. To achieve these purposes, simultaneous mea-
sures of the intratracheal air pressure, air flow rate, and
sound pressure during vocal fry and modal range phonation
were obtained. Recordings were made during sustained pho-
nation of two vcweis at tree vocal fry and two modal range
fundarme tal frequencies. The subgLcttal pressure sas measured
directly through a hypodermic needle inserted between the
first and second tracheal rings, air floA was measured through
a pneumotachograph inserted into the mouth, the sound pres-
sure was measured with probe tube microphone inserted into
the mouthpiece of the pneu-,otachograph. The voice signal,
subglottal pressure, and air flow were recorded simultaneously
on an oscillographic writer and FM magnetic tape recorder.
The subglottal air pressure and air flow rate recorded on
the multritrace oscillographic writer were sampled at 100 ms
intervals and converted into numerical values to estimate
the actual subglottal air pressure and air flow rate. Fun-
damental frequency was obtained from wave-to-wave measurements
of the voice signal. The data were submitted to statistical
analyses in order to test differences in pressure and flow
as they relate to fundamental frequency, vowel, and addition
of the mouthpiece to the recording apparatus.
From this investigation the following findings were
1. The rate of air flow produced during vocal
fry is less than that produced during modal
range phonation. Although flow rate is
variable, there is no overlap in the modal
and vocal fry frequency ranges for individ-
2. Air flow rate does not seem related to
frequency of phonation in vocal fry or
modal range phonation.
3. Subglottal air pressures accompanying vocal
fry phonation were found to be greater than
those accompanying low frequency modal range
4. There is a tendency for the subglottal pres-
sure to increase as vocal fry frequency
5. Subglottal pressure decreases as fundamental
frequency increases between the 10 and 30
percent frequencies in the modal range.
6. During vocal fry and modal range phonation,
the subglottal pressures for /i/ are greater
than those for /a/.
7. The addition of experimental apparatus for
recording air flow reduces*the subglottal
pressures during vocal fry and modal range
8. Increases in subglottal pressure during
vocal fry are closely related to increases
in relative intensity; however, the rate of
increase appears to vary individually.
It appears, therefore, that sustained vocal fry pho-
nation results from a relatively low mean rate of air flow
compared to that in modal phonation. This conclusion, which
is also supported by data obtained using a respirometer
(McGlone, 1967),may relate to the characteristically long
closed phase of the vocal fry glottal wave form during which
time there is no air flow. It might also be expected that
flow rates accompanying vocal fry during any type of phona-
tion would also be relatively low since fry is most often
heard after lung volume is reduced, for example, at the ends
The subglottal pressures found during sustained vocal
fry phonation were relatively high in comparison with the
pressures in adjacent low modal range frequencies. However,
these pressures were of the same general magnitude as those
reported by others (Kunze, 1962; Isshiki, 1964) for phonation
at the upper portion of the phonational range. This appears
to suggest that as one departs from low frequency modal pho-
nation, glottal resistance increases and there is a need to
generate higher pressures to sustain vocal fold vibration.
This relationship observed for sustained phonation, however,
need not be the case during other types of phonation. For
example, vocal fry usually occurs briefly at the ends of
sentences having downward inflections when both lung volume
and intensity also decrease. Thus, it might be hypothesized
that relatively high pressures may not be required to produce
vocal fry. Indeed, it might be expected that these subglottal
pressures would not differ greatly from those directly pre-
ceding them in modal phonation
SC- 3C Gml a(
co0 N N ci
0 0 c
l 0 a)
7 -1 H
0 0) -
:5a > =a
0 0 C
C > i
3 3 0 4
M O aD
H 0 4 3E
ft 0 0 e
00 0 n
Snj U- )
ft) a) a)
1 0 z
O ) 4
o 0 0 -4 3
c D 0 C
- = -0 r
3 0S 0
D > c. 0)
ft a a>
3) 3) 0 -
03 M 03
0, 43 0
d 0 0-
0 CD f .
SC m m
> -4 44
Recording sequence followed throughout the experiment
A. Needle inserted; air flow recording apparatus disconnected.
Pressure recordings only
1. Slow fry* /a/
2. Slow fry /i/
3. Medium fry /a/
4. Medium fry /i/
5. Fast fry /a/
6. Fast fry /i/
7. Modal 10% /a/
8. Modal 10% /i/
9. Modal 30% /a/
10. Modal 30% /i/
B. Needle inserted; air flow recording apparatus inserted.
Pressure and flow recordings
1. Slow fry /a/
2. Medium fry /a/
3. Fast fry /a/
4. Modal 10% /a/
5. Modal 30% /a/
*The terms, slow, medium, and fast refer to the reference
signals. The samples which the subjects produced were later
grouped tc fit these categories.
PROCEDURES FOR MEASURING SUBGLOTTAL
AIR PRESSURE AND RATE OF AIR FLOW
PROCEDURES FOR MEASURING SUBGLOTTAL
AIR PRESSURE AND RATE OF AIR FLOW
Measurement of the subglottal air pressure and rate
of air flow'were obtained from the analog output of a Multi-
trace Oscillographic Writer. The analog output is presented
schematically in Figure 8. This figure shows a reference
line (set to atmospheric pressure), the subglottal pressure
trace (Line A), the air flow trace (Line B), and the vertical
time lines placed at 100 ms intervals. To obtain the mean
subglottal pressure from this oscillogram, the distance from
the Reference Line to Line A was measured at 100 ms intervals.
The distance in centimeters at each point was then converted
to cm H20 by multiplying it by the calibration factor. The
average of 30 such measures resulted in the mean subglottal
air pressure for that particular sample.
The mean rate of air flow was obtained similar to
that for the pressure. The distance from the Reference Line
to Line B was measured in centimeters at 100 ms intervals and
multiplied by the appropriate calibration factor to produce
a flow rate in ml/sec.
-. (t- T
3 -* .c m
co 3 0 -1
INDIVIDUAL AND COMBINED MEANS OF THE FUNDAMEN-
TAL FREQUENCY, SUBGLOTTAL AIR PRESSURE, RATE
OF AIR FLOW, AND RELATIVE INTENSITY FOR PHONA-
TION OF /a/ AND /i/ AT FIVE FREQUENCY REGIONS
c- inco C)o00o ) 0)o
-4 -4 4 -i i i -4
I00n 0 C) c C n c N I
0 0> IT m 0
I0 I N V C )D
-t l c c) CO0D i C)0
In CIsn C)n~vC C 0C
N, > N >4 >4N >
3 .4 ...I .- I + 4-'
F- o i c3 c
to 0 a h
00 0 LC- 00 0 0 m
co-( -4m-oo4 m-
f-4 r-4 C ^ C Ul IC L ) m (0 ( (N 0) m o
1- N -4 - N-
CO V cO C M C 00 C C l
v o U, E- C-
^r ^ ^ iO t ^- c
0 0 -jt T-0-
oa 0 a :9a i
a '-) a
N N NS Nn N
0 OCOC (a)))
-4-4 a~a ( (
M CM CO N-
(a -S C,
0l 0 -4
Cl Cl CQ
^ C C cO CC) CO N Coc G) 0 Cli Cl T
- i v
c" Cj C1
N 00 0
t-O t- 0 C 0000
Cl l C
P ^ 0
coo C) 0 o co
cr C- V 0 co r u ) c v -0
0Cl c n cc n 0 xc
-4 co in u c -o
,a Ca ss a 0-
w 0 0 0
L, Co 'n 0 C C00 -4 cl ClC
(e ) -r:
i a)Ci ) Ca
lo ^ o ooo
-4 0o c0 0Q 0 0n 0 m (C
M C C C CJ M 0 C
h -H h
0 0 i 0 0 0
S ( 0o o0
SC CM 0o
N N >r
-4 ->61 a)
3 3 :;- l
ro o 4,
4 -1 a. a
4,( M 3 3
0 n 0
.-4 -4 -4
N N N
,- s 0s
00N '0(CN0 0n -I00~'
O 0m 0 om oW o 0
t~mO ^r CD~O- -
rlrlc O ^ o ls o c~co o t]c
C] 0 Q 0 C to 0 U0 0
h N N N 0 NM I C' C') 0 ) o o Cn)
r i f < o oi
* . . t . . .^ c M
- 0 (C 0 0 0 -4 0 N) t- m']
SC 0 ) o) -) E- o 00 0
.-( ^ (O o t D 'ia -
^ ^ y in m -'f-a o '
>4 >4 N> N
-3 -H n 4n
0 0a r)r
Cl c) V)
.-4 NC] CM
10 lr- C 33
0 0^ N~ 411
*3 0 a) 6) ^
a 4 (
.N M iD* 000 Mi V
0 00 0 0
o0, o oo
N0 N X-o
00 0 C )
Cn 0 COi (o
cd *H r H '
H H ) ( 0o
Ho H c Ho o H o '
C, cc* M* c"I M, V
v n v 0 n
3 00 0 0 0 0 to LO o L- ^1 X L 7
m m ~ o D OO 'H
CO OOCOC> ~lO 5^0 lr .0 |^ r 3'Hn '
c 0 IX
Brackett, I., An analysis of the vibratory action of the
vocal folds during the production of tones at se-
lected frequencies. Ph.D. dissertation, North-
western University, 1947.
Coleman, R., Decay characteristics of vocal fry, Folia
Phoniatrica, 15, 256-263 (1963).
Coleman, R., Pulse patterns of vocal fry, Paper presented at
Forty-Fourth Annual Convention of the American Speech
and Hearing Association, Denver, 1968.
Damste, H., Hollien,. H., Moore, P., and Murry, T., An x-ray
study of vocal fold length, Folia Phoniatrica, in
Draper, M., Ladefoged, p., and Whitteridge, D.. Expiratory
muscles involved in speech, Journal of Physiology,
138, 17p- S1 p (1957).
Draper, M., Ladefoged, P., and Whitteridge, D., Respiratory
muscles in speech, Journal of Speech and Hearing
Research, 2, 16-27 (1959).
Fant, G., Personal communication to H. Hollien and R. Wendahl
(1964), in Journal of Acoustical Society of America,
43, 509 (1968).
Flanagan, J., and Langraf, L., Self-oscillating source for
vocal-tract synthesizers, Preprints of 1967 Conference
on Speech Communication and Processing, 54-62 (1967).
Fry, D., Stead, W., Ebert, R., Lubin, R., and Wells, H., The
measurement of the intraesophageal pressure and its
relationship to intrathoracic pressure, Journal of
Laboratory and Clincical Medicine, 40, 664-673 (1952).
Hardy, J., Air flow and air pressure studies, American Speech
and Hearing Association Reports, 1, 141-152 (1965).
Hays, W. L., Statistics for Psychologists, New York: Holt,
Rinehart, and Winston, Inc. (1963).
Hollien, H., Personal communication, December, 1968.
Hollien, H., Vocal pitch variation related to changes in
vocal fold length, Journal of Speech and Hearing
Research, 3, 150-156 (1960).
Hollien, H., The relationship of vocal fold length to vocal
pitch for female subjects, Proceedings of the XII
International Congress of Logopedics and Phoniatrics,
Scoula Tipografica dell'Ente Nationale Sordomuti,
Padua, 38-43 (1962).
Hollien, H., The relationship of vocal fold thickness to
absolute fundamental frequency of phonation, Pro-
ceedings of the Fourth International Congress of
Phonetic Sciences, The Hague: Mouton and Company,
Hollien, H., Vocal fold thickness and fundamental frequency
of phonation, Journal of Speech and Hearing Research,
5, 237-243 (1962b).
Hollien, H., Coleman, R., and Moore, P., Stroboscopic lami-
nagraphy of the larynx during phonation, Acta Oto-
laryngologica, 65, 209-215 (1968).
Hollien, H., and Curtis, J., A laminagraphic study of vocal
pitch, Journal of Spee'ch and Hearing Research, 3,
Hollien, H., Damste, H., and Murry, T., Vocal fold length
during vocal fry phonation, in press, Folia Phoniatrica,
Hollien, H., and Michel, J., Vocal fry as a phonational
register, Journal of Speech and Hearing Research, 11,
Hollien, H., and Moore, P., Measurements of the vocal folds
during changes in pitch, Journal of Speech and Hear-
ing Research, 3, 157-165 (1960).
Hollien, H., Moore, P., Wendahl, R., and Michel, J., On the
nature of vocal fry, Journal of Speech and Hearing
Research, 9, 245-247 (1966).
Hollien, H., and Wendahl, R., Perceptual study of vocal fry,
Journal of the Acoustical Society of America, 43,
Hubbs, J., The new pulse, Journal of the Precision Measure-
ments Association, 39, 109-116 (1966).
Isshiki, N., Regulatory mechanism of the pitch and volume of
the voice, Fibi Inkoka Rinsho Kyoto, 52, 1065-1094
Isshiki, N., Regulatory mechanism of voice intensity varia-
tion, Journal of Speech and Hearing Research, 7,
Irwin, J., A study of the relationships between certain
laryngeal mechanisms and vocal pitch. M.A. thesis,
The Ohio State University, 1940.
Kunze, L., An investigation of the changes in subglottal
pressure and rate of air flow accompanying changes
in fundamental frequency, intensity, and vowels and
voice registers in male speakers, Ph.D. dissertation,
State University of Iowa, 19b2.
Kunze, L., Evaluation of methods of estimating subglottal
air pressure, Journal of Speech and Hearing Research,
7, 151-163 (1964).
Ladefoged, P., Subglottal activity during speech, Proceedings
of the Fourth International Congress of Phonetic Sci-
ences, The Hague: Mouton and Company, 73-91 (1962).
Ladefoged, P., Draper, M., and Whitteridge, D., Syllables and
stress, Miscellanea Phonetica, 3, 1-14 (1958).
Ladefoged, P., and McKinney, N., Loudness, sound pressure, and
subglottal pressure in speech, Journal of the Acoustical
Society of America, 35, 454-460 (1963).
Lindquist, E. F., Design and Analysis of Experiments in
Psychology and Education, Boston: Houghton-Mifflin
Lieberman, P., Direct comparison of subglottal and esophageal
pressure during speech, Journal of the Acoustical
Society of America, 43, 1157-1165 (1968).
McGlone, R., Air flow during vocal fry, Journal of Speech
and Hearing Research, 10, 299-304 (1967).
Mead, J., and Whittenberger, J., Physical properties of the
human lungs measured during spontaneous respiration,
Journal of Applied Physiology, 5, 779-796 (1953).
Michel, J., and Hollien, H., Perceptual differentiation of
vocal fry and harshness, Journal of Speech and Hearing
Research, 11, 439-443 (1968).
Moore, P., and Von Leden, H., Dynamic variations of the
vibratory pattern in the normal larynx, Folia
Phoniatrica, 10, 205-238 (1958).
Morner, J., Franeson, P., and Fant, G., Voice register
terminology and standard pitch, Speech Transmission
Laboratory Quarterly Progress Beport, No. 4 (1963).
Perkins, W., and Yanagihara, N., Parameters of voice pro-
duction 1. Some mechanisms for the regulation of
pitch, Journal of Speech and Hearing Research, 11,
Rossier, P., Buhlmann, A., and Wiesinger, K., Respiration
Physiology Principles and Their Clinical Applications,
translated by P. Luschinger and K. Moser, St. Louis:
C. V. Mosby Co. (1954).
Smith, S., Remarks on the physiology of the vibrations of the
vocal cords, Folia Phoniatrica, 6, 166-178 (1954).
Sonninen, A., Is the length of the vocal cords the same at
all different levels of singing? Acta-Otolaryngologica,
Supplement 118, 219-231 (1954).
Sonninen, A., The role of the external laryngeal muscles in
length adjustment of the vocal cords in singing, Acta-
Otolaryngologica, Supplement 130, 102 (1956).
Timke, R., Von Leden, H., and Moore, P.., Laryngeal Vibra-
tions: measurements of the glottic wave, Part 2,
Physiologic variations, Archives Otolarvngologica,
68, 438-414 (1958).
Van den Berg, J., Direct and indirect determination of the
mean subglottal pressure, Folia Phoniatrica, 8, 1-24
Van den Berg, J., Myoelastic-aerodynamic theory of voice
production, Journal of Speech and Hearing Research,
1, 227-240 (1958).
Van den Berg, J., Physiology and physics of voice production,
Acta-Physiologica et Pharmacologica Neerlandica, 5,
Wendahl, R., Moore, P., and Hollien, H., Comments on vocal
fry, Folia Phoniatrica, 15, 251-255 (1963).
Wendler, Von, Zur messung der stimmlippenlange, Zeipschlift
fur Laryngologie, Rhinologie, Otologie und Ihre
Grenzgebiepe, 44, 162-173 (1964).
Winer, B. J., Statistical Principles in Experimental Design,
New York: McGraw-Hill Book Co. (1962).
Thomas Murry was born September 10, 1943, at
Sewickley, Pennsylvania. In June, 1961, he was graduated
from Ambridge Senior High School, Ambridge, Pennsylvania.
He received a Bachelor of Science degree with a major in
Speech Pathology in May, 1964, from Indiana University of
Pennsylvania. In September, 1964, he enrolled in the
Graduate School of The Ohio State University where he was
a recipient of a Neurological and Sensory Disease Fellow-
ship. In May, 1965, he was graduated from The Ohio State
University with a Master of Arts degree with a major in
Voice Science. He enrolled in the Graduate School of the
University of Florida in September, 1966. He held a re-
search assistantship there until September, 1968, followed
by a faculty appointment as Interim Instructor until De-
This dissertation was prepared under the direction
of the candidate's supervisory committee and has been ap-
proved by all members of that committee. It was submitted
to the Dean of the College of Arts and Sciences and to the
Graduate Council, and was approved as partial fulfillment
of the requirements for the degree of Doctor of Philosophy.
March 25, 1969
Dean, Colle o A ts and Sciences
Dean, Graduate School