ABSTRACT
AN INVESTIGATION OF THE TRANSMISSION
OF RELATIVE SOUND PRESSURES OF
PHONATED VOWEL SOUNDS AT

SELECTED LOCATIONS OF
THE HEAD AND NECK

BY

Charles Lonegan

Many investigators have examined the nature of
vibration through tissues of the body of a speaker.
Research in this area indicated that vibrations were more
intense at some anatomical locations than others. There
has not been consistent agreement in results of investiga-
tions.

The purpose of this investigation was to examine
the relative sound pressures of vowel sounds uttered
by a speaker as they were transmitted through the tissues
of his head and neck. The study was directed at investigat-
ing, experimentally, the difference in the transmitted
relative sound pressures among twelve vowel sounds picked
up at sixteen anatomical locations, employing subjects with
three distinctly different body types.

Six adult male speakers were employed. Two subjects

were selected to represent each of three different body

Charles Lonegan

physiques: ectomorphs, mesomorphs, and endomorphs. The
subjects all utilized the General American English dialect.
Twelve vowels were selected for this study. Sixteen
anatomical locations were specified as sites to be investi-
gated. These were at various distances from the larynx,
representing differences in tissue composition. Sound
pressure levels and fundamental frequencies of all utterances
were held constant by all speakers, monitored at 12 inches
from the mouth.

A helmet and shield which had minimal contact with
the surface of the head and neck was designed to secure
four accelerometers brought into contact with the anatomical
locations.

The accelerometers employed were integrated-circuit-
piezoelectric quartz transducers sensitive to subtle
vibrations, with an almost linear frequency curve response
throughout the speech range. The accelerometers transmitted
the vibrations picked up to a tape recorder; simultaneous
responses from four anatomical locations were recorded.

The four accelerometers were shifted to other positions
in the helmet and the procedure repeated until sixteen
locations had been recorded.

Relative sound pressures of tissue transmitted
vibrations were analyzed by means of a power level
recorder; a graphic display of each sample was obtained

from which mean relative sound pressures were determined.

C»

SC

. .
«Q

in
7 so

 

Charles Lonegan

These data were treated by a three factor analysis
of variance in which the factors were: vowels, positions,
and body types to test for significances. In addition,
the resonance characteristics of the helmet and shield and
the frequency response curves of the accelerometers attached

to the helmet and shield were obtained experimentally.

Conclusions

 

l. The vowels /i/ and /u/ were transmitted through
the tissue of the head and neck with higher relative sound
pressures than any other vowels. .The vowels /a/ and /a¥
were transmitted through these tissues with lower relative
sound pressures than any other vowels.

2. The anatomical positions at the top-center of
the skull, the forehead, the neck near the larynx, and
the mandible transmitted vowel sounds with higher relative
sound pressures than any other locations.

3. On the basis of the relative sound pressures
transmitted, the sixteen anatomical locations fell into
two basic groups: Group I--locations toward the midline
and front of the skull; Group II--sites remote from the
larynx.

4. Highest relative sound pressures were transmitted
for the vowels /i/ and /u/ at the positions on the top-
center of the skull and on the forehead.

5. Differences between results obtained by this
experiment and previous studies may be attributed to refine-

ments in instrumentation and procedures employed, as well

Charles Lonegan

as to poSsible interactions between transmitted signals.
and resonance characteristics of the supporting system
which held the accelerometers.

6. Results suggest that contact microphones are
questionable as efficient means of transmitting speech
signals. However, further research is needed to determine
the interactions of signals with the supporting system and
spectral characteristics of transmitted signals through
various tissues. Intelligibility of tissue transmitted

signals should be also studied in the future.

 

AN INVESTIGATION OF THE TRANSMISSION
OF RELATIVE SOUND PRESSURES OF
PHONATED VOWEL SOUNDS AT
SELECTED LOCATIONS OF

THE HEAD AND NECK

BY

Charles Lonegan

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Audiology and
‘ Speech Sciences

1972

06" ACKNOWLEDGMENTS

The writer is indebted to several people whose
guidance, cooperation, and encouragement made the
completion of this study possible. Their interest in
this study is gratefully acknowledged.

To Drs. Herbert J. Oyer, Oscar Tosi and William
Bradshaw Lashbrook, staff members of the Department of
Audiology and Speech Sciences at Michigan State University,
for their guidance.

To Mr. Donald Riggs, technician in the Department
of Audiology and Speech Sciences at Michigan State University,

for assistance in the instrumentation.

ii

TABLE OF CONTENTS

ACKNOWLEDGMENTS . . . . . . . . .

LIST OF TABLES. . . . . . . . V. .

LIST OF

FIGURES O O O I O O O O 0

LIST OF APPENDICES . . . . . . . .

Chapter

I.

II.

III.

IV.

INTRODUCTION 0 O O O O O O 0
Purpose of the Study . . . . .
Importance of the Study . . . .
Definitions . . . . . . . .

REVIEW OF THE LITERATURE. . . .

Chest Resonance. . . . .

Transmission of Sound through Body Tissues.

Development of Instrumentation. .
Summary . . . . . . . . .

EXPERIMENTAL PROCEDURES . . . .

Subjects . . . . . .
Equipment and Materials . . .
Construction of Helmet and Shield.

Vowel Sounds and Anatomical Locations

Recording of Experimental Data. .
Analysis of Experimental Data . .

RESULTS AND DISCUSSION . . . .
Descriptive Analysis of Results .
Results of Statistical Tests . .

Retest Results . . . . . . .
Discussion . . . . . . . .

iii

Page,

ii

vi

vii

15
23
25

32

32
34
36
42
42
48

51

51
56
62
65

Chapter Page
V. SUMMARY AND CONCLUSIONS . . . . . . . . 71
Summary . . . . . . . . . . . . . 71
Conclusions . . . . . . . . . . . . .74
Suggestions for Future Research . . . . . 76

BIBLIOGRAPHY O I O O O O O I O O 0 O O O 80

APPENDICES O O O O O O O O O O O O O O O 84

iv

LIST OF TABLES

Table Page

1. Outline and summary of early research in
tissue-transmitted sound. . . .' . . . . 28

2. Comparison of results of pertinent studies

which investigated relative amplitude of

vibration of skin during speech production. . 30
3. Resonance responses of accelerometers

attached to helmet. . . . . . . . . . 4O

4. Anatomical locations examined . . . . . . 43
5. Fundamental frequency for each subject . . . 48
6. Summary of analysis of variance comparing

vowel, positions, and body type . . . . . 57

7. Means of twelve vowels ranked in descending

order of relative sound pressure level . . . 59
8. Display of transmitted sound pressure of

each vowel as compared to every other vowel . 60
9. Means of sixteen positions ranked in

descending order of relative sound pressure . 61
10. Display of each position compared to every

other position for differences between their

means 0 o o o o o o o I o o o o o 0 ~ 63
11. Summary of analysis of variance comparing

vowel, positions, and body type for retest
data 0 O O O O O O O O O O O O O 6 4

12. Comparison of the results of the Moser and
Oyer study to the results of the present
study. Position ranked in descending order
of relative sound pressures. . . . . . . 69

Figure

1.

LIST OF FIGURES

Page
Photograph of helmet and shield with
accelerometers in position . . . . . . . 39
Block diagram of system used to test
response of accelerometers . . . . . . . 41
'Schematic representation of anatomical
locations. . . . . . . . . . . . . 44

Block diagram of accelerometer and recorder
equipment. . . . . . . . . . . . . 47

Mean relative sound pressure levels for each
vowel for the groups of ectomorphs, endomorphs,
mesomorphs, and all subjects, regardless of
position . . . . . . . . . . . . . 53.

Mean relative sound pressure levels for each
position for the groups of ectomorphs,

mesomorphs, endomorphs, and all subjects,
regardless of vowel . . . . . . . . ..- 55

vi

 

Appendix
A.

B.

LIST OF APPENDICES

Sheldon's Classification System . . . .

Descriptive Data for the Six Subjects
Employed in the Study. . . . . . . .

Mean Age, Height, Weight for Each Body
Classifications. . . . . . . . . .

Frequency Response Curves of Tape Recorders

Experimentally Derived Curves for Acceler-
ometers . . . . . . . . . . . .

Statistical Methods Employed . . . . .
Original and Retest Scores . . . . . .
Mean Scores for Each Body Type. . . . .

Mean Scores for All Subjects Regardless of
Body Types . . . . . . . . . . .

vii

Page

85

88

90

92

94
99
105

112

116

CHAPTER I

INTRODUCTION

Speech scientists have long been interested in the
characteristics of speech sounds and the contribution of
resonation and forced vibration to the overall character-
istics of speech. In general, most of the research has
been focused upon the measurement of communication systems
such as the telephone and the radio.

Speakers often note that certain structures of the
body vibrate during speaking and singing. Singing
teachers, speech pathologists, and teachers of the deaf
have instructed their pupils to use these vibrational-
sensations as a feedback aid in achieving specific sounds
and vocal qualities. This has generally been attempted
with little understanding of the nature of the phenomenon
of tissue vibration.

Simon and Keller note that many investigators in
the last one hundred years became interested in the nature

of tissue vibration during speech.1 This interest stemmed

 

lClarence T. Simon and Franklin Keller, "An Approach
to the Problem of Chest Resonance," Quarterlijournal of
Speech Education, XIII (1927), p. 433.

 

 

from early research into the nature of vocal cord vibra-
tion and the resonance characteristics of the cavities
of the vocal tract. ,

Early investigators believed that the vibration
of bones and skin surfaces during voice production resulted
from resonance factors in the vocal tract. It was
generally concluded that vibration of tissue during voice
production contributed to the overall quality of the voice.
Bones and skin of the thorax, neck, and head were thought
to vibrate because the cavities underlying these structures
had resonant frequencies similar to the frequency of
vibration of the vocal cords.

As refinements in measurement procedures and
instrumentation were made, near the turn of the century,
investigators began to suspect that tissue vibration
during speech was the result of forced virbation and not
resonance. These investigators supported the View that
the vibration pattern of bone and skin tissue probably
added little to the overall quality of the voice. The
pattern of tissue vibration of the thorax, neck, and head
during voice production appeared to be different in
nature from the pattern of vibration of the laryngeal
source, thus indicating that tissue near the larynx was
forced to vibrate because of its proximity to the original

source and not because of similar resonance frequencies.

The present study was designed to investigate
further the transmission of the relative sound pressure
of vowels through various tissues of the head and neck
using more refined procedures and equipment than was

available to former researchers.

Purpose of the Study

 

The purpose of this study was to examine the overall
relative sound pressures of vowel sounds produced in the
vocal tract as they were transmitted through tissues of
the head and neck. The study was directed at investigating ,
experimentally, the differences in relative overall intensity
among twelve vowels recorded at sixteen anatomical loca-
tions, employing subjects with three different body types.
The sound pressure level and the fundamental frequency of
all vocal utterances were kept constant. The following
questions were posed for this investigation:

1. Is the relative overall sound pressure trans-
mitted through tissue for each vowel dependent
on body type and anatomical location?

2. Is relative overall sound pressure trans-
mitted through tissue for each body type
dependent upon the vowel and anatomical location?

3. Is the relative overall sound pressure trans-
mitted through tissue for each anatomical
locations dependent on specific vowel and

body type?

Importance for the Study

 

The present study examines the differences among‘
the relative sound pressure of twelve vowels as they are
transmitted through sixteen anatomical locations of sub-
jects representing three different body types. This study
could provide additional information concerning the use
of contact microphones. 'Indeed, Snidecor notes that in
some military, government, and scientific operations, the
use of a contact pick up transducer would be more desirable
than the conventional type of microphone held in front of
the mouth.1 In aviation, space exploration, and ocean
diving, conventional microphones are likely to pick up
and transmit not only speech signals but also competing
noises such as the hissing of the oxygen supply equipment.
The speech signal is often masked out or distorted by this
noise. A contact microphone would be sensitive only to
tissue vibration and not to competing environmental noises.

A contact microphone would also free the mouth area
from obstruction, allowing greater movement of the
articulators. In addition, with the mouth no longer

blocked by a microphone, observers would be able to read

1John C. Snidecor, Irving Rehman, and David D.
Washburn, "Speech Pickup by Contact Microphone at Head and
Neck Positions," Journal of Speech and Hearing Research,
II (1959), pp. 277-281.

 

lips as well as to employ sound cues in situations, such
as extreme noise, in which both visual and auditory cues
would aid communication.

Neely and Forshaw suggested that in diving
operations, for example, a contact microphone would be
less sensitive to damage by sudden changes in air
pressure or shock waves.1 Also, a contact mircophone
would be less likely to damage the mouth, teeth, and
other facial structures in the case of sudden shock
resulting from falls, crashes, etc.

Moser and Oyer summarized previous investigations
which had been conducted concerning the relative intensity
of speech sounds transmitted through human tissue.2 Work
in this area has indicated that some anatomical locations
of the head and neck are more efficient than others in
transmitting speech sounds. However, much of the early
work in this area employed contact transducers that were
imprecise for speech transmission. In addition, some
early procedures employed relied upon subjective evalua-

tion, poor control of application of the transducer to the

 

1Keith K. Neely and S. E. Forshaw, "Speaking and
Listening through the Head: I. The Intelligibility of
Speech Recorded in Quiet and at Different Positions of
the Head and Throat," Journal of Auditory Research, V
(1965): pp. 151-157.

 

2Henry M. Moser and Herbert J. Oyer, "Relative
Intensities of Sounds at Various Anatomical Locations of
the Head and Neck during Phonation of the Vowels,"
Journal of the Acoustical Society of America, XXV (1958),
pp. 275-278. - '

surface of the skin, investigation of few anatomical sites,
use of one or two subjects without regard for differences
in body physique. This has resulted in considerable lack
of agreement in results.

Early investigations have indicated that certain
refinements in instrumentation and procedures may result
in more valid and reliable data. 'It would appear that a
sensitive transducer attached to the proper location
could provide an efficient method of transmission of speech
sounds from a speaker to a listener.

The present study is an attempt to control the
variables indicated above and to obtain data resulting
from refinements in instrumentation and procedures

previously emploYed.,

Definitions

 

Certain terms employed in this report are defined
in this section for purposes of clarity. Attempt is
made to indicate how each of these terms is used in regard
to the procedures and instrumentation involved in the
present study.
2212.1.

The vowels employed in this study are the twelve
vowels /i/. /I/. /e/. / /. /a/. / /. /A/. / /. /e/. /o/.
/u/, and / / which are found in the General American

English dialect. No diphthongs were included.

Transducer

 

A transducer is a device used to convert the
mechanical vibrations of the head and neck into electrical
signals. In the present study, the transducer employed
was an integrated-circuit-piezoelectric quartz transducer

referred to herein as an accelerometer.

 

Relative Sound Pressure

 

The sound pressure of a signal determined by its
relationship to a calibration tone which has a known
sound pressure. The calibration tone is referred to as
zero decibels and acts as the reference point for compari—
son. The reference tone employed in this study was a
1000 Hz pure tone of 10 milivolts. In the present study,
relative sound pressure could range from -30.0 dB to
+20.0 dB because the graph paper employed for graphic
analysis had a grid whose range was 50 dB.

Body Type

 

Body type refers to classification of the physique
of each subject according to the somatotyping system
devised by Sheldon.1 Subjects were classified into the

general body types of ectomorph (thin, frail physique),

 

1W. H. Sheldon, Atlas of Men (New York: Harper
and Brothers, 1954). W. H. Sheldon, The Variations of
Human Physique (New York: Hafner PublisHIng Company,

 

 

 

mesOmorph (muscular, solid physique), and endormorph

(soft, round physique).

CHAPTER II

REVIEW OF THE LITERATURE

This chapter represents a review of the theoretical
and research literature relevant to this investigation.
A discussion of early considerations of the concept of
"chest resonance" is presented initially. A brief
history of the contemporary interest in the transmission
of sound through living human tissue is included. The
discussion then focuses on recent research into the
precise measurement of vocal sound picked up at various
anatomical locations of the human body. 'This review also
includes a consideration of refinement in pick up, filter,
and amplification equipment that led to the present
investigation. The chapter concludes with a summary of
the review in terms of the development of the present

investigation.

.CHEST RESONANCE

 

During the early part of the twentieth century,
voice and speech scientists were interested in the
investigation of the concept of "chest resonance." The

relationship between this phenomenon and the intelligibility

10

of speech and the quality characteristics of speech was
of particular interest.

Throughout the nineteenth century, voice scientists
in general had attributed great importance to the contri-
bution of vibration of the walls of the chest to voice
quality. It was not until the work of Helmholtz that
researchers began to divide into opposing camps regarding
the importance as well as the existence of chest resonance.1
Simon and Keller reported that

After these many years of bandying terms, the
division of opinion is rather sharply marked; writers
either accepted chest resonance enthusiastically as
a fortunate and much-to-be-desired phenomenon, or
else reject it flatly as an anatomical and acoustical

impossibility.

Early Support

 

Many early investigators vigorously supported the
existence and importance of chest resonance. In 1909,
Curtis supported the view that the chest cavity helped to
fortify the original vocal tone and to improve overtone

production.3 Fillebrown, 1911, believed thatthe largest

 

1Simon and Keller, ". . .the Problem of Chest
Resonance,“ p. 433. '

21bid., p. 433.

3H. H. Curtis, Voice Training and Tone Placement,
(New York: _Appleton Company, 1909), p. 66.

11

air chamber in the body was the chest which also acted as
a principal resonating cavity.1 In theface of consider-
able controversy, Scott concluded in 1915 that while
chest resonance wasdifficult to prove scientifically, his.
own senses of "hearing and feeling" testified to its

. f 2 '
existence.

Early Opposition

 

Many investigators argued that chest resonance
was nonexistent. Muckey felt that the chest was a closed
cavity during voice production and could not act as a
resonator.3 Aiken supported the same ndtion and stated
that only open cavities such as the mouth, nose, and
pharynx could act to resonate vocal sound.4 .Woolbert and
Weaver in 1922 stated that chest resonance was impossible
because the function of the thorax was to provide the
breath stream for Speech and could not, at the same time,
contribute toward vocal sound resonation.5 Early investi-
gations into chest inhalation lead later researchers to
examine the vibration of bones and tissue remote from the

chest.

 

1Thomas Fillebrown, Resonance in Singing and
Speaking (Boston: Oliver Diston Company, 1911), pp. 49-50.

2John R. Scott, The Technique of the Speaking
Voice (Missouri: Stephens Company, 1915f: p. 214.

3Floyd S. Muckey, Natural Method of Voice
Production (New York: Scribner's Sons, I915)} P. 59.

4W. A. Aiken, The Voice (London: Longmans, Green,
and Company, 1920), p. 36.

5M. Woolbert and A. Weaver, Better Speech (New York:
Harcourt, Brace, Company, 1922), p. 83.

 

 

 

12

Early Experimental Studies

 

In 1927, Simon and Keller conducted the first‘
experiment designed to pick up and record the patterns of
vibration of the bones during voice production. The
experiment was designed primarily to settle the long-standing
controversy concerning chest resonance.

Simon and Keller felt that much of the controversy
stemmed from ambiguity and misunderstanding of the term
"resonance."l (Physicists do not mean that resonance
occurs only in a cavity or chamber, but in any elastic
system such as air or elastic fluid. Resonance requires
an elastic system which is tuned to the sound being amplified,
whereas forced vibration occurs with a solid mass which
has no tuned relationship to the sound being amplified.

Simon and Keller suggested that while the chest probably
does not act as a resonator, it might be an area of forced
vibration.2

Simon and Keller used a special carbon button
stethosc0pe which was constructed to be sensitive to
vibrations picked up from contact with a surface and not
to be sensitive to air-conducted vibrations. The

stethosCoPe was placed against seven anatomical locations

 

1Simon and Keller, ". . .the Problem of Chest
Resonance," pp. 432-439. '

31bid., p. 435.

13

(right wing of thyroid cartilage, sternum at second rib,
medial sections of the left and right seventh ribs, sternum
at the sixth rib, medial portion of the clavicle on left'
and right sides) while the subject intoned the vowel /o/.
The sound was electrically converted to light waves which
were photographically recorded. These photographs were
measured carefully and plotted on graphs. From the
graphs the following general conclusions were made.
Chest vibration
consists of forced vibrations, caused by mechanical
shaking of the vocal cords; forced vibrations whose
frequency corresponds to the frequency of the
fundamental of the cord tone, and which are in no
way to be considered as indicative of any resonance
occurring with the cavity.1
In 1931, Wise became concerned with the problem of
chest resonance and set out to determine how the thoracic
cavity was caused to vibrate during voice production.
Preliminary studies were concerned with the conductivity of
various forms of body tissue and the routes that sound
waves travelled through the body. Wise employed a form
of the stethosc0pe called a phonendosc0pe which picked up
only vibrations of the surface it was placed against. By
sounding a tuning fork and placing it at one point of the

body (or by having the subject sing vowel sounds) and

placing the phonendosc0pe at another point, it was found

 

lIbid., p. 439.

14

that vibrations were conducted for long distances-through
the body. All forms of tissue were found to conduct
sound but with varying degrees of efficiency.

- On the basis of subjective aural judgments, the
conductile efficiency of various tissues was ranked-ordered
from most efficient to least efficient as follows:

(1) bone, (2) tendon, (3) tense muscle, (4) lax muscle,
(5) non-muscular tissue.

Wise concluded that sound vibrations reach the
thoracic cavity through several diverse paths, the
vertebral column being the most significant. ‘Each sound
path adds to the total vibration which is given off in
the form of measurable sound. The overall pattern of
vibration of the chest cavity results from forced vibration
and not fromresonance.l

Lindsley investigated the psychOphysical deter-
minants of voice quality. A vibration microphone was
employed to pick up signals from the chest. These signals
were then amplified, recorded, and superimposed upon the
subjects' own air-conducted speech sound, thus adding
chest vibration to normal speech. On the basis of sub-

jective judgments of voice quality, it was discovered

 

 

1Claud M. Wise, "Chest Resonance,“ QuarterlyJournal
of Speech, XVIII (1932), pp. 446-452.

15

that some voices were improved by the technique. The
conclusion reached was that the pattern of thoracic
vibration was the result of forced vibration.l

Early research in the area of chest resonance and
the transmission of sound through the body tissue indicates
that the walls of the chest as well as other more remote
body structures vibrate during the production of vocal
sound. The research cited indicates that the thoracic
cavity vibrates as the result of forced vibration rather

than resonance.

Transmission of Sound through Body Tissue
Over a period of several decades, Georg von Bekesy

conducted many investigations into the cOnductile
efficiency of living body tissue. In 1932, Bekesy noted
that the bones of the head vibrate in three different
patterns depending upon the frequency of the tone applied
to the bone. Below 200 Hz, the skull behaves as a rigid
body and is displaced in a direction parallel to that of
the vibrating body which initiated the displacement.
Various parts of the skull vibrate in simple parallel
movements. At 800 Hz, the forehead vibrates in the

Opposite direction from the back of the skull because of

 

1C. F. Lindsley, "PsychoPhysical Determinants of
Voice Quality," Unpublished Ph.D. Dissertation, University
of Southern California, 1934.

16

the inertia of the back of the skull. The middle of the
head becomes compressed, creating a system that resonates
at a certain frequency. Thus, the forehead moves forward
as the back of the head moves backward, and vice versa.
At 1500 Hz, the skull vibrates in sections separated by
nodal lines; this vibration pattern is similar to that of
a bell.1

In 1937 Bekesy investigated the relative amplitude
of the surface of the skin during phonation. He concluded
that skin throughout the body vibrates as the result of
phonation, but amplitude of vibration of the skin decreases
as the distance fromt he laryngeal source increases. The
displacement amplitude of the skin in the area of the ear,
for example, is only one-twentieth of the amplitude
recorded near the larynx.2

Bekesy felt that it was important to know the
magnitude of the vibration of various parts of the head so
that an understanding of air-conducted and bone—conducted
sound could be applied clinically to cases of partial
deafness. He found that it was difficult to measure
vibration of bony structures because soft tissue covering

them is an inefficient transducer of vibration from the

 

lGeorg von Bekesy and Walter A. Rosenblith,
Mechanical Properties of the Bar," in Handbook of Experi-
mental Psychology, ed. by S. 8. Stevens (New York: Wiley
and Sons, 1963), pp. 1108-1109.

2

 

Ibid., p. 1111.

1?

bone to the pick-up device. Investigation revealed that
the vibrations of the back of the head were about as large
as those of the front of the head at 1800sz. It was
concluded that 1800 Hz appears to be the resonant
frequency of the skull.1
Bekesy determined that it was possible to measure

the amplitude of vibration of different parts of the body.
He employed a vibration pick-up device whose output was
prOportional to acceleration and which registered
vibration only of the surface against which it is placed.
Bekesy concluded that laryngeal sound produced vertical
vibrations and rotation of the bones of the skull.

. . . the vibration component of the skull is caused

not only by vertical vibration of the edges of the

vocal cords, but also by sound pressure in the mouth.

This sound pressure does not produce a lateral symmetry

of the head, but it produces vertical vibrations of

the lower jaw and skull.3

In 1967 Bekesy indicated that loss of vibrational

amplitude of the skin may result from the skin being
composed of several soft, fatty layers, whose structure

varies from individual to individual and which vary at

different anatomical locations. Because of tissue

 

1Georg von Bekesy, "Vibration of the Head in a
Sound Field and its Role in Bone Conduction," Journal of
'the Acoustical Society of America, XX (1948), pp. 749-760.

2Georg von Bekesy, "The Structure of the Middle
Ear and the Hearing in One's Own Voice by Bone Conduction,"
Journal of the Acoustical Society of America, XXI (1949),
pp. 217—232.

31bid., p. 221.

l8

composition, vibration of the skin occurs in complex and
irregular patterns.1

Mullendore2 reported relative amplitudes of vibration
of three vowels at ten locations of the body for five
male subjects with similar physiques.. He employed a
contact microphone which was held by hand against the
selected locations by the subjects. The signal was
picked up, converted into electrical impulses, and recorded
on magnetic tape. Tape 100ps for each vowel were
analyzed by a frequency analyzer for relative intensity
levels. The vowels employed were /i/, /a/, and /u/.

In addition to these, the six vowels /o/, /I/, / /,'/ /,
/ /, and / / were used to investigate the spectrum
characteristics of vowel sounds picked up at various
locations of the human body.

Analysis of 440 vowel samples revealed differences
of relative intensity (amplitude) of vibration among the
ten anatomical locations. Mullendore rank-ordered the
locations from greatest amplitude to least amplitude of
vibration as follows: (1) thyroid cartilage, (2) mandible,

(3) nose, (4) top of skull, (5) clavicle, (6) vertebrae,

 

lGeorg von Bekesy, Sensory Inhibition (New Jersey:
Princeton University Press, 1967), p. 119.

 

2James M. Mullendore, “An Experimental Study of the
Vibrations of the Bone of the Head and Chest During
Sustained Vowel Sounds," Speech Monographs, XVI (1949),
pp. 163-176.

 

l9

(7) sternum--superior end, (8) sternum--inferior end,
(9) mastoid, (10) fifth rib.
Mullendore summarized by stating that

. .‘. measurable Variations in intensity of the
framework vibration were present among the various
locations studied for all subjects, with the head
vibration more intense than the thoracic . . . there
was a significant variation in the absolute intensity

of the bone-conducted sound even though the air-
conducted sound intensity was held constant for all . . .

1

The following overall conclusions were reached
regarding the spectral characteristics of the eight vowels
employed when picked up at each location: (1) all points
tested vibrated during production of vowel sounds, (2)
the nature of vibration was complex enough to provide a
reasonable basis for discrimination of the various vowels,
(3) the pattern of vibration varied from location to
location, (4) the wave composition differences in bone-
conducted sound varied considerably from those of air-
conducted sound. Bone-conducted sound has less total
energy.2

Moser and Oyer investigated the relative intensities
of twelve sustained vowels at sixteen locations ranging

from the larynx to the top of the skull.3 The investi-

gators used three male Speakers selected to represent

 

lIbid., p. 174.

21bid., p. 172.

3Moser and Oyer, "Relative Intensities of Sounds
0 o '0’" pp. 275-2780

20

three distinct body types: heavy, medium, and thin build.
It was assumed that differences in body types would
influence relative intensity measurements. The subjects
hand-held a bone oscillator while producing the vowels
/i/. /I/, /e/. / /. /€/. /a/. / /. /A/, / /. /0/, / /.
and /u/. I
‘The signals were recorded on magnetic tape and
later analyzed by a power level recorder for overall

relative intensity. The investigators concluded that the

 

head and neck could be mapped into four distinct regions 3
according to relative intensity of vowel sounds. The
regions are presented in order from most intensive to
least intensive:
Region I. Neck--left, right; angle of mandible--
left, right; larynx.

Region II. On and below occipital protuberance;

 

nasalis; center of mandible.

Region III. Top center of skull.

 

Region IV. Sphenoid--left, right; temporal—-

 

left, right; frontal bone; above occipital protuberance.
The general conclusion was reached that

. . . except for the region at the tOp and center of
the head, the relative intensity of tissue-conducted
sound produced by the vocal mechanism is decreased as
distance from the larynx is increased, . . .

lIbid., p. 277.

21

Snidecor, Rehman, and Washburn1 investigated the
use of a pick-up micrOphone at various anatomical sites
to determine the relative power of the vowels /i/, /e/,

/ /, and /u/. At each location, five recordings were
made of each of the five vowels. Results indicated the
following rank ordering of locations from greatest to
least relative intensity level: (1) larynx, (2) mandible,
(3) chin, (4) nose, (5) mastoid, (6) forehead, (7) ear
canal, and (8) zygoma.

These results correlated closely with the findings
of Moser and Oyer. On the basis of power, judged
intelligibility, and listener preference, the forehead,
mastoid, and larynx appear to be the most suitable pick-up
positions for situations involving no-lip or free-field
microphone use. I

Using a sound-treated box to protect the ears from
air-conducted sound and by loading the ear drum with
metal, Kirikae, Sato, Oshima, and Hirose2 found that bone-
conducted sound contributes to a subject's hearing his
own voice with different levels of loudness for different
vowels. Air-conduction feedback of one's own voice

appears to play an important part in hearing the vowels

 

lSnidecor, Rehman, and Washburn, "Speech Pickup
0 o 0," pp. 277-281. '

21. Kirikae, T. Sato, H. Oshima, and H. Hirose,
"An Experimental Study of Hearing One's Own Voice,”
Practica Oto-Rhino-Laryngologica, XXIII (1961), pp. 56—71.

 

22

/a/ and /e/, whereas the bone—conduction element is most
important in hearing the vowels /i/, /u/, and /o/. This
would indicate that the bones and tissue of the head vary
in conductile efficiency of vowel sounds according to
variation in frequency distribution or formant components
of each of the vowels. '

Neely and Forshaw investigated the intelligibility
of bone-conducted speech picked up in quiet by transducers
at six locations of the head and neck, including the
frontal bone, the tOp and center of the skull, above and
below the external occipital protuberance, the temporal
bone, and the side of the neck. Four trained male
Speakers hand-held the pick-up transducer against the
selected locations while reading twenty-four multiple-
choice intelligibility lists. Four different types of
transducers were used for each Speaker at each location
for all lists. The lists were tape-recorded and judged
by sixteen listeners for speech intelligibility. Speech
signals were judged to be most intelligible from on the
forehead; speech recorded at the back of the head ranked
second in intelligibility.l

A similar study investigated the intelligibility

of bone-conducted speech picked up in white and pink noise

 

lNeely and Forshaw, "Speaking and Listening
through the Head: I . . .," pp. 151-157.

23

of 120 dB SPL by transducers placed on the upper lip and

on the throat near the larynx. Strong, Chant, Forshaw, and
Neely1 found that listeners rated speech recorded on the
lip to be significantly more stable and intelligible than
that recorded on the throat. Different pressures of
application of the transducer was found to have no

influence on judged speech intelligibility.

DevelOpment of Instrumentation

 

Research between 1900 and 1925 into the trans-
mission of vibrations through the human body tissue relied
heavily upon the tactile sense and subjective judgment
of the investigator to determine the nature and location
of tissue vibration.

Simon and Keller2 attempted the first instrumental
experiment in this area. They employed a carbon-button
stethosc0pe which was sensitive to tissue-transmitted
vibration but not to air-transmitted vibration. The
investigators were able to convert the original sound waves
into light waves. These in turn were recorded on
photographic film and measurements of the vibrations were

made from this source.

 

1R. A. Stong, V. G. Chant, S. E. Forshaw, and Keith
Neely, "Speaking and Listening through the Head: III. The
Intelligibility of Speech Recorded in Noise," Journal
of Auditory Research, VI (1966), pp. 385-391.

 

2Simon and Keller, ". . . the Problem of Chest
Resonance," pp. 432-439.

24

The investigation of Wise1 marked an advance in
instrumentation design. A stethosc0pe which permitted
vibration to be picked up and recorded electrically was
employed. This permitted more accurate and more sensitive
collection of data.

Research in this area between 1934 and 1966
employed a variety of contact crystal micrOphones or
refinements of bone conductor oscillators driven in
reverse to obtain the tissue-transmitted vibration signal.
As refinements in these varieties of instruments were made,
data which were more accurate were reported. All of these
instruments were incapable of transmitting all of the
complex components of the patterns of vibration picked up
from the skin. A certain amount of information was lost
because of limitations in the sensitivity of the pick-up
device.

In 1969, the Piezotronics Company developed a
quartz transducer (accelerometer) with an integrated
circuit designed for dynamic pressure measurements of
pulsations and vibration environments. The accelerometer
is a high frequency response, dynamic pressure, measuring
instrument. An amplifier mounted inside the transducer

converts the high impedance output from the crystals into

 

1Wise, "Chest Resonance," pp. 446-452.

25

a low impedance system of less than one-hundred ohms. The
instrument is sensitive to extremely small pressure changes.
The accelerometer is powered by a compact eighteen volt,
battery pack which provides a constant two milliampere
current with low system noise.1
The accelerometer and battery power unit provide
for extremely sensitive pick-up and transmission of
pressure changes due to vibration. The unit can be
connected directly to an oscillosc0pe, sound level meter,
or recorder. The advantages of compact size, sensitivity,

and accuracy resulted in the selection of this system

for use in the present investigation.

Summary

This review has indicated that in the last
one-hundred years a considerable amount of investigation
has been conducted into the transmission of sound through
the tissues of the human body. Early researchers in voice
science and music believed that the chest acted as a
resonating chamber and contributed to the overall quality
and tone characteristics of the human voice during speaking
and singing. The concept of chest resonance was derived from
empirical observation of the vibration of the thorax

during phonation.

 

l"ICP Transducer Data and Information Manual,"
(New York: PCB Piezotronics Incorporated, 1969).

26

During the late nineteenth century at about the
time of the work of Helmholtz into resonation, Opinion
began to differ concerning the importance of, and even the
existence of, chest resonance. Some investigators
believed that chest resonance was a real and measurable
phenomenon which contributed noticeably to voice “r
characteristics. Others believed that the chest was
incapable of acting as a resonating chamber and that

vibration felt during phonation was the result of forced

 

vibration and not resonation.

‘E'

Early eXperimental research relied on one or
another form of the stethosc0pe which was held against
the skeletal framework and transmitted sound to a listener's
ear or to a photographic or sound recording device.
Measurements were roughly quantitative but based mostly
on listener judgments.

Refinement in instrumentation resulted in the
development of bone conduction transducers, magnetic tape
recorders, and equipment for spectral and power analysis
of sound waves. Measurements of sound transmitted through
tissue became less subjective.

The research in this area over several decades
has consistently indicated that vibration of the thorax
during phonation is primarily the result of forced
vibration. In general, it has been concluded that

vibration of the bones and tissue of the head and neck

27

during phonation is primarily the result of forced
vibration, with some resonation from the mouth and nasal}
cavities contributing to the overall pattern of vibration
to a lesser extent. Table 1 presents an outline and
summary of early experimental investigations into tissue-
transmitted sound.

Researchers since the early days of Bekesy's
research have, in general, found that amplitude or
intensity of the vibration of tissues of the head and
neck during phonation decreases in proportion to the
remoteness of the structure from the sound source. Some
research, however, has indicated that the tOp and back
of the skull vibrates with greater intensity than some
other structures closer to the larynx. Table 2 presents
an outline and summary of more recent research into
relative intensity of vibration of the skin during speech
production.

Differences in recent research findings may result
from the following:

1. Lack of control for differences in body
physique, i;g., height, weight, fatty tissue
composition.

2. Differences in stimulus materials employed,
i;g., singing versus Speaking; vowels versus

connected discourse.

28

 

.caxm mo codpmunfl>

mo “magma mmsHEHOump
mammwu maps“ msflma
lumps: “mumuou Omam
cam MAHOOHuHm> mumupfl>
mmcon “um coma two:

«0 mocmsvmum unmc0mmn
“mmmmwuosfl xsmnma Eouw
possumflp mm sflxm mo
soaumunfl> mo mpsuHHmEm
HOHHmEm «mcumuumm

m CH mmumunfl> omms

memwmm mama pmcflmuu H

mmusoomflp
pmuomcsoo tam mam3o>

mnoumaaflomo
soapOSGsoo Onon

cowumunfl>

pmouom mo uHSmmH
mm mmumunfl>

umwno «hocmHOHmmm
mcamum>

sufl3 pan

pssom pospcoo
mmsmmflu Ham

Hmmsflm
mHmE Umsflmuu H

xuom madcap
mam30> msflmsflm

mmoomOOsmconm

coaumunfl> onoo
Hmoo> mo hosmsqmum
on paommmunoo
soanz coaumnnfl>
pmouow mo pasmmu
mm mmumupfl> ummno

memwmm
mama pmswmnp H
\O\

wmoommnumum
souusnlsonumo

mSOHmSHOGOU
mpomnnsm
Hmsmflm sowmmm

HOOSOmcmua
as xoam

 

Asmmaummmflv Smmxmm

“HmmHC,mmuz

Akwmac umaamsuqosflm

 

.USSOm pmuuflfimcmuulwsmmwu

a“ noummmmu maumm mo mumEESm was msHHuSOII.H mamma

29

Failure to secure the placement of pick-up
transducer against skin to assure that data
collected fo not contain noise and distortion
created by movement of the transducer.

Use of instrumentation that lacked sensitivity
to complex and subtle patterns of tissue

vibration.

 

30

 

mwuHm
ms xOHm ummn xGMHMH

monsom Bonn mass
Hosanna GOHumunH>

GOHu
IHmomEoo m>m3 UQSOm
pwuospcoo mammHu

 

pHoummE .pmmnmuow mo mpspHHmEm mmmH cH GOHuMHum> sons mconSHosoo
QHH sum
pHoummE
msomwn .H.H .Hmuomfimu HOHHmmcHIESGHmum
Hmsmo paw. .H.H .pHosmnmm HoHnOQSmIEDGHmum
pmmanOM Hstm mo» mmnnmpum>,
pHoummE mocmnmnsuoum OHUH>MHO GOHumunH>
mmos HmuHmHooo Hmmc Hstm mo mou mo mpsuHHmsd
cHno xSMHMH mmos ummmq on
mHancms .H.H .mHancms mHanams umoz pmxcmm
xshHMH panHIummH .xoms mmMHHuno pHoumsu ImsOHumooq
\ \ \.\ \ \
\ \ \.\
\0\ \ \ \m\
\ \ \<\ \H\ Hmamflm
\s\ \H\ \s\ \M\ \S\ \9\ \.\ \H\ nommam
pHmn cams png was:
“mnoumHHHomo UHOS was: “ocOQQOHOHE
mson uSOHOMMHp m “HoumHHHomo anon Hmummuo pomusoo Hmospmcmua
mmstmmsm mmsvamnm
mHmE H ucmHOMMHU m I mmHmE m HMHHEHm I mmHmE m muomflQSm
AmmmHv “oompHcm AmmmHv “pseuummoz Amemav muopamaazz mumsssm

 

.GOHuospoum.£ommmm.msHH5@ Sme_mo sOHumunH> mo mpsuHHmEm

m>HUMHmH pmummHumw>QH AOH£3 mmesum ucmsHuumm mo muHSmmH mo SOmHummEOUII.N mqmda

 

31

 

 

.uommmm

on pm: QOHHMOHHmmm

mo mwu5mmmum uGOHOMMHp
.mHH Momma um mHnH

UmmanOM
pm manflmflaamucw

 

ImHHHmpsH umos sommmw umoE Summon mSOHmsHocou
MOGOHOHMMO
Hmsmo HMO mHHuODUcoo sH mnm>
umounp pawn mo mOSmmHu
mocmumnduoum uHEmsmuu Mona SOHumunH>
AGOHumoHHmmm mo mmusm HmuHmHooo mmOGOSOH mo mHm>mH mo mpsuHHm84
Immnm ucmummep my HHDxm Housmo CH Hmmep mHm3o> ummmq op
xcmuMH ppm: NO xomn I I I I “no: pmxamm
mHH momma pmmzmuom cm>Hm nos ImGOHuBOOH
umHsv
mmHo: mans gH mumHH \O\ \m\
cH mumna .HHmuqfl .Hamuqfl moflogo Hmcmflm
OOHOBU mHmHuHsE OH mHmHuHsE «N \5\ \M\ \H\ nommmm
.OHE mcH
IHHmosmo mmHos H
.OHE .OHE uomucoo H
uomugoo usmHOMMHp v HouMHHHomo mcon N gm>Hm no: “monomcmua
mmHmS N mmHmE w mHmE H muommnsm
.mmmav mnoum AmmmHC mammz AHGmHC mmxflpflm Smuseum

 

.UmchuCOUII.N mnmde

CHAPTER III

EXPERIMENTAL PROCEDURES

This chapter includes a description of the subjects
participating in this study and the recorded samples of
the twelve vowel sounds at the sixteen anatomical
locations of the head and neck. The chapter also
describes the design and construction of the experimental
helmet and shield as well as a discussion of the equip-
ment employed to record and analyze the data. Pro-
cedures employed to obtain the data and instructions
given to the subjects are described. The method of

analysis of the recorded samples is also presented.

Subjects

 

The subjects were six male graduate students
ranging in age from 23 years 0 months to 35 years 0 months.
Subjects were selected from the Department of Audiology
and Speech Sciences at Michigan State University and had
a mean age of 27 years 8 months. During their under-
graduate careers, all subjects had taken a course in the
understanding and use of the International Phonetic

Alphabet. They were able to transcribe spoken language

32

33

into written phonetic symbols as well as being able to
produce Speech sounds when presented with written phonetic
cues. All subjects used the General English American
Dialect of speech.

'To account for differences in the conductibility
of sound through the bones and skin of individuals with
different body characteristics, the general body
morphology classification system devised by W. H. Sheldon
was employed in the selection of subjects.1

For the purposes of the present study, the broad
categories of endomorph, mesomorph, and ectomorph were
used for the selection of subjects. The specific
techniques and measurements devised by Sheldon to
determine body classification were not used in this study;
however they are discussed in Appendix A. Two subjects
were selected to represent each of the three body
classifications. Body classification, for this study,
was based upon the dimensions of height in inches and
weight in pounds, as well as on over-all suitablity
to one particular category according to the descriptive
terms (solid, fragile, square) provided by Sheldon and

discussed previously in this section.

 

1

 

34

Appendix B presents the pertinent descriptive
data relative to each of the subjects employed in this
study as well as the body classification for each.
(Appendix C presents the mean age, height, and weight for

each body classification.

Equipment and Materials

 

The equipment and materials reported below were
used in the various phases of this experiment to be
described in the following sections of this chapter.

Equipment

 

Tape Recorders: Speed 7% inches/second
(Ampex, Model AG 500-2)
(Ampex, Model AG 600-2)
(Scotch Brand Magnetic Tape 201)
Microphone:
(Bruel-Kjaer, Model 4145)
Sound Level Meter:
(Bruel—Kjaer, Model 2203)

Power Source:

(ICP Piezotronics Battery Power Unit 480A)
(Bruel-Kjaer, Model 2805)

Accelerometers (four):

(ICP Piezotronics Quartz Transducers,
Model 111A21)

Filters:

(ICP Piezotronics special 6000 Hz Low
Pass Filter)

 

35

Oscillator:
(Hewlett—Packard, Model 4204 A)
Power Level Recorder:

(Bruel-Kjaer, Model 2305)
(Bruel-Kjaer, Paper QP 1102)

Impedance Head:

(Bruel-Kjaer, Model 8801)
Helmet:

(Astro-Cell Helmet, Model RAC—HZ)
Shaker:

(Bruel-Kjaer, Model 4801)
Preamplifier:

(Bruel-Kjaer, Model 2625)
Amplifier:

(Bruel-Kjaer, Model 2107)
(Bruel—Kjaer, Model 2603)

Generator:
(Bruel-Kjaer, Model 1024)

Phonetic Material

 

Twelve American English Vowels

Frequencngesponse Curves of the
Transmission Instrumentation

 

The frequency response curves of the Ampex Tape
recorders were found to be reasonably linear and flat for
both recording and reproduction. These curves are pre-

sented in Appendix D.

36

The relationship between pressure inputs of
different intensities and the output measured in milli-
meters was shown to be linear by the manufacturer. Some
response curves of the accelerometers to variable frequencies
of constant r.m.s. accelerations was determined experimentally.
Results indicated nonlinear responses throughout the Speech
range. Four of the obtained curves are presented in

Appendix E.

Construction of Helmet and Shield

A standard football helmet was modified to provide
a stable means of attaching the four accelerometers to the
sixteen sites on the head and neck for each subject. The
inner, foam rubber padding of the helmet was removed leaving
the fiberglass shell. Four solid brass rods which were
one-eighth inch in diameter and six inches long were
threaded on the outside and inserted through threaded
holes in the helmet shell. This provided four legs which
rested on the Skull. The legs provided a stable means
of resting the helmet on the head with minimum contact
to the head by anything other than the accelerometers.
Thus the helmet itself did nto touch the head at any
point but was suspended above the head. Foam rubber
tips were placed on each leg for comfort. The legs could
loe adjusted to raise or lower the helmet to account for

individual differences in head size.

 

37

Twelve of the sixteen anatomical Sites to be
investigated were on the head and could be reached through
the helmet. Provision had to be made so that four
accelerometers at one time could be attached through
the helmet to the surface of the skull. Eight holes,
each one inch in diameter, were drilled through the
helmet at various locations which corresponded to the
cranial anatomical sites to be studied. By turning the
front of the helmet to the back of the head, thus
repositioning the relationship of the holes to the
anatomical sites, several holes could be used to pinpoint
more than one anatomical site.

Aluminum tubing with a three-quarter inch overall
diameter and a one-quarter inch core was threaded on the
outside and on the inside and cut into four, eight inch
lengths. Each of these tubes could be attached through
a hole in the helmet by means of nuts and washers. The
accelerometers were threaded into the inner core of each
tube and the tubes adjusted so that the accelerometer
touched the surface of the skull at the precise point
required. In this manner, four accelerometers could be
attached to the helmet at one time.

Because of the method of fixing the accelerometers
In) the helmet, it was found impossible to measure the

pressure of application of the transducers against theskin.

 

38

For this reason, it was decided that the accelerometers
would be positioned so that solid contact against the
anatomical location was made. No mechanical measurement
of the pressure of contact by the transducers to the skin
was attempted. The original holes in the helmet were

one inch in diameter and the aluminum tubes were three-
quarters of an inch in diameter. This provided additional
room to move the tubes containing the accelerometers.

This margin for adjustment plus the ability to adjust

the position of the helmet on the head by means of four
legs proved to be sufficient to allow the use of one helmet
to collect the data for all Six subjects.

The four remaining anatomical sites were located
on the mandible and neck. These regions could not be
reached easily with the use of the helmet. Therefore a
plexiglass shield molded to fit around the neck in a
fashion Similar to a neck brace was constructed and
fitted to the helmet. Four holes, each one inch in
diameter, were drilled in the plexiglass to provide
attachments for the remaining anatomical sites. The
accelerometers were attached to the Shield in the same
manner as they were attached to the helmet. Figure 1
presents a photograph of the experimental helmet and

shield and together comprised the supporting system.

39

 

FIGURE l.--Photograph of helmet and shield with accelero—
meters in position.

Frequency Responses of Accelerometers
Attached to Helmet

The transmitting instrumentation (tape recorders
and accelerometers) were found to have reasonably flat
frequency response curves; however it was decided to
investigate experimentally the response curve of the
accelerometers when attached to the helmet. It was found
that the helmet itself present relative peak resonances
up to 20 dB at 80 Hz and at 120 Hz following the transverser

and posterior-anterior axis. These resonances were found

40

by shaking the helmet with input of constant r.m.s.
acceleration and different frequencies in several points
of the helmet and recording the output from opposite
directions through a linear read—out system.

The responses of the accelerometer to inputs of
different frequencies and constant r.m.s. acceleration
were also obtained. These measurements were performed
by shaking the base of the accelerometers, which were

mounted in the helmet, with a shaker yielding a constant

 

r.m.s. acceleration of a variable frequency, and recording
the electrical output of the accelerometers. A block
diagram of the system used for the measurement is preSented
in Figure 2.

The results of these measurements are presented
in Table 3. It will be seen that Significant resonances
occurred in the speech frequency range. Thus, the response
of the accelerometers was no longer linear when they were

attached to the helmet.

TABLE 3.--Resonance responses of accelerometers attached

 

 

to helmet.
, Input in Tenths Peak Output Response Frequency of
of r.m.s. in Relative Decibels . Peak Output
acceleration (g) of Sound Pressure Level Response

0.1 g 9 dB 850 Hz

0.3 g 13 dB 380 Hz

15 dB 900 Hz

0.5 g 13 dB 380 Hz

16 dB ' 850 Hz

 

41

.mumumsoumeoom mo omsommmu umwu on paws Emummm mo SOHOMHU xUOHmII.m musmHm

SEE

 

 

 

 

 

 

 

     

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Home Mam
Ill], mewnm
Hoom .
NOHN meN Mam Ummm
.lMfilll:Mwm Mam mosmpmmEH
. sou
VNOH Mam Q5 HNANHHH
Li
M H.Imuuqmmwz
C _
EMF; momm
momN Mam , mewwm
Hmpuoomm Hm3om
Hw>mq Hm3om i [
._s v mo
Hamsm .
#Hosom
IIIIII momN Mam I ,

“OHMHHQS<

 

 

 

42

Vowel Sounds and Anatomical Locations
The twelve vowel sounds selected to be recorded
and later analyzed for the investigation were: /i/, /I/,

/e/I /€/I / /l /a/I / /I /A/I / /I /O/l /u/l / /- These

sounds are the same as those used in the Moser and Oyer
study.1

The sixteen anatomical locations (also duplicates
of those used by Moser and Oyer) examined in this study
are listed in Table 4 and are coded according to the
order in which each of the locations was recorded for the
six subjects. This order was determined primarily by I
the limitations imposed by the predrilled holes in the

fiberglass helmet and shield. Figure 3 presents an

illustration of the sixteen anatomical locations.

Recording of Experimental Data

The data were collected in a sound treated room
(IAC 1400 Series). Six adult male speakers acted as
subjects. At a later time, all subjects were re-recorded
so that measurements of reliability could be made. The
sound pressure level and the fundamental frequency of all
vocalized utterances were kept constant by all speakers.
At a later time, all subjects were re—recorded so that

measurements of reliability could be made.

 

lMoser and Oyer, "Relative Intensities of
Sounds . . . ," pp. 275-278.

‘13-».

43

TABLE 4.--Anatomical locations examined.

 

 

 

Code Location

A 1 One inch below occipital protuberance

A 2 Two inches above occipital protuberance

A 3 Frontal bone, one inch above nasalis

A 4 Coronal suture, on tOp and center of Skull

B 1 Nasalis bone

B 2 External occipital protuberance

B 3 Left wing sphenoid

B 4 Right wing sphenoid

C 1 Lower border, left ramus of mandible

C 2 Lower border, right ramus of mandible

C 3 Right temporal, one inch above auditory meatus
C 4 Left temporal, one inch above auditory meatus
D 1 Mental protuberance of mandible

D 2 Below notch of thyroid cartilage

D 3 Left side neck, on plane with larynx

D 4 Right side neck, on plane with larynx
Practice

Each subject practiced sustaining each vowel for a
minimum of five seconds at a constant level of 76 dB
which he monitored visually on scale C of the sound
level meter. Subjects sat with their lips approximately
twelve inches from a table-mounted micrOphone connected
to the sound level meter. Each subject practiced maintaining
a constant fundamental pitch for all utterances of all
vowels investigated. Only after each subject was able
to sustain the same fundamental pitch at a constant sound

pressure level of 76 dB, were the data collected.

.. .L”‘I fl‘A-UJ

 

44

A4
W \\
B3 A2
2“; A‘"% (.B4
4 c3 . B2
Bl {E c4. 3

\ A1
”3 c1 ....
\ 02.
D3
OD

D2 D

\
l

\\

//\

Figure 3.--Schematic representation of anatomical locations.

 

45

Equipment

 

The modified helmet was placed on the head of each
subject so that the four adjustable legs touched the
skull but the surface of the helmet did not. The legs
were adjusted so that the helmet rested confortably and
securely on the head and so that the first four anatomical
sites were located under the appropriate holes. The
helmet was repositioned so that the holes necessary for
the second and third sets of cranial recordings could be
made. For the fourth set of recordings, the plexiglass
shield was attached to the helmet and the four holes it
contained were aligned over the required anatomical
locations.

Four integrated-circuit-piezoelectric quartz
transducers were used to pick up tissue vibration during
voice production. These transducers were designed to
obtain dynamic measurements of vibration environments,
and they contained a tiny integrated circuit amplifier
which converts a high impedance from the quartz crystals
into a low impedance signal. Each accelerometer was
attached to its own battery-supplied electrical source
and to one of four available tape recorder channels.
Attached to the output of each accelerometer system prior‘
to the tape recorder input was a specially designed 6000 Hz
low pass filter that prevented passage of high frequency .
noise in the system. A complete description of the

transducers was presented earlier in this report.

46

Procedures

 

The four accelerometers were affixed to the
helmet (or Shield) until recordings were obtained from
each of the Sixteen locations. The accelerometers were
placed with sufficient pressure to depress the surface
of the skin slightly at each location. Each time the
subject uttered the vowel sounds, the signals received
by each of the four accelerometers were picked up and fed
to the four channels of two tape recorders. Recordings of
vibrations at four locations were made simultaneously.

The positions of the four accelerometers were changed
four times until sixteen anatomical locations had been
recorded. Figure 4 presents a block diagram of the
equipment employed in recording the data.

Prior to recording, a 1000 Hz pure tone was fed
into each channel at a level of 100 milivolts. Input gain
was set to read —10 on each VU meter. Thus a constant gain
setting was established. This same tone acted as the
reference tone for the determination of relative intensity
for each recorded Signal.

Each subject was read the following instructions
prior to recording the data:

There will be four recording periods, each

separated by a rest period. These procedures will

be followed during each recording period. When you
are signalled to do so, intone the vowel sound which
you will see printed on a card placed in front of you.

You are to sustain the vowel for five seconds and you
will be signalled when five seconds have elapsed.

47

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

{Microphone
. _ — 12 inches— _ E {2 Sound Level
ubject
B&K Meter
4107 B&K
2203
Accelerometer
Power Level
Recorder
B&K 2305
Tape Efle
Recorder
gower Ampex
ource 600-2
PCB 4804 500_2
0 O——@
PCB Low Pass
Figure 4.--Block diagram of accelerometer and recorder

equipment.

the

48

Please attempt to keep the level of your voice at

the point indicated by the sound level meter in

front of you (76 dB). Also try to maintain the

same fundamental pitch for each utterance. There

will be a five second pause after which you will be
presented with a new vowel sound to be produced in the
same manner. This procedure will be continued until
twelve vowels have been recorded from Sixteen different
points on the head and neck. Are there any questions?

Table 5 presents the fundamental frequencies of

six subjects. It can be seen that none of the

fundamental frequencies differs significantly from the

mean of 130.8 Hz.

TABLE 5.--Fundamental frequency for each subject.

 

 

Subject Fundamental Frequency
S l 130 Hz
S 2 135 Hz
S 3 130 Hz
S 4 125 Hz
S 5 135 Hz
S 6 130 Hz
Mean Fundamental Frequency 130.8 Hz

 

Analysis of Experimental Data

 

Magnetic tape recordings were made of each of the

vowel samples obtained in the manner described in the

previous section. Twelve vowels for six subjects at

sixteen anatomical locations were recorded for a total of

 

49

1152 vowel samples. Each vowel Signal was fed into the

power level recorder directly from the magnetic tape.

The power level recorder plotted a graphic display of
relative sound pressure level for each vowel on the graph
paper type OF 1102. The grid on this paper consists of

fifty parallel lines. The space between each line represents
one decibel of pressure so that a range of fifty decibels

can be plotted.

I ‘ -

Prior to recording the signals on magnetic tape, a
1000 Hz pure tone was recorded on each channel at a level
of 100 millivolts at a constant gain setting. This tone
was fed into the power level recorder directly from the
magnetic tape. The tracing pen of the power level recorder
was set to rest on the thirtieth line of the graph paper
when the reference tone was fed into it. The thirtieth
line of the graph paper was then considered to be zero
decibels and became the reference point for future graphic
tracings.

All the 1152 vowel signals were plotted on the
graph paper. Each sound pressure graph was identified
according to subject, vowel, and recording site. Each
vowel was represented graphically as a series of peaks
and valleys which represented changes in relative pressure

of the Signal transmitted.

50

An average relative sound pressure score in decibels
was obtained for each of the 1152 vowel graphs. For
each sample, the relative sound pressure level of each peak
was determined, and an arithmetic mean of the relative sound
pressures of all peaks for the sample was taken. These.
mean relative sound pressures became the data to be

' analyzed statistically for the investigation.

CHAPTER IV

RESULTS AND DISCUSSION

There were three independent variables of interest
in the present study. These included body type, vowel,
and anatomical recording site. The dependent variable
under consideration in the experiment was relative sound
pressure transmitted through tissue.

The raw original and retest scores for all subjects
and vowels are found in Appendix G. The original scores for
all subjects and vowels were averaged according to body
type; these data can be found in Appendix H. Appendix I
contains the mean scores for all subjects and each vowel,
regardless of body type. Examination of these data provided
descriptive information relevant to the goals of the study.

In addition, statistical tests were performed to
examine the Significance of these data. The results of
these tests are included in Tables 6 through 11. A discussion
of the statistical tests employed in this study is included

in Appendix F.

Descriptive Analysis of Results

 

Observation of the data yielded by this study
suggested that the vowel sounds /i/ and /u/ are transmitted

51

U.‘

 

52

through tissue with the greatest relative sound pressure
through all locations in all body type groups of subjects.
The vowels /aw/and /a/ are transmitted with the least
relative sound pressure. The different Spectral character-
istics of each vowel as well as the resonance character-
istics of the body structures the helmet and accelerometers
may have been responsible for these differences encountered
in the overall sound pressure transmitted through tissue,~
by each vowel.

Of major interest was the interaction of body type
and vowel, body type and location, vowel and location,
and body type-vowel-location in the relative sound pressure
transmitted through tissue.

The mean relative sound pressures for each vowel
used in this study, regardless of position, were plotted
for the three subject groups of ectomorphs, mesomorphs,
and endomorphs as well as for all subjects combined.

These plots are presented in Figure 5. Vowels have been
ranked in descending order according to mean relative
transmitted sound pressure.

Examination of Figure 5 reveals that the mean
scores from each vowel for ectomorphs and endomorphs
were very close. The mesomorph subjects, Show mean relative
sound pressure scores which are consistently below the
score for either of the other groups, although the analysis

of variance test indicated that there was no Significant

Relative Intensity in Decibles

10
12
14

16

18

53

 

 

 

i u i a: e o If e 3 A CL
Vowels
Ectomorphs — - - -
Mesomorphs

Endomorphs ........
All Subjects

 

FIGURE 5.--Mean relative sound pressures transmitted for
each vowel for the groups of ectomorphs, endomorphs, meso-
morphs and all subjects, regardless of Position.

L—fld

54

difference among the three groups. However, a trend can
be seen in which mesomorphs appear to transmitt lower
relative sound pressure scores than other subjects. Mean
relative sound pressure scores at the sixteen locations
failed to Show any trends among the other two body types.
Figure 6 presents the mean relative sound pressures
transmitted for each Of the groups of ectomorphs, mesomorphs,
and endomorphs as well as for all subjects regardless of
vowel. The positions are ranked in descending order accord-
ing to relative sound pressure. Examination of Figure 6
reveals that there is considerable intra-body-type-group of
subjects variability and inter-body-type-group of subjects
variability. None of the three body type groups approached
the smooth descending slope of the means of all subjects.
Each body type group is represented by a markedly jagged
curve. There does not appear to be any trend for any
particular body type with respect to mean relative sound
pressures transmitted at the various anatomical locations.
It would appear that when vowel sounds are uttered
by male subjects and vibrations picked up by transducers
placed at a variety of anatomical locations of the head
and neck, the sound pressure is transmitted with the great-
est intensity through the top-center of the skull (A4), the
frontal bone (A3), the neck near the larynx (D4), and the

mandible (D1).

[—4
H
J

[—1
U

I"
U1

1

Relative Intensity in Decibles

H
\l

[.4
\O

21

J

L

1

L

 

55

 

 

 

A4 A3 '34 0'1 153 ci B1 c2 D2 c4 C3 A1 B4 B3 132 AZ
Location (Locations are identified in Table 4)
Ectomorphs — — - —
Mesomorphs

Endomorphs .......
All Subjects

 

FIGURE 6.--Mean relative sound pressures transmitted for each
position for the groups of ectomorphs, endomorphs, mesomorphs
and all subjects, regardless of vowel.

56

Results of Statistical Tests

 

All of the statistical computations for this study
were conducted on the Control Data Corporation 3600 Digital
Computer at the Computer Center of Michigan State University.
An analysis of variance of the three main levels used in
this experiment and four of their interactions was con-
ducted according to the STAT Series Program provided by
the center.1

Table 6 provides a summary of this analysis of
variance and shows that three of the seven F-statistics
are significant, at the level of confidence of 0.05%,
namely:' vowels, positions, and interactions of vowel and
positions. All other levels tested (body type and inter-
actions of body type with vowels and positions) showed no
significant differences.

These outcomes suggested that the questions asked
in Chapter I should be answered as follows:

1. The relative overall sound pressure transmitted
through tissue for each vowel is independent of body type
and anatomical location.

2. The relative overall sound pressure transmitted
through tissue for each body type is dependent upon each

vowel and anatomical location.

 

1Agricultural Experiment Station, "STAT Series
Description No. 14: Analysis of Variance with Equal
Frequency in each Cell," (East Lansing: Michigan State
University, 1968).

57

Hm>mH mooo.o paosmn unmoAMHamAm I *

 

 

HmHH Hs~.~memoa Hmuos
mom.» mas vmm.masm nmm+nom<uu
Hmo.H mmm.s omm meo.omm~ muse spomucoHpHmomuamzo> noxmxmv
*HHo.~ moH.mH mes va.mmv~ mcofluflmomuamso> xmxev
smm.moma m me.mva ao+nuu
who.o mHm.-~H m mmm.mqe~ mass spam loo
. NmH.mme ms svm.mmsom om+aomum
mam.a emH.mom om smm.ms~ma mass mpomumcoHpHmom ioxmc
swam.m mo¢.mms~ ma mmo.aommm . mcoHpHmom Ame
mom.mH mm mo~.mmm o¢+ao<us
mom.o smo.ea mm omm.~om mass spomuamzo> loamy
smmm.mm oom.s~aa HH so~.eemma Hmso> Ame
OHumHumum mumsvm mmnmsvm .
m sum: mo mo Esm mosmHHm> mo mousom

 

.mm»u mpOQ paw .mcoHuHmom .Hm30> msHHmmEoo OOSMHHM> mo mHmmHmnm mo.MHQEESmII.o mamma

58

3. The relative overall sound pressure transmitted
through tissue for each anatomical location is independent
of vowel and body type. In any anatomical position, the
vowels can be similarly ranked.

4. There is not any significant interaction among
vowel, anatOmical location, and body type with respect to
overall sound pressure transmitted through tissue.

The result of this analysis suggested that vowels
could be ranked from highest relative transmitted sound
pressure to lowest in the following manner: /i/, /u/, /I/,
/37. /e/. /0/: /V/. /€/, / /. /A/. /a/. /aB/- See Table 7).
This was valid for all subjects regardless of anatomical
position. Table 8 presents the array of comparisons of
mean sound pressure transmitted of each vowel to all the
others. This was determined by the use of Duncan's

Multiple Range Test.1

The (x) symbol represents entries
in the rows which differ' significantly from the entries
listed in the columns. It Should be noted that the
relative sound pressures of the two most intensly trans-
mitted vowels, /i/ and /u/, are Significantly different
from all the remaing vowels, with /i/ having the highest
relative transmitted sound pressure.

Further, Duncan's test suggested that positions
could be ranked from highest relative transmitted sound

pressure to lowest. (See Table 9). The position at the

topecenter of the skull had the highest relative sound

 

1Agricultural Experiment Station, "AFS MISC
Series . . ."

fitq

IT?
I. .

59

TABLE 7.--Means of twelve vowels ranked in descending order
of relative transmitted sound pressure level.

 

Mean Relative
Intensity of dB

 

Vowel re: 0 Reference
i -1.833
u -4.385
I -7.149
C7 -7.959
e -8.038
0 -9.165
v -9.725
e -ll.354
3 -ll.567
A -12.262
a -12.601

ae -12.773

 

60

TABLE 8.—-Display of transmitted sound pressure of each vowel
as compared to every other vowel.

 

 

Comparison

Vowel as a A e v o e I u i
E3 X X X X X X X X
a X X X X X X X
A X X X X X X X
3 x x x x x x X
e X X X X X X X X
V X X X X X X X X X
0 X X X X X X X X
e X X X X X X X X
3' x x x x x x x x
I X X X X X X X X X
u X X X X X X X X X X X
i X X X X X X X X X X X

 

X = difference between mean transmitted sound pressure was
Significant at 0.05 level.

.— — —_._.
. u-

61

 

«Hm.HH-
mmm.HH-
mmm.vau
mH~.eH-
44H.4H-
oos.~a-
HSH.HHI
mmm.oau
mmm.mu
msm.sn
Hmm.s-
omo.eu
mas.m-
omm.m-
mmm.~-
msm.o

mmcosH 03» m>OQMImoamumn5poum HmuHmHuoo
wocmnmnsuoum HmuHmHooo

mcHz pumHupHoamnmm

mGH3 uanHIpHocmnmm

sonn SUCH mcoImocmumnsuoum HmuHmHooo
uanHIHmuomEma

pmmHIHmuomsmB

Spawnanousnu mo nouoz
uanHImHQHUGmS mo mSEmm

econ mHHMmmz

pmmHImHansmE mo mafimm

ummHIxomz

mHancmE mo mocmnmnsuonm Hmucmz
panHIxomz

Odom Hmucoum

Hstm mo wouSOOImoa

Nd
mm
mm
vm
H4
m0
v0
NO
NO
Hm
HU
mo
HQ
v0
mm
v4

 

mocmummmn 0 "0H
mp aH suHmcmch
0>HuMHmm dam:

m>HMMHmH mo Hmpuo

SOHuHmom

.OHSmmmum canon

mchcoommp SH Umxcmu mSOHuHmom Smmume mo mamszI.m mqmde

62

pressures transmitted, while the position below the
occipital protuberance in the back of the skull had the
lowest relative sound pressure transmitted. Table 10
presents the array of comparisons of each position to

all other positions. The (x) symbol represents entries

in the rows which differ Significantly at the 0.05%

level of confidence from the entries listed in the columns.
The mean relative sound pressures transmitted at the
top—center of the Skull is significantly different from

nine of the remaining fifteen positions.

Retest Results

 

Analysis of variance of data yielded by retest of
every subject was conducted to determine the reliability
of the original measurements. An analysis of variance
test of the retest data helped to determine the significance
of the main levels tested and their interactions. Table 11
presents the summary of this analysis of variance. It
confirmed the significance of the same two factors, vowel
and position, and the interaction between these two factors
as Encountered in the original test. The probability of
the F-statistic was, once again, less than 0.05% for all
three significant statistics.

A Pearson product—moment correlation procedure was
conducted to test the correlation between the original scores

and the retest scores forthe six subjects. A positive

1:17. “_'

I. e vnlfiifi

63

 

.HmbmH mo.o um uSMOHchmHm cums smmsumn OOSOHOMMHO u x

 

><><><><><><><><><

XXXNNN

NNXNNV!

><><><><><><

NNXNNX

NNNNXX

N N x

NNNXNX
><><><><><><><
NNNNXNN
XXNXNXN
NXXNNNX
X><><><><><><

«4.
ma
3
8
me
Ho
3
mu
3
3
mo
.2
em
mm
3
.2

 

“I
4

mfl,

vQ

HQ

ma.

HU

mxxxxx

ND ND v0 mo Hd ¢m mm _Nm Nd

mSOHpHmom
comHHmmEoo

 

.msmmE HHmnp swmsvmn mmocmummme

How SOHuHmom nonpo mum>m on pmummfioo SOHuHmom some mo MMHmmHoII.oH mqmss

 

.Ho>ma mooo.o pcosmpupamoemmcmHm u *

 

64

 

HmHH mmn.Hm>mw Hmuoa
mom.w mmv «Hm.HmNN omm+oomsuu

emo.H mNh.v omm oem.mmmH mama StomISOHuHmomIHmso> Auxmxdv

*mmN.N mom.0H mmH mmm.mm>H mSOHpHmomIHmzo> Amxdv
hmv.mmmH m mm¢.mmov oo+onN

mmH.N «0H.mmmN N mom.momm make atom HOV
mum.th mv mNm.mmNmH Qm+oomuN

mon.o hmH.HNm om NOH.vmmm mama NpomISOHuHmom Auxmv

#«Hn.v va.hHON mH mmm.vaom coHuHmom amv
Hmo.NN mm Nvo.>N> o<+ou¢uN

mms.o eam.HH mm mam.omm mass spomuamzo> Auxac

*nmm.vv mmv.mmm HH mmv.mmeH Hmso> Amv

OHHmHumum mumsvm mo mmumswm
m cam: Ho saw mosmHHm> mo mousom

 

.mump ummumu How
mama upon tam .mSOHUHmom .Hm30> mSHummEoo mocmHHm> mo mHmmHmsm mo mnmEEsmII.HH mqmde

65

correlation of 0.546, significant at the 0.01 level, was
obtained. Thus, the scores obtained on test and retest

sessions were found to be statistically similar.

Discussion

 

The results of examination of the data collected
for this experiment indicate both specific conclusions
and trends which are note-worthy. The analysis of variance
treatment points out that there are statistically significant
differences among the mean relative transmitted sound
pressures for each vowel, anatomical position, and for
the interaction between vowels and anatomical positions.

Further statistical analysis indicates that the
vowels /i/ and /u/ were transmitted through tissue
significantly more intensly than any of the other vowels
employed in the study. The least intensly transmitted
vowels were the /A/, /a/, and /ay£ It Should be noted that
the /i/ and /u/ are both high vowels, whereas the /A/,

/a/, and /ae/are central and low vowels. Apparently

the high front and back vowels are transmitted more easily
through the tissues of the head and neck than any of the
other vowels.

Re-examination of Table 9 indicates two distinct
general groupings of the mean relative transmitted Sound
'pressures through the different anatomical sites, when
the means for all locations are compared to each other.

Group I, the high relative sound pressure region, consisted

66

of the eight locations showing the highest mean relative
sound pressures transmitted. Included in this group are
the following positions: top—center of the Skull,
frontal bone, nasalis bone, left and right side of neck,
and left and right ramus of the mandible.

Group II, the low relative sound pressure region,
consisted Of the eight locations showing the lowest mean
relative intensities. Included in this group are the
following positions: below the notch of the thyroid, left
and right temporal bone, left and right wing of the sphenoid,
and the area near the occipital protuberance. Although
there are significant differences among the means of
transmitted sound pressures within each group, the grouping
suggested above appears to be a meaningful one. An interest-
ing finding was that the relative sound pressure at the
thyroid notch of the larynx was not in the high intensity
group. This may have been due to the vertical movement
of the larynx during phonation resulting in unstable
positioning of the accelerometer.

Interactions between vowel and position proved to
be, for the most part, a random function. It was observed
that the mean relative sound pressures of the interactions
of the vowels locations with the highest relative sound
pressures transmitted were ranked highest in the list of
such interactions. Thus, the vowels /i/ and /u/ when

transmitted through the top-center of the skull and the

5-01—1M__ - —.

67

frontal bone had greater mean relative sound pressures

than any other vowels at any other locations. It was also

noted that the vowels /a/ and /ay’when transmitted through

the occipital protuberance region had lower mean relative A

sound pressures than any other vowels at any other locations.
Additionally, it would be recalled that there may

possibly have been an interaction of the resonances created

by the helmet with the sound pressures transmitted. It is

.“Mn—T _ ~

possible that some of the differences of the transmitted
sound pressures obtained may have been due to these
resonance factors rather than to the anatomical character?
istics of the positions from which the measurements were
taken.

Further statistical analysis of the positions indicates
that the anatomical sites of the top-center of the Skull and
the frontal bone had significantly higher transmission of
relative sound pressure than any other positions. The I
positions through which the relative sound pressures of the
vowels were transmitted less efficiently were the occipital
protuberance and the point two inches below the occipital
protuberance; both of these positions are located in the
back of the-skull.

Results of previous research in this area do not
Show close agreement concerning which anatomical sites
transmit speech sounds most efficiently with respect to

relative sound pressure. Mullendore1 found that the thyroid

 

1Mullendore, "An Experimental Study . . . ," pp. 163-
176.

68

cartilage, the mandible, the nose, and the top of the
skull transmitted vowels with greatest intensity.
Snidecore, Rehman, and Washburnl found that thelarynx,
mandible, chin, and nose were most efficient; and Neely
and Forshaw2 found the forehead, back of head, and top of
Skull to be the most efficient Sites.

The study by Moser and Oyer3 employed the same
anatomical sites and the same vowels as the present study.
Moser and Oyer used three subjects who hand-held a bone
oscillator to the anatomical locations while phOnating.
Table 12 presents a random ordering of the results of the
Moser and Oyer study and the results of the present study.

The Moser and Oyer results were grouped into four
different anatomical sites according to relative intensity.
It was concluded that "except for the region at the top
and center of the head, the relative intensity of tissue-
conducted sound . . . is decreased as the distance from
the larynx is increased, . . ."4 The position at the top

and center of the skull is found in Group III which is a

relatively low group.

 

lSnidecor, Rehman, and Washburn, "Speech Pickup
by . . . ," pp. 277-281.

2Neely and Forshaw, "Speaking and Listening . . . ,"
pp. 151-157.

3Moser and Oyer, "Relative Intensities of . . . , "

41bid., p. 178.

 

69

TABLE 12.--Comparison of the results of the Moser and Oyer

Study to the results of the present study.

Positions ranked

in descending order of relative sound pressures.

 

Moser and Oyer Results

Present Results

 

Angle of Mandible-Right
Larynx

Angle of Mandible-Left
Neck-Right

Neck-Left

 

Mental Protuberance
Nasalis Bone
Occipital Protuberance—Below

Occipital Protuberance

 

Top-Center of Skull

 

Occipital Protuberance-Above
Temporal Bone-Left

Sphenoid Bone-Right

Temporal Bone—Right

Frontal Bone

Sphenoid Bone-Left

Top and Center of Skull

Frontal Bone

‘~

Neck-Right

Mental Protuberance

Neck-left

Angle of Mandible-Left
Nasalis Bone

Angle of Mandible-Right
Larynx .
Temporal Bone-Left
Temporal Bone-Right
Occipital Protuberance-Below
Sphenoid Bone Right
Sphenoid-Left

Occipital Protuberance

Occipital Protuberance-Above

 

Solid Underling indicates divisions between groupings

achieved in each study.

70

The present study indicated that the position at the

top and center of the Skull was the most efficient transmitter

of vowel sounds. The anatomical sites fell into two groups

on the basis of their relative sound pressures Group I

consisted of locations toward the top and front of the Skull.

Group II consisted of those Sites on the sides and back )

of the skull, except for the larynx itself. {1
Examination of Table 12 indicates that while there i

is not a one-to-one correspondence in the rank orderings ;

of the results of the Moser and Oyer study and the present

study, the findings of the two studies basically agree.

Inspection of Table 12 reveals that there is a seventy-five

percent agreement between the two studies with respect to

the transmitted relative sound pressure of vowels at those

anatomical locations that appeared above or below the mid-

point in the ranking. Both studies indicate that as

distance from the larynx increases, the transmitted relative

sound pressure of the vowel decreases. An exception was

noted in the case of the poSition beneath the thyroid notch

on the larynx. The relative sound pressure of this position

was found in the low pressure group. This may have resulted

from unstable placement of the accelerometer against the

larynx which tended to move during phonation. The present

study suggests that locations near the larynx and overlying

broad, flat bones transmit vowel sounds with greater pressure

than locations remote from the larynx and overlying irregularly

shaped bone or no bone at all.

CHAPTER V

SUMMARY AND CONCLUSIONS

This study was concerned with the investigation

of the relative sound pressures of vowel sounds as they are

.vfi'hmEW’W
-4

intoned by male subjects of different body types and

transmitted through tissues of the head and neck.

Summary

Speech scientists have recognized that tissue
which is remote from the larynx vibrates during voice
production. Tissue vibration was thought to contribute
to the overall quality of the voice during Speech and
singing. Later investigators, aided by the development
of electronic instrumentation for sound analysis, concluded
that tissue vibration during voice production was the
result of forced vibration, and probably did not contri-
bute significantly to voice quality.

Many investigators have examined the nature of
vibration of Speech sound through tissues of the human
body. Research in this area indicated that sounds are
more intense at some anatomical locations of the body

than others. However, there has not been consistent

71

72

agreement in results. The early investigators have
indicated that certain refinements in instrumentation and
procedures might provide more precise and consistent
measurements and might result in data which are more valid
and reliable.

The purpose of this investigation was to examine
the relative sound pressures of vowel sounds as they were
uttered and transmitted through the tissues of the head and
neck. The study was directed at investigating, eXperimentally,
the difference in relative sound pressure among twelve vowel
sounds picked up at sixteen anatomical locations employing
subjects with three distinctly different body types.

Six adult male speakers were employed in this
investigation. The subjects were selected to represent
three different body physiques: ectomorph, mesomorph, and
endomorph. Two subjects were placed in each body
_c1assification. The subjects all utilized the General
American Dialect and were familiar with the International
PhOnetic Alphabet. Twelve vowels were selected for the
investigation. Sixteen anatomical locations of the head
and neck were specified as sites to be investigated that
were at varied distances from the larynx and that represented
differences with respect to tissue compositions. In some
instances positions on both sides of the head were

included.

fF—Tm ‘-—-

73

A Special helmet and shield which had minimal
contact with the surface of the head and neck were
designed. The helmet and shield provided holes through
which four accelerometers could be anchored. In this way,
control was exerted over the security of application of the
accelerometers against the surface of the head and neck.

The accelerometer employed was an integrated-

circuit-piezoelectric quartz transducer that was sensitive

I
a

to very subtle vibrations and had a linear frequency curve
throughout the speech range. The four transducers were
applied to the anatomical site under investigation through
the helmet and Shield, and the subjects intoned the vowels
while monitoring the intensity of their voices. Through
the use of a sound level meter, the sound pressure level
of all utterances was held constant by all speakers.
Subjects also practiced maintaining a constant fundamental
pitch for each utterance. The transducers transmitted the
signals picked up at the skin surfaces to a tape recorder.
In this way, responses from four anatomical locations were
recorded simultaneously. The four transducers were moved
and the procedure repeated until the output from all Sixteen
locations had been recorded.

The investigator analyzed the tape recorded vowel
signals by means of a power level recorder according to
relative sound pressure patterns, and a graphic display of

each sample was obtained. Mean relative sound pressure

74

measurements were made from each of the graphic displays;
these measurements provided the data for this investiga-
tion.

The dependent variable under consideration was
relative sound pressure. The data obtained from the subjects
were treated by a three factor analysis of variance in which
the factors were (1) vowels, (2) positions, and (3) body '
type. A Pearson product-moment correlation was used to

compare original data to retesthdata for all subjects.

Conclusions

 

Within the limitations of this experiment, there
are certain conclusions that can be drawn from the
hypotheses tested and from observations made of the
outcomes of the research.

1. The vowels /i/ and /u/ are transmitted through
the tissue of the head and neck with significantly higher
relative overall sound pressure than any other vowels. The
vowels /a/ and / / are transmitted through these tissues with
significantly lower relative overall sound pressure than
any other vowels.

2. The anatomical positions at the top-center of
the skull, the forehead, the neck near the larynx, and the
mandible transmit vowel sounds with higher relative than

any other locations.

75

3. On the basis of relative sound pressure the
sixteen anatomical locations fell into two basic groups.
Group I-The Central Region--consisted of locations toward
the midline and front of the Skull, near the larynx. Group.

II-The Peripheral Region--consisted of those sited remote

~

from the larynx.
4. Highest relative sound pressures are obtained

for the vowels /i/ and /u/ at the positions on the tOp-center

l.)

of the Skull and on the forehead. It is possible to hypo-
thesize that these vowels at these positions were better
transmitted than any other vowels at any other positions
because of the Spectral characteristics of the vowels as
well as the density and solidity of the anatomical structures.
The spectral characteristics of the vowels /i/ and /u/ may
be such that they are more easily transmitted through
the bones of the skull than any of the other vowels
investigated.

5. When the accelerometers were attached to
the supporting system, it was found that a non-linearity
in frequency response of the transducers occurred. With
a constant pressure of application of 0.1 gravitational
acceleration there was a 10 dB increase in responses of
frequencies below the speech range. This did not appear
to have a Significant effect on the relative sound pressure

levels of the ten vowels. However, as pressure of application

76

was increased, the non-linearity of transducer response
curves increased so that relative sound pressure levels,
of transmitted vowels, of 10 to 16 dB were found in the
speech frequencies. Thus vowels at various anatomical
locations may have been transmitted with different sound
pressures because of the interaction within the supporting
system. The obtained data must be considered with this
possibility in mind.

6. No statistically significant differences were
found among relative intensities among body types. However,
mesomorph subjects showed a trend toward lower relative
sound pressures across all vowels than did nother subjects.
This may be related to the solid muscles and tissue
structure associated with the mesomorph physique.

7. The measurements obtained by the procedures
employed in this eXperiment were shown to be reliable by
results of comparison to retest data.

8. Differences between results obtained by this
exPeriment and previous studies may be attributed to

differences in instrumentation and procedures employed.

Suggestions for Future Research

 

From the results of the study and from observations
of the procedures used, several suggestions for future

investigation can be made.

77

1. While the helmet and shield provided a secure
means of attaching the transducers to the head and neck,
they also introduced considerable subject discomfort and
difficulty in manipulating and changing the location of
the transducers, and nonlinearity of the frequency response
curves of the accelerometers.

Further research in this area may attempt to
create a method of attaching the transducers in such a
way that the entire surface of the head and neck is
uncovered. The transducers could, for instance, be
attached to arms which could be moved toward the surface
of the head and neck from the sides. In this way,
adjustments in location site and pressure of application
could be made more easily and with greater precision.

This would also reduce the discomfort experienced by the
subjects.

In addition, refinement in attaching the transducers
to the body surface might overcome the nonlinear responses
when the accelerometer was attached to the supporting
system.

2. It would seem that several studies similar to
the present one could be replicated meaningfully using the
same equipment, but with certain refinements in procedure.
Controlling for differences in pressure of application
of the transducers to the surface of the head and neck
would provide a clearer understanding of the effect of

varied pressure on the measurements.

Th7”

78

3. In addition, any replication of this study
should attempt to investigate larger groups of ectomorphs,
mesomorphs, and endomorphs to determine whether the trend
indicated by the data is a significant one.

4. A similar study to the present one could be
developed in which the obtained samples were judged by a
panel of listeners for intelligibility. In conjunction
with this, a spectral analysis of the frequency components
of each sample could be made. In this way, significant
relationships may be found between intelligibility,
frequency characteristics, and anatomical location.

5. Significant information could be obtained
through a study similar to the present one but employing
subjects with widely different fundamental pitches.

6. Initially it was stated that contact micro-
phones used in military, government, and scientific Opera—
tions might be better employed if optimal anatomical
positioning was discovered. The present investigation
suggests that the tOp-center of the skull and the forehead
were areas at which certain vowels were best transmitted.
Various contact micrOphones should be examined carefully
to determine if their frequency response curves may tend
to amplify vowel sounds, consonant sounds, and even
connected speech in different ways. A contact microphone
may possibly be designed which, when placed on the tOp-

center of the skull or on the forehead would pick up

79

tissue transmitted Speech signals whos intelligibility
would be apprOpriate for communication purposes. Thus
a contact micrOphone might be employed in place of an
air conduction micrOphone and result in safer as well as
more efficient transmission of speech of astronauts, deep
sea divers, mine workers, and aircraft personnel.
It is not clear what effect, if any, the nonlinearity

of response curves of the supporting system may have upon

 

the transmission of vowel sounds. However, the nonlinearity
suggests that perhaps contact microphones may not be an
accurate means of transmitting Speech sounds because
of potential distortion. Future investigation should
consider the use of supporting systems which will not
affect the intelligibility of the transmitted signal,
or the use of air conduction microphones.

Comparisons of intelligibility of air conducted
sound with intelligibility of tissue conducted sound may
provide a means of determining the overall usefullness

and efficiency of contact microphones.

BIBLIOGRAPHY

80

CIT—Til

BIBLIOGRAPHY

Books

 

Aiken, W. A. The Voice. London: Longmans, Green, and
- Company, 1920.

Bekesy, Georg von. Experiments in Hearing. New York:
McGraw-Hill Book Company,'l960.

and Rosenbligh, Walter A. "Mechanical Properties
of the Bar," in Handbook of Experimental Psychology,
S. S. Stevens, ed., New York: John WiIey and

Sons, Inc., 1951.

. Sensory Inhibition. New Jersey: Princeton
University Press, 1967.

Blalock, Hubert, M. Social Statistics. New York:
McGraw-Hill Book Company, 1960.

Curtis, H. H. Voice Training and Tone Placing. New York:
Appleton Company, 1909.

Fillebrown, Thomas. Resonance in Singing and Speaking.
Boston: Oliver DiSton Company, 1911.

Frost, H. M. The Laws of Bone Structure. Illinois:
Charles C. Thomas Company, 1964.

Hays William L. Statistics for Psychologists. New York:
Holt, Rinehart, afia Winston, 1964.

Judson, Lyman and Weaver, Andrew. Voice Science. New York:
Appleton-Century-Crofts, 1965.

Kerlinger, Fred N. Foundations of Behaviorial Research.
New York: Holt, Rinehart and Winston, 1966.

Lindquist, E. F. Desigg and Analysis of Experiments in
Psychology and Education. Boston: Houghton
Mifflin Company, 1953.

McLean, H. and Urist, M. Bone. Chicago: University of
Chicago Press, 1961.

81

82

Muckey, Floyd 8. Natural Method of Voice Production.
New York: Scribnerjs Sons, 1915.

 

Newby, Hayes A. Audiology. New York: Appleton—Century-
Crofts, 1964}

 

Scott, John R. Technique of the Speaking Voice. Missouri:
Stephens Company, 1915.

 

 

 

 

Sheldon, W. H. Atla§_of Men. New York: Harper and
Brothers, 1954. L
1
. Varieties of Human Physique. New York: Hafner
Publishing Company, 1963. E

Woolbert, M. and Weaver, A. Better Speech. New York: L4
Harcourt, Brace Company, I922.

 

Articles and Periodicals

 

Bekesy, Georg von. "Vibration of the Head in a Sound
Field and its Role by Bone Conduction," Journal
of the Acoustical Society of America, XXI(1948),
pp. 749-760.

 

. "The Structure of the Middle Ear and the Hearing
of One's Own Voice by Bone Conduction," Journal

of the Acoustical Society of America, XXI(1949),
pp. 217-232.

 

 

"Human Skin Perception of Travelling Waves
Similar to those on the Cochlea," Journal
of the Acoustical Society of America, XXVII(1955),
pp. 830-891.

 

 

Kirikae, 1.; Sato, T.; Oshima, H.; and Hirose, H. "An
Experimental Study of the Hearing One's Own Voice,"
Praggica Oto-Rhino—Lagynqologica, XXIII(1961),
pp. 56-71.

 

Dhaser, Henry M. and Oyer, Herbert J. "Relative Intensities
of Sounds at Various Anatomical Locations of the Head
and Neck during Phonation of the Vowels," Journal
of the_Acoustical Society of America, XXX(1958),
pp. 275—277.

 

buillendore, James M. "An EXperimental Study of the Vibration
of the Bones of the Head and Chest During Sustained
Vowel Sounds," Speech Monographs, XVI(1949),
pp. 163-176. ’

 

83

Neely, Keith K. and Forshaw, S. E. "Speaking and Listening.
through the Head: I. the Intelligibility of
Speech Recorded in Quiet at Different Positions
on the Head and Throat,“ Journal of Auditory
Research, V(1965), pp. 151—157.

Simon, Clarence T. and Keller, Franklin. "An Approach to
the Problem of Chest Resonance," Quarterly Journal
of Speech, XIII(1927), pp. 432-439.

 

Snidecor, John C., Rehman Irving, and Washburn, David D.
"Speech Pickup by Contact MicrOphone at Head and
Neck Positions," Journal of Speech and Hearipg
Research, II(1959), pp. 277-281.

Stong, R. A., Chant, V. G., Forshaw, S. E., and Neely,
Keith. "Speaking and Listening through the Head:
II. the Intelligibility of Speech Recorded in
Noise," Journal of Auditory Research, VI(1966),
pp. 385—391.

Wise, Claud M. "Chest Resonance," Quarterly Journal of
Speech, XVIII(1932), pp. 446-452.

Uppublished Materials

Agricultural EXperiment Station. "STAT Series Description
No. 14: Analysis of Variance with Equal Frequency

in each Cell." East Lansing: Michigan State
University, 1968.

. AFS MISC Series Description 108, "Duncans

Multiple Range Test." East Lansing: Michigan
State University, 1968.

"ICP Transducer Data and Information Manual." New York:
PCB Piezotronics Incorporated, 1969.

Lindsley, C. F. "Psychophysical Determinants of Voice
Quality." Unpublished Ph.D. Dissertation,
University of Southern California, 1934.

 

APPENDICES

84

 

APPENDIX A

SHELDON'S CLASSIFICATION SYSTEM

85

 

.1. _ 3g

 

Sheldon attempted to find a first-order criterion
for the classification of human physique.t The Sheldon
body classification system resulted from a study of the
physiques of over thirteen—hundred undergraduate males.
He photographed his subjects from three angles: back,

front, and side. In addition to obtaining the height

 

and weight for each subject, he made seventeen precise
measurements from the photographs. Sheldon found three
basic components of body constitution which could be
used to classify the body structure of every subject.
These were ectomorph, mesomorph, and endomorph.

In endomorphy, there is a predominance of
softness and roundness with massive digestive viscera.
In mesomorphy, there is a predominance of muscle and
connective tissue. The mesomorphic structure is hard and
rectangular. In ectomorphy, there is a predominance of
linearity and fragility. The ectomorphic structure has
the greatest surface area in pr0portion to its mass and

'the largest brain and central nervous system.

 

1Sheldon, Atlas of Men; Sheldon, The Varieties of
Human ngsique.

 

 

86

87

Sheldon concluded that every physique had
components of all three morphologies. He devised a
systemfor classifying each physique by three numbers.
The first number represented the predominance of
endomorphic component; the second, mesomorphic, and the
third ectomOrphic. A scale of seven points was used
with the number one representing a minimal degree of a II
component and number seven representing a maximal degree a
of a component. This patterning of three body components a
was called somatotyping. Sheldon was able to demonstrate
seventy-Six different somatotypes.
The pure endomorph was classified as 7 1 l; the
pure mesomorph was classified as l 7 1; and the pure
ectomorph as l l 7. A somatotype with an average degree

of all components was classified as 4 4 4.

APPENDIX B

DESCRIPTIVE DATA FOR SUBJECTS

88

 

APPENDIX B. —-Descriptive data for the six subjects employed

in the study.

 

 

Subject Age Height Weight Body
(Years—Months) (Inches) (Pounds) Type

1 28-0 72 130 Ectomorph

2 23—0 72 145 Ectomorph

3 25-9 71 175 Mesomorph

4 27—9 71 190 Mesomorph

5 29—5 73 205 ,Endomorph

6 35-0 70 185 Endomorph

 

89

 

APPENDIX C
MEAN AGE, HEIGHT, WEIGHT FOR

BODY CLASSIFICATIONS

90

APPENDIX C.—-Mean age, height, and weight for each body

 

 

classification.
Body Type Mean Age Mean Height Mean Weight
_ (Years-Months) (Inches) (Pounds)
Endomorph 24-6 72 137 . 5 r1
Mesomorph 26-7 71 182.5
Ectomorph 32-3 72 195.0 J

 

 

91

APPENDIX D

FREQUENCY RESPONSE CURVES OF TAPE RECORDERS

92

.. 'nly

IICOID

   

IIPIODUCI

 

Frequency response curves of tape recorders.

93

 

APPENDIX E

EXPERIMENTALLY DERIVED CURVES

FOR ACCELEROMETERS

94

Relative Sound Pressure in dB

15

10

95

 

1 l I 1 l l I l l

 

10 20 50 100 200 500 1000 2000 5000
Frequency in H2

Output from accelerometer mounted in helmet:
Coronal Location
Constant Input of 0.1g

Relative Sound Pressure in dB

15

10

l

 

10

96

l I l l l I I

20 50 100 200 500 1000 2000
Frequency in H2

Output from accelerometer mounted in helmet:
Frontal Location
Constant Input of 0.1g

1

5000

I

Relative Sound Pressure in dB

97

15

7

10-

U1
T

 

Otl I I l I I l I

10 20 50 100 200 500 1000 2000
Frequency in H2

Output from accelerometer mounted in helmet:
Frontal Location
Constant Input of 0.3g

I
5000

Relative Sound Pressure in dB

 

98

 

15—

10-

5»—

OI'I I I I I L A I l
10 20 50 100 200 500 1000 2000 5000

Frequency in H2

Output from accelerometer mounted in helmet:

Frontal Location
Constant Input of 0.59

APPENDIX F

STATISTICAL METHODS EMPLOYED

99

 

Assumptions underlying statistical tests: Before
selecting and proceeding with the several statistical
analyses to be attempted, it was necessary to consider the
extent to which the assumptions underlying the use of
the statistics had been met. Analysis of variance,
multiple comparison tests, and a Pearson product—moment
correlation were the statistics employed.

The assumptions underlying the use of the analysis
of variance were (1) normality: the samples selected were
drawn from populatibns that are normally distributed;

(2) homogenity of variance: the variances are assumed to
be statistically the same from group to group, within

the bounds of random variation, over the observations made;
and (3) continuity and equal intervals of measure: the
measures to be analyzed are continuous measures with equal
intervals.l

Examination of the data obtained from the six
subjects reveals that the distribution of scores was
mildly skewed but remained as a good approximation to a

normal distribution. Lindquist points out that the F-test

 

lFred N. Kerlinger, Foundations of Behavioral
Research (New York:. Holt, Rinehart and Winston, 1966),
pp. 258-260.

 

100

 

101

is quite insensitive to the shape of the distribution, even
when the distribution is moderately to markedly skewed.
He has stated that the F-test is so insensitive

. . . to the form of the distribution of criterion
measures that it hardly seems worthwhile to apply

any statistical test to the data to detect non-
normality, even though such tests are available.
Unless the departure from normality is so extreme that
it can easily be detected by mere inspection of the I
data, the departure from normality will probably have 1 3
no appreciable effect on the validity of the F-test. . . "

 

Since examination of the data reveals only mild Skewedness, L
it would appear that the assumption of normality is
satisfactorily met.

Regarding the second assumption of homogeneity of
variance across treatment groups, Kerlinger points out
that this assumption along with the assumption of normality
are generally overrated. He indicates that unless there
is good evidence to believe that homogeneity of variance
is violated, a parametric statistic such as the F-test
should be employed.2 Hays supports the View that so long
as sample sizes are equal for all conditions of the
analysis, violation of the homogeneity assumptions have

little effect upon the F-distribution.3 In the present

 

1E. F. Lindquist, Design and Analysis of Experiments
in Psychology and Education (Boston: Houghton Mifflin
Company, 1953), p. 87.

 

2Kerlinger, Foundations . . ., p. 259.

 

3William L. Hays, Statistics for Psychologists
(New York: Holt, Rinehart and Winston, 1964), p. 408.

 

102

study, sample sizes for all conditions of the analysis
were equal. The assumption of homogeneity of variance
Seems to be satisfied for this study.

The third assumption of continuous measures with
equal intervals seems to be met also. Blalock states
that an "interval scale level of measurement requires a
physical unit of measurement which can be agreed upon as
a common standard and which is replicable."1 The unit of
the decibel, which is the unit of measurement in this
study, satisfies all of these requirements.

The three-factor analysis of variance for repeated
measures on two of the three factors was selected as the
most apprOpriate statistic to test the null hypotheses
derived from the questions concerning relative intensity.
The variable of body type (subject) did not have repeated
measures, where-as the variables of vowel and anatomical
location did. Although a sample size of two in each body
group was small, the large amount of data collected for
each subject allowed the use of analysis of variance
technique.

Duncan's Multiple Range Test2 was selected to

determine the significant differences between the array of

 

1Hubert M. Blalock, Jr., Social Statistics (New York:

 

McGraw-Hill Book Company, 1960), p. 14.

2Agricultural EXperiment Station, AES MISC Series
Description 108, Michigan State University, "Duncan's
Multiple Range Test", 1968.

j;

103

means generated by the analysis of variance because of the
large number of means which were compared and the large
number of degrees of freedom associated with them. This
statistic is a multiple comparison test which calculates

the Shortest significant range using Harter's table which is
associated with degrees of freedom, number of means being
compared, and the confidence level desired. Each mean is
compared to every other mean to determine whether their
differences exceed the shortest significant range.

The Pearson product-moment correlation coefficient
was used to assess the correlation between the subjects'
test-retest measurements. This correlation is based on
the concomitant variation of members of sets of ordered
pairs. Each member of the pair must be measured in an
ordinal scale or a higher scale.1 The present measurements
were made in an interval scale which is higher than the
ordinal scale required. No further assumptions need be
made in the use of this statistic.

The Pearson product—moment correlation treats the
paired scores as test-retest data because the same subjects
were used.

The presentation of the results of the statistical
treatment is organized on the basis of the two-way and

three—way comparisons of the independent variables,

 

1Blalock, Social Statistics, p. 286.

 

 

104

followed by test-retest correlation analysis. Because
meaningful observations can be made by examination of the
original data, a section discussing the descriptive analysis
of the data is included. A discussion of the obtained

results concludes this chapter.

 

The null hypotheses generated by the seven questions
concerning the relative intensity measurements obtained L“
from the subjects were tested by means of three factor
I...

analysis of variance for repeated measures. The repeated
measures entered the analysis as the factors of VoWelS,
Positions, and Body Types.

The three factors of the analysis could be
imagined to be arranged as a three dimensional table of a
12 x 16 x 3 design. Twelve columns would represent the
twelve vowels, and sixteen rows would represent the
sixteen anatomical positions. The third dimension could
be conceptualized as three layers of the 12 x 16 matrix
(Vowels x Positions). The layers would represent the
three groups of subject body types--ectomorph, mesomorph,
and endomorph. Relative intensity values, the dependent
variable, would be entered for each of the six subjects

in the apprOpriate cell of the 192 cell matrix associated

with this body type.

APPENDIX G

ORIGINAL AND RETEST SCORES

105

 

106

.muoom ummumu mmumoHpsHa

 

 

 

 

o.VI o.NH H.mI h.MHI m.ml ¢.HHI H.mI o.mHI o.ml N.vHI o.mI m.OHI «Q
m.ml v.0HI o.ml N.NHI m.hI N.ml o.ml o.mHI m.ml m.HHI H.hl N.wl mo
H.vl o.ml o.mHI o.NNI N.mHI o.mHI o.MHI o.HNI N.HHI m.mHI m.VI o.OHI ND
m.NI N.H m.oHI m.NHI m.ml o.HHI o.ml Nm.l o.NHI o.0HI o.NHI w.mHI HQ
N.ml m.N m.0HI o.¢l m.HHI N.vI m.ml ¢.NI o.mHI m.ml o.vHI m.vl wU
N.ml m.mHI w. «H! m.NNI m. VHI m. NNI o.MHI w.HNI 0.0HI m. vNI o.VHI N.MNI mU
N.ml m.¢ m. MHI N.wI h. HHI m. VI m.¢I m.NI m.mHl w. HHI w.MHI m.hl NU
m.m m.mI N.NI m.HI o. NI o.NI o.v m.H H.ml e. ml m.vl w.mI HO
m.ol o.OHI m.HI m.0HI N.HI m.mI N.NI m.NHI m.¢l o.MHI m.NI m.ml vm
¢.MHI h.mHI 0.0HI N.0NI H.hHI m.mHI w.mHI m.hHl H.mHI «.mHI o.mHI w.mHI mm
N.mHI v.HNI v.0HI o.MNI m.mHI m.HNI o.mHI m.0NI m.hHI H.vNI m.mHI m.MNI Nm
o.th o.HHI m.mNI m. mHI o.mNI m. NHI w.hNI m. mHI m.mNI m.NNI N.mNI m.0NI Hm
o.HI o.m . N.¢I o. m m.ml N. v N.NI m. v m.hl o.N o.wl m.N v<
o.ml v.MI H.wI N.ml m.MI 0. ml m.MI m.ml H.mHI w.0HI v.HHI m.wl Md
h.mHI m.mHI N.vHI m.mHI m.MHI m.mHI o.NHI m.mHI b.0HI m.mHI N.mHI v.0HI N<
o.mHl o.mNI N.HNI >.NNI o.mHI m.NNI o.mHI o.MNI m.NNI 0.0NI m.NNI m.MNI Hd
« 5 a s O a a «
h.ml o.HHI N.HHI m.mHI N. OHI m.ml h.hl H.mHI m.bl v.NHI N.NI m.wl vQ
o.ml v.ml b.9HI o.vHI H. OHI m.NHI w.hl m.HHI b.ol o.ml N.HI o.ml mo
m.HHI N.HHI w.mHI m.HNI H.ml m.mHI m.mI m. NHI o.wl h. mHI o.m o.mI ND
H. HHI m.mHI m. CHI o.MHI o.¢l o.ml H.wl 0. VI m.ml m. HI m.mI N.m HQ
N. mHI o.ml o. VHI o.mI h.NHI m.ml N.HHI h.MI b.0HI N.ml w.vl H.v ¢U
N. ¢HI H.mNI m. «H! m.MNI N.MHI o.vNI Io.HHI N.HNI m. HHI H.0NI o.HI m.OHI mu
m. mHI o. vHI o.vHI h.ml m. NHI m.mI m.hl w.ol 0. 5| o.N N.H N.h NU
m. ml N. CHI o.mI o.NI m. NI v.wI m.H o.m o.N ¢.m m.m N.vH HU
H.ml o.HHI N.VI m. HHI m.ml m.NHI o.vl m. MHI H.MI ¢.MHI o.N o.mI vm
N.mHI m.mHI ¢.HNI m. HNI o.hHI ¢.mHI m.vHI m. mHI N.MHI ¢.MHI m.ml m.HmI mm
H.wHI N.MNI o.hHI m. VNI H.hHI N.MNI w.mHI v. NNI o.mHI m.MNI N.MHI v.0NI Nm
o.le m.NNI m.bNI m. MNI m.mNI m.mHI o.mNI o.hHI v.VNI m.MHI m.mHl o.hI Hm
v.5l m.HI m.hl m. HI m.ml N.H m.vl o.v o.ml H.m N.H m.h v4
m.OHI N.HHI m.OHI o.HHI H.oI m.ml o.VI N.vI o.ml m.0I o.N m.hl m4
N.©HI h.mHl N.wHI «.mHI o.VHI H.hHI H.mHI m.vHI w.NHI m.vHI H.mI N.NHI Nd
o.MNI m.th m.MNI m.mNI m.mHl N.mNI H.5HI o.mNI o.hHI v.vNI h.¢HI H.mNI H<
.Iw m a « a m * H a H mcoHuHmom
MHm30>

 

.nmuoeouomIIH uomnosm

"mmuomm ummumm pom HmchHHoII.0 xHozmmm¢

«I!

 

.muoom ummumu mmumoHch«

.107

 

 

 

 

 

o.o o.w o.HI m.h m.HI o.m m.NI o.N m.NI o.m N.o v.m va
m.HI m.0H m.NI o.N m.MI m.m o.vI m.h m.mI o.m m.vI v.m mm
H.@ m.m m.H m.m m.v H.N o.m v.v N.NI m.o m.mI m.0I No
o.HH m.h o.m N.VI m.h o.mI 0.5 m.mI m.w m.mI o.m H.MI HQ
m.mHI o.m m.vHI H.m m.mHI o.N o.HHI m.m m.oHI o.HI m.mHI H.HI v0
o.mHI N.MI N.nHI H.vI m.HNI o.NI h.vHI H.HI N.NHI m.HHI m.mHI m.mI mo
h.m o.N m.HI m.>I m.MI o.mI o.v m.¢I N.NI m.NHI m.mI N.0HI NU
m.OH m.¢I N.H N.NHI m.m o.HHI H.m H.mI m.MI m.mHI o.HI H.NHI HO
o.NI N.0I H.vI N.NI h.mI m.MI v.HI N.HI N.mI N.wI h.mI m.mI vm
o.mI N.HHI m.HHI o.NHI m.HHI >.mHI m.mI N.MHI o.mHI H.0NI o.MHI N.NHI mm
o.HHI m.nHI o.MHI m.mHI N.MHI m.mHI N.OHI m.mHI m.vHI m.omI m.MHI o.HNI Nm
w.mH N.m m.mH N.H o.VH o.HI «.mH m.q m.OH m.vI o.HH m.mI Hm
o.mI N.MI o.NI m.mI H.NI m.mI H.NI N.HI h.oHI m.mI m.oHI m.mI ed
m.m N.OH m.N N.m m.OI o.N m.N m.> m.NI o.mI m.NI N.mI m4
H.OHI N.NI o.mHI N.NHI o.MHI o.mHI n.oHI v.vHI n.vHI o.NHI N.VHI m.MHI Nd
m.mI m.hI o.vI N.oHI o.mI o.HHI N.vI m.mI o.NI m.OHI N.mI m.oHI Hd
« S a a O a a. < a.
o.HI m.m H.HHI N.m H.mHI o.m m.NI n.m m.vI N.v o.N o.NH v0
m.oI m.v n.vHI o.o m.OHI o.o m.0HI m.OI v.wI v.N m.0I 0.0H mo
o.HI m.NI o.HI o.mI o.mI 5.0 o.NI h.HI m.n m.H m.m o.v No
o.m m.o o.N m.m m.m 0.0I m.H N.HI m.v N.H o.mH m.HI HQ
m.mHI m.MI o.NNI N.NI H.mHI N.HI o.mHI m.m m.MHI m.m o.mHI o.N v0
m.mNI w.mHI H.mNI m.oHI o.mHI m.mI N.mHI H.mI m.MHI o.NI o.HHI v.HI mo
m.mI m.mI N.mI m.HHI N.NI N.NI m.mI >.HI o.HI m.m o.m m.m No
n.0I m.HHI h.OI N.0HI m.0I o.NI N.H v.qI m.m w.MI o.m m.NI HO
H.mI H.mI m.HHI m.0HI «.mI m.wI N.NI m.mI o.H v.0 H.m m.v em
N.VHI o.ONI H.vHI n.0NI m.vHI H.oNI o.MHI o.mHI m.NHI o.NHI v.5I m.0HI mm
o.mHI m.HNI m.mHI N.NNI v.mHI o.NNI o.mHI o.ONI N.mHI N.NHI o.HHI N.NHI Nm
m.> m.vI N.m m.hI o.OH N.mI N.NH o.HI m.MH m.o 0.0N o.OH Hm
N.NI m.mI N.HHI m.mI H.NI m.nI m. I «.mI H.mI o.VI m.o o.H vd
m.mI m.MI m.oI o.VI m.mI m.mI m.MI H.m m.NI o.v 0.5 c.0H m<
H.NHI w.mHI o.MHI m.mHI m.mHI m.mHI H.0NI w.mHI n.mHI o.mHI o.HHI m.oHI N<
w.mI N.HHI o.vHI H.mHI n.0HI N.NHI H.mI o.mHI o.mI m.mI o.HI N.H H<
« m « x m « m « H « H mcoHuHmom
mHm30>
.nmuosouomuum pomHnsm .pmscHuaoouu.O xHozmmm<

2108

 

 

 

 

 

IIIILF «I h.

. QHOUM “mmgmcfl MGHMUflUCHfl
m.ml o.NH 0.MHI 5.0 0.MHI 0.0 m.MHI 0.5 m.MHI 0.0 0.mHl N.m VG
o.NI m.VI 0.0I H.0I m.vl m.hl m.ml v.0I 0.0I 0.ml 0.hl N.NI m0
0.bl 0.NI .0.MHI m.ml o.NHI N.ml 0.vHI 0.0I m.VHI «.ml N.HHI b.0l ND
0.0 N.NI 0.0 ¢.vHI N.H h.MHI o.N v.HHI N.NI N.0NI m.MI o.NNI HQ
0.0HI o.NNI m.0NI o.NNI 5.0HI o.NNI m.mHl o.NNI 0.HNI o.NNI N.HNI 0.5Nl e0
o.NI m.0l 0.mHI m.ml 0.mHI 0.5! 0.MHI ¢.ml N.NHI N.NHI 0.0HI H.0I MU
0.H 5.0! 0.hl m.ml m.hl N.MI 0.5! 0.NI N.NHI h.mI 0.MHI N.ml NU
m.m N.h m.NI N.NI m.NI 0.NI o.NI 0.0I N.NI b.0HI 0.0I 0.mI HU
0.0HI H.MNI 0.MNI H.NNI 0.0NI 0.vNI 0.0NI m.MNI 0.HNI 0.MNI N.HNI m.MNI em
0.ml H.vHI m.HHI 0.wHI o.HHI N.NHI N.0HI v.0HI N.NHI 0.0HI m.MHI m.0NI mm
m.NHI 0.MI N.NHI 0.0I o.NHI h.ml m.hHl m.ml N.NHI o.NHI N.NHI o.NH! Nm
N.HI m.MI m.ml H.0Hl N.0H! N.NHI N.ml H.0I N.NHI N.HHI m.MHI 0.0HI Hm
0.m m.N 0.HI H.NI 0.NI m.HI H.0I m.0l 0.0I v.5! 0.HI v.MI vd
0.m 0.NI N.NI N.NI m.HI N.NI N.ml m.HHI N.NI 0.NHI N.NI N.0HI md
m.ml h.ml 0.HHI b.0HI 0.0HI v.mHI 0.0HI 0.0HI N.NHI 0.0NI m.VHI «.mHI Nd
m.mHl 0.0HI m.HNI m.mHI m.0NI 0.0HI N.NNI N.NHI m.¢NI 0.MNI N.mNI N.0H! Hd

« 5 a a O a < a «

o.NHI h.N o.NHI N.N m.0HI 0.v N.mHI 0.0 o.HHI h.m m.ml 0.HH vQ
b.0HI m.0HI m. I m.NHI 0.0I m.0HI N.NI m.ml H.ml H.NI h.el h.vl MD
m.MHI 0.0HI 0.0HI o.HHI 0.0HI m.hl m.0HI 0.vl 0.0I 0.¢I 0.0 N.HI ND
m.HI N.mNI m.H h.mNI o.N 0.MNI m.m N.HNI H.N m.mNI 0.MH N.NHI HQ
0.HNI o.NNI m.0NI o.NNI m.mHI 0.5NI 0.mHI o.NNI 0.0HI N.NNI 0.mHI N.NNI v0
N.NHI 0.0I m.mHI N.HHI 0.vHI m.hl m.HHI N.ml m.ml m.NI 0.ml N.N MU
0.VHI m.mI N.NI N.NI 0.0I m.HI o.NI m.m 0.HI 0.0 m.v N.HH NU
o.HHI o.NI w.ml 0.0I m.NI m.N- o.N H.m v.m 0.m v.m N.0H HU
N.NNI N.NNI N.NNI m.NNI 0.0NI v.MNI N.mHI o.NNI o.NHI m.mHl N.NHI 0.HNI vm
0.0HI m.mHI 0.0HI m.0NI m.MHI m.mHI m.HHI «.mHI m.0HI 0.HHI m.HHI m.MHI mm
0.0NI N.NHI N.HNI o.NHI «.mHI m.0HI v.0HI 0.ml m.MHI 0.wl m.HHI 0.mI Nm
N.NHI N.NHI N.HHI 0.0HI m.ml m.vl m.ml m.HI v.VI 0.H 0.0 N.H Hm
m.ml 0.0I m.ml m.vI 0.0 N.0I o.N h.m H.0H 0.HH 0.0 v.5 vd
n.0HI 0.mHI N.NI N.NHI N.0I 0.0I m.m1 N.mI 0.m m.NI m.m m.0 md
0.0HI N.HNI m.MHI 0.0HI 0.0HI N.mHI m.ml m.HHI 0.0HI 0.0HI 0.ml 0.0I Nd
¢.0NI N.NNI h.mNI m.mHI N.NNI N.mHI N.0NI m.MHI N.NHI N.NI 0.mHI m.0I Hd.

« m a a m a m a H a H mcoHuHmom

mHm30>

 

.tmuosommZIIm ummnndm

.omscHucooII.0 xHozmmmd

 

 

109

 

willlilflh.
.mHOUm HWQUNH mmUMUHUCHi
m.vl o.NI m.mI N.¢HI m.ml o.NHI m.>l o.NHI o.HHI m.mHI 0.mI 0.0HI 0Q
N.HI m.ml N.0I 0.HHI o.vl 0.0I m.MI m.mI m.hl N.NHI H.mI N.0I mo
0.HNI m.vHI n.0NI N.NNI 0.mNI 0.0NI 0.0NI m.0NI m.HNI m.mNI 0.mHI 0.0NI ND
m.N m.m o.N 0.¢I 0.m 0.0 0.v N.H m.NI m.HHI 0.0 o.NI HQ
m.HHI N.NHI h.mHI 0.0HI m.MHI H.NHI 0.¢HI N.vHI 0.0HI 0.0HI 0.0HI N.NHI v0
0.MHI m.0HI o.NHI N.vHI h.mHI m.mHI 0.mHI N.HHI H.0NI m.hHI 0.0HI m.oHI MU
o.HI m.ml m.HHI m.hHI N.0I m.MHI v.ml 0.MHI 0.mHI 0.0HI m.vHI o.NHI NU
N.N 0.h 0.mI m. I 0.VI 0.HI m.0I 0.0 w.MHI 0. I m.NHI 0.ml HU
m.NI H.NHI H.NI m.mHI N.vl m.mHI m.VI N.NHI m.0HI 0.0NI N.NI 0.HNI vm
m.NHI m.hHI 0.0NI N.NHI 5.0HI o.NHI m.wHI 0.0HI N.MNI m.wHI 0.HNI 0.0HI mm
0.HHI 0.0HI m.mHI m.MNI H.HHI N.HNI 0.MHI H.NNI o.NHI H.mNI 0.mHI m.VNI Nm
m.m v.m 0.0 o.HI m.m 0.v m.0 0.0 H.mI 0.hl 0.vI b.0l Hm
m.m 0.0H m.H m.0 N.v m.v m.NI h.v N.mI N.HI m.MI h.MI vd
N.¢I m.mI H.0I 0.MI m.mI o.HI m.ml o.NI b.0HI o.NI m.HHI 0.ml md
N.H v.0NI m.mI 0.0NI m.MI m.mNI 0.VI 0.0NI 0.0HI 0.0mI 0.0HI 0.0mI Nd
0.0HI 0.0I m.mHI o.NHI 5.0HI N.0HI m.hHI m.mHI m.0NI 0.HNI N.NHI w.mHI Hd
a 5 a t O s « < «
N.0HI m.HHI m.MHI 0.0HI m.HHI v.mHI o.NI 0.0I m.mI v.NHI o.NI m.vl «Q
o.NHI N.NI m.MHI b.0HI h.MHI m.HHI 0.MI 0.ml m.hl 0.mI 0.0 0.0 m0
m.mHI 0.HNI N.0HI m.MHI 0.mHI 0.MHI 0.ml o.NI N.0HI H.mHI m.0I m.ml ND
0.mI m.ol 0.hI 0.0HI m.ml N.NHI m.N m.mI H.0 o.HHI m.m H.H HQ
m.mHI o.NHI m.mHI N.NHI m.mHI N.0NI m.vHI v.0HI m.HHI m.hHI 0.mI 0.mI v0
m.mHI m.mHI N.NHI m.wHI o.NHI «.mHI m.mHI v.MHI 0.MHI m.VHI m.wl 0.mI mu
m.MHI N.NHI v.HHI N.0H! N.HHI b.mHI m.ml N.HHI m.ml N.MHI H.0 N.NI NU
N.HHI N.NI m.ml m.mI N.NI v.ml 0.mI v.0 m.vI o.NI 0.0 N.NH HU
H.0HI 0.HNI N.HHI N.HNI o.HHI 0.HNI N.ml m.0NI h.ml H.NNI H.N m.mHI vm
N.NNI v.0NI m.mml 0.0NI 0.MNI m.HNI m.mHI m.mHI m.0NI 0.HNI 0.ml m.mHI mm
m.0HI h.mNI N.NHI m.mNI 0.0NI v.mNI 0.hHI m.MNI H.0NI H.mNI H.0I 0.HNI Nm
0.mI v.mI o.NI 0.5I m.m 0.0HI N.N m.H N.NI m.ml o.HH 0.0I Hm
m.vl v.vI m.VI 0.mI v.m N.HI m.H N.m m.o: m.N 0.0H o.NH vd
0.0HI m.mI o.HHI v.0HI N.0HI h.ml m.hl m.0l m.0HI N.mI m.MI 0.v md
0.0HI 0.0MI m.0HI 0.0MI N.NI 0.0ml m.v m.©NI o.NI 0.0MI 0.m N.mNI Nd
o.NNI 0.0NI N.0NI v.HNI 0.HNI m.mHI m.mHI N.HHI h.mHI 0.0HI N.NI m.ml Hd
« m x s w a w a H ¥ H mCOHUHmom

 

 

 

ill

.tQMOEOmmEIIv uomflosm .UmScHuSOUII.U xHszmmd

110

 

.muoom ummumu

mmumoHtcH«

 

 

 

 

0.H 0.m m.HI 0.MI m.HI m.HI v.NI N.NI H.vI N.0I h.ml N.NI v0
m.vl m.m 0.NI m.m v.5l 0.m 0.mI 0.0 0.0HI N.v H.0I h.m mo
m.HH N.0HI 0.0 m.mHI 0.0 H.NHI N.N w.MHI m.m m.mHI h.m m.vHI N0
m.m m.N 0.m w.mI m.m ¢.ml o.m m.0I N.v H.0I m.N 0.ml HQ
m.mHI 0.HHI 0.0NI H.0HI H.NHI h.vHI H.NHI H.¢HI H.NNI H.NHI 0.HNI 5.0HI vU
m.0HI v.HHI m.VHI m.mHI 0.VHI m.mHI 0.MHI m.¢HI 0.0HI 0.0HI o.NHI o.NHI mU
m.m 0.mHI N.MI 0.HNI .MI 0.NNI o.HI H.0NI 0.NI 0.MNI 0.0I N.MNI NU
m.w 0.NI o.HI N.HHI 0.0I 0.0HI H.N m.h H.mI v.MHI 0.mI N.NHI HU
0.0NI m.MHI m.hHI N.NHI m.0HI o.HHI o.NHI b.0HI 0.0NI h.mHI 0.0NI m.ml vm
o.vHI 0.0HI o.HHI 0.¢HI m.0HI 0.MHI n.0HI 0.0HI h.MHI H.mHI h.VHI v.0HI mm
0.mHI 0.NHI 0.0HI o.NHI 0.0HI o.NHI m.0HI 0.0HI 0.HNI N.0NI m.NNI N.0HI Nm
m.m m.v m.N H.N m.N N.N N.N w.m m.vI v.0I 0.NI v.vl Hm
m.NI 0.0H 0.NI 0.H 0.NI m.N m.HHI o.v o.NHI m.0I N.NHI 0.NI wd
o.N v.0 m.NI 0.H 0.MI o.v m.0I o.v h.vI 0.NI m.mI v.5I md
0.0 N.MI H.MI 0.0HI 0.MI N.0HI 0.HI 0.mI m.vI m.NHI H.mI m.MHI Nd
m.N 0.0HI H.NI 0.NHI m.HI N.HHI o.HI N.0HI 0.mI o.NHI N.vI m.hHI Hd
« D a a O a a < «
H.vI m.mI 0.VI 0.0I 0.VI m.hI 0.HI N.MI 0.0I 0.m 0.0 o.m v0
o.HHI 0.m m.HHI m.m o.HHI m.v 0.NI 0.0 0.NI 0.0H 0.NI m.mH mo
0.v h.vHI 0.m m.mHI o.m 0.0HI m.m «.mHI m.m N.¢HI m.0H 0.mI ND
m.v 0.mI m.o m. I 0.h 5.0 m.m ©.m m.m N.m m.hH N.0H HQ
0.HNI o.NHI 0.NNI m.mHI N.0NI «.mHI. o.mHI m.vHI 0.0HI o.NHI m.0HI m.ml NU
N.0HI 0.0HI H.NHI N.NHI m.mHI o.mHI m.NHI N.0HI 0.0HI ¢.VHI m.mI H.NI mU
0.NI N.MNI m.mI 0.HNI H.mI 0.NNI 0.0 N.0HI 0.H 0.mHI 0.0H N.NI NU
0.mI N.vHI m.mI 0.VHI H.MI 0.MHI m.N N.NI 0.v N.NI o.HH 0.H HU
0.0HI H.NHI N.0NI N.MHI 0.0NI w.MHI N.NHI m.NHI 0.NHI v.HHI 0.vHI N.NI vm
m.MHI m.mHI m.MHI o.NHI 0.MHI m.mHI 0.mI m.MHI m.ml N.NHI 0.NI o.HHI mm
0.HNI N.NHI 0.HNI m.mHI N.HNI H.mHI m.0HI o.NHI 0.0HI N.mHI m.mI N.HHI Nm
h.ml 0.VI m.mI h.MI h.MI N.HI m.H N.H m.N N.m 0.5 m.w Hm
w.MHI N.NI m.MHI N.NI N.NHI o.HI 0.mI m.N m.©l m.m 0.0 m.m vd
h.mI N.0I H.mI m.mI N.mI o.vl 0.0 m.H 0.H v.0 0.m m.¢ md
m.0I N.MHI 0.0I m.vHI m.ml m.MHI H.NI N.NI 0.0 ©.ml 0.5 o.HI Nd
m.mI o.NHI 0.0I N.NHI H.0I m.mHI 0.HI m.HHI 0.H m.0HI m.m v.hl Hd
« m i r a m « H x H mcoHuHmOm
mHm30>

.SQHOEoochIm uomnodm

.COSCHHCOUII.U xHszmmd

 

 

 

111.

 

 

.,LI 1.. r , -
.QHOUW UWQUTH mmUMUHUCHk
m.m m.el 5.H 0.0I m.v 5.NI «.m N.HHI m.N m.HHI H.m w.mI v0
0.H 5.0! 0.vI m.5l m.HI 0.5I 0.HI 0.0HI v.MI 0.NI 5.NI 5.MHI mo
0.0 m.mI 0.HI 5.VHI 5.HI N.vHI v.vl «.mHI 0.0 0.vHI v.0I 0.MHI NO
0.0H m.¢H 0.HH m.NH m.NH m.MH H.0 m.5H 0.0H m.MH H.HH 5.HI HQ
0.0I 0.NI 0.mHI 0.0HI 0.mHI v.0I 0.5I N.NI 0.0HI m.HHI H.NHI 0.0HI vU
0.0 0.HHI 0.H 0.mHI 0.N N.NHI 5.m N.mHI 0.0 m.mHI m.0I m.MHI mU
0.v N.NI m.ml 0.0HI N.vl 0.mHI m.0 5.NHI H.0I 0.NNI H.0HI ¢.VNI NU
0.m m.N 0.MI N.0I 0.H! 0.0HI 5.v m.vI 0.0I m.MHI 0.NI 5.mHI HU
0.vI 0.0HI m.5I 0.5HI 0.ml 0.0HI 0.vI 0.mHI m.5I 0.5HI H.ml 0.5HI vm
0.0HI 0.H 0.HNI 5.HI 0.0HI 0.NI H.0NI 0.0 m.HNI N.0I m.mHI 0.5I mm
v.NI 5.mI 0.mI 5.mHI 0.NI m.mHI 0.NI N.MHI 0.0I 0.5HI 0.0I 0.5HI Nm
0.5 m.N 0.0 N.HI m.H 0.0I m.m m.0I m.VI m.vI w.vl v.NI Hm
0.5 0.0 0.m 0.m m.m N.H m.m m.m m.v o.HI 0.m N.vl vd
0.HI 5.H 0.mI 0.VI 0.NI 0.HI 5.HI N.H w.vI o.HI m.mI 0.0 md
0.NI H.NNI 0.5I H.0NI 0.vI 0.mNI 0.NI ¢.vNI 0.mI 0.0NI 0.5I 5.mNI Nd
0.0 0.vl 0.¢ m.¢l 0.NI N.vI m.mI 0.NI m.5I 0.NI H.0I 5.5I Hd
« s a a O a a < «
m.m m.ml m.N 0.NHI 0.H 0.NHI m.0I 0.0HI H.H N.0HI m.m 0.NI v0
m.mI H.NHI m.5I 0.0HI H.vI 0.mHI 0.NI 0.NHI o.HI 0.0HI o.N v.NI mo
H.NHI 0.5HI N.NHI 0.mHI 0.MHI 0.mHI 0.5I o.HHI N.mI m.5I m.H 0.m ND
0.0H 0.0 m.m . 0.5H H.0H 5.NH H.MH m.MH 0.MH m.mH 0.5 m.mH HQ
w.mHI m.HHI 0.0HI v.HHI 0.NI m. I 0.5I m.mI 5.0I N. I m.H m.N vU
0.0 0.MHI 5.H m.HHI 0.v m.HHI 0.v 0.NI o.m m.5I o.NH 0.HI mU
0.0HI 5.MNI m.ml v.vNI m.vI N.0HI 0.HI 5.0HI 0.0 N.mHI 5.0 N.m NU
H.0I m.5HI m.NI 5.NHI 0.H 0.0HI o.v 5.NI 0.m 0.0 o.NH m.MH HU
0.0HI 5.0HI 0.0HI m.5HI 0.NI 0.0HI 0.NI 0.mHI v.mI v.NHI H.MI 0.mI vm
H.0HI v.5l 0.5HI v.0I H.0HI 0.ml 0.5HI 0.NI m.5HI 5.HI 0.mHI 0.m mm
m.5I N.0HI 0.0HI m.mHI 0.ml 0.5HI m.5I N.mHI 0.mI 5.5I N.NI m.NI Nm
m.ml 0.MI H.5I N.vl 0.NI 0.NI N.0I v.NI m.N 0.0I 0.0 m.HI Hm
0.H m.vl m.0 m.MI m.H m.0 m.m 0.H m.m m.N m.m m.m vd
H.0I N.HI 0.5I 5.VI m.VI 0.0 H.HI m.N N.H m.m m.N 0.0 md
m.5l H.5NI 0.NI N.0NI v.0I m.wNI m.mI m.VNI m.mI 0.0NI o.HI H.HNI Nd
m.0HI N.mI m.5I 0.NI 0.0I N.5I m.0I m.mI 0.m N.vl m.0I H.0 Hd
« m a a a O « H a H mCOHUHmOm
mHm30>

 

.EQHOEOOSMIIG ummnnsm .UmocHuGOOII.o xHQZMde

APPENDIX H
MEAN RELATIVE SOUND PRESSURE SCORES

FOR EACH BODY TYPE

112

 

113

 

 

 

 

0.NI m.NI N.MI N.mI H.vI 5.0I 5.NI N.0I 5.0 m.HI H.vI N.N v0
H.0 m.vI N.0I m.mI 5.NI ¢.HI N.NI 0.5I N.0I m.mI m.mI m.N mo
m.N 5.5I m.5I N.NI m.5I N.mI 0.5I 5.MHI HWOHI H.0HI m.mI m.NI ND
m.¢ m.mI 0.0HI 5.NI m.mI m.mI m.5I m.VI m.vI 0.NI m.HI m.m HQ
m.m o.vI H.HI m.H 5.mI o.MI 5.mI m.MI H.vI 5.N N.mI m.m so
H.0HI m.MHI H.mHI m.HHI m.5HI m.mHI m.0NI v.0NI v.0HI H.MHI m.HHI m.ml mU
v.m m.mI H.5I m.MI m.HHI o.mI v.HHI m.mI H.0I H.HI 5.0I 0.0 NU
m.NI H.5I m.mI m.NI m.0HI m.5I 5.0HI H.0I 5.0I 5.0I 0.0 m.m HU
H.mI N.mI H.0I m.mI w.mI m.5I H.0HI m.HHI H.0HI m.mI m.mI N.NI vm
N.mHI 0.NHI 0.NHI N.mHI m.mHI o.mHI 5.mHI H.HNI 5.mHI m.mHI 5.NHI 0.5HI mm
m.mHI v.HNI H.mHI 0.0HI N.NNI m.NNI N.MNI «.mNI 0.NNI N.HNI m.0NI m.mHI Nm
v.HI m.mI «.ml 5.0I 5.mHI N.NHI m.MHI m.mHI 0.0HI m.mI m.oI m.H Hm
m.N m.HI 5.0I v.H m.VI o.MI v.mI N.mI o.m 5.0I m.o ¢.v vd
v.m o.HI m.OI m.H m.5I m.mI m.5I m.5I H.0I m.OI 5.H 0.NI md
N.NHI m.¢HI N.¢HI m.vHI 0.0HI m.vHI H.5HI N.0HI m.mHI v.mHI 5.¢HI m.HHI Nd
v.0HI m.mHI 5.0HI m.¢HI v.mHI H.5HI m.mHI N.HNI 5.NHI 0.0NI m.mHI m.HHI Hd
5 o < m m m H H meHuHmom
mHmBO>
.msmHOEouom MOM mmuomm musmmmHm chom m>HumHmH CmmZII

. m xHQmemd

 

 

114

 

.IIlmIIIIIIIr
H.mu m.~HI H.mu H.4I H.NI 4.4: N.H- H.mn H.H m.HI H.m ea
m.mn H.ou m.mu m.HHI H.HI m.mu o.HHI m.oa- H.mu H.mu m.~I ma
m.eHI m.~HI ~.mHI H.mHI o.mHI m.mHI H.NHI N.0HI m.m: m.mn s.mI . an
~.mu H.mu m.oI s.mHI o.HHI o.oHu «.mHI m.o~u «.mHI e.mHI e.mu Ha
m.-u H. mm: O. can 0. mm: o.NNI m.mmu m.-u H. mm- H. Hm- N.NNI H. SH- so
m.HHI H. OH: m. m: m. «HI m.~HI m.mHu H.4HI m. «H: m. an m.mu m. on no
v.HHI m.mu o.N: o. NHI H.HHI ¢.HHI «.mu m.mn N.H- m.mu 5.4 No
m.mI m.HI e.ou m.mu m.mu H.HI H.HI m.mn H.H m.H e.HH Ho
~.om- N.HNI e.omn m.H~I N.NNI N.NNI m.H~I m.-u H.HNI H.omn. m.mHI em
H.mHI H.mHI m.mHu m. mHI m.mHI m.mHI v.o~n o.o~n m.HHI ~.eHI o.mHI mm
v.GHI ¢.mHI .m.mHI m. was ~.mHI «.mHI H.mHI m.HHI «.mHI m.oHI o.mHI mm
m.mu m.~HI m.eu H.mu H.mu I m.OHI «.mu m.nu H.ou v.~- H.o I Hm
m.ou m.H H.H m.eu m.mu e.mI m.mu m.o- m.e m.e c.0H we
H.GI m.eu H.HHI m.mn «.mu H.NHI m.HHI m.mI o.mn m.mu H.N m4
H.HNI m.o~n e.H~I o.m~u «.mmu m.m~- m.m~I o.NNI m.mHu m.o~I H.HHI ma
m.HHu e.oHI G.GHI 0.HNI v.HHI H.HNI m.omu o.mHI o.NH: m.NHI m.eu H4
0 < M w m H H mGOHHHmOQ

 

mHm30>

 

.mamHoEOmmz .omSGHudoolI.m xHazmmmd

 

 

115

 

m.¢l m.NI v.5l m.NI m.hl v.51 m.ml o.ml v.9: m.ml m.o «a
m.ol mwo m.ol m.NI o.wl H.vl h.ml H.ml h.NI v.ol m.w mo
H.mHI m.mal m.val b.ml m.mal m.mH| m.mHI m.wHI N.MHI m.OHI m.HI .ND
v.m H.v m.m h.m m.MI m.H N.N h.m h.m m.oa H.mH HQ
H.mHI H.NHI H.HHI m.val m.mHI m.¢al m.¢HI m.HHI m.ml m.ml m.HI v0
m.mHl m.mHI m.mal m.mal N.mHI m.mal m.mHI b.¢HI m.NHI m.oan m.¢l m0
m.mHI m.mHI v.mHI m.mml m.mml v.mml H.m~| H.0NI m.MHI H.mal 5.HI NU
m.ml c.0HI o.ml m.mal v.val m.mHI N.NI m.ml N.NI N.h 0.H H0
m.¢HI N.vHI m.NHI m.mHI v.MHI w.vHI o.mHI N.mHI r.mHL m.HHI m.ml vm
m.hl o.ml «.ml m.ml 5.HHI m.HHI N.mal ¢.oal m.hl N.mal o.MI mm
m.mHI N.mHI m.val m.mHI m.mHI h.mal m.mHI m.mHI H.mHI m.HHI 0.5! mm
v.0 m.o 0.H m.m| v.mt . m.MI m.MI H.NI m.on N.H o.N Hm
m.N H.N m.m N.H: v.MI N.MI N.NI N.0| H.N H.m m.v 04
m.HI m.H m.N m.NI o.vl m.m1 H.ml m.NI o.N m.N m.w Md
m.mHI m.bal n.mal N.¢HI m.mal H.0NI N.HNI 0.0NI H.5HI m.mal H.HHI Nd
N.m| m.hl m.ml m.NHI m.NHI v.val m.NHI m.HHI H.0HI m.hl 5.0! H¢
O < m m m H H mCOHUHmOm

 

mam30>

 

.mnmuofiowcm .UmsaflucOUIl.m XHszmm<

APPENDIX I
MEAN RELATIVE SOUND PRESSURE SCORES

FOR ALL SUBJECTS

116

 

 

”L .—
‘un

 

117

 

h.ml v.mn H.5I m.ml m.vl m.¢| 5.5I o.ml m.~l H.¢I m.N qo
m.OI H.NI m.vl «.ml N.¢I H.ml v.ml v.hl h.ol m.ml N.N .mD
¢.NHI H.NHI o.NH! 5.NHI m.oa1 m.NHI m.mal N.HHI m.ml H.5I v.NI ND
h.¢l H.vl N.5I N.ml o.ml N.NI m.ml N.MI H.HI v.NI m.v HQ
o.NH! N.HHI o.oal m.mHI o.mHI m.HHI m.mal N.mHI m.ml «.m! c.vl v0
v.MHI o.NH: H.HHI m.mHI m.¢HI m.mHI 0.0HI m.¢HI m.HHI m.ml H.ml mu
m.HHI m.HHI NJmI m.mHI o.vHI m.mHI m.mHI v.NHI H.ml m.bl m.m NU
¢.ml o.ml m.NI m.HHI H.HHI m.HHI m.ml N.5I H.0 . ¢.0I ~.m HU
h.mHI m.mal v.mHI m.mHI N.vHI m.mHI N.wHI m.mal b.vHI m.mHI H.0HI «m
m.val m.vHI .m.NHI H.@HI h.mHI m.mHI N.mHI 5.0HI m.man m.HHI m.m| mm
o.m| N.NHI m.mHI m.mHI b.mal H.0NI h.oNI m.mHI m.bHI H.mHI H.NHI Nm
h.¢l N.v| m.NI v.ml H.ml m.ml v.ml m.hl m.mn m.NI h.m Hm
H.0: m.o v.m H.ml m.ml m.vl H.VI m.HI H.N v.m h.¢ ¢¢
o.HI N.H! m.ou o.wl m.m| m.hl H.ml m.ml m.0I H.OI m.m mfl
H.mHI v.mHI m.hal 0.0NI N.mal H.HNI m.oml h.mal H.NHI m.mHI N.MHI Nfi
v.vHI h.mal o.mHI o.NHI m.mHI m.hal m.mHI H.mHI m.vHI H.NHI h.ml ad
0 < m w m H M mcofluwmom

 

mHm3o>

 

.muomnndm Ham How mmuoum musmmmum wcdom m>HumHmH :mmzun.H xHozmmma