An Introduction To Speech Sciences (Acoustic Analysis Of Speech)

Abstract
peech sciences deal with the
acoustical characteristics of speech
by means of sophisticated soft wares as
well as hard wares. Although, a speech
science is a well known science in the
developed countries, especially the west-
ern societies, however, it has been
remained almost unknown in Iran, though,
in recent years a group of scholars have
been involved in this branch of science.
The application of the speech sciences in
the field of 'Speech and Language
Therapy' has been widely accepted
throughout the world. The main focus of
these applications is the therapeutic
aspects of the program, which is used for
both adults and children. However, the
non-therapeutic applications are mainly
used for research purposes only.
In this article, a brief introduction about the
nature as well as the application of speech
analysis for both normal and abnormal
speeches is given. In addition, examples
of the acoustical characteristics
fconsonants and vowels such as, voice
onset time (VOT), formant frequencies
and spectrograms will be presented.
Key words:
Acoustic analysis, VOT, Spectrogram,
Formant frequency, Disordered speech
Introduction
Speech science is the study of all the fac-
tors involved in producing, transmitting,
perceiving and comprehending speech,
including all relevant aspects of anatomy,
physiology, neurology and acoustics, as
well as phonetics. Speech analysis began
in 1940s (i.e., during the Second World
War) in the United States of America. The
study of speech production from an
acoustical point of view provides the
means for looking at a very complex
process in a simpler way. Speech science
is the study of all the factors involved in
producing, transmitting, perceiving and
comprehending speech, including all
An Introduction to Speech Sciences
(Acoustic Analysis of Speech)
Rezaei Nasser , BSc, P.G.D., MA, PhD
Assistant Professor; University of Social Welfare and Rehabilitation Sciences,
Salehi Abolfazl , BSc , MSc
Lecturer , University of Social Welfare and Rehabilitation Sciences, Tehran.Iran
vol.4 - No .4 5
S

relevant aspects of anatomy, physiology,
neurology and acoustics, as well as pho-
netics.
Acoustics is the branch of physics that
deals with sound, and Psychoacoustics
is the study of the psychological
response to sound; it is a division of psy-
chophysics, or the general study of psy-
chological responses to physical stimuli.
The study of speech acoustics has both
a physical and psychological side. The
physical side pertains to the physical
structure of the sounds of speech. The
psychophysical side is concerned with
the perception of these sounds. A proper
understanding of speech requires knowl-
edge of both of these aspects of speech
acoustics. A well known riddle asks, "If a
tree falls in the forest, but there is no
one to hear it, does it make a sound?"
Of course, the answer depends on the
definition of terms. If sound is defined
with respect to human perception, then
no sound could be verified. But if sound
is defined as a physical disturbance in
the air, then sound must have occurred.
A riddle more regarding speech might
be: "If speech is made visible as pat-
terns on paper (or a video monitor), but
no one hears it, is it really speech?"
Sound is vibration. Vibration is a repeti-
tive to-and-fro motion of a body. Usually
we do not directly hear the vibration in a
sound source such as
an engine, but rather we hear the vibrations
that are broadcast, or transmitted, in a medi-
um like air. Figure B-1 shows a simple physi-
cal arrangement to demonstrate the nature
of sound. The sound source is a stretched
rubber band that can be pulled to set it into
vibration. The band undergoes a series of to-
and-fro vibration after it is pulled. The initial
vibrations are of large amplitude, meaning
that the to-and-fro swings have a relatively
great motion. The amplitude diminishes until
motion eventually stops altogether. The
reduction in amplitude reflects damping, or
the loss of energy.
Oscillograms:
The first major advance in the acoustic
analysis of speech began with Oscillograms
(waveforms, or graphs of amplitude over
time) which was an important advance. In
this method, the sounds selected for analy-
sis were often vowels, because they are rel-
atively easier to analyse than most conso-
nants.
The Henrici Analyser:
Is one of the earliest tools for spectral
analysis. It is a mechanical device consisting
of five rolling integrating Units (glass
spheres). They could obtain Oscillograms,
tracing on a surface,
6
Iranian Rehabilitation Journal

calculating the values of amplitude, and
pressure.
Filtering:
A filter is a frequency-selective trans-
mission system. That is, a filter will pass
energy at certain frequencies but not at
others. A filter allows a selective look at
the energy in various frequency regions.
The Spectrograph:
Provided major advantages to the
study of speech and was developed in
1940s. It was fast, provided a better
description of the energy concentration in
speech, and finally, it produced a running
short-term spectrum, enabling the scien-
tists to visualise how energy concentra-
tions changed in time. The display of the
running short-term spectrum is called
spectrogram.
Digital Signal Processing of Speech:
The dominance of the spectrograph
was seriously challenged only with the
introduction of digital computers. The
challenge has intensified with the contin-
ued refinement of computers (hardware)
and analysis programs (software). Some
developments in the use of digital meth-
ods for speech analysis are: waveform
display; waveform editing;
playback of original or edited signals; F0;
Jitter or Shimmer analysis; spectrogram;
FFT' LPC; and speech synthesis.
The Speech Chain:
Understanding of subject of spoken com-
munication requires knowledge of many dis-
ciplines. The Speech Chain provides an
easy-to-understand introduction to these
topics and explains how events within these
domains interact to bring about communica-
tion by speech. A convenient way of exami-
nation what happens during speech is to
take the simple situation of two people talk-
ing to each other. For example, you as a
speaker want to transmit information to
another person, the listener. The first think
you have to do is to arrange your thoughts,
decide what you want to say and then put
what you want to say into linguistic form.
The message is put into linguistic form by
selecting the right words and phrases to
express its message, and by placing these
words in the order required by the grammat-
ical rules of the language. This process is
associated with activity in the speaker's
brain, and it is from the brain that appropri-
ate instructions, in the form of impulses
along the motor nerves, are sent to the mus-
cles that activate the vocal organs (the
lungs, the vocal cords, the tongue, and the
lips). The nerve impulses
vol.4 - No .4 7

set the vocal muscles into movement,
which in turn, produce precise pressure
change in the surrounding air. We call
this pressure change a sound wave.
Sound waves are often called acoustic
waves, because acoustics is the branch
of physics concerned with sound.
The movement of the vocal organs gen-
erate a speech sound wave that travels
the air between speaker and listener.
Pressure changes at the ear activate
the listener's hearing mechanism and
produce nerve impulses that
travel along the acoustic nerve to the listen-
er's brain. In the listener's brain, a consider-
able amount of activity is already taking
place, and this activity is modified by the
nerve impulses arriving from the ear. This
modification of brain activity, in ways that are
yet fully understood, brings about recognition
of the speaker's message. We see, there-
fore, that speech communication consists of
a chain of events linking the speaker's brain
with the listener's brain. We shall call this
chain of events the speech chain.
Acoustic Analysis of Speech
Acoustic analysis of speech is the
study of the acoustical characteristics
of speech, both normal and abnormal
speech. It involves the physical aspects
of spoken language.
These include; waveform analysis, FFT or
LPC analysis, voice onset time (VOT) meas-
urements, formant frequency measurements,
and so on.
8

Some of these properties will be reviewed
briefly.
Voice Onset Time (VOT):
Voice Onset Time (VOT) for instance, is the
duration of the period of time between the
release of a plosive and the beginning of
vocal fold vibration. This period is usually
measured in milliseconds (ms), which is a
microscopic analysis. It is useful to distin-
guish at least three types of
VOT which are shown in the schematic
diagram below:
-Zero VOT: where the onset of vocal fold
vibration coincides (approximately) with
the plosive release.
-Positive VOT: where there is a delay in
the onset of vocal fold vibration after the
plosive release.
-Negative VOT: where the onset of vocal
fold vibration precedes the plosive
release.
Formant frequency
Formant is the vocal tract resonance.
Theoretically, there is an infinite number of
formants, but for practical purposes only the
lowest three or four are of interest.
Formants are identified by formant numbers
(e.g., F1, F2, F3, and F4, numbered in suc-
cession beginning
with the lowest frequency formants). Each
formant can be described by two charac-
teristics: centre frequency (commonly
called formant frequency) and bandwidth
(formant bandwidth, which is a measure
of the width of energy in the frequency
domain.
vol 4 - No .4 9

Acoustical Characteristics of Vowels
Producing different vowels is like altering
the size and the shape of the vocal tract,
which is like a tube. The air in this tube is
set in vibration by the pulse of air from
the vocal folds. Every time they open and
close, the air in the vocal tract above
them will be set in vibration. Because,
the vocal tract has a complex shape, the
air within it will vibrate in more than one
way. Usually we can consider the body of
air behind the raised tongue (i.e., in the
throat) to vibrate in one way, and the
body of air in front of the tongue (i.e., in
the mouth) to vibrate in another way. In
the vowel /i:/ as in
deed, the air behind the tongue will vibrate
at about 250 Hz, and the air in front of it at
about 2100 Hz.
The relationship between the tongue posi-
tion and formant frequencies
The frequency of F1 is inversely related to
tongue height (e.g., high vowels have a low
F1 frequency), and the frequency of F2 is
related to tongue advancement (e.g., F2
frequency increases as the tongue position
moves forward in the mouth).
F1: High vowels such as /i/ and /u/ have a
relatively low frequency of the first formant
(F1), whereas the low vowels /a/ and /ae/
have a relatively high frequency of this for-
mant, that is, the frequency of
10

F1 varies inversely with tongue height of
the vowel.
F2: On the other hand, back vowels /u/
and /a/ share a low frequency of F2,
whereas the front vowels /i/ and /ae/ have
a high F2, that is, the frequency of the
second formant varies with the posterior-
anterior dimension of vowel articulation.
This result points to an articulatory-acoustic
correspondence: the frequencies of the first
two formants, F1 and F2 can be related to
dimensions of vowel articulation.
Figure 4: Illustration of the tongue position for the production of the
vowels /i:/ and /u/.
Figure 5: The relationship between the tongue position and formant
frequencies for different vowels.
vol. 4 - No .4 11

The Acoustic Characteristics of
Disordered Speech
The Characteristics of disordered speech
are somewhat different from those of the
normal speech. These include:
Howell and Vause (1986), in an attempt to
investigate the acoustic properties of stut-
tered speech, analysed the spectral prop-
erties of 8 English vowels in adult stutter-
ers in comparison with the subsequent
fluent vowels. They intended to examine
some of the previous researchers findings
such as Van Riper (1971) who had stated
that, the stutterer produces the schwa
vowel at a point where some other vowel
is required. In this experiment, words con-
sisting of a voiceless consonants in a
CVC syllable were investigated, where the
final consonant in the fluent word includes
nasals, fricatives (voiced and voiceless),
a)Vowels & Diphthongs
b)Sonorants: nasals, glides, liquids
c)Obstruents: e.g., stops, fricatives
Dysfluent vs. Fluent Vowels and
voiceless plosives
The acoustic properties and formant fre-
quencies of stutterers for fluently and
dysfluently produced vowels, as well as
the duration and amplitude of vowels
were examined. The results revealed
that dysfluent vowels were shorter than
the fluently produced vowels. Other
acoustic data showed that the brief vow-
els in repetitions had formant locations
at the same frequencies as in the fluent
speech of the same speaker. Stuttered
vowels were low in amplitude and short
in duration, which caused stuttered vow-
els to be perceived as schwa.
12

Figure 6: Idealised spectrogram of a young male who stutters producing a
sound/syllable repetition (SSR) on the word "but" with one iteration of "bu-" in
the stuttered or repeated portion of the SSR ("bu-but"). The stuttered portion is
immediately followed by the fluent portion used for stuttered versus fluent com-
parisons (Yaruss and Conture, 1993, p. 888).
Figure 7: Spectrograms for the sentence "mama made apple jam" produced
by a a speaker with velopharyngeal incompetence (hepernality) at the top (A)
and by a normal speaker at the bottom (B).
vol.4 - No .4 13

References:
Ball, M. J. and Code, C. (1997).
Instrumental Clinical Phonetics. England,
Whurr Publishers.
Denes, P. B., and Pinson, E. N. (1993)
The Speech Chain: The Physics and
Biology of Spoken Language. New York,
W. H. Freeman and Company.
Fry, D. B. (1982). The Physics of
Speech. Cambridge: Cambridge
University Press.
Howell, P., and Vause, L. (1986).
Acoustic Analysis and Perception of
Vowels in Stuttered Speech. Journal of
the Acoustic Society of America, 79 (5),
1571-1579.
Kent, R. D., and Read, C. (1992). The
Acoustic Analysis of Speech. USA,
Singular Publishing Group, Inc.
Rezaei-Aghbash, N. (2003). Acoustic
Characteristics of Dysfluent Speech: A
Study of Persian and English stutterers.
An unpublished PhD dissertation. UK.
Van Riper, C. (1971). The Nature of
Stuttering. 2nd edition. NJ (USA),
Prentice-Hall, Englewood Cliffs.
Yaruss, J. S., and Conture, E.G. (1993).
F2 Transition During Sound/Syllable
Repetitions of Children Who Stutter.
14

An Introduction To Speech Sciences (Acoustic Analysis Of Speech)

More Related Content

Similar to An Introduction To Speech Sciences (Acoustic Analysis Of Speech) (20)

More from Jeff Nelson (20)

Recently uploaded (20)

An Introduction To Speech Sciences (Acoustic Analysis Of Speech)