U.S. patent number 3,914,550 [Application Number 05/388,029] was granted by the patent office on 1975-10-21 for artificial speech device.
This patent grant is currently assigned to Cardwell Associates, Inc.. Invention is credited to Gilbert I. Cardwell, Jr..
United States Patent |
3,914,550 |
Cardwell, Jr. |
October 21, 1975 |
Artificial speech device
Abstract
Disclosed is a device which generates sound in selected
waveforms and at selected frequencies via a transducer located in
the mouth for conversion into intelligible speech by a person who
has lost the use of the larynx or vocal cords. The device impulses
the diaphragm of the transducer with pulses of generally
rectangular waveform to provide the generated sound with a waveform
comprising a series of damped sinusoids with harmonics. Such
waveforms enable use of the device with practically any individual
regardless of the resonant and other acoustical characteristics of
the individual's mouth, and permits the transducer and an included
tube for directing the sound to be freely placed in and disposed at
a variety of different locations throughout the mouth. The
frequency of the generated sound may be varied by varying the pulse
repetition rate, while the amplitude of the sound can be varied by
changing the pulse width.
Inventors: |
Cardwell, Jr.; Gilbert I.
(Palos Verdes Peninsula, CA) |
Assignee: |
Cardwell Associates, Inc.
(Palos Verdes Peninsula, CA)
|
Family
ID: |
23532334 |
Appl.
No.: |
05/388,029 |
Filed: |
August 13, 1973 |
Current U.S.
Class: |
381/70;
623/9 |
Current CPC
Class: |
A61F
2/20 (20130101) |
Current International
Class: |
A61F
2/20 (20060101); A61F 001/20 (); G10L 001/10 () |
Field of
Search: |
;179/1AL |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Claffy; Kathleen H.
Assistant Examiner: Kemeny; E. S.
Attorney, Agent or Firm: Fraser and Bogucki
Claims
What is claimed is:
1. An artificial speech device comprising:
oscillator means for generating a succession of generally equally
spaced timing signals;
monostable means responsive to the timing signals for assuming an
unstable state for a selected period of time in response to each of
the timing signals and for thereafter assuming a stable state;
an electro-acoustic transducer;
means coupled to energize the transducer whenever the monostable
means is in the unstable state; and
a flyback damping circuit coupled between the monostable means and
the transducer and comprising the serial combination of diode means
and resistor means.
2. An artificial speech device comprising:
an electro-acoustic transducer adapted to introduce acoustic waves
into the mouth of a user, the transducer having a pair of input
terminals;
pulse generating means for repetitively generating pulses of
generally rectangular waveform at a pair of output terminals
thereof,
means coupling one of the output terminals of the pulse generating
means to one of the input terminals of the transducer; and
a flyback damping circuit coupled between the other one of the
output terminals of the pulse generating means and the other one of
the input terminals of the transducer, the flyback damping circuit
comprising the serial combination of a diode and a resistor.
Description
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to devices for generating sound for
conversion into speech by persons otherwise unable to speak
intelligibly due to the loss of the use of the larynx or vocal
cords, and more particularly to devices of this type which
introduce the generated sound directly into the mouth of the
user.
2. History of the Prior Art
An increasing number of persons have lost the ability to speak
intelligibly or to speak at all due to impairment of the larynx or
to the vocal cords contained therein. During normal speech, the
vocal cords vibrate in response to the expulsion of air from the
lungs to create sound which is converted by the mouth into
intelligible speech. Where use of the vocal cords is lost such as
by reason of surgical removal of the larynx, the person losses his
ability to speak. With considerable training and practice, the
person so handicapped can sometimes develop a technique in which
air is expelled from the stomach in such fashion as to create a
sound for conversion by the mouth to speech. Such technique,
however, has serious limitations making the use of artificial
speech devices generally preferable.
There are a variety of artificial speech devices designed to solve
the speech problem noted above. Most of these devices generate
sound which is introduced into the mouth to replace the sound
normally provided by the vocal cords. U.S. Pat. No. 3,291,912 of
Flanagan is an example of one such device in which the sound is
generated outside the mouth and is thereafter transmitted through
the throat and into the mouth. The Flanagan arrangement transmits
the sound through the throat by exciting a diaphragm adjacent the
throat wall with a train of pulses. A basic timing device is used
to determine the basic frequency of the generated sound, and a
plurality of pulses are then generated during each period so
defined and applied to excite the diaphragm. U.S. Pat. Nos.
2,862,209 of Cooper and 3,084,221 of Cooper et al illustrate
arrangements in which the sound is introduced directly into the
mouth. The cooper patent discloses a denture which acts as a horn
to direct sound from a transducer to the back of the mouth. The
Cooper et al patent shows a bite block which may be selectively
compressed between the teeth to control the introduction of sound
into the mouth. Still other devices are shown by U.S. Pat. Nos.
2,093,453 of Kellotat and 3,508,000 of Snyder. The Kellotat patent
discloses a denture which responds to exhaled air to produce sound
within the mouth. The Snyder patent discloses an arrangement which
generates an electromagnetic signal inside the mouth for conversion
into sound outside of the mouth.
While arrangements which provide the generated sound outside of the
mouth have the advantage that no device is required to be placed
within the mouth, such arrangements have proven to be generally
unsatisfactory in most instances, principally because of the
difficulty of producing sound within the mouth of sufficient
amplitude to be readily intelligible. Moreover the generated sound
is typically of such a nature that it does not resonate with or
otherwise compliment the acoustical characteristics of the user's
mouth, resulting in further reduction in the intelligibility of the
speech.
Arrangements which locate the transducer within the mouth or
otherwise function to introduce the generated sound directly into
the mouth have proven to be generally superior, principally because
the loss in volume is minimized in such situations. However, such
arrangements leave much to be desired from other standpoints. For
instance, most such arrangements use a signal of sinusoidal or
similar waveform to excite the transducer. The resulting acoustical
energy typically has a waveform which is itself sinusoidal and of
singular frequency and amplitude. The singular frequency of the
sound waveform typically requires that the transducer assembly be
placed at a fixed, critical location within the mouth in an attempt
to match resonant characteristics of the mouth to the generated
sound. The dentures and similar devices used to hold the transducer
assembly at a fixed location within the mouth as so required,
result in arrangements which are very uncomfortable and cumbersome
for the user. Moreover even in the relatively uncommon situation
where the singular frequency of the generated sound is closely
matched to the resonant characteristics of the mouth, the resulting
speech leaves something to be desired in terms of intelligibility,
among other reasons because the generally constant amplitude
characteristic of the generated sound is unlike that of normal
speech.
Accordingly, it would be desirable to provide an artificial speech
device for generating sound which is easily converted into highly
intelligible speech by virtually any user, regardless of acoustical
characteristics of the user's mouth or of the particular location
within the mouth at which the sound is introduced. Such an
arrangement in which a sound capable of resonating at any of many
different frequencies is produced would not only produce more
intelligible speech but would enable the transducer assembly to be
loosely placed in and readily moved about the inside of the user's
mouth, thereby freeing the user from annoying dentures and similar
arrangements.
BRIEF DESCRIPTION OF THE INVENTION
Artificial speech devices in accordance with the invention generate
pulses which impulse the diaphragm of the transducer to produce an
acoustical waveform which is essentially a damped sinusoid and
which therefor closely approximates the waveforms of normal speech.
The damped sinusoids include numerous harmonics, enabling them to
resonate in virtually any mouth independent of the acoustical
characteristics of the mouth. The transducer arrangement can be
placed in practically any location within the mouth, so long as the
sound produced thereby is directed over the tongue and into the
speech forming area of the mouth. Accordingly, the transducer
assembly consisting of the transducer and a hollow tube connected
thereto for directing the sound to the back of the mouth may be
naturally positioned so as to be held loosely within the mouth.
The basic frequency of operation is determined by an oscillator
which generates a succession of generally equally spaced timing
signals to define the periods. A monostable circuit responds to the
occurrence of each timing signal to change state for a selected
period of time and provide a driving pulse of generally rectangular
waveform to the transducer. The generated pulse impulses the
transducer diaphragm, then terminates for the remainder of the
period, leaving the transducer diaphragm free to vibrate in
response to the impulsing thereof. The amplitude of the generated
sound and thus of the resulting speech can be controlled by varying
the width of the pulses. The pitch or frequency of the sound can be
varied by varying the rate of occurrence of the timing pulses.
BRIEF DESCRIPTION OF THE DRAWINGS
The foregoing and other objects, features and advantages of the
invention will be apparent from the following more particular
description of a preferred embodiment of the invention, as
illustrated in the accompanying drawings, in which:
FIG. 1 is a perspective view of an artificial speech device
according to the invention;
FIG. 2 is a plan view of the inside of the mouth of a user showing
a typical positioning of the sound unit of the artificial speech
device therein;
FIG. 3 is a basic block diagram of the electrical portion of the
artificial speech device of FIG. 1;
FIG. 4 is a schematic diagram of the arrangement of FIG. 3;
FIGS. 5A through 5P are waveforms useful in explaining the
operation of the artificial speech device in FIG. 1; and
FIGS. 6A and 6B are waveforms which are also useful in explaining
the operation of the artificial speech device of FIG. 1.
DETAILED DESCRIPTION
A preferred form of artificial speech device 10 in accordance with
the invention is illustrated in FIG. 1. The device 10 includes a
control unit 12 carried in a convenient location such as the user's
shirt pocket and a sound unit 14 which can be placed in the user's
mouth as shown in FIG. 2. The sound unit 14 is coupled to the
control unit 12 by a cord 16 comprising a pair of wires.
The control unit 12 comprises a generally rectangular housing 18
which contains the power source in the form of a longlife battery,
the electronics for generating signals applied to the sound unit 14
via the cord 16, and various controls for the device 10. The
controls appear on the front panel of the housing 18 and include a
volume control 20 for adjusting the volume of sound produced by the
sound unit 14, a toggle switch 22 for turning the device on to
generate the sound at the sound unit 14, a pushbutton 24 which may
be used in lieu of the toggle switch 22 and which produces sound at
the sound unit 14 when depressed, and a pitch control 26 which
varies the pitch of the sound produced by the sound unit 14 by
varying the base frequency thereof. A plug and jack arrangement 28
couples the cord 16 to the inside of the housing 18.
The sound unit 14 includes an electro-acoustic transducer 30 and a
sound tube 32. The transducer 30 responds to electrical signals
generated by the control unit 12 and carried by the cord 16, to
generate acoustical energy which is then carried by the sound tube
32. A coil 33 within the transducer 30 responds to the electrical
signals delivered by the cord 16 to cause movement of a diaphragm
35 and thereby generate acoustical energy. The transducer 30 may
comprise an electro-acoustic transducer of the type commonly used
in hearing aids. Alternatively, the transducer 30 may comprise
other types of transducers such as a piezoelectric transducer or a
magnetostrictive transducer. The sound tube 32 which comprises a
piece of hollow plastic tubing rests on top of the tongue as shown
in FIG. 2 and serves to direct the sound to the speech forming area
of the user's mouth.
In operation the control unit 12 is placed in a convenient location
on the user's person such as a front shirt pocket or an inside coat
pocket and the sound unit 14 is placed within the user's mouth 34
in a convenient and comfortable location and so that the sound tube
32 is disposed over or on top of the tongue 36. The cord 16 extends
between the lips 38 of the mouth opening and to the control unit
12. The sound generated by the transducer 30 is passed by the sound
tube 32 over the top of the tongue 36 and toward the region of the
throat. As is known in the art such sound is converted into speech
by manipulation of the mouth, tongue, throat and lips much in the
same way as normal speech is produced. The only difference is that
the sound which is converted into speech is provided by the sound
unit 14 rather than by the vocal cords in the present instance.
In accordance with a feature of the invention to be described in
detail hereafter, the artificial speech device 10 produces
acoustical energy which is readily converted into intelligible
speech regardless of the particular acoustical and resonant
characteristics of the mouth. This is done by producing a sound
waveform having a variety of harmonic components, one or more of
which are at the appropriate frequency so as to resonate and
favorably interact with the particular characteristics of virtually
any mouth. In accordance with a further feature of the invention,
the sound waveform produced by the artificial speech device 10 is
very similar to that of normal speech so as to make the resulting
speech more intelligible. A still further feature of the invention
results from the fact that the sound produced by the artificial
speech device 10 need not be critically adjusted to the particular
oral characteristics of the user. For this reason the sound unit 14
may be placed in virtually any convenient and comfortable position
within the mouth 34, so long as the sound tube 32 is positioned to
direct sound in such a fashion that it reaches the back of the
mouth. This provides for a "natural positioning" type of use in
which the sound unit 14 can be allowed to float or move freely
throughout the mouth so as to be used in a location most convenient
and effective for the individual. In addition to eliminating
dentures or other inconvenient or uncomfortable arrangements
required by those prior art systems in which the location and
direction of the sound is critical, the natural positioning
arrangement of the sound unit made possible by the present
invention makes the device much more pleasant to use in terms of
comfort within the mouth and reduction of nervousness and anxiety
which frequently occur when the user becomes conscious that he must
retain the sound unit continually within the mouth during prolonged
conversational periods and must maintain the sound unit in a
specific position or location if satisfactory results are to be
obtained. The physical configuration of the sound unit 14 aids in
the comfort and ease of use. Because the transducer 30 has a
substantially larger diameter than the tube 32, the transducer 30
tends to stay outside the teeth and tends to prevent the entire
sound unit from falling or slipping into the central area of the
mouth.
FIG. 3 comprises a basic block diagram of the electronic portion of
the artificial speech device 10. As shown in FIG. 3 the device 10
includes a power source 50 feeding an oscillator 52. The oscillator
52 which is made variable in frequency so as to be able to vary the
pitch of the generated sound produces a series of generally equally
spaced timing signals or pulses which define the base or nominal
operating frequency of the device 10. This frequency can be
selected so as to compliment the voice and other characteristics of
the user. For example, a base frequency on the order of 200 hertz
may typically be chosen in the case of a male user.
The timing pulses produced by the oscillator 52 are applied to a
power amplifier 54 which together with a portion of a pulse
generator and driver 56 comprises a monostable circuit 58. The
power amplifier 54 responds to each timing pulse from the
oscillator 52 to saturate a transistor and thereby switch the
monostable circuit 58 into its unstable state. With the monostable
circuit 58 in the unstable state, the driver portion of the pulse
generator and driver 56 provides a pulse of generally rectangular
waveform to the transducer 30. Generation of the pulse is continued
so long as the monostable circuit 58 remains in the unstable state
as determined by the time constant of an RC circuit within the
pulse generator and driver 56. The RC circuit maintains the
monostable circuit 58 within the first state for a selected period
of time, then switches the monostable circuit into its stable state
by turning off the transistor within the power amplifier 54 and at
the same time biasing a transistor within the pulse generator and
driver 56 into conduction. This terminates the pulse of rectangular
waveform provided to the transducer 30, and no further signal is
provided to the transducer during the remainder of the period
defined by the timing pulse from the oscillator 52. Upon the
occurrence of the next timing pulse from the oscillator 52 so as to
define the beginning of a new period, the process is repeated with
the monostable circuit 58 again changing state to provide a pulse
of generally rectangular waveform to the transducer 30. The time
constant defined by the RC circuit within the pulse generator and
driver 56 is made variable so as to vary the widths of the
rectangular pulses and thereby the amplitude of the sound produced
by the transducer 30.
Each of the rectangular drive pulses impulses the diaphragm of the
transducer 30 by driving the diaphragm in a selected direction and
by a selected amount and thereafter allowing the diaphragm to
undergo free vibratory motion in response to the impulsing thereof.
The resulting acoustical energy generated by the transducer 30
comprises a succession of damped sinusoidal waveforms with
harmonics.
A preferred arrangement of the artificial speech device 10 of FIG.
3 is schematically illustrated in FIG. 4. The power source 50
comprises a conventional 9 volt battery 60 coupled in parallel with
a capacitor 62 across the input terminals 64 and 66 of the
oscillator 52. Closure of either the toggle switch 22 or the
pushbutton 24 couples the power source 50 to energize the
oscillator 52 and feed the power amplifier 54. The oscillator 52
which is of conventional form responds by generating a plurality of
timing signals or pulses at a frequency determined by a variable
resistor 68 within the oscillator circuit. The timing pulses are
applied via a diode 70 to the base of a transistor 72 coupled to
control the conduction of a transistor 74.
The transistors 72 and 74 may be biased into conduction by the
action of a resistor 75 and a diode 76 when a transistor 77 is not
conducting to switch the monostable circuit 58 into its unstable
state. The monostable circuit is normally in its stable state in
which the transistors 72 and 74 are not conducting and the
transistor 77 is biased into conduction by a capacitor 78, a
variable resistor 80, and a fixed resistor 82. A resistor 84
coupled in circuit with the transistors 72 and 74 is also coupled
to a terminal 86 directly and to a terminal 88 through a flyback
circuit comprising the serial combination of a diode 90 and a
resistor 92. The terminals 86 and 88 are coupled via the wires of
the cord 16 to the transducer 30.
The manner in which the circuit of FIG. 4 functions to produce
sound in accordance with the invention can best be understood by
referring to the waveforms of FIGS. 5 and 6. FIG. 5A depicts the
waveform of the natural voice when the sound "ah" is made. This
waveform, which is an accurate representation of the waveform
actually appearing on an oscilloscope in response to the electrical
measurement of such sound, is divided into a succession of periods
which in this instance are approximately 7 milliseconds in length
denoting a base frequency of the voice of approximately 143 hertz.
The waveform 94 of FIG. 5A comprises a succession of damped
sinusoids with each sinusoid occurring within a different period.
The waveform components are not perfect sinusoids in that they
include harmonic components in a variety of different
frequencies.
The waveforms of FIGS. 5B through 5I illustrate the manner in which
the circuit of FIG. 4 is used in conjunction with the user's mouth
to produce sounds very similar to those of the natural human voice.
The variable resistor 68 within the oscillator 52 is varied to
adjust the base frequency to approximately 143 hertz. This means
that every 7 milliseconds the oscillator 52 generates a new timing
signal or pulse 96 to define the beginning of a new period as shown
in FIG. 5B. Each pulse 96 is applied via the diode 70 to
momentarily bias the transistors 72 and 74 into conduction.
Conduction of the transistor 74 causes current to flow through the
resistor 84, causing a reduction in the collector voltage of the
transistor 74 as shown in FIG. 5C. This is accompanied by
generation of an output pulse to the transducer 30 as shown in FIG.
5F. At the moment that the transistor 74 is biased into conduction
as a result of a pulse 96 from the oscillator 52, the transistor 77
is biased into nonconduction as shown by FIG. 5D which is a
representation of the base voltage thereof. At the same time, the
charge level of the capacitor 78 begins to change as shown in FIG.
5E.
With the transistor 77 being biased into nonconduction, current at
the collector thereof flows through the diode 76 to the base of the
transistor 72 to hold the transistors 72 and 74 in a conductive
state and thereby temporarily hold the monostable circuit 58 in the
unstable state. Discharge of the capacitor 78 as shown in FIG. 5E
is controlled by the resistor 82 and the variable resistor 80. When
the level of charge on the capacitor 78 reaches a certain value,
the transistor 77 is biased into conduction which diverts the
current from the diode 76 and thereby biases the transistors 72 and
74 into nonconduction, to return the monostable circuit 58 to the
stable state. As seen in FIG. 5F, termination of conduction by the
transistor 74 terminates the flow of current through the resistor
84 and thereby the output pulse to the transducer 30.
The flyback circuit comprising the diode 90 and the resistor 92
permits relatively rapid change of the current flowing through the
coil 33 of the transducer 30. When the pulse to the transducer 30
is terminated, the voltage thereof decreases through zero to a
negative value due to the inductive effect of the coil 33 and the
force returning the diaphragm to the initial position. As shown in
FIG. 5F, the voltage gradually decreases back to zero from its
negative value as the inductive effect of the transducer coil 33 is
overcome. The minor variations in the voltage waveform in this
region are caused by movements of the transducer diaphragm 35 in
response to the behavior of the sound waves within the sound tube
32. Such behavior of the transducer diaphragm 35 and the resulting
signal aberrations are directly related to the advantageous use of
the transducer to generate desired acoustical waveforms in
accordance with the invention.
In the present example, the variable resistor 80 in the arrangement
of FIG. 4 has been set to deliver output pulses which are of about
20 microseconds duration. Such pulses are of relatively short
duration and result in an acoustical waveform of relatively small
amplitude as shown in FIG. 5G. The amplitude of the acoustical
waveform is increased by widening the rectangular pulses applied to
the transducer 30. This is accomplished by adjusting the variable
resistor 80 so as to provide for a longer time constant when
combined with the capacitor 78. As shown in dotted outline in FIGS.
5C through 5F, proper adjustment of the variable resistor 80
provides a longer time constant and thereby an output pulse of
greater duration. In the example shown in dotted outline, the time
constant of the RC circuit has been changed to provide for a 40
microsecond interval after pulsing of the transistor 72 by the
oscillator 52 before the capacitor 78 charges to a level sufficient
to bias the transistor 77 into conduction and switch the monostable
circuit 58 from the unstable state back to the stable state. The
variable resistor 80 is coupled to be controlled by the volume
control 20 on the outside of the housing 18 of the control unit
12.
The actual voltage of the transducer 30 in the case of a 20
microsecond output pulse is shown in greater detail in FIG. 6A. As
previously noted, termination of the pulse results in a reverse
voltage due to the flyback circuit comprising the diode 90 and the
resistor 92, the effects of the driving coil 33 within the
transducer 30 and the effects of the sound waves within the tube
32. In the case of the 20 microsecond pulse of FIG. 6A, the reverse
voltage decays to zero about 100 microseconds after the current
pulse was begun. FIG. 6B shows the case where the variable resistor
80 of FIG. 4 has been adjusted to provide for a relatively long
voltage pulse of 200 microseconds duration. In that instance, the
reverse voltage decays in a manner similar to that in the example
of FIG. 6A except that the decay time is much longer and the
voltage waveform is much more erratic due to the substantially
greater impulsing of the transducer 30 by the much longer voltage
pulse. In the example of FIG. 6B, the reverse voltage does not
decrease to zero until approximately 3,000 microseconds after the
initial pulse was begun.
As previously noted, FIG. 5G depicts the output of the sound unit
14 comprising the transducer 30 and the sound tube 32 when the
driving pulses are of 20 microseconds duration. FIG. 5H depicts the
output of the sound unit 14 in response to current pulses of 200
microseconds duration as shown in FIG. 6B. It will be noted that
the longer drive pulses produce an accoustical waveform which is of
greater amplitude and greater duration.
It will be noted from FIGS. 5G and 5H that artificial speech
devices 10 in accordance with the invention produce sound, the
waveform of which comprises a succession of damped sinusoids with
harmonics. This results from the use of voltage pulses of selected
duration to energize the transducer 30 in such a fashion as to
create an impulse function. The electrical impulsing as provided by
the voltage pulses is translated into an acoustical impulse
function by the transducer 30. It is known by Fourier and other
waveform analysis techniques that impulse functions have relatively
broad frequency spectrums. This helps to explain the resulting
acoustical waveforms as shown in FIGS. 5G and 5H where many
harmonic components are present. As previously noted, the mouth of
the user as well as the sound tube 32 have particular acoustical
characteristics. Among other things, the mouth and the tube
resonate at a certain frequency or frequencies characteristic of
the particular mouth and tube in question. In the case of natural
speech, the sound produced by the vocal cords contains many
frequency components enabling it to resonate in a favorable manner
within the mouth as the mouth is manipulated to convert the sound
into speech. However a similar effect has heretofore not been
possible with artificial speech devices because of the nature of
the sound produced by such devices. For example, where a signal of
essentially sinusoidal waveform is used to excite the transducer as
in the case of many prior art devices, the resulting acoustical
energy produced by the transducer is itself of generally sinusoidal
waveform. Very few, if any, harmonic components are present in the
resulting signal. As a result, the frequency of the generated sound
and the location at which such sound is introduced within the mouth
must be very critically selected in an attempt to maximize the
resonate action as the mouth converts the sound into speech.
However, even where the frequency characteristics and place of
introduction of the sound are carefully chosen, the results are
often marginal or unacceptable due to the variations in the
characteristics of the mouth, variations in the frequency or
location of the generated sound and other factors.
In the present invention, the drive pulse generated at the
beginning of each period acts to move the diaphragm 35 of the
transducer 30 in a selected direction and by a selected amount, at
which point the pulse is terminated and the diaphragm 35 is
thereafter free to move in response to the impulsing thereof for
the remainder of the period. Where the pulses are of relatively
short duration as in the case of the 20 microsecond pulse of FIGS.
5F and 6A, the transducer diaphragm 35 is driven in the selected
direction as determined by the polarity of the pulse over only a
relatively small extent of its possible travel. In the case of a
very large pulse such as the 200 microsecond pulse shown in FIG.
6B, the transducer diaphragm 35 is driven to or almost to its limit
of travel. This is shown by the relatively large initial value of
the reverse voltage which is depicted in FIG. 6B and which is due
to the very strong attempt by the diaphragm 35 to return from a
greater displacement to its neutral position.
The digital rectangular technique employed to energize the
transducer also results in relatively efficient operation of the
electrical circuitry, when compared with sine wave energizing
techniques. This permits a relatively long period of operation with
a compact power source.
FIG. 5I illustrates the waveform of sound which actually results
when the waveform of FIG. 5G is introduced into the mouth and
converted into speech by the mouth. The waveform shown is an
accurate reproduction of an oscilloscope waveform which resulted
from the user making a simple "ah" sound as in the case of FIG. 5A.
It will be noted that the waveform of FIG. 5I is of considerably
greater amplitude and duration than the corresponding waveform of
FIG. 5G. This is principally due to the increased resonant effect
which the mouth has on a signal of multiple frequencies as in the
case of FIG. 5G. Like the waveform of FIG. 5A which is produced by
the natural human voice, the waveform of FIG. 5I comprises a
succession of damped sinusoids with harmonics.
As in the case of FIG. 5A, the waveform of FIG. 5I is divided into
equal periods of 7 milliseconds duration. This is a further feature
which makes speech produced with the aid of artificial speech
devices according to the invention more intelligible and which is
not found in prior art devices. For example, in those devices where
the transducer is excited by a signal of generally sinusoidal
waveform, the amplitude of the resulting acoustical wave of
generally sinusoidal form at the output of the transducer is of
generally constant amplitude and does not decay or decrease over
the course of each period as in the case of the present
invention.
A further feature provided by the invention is the relatively high
frequency of the output sound. It will be seen in FIG. 5I that the
frequency of the sinusoid thereof is considerably higher than the
frequency of the sinusoid in FIG. 5A. Thus while the base frequency
of 143 hertz is preserved, the frequency of the sinusoid or other
basic waveform as produced by the present invention is increased.
This frequency increase has been found to greatly enhance the
intelligibility of speech produced using artificial speech devices
according to the invention. Accordingly, for a given amplitude
sound generated using the devices of the invention is more clearly
understood than that of the natural human voice. This has been
clearly demonstrated by actual listening tests in which speech
produced with the help of devices according to the invention is
compared with the natural human voice in terms of its ability to be
understood over substantial distances at relatively low
volumes.
As previously discussed, lengthening of the drive pulses from 20
microseconds to 200 microseconds produces an output sound as shown
in FIG. 5H in which the sinusoids are of greater amplitude and
duration than those of FIG. 5G. The waveform of speech produced
using the sound of FIG. 5H is very similar to that shown in FIG. 5I
except that it is about an order of magnitude larger in
amplitude.
As previously described the volume of the sound produced by the
artificial speech device 10 is changed by varying the pulse width
via the variable resistor 80 using the control 20. The pitch or
tone of the output sound can be varied by changing the base
frequency. In this respect, however, the frequency is varied only
to the extent that the pulse repetition rate is varied so as to
change the period. The pulse width remains the same. The base
frequency is varied via the pitch control 26 which adjusts the
variable resistor 68 within the oscillator 52.
FIGS. 5J through 5P provide one example in which the resistor 68 is
adjusted to shorten the period and thereby increase the base
frequency. As shown in FIG. 5J, the oscillator pulses 96 are in the
same size but occur with greater frequency than in the case of FIG.
5B. The RC time constant remains the same, and accordingly, the
transistor 74 is turned on for the same amount of time as in the
case of FIG. 5C as shown in FIG. 5K. Likewise the transistor 77 and
the capacitor 78 behave in similar fashion as seen in FIGS. 5L and
5M to produce 20 microsecond current drive pulses as shown in FIG.
5N. The resulting output of the sound unit 14 is shown in FIGS. 5-O
with the speech produced in response thereto being shown in FIG.
5P. The acoustical waveform of FIGS. 5-0 is the same as that of
FIG. 5G except that the sinusoids occur more frequently since the
frequency is higher and the periods are correspondingly shorter.
Likewise, the acoustical waveform of FIG. 5P is similar to that of
FIG. 5I except that the sinusoids are shorter.
In actual practice, the pitch control 26 is typically used to vary
the base frequency until a frequency or range of frequencies is
found which seems to best compliment the user's characteristics.
Also the control 26 may be varied during the course of conversation
to reduce the monitone effect of speech generated at a constant
base frequency. For example, the variable resistor 68 may be
coupled via internal wires to a hand-held unit which includes an
on-off switch and a trigger control.
While the invention has been particularly shown and described with
reference to a preferred embodiment thereof, it will be understood
by those skilled in the art that various changes in form and
details may be made therein without departing from the spirit and
scope of the invention. For example, the control unit 12 can be
miniaturized such as with the help of integrated circuits so as to
be combined with the sound unit 14 and contained completely within
the mouth.
* * * * *