U.S. patent number 3,784,750 [Application Number 05/229,322] was granted by the patent office on 1974-01-08 for apparatus and prosthetic device for providing electronic correction of auditory deficiencies for aurally handicapped persons.
This patent grant is currently assigned to Shalako Resource Systems, Inc.. Invention is credited to John K. Lauchner, William P. Stearns.
United States Patent |
3,784,750 |
Stearns , et al. |
January 8, 1974 |
**Please see images for:
( Certificate of Correction ) ** |
APPARATUS AND PROSTHETIC DEVICE FOR PROVIDING ELECTRONIC CORRECTION
OF AUDITORY DEFICIENCIES FOR AURALLY HANDICAPPED PERSONS
Abstract
There is disclosed herein improved methods and apparatus for
measuring and describing auditory deficiencies for purposes of
prescribing compensatory amplification, and improved methods and
apparatus for providing such compensatory amplification in a
prosthetic device in practical wearable form, and particularly
electronic circuits which may be used both in measuring and
correcting such deficiencies.
Inventors: |
Stearns; William P.
(Scottsdale, AZ), Lauchner; John K. (Phoenix, AZ) |
Assignee: |
Shalako Resource Systems, Inc.
(Scottsdale, AZ)
|
Family
ID: |
22860718 |
Appl.
No.: |
05/229,322 |
Filed: |
February 25, 1972 |
Current U.S.
Class: |
381/320; 381/98;
381/106; 381/321 |
Current CPC
Class: |
H03G
9/02 (20130101); H04R 25/505 (20130101); H04R
25/356 (20130101) |
Current International
Class: |
H03G
9/00 (20060101); H04R 25/00 (20060101); H03G
9/02 (20060101); H04r 029/00 () |
Field of
Search: |
;179/1N,17R,1F,1D,1FS,1A |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Cooper; William C.
Assistant Examiner: Leaheey; Jon Bradford
Attorney, Agent or Firm: Lyon & Lyon
Parent Case Text
CROSS-REFERENCE TO RELATED APPLICATIONS
The present application is directed to inventive concepts which are
improvements over those described in copending U.S. Pat.
application Ser. No. 133,229, filed Apr. 12, 1971, by William P.
Stearns and entitled, "Method and Apparatus for Providing
Electronic Sound Clarification for Aurally Handicapped Persons."
The present application also is related to copending U.S. Pat.
application Ser. No. 229,398 filed concurrently herewith in the
names of William P. Stearns and Barry S. Elpern entitled, "Method
for Providing Electronic Restoration of Speech Discrimination in
Aurally Handicapped Persons," which describes and claims methods
and apparatus disclosed in the present application.
Claims
What is claimed is:
1. Apparatus useful for measuring human auditory deficiency and/or
providing compensatory amplification for aurally handicapped
persons comprising:
input circuit means for receiving complex phonic signals to be
selectively amplified over a plurality of pass bands,
selective audio amplification means comprising a plurality of
independently adjustable filter means having adjacent pass bands
for enabling independent adjustment of amplification within each of
a plurality of adjacent audio pass bands, said amplification means
being coupled with said input circuit means for receiving signals
therefrom and each of said filter means providing output
signals,
a plurality of compression means, each compression means coupled to
the output of a corresponding filter means for providing output
signal compression when the output signals from said corresponding
filter means exceeds a predetermined level, and
summation means for combining the output signals after the output
signals have been acted on by said plurality of compression
means.
2. Apparatus as in claim 1 wherein said filter means are active
filter means.
3. Apparatus as in claim 2 wherein each of said active filter means
includes variable amplitude control means coupled with said active
filter means.
4. Apparatus as in claim 3 wherein said input circuit means
includes microphone means coupled with an amplifier, said amplifier
being coupled with said amplifier means.
5. Apparatus as in claim 3 wherein each of said active filter means
includes operational amplifier low pass, high pass, and band pass
filters, said operational amplifier filters being formed of
integrated circuits.
6. Apparatus as in claim 1 wherein the complex phonic signals are
continuous discourse.
Description
BACKGROUND OF THE INVENTION
This invention relates to the sound amplification arts, and to
their application in the amelioration of auditory deficiencies
resulting from damage to the sensori-neural structures of the human
ear. It relates particularly to methods and apparatus for detecting
and specifying deficiencies in a persons ability to perceive and to
comprehend spoken language, and to methods and apparatus for
correcting such deficiencies.
Sensori-neural hearing loss is generally considered to be the most
prevalent type of auditory handicap found in the United States as
well as in other civilized cultures. It constitutes a significant
barrier to adequate communication in 5 to 10 percent of the total
United States population, and in more than 50 percent of the
population over 60 years of age. Furthermore, these proportions are
expected to increase in conjunction with ongoing increases in
ambient noise levels and life expectancy in our society.
Sensori-neural impairment may result from any one or more of a
number of causes, including, but not limited to genetic and
congenital factors, viral diseases, specific toxic agents,
circulatory disturbances, specific physical trauma and excessive
exposure to noise. Irrespective of the primary cause, however,
sensory cells within the organ of hearing or their associated
neural units suffer some degree of damage and are rendered
partially or totally incapable of fulfilling their respective roles
in the processing of auditory information. This form of damage
cannot be repaired by means of currently known medical or surgical
techniques, and the probability of discovery of effective
techniques within the foreseeable future appears rather remote.
Thus, in virtually all cases of sensori-neural hearing loss,
amplification of incoming sounds represents the only possible means
for restoring adequate hearing ability.
Hearing loss resulting from sensori-neural damage is usually
irregular with respect to frequency, being selectively greater for
particular portions of the audible frequency range. The ability to
hear sounds in the range above 1,000 Hz is often affected more than
the hearing of sounds below 1,000 Hz, although this is by no means
a universal observation. The ultimate consequence of irregular
hearing acuity for various portions of the audio frequency spectrum
is distortion in the perception of complex sounds, i.e., sounds
composed of a number of different frequencies.
A certain amount of distortion in complex sounds may be tolerable,
but current information does not permit precise specification of
the maximum amount of each type of distortion which may exist
without interfering materially with accurate sound recognition.
Many gross sounds, for example, do not demand a great deal of
analytic power in the auditory system, so even a rather severely
impaired system may function adequately in the interpretation of
such sounds.
In audiologic parlance, the term "discrimination" denotes the
capacity of the ear to analyze incoming acoustic patterns and
interpret them appropriately. Analytic power may fail at any of
several stages in the auditory process, commonly in the organ of
hearing or first order neurons due to damage to these structures.
Since the ear may be required to perform many degrees of
discrimination, varying from extremely coarse to extremely fine,
its analytic power may be measured through the use of tests which
demand auditory discriminations of progressive difficulty until
failure occurs.
Among the most difficult discriminations required of the human ear
are those necessary for accurate interpretation of speech,
particularly speech in the presence of noise. Because of the
fundamental importance of spoken communication, it is obvious that
chronic inability to understand what people say could profoundly
influence an individual's social, economic, and cultural
well-being. Tests of speech discrimination are commonly employed,
therefore, to derive a realistic estimate of a person's everyday
functional adequacy in hearing.
Each of the phonic units of a spoken word is a complex sound,
composed of several frequencies clustered in a more-or-less
definable range. When the acuity of the ear has been selectively
impaired in a specific frequency range, speech sounds or their
components falling in that range may be heard at reduced intensity
or not at all. Impairment in several frequency ranges compounds the
difficulty and is probably responsible in large measure for the
primary complaint of the individual with sensori-neural hearing
loss, that he can hear a speaker's voice but cannot understand what
is said. The mechanism for inhibiting such understanding may be the
non-linear responses that result in intermodulation products and
harmonics which could cause interference with the desired spectral
components of speech.
On the basis of the foregoing information, it would seem quite
reasonable to deal with sensori-neural hearing loss by selective
spectrum amplification; that is, providing amplification only in
those frequency ranges or bands in which acuity is deficient, and
only in the amount of the deficiency. Thus, the ultimate value of
selective spectrum amplification rests on the application of
appropriate methods for measuring the degree of auditory deficiency
as a function of various frequency bands, and also on the
construction of a wearable device which is fully capable of
producing amplification to compensate for the measured
deficiencies. Because of existing inadequacies in both respects,
the principle of selective amplification has fallen into disrepute,
for the hearing aid industry has adopted the pure tone (single
frequency) threshold audiogram as the criterion measurement, and
has produced hearing aids with inadequate capabilities for
providing proper acoustic output at each portion of the audio
band.
The threshold audiogram curve represents an individual's measured
absolute auditory threshold for a series of pure frequency tones,
usually in the range of 250 Hz to 8,000 Hz sampled at octave
intervals on the assumption that intra-octave tone thresholds
follow the general audiogram contour. However, it is demonstrable
that fairly marked departures from the overall pattern may exist at
intermediate frequencies, i.e., frequencies between pure tones, one
octave apart.
The rationale for utilizing threshold measurements is shrouded in
history, but it is exceedingly interesting to note that the
analogous procedure of measuring visual thresholds for
monochromatic (single color) lights is never performed to measure
the visual acuity of the eye or to prescribe eyeglasses. In fact,
careful consideration of the types of measurements which are
genuinely helpful in guiding the design of particular hearing aid
features suggests that the pure tone threshold curve is virtually
useless for several reasons:
A. under everyday circumstances, individuals react only to
supra-threshold sounds, as these are the sounds of primary
significance. For practical purposes, threshold sounds remain
unnoticed.
B. the contour of an individual's threshold curve is observably
different from the contour of his supra-threshold equal loudness
curves or comfortable listening level curves.
C. an individual's recognition of complex phonic units or their
combination into spoken words is essentially unrelated to his
acuity for individual pure tones.
Control of acoustic output in current hearing aids is ordinarily
achieved through manipulation of frequency response, which refers
to the acoustic output of a sound transmission system at each of
the frequencies within its pass band when the input level is
maintained constant for all frequencies. A graphic representation
of a system's frequency response is referred to as a response
characteristic, curve or contour. Manufacturers commonly claim that
they are able to build hearing aids to yield any required frequency
response; but this does not appear to be the case in practice
because there are definite limitations on the bandwidths and
response curves available in prsent day aids. In practice,
manufacturers use combinations of components which produce a
limited choice of response patterns and simply select one which
most closely corresponds to the criterion, which, as mentioned
earlier, usually is a threshold audiogram curve.
One additional comment is relevant as a preface to the innovative
concepts to which the present invention is particularly addressed.
It is generally recognized that the ear with sensori-neural hearing
loss is excessively susceptible to overloading, which is to say
that, although it may be relatively insensitive to sounds of low or
moderate intensity, it is hypersensitive to sounds of higher
intensity (e.g., non-linear response characteristics). This
condition restricts the useful operating range of the ear, referred
to as the dynamic range; that is, the decibel difference between
the lowest intensity at which a sound is reliably detected
(absolute threshold) and the upper limit of comfortable loudness
for that sound (discomfort threshold).
Whereas the dynamic range of the normal ear is of the order of 100
dB, the range of a sensori-neurally impaired ear may be as little
as 10 or 15 dB, generally over a limited frequency spectrum range.
Thus, for an impaired ear to function with any degree of adequacy,
the full intensity range of the outside acoustic world must be
restricted in some way to fit through an abnormally small sound
window and such restriction must cause minimal intermodulation
products, harmonics, and so forth which would result in distortion.
Without such restriction, the ear is readily overloaded, leading to
psychologic or physical annoyance and distortion of incoming
acoustic patterns.
The consequences of overloading have been appreciated for many
years, and output compression devices are widely used in today's
hearing aids. Without exception, however, these devices operate on
a broad frequency band, so that when any frequency component of a
signal reaches a predetermined critical level, the entire pass band
of the hearing aid is compressed. Consequently, the components
which are not at a critical intensity are needlessly
attenuated.
Our evaluation of relevant factors has led to the evolution of
several innovative concepts concerned with improved methods and
apparatus for measuring and describing auditory deficiency for
purposes of prescribing compensatory amplification, and with
improved methods and apparatus for providing such compensatory
amplification in practical and wearable form.
SUMMARY OF THE INVENTION
While the present application and said co-pending appliation filed
in the names of Stearns and Elpern include similar disclosures, and
both teach the same inventive concepts for the sake of
completeness, the claims of the present application are
particularly directed to electronic circuits of the nature
disclosed herein and equivalents thereof for enabling the
objectives set forth herein to be accomplished. Accordingly, it is
an objective of the present application to provide an electronic
measurement system with the following capabilities:
a. division of the audible frequency spectrum into a series of
adjacent frequency bands through the use of filter networks, the
width of any band in the series being of such magnitude as may be
deemed appropriate;
b. specific and individual intensity or volume control associated
with each of the filter networks defined in (a) above;
c. specific and individually adjustable output compression
associated with each of the filter networks defined in (a)
above;
d. activation of the filter networks individually or in concert
(all pass);
e. introduction and transmission of recorded material such as from
tape, or electronically generated signals through the system of
filter networks;
f. electro-mechanical transduction of electronically processed
signals, such transduction occurring within the external auditory
canal of the test subject;
g. pre-amplification and mixing of input signals and for broadband
intensity control;
h. monitoring voltage and/or recording the swept frequency spectrum
across the output transducer.
Through the use of the foregoing concepts, further objectives of
this invention are:
a. to facilitate the use of narrow band signals as test stimuli
inasmuch as they represent the most satisfactory compromise between
the precise physical describability of pure tones and the complex
acoustic composition of speech sounds;
b. to facilitate the use of continuous conversational speech as
test stimulus;
c. to facilitate the measurement of comfortable listening level or
other supra-threshold response to various frequency bandwidths as
an index of auditory deficiency.
Another object of this invention is to enable the provision of
sufficiently miniaturized hearing aid apparatus for wearing by
aurally handicapped persons. Such miniaturization can be
accomplished by electronic techniques, and the apparatus is
intended to implement the amplification features determined by the
electronic measurement techniques.
BRIEF DESCRIPTION OF THE DRAWINGS
The foregoing and other objects and features of the present
concepts will become better understood through a consideration of
the following description taken in conjunction with the drawings in
which:
FIG. 1 is a block diagram of test equipment used in testing the
speech discrimination ability of a subject according to the present
concepts;
FIGS. 2a and 2b are curves illustrating speech discrimination
scores associated with various spectra and test conditions to be
discussed later;
FIGS. 3 and 4 are curves illustrating the response characteristics
of a master hearing aid after such response characteristics have
been set to obtain the best speech discrimination ability for the
patient under evaluation;
FIG. 5 is a block diagram of a wearable hearing aid according to
the present invention;
FIG. 6 is a more detailed block diagram of a hearing aid of the
nature of that illustrated in FIG. 5;
FIG. 7 is a curve indicating typical response of a single filter of
the hearing aid of FIG. 6; and
FIGS. 8 through 11 are specific circuit diagrams of the aid of FIG.
6.
Briefly, in accordance with the present concepts, the format for
measuring the auditory deficiency of a subject involves testing
with a master hearing aid device in accordance with steps as set
forth below.
1. The hearing aid receiver is inserted into the external canal of
the test ear and secured with a packing of earmold impression
material to provide an acoustic seal.
2. The non-test ear is occluded by an insert earplug and a
circumaural muff to block any auditory perception by that ear.
3. The patient is seated comfortably opposite a loudspeaker within
a sound-treated enclosure, the loudspeaker having been acoustically
equalized to produce a flat frequency-response characteristic at
the position of the patient's head. The test signals used to
accomplish such equalization are narrow bands of noise.
4. Tape recorded continuous discourse is delivered through the
loudspeaker at a level approximating that of normal conversation,
i.e., 60 dB SPL.
5. The patient is instructed to listen carefully to the tape
recorded speech and to adjust the various filters and the main
volume control to achieve maximum clarity and comprehension of the
material.
6. When he feels he can attain no further improvement in clarity
and comprehension of the recorded speech, the patient notifies the
examiner and the recording is stopped.
7. Standardized tape recorded tests of speech discrimination
ability are then delivered at the same level as the continuous
discourse. The speech discrimination score is noted for each test
given.
8. Relative performance following the present concepts may be
assessed through comparison with unaided discrimination scores or
scores obtained under identical listening conditions with other
aids. Furthermore, the voltage level across the hearing aid
receiver may be measured and recorded for each filter such as
illustrated in FIGS. 2A-2B. These measurements when obtained for
each filter network in the system, provide the response curve
necessary to restore the subject's discrimination ability, and this
response curve can then be used for prescribing the appropriate
correction. If necessary, the response curve, which the subject has
initially adjusted, may be revised by the subject or by the
examiner to effect further improvement in speech discrimination
ability. Additional standardized speech discrimination tests are
then required to assess the effect of such revisions. Such
adjustments may be based, for example, on whether the subject is
missing vowels or consonants and so forth. Furthermore, the
examiner may adjust the response to smooth large variations
therein. Also, a pink or white noise test curve may be run and/or
noise inserted during the test process.
Before giving further examples of measurement procedures and
results, suitable test apparatus will be discussed. Turning to the
drawings, FIG. 1 illustrates an organization of test equipment used
in testing speech discrimination of a subject to the extent
necessary to practice the method of the present invention. Speech
from an audio tape recorder and playback apparatus 10 is applied to
an audio mixer network 11. A pink or white noise generator 12 also
may be coupled to the audio mixer network 11 to combine speech and
noise. The speech and/or audio noise is thus fed to a filter
network 13 which comprises a plurality of filter networks F.sub.1
through F.sub.n. Each of the filter networks F.sub.1 through
F.sub.n has a discrete pass band and the entire combination 13
preferably covers a frequency range from approximately 125 Hz to at
least 6,300 Hz. The filters F.sub.1 through F.sub.n divide the
audible frequency spectrum into adjacent pass bands. The pass bands
provided by each filter F.sub.1 through F.sub.n may be as wide or
as narrow as desired in obtaining appropriate audio noise and/or
speech recognition characteristics, and need not be related in an
octave relationship as is frequently the case with filter networks.
Filter set 13 may also be implemented with a combination of
adjustable band reject filters arranged in a series configuration
rather than parallel filters.
Furthermore, the signal amplitude variation is adjustable over a
suitable range. The output of the filter set 13 is passed through
an adjustable gain broadband audio amplifier 14, and hence to an
audio receiver 15 positioned in a human ear 16. A vacuum tube
voltmeter 17 can be used to measure the amplitude of the signal
voltage impressed across the receiver 15.
By way of example, the noise generator 12 may be a HP (Hewlett
Packard) 8057A precision noise generator, the tape recorder and
playback 10 may be a Craig Model -2704 cassette recorder and
playback unit, the mixer 11 may be a Shure .MG7 microphone mixer,
the filter network 13 may be a HP 8056A, the audio amplifier 14 a
McIntosh MC2505, the receiver 15 a Tibbetts model 102-10G hearing
aid receiver, and the vacuum tube (RMS) voltmeter 17 a Ballentine
model 320.
The response curves for two test subjects are shown in FIGS. 2a and
2b together with discrimination test scores. FIG. 2a pertains to
one ear of one test subject, and FIG. 2b pertains to one ear of
another test subject, the ordinate being logarithmic and indicating
voltage across the receiver 15 in millivolts. The abscissa is
frequency. Curve (a) in both figures represents the most
comfortable listening level set by the subject for individual
one-third octave bands of noise; curve (b) represents the
comfortable listening levels set by each subject for maximum
intelligibility of running speech; and curve (c) is an examiner's
revision of curve (b) for example, to minimize response peaks.
Curves (d) and (e) in FIG. 2a respectively represent frequency
response of 3 dB per octave and 4 dB per octave. These latter two
curves are simply a plot of frequency versus sound pressure level,
but with a constant dB per octave change as compared with the
random response plotted from the subject's own response setting,
modified in some cases by the examiner as indicated above with
respect to curve (c).
The percentages indicated at the upper right edge of the curves of
FIGS. 2a and 2b indicate speech discrimination scores achieved by
the respective two subjects. This refers to a test wherein the
subject is provided with hearing aid apparatus having a response
curve set according to the respective curves indicated in FIGS. 2a
and 2b and the subject listens to speech supplied to the input of
the hearing aid apparatus. The test material consists of
standardized phonetically balanced lists of words, each list
comprising 50 words to which the subject must respond by repeating
each word immediately after it is presented by a tape recorded
speaker. Considering FIG. 2a in this regard, it will be seen that
the speech discrimination score was only 4 per cent (4 percent) for
response curve a, that is, merely a comfortable listening level set
by the subject. On the other hand, a speech discrimination score of
76 per cent was obtained when the hearing aid apparatus was
adjusted to response curve ba significant improvement; and
increased to 92 per cent for response curve c (the examiner's
revised curve b). As shown by FIG. 2b, the original score was 20
.sub.per cent for curve a, but increased to 54 per cent for curve
b, and to 88 per cent for curve c. It will be understood that
another test subject's response may be significantly different from
that shown in either FIGS. 2a or 2b and, in fact, each ear of a
subject may differ significantly in its uniqueness.
It should be noted that the final response adjustment (curve c) for
optimum discrimination was made by an examiner using a spectrum
shaping technique that has been developed as part of this
invention, and principally involves smoothing of response peaks of
curve b as noted earlier. In the case of FIG. 2a, the individual's
ability to distinguish the spoken words of the speech
discrimination tests increased to 9 percent (curve c), which is a
significant improvement over the 76 percent for curve b and a
significant improvement over the scores for the constant dB per
octave change curves d and e.
Several specific examples and results in accordance with the
present test method are described below. Subject X was tested with
three hearing devices. In each instance the left ear was tested and
the right ear was covered. Test No. 1 was made while the subject
used her personal, commercially purchased, hearing aid. The test
involved use of Comm Tech Auditory Test N-1 sentences with a
competing signal of two female talkers in the background. This is a
relatively difficult discrimination test as compared to mere use of
a word list. The resulting speech discrimination score (SDS) was 22
percent. Test No. 2 involved a wearable hearing aid with adjustable
filters using the circuit shown in FIG. 6 hereof and the same test
as in Test No. 1, and the speech discrimination score was 61
percent after adjustment of the wearable hearing aid in accordance
with steps 5 and 6 of the test noted earlier. Test No. 3 was
conducted several days later and involved the use of CID Auditory
Test W-22, List 4D. The speech discrimination score with the
subject's personal aid was 68 percent. Test No. 4 involved CID
Auditory Test W-22, List 2F, and the use of a master hearing aid
with a circuit as illustrated in FIGS. 5 and 6. The wearable aid
and master aid are the same electrically and use essentially the
same components, but the master aid is physically larger and has
more easily adjustable knobs for filter settings. The speech
discrimination score in this case was 88 percent. FIG. 3
illustrates the frequency response of the master hearing aid after
it was set by Subject X in the test.
Test No. 5 involved an audiological evaluation of Subject Y
conducted by a university Speech and Hearing Clinic. The
phonetically balanced speech discrimination score was 70 percent.
This same subject was tested in accordance with the present method
after the subject set the response of the master hearing aid. The
subject was tested with CID Auditory Test W-22. List 4D, and had a
speech discrimination score of 92 percent in the same ear. A
similar test of Subject Y was conducted with CID Auditory Test
W-22, List 2F, wherein the subject adjusted the response of the
master hearing aid and then the response was trimmed by an
audiologist in the manner noted earlier, and the speech
discrimination score was improved to 96 percent. FIG. 4a is an
oscillographic waveform, similar to the curve of FIG. 3, indicating
the response of the master hearing aid after being set and trimmed
by the examiner.
Test No. 6 involved Subject Z whose unaided phonetically balanced
speech discrimination score was 66 percent. He was tested in
accordance with the present method after he had adjusted the
response of the master hearing aid. The master hearing aid in this
case differed from that previously employed in Test No. 5, in the
following respect: six adjacent filter networks divided the overall
speech spectrum into unequal bandwidths, whereas, for previous
tests, the filter bands were each one octave in width. The unequal
bandwidths were selected on the basis of their relative
contribution to overall speech intelligibility. Such bandwidths are
often referred to as "equal intelligibility bands." With the
aforesaid master hearing aid adjusted by the subject (See FIG. 4b)
for optimum speech intelligibility, he achieved a speece
discrimination score of 96 percent on CID Auditory Test W-22, List
3-D. A control test under identical acoustical and procedural
conditions with a conventional hearing aid yielded a speech
discrimination score of 84 percent on CID Auditory Test W-22, List
3-F.
Turning now to an exemplary aid, a block diagram of a wearable
hearing aid is shown in FIG. 5. The Figure illustrates the
practical miniaturized circuitry for a hearing aid which can be
adjusted to duplicate the response curve obtained with the test
apparatus shown in FIG. 1. A microphone and FET (field effect
transistor) amplifier 21 feeds the received input signals to a
broadband audio IC (integrated circuit) amplifier 22 which has a
volume (amplitude) control 23. Driver amplifier 24 provides a low
impedance source for filter network 25, including plural amplitude
controls 26 and plural active IC bandpass filters 27, the outputs
of which are fed to a summation, or all-pass, network 28. As will
be apparent to those skilled in the art, the bandpass filters 27
each have a bandpass and amplitude control 26 suitable for closely
providing or approximating the desired response curve (e.g., curve
c in FIGS. 2a and 2b) and thus are selected to provide suitable
speech recognition. These filters 27 accordingly may each provide a
portion of the total pass band of the filter network 25. The pass
band provided by each filter may be as wide or as narrow as
required ot obtain optimal speech discrimination and need not be
related in any octave relationship or fractional combination
thereof. An integrated circuit amplifier 29 having a pass band
commensurate with that of filter network 25 provides the final
signal amplification prior to the signal being applied to a
receiver 30. Automatic saturation elimination control 31 provides
signal compression when the signal exceeds a predetermined level.
The foregoing hearing aid configuration offers the following
advantages: Independent control of pass band amplitude for each of
several portions of the spectrum; separate response control for
each ear (binaural); ease of readjustment as the patient's
requirements change with time; this may be accomplished by the
replacement of filter elements having different pass bands and
adjusted for different amplitudes. In desired cases a narrow band
notch rejection filter may be added after the summation network to
alleviate narrow band resonance problems observed in some patients.
The concept is readily adaptable to MSI (medium scale integration)
integrated circuit techniques. This permits a substantial size
reduction in hearing aid models. Rechargable or long life batteries
may be used as desired. Ease of repair, ruggedness, and waterproof
sealing of the electronic circuits can be readily accomplished.
Attractive and compact packaging can be provided.
FIG. 6 is a detailed block diagram of a hearing aid of the nature
of that illustrated generally in FIG. 5, and may be manufactured in
a miniaturized wearable form. Also, the circuit of FIG. 6 can be
used in the master hearing aid which, as noted earlier, is
preferably a larger test instrument having larger and more readily
adjustable knobs for varying the response characteristics thereof
during testing. The wearable aid may be as small as practical. A
prototype aid having thumbwheel adjustments for the filter circuits
has been constructed and packaged with outside dimensions of five
by three by one and one-eighth inches, but obviously smaller sizes
can be manufactured. An exemplary master aid has been constructed
with outside dimensions of 15 by 10 by 4 and 1/2 inches. The basic
hearing aid shown in FIG. 6, includes an integrated
microphone/low-noise FET amplifier stage 50 followed by a low-noise
amplifier section 52 which drives a bank of parallel and
independently adjustable bandpass filters indicated generally at
54. The filters are side-by-side in frequency and are adjustable in
gain only after initial frequency alignment. FIG. 7 is a scope
trace showing a typical single filter response at a center
frequency, fo, of 1 KHz. A summation circuit 56 adds all of the
filter outputs on a common bus in a linear summation. The summed
signal then is applied to a linear amplifier and drive circuit 58
which, in turn, drives a miniature magnetic receiver 60 of the
hearing aid.
In order to prevent overdriving the receiver into the nonlinear
region, an automatic saturation elimination (ASE) circuit 62
provides a gain controlled loop back to the front end circuits. The
ASE feedback signal can either be sensed at the signal line 64 to
the filter bank 54 or at the receiver drive point 66 in the hearing
aid output. A volume control 68 ahead of the filter bank 54 permits
the overall hearing aid gain to be set at any desirable quiescent
value.
Low pass filter circuits 70 and 72 are used for B+ and B- noise
filtering and decoupling at various points throughout the system as
desired. The hearing aid is designed to operate from hearing aid
batteries providing a balanced plus and minus voltage with respect
to the common bus 64.
Turning to a more specific discussion of the system of FIG. 6, a
miniature ceramic microphone with a built-in low-noise FET
amplifier may be used as the input transducer stage 50. Units
similar to the Knowles BL-1671 may be employed. This unit has a
response from less than 100 Hz to greater than 8,000 Hz as measured
by standard hearing aid microphone measurement techniques. With the
system as shown in FIG. 6, a 1.3 volt dc supply voltage is used to
power the build-in FET amplifier of the input transducer. Control
of this voltage to lower levels is one way of controlling the front
end gain of the hearing aid by such means as the ASE control loop
62. A dc voltage on the output leads, in combination with the audio
signal, requires a decoupling capacitor prior to feeding the low
noise preamplifier in the hearing aid front end.
To minimize front end noise contribution, a dual Darlington
connected amplifier pair (such as a Motorola 2N5089 NPN low-level,
low noise device) operating at low current levels and with a large
input current limiting resistor, is included following the
microphone circuit as more specifically illustrated in FIG. 8 in
the low noise two stage preamplifier 80. The dual low-noise
amplifier is connected to the volume control potentiometer 68 that
sets the quiescent gain of the overall hearing aid.
Additional front end gain is provided by two operational amplifiers
82 following the low noise preamplifier 80 and volume control 68.
These amplifiers may incorporate very low current drain IC
operational amplifiers, such as the Solitron UC 4252 dual unit.
Feedback resistors around each operational amplifier permit the
gain to be set at any desired value within the operating range. A
complementary pair driver stage 84, including devices such as
Motorola 2N5089 and 2N5087 transistors, provides a push-pull drive
signal to the signal line 64 which feeds the filter bank circuits
54. This same bus is an alternate source for feeding the ASE
automatic gain control feedback loop 62 as noted earlier.
An exemplary filter bank is illustrated in FIG. 9 and includes six
parallel filter networks numbered 1-6 of adjacent frequency bands,
and each has independent gain control. However, it is to be noted
that different numbers and types of filter networks may be used as
desired. Active three pole filters are included which incorporate
operational amplifiers such as the Solitron UC 4253C triple
operational amplifier, in integrated circuit configuration. Each
amplifier draws microamperes of current, which is of prime
importance in minimizing battery drain for longer operating
life.
As illustrated in FIG. 9, each filter band is made up of three
operational amplifiers 90-92 in active filter circuit
configurations. The first filter section 90 is a low pass filter
followed by a high pass filter 91 and then a band pass filter 92.
Selection of the proper resistors (R's) and capacitors (C's)
determines the center frequency, band pass, ripple and gain of each
three pole filter section. A gain potentiometer 94 is included at
the input to each filter section to provide independent gain
control for the particular frequency band represented.
The selection of the band limits is flexible during the initial
alignment. Possible alignments include octave bands, one-third
octave bands, unequal bands adjusted for optimum speech
discrimination, and band with frequency gaps in special areas for
selective sound elimination purposes. The operational amplifiers
operate between a balanced positive and negative battery supply
with a quiescent output level of zero volts. This permits maximum
voltage swing of the output waveform prior to reaching saturation,
as well as minimum quiescent current drain during absence of
signal. The six filter outputs are linearly summed in a resistive
summation network 56 prior to feeding the post amplification
circuits 58 of the hearing aid.
In order to prevent distortion in the hearing aid during the
presence of large audio signal levels, an automatic saturation
elimination circuit (ASE) 62 is included as noted earlier. This
circuit samples the audio signal either at the signal line 64 of
the filter band 54 or at the drive point 66 for the hearing aid
receiver. As illustrated in FIG. 10, the audio signal is detected
in a voltage doubler circuit 96, passes through a low pass filter
97, and then feeds an NPN transistor common emitter driver stage
98. The latter stage 98 incorporates a device such as the Motorola
2N5089 transistor in a low current drain circuit. The output of
this stage supplied B+ for the FET amplifier in the microphone
assembly.
A large signal at the input of the ASE circuit 62 results in a drop
in the amount of voltage supplied to the microphone FET and thus
reduces the gain of the signal into the receiver front end. The
response time of the circuit 62 is in the range of a few
milliseconds and can be adjusted to other values if desired. A
large filter capacitor 1 .mu.f at the collector of the driver
transistor in stage 98 minimizes the noise applied to the FET
amplifier B+ supply and also provides a time constant needed in the
ASE loop to prevent loop oscillations. The diodes used in the
doubler 96 may be types such as 1N914 low cost silicon units
available from several manufacturers. The doubling action permits a
smaller signal to activate the ASE loop 62 without the addition of
transistor gain stages and the corresponding power dissipation.
A dual operational amplifier and complementary pair transistor
driver similar to the corresponding circuits ahead of the filter
bank 54 are used as a post amplifier circuit 58 to drive the
miniature magnetic receiver assembly 60. This post amplification
circuit is illustrated in FIG. 11. Like circuit components are
used, including the dual integrated circuit operational amplifier
102 (like amplifier 82 of FIG. 8) and an NPN/PNP complementary
driver transistor circuit 103 (like driver 84). The gains of the
operational amplifiers are set by means of feedback resistor
networks. Typical gain values of 10 dB per amplifier may be used in
the post amplifier stages. Current setting resistors in the
operational amplifier circuits permit quiescent operation with
microamperes of drain. The driver stage (complementary pair) have
the biases adjusted to provide a minimum current needed for driving
the receiver. A balanced positive and negative power supply with
respect to the signal line 64 permits low quiescent current drain
in the absence of a signal. An optional output signal connection to
the gain control loop 62 as described earlier permits the gain
control sensing to be supplied at the receiver input.
Finally, the system of FIG. 6 preferably employs a miniature
magnetic receiver. Various miniature magnetic receivers can be
connected to the driver circuit of the hearing aid depending on the
patient's requirements. For persons requiring more volume, larger
diaphragm receivers can be used. Smaller receivers capable of being
placed entirely within the ear canal can also be driven by the same
driver stage.
The present embodiments of this invention are to be considered in
all respects as illustrative and not restrictive, the scope of the
invention being indicated by the appended claims rather than by the
foregoing description, and all changes which come within the
meaning and range of equivalence of the claims therefore are
intended to be embraced therein.
* * * * *