Hearing aids

May 28, 1968

JEAN-CLAUDE LAFON

3,385,937

HEARING AIDS

Filed Jan. 29, 1964

<img file="US3385937A_D0001.tif" />

<img file="US3385937A_D0002.tif" />

<img file="US3385937A_D0003.tif" />

SUMMATION

<img file="US3385937A_D0004.tif" />

AMPLIFIER AND AMPLITUDE DE MOD

<img file="US3385937A_D0005.tif" />

<img file="US3385937A_D0006.tif" />

United States Patent Office

3,385,937

Patented May 28, 1968

385 937

HEARING AIDS

Jean-Claude Lafon, Saint-Rambert-I’Ile-Barbe, Rhone, France, assignor to Centre National de la Recherche Scientifique, Paris, France, a French Government administration

Filed Jan. 29, 1964, Ser. No. 341,004

Claims priority, application France, Feb. 14,1963,

924,885

Claims. (CI. 179—107)

ABSTRACT OF THE DISCLOSURE

An auditory prothesis device comprising a microphone, two band-pass filters connected to the output of said microphone for passing, respectively, frequencies ranging, on the one hand, from 1500' to 3500 c.p.s. and, on the other hand, from a lower limit ranging from 4500 to 6000 c.p.s. to an upper li.mit ranging from 7000 to 8000' c.p.s., means for detecting the amplitude of the signals passed through each of said filter means, two oscillators for producing each a distinct alternating voltage of a frequency ranging from 350 to 1000 c.p.s., a modulator for modulating the voltage supplied by each of said oscillators in accordance with the output of each of said detecting means, respectively, means for mixing the modulated voltages supplied by said two modulators, an amplifier connected to the output of said microphone in shunt with said band-pass filters, a .mixer having two inputs connected respectively with the output of said amplifier and with the output of said mixing means and an earphone having its input connected with the output of said mixer.

The present invention relates to hearing aids or auditory prosthesis devices.

The chief object of the present invention is to provide a device of this kind which is better adapted to meet the requirements of practice, in particular which permits of making the sounds of a speech more intelligible to a deaf person.

According to the essential feature of the present invention, such a device comprises, in combination, a microphone for transforming acoustic signals into electric signals, two band-pass filter means for passing only the electric signals delivered by said microphone the frequencies of which range substantially on the one hand from 1500 to 3500 cycle per second and on the other hand from a lower limit ranging from 4500' to 6000' cycles per second to an upper limit ranging from 7000 to 8000 cycles per second, respectively, amplitude detecting means for detecting the envelope of the signals passed through said filter means, compensating means for producing an alternating voltage of a frequency ranging from 350 to 1000' cycles per second, means for modulating the voltage supplied by said compensating means in accordance with the output of said detecting means, means for amplifying the modulated voltage supplied by said modulating means, and an earphone for transforming the voltage thus amplified into acoustic signals.

According to another feature of my invention the hearing aid device further comprises, between the microphone and the earphone, filter means for passing only the electric signals from the microphone having a frequency below 1000 cycles per second and means having their output connected with the last mentioned filter means for transmitting the signals therefrom to the earphone.

A third feature according to my invention consists in providing two separate chains respectively for the signals ranging from 1500 to 3500 cycles per second and for these ranging from 4500 (or 6000) cycles per second to 7000 (or 8000) cycles per second, each of these chains comprising a filter, an amplitude detector and a modulator for mixing the detected signals with a compensating reciprocating voltage of a frequency chosen within the range from 350 to 1000 cycles per second, the compensating frequencies being different for said two chains respectively.

Preferred embodiments of my invention will be hereinafter described with reference to the appended drawing, given merely by way of example, and in which:

FIG. 1 is a diagram illustrating the principle of my invention;

FIG. 2 is a diagrammatical view of a hearing aid device made according to a first embodiment of the invention;

FIG. 3 is a view similar to FIG. 2 relating to another embodiment of the invention.

It should first be reminded that seriously deaf persons perceive only sounds of a relatively low frequency, ranging for instance from 20 to 1000 cycles per second.

In order to improve the hearing of such persons and in particular to give them information relative to the remainder of the frequency band corresponding to speech (that is to say from 1000 to 7000-8000 cycles per second), it has already been proposed to filter high-pitched sounds and to modulate with the filtered signals thus obtained a low-pitched sound of a frequency ranging within the band that can be heard by the deaf person.

But, up to the present time, on the one hand the frequency band chosen for the low-pitched sound was lower than 300 cycles per second so as to correspond to the frequency band best audible for deaf persons and possibly to provide several harmonics and on the other hand, the filtered sounds were distributed over the whole partial band in question.

Now the first of these features practically prevents the deaf person from distinguishing the loud sounding consonants from muffled consonants and the second of said features has for its effect to produce multiple signals a great number of which is unnecessary and prevents a good understanding of the message that has been formed, while complicating the apparatus.

I have found by studying the components of human speech: that on the one hand the utilization of a compensating low-pitched sound of a frequency lower than 350 cycles per second masked the low-pitched components (of frequencies ranging from 20 to 100 cycles per second) of loud sounding consonants b, as in “boat,” d, as in “day,” g, as in “go,” and chiefly z, as in “zeal” and v, as in “vote” and j as in “jazz,” so that these components can no longer be distinguished from the corresponding muffled consonants, such as p as in “pay,” t as in “two,” k as in “case” and chiefly s as in “sun,” / as in “foot” and sh as in “show,” respectively; and that, on the other hand, the only phonemes of frequencies higher than 3500 cycles per second that are useful in practice are the muffled spirants s, as in “sun” and f, as in “foot,” and the loud spirants z, as in “zeal” and v, as in “vote” an important amplitude of which corresponds to the frequencies ranging from 6000 to 8000 cycles per second and possibly phoneme y, as in “you,” the portion of which sufficient for a good intelligibility corresponds to frequencies ranging from 4500 to 5500 cycles per second.

Furthermore it was found that the range of frequencies between 1000 and 1500 cycles per second essentially comprised, as sound elements necessary for understanding speech, those relating to vowels identifiable by mere lip reading.

Starting from these facts and also from the fact that a person suffering from serious deafness perceives sounds of frequencies lower than 1000 cycles per second, I trans70

3,385,937 pose, according to the present invention, the information portions of speech having frequencies ranging on the one hand from 1500 to 3500 cycles per second and on the other hand from 4500-6000 to 7000-8000 cycles per second by replacing them by at least one low-pitched sound of constant frequency ranging from 350 to 1000 cycles per second modulated by the curve forming the envelope of the amplitudes of said portions of speech.

This transposed low-pitched sound preferably is a pure sinusoidal sound but it might also be of any other desired nature, for instance an aleatory noise, either filtered or not.

Its frequency may be chosen as low as 400 cycles per second in apparatus to be used for deaf persons which do not perceive any sound having a frequency exceeding 500 cycles per second. However preferably its frequency is chosen in the range from 800 to 1000 cycles per second.

I thus obtain, with a relatively simple apparatus, an intelligible reproduction of human speech, owing to the fact that the artificial signals on the one hand do not mask the low-pitched components of loud sounding consonants and on the other hand comprise but a minimum number of parasitic sounds.

The principle of the invention is diagrammatically illustrated by FIG. 1, where the frequencies are plotted in abscissas and the amplitudes in ordinates.

The curves A and B are the curves of response of the band-pass filters corresponding respectively to the frequency bands of from 1500 to 3500 cycles per second and from 6000 to 7500 cycles per second. The dot-anddash lines indicate the transposition of the filtered sounds from these frequency bands into a compensation sound C the constant frequency of which is within frequency band D ranging from 350 to 1000 cycles per second. FIG. 1 further shows at M the response curve of another bandpass filter corresponding to the frequency band ranging from 4500 to 7500 cycles per second, this filter being used instead of that corresponding to the frequency band from 6000 to 7500 cycles per second if it is desired to take phoneme y into account as above stated.

FIG. 2 shows a first embodiment of the invention.

This embodiment comprises:

A microphone 1 transforming the acoustic signals it receives, produced by human speech, into electric signals, A preamplifier 2 fed with these electric signals, Two band-pass filters 3_X and 3₂ adapted to pass among the preamplified signals, only those corresponding respectively to frequencies ranging from 1500 to 3500 cycles per second for 3_X and from 4500 (or 6000) to 7000 (or 8000) cycles per second, for 3₂,

System 4 for amplifying and detecting or demodulating the signals filtered by 3_X and 3₂, which may include a conventional amplitude or envelope detector that develops a varying D C. signal which follows the envelope of the input signal components,

A source 5 capable of producing a sinusoidal voltage of constant frequency ranging from 350 to 1000 cycles per second (preferably from 800 to 1000 cycles petsecond) if the deaf person for which the apparatus is intended is capable of perceiving sounds corresponding to this 800-1000 band,

An adjustable gain amplifier 6 energized both by the output of detecting system 4 and by the sinusodial voltage from source 5, so as to have said voltage modulated by the curve forming the envelope of the two portions filtered at 3_X and 3₂, of the speech received at 1,

An amplifier 7 for the voltage thus modulated, and

An earphone 8 capable of transforming the electric signals at the output of amplifier 7 into acoustic signals fed to the ear or ears of the deaf person.

The low frequency portion of the speech, which can be heard by said person may be directly received by him after amplification.

However it is preferred to filter and to amplify this last mentioned portion in the same apparatus as the other speech portion so as to send to the ear of the deaf person only the sounds that may be useful.

For this purpose, I connect with the output of pre5 amplifier 2 a low-pass filter 9, for instance the cutoff frequency of which is 1000 cycles per second and the output of this filter 9 is mixed (summed) with that of the adjustable gain amplifier 6 in a mixer (summation device) 10 the output of which is connected to the input of amplifier 7. It will be understood that the term “mix” or “mixer” as used throughout should be taken in the sense of addition or summation of signals and not in the sense of nonlinear modulation with production of side bands and the like.

FIG. 3 shows another embodiment of the invention.

In this embodiment every element the function of which is analogous to that of the corresponding element of the system of FIG. 2 is designated by the same reference character.

This second embodiment differs from the first one merely in that it uses two compensation signals instead of a single one.

Source 5 is therefore replaced by two distinct sources 5, and 5₂ each producing a sinusoidal voltage the constant 25 frequency of which ranges between 350 and 1000 cycles per scond, the frequencies of said sources and 5₂ being different from each other.

The amplifying and envelope detecting system 4 of the first embodiment is in this second embodiment replaced 30 by two systems 4, and 4₂ respectively, each of which treats one of the signals issuing from filter 3, and filter 3₂ respectively. The outputs of said systems 4, and 4₂ are applied to two adjustable gain amplifiers 6, and 6₂ further energized by voltages supplied by 5_X and 5₂ re35 spectively.

The modulated signals obtained at the outputs of amplifiers 6j and 6₂ are mixed (added) together at 11 before being further mixed (added) with the signals coming from filter 9.

⁴⁰ Of course use can be made of miniaturized circuits including semi-conductors.

With the arrangement according to the present invention, I apply to the ear of a deaf person an information comprising, in addition to the low frequency sounds of 45 speech that can be directly heard by him, a low frequency image of the high-pitched portion of this speech, this image being free from unnecessary parasitic sounds and being adapted not to mask the low frequency components of the loud sounding consonants.

In a general manner, while I have in the above description disclosed what I deem to be practical and efficient embodiments of my invention, it should be well understood that I do not wish to be limited thereto as there might be changes made in the arrangement, disposi55 tion and form of the parts without departing from the principle of my invention as comprehended within the scope of the appended claims.