US3539726A - System for storing cochlear profiles - Google Patents

Description

Nov., w, E? v. w. @OLIE SYSTEM FOR STORING COCHLEAR PROFILES Filed Dec. 16, .1968

United States Patent Office SYSTEM FOR STORING COCHLEAR PROFILES Victor W. Bolie, Stillwater, Okla., assignor to Oklahoma State University, Stillwater, Okla., a corporation of Oklahoma Filed Dec. 16, 1968, Ser. No. 783,911 Int. Cl. Gl 1/ 00 U.S. Cl. 179-1 12 Claims ABSTRACT OF DISCLOSURE This invention relates to a system for storing cochlear profiles. More particularly, the invention relates to a system for storing cochlear profiles of phonemes including a frequency dispersing circuit receiving at its input the full spectrum of frequencies making up a phoneme and providing multiple individual outputs of separate narrow frequency bands, each of the bands representing an analog signal of a discrete portion of the phoneme full spectrum and which multiple analog signals define a composite cochlear profile of the phoneme, a sampling step switchv providing sequentially.lv an output signal of each of the analog signals of the frequency dispersing circuit, an analog-to-digital converter receiving the sequential analog signal outputs of the sampling switch and converting each signal sequentially into a digital signal, a memmory storage means connected to receivev and store the digital signals, and means of automatically sequentially advancing in synchronization the sampling switch and the memory storage means until each of the discrete portions of the phoneme full spectrum is stored in digital form.

CROSS-REFERENCES This disclosure is not related to any pending United States or foreign patent application.

BACKGROUND AND lSUMMARY on THE INVENTION It has been recognized that the ultimate application of computer technology will include means of direct vocal communication of human operators with the computer devices. Only in this way can immediate and direct communication from the human'mind to the memory storage and recall devices of computer apparatus be established in a way not requiring requests and commands to be translated into code form. In addition, there are many practical reasons why it is important to be able to recognize orally transmitted intelligence.

A first step in a means of oral communication with computer and memory storage devices includes the storing of the various portions of speech of any given language in an arrangement so that such stored speech portions may be subsequently scanned for use in implementing automatic speech recognition techniques. Speech sounds may -be generally classified as voiced and unvoiced. Of the voiced sounds in the English language the vowels form the larger part. These are sounds like a, e, i, o, u, and sometimes y. In addition there are vowel-like consonants including m, n, g, W, and l, which are all characterized by a musical type sound. Other sources of oral word components are those which may be considered unvoiced sounds, that is, those which are generated without use of the vocal cords. The consonants t, p, f, s, and k are examples of unvoiced sounds in which the vocal cords are not used. Such sounds are generated by the interruption of the speakers breath outow by action of the lips, tongue, and soft palate in such a way as to cause momentary buildup of air pressure, with the sudden release of the breath pressure causing characteristic sounds l 3,539,726 Patented Nov. 10, 1970 of ythe particular consonants. Other consonants are provided by partially restricting the free flow of airyfrom the mouth producing the hissing sounds characteristic of the letters s, ch, f, h, th, soft c, v, and z. Such are called sibilants. Other consonants, such as b, d, and g, are formed by the generation of a subdued buzz combined with the interruption of the breath ow by the action of the lips, tongue, and soft palate. In addition, there are voiced sibilants such as the change of an s sound to a z sound when the vocal cords are brought into place.

Language is made up of words, and words are made up of a combination of small sound increments, such increments being termed a phoneme There is some area of disagreement as to the total number of phonemes utilized in any given language. For instance, inthe English language the number of phonemes is said by some to be as few as twenty-live. Others say that as many as forty different phonemes are required to formulate the words of the full range of the commonly spoken English language. In any event it is recognized that a relatively few sounds (25 to 40) are all that is required to form words and that the combination of these sounds into words and the words into sentences results in the tools by which all human intelligence is vocally transmitted and received. Generally speaking the ear of a human being functions in conjunction with the logic system of the brain to recognize each phoneme of sound and thereafter to recognize the combination of phonemes which make up words. Su-bsequently a higher order of intelligent' activity recognizes combinations of words as expressions of thought. Thus it can be seen that oral communication begins with a number of phonemes in the range of approximately 25 to 40. Any computer arrangements for utilizing the spoken language must therefore begin with means of storing for subsequent scanning and recognizing the phonemes which make up human speech.

It has also ,been learned that the ear apparently functions to recognize phonemes by an arrangement in which the cochlea distinguishes the separate but sometimes overlapping frequency bands making up each phoneme of sound.

In response to this knowledge others have provided simulators for mechanical action of the human ear which i provides electrical signal outputs of space-time activity "i patterns corresponding to those communicated from the inner ear through the acoustic nerveto the brain as a result of sound stimulation. Generally speaking, such devices may be termed an analog ear or a frequency dispersive audio delay line, or a frequency dispersive circuit. By any of the names to which they are referred such circuit arrangements are primarily spectrum analyzers, that is, each circuit receives an audio reproduction of the full frequency range of a phoneme', such as from a microphone and audio amplifier, and provides, at a series of output terminals, narrow bands of discrete portions of the composite frequency spectrum. Such bands may be overlapping or non-overlapping, however, in the preferred arrangement the bands are overlapping. The total of the composite signal output from the frequency dispersing circuit defines an envelope which is characteristic of the phoneme received. By storing the envelopes of the various phonemes in digital form in a memory storage device, the device may be subsequently scanned for comparison of envelopes to identify phonemes. This invention relates to a system for storing the envelopes of discrete signal outputs from a frequency dispersive circuit, which may be termed a cochlear profile, in a memory storage means.

It is therefore an object of this invention to provide a system for storing cochlear profiles of sound phonemes.

More particularly, an object of this invention is to provide a system for establishing cochlear profiles of phonemes and to code such profiles in digital form into a memory device.

Still more particularly, an object of this invention is to provide a system wherein a phoneme may be introduced into a sound receiving means and, by a single switch actuation, a series of discrete analog signals forming a composite envelope characteristic of the phoneme is sequentially converted from analog-to-digital form and stored into a memory device, the memory device and means of sequentially selecting discrete portions of the composite signal being automatically advanced in synchronization.

These as well as many other objects of this invention will be understood by referring to the description and claims taken in conjunction with the drawings.

DESCRIPTION OF THE VIEW The drawing is a circuit diagram of a system for storing cochlear profiles illustrating an embodiment of the invention.

As previously stated the invention relates to a system for storing cochlear profiles of phonemes. When a profile of a phoneme is to be stored a sound, such as from a speaker, is directed into a microphone 10. As an example, if it is desired to store in digital form a profile of the phoneme ah this sound is expressed into the microphone 10. The sound is amplified in amplifier 12 and fed into a frequency dispersive audio delay line 14 which may also be described as a frequency dispersive circuit. The frequency dispersive audio delay line 14 provides a plurality of separate outputs 16 each responsive to pass a separate narrow frequency band. The sum of the frequency bands passed through outputs 16 provides a composite envelope of substantially the full spectrum of the phoneme to be stored. While the number of frequency bands delivered out of the frequency dispersive audio delay line 14 may vary, a typical arrangement for practicing the invention includes thirty-two output points 16 each of which pass only those frequency components in a selected narrow band, the width of the band being typically about ten to twenty percent of the center frequency of the band. The frequency dispersive audio delay line 14 includes a series of rectifers and RC smoothing filters which provide a corresponding series of measures of the intensities of the various frequency components of the speech signal. The frequency-selective characteristics of the audio delay line may be achieved by means of a sequence of low-Q series- RLC-shunts interleaved with a corresponding sequence of low-Q series-RL-segments with the inductances of both the series and shunt inductors having numerical values proportional to the antilogarithm of their normalized distances from the receiving end of the line, and with the resonant frequency of each series-RLC-shunt being made inversely proportional to the antilogarithm of its normalized distance from the receiving end of the line. The two dimensional patterns which would be seen from a plot of a tapped-voltage amplitude versus the tap-number can be designated as a cochlear profile in analog with the prole of vibration-amplitudes existing on the basilar membrane of the inner ear in response to a composite sound wave impinging on the eardrurn. The thirty-two outputs, only three of which are shown to conserve space on the drawing, are fed into the input of an addressable sampling switch 18. In the illustrated embodiment the addressable sampling switch 18 has thirty-two analog signal inputs 18A connected to the thirty-two outputs 16 of the frequency dispersive audio delay line. In addition, the addressable sampling switch 18 has a multi-bit address input 18B by which the switch is sequentially addressed to each of said inputs 18A. In the illustrated arrangement the address input is made a five-bit binary signal, to provide any one of thirty-two different addresses. The sampling switch 18 further has an analog signal output 18C to which each of the frequency dispersive audio delay line outputs 16 is sequentially connected. The sampling switch in addition includes a terminator signal output 18D at which a signal in the form of a single digital pulse appears when the sampling switch has completed a sweep of each of the frequency inputs 18A.

The analog signal output 18C of the sampling switch 18 is connected through a logarithmic transductor 20 to an analog-to-digital converter 22. Since the logarithmic transductor 20 is optional it can be said that the Isampling switch analog signal output 18C is connected to analog signal input 22A of an analog-to-digital converter 22. The analog-to-digital converter 22 further includes an initiator signal input 22B in the form of a single digital pulse, and a multichannel binary signal output collectively identified as 22C. In the illustrated embodiment the analog-to-digital converter 22 is a six-bit converter so there are six output circuits 22C as illustrated. In addition the analog-todigital converter 22 includes a terminator signal output 22D at which a signal in the form of a single digital pulse appears each time the analog-to-digital converter has completed the conversion started by the initiator pulse applied to terminal 22B.

The multichannel binary signal output 22C from analog-to-digital converter 22 is fed to the multichannel binary signal input 24A of a write-gate array 24. The write-gate array 24 further includes a multichannel binary signal output collectively identified as 24B and a writecommand input 24C. The write-command input 24C is connected to the terminator signal output 22D of analogto-digital converter 22. The cochlear profile is stored in a memory storage means 26 which may be in the form such as a memory drum. The memory storage means 26 includes a multichannel signal input 2'6A which is connected to the write-gate array output 24B. The memory storage means 26 includes an address control matrix portion 26B having a multibit address input collectively identified as 26B. In the illustrated arrangement the address control matrix of the memory storage means has a five-bit address signal input 26C plus a six-bit address-advance signal input 26D totaling an eleven-bit address control matrix.

A multibit counter 28 is provided having a multichannel output collectively identified as 28A which is connected both to the memory storage means address signal input 26C and to the multibit address input 18B of the addressable sampling switch 18. In addition the multibit counter 28 includes a count stepping signal input 28B. In the illustrated arrangement the counter 28 is a ve-bit counter corresponding to the five-bit input 18B of the addressable sampling switch 18 and the ve-bit input 26C address control matrix of memory storage means 26. Multibit counter 28 also furnishes a terminator signal output 28C in the form of a single digital pulse which appears each time counter 28 has been stepped through a full sweep of its maximum available steps.

A second multibit counter 48 is provided having a multichannel output which furnishes the address-advance signal 26D of memory storage means 26 and also having a stepping signal input 28C received from the terminator output of multibit counter 28. In the illustrated arrangement counter 48 is a six-bit counter. The total eleven-bit input to address control matrix 26B thus allows the storage in memory storage means 26 of a total of 26:64 cochlear profiles each consisting of a sequence of 25:32 numerically represented data samples obtained from the frequency dispersive audio delay line 14.

Storage of a cochlear profile is initiated when storing signal switch 30 is depressed. By means of a battery 32 a voltage pulse is transmitted through capacitor 34 and diodes 36 and 46 to the initiator signal input 22B of the analog-to-digital converter 22. In addition, a flip-Hop circuit is provided having a set signal input 38A, a reset signal input 38B, and a control signal output 38C. The storing signal passing through diode 36 is applied to the set signal input 38A of the flip-flop circuit 38. One output signal 38C of flip-flop circuit 38 is fed as a control signal 40A to gate circuit 40, the digital pulse input 40C of which is transmitted as a digital pulse output control signal 40B only during the time interval beginning when flip-flop 38 is set by input signal 38A and ending when flip-flop 38 is reset by input signal 38B. The output pulse signal 40B from gate 40 is connected to the initiator signal input 22B of analog-to-digital converterl 22. Diode 46 prevents signal 40B from affecting flip-fiop 38. The terminator signal output 22D of the analog-todigital converter 22 not only lfurnishes the write-command signal 24C but also is connected by way of a first delay circuit 42 and a second delay circuit 44 (the function of 'which will be described subsequently) to the signal input 40C of gate 40. The

output pulse signal of the first delay circuit 42 also furnishes the input signal 28B to multibit counter 28.

OPERATION When it is desired to store a cochlear profile of a specific sound, such as a phoneme of speech, the sound, such as an ah sound, is directed into the microphone 10. The sound is amplified in speech amplifier 12 and is fed to the frequency dispersive audio delay line 14. There the sound is separated into separate analog signal outputs, each representing an analog signal of a discrete portion of the phoneme full spectrum. The multiple analog signals appearing at the output 16 of the frequency dispersive audio delay line 14 together define a composite cochlear profile of the phoneme. Each of these analog signals is fed to sampling switch 18. Assuming counters 28 and 48 are first initialized to their all-zero conditions, the analog signal of the first of the analog outputs of the audio dispersive line is fed by way of logarithmic transductor 20 to the analog input 22A of analog-to-digital converter 22. While the phoneme is directed into microphone 10 its cochlear profile can be stored by depressing switch 30. This not only sends a signal by way of diode 46 to the initiator signal input of analog-to-digital converter 22, but also enables gate 40 to transmitby reason of fiip-flop '38 being simultaneously set. By reason of such initiator signal input the analog signal representing the first of the thirty-two outputs of frequency dispersive audio delay line 14 is converted from an analog to a digital signal. Such digital signal appears across multibit output 22C of the analog-to-digital converter and at the input of the write-gate array 24. Upon completion of the conversion a termination signal appears at analogto-digit converter termination signal output;- 122D which is ifed to the write signal input 24C of the write-gate array. This signal causes a transfer of the multibit digital signal into the particular location in memory storage 26 determined by the initialized all-zero conditions Iof counters 28 and 48. Thus, a digital signal is stored in memory storage means 26 representing the analog signal appearing at the first of the thirty-two outputs of the frequency dispersing audio delay line. Shortly thereafter the termination signal at 22D is fed through delay 42 to the count stepping signal input 28B of counter 28. This advances the counter by one step which not only advances the sampling switch 18 so that the analog signal :from the second output of the frequency dispersive audio delay line 14 is fed to the analog signal output 18C but also advances the address of memory storage means 26 to the next location. First delay 42 is provided to insure that the digital signal from analog-to-digital converter 22 and write-gate array 24 appearing at digital signal inputs 26A is stored in the memory storage means 26 before the memory storage means is advanced to its next address by means of counter 28. In addition, the signal passing through delay circuit 42 passes through a second delay circuit 44 and to the signal input 40C of gate circuit 40. As long as the flip-flop circuit 28 is in the set position gate circuit 40 is biased to conduct so that such input signal 40C is passed through a gate output signal 40B to the initiation signal input 22B of the analog-to-digital converter 22. This then converts the analog signal from the second tap of the frequency dispersive audio delay line 14 and upon conversion and appearance of the terminator signal at 22D the write-gate array 24 is signaled to transmit the digital signal to storage means 26. The function of second delay 44 is to insure that the sampling switch 18 and the address control matrix 26B have-each advanced to their next positions following the signal at the output of first delay 42 before the signal 40C passes through gate 40 for the next conversion.

This sequence continues, automatically stepping the sampling switch 1 8 through thirty-two separate positions at each of which ppsitions the analog signal appearing at a different tap of the frequency dispersive audio delay line is converted in to a digital signal and stored. Upon termination of the thirty-second storage a reset signal appears at signal output 18D of sampling switch 18. This signal resets flip-flop circuit 38 so that the signal appearing at 38C and 40A'v turns gate circuit 40 to the otfcondition, in which condition no additional signal inputsl at 40C can pass therethrough to initiate further conversions by converter 22. Thus, a sweep of the full thirty-two taps of the frequency dispersive audio delay line is automatically completed upon the pressing at one time ofj'the storage signal switch 30. At the same time five-bit counter provides a count termination signal at output 28C to step six-bit counter 48 which in turn advances the address of the control "matrix for storing the next cochlear profile.

When it is desired to store the next cochlear profile such as the phoneme the sound is directed into microphone 10 and storage signal switch 30 is pushed. The automatic sequence is then repeated in which thirty-two analog signals are converted into digital form and stored in the next sequence of thirty-two locations in memory storage 26.

The cochlear profile described by the amplitude of the thirty-two output voltages of the frequency dispersive audio delay line hasua pattern shape which responds only slowly to changesv in speech sounds, provided thatjthe time constants of lthe- RC filters supplied by the associated rectifiers are suitably long. The switching delays of the various components of the system are of such brevity that the total time interval in which the flip-flop 38 maintains the gate 40 in the enabled condition is less than y1D0 milliseconds so that'the cochlear profile remains essentially stationary during the time required to read yand store the thirty-two corresponding voltage amplitudes.

The logarithmic transductor 20 is optional in the circuit, the purpose b eing to convert the analog signal output of each of the `taps of the frequency dispersive audio delay lines into ap-. .voltage proportional to the logarithm of each such signals so that the storage signal more nearly corresponds with the aural receptiveness of the human ear. It can be seen that the system functions the same whether or not the logarithmic transductor 20 is utilized.

The delay circuits 42 and 44 may each be in the form of a monostable multivibrator followed by simple RC differentiator circuit and diode to produce the appropriately delayed pulse.

The airborne sound directed to microphone 10 may include, in addition to aural phonemes, musical tones, underwater sounds, noises of burning fires, sound of approaching aircraft, or sounds from any other source. Once the various cochlear profiles are loaded into the datamemory, they can be recalled whenever needed as reference signals for more detailed analysis. In addition, such' storage profiles may be utilized in computer devices for automatic pattern recognition operation, for sound operated control systems, and multitude of other uses requiring rapid access to a set of cochlear profiles.

In the invention certain numerical relationships are shown such as counter 28 is illustrated as being a tive-bit counter, analog-to-digital converter is illustrated as being a six-bit converter, and so forth. This is by way of exemplilication only and the invention is not limited to such illustrated numerical embodiments.

It is understood that the invention has been described with a certain degree of particularity and that many changes may be made in the details of construction and the arrangement of components without departing from the spirit and scope of this disclosure. It is understood that the invention is not to be limited by the abstract herein, nor the circuit arrangement which is shown to exemplify one embodiment of the invention, but the invention is to be limited only by the scope of the attached claims, including the full range of equivalency to which each element thereof is entitled.

What is claimed:

1. A system for storing cochlear profiles of phonemes comprising:

a frequency dispersive circuit receiving at its input the full spectrum of frequencies making up a phoneme and providing multiple individual outputs of separate frequency bands, each of which bands is represented by a rectifier-derived signal voltage analog of the intensity of a discrete portion of said phoneme full spectrum and which multiple analog signals dene a composite cochlear profile of the phoneme;

a sampling step switch providing'sequentially an output signal of each of said analog signals from said frequency dispersive circuit;

an analog-to-digital converter receiving the sequential analog signal outputs of said sampling switch and converting each such signals sequentially into a digital signal;

a memory storage means connected to receive and store said digital signals from said analog-to-digital converter; and

means of automatically sequentially advancing in synclironization said sampling switch and said memory storage means until each said discrete portion of said phoneme full spectrum is stored in digital form.

wherein said analog-to-digital converter includes av termination signal output connected to said multibit counter step signal input whereby upon completion of each conversion by said analog-to-digital converter a stepping signal is provided to said multibit counter.

3. A system for storing cochlear profiles of phonemes according to claim 1 in which said analog-to-digital converter includes an initiator signal input and including:

a gate means having a signal output connected to said initiator signal input of said analog-to-digital converter and having a signal input connected to said terminator signal output of said analog-to-digital converter, and further having a control input and including means of providing an enabling control signal to said gate during said automatically sequentially advancing of said sampling switch and said memory storage means as said discrete portions of said phoneme full spectrum is stored and of providing a disabling control signal to said gate to terminate said automatically sequentially advancing of said sampling switch and said memory storage means upon the completion of the storage of all of the said discrete portions of the full spectrum of a phoneme.

4. A new system for storing cochlear profiles of phonemes according to claim 3 wherein said sampling switch provides reset signal output when said sampling switch has been sequentially stepped through the number of steps equal to said individual outputs of said frequency 8 dispersive circuit and wherein said means of providing said enabling control signal and said disabling control sign to said gate include:

a ip-op circuit having a set signal input, a reset signal input connected to said reset signal output of said sampling switch, and a control signal output connected to said gate means control input, said flip-Hop circuit providing an enabling signal to said gate means when a set signal is received and providing a disabling signal to said gate means when a reset signal is received from said sampling switch; and including a storing switch means providing, upon actuation, a storing signal to said flop-flop circuit set signal input.

5. A system for storing cochlear proles according to claim 4 including:

a delay means in said terminator signal output of said analog-to-digital converter whereby said terminator signal to said multibit counter which in turn steps said sampling switch and addresses said memory storage means is delayed to insure that the digital signal to said memory storage means is first stored before said counter is stepped.

6. A system of storing cochlear profiles according to claim 4 including:

a delay means in said gate circuit initiator signal input whereby the receipt of an initiator signal input to said analog-to-digital converter is delayed after the storing of each converted analog signal before the storing of the subsequent signal to permit sufficient time for said multi-bit counter to step said sampling switch and to address said memory storage control matrix.

7. A system of storing cochlear profiles according to claim 4 including;

a rst delay means in said terminator signal output of said analog-to-digital converter whereby said terminator signal to said multibit counter which in turn steps said sampling switch and addresses said memory storage means is delayed to insure that the digital signal to said memory storage means is first stored before said counter is stepped; and

a second delay means in said gate circuit signal input whereby the receipt of an initiator signal input to said analog-to'digital converter is delayed after the storing of each converted analog signal before the storing of the subsequent signal to permit sufficient time for said multi-bit counter to step said sampling switch and to address said memory storage control matrix.

8. A system of storing cochlear profiles of phonemes of sound comprising:

a microphone providing means for receiving phonemes of sound;

an amplifier having an input and an output7 the input being connected to said microphone;

a frequency dispersive circuit having an input connected to the output of said amplifier, and having a plurality of separate outputs each responsive to the intensity of a component of the output of said amplifier as passed in a separate frequency band, each represented as an analog signal, the sum of the analog signals covering substantially the full spectrum of each of the phonemes to be stored;

an addressable sampling switch having a plurality of analog signal inputs each connected to one of said outputs of said frequency dispersive circuit, and having a multibit address input by which the switch iS sequently stepped through each of said inputs, the sampling switch further having an analog signal output to which each of said outputs of said frequency dispersive circuit is sequentially connected, and the sampling switch further having a reset signal output at which a signal appears when the switch has completed a stepped sweep of each of said analog signal inputs; an analog-to-digital converter having an analog input, an initiator signal input, a terminator signal output, and a multichannel digital signal output, the analog input being connected to said analog signal output of said sampling switch; v

a write-gate array having a multichannel digital signal input connected to said multichannel digital signal output of said analog-to-digital converter, and having a multichannel digital signal output and a Write signal input;

a memory storage means having a multichannel digital signal connected to said multichannel digital signal output of said write-gate array, and including an address control matrix portion having a multibit address input and an address advance signal input;

a multibit counter having a multichannel output connected to said multibit address input of said memory storage address control matrix and to said multibit address input of said addressable sampling switch, the counter having a count termination signal output connected to said address advance signal input of said memory storage address control matrix, and having a count stepping signal input;

a storing signal switch providing, when actuated, a

Storing signal pulse, said storing `signal switch connected to said initiator signal input of said analogto-digital converter;

a Hip-flop circuit having a set signal input connected to said storing signal switch, a reset signal input connected to said reset signal output of said sampling switch, and a control signal output; and

a gate circuit having a signal output connected to said initiator signal input of said analog-to-digital converter, and having a signal input connected to said terminator signal output of said analog-to-digital converter and further having a gate control input connected to said control output of said ip-op circuit, said Hip-flop circuit providing an ena-bling control signal to said gate circuit upon receipt of a storing signal and maintaining said enabling control signal until a reset signal is applied to said ip-flop reset signal input from said sampling switch upon the completion of the stepped sweep of each of said analog signal inputs, said count stepping signal input of said multibit counter being connected to said terminator signal outptu of said analog-to-digital converter whereby upon the completion of aconversion said counter delivers a next address signal to said memory storage address control matrix to automatically advance said memory storage to re ceive and store, in digital form, each of said separate analog signals at said plurality of outputs of said frequency dispersive circuit upon the initial actuation of said storing signal switch.,

9. A system for storing cochlear profiles according to claim 8 including:

a delay means in said terminator signal output of said analog-to-digital converter whereby said terminator signal which is communicated to said multibit counter which in turn advances the address of said sampling switch and advances the address of said memory storage means is delayed to insure that said signal to said write-gate array has rst stored the digital conversion of the previous analog signal.

10. A system of storing cochlear profiles according to claim 8 including:

a delay means in said gate circuit initiator signal input whereby the receipt of an initiator signal input to said analog-to-digital converter is delayed after the storing of each converted analog signal to permit suflicient time for said multibit counter `to step said sampling switch and to address said memory storage control matrix.

11. A system of storing cochlear profiles according to claim 8 including: f

a rst delay means in said terminator signal output of said analog-to-digital converter whereby said terminator signal to said multi-bit counter which in turn steps said sampling switch and addresses said memory storage means is delayed to insure that the signal to said write-gate array has first stored the digital signal of the previous analog signal; and

a second delay means in said gate circuit initiator signal input whereby the receipt of an initiator signal input to said analog-to-digital converter is delayed after the storing of each converted analog signal to permit sufficient time for said multibit counter to step said sampling switch and to address said memory storage control matrix.

12. A system for storing cochlear profiles according to claim 8 including a logarithmic transductor circuit between said sampling switch analog signal output and said analog-to-digital converter analog signal input whereby the analog signal of each output position of said frequency dispersive circuit is logarithmically scaled prior to conversion and storage.

References Cited UNITED STATES PATENTS 3,294,909 12/ 1966 Caldwell et al, 179-1 3,469,034 9/1969I Stewart 179-1 KATHLEEN H. CLAFFY, Primary Examiner C. W. JIRAUCH, Assistant Examiner i U.S. Cl. XR 324--77

Images