BACKGROUND OF THE INVENTION A practical problem has prevented the widespread use and availability of high fidelity hearing aids. Specifically, dampers, which are used to smooth the frequency response, often needed to be near the tip of the hearing aid outlet at a point where they are easily clogged with ear canal wax.

As a result, hearing aid manufacturers stopped using dampers near the eartip, and unpleasant peaks in the frequency response became commonplace. This problem was recognized by Killion et al. in U. S. Patent No. 5,812, 679 issued September 22, 1998 entitled"Electronic Damper Circuit for a Hearing Aid and Method of Using the Same"and in U. S. Patent No. 6,047, 075 issued April 4,2000 entitled"Damper for Hearing Aid. "These patents describe the use of electronic filtering to substitute for the acoustic damper. One of the realizations at the time was that by making the filter programmable, it could be adjusted to accommodate the different peak frequencies that are obtained when different lengths of tubing are used with the earphone to accommodate different lengths of ear canals and earmolds.

Although the electronic damping of Killion et al. was a substantial contribution, we now have realized additional problems in making a completely high fidelity hearing aid. Even though the response with different receiver"plumbing" arrangements can be adequately damped, the finished frequency response may not produce a full fidelity hearing aid. In other cases, the model of receiver that is chosen on the basis of power handling or other considerations, may have its peak frequency placed well below 2 kHz. In this situation, according to the prior art, a high fidelity response becomes nearly impossible, regardless of the choice of damping or plumbing.

To explain, a full fidelity hearing aid generally must have a frequency response matching one of the"CORFIG"responses described by Killion and Monser (CORFIG: Coupler Response for Flat Insertion Gain by Mead C. Killion and Edward L. Monser, IV, in Acoustical Factors Affecting Hearing Aid Perfonnance, Studebaker, G. A. and Hochberg, I. , eds. , pgs. 149-168,1980) (Appendix A) and by Killion and Revit (CORFIG and GIFROC: Real Ear to Coupler and Back by Mead C.

Killion and Lawrence Revit in Acoustical Factors Affecting Hearing Aid Perfonnafzce (2nd Ed.), Studebalcer, G. A. and Hochberg, I. , eds. , pgs. 65-86,1993). The adequately damped peak may, in a particular case, be at a different frequency than the approximately 2.5 kHz frequency of the open ear. In order to have a full fidelity frequency response, it may be necessary to replicate the response that would normally occur at the eardrum without a hearing aid in place. This response is described in the "CORFIG"curve for the type of hearing aid in question (behind-the-ear, in-the-ear, canal aid or completely-in-the-canal aid) as described in Appendix A and in Killion and Revit (mentioned above).

In addition, the microphone response often rolls off above 3 or 4 kHz, making it desirable to further equalize the microphone. This was recognized by Killion et al. in the early application notes for the"K-AMP"integrated circuit chip (as described in ER-101-28D Data Sheet dated 92/7/2) (Appendix B). Capacitor C2S, as described in Note 2 of Appendix B, provided a high frequency boost to compensate for the loss of high frequency response in typical microphones, just as capacitor CHFB produced a high frequency boost to compensate for the loss of high frequency output in typical receivers (Appendix B). A problem arises because microphones must sometimes be mounted at a distance behind the faceplate of the hearing aid and connected to the opening in the faceplate with a section of tubing. Different hearing aids in the same nominal family of hearing aids, therefore, may require different amounts of high frequency correction for the microphone and/or receiver.

Further limitations and disadvantages of conventional and traditional approaches will become apparent to one of skill in the art, through comparison of such systems with the present invention as set forth in the remainder of the present application with reference to the drawings.

BRIEF SUMMARY OF THE INVENTION Aspects of the present invention are found in a hearing aid that has a microphone, a filter with a response curve defined to be one of the CORFIG response curves, and a receiver. The microphone converts the received sound energy into an electrical signal that is then sent to the filter. The filter modifies the electrical signal to achieve a hearing aid frequency response that corresponds to a CORFIG response curve, and the receiver converts the modified electrical signal back into sound.

In one embodiment, the response curve of the filter can be defined to include the high frequency boost needed to compensate for the high-frequency roll-off of the microphone response. In a further embodiment, the filter characteristics could include equalization to modify the response curve of a directional microphone into that of a non-directional microphone. In an additional embodiment, the hearing aid can be programmed to apply filtering to remove one or more peaks in the response curve of the microphone.

In yet another embodiment, the filter could be configured to apply a bandsplitting filter, a set of compression amplifiers, and a combiner, in order to compensate for the frequency-dependent hearing loss of the user. The bandsplitting filter segments the spectral content of the sound received by the microphone into a number of sub-bands. A separate compression amplifier processes each of those sub- bands, and the outputs of the compression amplifiers are then combined.

An embodiment of the present invention may also have the filter programmed in order to reduce one or more peaks in the response curve of the receiver. An additional embodiment could have the filter arranged to provide the high-frequency boost needed to compensate for the high-frequency roll-off of the receiver. An embodiment of the present invention may also allow the characteristics of the filter to be programmed after completion of manufacture of the hearing aid.

In another embodiment of the present invention, the electrical signal from the microphone is transformed into a digital representation by an analog-to-digital converter, and the filtering is performed using the digital representation. The result of the filtering operation is converted into a second electrical signal by a digital-to- analog converter.

An additional aspect of the present invention relates to a method of programming a digital hearing aid. The method illustrated includes the steps of equalizing one or more of the response characteristics of the microphone and receiver, removing at least one peak in the response curve of the microphone, removing at least one peak in the response curve of the receiver, and modifying the resulting frequency response characteristics of the digital hearing aid to correspond to the CORFIG response curve for the type of hearing aid being programmed. The method may also include programming the hearing aid frequency response curve to compensate for the frequency-dependent hearing loss of the hearing aid user.

Another aspect of the present invention relates to the operation of a digital hearing aid. A digital hearing aid operating according to one embodiment of the present invention converts received sound into an electrical signal, modifies the electrical signal in order to produce a CORFIG response for the particular type of hearing aid, converts the modified electrical signal into sound, and transmits the resulting sound into the ear canal of the hearing aid user.

In one embodiment, the present invention may operate so as to further modify the electrical signal to equalize the effects of the microphone and receiver. In a further embodiment, the operation of the digital hearing aid may effect a frequency response curve in order to compensate for the frequency-dependent hearing loss of the hearing aid user.

An additional embodiment of a method of operating a digital hearing aid comprises, for example, the steps of receiving sound from a sound field, generating a desired first frequency response, and subsequently modifying the first frequency response to achieve a desired second frequency response, in order that the desired second frequency response is perceived by the hearing aid user to be the desired first frequency response.

In one embodiment of a method of operating a digital hearing aid according to the present invention, the digital hearing aid generates a desired first frequency response that is an approximately flat frequency response. In still another embodiment of a method of operation, a digital hearing aid of a particular type modifies the desired first frequency response so that desired second frequency response is the CORFIG frequency response corresponding to the type of hearing aid being operated. Yet another embodiment of operation according to the present invention further modifies the desired first frequency response so that the desired second frequency response also compensates for the frequency-dependent hearing loss of the hearing aid user.

These and other advantages and novel features of the present invention, as well as details of an illustrated embodiment thereof will be more fully understood from the following description and drawings.

BRIEF DESCRIPTION OF SEVERAL VIEWS OF THE DRAWINGS Fig. 1 shows a block diagram illustrating one embodiment of the present invention.

Fig. 2 shows a block diagram of the programmable digital circuit of Fig. 1, in accordance with one embodiment of the present invention.

Fig. 3 illustrates an example of an uncorrected frequency response curve obtained with an undamped hearing aid, and a frequency response curve incorporating basic correction.

Fig. 4 shows the composite frequency response characteristics of a set of filters adjusted to remove the two peaks in the uncorrected response curve shown in Fig. 3, in accordance with the present invention.

Fig. 5 illustrates the desired flat frequency response resulting from the application of the filters with response characteristics shown in Fig. 4 upon the frequency response characteristics of the undamped hearing aid as shown in Fig. 3, in accordance with the present invention.

Fig. 6 shows the hearing aid frequency response after the application of a CORFIG response curve to the flat response characteristic shown in Fig. 5, resulting in hearing aid performance in accordance with the present invention.

Fig. 7 is a flow diagram illustrating a method of programming a hearing aid in accordance with one embodiment of the present invention.

Fig. 8 is a flow diagram showing hearing aid operation for one embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION One embodiment of the present invention comprises a method that uses seven "programmable bi-quad"filters in a particular digital hearing aid circuit. Two of the filters may provide, for example, peak damping as described in U. S. Patent No.

5,812, 679 and U. S. Patent No. 6,047, 075, which patents are hereby incorporated herein by reference in their entirety. These patents generally describe, for example, a shelving filter and a microphone compensation filter. Unlike the previous approach, however, in one embodiment of the present invention, filters (four, for example) are used to completely flatten the response of the hearing aid. Thus, it no longer matters if the peak frequency was at 2 kHz instead of at 2.5 kHz ; the peak is completely flattened. Two additional filters, for example, are then used to reinsert the desired "CORFIG"frequency response shaping. The entire tuning process can be automated or is readily accomplished even without automatic computer selection of all of the filter characteristics, by watching an ongoing frequency response such as obtained from the Frye 6500 hearing aid test box"composite"signal, and adjusting it to a straight line on the computer screen. After that has been accomplished, the preprogrammed"CORFIG"equalization corresponding to the type of hearing aid being built is inserted. Alternately, the"CORFIG"responses can be built in the computer program, and the second step can be another flattening step resulting in a straight line on the computer screen where the proper hearing aid frequency response has the required"CORFIG"subtracted from it before presentation, meaning that a perfectly flat line would represent a hearing aid that had exactly the right"CORFIG" response.

Fig. 1 shows one embodiment of a hearing aid in accordance with the present invention. Hearing aid 100 may be any type of hearing aid (e. g. , BTE, ITE, ITC, or CIC. ) Hearing aid 100 comprises a microphone 101, a programmable digital circuit 103, a receiver 105, an optional microphone sound tube 107 and an optional receiver sound tube 109. Sound is picked up from sound tube 107 by the microphone 101, and transduced into an electrical signal. The electrical signal is fed to programmable digital circuit 103, and the output of programmable digital circuit 103 is fed to the receiver 105. The receiver 105 transduces the signal back into sound, which is then is fed into the ear canal via an optional receiver sound tube 109.

Fig. 2 illustrates a block diagram of one embodiment of the programmable digital circuit 103 of Fig. 1. The output of the microphone (e. g. , microphone 101 of Fig 1) is fed to an analog to digital converter 201, the output of which is fed to filter 203, the first of five bi-quad filters. Filter 203 comprises, for example, a 20 Hz cut off frequency high pass filter for dc blocking and a 16 kHz boost. Filter 205 comprises, for example, a broad notch filter for damping a microphone response peak.

Filter 207 (optional) inserts a low frequency gain boost to equalize a directional microphone response to a non-directional microphone response. Filter 209 comprises, for example, a broad notch filter for removing or damping a primary receiver response peak. Filter 211 likewise comprises a broad notch filter to remove or damp a second receiver response peak. The output of these filters is fed to a band-splitting filter 213, which operates in conjunction with programmable compressors 215,217, 219 and 221. The programmable compression amplifiers 215,217, 219 and 221 are programmed to act to compensate for the frequency-dependent hearing loss of the person to be fitted with the hearing aid. A volume control 223 operates in a normal manner to adjust the gain of the hearing aid.

In the embodiment of Fig. 2, two additional bi-quad filters follow the summation of the four compressor channels. Filter 225 inserts the desired frequency response peak according to the appropriate CORFIG curve, and filter 227 produces the desired high frequency response boost to compensate for the high frequency roll- off of the receiver. Filter 227 may include additional response compensation to assist in meeting the exact CORFIG curve depending on hearing aid model type (i. e. , ITE, ITC, etc.).

A portion of Fig. 2, namely those filters that adjust for the response of the receiver, microphone, plumbing, etc. (e. g. , filters 203,205, 207,209, 211,225 and 227) may be programmed during the manufacturing process, and may be set so they are not modifiable by a hearing aid dispenser. In some cases, however, it may be desirable to allow a hearing aid dispenser to modify this portion to match the external acoustic characteristics of an individual ear (i. e. , an individually measured CORFIG).

Another portion of Fig. 2, namely splitting filter 213 and programmable compressors 215,217, 219 and 221, may be programmed after manufacture by the hearing aid dispenser, depending on the characteristics of the hearing loss of a patient.

Fig. 3 illustrates two frequency responses obtained in an undamped hearing aid. Curve 301 shows the"as is"frequency response obtained without correction, and curve 303 shows the"as is"frequency response obtained using the digital hearing aid amplifier 103 with only the basic correction of filter 227. The amplifier in this case is a Gennum GB3210. Filter 227 may comprise, for example, a digital version of the circuit shown in Appendix B.

Fig. 4 shows the response characteristics produced by the combination of filters 209 and 211 after they have been adjusted for the undamped peaks in curve 303 of Fig. 3. Curve 401 of Fig. 4 illustrates two notches, namely, notch 403 that results from application of filter 209, and notch 405 that results from application of filter 211.

Filters 209 and 211 may comprise, for example, filters as described in incorporated U. S. Patent No. 5, 812, 679 and U. S. Patent No. 6,047, 075.

Fig. 5 shows the same hearing aid of curve 303 of Fig. 3 after the response has been flattened using filters 209 and 211, as shown in Fig. 4, and filters 203,205 and 207 have been applied. Fig. 5 illustrates a desired flat response.

Fig. 6 shows the same hearing aid of Fig. 5 after filter 225 has been used to reintroduce a"CORFIG"response. Fig. 6 illustrates a frequency response of a hearing aid in accordance with the present invention, such that a listener perceives a high fidelity sound free of the unnatural coloration frequently found in present day digital and analog hearing aids. In other words, the hearing aid produces the response of Fig. 6, but the listener perceives the response of Fig. 5. Fig. 5 thus illustrates the effective frequency response as perceived by the listener, and shows nearly perfect fidelity. To our knowledge, no hearing aid has ever had this high fidelity a frequency response.

Fig. 7 is a flow diagram of a method of programming a hearing aid in accordance with one embodiment of the present invention. With this method, at block 700, the hearing aid audio response is modified for gross frequency response characteristics such as the high-frequency roll-off and directionality of the microphone and high-frequency characteristics of the receiver. At the next block, 702, additional compensation is provided to reduce or damp a peak that may be present in the microphone response due to the microphone itself, or the mechanical components coupling the sound energy to the microphone. At block 704, programming is provided to allow peaks in the response curve of the receiver to be minimized. At the last block in the illustrated embodiment, 706, hearing aid performance is modified to apply the CORFIG response curve for the type of hearing aid being programmed. This example of a method of adjusting the operation of the hearing aid results in the perception by the hearing aid user of the high fidelity frequency response shown in Fig. 5.

Fig. 8 is a flow diagram illustrating a method of operating a hearing aid in accordance with one embodiment of the present invention. The process begins with block 802, at which sound energy is received from the environment and directed to the microphone of the hearing aid. The microphone converts the sound energy into an electrical signal at block 804. The spectral content of the electrical signal representing the sound is then modified at block 806 to compensate for microphone directionality and for peaks and/or high frequency roll-off in the response curves of the microphone and the receiver. The electrical signal is further modified at block 808 in order to produce a hearing aid response curve that corresponds to the CORFIG response curve for the particular type of hearing aid in operation. At block 810 the receiver of the hearing aid converts the resulting electrical signal back into sound energy, and at block 812 the sound energy is conveyed into the ear canal of the hearing aid user.

While the invention has been described with reference to certain embodiments, it will be understood by those skilled in the art that various changes may be made and equivalents substituted without departing from the scope of the invention. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from its scope. Therefore, it is intended that the invention not be limited to the particular embodiment disclosed, but that the invention will include all embodiments falling within the scope of the appended claims.

APPENDIX A A Volume In the Perspectives in Audilogy Serles CHAPTER 8 ACOUSTICAL CORFIG : FACTORS AFFECTING COUPLER RESPONSE FOR FLAT INSERTION-GAIN Mead C. Killion and Edward L. Monser, IV PERFORMANCE CONTENTS RECENT CORFIG DATA ISO Definitions ISO Measurement Techniques.......... 151 Y Effect of Sound Field t53 Effect of Microphone Location * 156 Gerald A. Studebaker, Ph. D. Effect of thy 159 Distinguished Professor Effect of Earmold 160 Effect of ndividuel Dilierences 161 Department of Audiology and Speech Pathology HISTORY, 16Z Mem his State Universit 162 Hearing Aid Research 163. and SUMMARY,.,.,,,, 166 . 166 Irving Hochberg, Ph. D. Peofessor and Executive Officer The acronym CORFIG, which literally stands for"Coupler Response for Doctoral Program in Speech and Hearing Sciences Fu InsertionGain,"canbethought ofastheanswertothisquestion : What Graduate School, City University of New York would the coupler frequency response of a hearing aid look like if the hearing aid produced a perfectly flat frequency response for the average user ? Another interpretation for CORFIG might be CORrection FIGure, since knowledge of the CORFIG corresponding to a given hearing aid design allows one to estimate the insertion gain that will be provided to a user of that hearing aid. The use of such"coupler correction curves"has a long history. The relationship between the coupler response of a hearing aid and its in situ response can be reasonably well specified on the average, but such a specification must include the following information : r 1, The type of sound field 2. The orientation of the listener in the sound field University Park Press 3. The location of the microphone on the listener's head Baltimore 149 Copyright 1980 by University Park Press, Reprinted with permission CORFIG 151 150 Killion and Monscr 4. The exact construction of the earmold used to measure the in situ ing aid measured on a KEMAR manikin and the average functional gain response measured on a group of subjects with sensorineural hearing loss. 5. The exact construction of the earmold used to measure the coupler The informal phrase coupler response of a hearing aid is used in this response chapter as a shorthand for the more elaborate phrase the response of a 6. The type of coupler used to obtain the coupler response hearing aid whose inpul SPL is held constant and whose output SPL is . measured into a specified coupler. The coupler response as used here is Each of these factors is discussed in conjunction. with the presentation 53. 22 (1976) when the coupler and earmold are chosen appropriately. of recentlyobtained average CORFIG data. S3. 22 (1976) when the coupler and earmold are chosen appropnately. Measurement Techniques RECENT CORFTG DATA There are several ways to obtain the CORFIG curves. The simples intui- Definitions tively would be to construct a hearing aid that produced an exactly flat insertion gain frequency-response curve on the KEMAR manikin and then The insertion gain of a hearing aid on a given user is equal to the ratio of measure its coupler response. the eardrum pressure produced by the hearing aid to Ihe unaided eardrum the eardrum pressure produced by the heanng a<d to the unatded eardrum,......,-.-.-. t.-. DnTr'...., pressure that would have been produced by the same sound field. Unless Less intuitive but more simple in practice, the required CORFIG curve can be predicted from direct measurements of the sort described by Kuhn (Chapter 4, this volume) and Shaw (Chapter 6, this volume). As discussed point source located directly in front of the listener at a distance I m from by Knowles (1959) and others, the CORFIG cUrve corresponding to a given a hypothetical line drawn between the two ear canal openings. hearing aid and given sound field condition can be predicted from : The term funclional gain was introduced by Pascoe (1975) to describe the subjective threshold-difference method of determining the in situ gain 1. The unaided sound pressure level normally produced at the eardrum of a hearing aid. A heanng aid that improved a subject's threshold by 20 by that sound field dB at 1 kHz, for example, was said to have a functional gain of 20 dB at 2. The sound pressure level produced at the in situ hearing aid micro- 1 kHz. To avoid standing wave problems in his test room, Pascoe used phone inlet by the same sound field '/6-octave bands of noise as stimuli. 3. The difference between the eardrum and coupler response of the ear- In most instances, the objective and subjective measures might be phone-earmold combination used with the hearing aid expected to give the same estimate of the in situ gain of the hearing aid, but. The difference (3) should be close to zero on the average when the a variety of experimental pitfalls can plague either type of measurement. coupler response of the hearing aid is measured into a Zwislocki coupler and Background noise (Walden and Kasten, 1976), hearing aid-microphone the same earmold construction is used for both coupler response and inser- noise (Killion, 1976a), standing waves in the test room with pure tones or tion gain response measurements. Thus a good estimate of the CORFIG can insufficiently steep filter slopes with narrowband noise stimuli (Orchik and be obtained with the aid of a KEMAR manikin by subtracting the sound Mosher, 1975), inadequate specification of sound field incidence, variations pressure level available to the in situ hearing aid microphone from the sound in earmold construction and fit, and hearing aid overload (a particular, pressureleveldevelopedattheeardnum-positionmicrophoneoftheun aided problem at 4 kHz with subjects having noise-induced hearing loss with a manikin. The result is the difference that the hearing aid. must make up in "4-kHz notch") have been implicated in producing unusual results during order to compensate for the loss of external ear resonance and the out-of-ear . functional gain determinations. Most ofthese problems, plus problems with microphone location. (In a facetious sense, all one has to do to obtain probe-tube location and probe-tube microphone ca ! ibration (Burkhard and Zwislocki coupler CORFIG curves is subtract everything in Kuhn's chap- Sachs, 1977), can plague attempts to obtain accurate insertion gain meas-ter from everything in Shaw's chapter.) urements. When everything goes well, however, it is possible to obtain good This procedure is illustrated in Figure 1, which is based on measure- agreement between functional gain and insertion gain measurements, ments made on a KEMAR manikin several years ago in the anechoic Causey and Beck (1977), for example, found less than a 2-dB difference at chamber at Industrial Research Products Inc. (IRPI). The curve on the left most frequencies between the insertion gain of an over-the-ear (QTE) hear-shows the pressure developed at the eardrum-position microphone on an '\, EAR SIMULATOR I PD IFIELD) pmlc. , .- i -i . 0 INCIDENCE SOUND FIEL -0 INCIDENCE SOUND FIELD"'_ 20dB 20d8 20dB 0. 4--R-H p p 0 100 IkKz lOkHz 100 IkHt 10 kHz ) 00 H ! tO) <Hl ? 0 tH-tt tOkHl) 00) t Ottt-tl EARDRUM PRESSURE OTE MICROPHONE PRESSURE-DIFFERENCE FUNCTION (REF. COUPLER RESPONSE) Figure 1. Derivation or coupler rcsponse requircd for unity insertion gain (CORFIG) for an ovcr-thecar (OTE) hearing aid. (Reprinted with permission from Killion, 1976b.) unaided KEMAR manikin in a 0'incidence sound field, taken from Burk- hard and Sachs (1975). The curve in the middle shows the pressure devel- oped at the side of the head at a location corresponding to the microphone inlet of an OTE hearing aid. This curve was one of the 23 curves obtained by Madaffari (1974) at possible hearing aid microphone locations on the side of the head. The curve on the right shows the difference between the first two curves, and represents an early estimate of the coupler response required of an OTE hearing aid if it is to produce a flat insertion gain frequency-response curve. In other words, the curve on the right is what we now call the CORFIG curve, or more precisely, the Zwislocki coupler CORFIG curve. A different CORFIG curve results each time the microphone location, sound source location, type of sound field, earmold, or type of coupler is changed. Effect of Sound Field The effect of different sound field measurement conditions is illustrated in Figure 2 for three sound field conditions. These curves were obtained with an OTE hearing aid utilizing a"forward-looking"microphone whose inlet was directly over the pinna and slightly forward of the vertical midline of the earcanal. The 0'and 90'incidence CORFIG curves are similar to the (inverted) correction curves shown by Burkhard (1978), although refine- ments in our measurement techniques (use of more nearly a"point source" loudspeaker) have resulted in minor changes at high frequencies. A Problem with 0'Incidence ifeasurements The 8-kHz notch in the 0'incidence CORFIG curve of Figure 2 is a result of a sharp concha antiresonance, which is discussed by Shaw (Chapter 6, this volume). This antiresonance produces a substantial null in eardrum pressure at approxi- mately 8 kHz. This antiresonance disappears when the concha is filled with an earmold. In order for a broadband OTE hearing aid to have a flat insertion gain in the 5-to 10-kHz region for 0'incidence sound, either the microphone inlet must be located in front of the ear canal entrance in an unoccluded concha, or an 8-kHz notch filter must be included in the hearing aid in order to duplicate the effect of the concha antiresonance. Viewed another way, an OTE hearing aid that has a smooth coupler response in the 5-to 10-kHz region will have an insertion gain curve that shows a sharp peak at about 8 kHz when the insertion gain is measured in a 0'incidence sound field on a KEMAR manikin. On real subjects, one would obtain a similar peak from either objective insertion gain or subjective functional gain measurements in a 0'incidence field, but the exact frequency and magnitude of the peak would vary among subjects (see discussion below and Chapter 6). - :.. p'90°,. '. RANDOM, j ' 0 goo RANDOM -10 0 z... r I __ zu J -'. L 20 50 100 200 500 IkHz 2kH : 5kli lokHz 20kHz FREQUENCY Figure 2. Comparison between 0'incidencc (broken linc), 90'incidence (dottcd linc), and random incidence (solid line) Zwislocki couplcr CORFIG : for an OTE hearing aid. The 8-kHz CORFIG notch (insertion gain peak) presents a dilemma to the designer of a broadband hearing aid. Insertion gain is now commonly measured at 0'incidence in an anechoic chamber, as discussed in Manikin MeasarementsConferenceProceedings (Burkhard, 1978) or the 1978 Y. A. Handbook HearingAid ASeasilrement (Beck, 1978), but listening is sel- dom done in an anechoic chamber. In most real-life situations, sufficient reflected energy arrives at the listener's ear to effectively fill in the notch caused by the concha antiresonance. Even in close face-to-face conversa- tions, 10% to 20% of the energy arriving at the listener's ears is typically reflected energy. When listening to music, the sound sources are generally located much farther than a critical distance away from the listener, so the majority of the energy arriving at the listener's ear is reflected energy. With high fidelity loudspeakers in a typical living room, for example, the curve given by Olson (1967, p. 285) indicates that 90% of the energy arriving at the listener's ear can be reflected energy. The sound arriving after reflections from the walls, ceiling, and floor more nearly represents a diffuse or ran- dom-incidence sound field than a 0'incidence sound wave (which, after all, can only be achieved under anechoic conditions). These comments should not be misinterpreted to mean the notch caused by the concha antiresonance has little significance. The perceptual importance of the concha antiresonance in vertical localization has been studied quite extensively (see, for example, Butler and Belendivk, 1977). For judgments of overall spectral balance in the perceived sound, however, the experiments performed by Schulen (1975) indicate that random-inci- dence sound plays a dominant role. As a result of these considerations, we undertook to obtain random- incidence CORFIG curves. Random-Incidence iYfeasurements The estimate of the random-inci- dence CORFIG shown as the solid curve in Figure 2 was obtained in the IRPI reverberation chamber (width, length, and height of 2. 42 mm, 3. 58 mm, and 3. 64 mm, respectively) using a warble tone with i50 Hz deviation and a 10-Hz triangular wave modulation (modulation index of 10). As with the 0° and 90° data, the random-incidence curve represents the difference between the sound pressure level (SPL) measured first at the unaided mani- kin eardrum location and then at the in situ hearing aid microphone inlet position. (For the latter measurement, the hearing aid was placed on the manikin and attached to an earmold filling the manikin concha.) The inlet pressure was measured with a probe microphone located 1-2 mm from the blocked microphone inlet. The random-incidence curve shown in Figure 2 represents the average of several measurements made with various locations of the manikin in the reverberation room. In some cases, the pressure at the eardrum location was measured with the same probe microphone used to 156 Killion and Monser measure microphone inlet pressure, while in others a Briiel & Kjaer 4134 h-inch condenser microphone was employed. (Both measurements gave essentially similar results after corrections for the pressure frequency re- sponse of each microphone.) Since there appears to be few random-incidence data available in the literature, we were pleased at the close agreement between the eardrum pressure data we obtained and the data calculated by Shaw (1976) from a 52-source position experiment using a KEMAR ear mounted in a small baffle. That comparison is shown in Figure 3. The 3-dB difference shown at low frequencies is the result of a small baffle effect in Shaw's experiment. Choosing the Reference Sound Field The choice facing the hearing aid designer may not be as important in practice as it first appears. Note that all the OTE CORFIG curves in Figure 2 are similar everywhere below 4 kHz. When the hearing impairment is severe enough so that maximizing speech intelligibility in face-to-face listening situations is the dominant consideration, a suitable hearing aid may well have a limited enough band-. width so that it will make little difference which sound field condition is chosen for reference. When the hearing impairment is mild enough so that sound quality, is the major consideration, the random-incidence sound field might be a more suitable reference. The main problem with random-incidence measure- ments is that they require either a reverberation chamber or the averaging of a large number of anechoic chamber measurements. An alternate possi- bility might be to average the 0'and 90'incidence anechoic chamber data for hearing aids whose response extends into the 5-to 10-kHz region. This would provide a rough approximation to random-incidence measurements without their complications. Effect of Microphone Location If the hearing aid microphone is located in the ear instead of over the ear, the location of the sound source becomes much less important. This is illustrated in Figure 4, where representative CORFIG curves are shown for three different microphone locations : over-the-ear (OTE), in-the-ear (ITE), and in-the-concha (ITC). The OTE curves in Figure 4 are identical to those shown in Figure 2, and are included here for comparison. The ITE microphone location corresponds-to a microphone approxi- mately centered in the face of an ITE aid that fills the concha, so that the microphone is located approximately flush with the plane of the external pinna. The ITC microphone location is similar to that diseussedrby Berland and Nielsen (1969), in which the microphone inlet of an OTE hearing aid was extended down into the unoccluded concha in front of a"phantom" z 0 0 LL zu . Z, < 10 C tL SEE TE) : T > _ 4 |X I °S-100zo 0 : OD 2 c"Cß a ZkHz 5kPL DkHz ZOk D-2 LLI-i 0 LU SEE TEXT u. u M 0 a LJ J 1 I O I J-IO J J N - 202p 50 100 200 500 IkHz 2kHz 5kHz IOkHx 20kHz FREQUENCY Figure 3 R. andom-incidcnce eardrum pressure response of KEMAR manikin. Solid line represents IRPI estimate, average of left and right ears ; dotted linc rcprescnts Shaw estimaye, ICEMAR ear mounted in small bafflc. (Reprinted with permission from Killiont 1979.) h u 158 Killion and Monser \- : ; : " on O* 90'fttNDOM 20 0 RANDOM OTE A-^ m OT) 0E { lu 20 OTE 0 CL in ut w 1 w ;. u ive o. 0 1 :/. ITE a w 20- io Q W 0 ETC Z0 50 loo ZOO BOO 1 kH, SAHI IOkH-20kHs FREQUENCY Figure 4. CORFIG curvcs for thrce hearing aid types and three sound field conditions. or"canal-lock"earmold. In this case, nearly all of. the directional properties of the ear are obtained, as discussed by Shaw (Chapter 6, this volume). In a sense, the curves in Figure 4 simply illustrate that one loses more and more of the directional effects of the external ear as the microphone is moved farther from the ear canal entrance, a point that has been made frequently in recent advertising literature for ITE hearing aids. Conversely, CORFIG 159 the closer the hearing aid microphone is located to the ear canal entrance, the less one has to worry about which sound field condition is used to measure insertion gain. Effect of the Coupler So far we have discussed only Zwislocki coupler CORFIG curves. The greater pressure developed by a hearing aid earphone-earmold combination in real ears compared with the 2-cc coupler was shown by Sachs and Burkhard (1972b). As a rough rule of thumb, this amounts to 3. 5 dB at low frequencies, 5 dB at 1 kHz, 10 dB at 3 kHz, and IS dB at 6 kHz, and is nearly independent of the earphone-earmold combination as long as a good earmold seal is obtained. As has been often observed, the error introduced by the 2-cc coupler tends to compensate for the error introduced by ignoring head diffraction and loss of external ear resonance. The comparison shown in Figure 5 illustrates this relationship. The 0-incidence CORFIG for an OTE aid-in other words, the increased response required of the hearing aid to compensate for. the loss of external ear resonance and the out-of-ear microphone location-is shown by the solid curve in Figure 5. The Sachs and Burkhard data on the increase in actual eardrum pressure over the 2-cc coupler pressure developed by hearing aid earphones is shown by the bro- ken curve, which was drawn to have the same low frequency asymptote as the solid curve. The net result of these compensating errors is that the traditional 2-cc coupler frequency response curves would not have been too badly in error (Lybarger, 1978) except for yet another (earmold-related) error, which is discussed in the next section. cl err - _ _ U) UT \/ w ut zu LEU Q J 200 500 I kHz 2 kHz 5 kHz 10 kHz FREQUENCY- Figure 5. Comparison between increased response required of an OTE hearing aid to com- pensante for the loss of external ear resonance and the out-of-ear microphone location. Solid line represents 0'Zwislocki coupler CORFIG ; broken line represents difference between real ears and the 2-cc coupler, drawn to the same low frequency asymptote. 160 Killion and Monser The difference between the two curves in Figure 5 represents the 2-cc coupler CORFIG for an OTE hearing aid in a 0'incidence sound field. That difference is shown as a solid curve in Figure 6, a curve sensibly equal to that shown by Cole (1975), Knowles and Burkhard (1975), and Lybarger (1978), except for a 3. 5-dB level shift (resulting from our equating the low frequency asymptotes in Figure 5) to emphasize the shape of the difference curve. An OTE hearing aid whose frequency response matched the solid curve of Figure 6 when measured in accordance with ANSI-S3. 22 (1976) or IEC-I 18 (1959) should produce a flat insertion gain frequency-response curve when measured on a KEMAR manikin in a 0'incidence sound. field, provided that : 1) the same earmold configuration is used on the KEMAR manikin as is used for the 2-cc coupler measurements, and 2) the OTE hearing aid has the same microphone location used to obtain these COR- FIG curves. The former is the more important provision, and is discussed in the following section. Effect of Earmold The earmold simulator specified for OTE hearing aids in the ANSI and IEC standards comprises 25 mm of 2-mm diameter tubing followed by 18 mm of a 3-mm diameter hole leading into the 2-ce coupler volume. This particu- lar earmold-coupler combination is actually one of several so-called"2-cc couplers."In the United States, it is more properly called an"HA-2 ear- phone coupler with entrance through a rigid tube" (ANSI-S3. 7, 1973) ; in Europe, it is called a"2 cm'coupler with earmold substitute" (IEC-126, 1973). o w 10 du .. 4 J DB 200 500 1 kHz 2 kHz 5 kHz 10 kHz FREQUENCY Figure 6. Difference between the two curves of Figure 5. Solid line rcpresents 2-cc coupler CORFIG for an OTE hearing aid in a 0'incidence ficid. Thc dottcd line reprcsents the additional high frequency gain required to offset the use of a"conventional"carmold rather than the"HA-2"configuration for one hearing aid ; namely, the one illustrated in Figure 5b of Beck (1978). CORFfG The typical eannold supplied to a hearing aid purchaser, however, has the flexible tubing extending nearly to the tip of the earmold. The loss of high frequency response resulting from such a constant bore earmold con- struction was shown by Lybarger (1972) and by Dalsgaard and Jensen (1976). The dotted curve in Figure 6 illustrates, for one particular hearing aid, the additional high frequency gain required to offset the use of a conventional eanmold (l. 9 mm tubing extending nearly to the tip of the earmold) rather than the"HA-2"configuration. The amount of additional gain required depends on the particular hearing aid design (compare, for example, Figures Sa and 5b of Beck, 1978). Nonetheless, the dotted curve of Figure 6 is similar to the difference between functional gain and coupler gain obtained by Pascoe (1975, Figure 9), who measured functional gain with the subject's custom earmolds and 2-cc coupler gain with the HA-2 coupler, and to the differences between insertion gain and coupler gain obtained by Dalsgaard and Jensen (1976, Figures 8 and 9) under similar measuring conditions. With broadband earphones and/or the use of earmold venting, the variation in hearing aid frequency response brought about by changes in earmold construction can be much larger than the net correction shown by the solid curve in Figure 6. Stated differently, the error incurred by ignoring the CORFIG corrections may be small compared to the error incurred by ignoring variations in earmold constructions. (Indeed, it is possible to use a large-tube earmold whose acoustics provide a rough "built-in correction"to the solid CORFIG curve of Figure 6 ; i. e., whose transfer characteristic, when compared with an HA-2 earmold, provides the required boost at 2. 7 kHz). These comments are not meant to mini- mize the importance of the CORFIG determinations, but to reemphasize the large error that can occur when earmold construction is not well defined and controlled. (Extensive discussions of earmold acoustics are found in other chapters of this volume.) Effect of Individual Differences All the data discussed so far are average data as represented by the acoustic characteristics of the KEMAR manikin measurements. If a separate hear- ing aid were to be designed for each user, it would presumably be possible to take into account individual. eccentricities in external ear ("ear canal") resonances and eardrum impedances. To be economically practical, how- ever, most hearing aid designs must be based on average data. Under those circumstances, individual variations in external ear resonance and eardrum impedance may cause the (insertion) gain and the (insertion gain) frequency response of a hearing aid to deviate substantially, for a given individual, from the design nominals. This comes about for two reasons : individual 162 Killion and Monser differences in external ear resonances and individual differences in ear canal and eardrum impedances. External Ear Resonances An estimate of the individual differences in external ear resonance was provided in the data of Filler, Ross, and Wiener (1945). In that report, individual sound field-to-eardrum pressure curves for 1. 2 male and 2 female subjects were given, which can be compared with the overall average curves for the same subjects shown by Wiener and Ross (1946). The standard deviation (from the average curve) of the individual curves ranged from I to 2 dB below 1800 Hz up to from 4 to 7 dB in the 5-8 kHz region, with peaks at 2. 1 and 3. 3 kHz. The peak deviations oc- curred mostly because individual external ear resonance frequencies were lower or higher than the average. No individual eardrum pressure curve deviated more than 7. 5 dB from the average curve below 5 kHz, but the majority deviated by at least 5 dB at some frequency below 5 kHz. Eardrum Impedance As part of a study leading to a validation of the modified Zwislocki coupler, Sachs and Burkhard (1972a) reported the probe-tube microphone measurement of the sound pressures devel- oped in 11 occluded ears (6 male and 5 female) by subminiature hearing aid earphones. The standard deviation of the pressure developed ranged from approximately t dB at 1 kHz to 5 dB in the 6-to 8-kHz region. Greater pressures were developed in female ears (by 3 to 5 dB at the higher. frequencies). The Nef Result Although the variations in outer ear resonance and eardrum impedance are only partially independent variables, it is clear that no hearing aid designed for the average ear can be expected to produce an insertion gain, in the majority of individual cases, that does not exhibit one deviation of perhaps 7 dB at some frequency. The subjective importance of such deviations to a long-term hearing aid wearer is not known, but even larger deviations in unaided frequency response can occur because of the accumulation of earwax in the canal, deviations that occur with such a gradual onset that they often go unnoticed by the sufferer until the canal becomes almost completely blocked. Thus, in most cases, it seems reason- able to assume that satisfactory adaptation to a slightly inaccurate insertion gain frequency response makes it unnecessary to provide modification for individual eccentricities. Experimental evidence one way or the other is lacking. HISTORY Cotipler Corrections The 2-cc coupler was first described by Romanow (1942), who emphasized that it was not a real-ear simulator, but simply a convenient, easily fab-. CORFIG 163 ricated coupler into which readily reproducible hearing aid measurements could be made. Romanow provided tentative correction curves, to be ap- plied to the 2-ec coupler response curves for body-worn hearing aids, in order to estimate the field-referenced response of the hearing aid. In 1944, LeBel reported a series of studies in which he obtained 2-cc coupler correction curves similar to those described by Romanow. To illustrate the importance of the corrections, LeBel showed a comparison between two frequency responses : one uncorrected, which looked fairly flat, and one corrected, which showed a 15-dB high frequency loss. Knowledgeable hearing aid designers have been using these corrections for years. Lybarger described one such hearing aid design in 1947, for example. With head-worn hearing aids a totally new set of corrections was required, since the hearing aid microphone was now located on the head instead of the chest. Thus in 1959 Knowles described a tentative set of coupler correction curves, based on the data provided by Wiener and Ross (1946) on the (unaided) eardrum pressure produced by a progressive sound field. The refined estimate of the difference between normal eardrum pres- sure and the pressure available to the microphone of an ITE hearing aid, as given by Knowles in 1967, is essentially similar to the best estimates of today. It was reproduced by Killion and Carlson (1970). The remaining question was the difference between the eardrum pres- sures produced in real ears and in the 2-cc coupler. Although several laboratories had reported ear canal plus eardrum impedance values substan- tially higher than those presented by the 2-cc coupler, no alternate coupler had gained general acceptance until Zwislocki (1970) reviewed the problem and developed a realistic ear simulator. The suitability of this ear simulator (coupler) was verified by Sachs and Burkhard (1972a), who determined the 'eardrum pressure developed by insert receivers on 11 real cars. A slightly modified version of the original Zwislocki coupler (IRPI DB-100) was described by Sachs and Burkhard (1972b). Most of the modifications were for fabrication convenience, although a small change from Zwislocki's origi- nal design was made in one of the four acoustic branches, a change which smoothed the transfer impedance in the 500-to 1000-Hz region by approxi- mately 1 dB. In 1972 the KEMAR manikin (Burkhard and Sachs, 1975) was intro- duced, and it became practical to obtain a physical measurement of the insertion gain of hearing aids on a routine basis. Hearing Aid Research Before leaving the history of coupler correction curves, it is interesting to note that although those individuals involved with the standardization of hearing aids were acutely aware of the need for such corrections to obtain 164 Killion and Monser the true frequency response of a hearing aid as perceived by a wearer, these corrections were almost universally ignored in audiologic research. This was true even when the effect of the frequency response of a hearing aid on aided performance was the experimental question. Figure 7, for example, shows the effective or"orthotelephonic"frequency response calculated by Fletcher (1953) for the master hearing aid used in the Harvard study (Davis et al., [947). It is evident that the frequency response that was labeled"fiat" was not what one would normally consider a flat frequency response. Simi- larly, it is perhaps not surprising that the frequency response labeled"high- pass 6"gave almost universally better results. (The"high-pass 6"response was not flat either ; but except for the peak at 7 kHz it looks like a perfectly sensible real-ear response for a hearing aid.) Lybarger (1978) gives an expanded discussion of these issues. The first systematic attempt to relate the speech discrimination ob- tained with hearing aids to their true in situ frequency response was re- ported by Fournier (1965), who used free field Bekesy audiometry to obtain what we now call the functional gain of the hearing aids under evaluation. The functional gain of the hearing aid was obtained from the difference between the aided and unaided threshold tracings for each subject. (This same technique was used a few years later in the United States by Green (1969) to obtain the functional gain frequency-response provided by open canal hearing aid fittings.) Not until the early 1970s did comprehensive research studies begin to appear in which the experimental design took into account the real-ear frequency response of the hearing aid. Excellent studies were conducted by Villchur (1973), Pascoe (1975), Skinner (1976), and Lippman (1978). An Aside on Terminology In 1975, Mahlon Burkhard and I became dissatisfied with the term or- Ihotelephonic for reasons discussed by Burkhard (1978, p. 17). We were looking for a term that could be unambiguously defined to mean the objec- tive measurement (with a probe-tube microphone at the eardrum, for exam- ple) of the in situ gain of a hearing aid. Professor R. V. Schoder, a classics scholar at Loyola University, coined the term etymotic (pronounced et-im- oh'-tik ; literally, real ear), which he assured us met all the rules for coining "new ancient Greek words." At about the same time, Dalsgaard (1974) had begun using the term insertion gain, a well-defined engineering term that appears to have been first applied to hearing aids by Ayers (1953). Both terms have the same meaning. The term insertion gain-which is self-explanatory-appears to be gaining favor over the term etymolic in the United States. Both terms are used in Europe.

FREQUENCY<BR> Figure 7. Real-ear (orthotelephonic) response of master hearing aid used in the Harvard study. (After Fletcher, 1953.) 166 Killion and Monser SUMMARY As stated earlier, the relationship between the coupler response of a hearing aid and its in situ response can be reasonably well specified on the average, but such a specification must include the following information : 1. The type of sound field 2. The orientation of the listener in the sound field 3. The location of the microphone on the listener's head (or body) 4. The exact construction of the earmold used to measure the in situ response 5. The exact construction of the earmold used to measure the coupler response 6. The type of coupler used to obtain the coupler response When the bandwidth of the hearing aid is limited, the first three factors may be safely ignored for most practical cases, but they become increasingly important as the hearing aid response is extended above 4 kHz. In the latter case, the use of random-incidence sound field measurements has some appeal. The large number of variables that must be specified before reproduci- ble CORFIG determinations can be made indicates that the choice of standardized coupler corrections may be a complicated business. 1 REFERENCES American National Standards Institute. 1973. Method for Coupler Calibration of Earphoncs, ANSI-S3. 7-1973. American National Standards Institute, New York. American National Standards Institute. 1976. Specification of Hearing Aid Charac- teristics, ANSI-S3. 22-1976. American National Standards Institute. New York. Ayers, E. W. 1953. A discussion of some problems involved in deriving objective performance criteria for a wearable aid from clinical measurements with labora- tory apparatus. Proceedings of the First ICA Congress, Delft, pp. 141-143. W. D. Meinema, Delft. Beck, L. B. (ed.). 197B. Handbook of Hearing Aid Measurement 1978. Veterans Administration, Washington, D. C. Berland, O., and E.. Nielsen. 1969. Sound pressure generated in the human external car by a free sound ficld. Audccibcl 18 : 103-109. Burkhard, M. D. (cd.). 1978. Manikin Measurements-Conference Proceedings. Industrial Research Products, Elk Grove Village, Ill. Burkhard, M. D., and R. M. Sachs. 1975. Anthropometrie manikin for acoustic research. J. Acoust. Soc.. Am. 58 : 214-222. Burkhard, M. D., and R. M. Sachs. 1977. Sound pressure in insert earphone couplers and real cars. J. Speech Hear. Res. 20 : 799-807. Butler, R. A., and K. Belendivk. 1977. Spectral cues utilized in the localization of sound in the median sagittal plane. J. Acoust. Soc : Am. 61 : 1264-1269. Causey, G : D., and L. B. Beck. 1977. The relation of measurements on KEMAR to CORFIG 167 behavioral performance. Paper presented at the annual convention of the Ameri- can Speech and Hearing Association, November 2-5, Chicago. Cole, W. A. 1975. Hearing aid gain-A functional approach. Hear. Instr. 26 : 22-24. Dalsgaard, S. C. 1971. On the frequency response of hearing aids. In Proceedings of the 7th International Congress on Acoustics, pp. 477-480. Akademiai Kiado, Budapest. Dalsgaard, S. C. 1974. Measurement of insertion gain of hearing aids. Proceedings of the Eighth International Congress on Acoustics, London, p. 205. Goldcrest Press, Trowbridge. Dalsgaard, S. C., and O. D. Jensen. 1976. Measurement of the insertion gain of hearing aids. J. Audiol. Tech. 15 : 17fS183. Davis. H.. S. S. Stevens, R. H. Nichols, C. V. Hudgins, R. J. Marquis, G. E Peterson, and D. A. Ross : 1947. Hearing Aids : An Experimental Study of Design Objec- tives. Harvard University Press, Cambridge. Filler, A. S., D. A. Ross, and F. M. Wiener. 1945. The pressure distribution in the auditory canal in a progressive sound field. Harvard University Report PNR-5 (Contract N5ori-76) to Office of Research and Inventions, U. S. Navy. Fletcher, H. 1953. Speech and Hearing in Communication. D. Van Nostrand Co., New York. Fournier, J. E. 1965. La correction prothétique des curdites ; probiemcs techniques et medico-sociaux. Pract. Oto-Rhino-Laryngol. 27 : 180-198. Fournier,) : E. t968. Hearing aid evaluation by pure tone automatic audiometry. Audecibel 17 : 99-108. Green, D. A. 1969. Non-occluding canmolds with CROS and IROS hearing aids. Arch Otolaryngol. 89 : 96-106. IEC-118. 1959. Recommended methods for measurements of the electro-acoustical characteristics of hearing aids. Bureau Centre de la Communications Elee- trotechnques Intemationaies, Geneva, Switzerland. IEC-126. 1973. IEC reference coupler for the measurement of hearing aids using earphones coupled to the ear by means of ear inserts. Bureau Centre de la Communications Electrotechniques Internationales, Geneva, Switzerland. Killion, M. C. 1976a. Noise of ears and microphones. J. Acoust. Soc. Am. 59 : 424 433. Killion, M. C. 1976b. Experimental wide-band'hearing aid. J. Acoust. Soc. Am. 59 (suppl. I) : S62 (A). (Available from Knowles Electronics, Franklin Park, 111.) Killion, M. C. 1979. Equalization filter for eardrum-pressure recording using a KEMAR manikin. J. Audio Eng. Soc. 27 ; 13-16. Killion, M. C., and E. V. Carlson. 1970. A wideband miniature microphone. J. Audio Eng. Soc. 18 : 631-635. Knowles, H. S. 1959. Physical aspects of hearing aids. Paper presented at Allerton House, University of Illinois, Urbana. Knowles, H. S. 1967. Some problems in head-worn aid response measurements. Paper presented at the 33rd Convention of the Audio Engineering Society, Octo- ber 16, New York. Knowles, H. S., and M. D. Burkhard. 1975. Hearing aids on KEMAR. Hear. Instr : 26 : 19-21. LeBel, CJ. 1944. Pressure and field response of the ear in hearing aid performance determination. J. Acoust. Soc. Am. 16 : 63-67. Lippman, R. P. 1978. The effect ofamplitude compression on the intelligibility of speech for persons with sensorineural hearing loss. Doctoral thesis, Massachu- 168 Killion and Monser setts Institute of Technology, Cambridge. (Available from University Microfilms, Ann Arbor, Mich.) Lybarger, S. F. 1947. Development of a new hearing aid with magnetic microphone. Electric. Manufact. 40 : 104-108, 162. Lybarger, S. F. 1972. Ear moids. In J. Katz (ed.), Handbook of Clinical Audiology. pp. 602-623. The Williams & Wilkins Co., Baltimore.. Lybarger, S. F. 1978. Selective amplification-A review and evaluation. J. Am. Audiol. Soc. 3 : 258-266. Madaffari, P. L. 1974. Pressure variation about the ear. J. Acoust. Soc. Am. 56 : S3 (A). (Available from Knowles Electronics, Franklin Park, III.) Olson, H. F. 1967. Music, Physics and Engineering. Dover Publications, New York. Orchik. D. J., and N. L. Mosher. 1975. Narrow band noise audiometry : The effect of filter slope. J. Am. Audiol. Soc. 1 : 50-53. Pascoe, D. P. 1975. Frequency responses of hearing aids and their effects on the. speech perception of hearing impaired subjects. Ann. Otol. Rhinol. Laryngol. 84 (suppl. 23) : 1-40. Romanow, F. F. 1942. Methods for measuring the performance of hearing aids. J. Acoust. Soc. Am. 13 : 294-304. Sachs, R. M., and M. D. Burkhard. 1972a. Earphone pressure response in ears and couplers. J. Acoust. Soc. Am. 51 : 140 (A). (Available from Knowles Electronics, Franklin Park, 111.) Sachs, R. M., and M. D. Burkhard. 1972b. Zwislocki coupler evaluation with insert earphones. Industrial Research Products Report No. 20022-1 to Knowles Elec- tronics, Franklin Park, Ill. Schulen, R. B. 1975. In situ : measurement and equalization of sound reproduction systems. J. Audio Eng. Soc. 23 : 178-186. Shaw, E. A. G. 1976. Diffuse field sensitivity of external ear based on reciprocity principles. J. Acoust. Soc. Am. 60 : S102 (A). Skinner, M. W. 1976. Speech intelligibility in noise-induced hearing loss : Effects of high-frequency compensation. Doctoral thesis, Washington University, St. Louis. (Available from University Microfilms, Ann Arbor, Mich.) Villchur, E. 1973. Signal processing to improve speech intelligibility in perceptive deafness. J. Acoust. Soc. Am. 53 : 1646-1657. Walden, B. E., and R. N. Kasten. 1976. Threshold improvement and acoustic gain with hearing aids. Audiology 15 : 413-420.... Wiener, F. M., and D. A. Ross. 1946. The pressure distribution in the auditory canal in a progressive sound field. J. Acoust. Soc. Am. 18 : 401-408. Zwislocki, J. J. 1970. An acoustic coupler for earphone calibration. Report LSC-S-7, Laboratory of Sensory Communication. Syracuse University, Syracuse, N. Y.

APPENDIX B ER101-28D TM gRIED CAPACITaR r--T-\) -r-) rt-r-Art t BURtEDCAPACtTOR ETY M OT I C R E S EA R C H su BSTRATE (BCS) - K-AMP HYBRID ! C crs c6 CHrs vs f ... '.. +MIC CF2 C6 CHFB VBAT . itSZY V : yj ; t.. ff'^ : ji', ; : : AUX. IN doux in gond AUX. IN. uaur'"'v : t'-= : "ry ; r ;,. :. 4 : : : . : = : tv : :, CT (C', CTOI.. w-.. i \ I _'16Q -. 0061 ME cite CTOL lzS +. 006" MAX VGOUT VCR CF AMPIN LOGIN-TENTAToVE LEVEL-DEPENDENT FREQ. RESP. OF PREAMP ESTIMATED CANAL A) 0 FREQUENCY RESPONSE E9tQQ65 ES0127 aF1746, CHFas1QpF I I ; Z o, , I I., I NPUT SPL1N dA : I p. y0 _ W . . i. I 50 I t0s000s O. s < l. i I ! I tau to I i i'."o .-., a I 125 250 Sa0 1QQ0 MM 4QQQ. SOOO 160QCt, lzs 2so sac low'low SOOQ MM aooo coo FRECUENCY IN HZ. FREOUENCYIN HZ NOTES : The important changes in using the 28D hybrid compared to the 19D and 24D hybrids are: 1. On the 28D hybrid, the leads are soldered on the sides to the. 056"x. 020"gold contact lands.

2. The 28D hybrid eliminates the need for two components normally required to meet the 16 kHz response goal of the High Fidelity K-AMP hearing aid. (The 19D and 24D hybrids require150 pF and 47k in parallel with the volume control for that purpose. ) The 28D hybrids contain a 630 pF capacitor (C2S) in shunt with VCR. That capacitor provides a high-frequency boost which disappears for quiet sounds (see graph). Since even the normal ear has a loss of bandwidth for quiet sounds, the subjective result is quite satisfactory. An advantage of this type of boost is a reduced tendency to feedback in quiet environments.

3. The 28D hybrid contains a 1000 pF from VGOUT to ground, eliminating the need for an external capacitor to prevent 1 MHz oscillation (see'The Mostly Inaudible Squeek"application note).

4. The 28D hybrid contains a 3 nF capacitor from MIC+ to ground. This increases the power-supply feedthrough isolation at high frequencies, and may sometimes make additional filtering unnecessary when a. high power Class D receiver is used.

With regard to note &num 2, the HLHFPG (High level high frequency preamplifier gain) test limits have been increased 5 dB to accommodate the effect of C2S. That test is performed at 10 kHz with a 10 mV input (approximately 90 dB SPL equivaient). 4 n u P 1 n fi SH2-B Ft 14 ZOAT RIB . In 3 e lO ln 40daatttnuatQf.' PREAIiP - I INPUT V/N R10 R12 1 3'12 ECEV£R LOOK Ril 2*-CA N 1 SU2 A to l I SU3 I. n IN L 10 ion VCOUT RS 15gon F Vr'R ils PREAMP su aux Sua RJ. t UK ! liK,. L NOTES : RE5 ? STORS e 70LERANCE TEST CIRCUIT 2 SU2 15 FOR STaan. : rYUMOERLoao UXTH LS OHn 3ATTERY Ir1PEDaHCE 3. ALt. RESISTOR VALUES Aqr-IN OHMS 4. VALUES GIVEN FOR. 1n ON OWC. ER10L-69 iIRE MEA5UREO AT THE HYARID INPUT ETYMOTIC RESEG1HCH VOLTPGC AT INPUT TO 4aA3 ATTEHQETOR SHOUID AE 104 THI$ VRIUE E. G. : FOR G60 fIEHSUFEHEHT, 1n o 1mV ElK GROVE VILLC1CE Il 60Q07 INPU1 TO pTTEN. = lGamV. Tltle TEST CIRCUIT ie aCUmen Humbr R ER'Û L63 SCH 9 at. : et 1 0 1 RESPONSE OF OUTPUT & RCVR EQ. AMP RESPONSE OF OUTPUT & RCVR EQ. AMP aa ao CHFB (C7) =. 01 uF CHFB (C7) 2. 2 nif m.-I----- V. C = o K a V. G. = a z V. 30 a I ao 3 K 10 K Cl : Lu (i u w eto X 2025 250 SOQ 1040 2400 4000 8000 56000'at25 254 540 1Q00 2a00 4040 840 15 (104 t2S 2SO SOO) OCO MM 4000 3000 tSOOO) 2S 2SQ 500 tOM 2000 4000 MM tECCO FREQUENCY IN HZ'''FREQUENCY IN HZ