WO2000078096A2 - Hearing aid with an acoustical format - Google Patents

Abstract

Hearing loss types are segregated into a predetermined number of classes. For each class, an approximate hearing loss is determined and a corresponding fixed acoustical format, e.g. frequency response, is assigned to correct a corresponding type of hearing impairment. Hearing aids are then classified in a matrix based on characteristics of a corresponding frequency response. An appropriate frequency response of a hearing aid is achieved using analog filters, optionally employing switched capacitor forms.

Description

HEARING AID WITH AN ACOUSTICAL FORMAT

RELATED APPLICATION(S)

This application is a continuation-in-part of and claims priority to U.S. Application No. 09/524,043 filed March 13, 2000; which itself claims the benefit of U.S. Provisional Application No. 60/139,204 filed June 15, 1999; and U.S. Provisional Application No. 60/157,973 filed October 6, 1999; the entire teachings of all of which are incorporated herein by reference.

This application is related to copending U.S. Applications: 09/524,666; 09/524,040; 09/524,501; 60/188,997; 60/188,996; 60/188,721; 60/188,857; 60/188,736; all of which were filed March 13, 2000, the entire teachings of all of which are incorporated herein by reference.

BACKGROUND OF THE INVENTION

One form of hearing loss is caused by reduced sensitivity at high frequencies. This type of hearing impairment is sometimes corrected using frequency equalization. For example, the range of frequencies in which a listener has reduced sensitivity is amplified so that there is an even sensitivity of all frequencies across the audible range. In older hearing impaired patients, reduced sensitivity is more predominant at the high end of the audible frequency range. As a result, frequencies at the high end of the audible spectrum are usually amplified to support equalization of sound across the audible spectrum for impaired listeners.

Such hearing aids are typically custom fitted to an individual based upon his or her audiological and physical needs. Accordingly, an acoustical format is created to compensate for the hearing loss of the individual.

A common method of providing a desired acoustical format is to create a custom ear mold or shell, which is made to fit the ear and/or ear canal of the individual. It can be a painstaking process to correct hearing using such methods because types of hearing impairments vary from one individual to another as do the physical characteristics of one ear canal versus another.

More particularly, to create an appropriate acoustical format, a basic hearing aid style such as behind-the-ear (BTE), in-the-ear (ITE), in-the-canal (ITC), or completely-in-the-canal (CIC) is selected for fitting. The corresponding acoustical format is then tailored at the audiologist's office. Depending on the method, this process might include adjusting mechanical features of the hearing aid such as switches or rotating trimmer controls. Moreover, the process might involve adjusting the characteristics of the device electronically using a programming device such as a handheld programming unit or computer interface. Whether proper fitting of an earpiece requires adjusting programmable circuits or the physical characteristics of the earpiece itself, the additional time to properly fit an individual with such devices results in an increase in cost without a substantial increase in benefit to the hearing impaired patient.

Analog signal processing for hearing aid applications initially consisted of frequency-independent linear amplification. Later, frequency compensation and compression circuits were included in the signal processing functions. In some hearing aid circuits, the frequency spectrum is split into two channels, a low- frequency channel and a high-frequency channel, with the gain and compression of each channel independently controlled.

Each channel includes filters. Filters for analog signal processing are often implemented using op-amps. For hearing aids, the constraints of small size, low power and low operating voltage have driven the signal processing filters to rather simple filter designs. For example, the basic dynamic filter in a K-amp hearing aid circuit is implemented as a single op-amp, first order filter. Another example is the Gennum DynamEQ hearing aid circuit in which the filter is implemented using a two op-amp first-order filter.

While simple first-order filters used in hearing aids do provide benefit to the hearing aid user, higher order filters can match the characteristics of the hearing loss much better. High order filters should provide the hearing aid user higher benefit, particularly in terms of a more natural sound since high order filters offer more spectral control. For example, high order filters can be tuned not to over-amplify frequencies where the user's residual hearing is still acceptable.

Theoretically, high order filters can be realized with a single op-amp, but for practical reasons, are often not. For example, a second-order filter section is often implemented using the well-known biquad configuration having three op-amps. A typical, continuous-time, biquad, second-order filter section uses three op-amps, while a switched-capacitor version of the same biquad filter section can be implemented using only two op-amps. Therefore, analog signal processing with high-order filters often require many op-amps. Thus, using a bi-quad, three op-amp filter arrangement, signal processing in a hearing aid application using sixth-order filters may use as many as eighteen or more op-amps.

It is well known that hearing aids provide amplification to compensate for hearing loss. Under noise-free conditions, simple amplification and frequency compensated amplification systems provide acceptable performance. However, under noisy conditions and particularly conditions where the noise contains higher low-frequency components, the amplifiers of certain hearing aids are often driven into saturation (clipping). When clipping occurs, the high frequency components essential to speech intelligibility are sometimes so distorted or attenuated that high frequency components of the original signal disappear altogether.

In an attempt to eliminate distortion due to loud low- frequency signals, some hearing aids include compression circuits. When loud low-frequency signals are present, these compression circuits reduce the gain of the hearing aid. This reduces the strength of the high-frequency signal components critical to speech intelligibility. Experience has shown, therefore, that these types of hearing aids are often ineffective in noisy environments such as restaurants, automobiles, trains and airplanes where low-frequency noise is predominant.

SUMMARY OF THE INVENTION

The present invention is generally directed towards certain aspects of hearing aid devices. One aspect of the present invention involves segregating types of hearing impairments into classes or ranges. For each range, an approximate or average level of reduced sensitivity or hearing loss is determined. Accordingly, an acoustical format having a defined frequency response is then calculated to remedy hearing loss for the approximated types of hearing impairments for each range. In one embodiment, each hearing loss range covers a span of about 10-12 dB, reducing the number of ranges to a reasonably manageable number. Hearing aids are preferably programmed at a factory with a fixed acoustical format and are prescribed to a patient for correcting a corresponding type of hearing impairment.

In accordance with other aspects of the present invention, a set of predetermined parameters are used to define a matrix of hearing aids in which each of the hearing aids has one fixed acoustical format. One parameter classifies a hearing aid according to a relative change of gain in a predetermined frequency range of an acoustical format. A second parameter is used to further classify a hearing aid based on, for example, a maximum or peak gain of a particular acoustical format at a predetermined audible frequency. In one embodiment, the hearing aids are classified to produce a two dimensional matrix. Each type of hearing aid in the matrix has a pre-programmed acoustical format or frequency response to remedy a hearing impediment depending on a severity of the corresponding type of hearing loss being corrected.

In accordance with still another aspect of the present invention, a transfer function provides basic characteristics of a 2-channel system. An analog filter provides the flexibility to easily select the maximum slope of the transfer function between the low-frequency and high-frequency ranges. Slope selection flexibility facilitates, for example, hearing aid designs that can be tuned to offset the various hearing loss characteristics.

In accordance with yet another aspect of the present invention, analog filter performance is improved and physical size of the high order filters is reduced. In particular, second order filters are implemented using a combination of two op-amps and switched-capacitors. Using the preferred second order filter sections, higher order filters, which include multipliers, produce the characteristics of the transfer functions operating in the hearing aids, for example. In an alternate embodiment of the high order filter, a variable gain op-amp circuit is configured to act as a multiplier, which eliminates multiplier circuits to improve power savings further.

In an application such as hearing aids, it is useful to implement the signal processing design using a minimum number of op-amps. The benefits of having a minimum number of op-amps include smaller size, lower noise, lower power, and lower cost when compared to implementations using a higher number of op-amps.

One aspect of the present invention is directed towards reducing the effects of background noise of an audio signal. In particular, frequency components of an audio input signal are segregated and processed in separate channels. A summer circuit is used to recombine the separated components into the output signal for a listener. Since separate channels are used to process corresponding frequency bands, clipping due to background noise in one channel does not completely deteriorate the integrity of frequency components of the other channels of the input signal. In one embodiment, the amplification of frequency ranges of a given channel is adjusted to compensate for the hearing loss of a hearing impaired patient.

Each channel includes a filter of which a corresponding output is fed into a non- linear amplifier. The output of the amplifier is then fed into a second filter whose output is recombined at a summer circuit with the channel outputs of other frequency components of the original audio input signal. In a preferred embodiment, the filter circuits for each channel are chosen so that the range of frequencies are contiguous across the audible spectrum. Based on this topology, it is possible to process an input signal to produce a related output signal at the summer circuit to compensate for the hearing loss of an individual. The previously discussed features associated with segregating an input signal into separate channels is advantageous over the prior art because simple software algorithms can be sued to produce an output signal with low distortion even in the presence of noise. These simple algorithms can be easily implemented in low-power consuming devices such as digital signal processors.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a graph of hearing loss versus frequency which is representative of a range of hearing loss generally considered as mild to moderate.

Fig. 2 is a graph as in Fig. 1 showing the representative range of hearing loss segregated into a limited number of regions.

Fig. 3 is a graph as in Fig. 1 of the nominal hearing loss for each of the representative regions of hearing loss useful in understanding the principles of the present invention.

Fig. 4 is a graph plotting the gain in decibels versus frequency of representative target responses for a family of linear hearing aids which could compensate for the hearing loss plotted in Fig. 3 according to the present invention.

Fig. 5 is a graph as similar to Fig. 4 of the representative target responses for a family of non-linear hearing aids according to the present invention.

Fig. 6 is a block diagram illustrating functional components of the present invention.

Figs. 7 - 12 are magnitude Bode plots of transfer functions describing analog filters for use in hearing aid applications.

Fig. 13 is a block diagram of a generic analog filter for implementing a transfer function of the form depicted in Figs. 7-12.

Fig. 14 is a schematic diagram of a continuous-time biquad band-pass filter.

Fig. 15 is a schematic diagram of a switched-capacitor biquad band-pass filter.

Fig. 16 is a block diagram of circuitry to implement signal processing transfer function.

Fig. 17 is a block diagram of optimized circuitry to implement signal processing transfer function.

Fig. 18 is a schematic diagram of preferred implementation of this invention.

Fig. 19 is an example of a matrix of hearing aids classified according to the principles of the present invention.

Fig. 20 is a schematic diagram of a hearing aid utilizing separate channel circuits to process corresponding frequency components of an audio input signal. The foregoing and other objects, features and advantages of the invention will be apparent from the following more particular description of preferred embodiments of the invention, as illustrated in the accompanying drawings in which like reference characters refer to the same parts throughout the different views. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention.

DETAILED DESCRIPTION OF CERTAIN EMBODIMENTS

Fig. 1 is a graph depicting a range of hearing loss considered as mild to moderate. For example, line A represents a reduced sensitivity of a hearing impaired individual having a mild case of hearing loss. Line B represents a reduced sensitivity of a hearing impaired individual having a moderate case of hearing loss. Shaded region 100 represents a continuum of hearing loss somewhere between the severity of hearing loss as represented by line A and line B.

Notably, line A and line B generally have a common shape to the extent that hearing loss is greater at high frequencies. However, sensitivity of a patient having a hearing impairment as depicted by line B has a steeper hearing loss drop-off in the mid-range of audible frequencies as shown in Fig.l.

Fig. 2 is a graph illustrating ranges of hearing loss according to the principles of the present invention. A first region 210 depicts a milder form of hearing loss and is defined as the region between line A and line C. A second region 220 is the next more severe range of hearing loss and is defined as the region of hearing loss between line C and line D. Lastly, a third region 230 is the most severe range of hearing loss in our example and is defined as the region between line B and line D. Each region, e.g., the first region 210, second region 220, or third region 230, defines a class of hearing impaired individuals. Those within a class are considered to have very nearly the same type of hearing impairment.

Fig. 3 illustrates the same features as found in Fig. 2, however, several lines have been added depicting the average or approximated hearing loss for a specified region. For example, the approximate hearing loss for the first region 210 is defined by line X. In a similar manner, line Y is the approximated hearing loss for the second region 220, while line Z is the approximated hearing loss for the third region 230.

An average listener generally cannot detect a 1 dB change in sound. However, sound level changes on the order of 3 dB are typically detectable. A sound level increase of +10 dB is considered twice as loud to the average listener. Accordingly, to reduce the number of hearing loss regions to a manageable number, each region is separated by approximately 10-12 dB. That is, each region roughly covers a hearing loss range of about 10-12 dB. This wide coverage defining a range ensures that there is not an unnecessary number of acoustical formats tracked for a family of hearing aids of the present invention. Otherwise, there are potentially an infinite number of hearing loss types and corresponding number of corrective acoustical formats that must be maintained.

Fig. 4 is a graph illustrating a desired hearing aid response for each of the hearing impediment ranges as defined by line X, line Y, and line Z of Fig. 3. As shown, the various frequency responses define a family of linear hearing aids. Acoustical format 410 provides appropriate gain at coπesponding frequencies of a received acoustical signal to restore hearing for individuals diagnosed with hearing impediments as defined by line Z.

Based on a close proximity to line Z, individuals with hearing impediments that fall within the third region 230 are provided with acoustical format 410 to restore their hearing. In a similar manner, individuals diagnosed with hearing impediments in second region 220 are assigned an acoustical format 420 to correct their hearing. Likewise, individuals diagnosed with hearing impediments in the first region 210 are assigned an acoustical format 430 to correct their hearing. In one embodiment, the gain at a particular frequency as illustrated by the acoustical formats is about half that of the hearing loss. Accordingly, each acoustical format is about 5-6 dB apart from each other.

In one embodiment, the hearing aids are accurate within 3 dB of the desired response. That is, the frequency response of an acoustical format due to manufacturing tolerances is off by no more than 3 dB from the ideal response as shown in Fig. 4 or Fig. 5. Fig. 5 is a graph illustrating prefeπed frequency responses for a family of non-linear hearing aids. For example, a desired frequency response of several acoustical formats is shown for each of the hearing impediments, as defined by line X, line Y, and line Z of Fig. 3. An acoustical format 510 provides an appropriate gain for restoring normal hearing to individuals diagnosed with a hearing impediment as defined by line Z. Based on a close proximity to line __, individuals diagnosed with hearing impediments that fall within the third region 230 are assigned acoustical format 510 to restore their hearing. Similarly, individuals diagnosed with hearing impediments within the first region 210 and the second region 220 are assigned acoustical format 530 and acoustical format 520, respectively, to coπect corresponding types of hearing loss.

Fig. 6 is a block diagram illustrating the functional components of one embodiment of the present invention. Generally, acoustical vibrations 605 in air are processed by a hearing aid 600 to provide restoration of hearing for a patient at ear canal 660.

More particularly, acoustical vibrations 605 are detected at a microphone 610 that, in turn, produces a low-level electrical signal 615 coπesponding to the detected sound. A pre-amplifier and compressor 620 amplify the low-level signal and compress it to fit within the dynamic range of audible hearing. Amplified signal 625 at the output of the pre-amplifier and compressor 620 is then fed into an amplifier/filter circuit 630 for further processing. The electrical characteristics of the filter/amplifier circuit 630 depend on an acoustical format type 627 selected.

One format type, for example, that can be selected for hearing aid 600 is acoustical format 530 as previously discussed. When formatted accordingly, the filter/amplifier 630 processes amplified signal 625 to produce an output signal 635, which is fed to an output driver 640 for driving a speaker 650. Sound output 655 from the speaker 650 is then directed to a patient's ear canal 660. In this way, a single hearing aid 600 can be formatted to correct a particular type of hearing loss based on a selected acoustical format.

As noted previously, a limited number of acoustical formats are tracked and maintained to coπect hearing impairments of, for example, the general population. In furtherance of this goal and according to the principles of the present invention, two or more parameters are used to track different types of acoustical formats. For example, an acoustical format is identified by a combination of parameters such as the shape of the frequency response and the maximum or peak gain of a particular acoustical format range. Classifying hearing aids in this manner simplifies tracking multiple types of hearing aids.

The shape of the frequency response, as mentioned, is one aspect of an acoustical format and is identified by a code. In one embodiment, the code is a letter of the alphabet. Preferably, the code defines the steepness of the gain profile in a portion of the audible frequency band.

Peak gain information such as maximum gain at a certain frequency is another aspect of a particular acoustical format and is optionally identified by a number. For example, if a code indicates a particular shape of the frequency response curve, a number is optionally used to indicate peak gain information of an acoustical format type or possibly the range of gains of a particular acoustical format. In one embodiment, the number refers to the maximum or peak gain in decibels (dB).

In one embodiment, hearing aids are classified based on steepness of a frequency response and degree of hearing loss at a particular frequency. Preferably, the classes of different types of hearing loss are equally spaced. That is, the hearing loss ranges are preferably separated by equal spacings such that the hearing loss of different classes at a predetermined frequency is 28, 40, 52, 64, and 84 dB.

A combination of these parameters (curve shape, gain range or peak gain) are sufficient to accurately describe many different types of acoustical formats. This renders it possible for a hearing aid provider to maintain a limited number of acoustical formats while providing hearing aids for many different types of hearing loss.

After a patient is diagnosed and a type of hearing loss is identified, the audiologist need only determine which acoustical format is appropriate to remedy the hearing loss. A pre-programmed hearing aid with a fixed frequency response to coπect a coπesponding type of hearing loss is then prescribed to the patient. In one embodiment, each hearing aid is programmed at a factory with a predetermined acoustical format and is not re-programmable. A matrix of different types of hearing aids each having a different acoustical format are then maintained at, for example, a local pharmacy. Preferably, hearing aids in the matrix are low- cost and disposable.

After an appropriate hearing aid is selected from the matrix of hearing aids and prescribed by an audiologist, the patient need only pick-up the prescription at a local pharmacy supplying such devices. Based on this method, there is no need to make adjustments to the hearing aid at the audiologist's office. Rather, a patient's type of hearing loss is identified by the audiologist and the corresponding type of hearing aid is prescribed to remedy the hearing impairment.

Accordingly, it is not necessary to maintain a matrix of hearing aids at the audiologist's office because there is no need to reprogram or custom fit a hearing device every time a patient needs a new hearing aid. Instead, the matrix of hearing aids is maintained at an outlet such as a pharmacy, reducing the time associated with handling large volumes of hearing aids at the audiologist's office. This results in lower overall costs for a hearing aid because the overhead to support the sale of hearing aids at a pharmacy is typically less than if the sale of hearing aids was out of an audiologist's office. Moreover, the time and expense associated with custom fitting a hearing aid to a patient is also reduced or eliminated because a hearing aid having fixed frequency response is prescribed to a patient, providing a remedy for the coπesponding type of hearing loss.

One aspect of the invention describes an analog filter suitable for hearing aid applications. The analog filter is described in the s-domain with the following normalized equations:

U(s) = ((s²) / (s²+γ-s+l))ⁿ (1)

V(s) = ((s-l) / (s+l))ⁿ (2)

X(s) = N(s) - U(s) (3) where s is the complex frequency jω, γ controls the resonance of the second-order filter section U(s), and n selects the number of sections and, hence, the maximum slope of the filter. Parameter n is an integer normally in the range of 1 to 4. U(s) defines a high-pass filter; N(s) defines an all-pass filter; and X(s) defines a low-pass filter. U(s) and X(s) are combined in various ratios to produce the desired transfer function of the hearing aid as follows:

T(s) = (1+α-β) X(s) + (1+ ) U(s) (4)

where (1+α) is the high-frequency gain and (1+ α-β) is the low frequency gain. The parameter α controls the filter and, in particular, the high-frequency gain while parameter β controls the amount of low- frequency gain relative to the high- frequency gain. The transfer function can be frequency scaled by letting s=jω / ω_c = jf / f_c, where ω_c (f_c) is the comer frequency of the high-pass transfer function U(s).

Equations 1-4 describe a family of transfer functions suitable for hearing aid applications. By varying f_c, γ, n, a and β, a wide range of transfer functions suitable for hearing aid applications can be achieved. Figures 7-12 show representative frequency responses for transfer function T(s) for some different values of f_c, γ, n, a and β.

Referring to Fig. 7, the independent variables, α, β, n, defining the transfer function, T(α, β, n), produce the family of frequency responses. Specifically, the family of frequency responses results from n = 1, 2, 3, and 4; fc = 2000 Hz; γ = 1.33; α = 64, and β = 0. More steeply sloped curves represent frequency responses of higher order filters (i.e., higher values of n).

In Fig. 8, the value of n remains constant (n = 1) and other independent variables remain the same as in Fig. 7, but the value of a = 0, 1, 2, 4, 8, 16, 32, and 64 controls the high frequency gain. Fig. 9 has the same family of curves as in Fig. 8, but with the value of β = 0.066, which accounts for the low-frequency gain relative to the high-frequency gain. Note that α-β controls the gain over a wider portion of the frequency spectrum.

In Fig. 10, the family of frequency responses differs from the family of frequency responses of Fig. 8 from changing n - \ X n = 2. Fig. 11 combines Fig. 10 with independent variable α = 0.033 rather than 0.0. Finally, the family of frequency responses in Fig.12 results from changing fc from 2000 Hz to 400 Hz and α = 8, but otherwise, all other independent variables are the same as those composing the family of frequency responses of Fig.8.

For convenience, the chart below summarizes the settings described above.

Fig. # n fc γ α β

7 1,2,3,4 2.0kHz 1.33 64 0.000

8 1 2.0kHz 1.33 0,1,2,4,8,16,32,64 0.000

9 1 2.0kHz 1.33 0,1,2,4,8,16,32,64 0.066

10 2 2.0kHz 1.33 0,1,2,4,8,16,32,64 0.000

11 2 2.0kHz 1.33 0,1,2,4,8,16,32,64 0.033

12 1 0.4kHz 8.00 0,1,2,4,8,16,32,64 0.000

A block diagram of the analog filter described by equations (1) through (4) above is shown in Fig.13. The signal path is shown as solid lines and the control signals are shown as dashed lines. An automatic gain control (AGC) circuit generates the control signal alpha (α) based on characteristics of the signal U(s). This analog filter includes three second-order high-pass filter sections and three first-order all-pass filter sections. It should be understood that the analog filter comprises analog components (e.g., resistors and capacitors) that relate to parameters α, β, γ and n.

For low-power hearing aid applications, it is desired to minimize the circuitry used. Therefore, in an alternate embodiment of the invention, the analog filter described above and defined by equations (1) through (4) is replaced with an analog filter that replaces the high-pass filter sections with band-pass filter sections and eliminates the all-pass filter sections. The alternate analog filter is suitable for hearing aid applications and its transfer function is described in the s-domain with the following normalized equations:

U(s) = (s² + 1) / (s² + δ-s + 1) (5)

X(s) = (δ-s)/(s² + δ-s+l) (6) T(s) = (1 + α-β-γ)-U(s) - (α-γ)-X(s)ⁿ (7)

where U(s) defines a band-reject filter (i.e., a notch filter), X(s) defines a band-pass filter, and T(s) defines the overall filter transfer function. Also, the parameter α controls the high-frequency gain, a controls the low-frequency gain relative to the high-frequency gain, δ controls the sharpness of the band-pass filter, and β controls the maximum high frequency gain in conjunction with α. The parameter n defines the number of cascaded band-pass filters, where n is normally in the range from one to three. Preferably, the δ parameter may take on a range from 1 to 2 and more preferably between 1.4 and 1.6. In the present embodiment of the invention, the δ parameter has a value of 1.538 (i.e., 1/0.65). While equations (5) through (7) have been normalized to a characteristic frequency of 1 radian sec, one skilled in the art will realize that a much higher characteristic is needed for a hearing aid application. Preferably, the characteristic frequency will be scaled to between 3000 Hz (18850 radian/sec) and 7000 Hz (43982 radian/sec). In the present embodiment of the invention, the characteristic frequency is scaled to 5000 Hz (31461 radian/sec).

The following description presents an electronic circuit providing a configurable high-order filter primarily for hearing aid applications, such as for generating transfer functions (5) - (7) described above. The electronic circuit embodies a filter that generally provides high-frequency amplification relative to low frequencies. The prefeπed embodiment of the invention uses less circuitry and, in particular, fewer op-amps (operational amplifiers) than non-prefeπed embodiments.

One circuit design uses a series of band-pass filter sections to implement the desired signal transfer function. To better understand the invention, one must understand the second-order biquad filter section and how this biquad filter can be implemented as a switched-capacitor filter. Fig. 14 shows a schematic diagram of the well known continuous-time, second-order biquad band-pass filter. As shown in the figure, two band-pass outputs exist. One is a non-inverting band-pass output while the other is an inverting band-pass output. Fig. 15 shows a schematic diagram of the same biquad band-pass filter implemented using switched-capacitor resistors. Depending on the configuration of the switched-capacitor resistor, the resistance may be either positive or negative (i.e., inverting switched-capacitor resistor). Since negative resistors can be implemented, the switched-capacitor biquad band-pass filter can use one fewer op-amp than the continuous-time filter of Fig. 14. Also, for the switched-capacitor biquad filter, only one band-pass filter output is available. To make either an inverting or non-inverting band-pass filter, the input resistor can be made either non-inverting or inverting, respectively. This is shown in Fig. 15 as resistors Rl-A (negative resistance for non- inverting band-pass output) and Rl-B (positive resistance for inverting band-pass output).

Fig. 16 shows a block diagram of the desired signal-processing algorithm described by equations (5) - (7) listed and described above. The circuit includes a band-reject filter and three band-pass filters. The output of the band-pass filters are X¹, X², - dX³, respectively. A selector (i.e., multiplexer) selects one of the bandpass filter outputs based on a control signal designated n. The output of the selector is designated X". The output of the selector goes to an automatic gain control (AGC) control circuit, which develops a control signal designated alpha. The signal alpha is multiplied by a constant factor gamma to generate the control signal designated alpha*gamma. The output of the band-reject filter is multiplied by a constant factor beta to generate a signal designated beta*U. The signal X" is subtracted from beta* U and multiplied by alρha*gamma to create a signal designated alpha*gamma*(beta*U-X") + U. Finally, the band-reject signal -7 is added to alpha*gamma*(beta*U-X) to create the output signal alpha*gamma*(beta*U-Xⁿ) + U. Given that X is the transfer function of the band-pass filters, each having the same center frequency, and Uϊs the transfer function of the band-reject filter having the same center frequency as X, then it can be shown that the final output signal, alpha*gamma*(beta*U-X") + U, provides maximum gain at the center frequency when alpha is large.

A straightforward implementation of the system shown in Fig. 16 uses at least 13 op-amps, excluding the AGC control circuit. This is assuming switched- capacitor filters using 2 op-amps for each band-pass filter and 3 op-amps for the band-reject filter. For low-power applications such as hearing aids, it is desirable to minimize the number of op-amps. By using fewer op-amps, three improvements are achieved: (1) less power is needed, (2) less silicon area is needed for a custom integrated circuit, and (3) less silicon area translates into lower cost. The invention describes a prefeπed embodiment of the signal-processing algorithm, shown in Figs. 17 and 18, in which only 9 op-amps are needed.

Fig. 17 shows a block diagram of the prefeπed embodiment of the signal- processing design. The circuit includes three band-pass filters. One band-pass filter is an inverting band-pass filter, while the other two band-pass filters are non- inverting. There is no discrete band-reject filter as in Fig. 16. The outputs of the three band-pass filters are designated - ^', -X², and -X respectively. A selector selects one of the band-pass filter outputs based on a control signal, designated n. The output of the selector is designated -X". The output of the selector goes to an AGC control circuit, which develops a control signal designated alpha.

An inverting summing amplifier sums the output of the selector -X", the output of the first (inverting) band-pass filter -X¹ weighted by a constant factor beta, and the input signal also weighted by a constant factor beta, to form an output signal that is weighted by another constant factor gamma and designated by - gamma* (beta*U-X"). The output of this first inverting summing amplifier goes through an inverting amplifier, with a gain factor controlled by alpha to generate an output signal designated alpha* gamma* (beta*U-X"). Finally, a second inverting summing amplifier, with equal weighting on all inputs, sums the output of said inverting amplifier with the output of the first (inverting) band-pass filter and the input signal to generate an output signal designated -(alpha*gamma*(beta*U-Xⁿ) + U).

Except for the negative sign, the output signal of the block diagram of Fig. 17 is identical to the output of the block diagram of Fig. 16. In an alternate embodiment of the invention (not shown), the final inverting summing amplifier in Fig. 17 is replaced with a non-inverting summing amplifier, and the output signal is designated (alpha*gamma*(beta*U-X") + U) (i.e., the leading negative sign is deleted). Fig. 18 shows a schematic diagram of the prefeπed embodiment of the present invention. The circuit implements the signal-processing design shown in Fig. 17. The first (inverting) band-pass filter comprises op-amps AR101 and AR102, resistors R101-R104 and capacitors C101 and C102. The second (non- inverting) band-pass filter comprises op-amps AR201 and AR202, resistors R201- R204 and capacitors C201 and C202. The third (non-inverting) band-pass filter comprises op-amps AR301 and AR302, resistors R301-R304 and capacitors C301 and C302. The selector comprises switches SI -S3. The first inverting summing amplifier comprises op-amp AR1 and resistors R1-R5, where resistor R3 sets the constant factor beta, and resistor R5 sets the constant factor gamma. The inverting amplifier comprises op-amp AR2 and resistors R6 and R6, where by varying the resistance of either resistor R6, R7, or both R6 and R7 varies the amplification factor alpha. Finally, the second inverting summing amplifier comprises op-amp AR3 and resistors R8-R11. The circuit of Fig. 15 comprises a total of nine op-amps. Although component values (resistors and capacitors) are not shown, one skilled in the art can easily determine a set of component values to achieve the desired signal- processing algorithm.

In another embodiment of the invention, op-amp AR3 and resistor Rl 1 of Fig. 18 may be eliminated. In this case, the final inverting summing amplifier is replaced with the resistive summing network comprising resistors R8-R10. This embodiment of the invention uses only eight op-amps. The output signal, taken at the junction of resistors R8-R10, is given by (113)* ((alpha* gamma* (beta* U-

In an embodiment as previously mentioned, a matrix of hearing aids is preferably defined by attributes of the hearing device. Preferably, defining a class of hearing aids includes separating types of hearing aids based on steepness of response gain in a range of frequencies and peak response gain at a predetermined frequency.

Fig. 19 is an example of a hearing aid matrix where hearing aids are classified according to their frequency response characteristics as described above. The 3 X 3 matrix classifies 9 different types of hearing aids. Each hearing aid is preferably pre-programmed at a factory with a unique acoustical format for remedying a certain type of hearing loss. In an alternate embodiment, a hearing aid is programmed with multiple types of acoustical formats, while only one of the acoustical formats is selected at a time.

Hearing aids in a column, such as F-20, F-26 and F-32, define a class of hearing aids having a similar frequency response characteristic but different peak gain values. "F" in the hearing aid identifier coπesponds to "flat," which describes the frequency response of a particular class of devices. For example, see acoustical format 510 as shown in Fig. 5. The steepness of the gain slope in mid-range frequencies 1000-1200 Hertz is relatively flat and, therefore, the acoustical format 510 would be classified accordingly as a class "F" type of hearing aid.

In a similar manner, the letter "S" in our exemplary 3 X 3 matrix stands for "steep," while the letter "P" stands for "precipitous" (very steep). Accordingly, classes of acoustical formats like acoustical format 520 (in Fig. 5) having a steep gain slope in mid-range frequencies 1000-1200 Hertz are assigned the letter "S." In a similar vane, classes of acoustical formats like acoustical format 510 (in Fig. 5) having a very steep gain slope in the mid-range frequencies 1000-1200 are assigned the letter "P."

It should be noted that different acoustical formats are generated as shown above using the desired filters to create a target response as shown in Figs. 7-12. For example, steeper responses as described for "P-type" hearing aids are created using higher order filters such as when n=3 as previously discussed. Conversely, hearing aids with flatter target responses are created using lower order filters such as when n=l as described above.

In furtherance of the method of classifying hearing aids in a matrix, a second parameter is used to further distinguish hearing aids having the same assigned letter. For example, hearing aids having a flat response, i.e., F-20, F-26 and F-32, as shown in Fig. 19 include respective numerals, i.e., 20, 26, and 32, coπesponding to the peak gain (in decibels) of the frequency response of a particular type of hearing device.

According to the classification method of the present invention, acoustical formats 510, 520 and 530 as shown in Fig. 5 could be classified as P-32, S-26 and F- 20 respectively. It should be noted that peak gain for this family of target responses corresponds with numerical value of the hearing aid, i.e., 32, 26 and 20. Consider acoustical format 510. It has a precipitous (very steep) gain slope in the middle of audible frequency range and a peak gain at 8000 hertz of 32 decibels. Hence, this hearing aid would be classified in the matrix as P-32. Hearing aids programmed with acoustical format 520 having a steep gain slope and a peak gain of 26 decibels would be classified in the matrix as S-26. In a similar way, hearing aids programmed with acoustical format 530 having a flat gain slope and peak gain of 20 would be classified in the matrix as F-20.

Fig. 20 is a schematic diagram of a hearing aid device utilizing separate channels to process frequency components of an audio input signal. Microphone 255 detects acoustical vibrations and produces an audio input signal 257 that is fed to each of multiple Channels 1 through N. Depending on the application, there are potentially two or more separate channels, i.e., N is an integer greater than 1.

Each channel includes a coπesponding bandpass filter 250-1 (Channel 1), 250-2 (Channel 2)...250-N (Channel N) to separate audio input signal 257 into bands of frequency components. For example, the bandpass filter 250-1 passes a band of lower frequencies such as 100-500 Hz (Hertz) for signal processing in channel 1, bandpass filter 250-2 passes frequencies such as 500-1000 Hz and bandpass 250-N passes a band of higher frequency components such as 10-12 KHz (Kilohertz) for signal processing in channel N. Based on this topology, frequency components of the audio input signal 257 are separated so that they can be processed individually. Accordingly, distortion caused by clipping in one channel will not effect the integrity of frequency components processed by the other channels.

The output of each channel's bandpass filters 250 is fed into a coπesponding non-linear amplifier 260 for a particular channel 1 through N. Each range of frequencies and is separately amplified and limited by the non-linearity of non-linear amplifier 260. Depending on the application, the non-linearity of the amplifier is optionally implemented using either hard-clipping or soft-clipping. When nonlinear amplifiers are utilized, soft clipping is prefeπed because distortions produced by the amplifier will generally be less damaging than when hard-clipping techniques are used.

In addition to programming the clipping characteristics of a channel, nonlinear amplifiers 260 are also programmed to provide the appropriate gain to compensate for the hearing loss of a hearing impaired patient. For example, if a patient has hearing loss in a particular frequency range, the gain of the amplifier 260 is adjusted for altering the components of the original audio signal so that an impaired patient hears as if he had more normal hearing.

In a addition to amplifier 260, each channel has a second bandpass filter 270 that matches the characteristics of the first bandpass filter 250. The inclusion of the second bandpass filter 270 is beneficial because it helps to reduce unwanted frequency components such as noise outside the bandpass of the channel, thus, producing a purer output. Output signals of the second bandpass filter 270-1 through 270-N are fed into summer circuit 285 to drive a sound producing device such as a speaker.

It should be noted that although the embodiment described above suggests the use of certain components, alternative components are optionally used to achieve the advantages according to the principles of the present invention. For example, linear amplifiers are optionally used to provide signal gain for a particular channel. Additionally, the components of fig. 20 are optionally embodied using analog circuitry rather than digital signal processors and related circuitry.

The previously described hearing aid has many advantages. For example, amplifier distortions caused by loud low frequency background noise will not effect high-frequency components of the audio input signal because they are separated by individually processed channels. Clipping in one channel, therefore, will not effect the performance of the other channel. Typically, low frequency background noise is responsible for clipping in amplifiers. Since the high-frequency channel is not clipped along with the low- frequency channel as found in previous applications, speech intelligibility is preserved. That is, high-frequency components are preserved for the listener even though there is distortion in one of the channels.

Another advantage is the minimal circuitry required to create a low-cost hearing aid. For example, the packaging of the circuit is minimal and therefore less invasive to a patient wearing the hearing aid. Also, it is possible to provide a signal processing algorithm implemented using low power analog (continuous-time) or digital (discrete-time) signal processing circuitry. Based on the simplicity of the algorithm, the circuit is optionally provided on a minuscule silicon chip powered by a coπespondingly small battery device.

Certain aspects of the present invention have been discussed in terms of a hearing aid application. However, such principles are optionally used in communication systems in which speech needs to be transmitted in the presence of noise and, in particular, low-frequency noise. For example, applications such as cellular telephones can potentially benefit from the principles of the present invention by reducing the effects of low-frequency "road" or background noise when using the telephone in a vehicle such as an automobile, bus or train.

While this invention has been particularly shown and described with references to prefeπed embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the invention encompassed by the appended claims.

Claims

CLAIMSWhat is claimed is:

1. A method of providing hearing aids to compensate for hearing loss comprising the steps of: segregating types of hearing loss into a predetermined number of classes; for each class, determining an acoustical format having a fixed frequency response to remedy a coπesponding type of hearing loss; and programming a hearing aid to have a fixed acoustical format to remedy a type of hearing loss.

2. A method as described in Claim 1, wherein the type of hearing loss is mild to moderate.

3. A method as described in Claim 1, wherein each predetermined acoustical format remedies a coπesponding type of hearing loss by providing various levels of gain across the audible spectrum.

4. A method as described in Claim 3, wherein gain of an acoustical format is higher for frequencies at the high end of the audible spectrum.

5. A method as described in Claim 1 further comprising the steps of: identifying a patient's type of hearing loss; and prescribing a hearing aid having the appropriate acoustical format to remedy a coπesponding type of hearing loss.

6. A method as described in Claim 1, wherein each acoustical format supports a range of gain in the audible spectrum, a difference in peak gain between formats being up to 12 decibels.

7. A method as described in Claim 1, wherein each acoustical format supports a range of gain in the audible spectrum, a difference in peak gain between formats is about 6 decibels.

8. A method as described in Claim 1, wherein hearing loss ranges from about 10 to 60 decibels.

9. A method as described in Claim 1 , wherein an eπor tolerance of manufactured hearing aids within a class is less than 3 decibels.

10. A method of tracking a set of hearing aids to remedy different types of hearing loss, the method comprising the steps of: segregating types of hearing loss into a predetermined number of classes; for each class, determining an acoustical format having a fixed frequency response to remedy the coπesponding type of hearing loss, each hearing aid being programmed to provide a coπesponding fixed frequency response for remedying a certain type of hearing loss; and classifying a hearing aid based upon attributes of a coπesponding acoustical format.

11. A method as described in Claim 10, wherein the step of classifying a hearing aid includes the step of: classifying a hearing aid based upon varying gain levels in a coπesponding frequency response.

12. A method as described in Claim 10, wherein the step of classifying a hearing aid includes the step of: classifying a hearing aid based upon a peak gain level in a coπesponding frequency response.

13. A method as described in Claim 10, wherein the step of classifying a hearing aid includes the step of: classifying a hearing aid based upon a degree to which gain changes in a coπesponding audible frequency range of an acoustical format.

14. A method as described in Claim 10, wherein each acoustical format supports a range of gain in the audible spectrum such that a difference in peak gain between is up to 12 decibels.

15. A method as described in Claim 10, wherein each acoustical format supports a range of gain in the audible spectrum such that a difference in peak gain between formats is about 6 decibels.

16. A method as described in Claim 10, wherein hearing loss ranges from about 10 to 60 decibels.

17. A system compri sing : a hearing aid having one of a predefined set of fixed acoustical formats to remedy a coπesponding type of hearing loss, the acoustical format providing a predetermined gain profile across an audible spectrum.

18. A system as described in Claim 17, wherein the coπesponding type of hearing loss is mild to moderate.

19. A system as described in Claim 17, wherein gain is higher for frequencies at the high end of the audible spectrum.

20. A system as described in Claim 17, wherein each acoustical format supports a range of gain in the audible spectrum, a difference in peak gain between formats is up to 12 decibels.

21. An analog filter, comprising: a first stage comprising a first channel and a second channel, the channels filtering respective all-frequency and high-frequency bands, the channels defined by respective transfer functions and processing an input signal and producing respective first and second, first stage, output signals; and a second stage comprising mathematical circuits having selectable coefficients for controlling a maximum slope between a low frequency band gain and a high frequency band gain, the second stage combining the first and second, first stage, output signals to produce a final, filtered, output signal.

22. The analog filter of Claim 21 wherein the first stage comprises at least one all-pass filter and one high-pass filter.

23. The analog filter of Claim 21 wherein the first stage comprises a plurality of all-pass filter sections and a plurality of high-pass filter sections.

24. The analog filter of Claim 23 wherein the first stage further comprises at least one selection circuit for selecting the number of all-pass and high-pass filter sections through which the input signal is processed.

25. The analog filter of Claim 21 wherein the second stage combines the first and second outputs in a third channel, the combination comprising a low frequency band.

26. The analog filter of Claim 21 wherein the mathematical circuits include at least one of the following circuit types: subtractors, adders, or multipliers.

27. The analog filter of Claim 21 wherein the second stage comprises an automatic gain control circuit providing an automatic gain control constant dependent on characteristics of the second, first stage, output signal.

28. The analog filter of Claim 21 wherein the transfer function of the analog filter is described in the s-domain with the following equation:

T(s) = (1 + α-β) X(s) + (1+α) U( ) , where:

U(s) = ((s²) / (s²+γ-s+l))ⁿ (high pass filter) V(s) = ((s- 1 ) / (s+ 1 ))^π (all pass filter)

X(s) = V(s) - U(s) (low pass filter)

U(s) and X(s) being combined in various ratios to produce the desired transfer function.

29. The analog filter of Claim 28 wherein the circuit comprises selectable components related to parameters a, β, γ and n.

30. An analog filter, comprising: means for filtering low- frequency and high-frequency bands, said means defined by a transfer function comprising the frequency bands and a region spanning between the frequency bands; means for providing flexibility for selecting a maximum slope of said region spanning between the frequency bands.

31. A method for filtering an analog signal, comprising: filtering an input signal into at least two frequency band outputs, a first frequency band signal comprising the input output comprising the input signal having about the same magnitude characteristics as the input signal and a second frequency band signal output having a higher magnitude at higher frequencies than at lower frequencies; and combining the first and second frequency band signal outputs through a combination of mathematical circuits having selectable coefficients for controlling a maximum slope between a low-frequency band gain and a high frequency band gain.

32. The method according to Claim 31 wherein filtering the input signal comprises filtering the input signal through a selectable number of filter sections.

33. The method according to Claim 32 wherein the filter sections comprise at least all-pass filters and high-pass filters.

34. The method according to Claim 31 wherein combining the first and second frequency band signals comprises forming a signal having a higher magnitude at lower frequencies than at higher frequencies.

35. The method according to Claim 31 wherein combining the first and second frequency band signals comprises determining an automatic gain control coefficient based on the second frequency band signal.

36. A method for filtering an analog signal, comprising: receiving an input signal; filtering the input signal by an analog circuit defined by the following transfer function:

T(s) = (1 + α-β) X(s) + (1+α) (s) , where:

U(s) = ((s²) / (s²+γ-s+l))ⁿ (high pass filter) N(s) = ((s- 1 ) / (s+ 1 ))ⁿ (all pass filter)

X(s) = N(s) - U(s) (low pass filter)

\J(s) and X(s) being combined in various ratios to produce the desired transfer function; and outputting the filtered input signal.

37. A hearing aid, comprising: a microphone converting an audible input signal into an electrical output signal; an analog circuit defined by the following transfer function: Υ(s) = (1 + α-β) X(s) + (1+α) υ(s) , where:

U(s) = ((s²) / (s²+γ-s+l))ⁿ (high pass filter) V(s) = ((s-1) / (s+l))ⁿ (all pass filter)

X(s) = V(s) - U(s) (low pass filter)

O(s) and X(s) being combined in various ratios to produce the desired transfer function, the analog circuit coupled to the microphone and processing the electrical output signal from the microphone; and a receiver coupled to the analog circuit for producing an audible output signal coπesponding to the processed electrical output signal, the audible output signal being an improvement in observable sounds over the audible input signal for a person using the hearing aid.

38. A device for implementing a signal processing transfer function, comprising: an input for receiving an input signal; an output for transmitting an output signal; and a circuit defined by a signal processing transfer function, said circuit comprising a cascade of second-order filter sections and other circuitry, said circuit coupled between said input and said output and processing said input signal to form said output signal.

39. The device of Claim 38, wherein an inverting filter section is a first section in the cascade of filter sections and non-inverting filter sections compose each of the remaining filter sections.

40. The device of Claim 39, wherein the filter sections are band-pass filters.

41. A device for implementing a signal processing transfer function in a hearing aid, comprising: in the hearing aid: a first circuit defined by a first transfer function; and a second circuit defined by a second transfer function, the second transfer function being determined by adding or subtracting the first transfer function with the input signal to produce a desired output signal.

42. The device of Claim 41, wherein the first transfer function comprises a bandpass filter.

43. The device of Claim 41, wherein the second transfer function comprises a band-reject filter.

44. A device for implementing a signal processing transfer function, comprising: a first inverting band-pass filter filtering an input signal; a second non-inverting band-pass filter filtering a first output signal from the first inverting band-pass filter; a third non-inverting band-pass filter filtering a second output signal from the second non-inverting band-pass filter; at least one amplifier; and means for selecting an output from at least one of the filters for amplification by the amplifier.

45. The device of Claim 44 wherein at least one band-pass filter consists of two op-amp circuits.

46. The device of Claim 44 wherein the amplifier comprises a first inverting summing amplifier.

47. The device of Claim 46, further comprising means for summing the input signal with the first output of said first inverting band-pass filter and weight the sum by a first factor as a second input to said first inverting summing amplifier.

48. The device of Claim 46, wherein the amplifier further comprises a second inverting summing amplifier.

49. The device of Claim 48, wherein the output of said first inverting summing amplifier is an input to said second inverting summing amplifier.

50. The device of Claim 49, further comprising means to weight the total sum of inputs to said second inverting summing amplifier by a second factor.

51. The device of Claim 48, wherein one input to said second inverting amplifier is the output of said first inverting summing amplifier, and the output of said second inverting amplifier is weighted by a third factor.

52. The device of Claim 51, wherein the amplifier further comprises a third inverting summing amplifier.

53. The device of Claim 52, wherein the output of said second inverting amplifier is an input to said third inverting summing amplifier.

54. The device of Claim 53, wherein a second and a third input to said third inverting summing amplifier comprises the input signal and the output of said first inverting band-pass filter, respectively.

55. The device of Claim 54, further comprising a means to weight the total sum of inputs to said third inverting summing amplifier by a fourth factor.

56. The device of Claim 55, wherein each summing amplifier consists of a single op-amp and passive circuit elements.

57. A method for implementing a signal processing transfer function, comprising: filtering an input signal by a first inverting band-pass filter to produce a first filtered signal; filtering the first filtered signal by a second non-inverting band-pass filter to produce a second filtered signal; filtering the second filtered signal by a third non-inverting band-pass filter; selecting the first, second, or third filtered signal to be amplified; and amplifying the selected filtered signal.

58. The method of Claim 57 wherein at least one band-pass filter consists of two op-amp circuits.

59. The method of Claim 57 wherein said amplifying is performed by at least one summing amplifier.

60. An apparatus for a hearing aid, comprising: a microphone converting an audible signal to an electrical signal; a filter for filtering the electrical signal, the filter defined by the transfer function:

U(s) = (s² + l) / (s² + δ-s + l) (5)

X(s) = (δ-s) / (s² + δ-s + 1) (6)

T(s) = (1 + α-β-γ)-U(s) - (α-γ)-X(s)" (7) where U(s) defines a band-reject filter, X(s) defines a bandpass filter, and T(s) defines the overall filter transfer function; and a receiver converting the filtered electrical signal into an output audible signal for a hearing aid user.

61. A method amplifying an audible signal by a hearing aid, comprising: converting an audible signal to an electrical signal; filtering the electrical signal, the filter defined by the transfer function:

U(s) = (s² + 1) / (s² + δ-s + 1) (5)

X(s) = (δ-s) / (s² + δ-s + 1) (6)

T(s) = (1 + α-β-γ)-U(s) - (α-γ)-X(s)" (7) where U(s) defines a band-reject filter, X(s) defines a bandpass filter, and T(s) defines the overall filter transfer function; and converting the filtered electrical signal into an output audible signal for a hearing aid user.

62. A hearing aid apparatus comprising: an audio input signal; a first channel for receiving the audio input signal and amplifying a first range of frequencies thereof to produce a first channel output signal; a second channel for receiving the audio input signal and amplifying a second range of frequencies thereof to produce a first channel output signal; and a summer circuit for combining the channel output signals.

63. A hearing aid apparatus as described in claim 62, wherein the amplification of the predetermined range of frequencies for a given channel is selected to compensate for a coπesponding hearing loss of a hearing impaired patient.

64. A hearing aid apparatus as described in claim 62 further comprising: a microphone that generates the audio input signal; and a speaker for generating an audio output based on the combined channel output signals.

65. A hearing aid apparatus as described in claim 62 further comprising: additional channels, each channel receiving the audio input signal and amplifying a selected range of frequencies thereof to produce a coπesponding channel output signal, the summer circuit combining channel outputs to produce an output audio signal.

66. A hearing aid apparatus as described in claim 65, wherein the range of frequencies for each of the channels is selected so that the ranges are contiguous with each other.

67. A hearing aid apparatus as described in claim 62, wherein each of the channels includes a bandpass filter circuit separating the audio input signal into frequency ranges.

68. A hearing aid apparatus as described in claim 67, wherein the output of the filters are fed into coπesponding non-linear amplifier circuits to amplify frequency components of the audio input signal.

69. A hearing aid apparatus as described in claim 68, wherein an output of each non-linear amplifier is fed into a coπesponding second filter circuit.

70. A hearing aid apparatus as described in claim 62, wherein the apparatus is implemented as a low power signal processing algorithm in a digital device.

71. A hearing aid apparatus as described in claim 62, wherein the amplifiers of each channel provide soft clipping.

72. A method of processing signals for a hearing aid, the method comprising: providing an audio input signal; receiving the audio input signal and amplifying a first range of frequencies thereof to produce a first channel output signal; receiving the audio input signal and amplifying a second range of frequencies thereof to produce a first channel output signal; and combining the channel output signals to produce an audio output signal.

73. A method as described in claim 72, wherein an amplification of the predetermined range of frequencies for a given channel is selected to compensate for a coπesponding hearing loss of a hearing impaired patient.

74. A method as described in claim 72 further comprising: receiving the audio input signal and amplifying additional selected ranges of frequencies thereof; and combining the channel outputs to produce an output audio signal.

75. A method as described in claim 72, wherein the range of frequencies for each of the channels is selected so that the ranges are contiguous with each other.

76. A method as described in claim 72, wherein each of the channels includes a bandpass filter circuit for separating the audio input signal into coπesponding frequency ranges.

77. A method as described in claim 76 further comprising the step of: feeding the output of the filters of each channel into coπesponding non-linear amplifier circuits to amplify frequency components of the audio input signal.

78. A method as described in claim 77, wherein an output of each non- linear amplifier is fed into a coπesponding second filter circuit.