EP0071845A2 - Apparatus for compensating hearing deficiencies - Google Patents

Abstract

The invention relates to a device for compensating for hearing defects, in which a parallel arrangement of a plurality of signal branches is arranged behind the element receiving the input sound signals (transducer, such as a microphone (1), hearing coil or an audio input connected to a radio, etc.), of which Each consists of a frequency-selective filter (4a to 4n), a level-dependent gain control (5a to 6n) and an arrangement for non-linear signal deformation (9a to 9n), followed by a summing amplifier (11) which summarizes the partial signals and which is connected via a power amplifier (12 ) is connected to an output signal converter (13). In this device, an arrangement is specified according to the invention, which allows multi-channel processing of the input signal with regard to space requirements and power consumption even in hearing aids to be worn on the head, in front of and / or behind the level-dependent gain control (5a to 6n) and the non-linear signal deformation (9a to 9n) lying filters (4a to 4n) are designed as time-discrete and amplitude-dependent filters. Devices according to the invention are particularly suitable for use as hearing prostheses.

Description

The invention relates to a device for compensating for hearing damage according to the preamble of claim 1. Devices of this type are e.g. described in Scand.Audiol. 8: 121-126, 1979, as "programmable hearing aid with multichannel compression" by S.Mangold and A.Leijon (see in particular page 121, right column, last paragraph including page 122, right column, paragraph 4).

In the known device, the electrical input signal, which is generated, for example, in a microphone or an induction pick-up coil, is fed to a plurality of filters, which each pass adjacent sections of the offered frequency range. The individual parts of the signal are then influenced with regard to the hearing loss, which must be compensated for, by compression and change in the amplitudes. Finally, the various signals from the so-called channels are brought together again and fed to the hearing impaired person via an output converter. The control of the filters as well as the compression and the volume control takes place via a memory that has been programmed with data about the hearing loss to be compensated for or with data derived therefrom, for example by inputting this data through an audiometer via a data input of the hearing aid. .

• 1. Soll die Hörhilfe auch schwergradige Hörstörungen ausgleichen können (z.B. starke Hochtonverluste), so werden Filterschaltungen notwendig, die viel Raum und Strom beanspruchen, so daß ein Einbau in Hinter-dem-Ohr-Geräte erschwert ist.
• 2. Es ergeben sich Genauigkeits- und Temperaturstabilitätsprobleme bei den Widerständen und Kondensatoren, insbesondere, wenn die Filter in integrierter Schaltungstechnik realisiert werden sollen.
• 3. Die Einstellung der Filtercharakteristik mit der für eine universell anwendbare Hörhilfe nötigen Variationsbreite und Genauigkeit erfordert sehr aufwendige Schaltungen (z.B. Digital-Analog-Wandler und Analog-Multiplizierer).

Although the analog signal processing implemented in the known device represents a basically simple method and has so far been used in hearing aid technology corresponding technology, there are the following disadvantages in the implementation:

• 1. If the hearing aid should also be able to compensate for severe hearing disorders (for example, strong high-frequency losses), filter circuits are required which take up a lot of space and electricity, so that installation in behind-the-ear devices is difficult.
• 2. There are problems of accuracy and temperature stability with the resistors and capacitors, especially if the filters are to be implemented in integrated circuit technology.
• 3. The setting of the filter characteristic with the range and accuracy necessary for a universally applicable hearing aid requires very complex circuits (for example digital-analog converter and analog multiplier).

The disadvantages mentioned under 2 and 3 are avoided if the signal processing is completely digital, i.e. time-discrete and amplitude-quantized, is carried out. Such a hearing aid working with integrated logic circuits is known from US Pat. No. 4,187,413. Because of the effort for the analog-digital converter at the input and the digital-analog converter at the output, the difficulty mentioned under 1. remains. In particular, the high current requirement of such circuits can be covered only with difficulty from the batteries which can be used in the installation space which is already limited by the circuit in behind-the-ear devices.

The invention is based on the object of specifying an arrangement for a device for compensating for hearing damage according to the preamble of patent claim 1 Visible space requirement and power consumption also enables multi-channel processing of the input signal in hearing aids to be worn on the head, which can be controlled from a memory. The above object is achieved according to the invention by the measures mentioned in the characterizing part of claim 1. Advantageous further developments and refinements can be found in the subclaims.

The use of discrete-time and amplitude-analog filters means that complex circuits are avoided, so that implementation in the size of commercially available pocket hearing aids or behind-the-ear hearing aids is made considerably easier. This is possible with the now known time-discrete integrated filter circuits, which have all the advantages of pure digital filters that are essential for hearing aid applications, but which do not have any analog-digital and analog signals because of the analog representation of the state variables. Digital-to-analog converters require more. These are preferably switched capacitor filters ("SCF"), chain storage filters ("bucked brigade devices" -BBD) and filters with charge-coupled devices ("charge coupled devices" -CCD). This makes it possible to equip small pocket hearing aids and behind-the-ear hearing aids with time-discrete filters. Because the filters mentioned can also be constructed in such a way that their coefficients can be changed very quickly by digital control signals, it is possible according to the invention to carry out a multi-channel adaptive optimal filtering in the hearing device. This also enables the targeted reduction of noise, as described in more detail in US-PT 40 25 721.

The output signals from time-discrete filters operating in an amplitude-analog manner and from digital-to-analog converters are in the form of a staircase curve. This means that their spectrum contains repetitions of the signal spectrum at multiples of the sampling frequency (known e.g. from A.B. Carlson, Communication Systems, McGraw Hill, New York, 1968, Sect. 7.1 - 7.2, pages 272 to 289). If parts of these repetition spectra fall within the audible frequency range, they become audible as distortions. For this reason, these repetition spectra are usually suppressed by an analog low-pass filter (a so-called "smoothing filter").

It has proven to be particularly expedient to select the operating clock frequency of the time-discrete filters higher than the sum of the upper limit frequency of the hearing ability and the limit frequency of the input amplifier, because the repetition spectra mentioned are then completely above the audible frequency range. The limit frequency is to be understood as the frequency at which a limit value of the frequency response (e.g. -60 dB) is finally undershot. In this way, the above-mentioned distortions can no longer be heard in a simple manner and there is no need for their screening out.

The discrete-time filters used have the advantage that they can also be produced as integrated circuits both in thick or thin film and in monolithic integration technology. This enables highly complex circuits to be implemented in a small space. The time-discrete mode of operation has the advantage that the problems with stability and temperature behavior known from integrated analog circuits can largely be avoided, and thus also the circuits with discrete components which are often required to stabilize the integrated circuits. Especially scarf Ter capacitor filters can be integrated particularly advantageously in complementary metal-oxide-silicon (CMOS) technology to form circuits which are characterized by a small space requirement, maximum time and temperature constancy, and very low supply voltages and currents.

The invention includes multi-channel hearing aids of any number of channels, ie devices with generally n parallel frequency-selective filters, whose pass bands overlap at most slightly in the falling edges of the frequency response, n ^Z 2 being selected. In view of the intended optimal compensation of as many practically occurring hearing disorders as possible, a desirable upper limit of the number of channels n at the current state of knowledge is the number of frequency groups ("critical bands") of the hearing, which is given as 24 (according to E. Zwicker, Scaling, in: WDKeidel and WDNeff (Ed.), Handbook of Sensory Physiology, Vol. V, Part 2, Springer, Berlin 1975, Section III.A, pages 409 to 414).

Such high numbers of channels are currently not feasible due to the space and power requirements of the necessary circuit elements. However, it has been found that three-channel devices already allow a much better adaptation than conventional hearing aids if the passband of the filters correspond to those frequency bands that the most important formants occupy on average. This would make the first range between the lower frequency limit of the sound transducers (approx. 50 Hz) and approx. 600 Hz, the second between approx. 600 Hz and approx. 2.5 kHz and the third between approx. 2.5 kHz and the through the transducers set upper limit (currently 8 to 10 kHz). With such devices, the hearing impairment can be compensated with sufficient accuracy in very many cases will; it also prevents strong low-frequency interference signals (eg traffic or machine noise) from adversely affecting the gain control in the higher-frequency channels that are particularly important for speech intelligibility, ie in particular at approximately 1 to approximately 8 kHz.

It has proven to be expedient to provide only one volume control, whose output signal influences the amplification of one signal amplifier in each subchannel. This avoids the installation of multiple potentiometers, which are problematic in terms of their space requirements and the difficult synchronization synchronization. At the same time, an individual actuator characteristic curve can be realized in each channel, determined by the type or presetting of the respective amplifier.

It has also proven to be advantageous, before or after the additive combination of the partial signals, to filter out distortion components, which may result from the nonlinear signal deformation due to the automatic gain control (AGC) and the peak value limitation (PC), from the partial signals or from the sum signal to effect. For this purpose, low-pass filters or band-pass filters can be used, whose frequency responses approximate those of the filters for channel separation described above. Depending on the degree of interference suppression required, simple passive RC filters, integrated active RC circuits or, in turn, time-discrete filters can be used.

The use of time-discrete filters makes it easy to change the filter characteristics (frequency limits and gains) over a wide adjustment range. This is done appropriately partially in that the setting parameters are digitally coded in an external device, most advantageously already in the audiometer, and transmitted serially to the hearing aid via a double line or in parallel via several lines. This data is stored in a programming circuit, which derives setting signals therefrom in a manner known in principle (above-mentioned publication by Mangold and Leijon; US Pat. No. 4,187,413) and feeds it to the filters. As also known in principle, it proves to be expedient to also set the parameters of the gain control and peak value limiting circuits (for example basic gain, control use, static and dynamic characteristic curve profile) by means of further data transmitted to the programming circuit.

The parameter memory of the programming circuit is expediently designed to be erasable, for example in the form of a programmable read-only memory that can be erased by ultraviolet light or electrical voltage (erasable programmable read-only memory (EPROM) or electrically alterable read-only memory (EAROM)). This makes it possible to later process the hearing aid data that has been permanently programmed for a longer period, e.g. in a further audiometric examination of the hearing device wearer in accordance with the change in hearing damage that has now occurred.

An extension of the programming circuit, which has proven to be expedient in many cases, can be obtained in that, in addition to the storage of predetermined basic data, a continuous change in the hearing aid data that is dependent on the input signal is made possible by the programming circuit itself, for example by implementing this circuit by means of a microcomputer circuit. This eliminates an adaptive noise signal pressure by optimal filtering possible, as is known from US Patent 40 25 721. However, the principle implemented there in only one channel can be expanded by the invention to multi-channel optimal filtering in all frequency channels.

Further details and advantages of the invention are further explained below with reference to the embodiment shown in the figure.

In the figure, a hearing aid equipped with filters is drawn in a schematic block diagram.

In the device shown, a microphone 1 is provided as an input converter, which is connected to a preamplifier 2, which, as indicated by 2 ', has a low-pass frequency response. The signal thus amplified is then divided at a point 3 to a plurality, i.e. a total of n time-discrete frequency filters 4a to 4n, distributed. Of these, the one designated 4a is a bandpass filter which passes frequencies from 50 to 600 Hz. The filter 4b also connected to point 3 is a bandpass filter, which is effective at frequencies from 0.6 to 2.5 kHz. If the frequency range of the filters 4a and 4b is reduced, several filters can then be provided, as indicated by points 4c. Finally, the filter 4n follows last, which is effective in the frequency distribution specified for 4a and 4b at 2.5 to approx. 8 kHz.

,

The filters are then followed by controllable amplifiers 5a to 5n which, together with controllers 6a to 6n, implement gain control in a manner known in principle. Here too, the arrangement of further control amplifiers is denoted by 5c and regulators by 6c. The signals then arrive at controllable amplifiers 7a to 7n, which are controlled make the volume adjustment using the output voltage of the volume control 8.

The signals are then subjected to a peak limitation in the nonlinear elements 9a to 9n in a known manner. Signal distortions caused thereby are reduced by post-filtering with filters 10a to 10n, which frequency response can correspond to filters 4a to 4n, for example. In the case of the control amplifiers 7a to 7n, the limiters and the distortion-reducing filters 10a to 10n, 7c, 9c and 10c also indicate that additional channels can be added.

The signals treated in this way are finally combined again at a point 11 and fed via a power amplifier 12 to a receiver 13 as an output converter.

The setting of filters 4a to 4n, controllers 6a to 6n. and peak limiters 9a to 9n are carried out by a programming circuit 14. The filters 4a to 4n receive their control signals via the lines 15a to 15n; The same applies to the controllers 6a to 6n via the lines 16a to 16n, to the limiters 9a to 9n via the lines 17a to 17n and finally to the filters 10a to 10n via the lines 18a to 18n.

The programming circuit 14 in turn receives the setting data from an external device (e.g. an audiometer) via one or more data lines 19, the transmission and the storage in the programming circuit 14 being controlled via a plurality of control lines 20 from the external device. The connection to the latter is mediated by a plug connection 21.

If the programming circuit 14 by a microcompu terschaltkreis realized, it can calculate the setting parameters completely or partially yourself, depending on the currently available input signal, which is supplied to it for this purpose via line 22.

The mode of operation of the device results from the fact that the input signal converter, i. in the microphone 1 or in its place induction pick-up coil for electromagnetic vibrations, generated electrical signal in the amplifier 2 is raised to such a voltage level that it is easily accessible to the subsequent signal processing. The low-pass frequency response 2 'contained in the amplifier 2 prevents signal portions and possibly coupled interference signals, which are above half the sampling frequency, from being folded back into the audible frequency range during the sampling process to be carried out in the discrete-time filters 4a to 4n.

The signal in the filters 4a to 4n is then sampled and frequency-selectively suppressed to such an extent that the respective parts of the signal belonging to the specified frequency ranges can be treated separately. In the control amplifiers 5a to 5n, which are controlled by the controllers 6a to 6n, a gain control that is dependent on the input or output level is achieved, whereby various known control principles can be used, for example the usual AGC circuits using the short-term average of these levels, but also instantaneous value compressors, as specified by Keidel and Spreng in DE-AS 15 12 720. This enables extensive compensation for disturbances in hearing dynamics (e.g. loudness compensation - recruitment).

By means of the actuator 8 and the control amplifiers 7a to 7n controlled thereby, the hearing aid wearer has the possibility of bringing the volume of the output signal into a volume range which is comfortable for him. In principle, any nonlinear signal deformation can be achieved with the nonlinear circuits 9a to 9n. In the normal case, a peak value limitation is carried out in a known manner and thus the occurrence of unpleasant or even hearing-damaging peak values of the output sound pressure level is prevented.

The distortion components caused by these nonlinearities are reduced in the filters 10a to 10n, but the useful signals are left unaffected as far as possible. Filters 10a to 10n can be dispensed with if the interference suppression due to the low-pass characteristics of power amplifier 12 and receiver 13 is sufficient. After the combination of the partial signals at the addition point 11, the further treatment of the sum signal takes place in the usual way, i.e. it is in the amplifier 12 to operate the output converter, i.e. in the present case, the receiver 13, brought the necessary intensity. A signal then appears on the receiver 13 which is suitable for compensating for the particular hearing loss.

In the case of hearing loss, for example where the hearing ability for high frequencies is mainly impaired and, in addition, loudness compensation (recruitment) essentially only occurs in this area, the (unregulated) basic amplification of the frequency channels on the amplifiers 5a to 5n must be set in a known manner that the pathological course of the patient's hearing threshold is compensated for on average as best as possible. The controllers 6a to 6n are now set so that the loss of dynamics in the respective Fre quenzband is balanced as well as possible, ie the controller 6n in the high-frequency channel will bring about a significant gain reduction at high levels, while the controller 6a in the low-pass channel remains almost without influence. Finally, the limiters 9a to 9n are to be set in a known manner in such a way that the patient's discomfort threshold is not significantly exceeded by the signal level at any frequency. Are the filters 10a _. installed up to 10n, they must be dimensioned so that distortion components are largely suppressed (for example, by designing the frequency response as duplicates of the corresponding channel separation filters 4a to 4n).

If the programming circuit 14 is a microcomputer circuit operating in the sense of an adaptive optimal filter, this will only maintain the basic setting described above if it is only in the input signal supplied via line 22 according to methods described in US Pat. No. 4,025,721 Speech, but no significant interference signal components. However, if noise components are detected, the gain in each channel is automatically reduced in the sense of the optimal filter function, the greater the ratio of noise level to speech signal level in the channel concerned.

The data which are fed to the programming circuit 14 via the plug connection 21 can be taken from an external device, for example an audiometer. For this purpose, it is necessary that the transmitting part of a data interface is installed in the external device, while the programming circuit 14 is designed so that it fulfills the function of the associated receiving part. The data transfer from the external device to the hearing aid can be according to the signal plan of a standardized cut position (e.g. CCITT-V.24 according to DIN 66020), only the signal levels have to be adjusted to the operating voltage of the hearing aid. After the transmission, an agreed data word or control signal causes the non-volatile storage in an EPROM or EAROM. Later reprogramming is easily possible by deleting the non-volatile memory (EPROM or EAROM) according to its design (using ultraviolet radiation or electrical voltages) and transferring a new data record.

Claims (12)

1. A device for compensating for hearing defects, in which a parallel arrangement of several signal branches is arranged behind the element receiving the input sound signals (transducer, such as a microphone, hearing coil or an audio input connected to a radio, etc.), each of which has a frequency-selective filter, there is a level-dependent gain control and an arrangement for non-linear signal deformation, followed by a summing amplifier which summarizes the partial signals and which is connected to an output signal converter via a power amplifier, characterized in that the filters lying in front of and / or behind the level-dependent gain control and the non-linear signal deformation act as are time-discrete and designed as amplitude-dependent filters. 2. Apparatus according to claim 1, characterized in that in each parallel signal branch between the arrangement for non-linear signal deformation and the summing amplifier combining the partial signals, a filter is additionally inserted in each case, which reduces the distortion components of the respective partial signal resulting from the non-linear signal influencing. 3. Apparatus according to claim 2, characterized in that the filters reducing the distortion components are designed as discrete-time and amplitude-dependent filters. 4. Apparatus according to claim 1, characterized in that the operating clock frequency of the discrete-time filter is higher than the sum of the upper Frequency limit of hearing and the upper limit frequency of the input amplifier. 5. Apparatus according to claim 1, characterized in that the filters are designed as integrated circuits. 6. Apparatus according to claim 1, characterized in that the arrangement comprises a programming circuit which contains data for influencing the coefficients of the discrete-time filter, the parameters of the gain controller and the non-linear circuitry and the coefficient of the distortion-reducing filter. 7. Apparatus according to claim 6, characterized in that the programming circuit is supplied with setting data from an external data generation device, such as an audio meter. 8. Apparatus according to claim 1, characterized in that the frequency-selective filters are switch-capacitor filters ("Switched Capacitor Filters"). 9. Apparatus according to claim 1, characterized by a microphone (1) as an input signal converter, a subsequent preamplifier (2, 2 ') having a low-pass frequency response, the output of which is connected to at least two (generally n) time-discrete filters (4a to 4n) connected in parallel, on the at least in one channel (generally in all n channels) a control amplifier (5a to 5n) with controller (6a to 6n) for automatic gain control (6a to 6n), each with a common volume control (8) whose gain can be changed Signal amplifier (7a to 7n), A non-linear element (9a to 9n) and a frequency-selective filter for attenuation of the distortion component (10a to 10n) each follow and that the outputs of these parallel circuits are brought together at a summing point (11) at which an output amplifier (12) and finally an output converter ( 13) and a programming circuit (14), via which the filter coefficients of the filters (4a to 4n and 10a to 10n) and the parameters of the controllers (6a to 6n) and the non-linearities (9a to 9n) can be set. 10. Apparatus according to claim 9, characterized in that the programming circuit (14) via a plurality of lines (19 and 20) and a plug connection (21) can be supplied with setting data from an external device. 11. The device according to claim 10, characterized in that the programming circuit (14) is an integrated microcomputer circuit. 12. Apparatus according to claim 11, characterized in that programs and data are stored and effective in the microcomputer circuit which enable a setting of the time-discrete filter (4a to 4n) as an adaptive optimal filter which is dependent on the input signal pending at point 3.

Images