EP1458216B1 - Device and method for adaption of microphones in a hearing aid - Google Patents

Abstract

The system has two microphone input signals, Mic In (2) and Ref Mic (7). The signals are fed to first (1) and second (5) microphones with given transmission functions (s/s-pol-ac1) and (s/s- pol-ac2). The output of the first microphone goes to a compensation filter (3') with a transmission function dependent on the first two transmission functions (s-pol-ac1/s-pol-ac2). The output from this filter goes to a first output Out Mic 1 (4). The output of the second microphone goes to a second output Out Mic 2 (8).

Description

The present invention relates to a method for the mutual adaptation of a plurality of microphones of a hearing device. Moreover, the present invention relates to a corresponding device for adapting the microphones.

Hearing impaired people often suffer from a reduced ability to communicate in noise. To improve the signal / noise ratio directional microphone arrays have been used for some time, the benefits to the hearing impaired is undisputed. Often, either first-order systems, i. H. used with two microphones, or higher order. The exclusion of backward received interference signals as well as the focus on frontally incident sounds facilitate a better understanding in everyday situations.

Directional microphones, however, are sensitive to detuning of the transfer functions of the microphones by amount and phase. The sensitivity to moods increases with the order of the directional microphone system and with decreasing frequency. At low frequencies, such directional microphone systems are the most sensitive.

In the document EP 0982971 A2 It is stated in this connection that a microphone at low frequencies can be determined by a first order high pass. Accordingly, according to FIG. 1 Characterize a first microphone 1 by a high pass with the transfer function s / s-pol_ac1. The microphone 1 receives a first input signal 2. This filtered with the high-pass filter of the microphone 1 input signal 2 is converted by means of a first compensation filter 3 in a first microphone output signal 4. The Compensation filter 3 has the transfer function s-pol_ac1 / s-pol_ideal. Both numerator and denominator can be represented as polynomial. The numerator polynomial of the compensation filter 3 is chosen to correspond to the denominator polynomial of the high-pass acoustic impedance of the microphone 1. The denominator polynomial of the compensation filter 3 corresponds to the denominator polynomial of the high pass of an ideal microphone. By multiplying the two transfer functions of the high-pass filter, which characterizes the real microphone 1, and the compensation filter 3, a normalization with respect to the ideal microphone results and the specific transfer function of the first microphone is compensated.

When looking at hearing aid microphones, it has been shown that, in a simplified approach, in particular the acoustic high-pass filter present at the lower edge of the usable frequency band must be examined with regard to detuning. Dirt, aging or altered environmental influences have a particularly strong effect on this high pass and thus change the amplitude and frequency response of the microphone in the particularly critical, middle and lower frequency range. One way to reduce such detuning is to force the same high-pass frequency in all microphone paths.

In the same way, the specific high-pass with the transfer function s / s-pol_ac2 of the second microphone 5 is compensated by a second compensation filter 6 with the transfer function s-pol_ac2 / s-pol_ideal, so that a corresponding second microphone output signal 8 is formed from the second microphone input signal 7. Again, the denominator polynomial of the high-pass filter 5 is eliminated by the numerator polynomial of the second compensation filter 6. With these two compensation filters 3 and 6, the fluctuations of the high-pass cutoff frequency from microphone to microphone can lead to phase and amplitude errors, especially at low frequencies would be compensated for by setting the same cutoff frequencies in all microphone paths.

In the further document US 6,272,229 B1 A method for relative, adaptive phasing of two microphones is roughly sketched. A general block diagram for an adaptive system is given. The system includes a block "acoustical delay compensation", which compensates in a kind of preprocessing the linear phase difference of the microphones, which is due to the signal propagation delay between the microphones. An adaptation rule is not specified.

Furthermore, from the document EP 1 191 817 A1 a hearing aid with adaptive microphone adaptation known. A processor of the hearing aid is capable of determining a difference in mean signal levels of two input signals. From this and with the aid of IIR or FIR filters, the frequency response of a channel can be corrected. Optionally, processing takes place in a frequency range from 100 Hz to 1 kHz.

Further internal implementations primarily address the input sensitivity difference of the microphones. A temporally averaged view of the input levels at the microphones can be drawn on the input sensitivity of the microphones. Assuming that the incident switching signals are received by all microphones with a time delay but at almost the same level, the amplitude of the input sensitivities can be adjusted by adjusting the average input levels at the microphones.

The object of the present invention is to simplify the compensation of microphone differences in hearing aids.

According to the invention, this object is achieved according to claim 1 by a method for mutual adaptation of a plurality of microphones of a hearing aid, by measuring a first amplitude of a first output signal from a first of the plurality of microphones in a predetermined frequency range, measuring a second amplitude of a second output signal from a second one of the plurality Microphones in the predetermined frequency range and filtering the first output signal in response to the first amplitude and the second amplitude, so that the difference between the two output signals is reduced. The filtering is done by multiplying by a denominator and counter polynomial, but only the numerator polynomial is varied by a closed-loop control. The predetermined frequency range for measuring the amplitudes of the two output signals of the microphones corresponds to several frequency bands below 150 Hz. In particular, the frequency band is between 40 and 60 Hz or 80 to 120 Hz. This is the range in which there are differences in the cutoff frequency of the high-pass filters make the microphones particularly noticeable.

Furthermore, a corresponding device according to claim 5 is provided according to the invention.

Compared to the state of the art FIG. 1 can be dispensed with by the invention to a compensation filter in a microphone path, the reference path. Each compensation filter is thus contained in each path except the reference path. This means that, for example, with three microphones in two microphone paths, a compensation filter is to be provided, while the third microphone path is used as the reference path.

The filtering may be adjusted by a control loop so that the first and second amplitudes correspond to each other. This makes it possible to change the timing of the transfer function the microphones, for example, by pollution or aging effectively counter.

The compensation filtering can be divided into two sub-filters. A first partial filtering is realized by a denominator polynomial that models the high-pass frequency of the reference path. A second sub-filter is realized by a counter polynomial adapted to minimize the average level difference between the microphone paths. The adaptation takes place by amount formation of the signals, whereby a phase dependence is eliminated. This can be dispensed with a unit such as the above-mentioned "acoustical delay compensation" block.

Preferably, the coefficients of the counter polynomial depend only on a single parameter. This leads to a low effort in the adaptation. If only the counter polynomial is adaptable, this will in principle not lead to identically identical microphone signals, since there may be an error between the characteristic of the reference microphone and the filter effect described in the denominator polynomial. However, the effect of this good approximate solution is sufficient to significantly improve the directivity with minimal effort.

An optimal adaptation of the two or more microphones to each other is possible, although the denominator polynomial is variable. This additional adaptation option also ensures a faster adaptation by the control loop.

Advantageously, the amount and / or phase of the first output signal can be modified by the filtering. This can be used to improve the directional microphone setting.

The advantage of an adaptation with the microphone model compared to an adaptation with a filter that can emulate any phase functions, lies on the one hand in the simplicity of realization. On the other hand, it is fundamentally advantageous to start from a simplified model presentation and to align the compensation specifically to the model.

FIG 1

ein Blockschaltbild zur Kompensation von Verschiebungen von Hochpasseckfrequenzen gemäß dem Stand der Technik;

FIG 2

ein Blockschaltbild zur Kompensation von Verschiebungen von Hochpasseckfrequenzen gemäß der vorliegenden Erfindung;

FIG 3

ein Schaltungsdiagramm einer Kompensationsschaltung gemäß einer ersten Ausführungsform der vorliegenden Erfindung; und

FIG 4

ein Schaltungsdiagramm einer Kompensationsschaltung gemäß einer zweiten Ausführungsform die zum Verständnis der vorliegenden Erfindung hilfreich ist.

The present invention will now be explained in more detail with reference to the accompanying drawings, in which:

FIG. 1

a block diagram for compensation of shifts of Hochpasseckfrequenzen according to the prior art;

FIG. 2

a block diagram for compensation of shifts of Hochpasseckfrequenzen according to the present invention;

FIG. 3

a circuit diagram of a compensation circuit according to a first embodiment of the present invention; and

FIG. 4

a circuit diagram of a compensation circuit according to a second embodiment which is helpful for understanding the present invention.

The aim is to match the two or more microphones in their electrical and acoustic behavior to each other. Each microphone can be described in the low-frequency range by a characteristic acoustic high pass whose corner frequency is about 50 Hz and a high-voltage electrical pass whose corner frequency is about 100 Hz. Both the acoustic and electrical high passes of the plurality of hearing aid microphones are slightly different and can be adapted to each other in the following manner.

According to the block diagram of FIG. 2 If a compensation according to the invention of the microphone differences is that first as in the prior art according to FIG. 1 the microphone input signal 2 is filtered with an acoustic high pass 1 of the first microphone 1 with the transfer function s / s-pol_ac1. The subsequent compensation filter 3 'has the transfer function s-pol_ac1 / s-pol_ac2. With this transfer function, the second microphone path, which is in FIG. 2 below is taken into account. In this second microphone path, as in the prior art, the signal 7 of a reference microphone 5 is subjected to high-pass filtering in accordance with the transfer function s / s-pol_ac2. The denominator polynomial of the second high-pass acoustic pass of the second microphone 5 is used to normalize the compensation filter 3 'in the first microphone path. With this standardization, the compensation filter 3 'does not have to be normalized to an ideal microphone in order to obtain the first microphone output signal 4. In the second microphone path can be dispensed with a compensation filter to obtain the second microphone output signal 8.

The compensation filter 3 'has a transfer function with a counter polynomial s-pol_ac1 and a denominator polynomial s-pol_ac2. With simplified compensation, only the numerator and not the denominator and the counter are adjusted. The denominator of the compensation filter 3 'is set at a nominal frequency. In the acoustic case the nominal frequency is 50 Hz and in the electrical case 100 Hz. However, with this fixed nominal frequency only approximate compensation is possible. This approximate compensation is, as mentioned, sufficiently good, for example, to improve the directivity of a directional microphone.

The transformation of such a compensation filter from the analog to the digital domain leads to a simple IIR filter of the first order, which can be represented as follows: $$\frac{p_1\left(X_p\right)\cdot z+p_0\left(X_p\right)}{z+q_0}$$

The functions p ₁ and p ₀ and the parameter q ₀ result from the above-mentioned European Patent Application EP 0982971 A2 , The variable z represents the frequency variable of the microphone input signal. The parameter Xp corresponds to a manipulated variable of the compensation filter. The denominator can not be varied in this simplified approach.

This results in an improved adaptation of the compensation filter in that the denominator in its transfer function can also be varied by a parameter Xq as follows: $$\frac{p_1\left(X_p\right)\cdot z+p_0\left(X_p\right)}{z+q_0\left(X_p\right)}$$

An implementation for adapting the high pass of a microphone according to the first embodiment, in which the denominator of the transfer function of the compensation filter is fixed, is in FIG FIG. 3 shown as a block diagram. The input unit forms the compensation filter 3 ', which is already in connection with FIG. 2 was explained. Input signal here is the signal 2 of a first microphone, in contrast to FIG. 2 on the playback of an acoustic high pass, which represents the microphone has been omitted. Output signal of the compensation filter 3 ', which performs the low-frequency microphone matching in the present case of the acoustic high-pass at 50 Hz, in turn, the signal 4. This is a multiplication unit 10, in which the signal with a corresponding compensation factor 11 broadband corrected in amplitude can be.

In a subsequent bandpass filter 12, a frequency range between 40 and 60 Hz is cut out of the output signal of the multiplication unit 10 and fed to a level meter 13. There, the level of the frequency range to be analyzed determined from the signal of the first microphone 2.

In parallel, the resulting from a second microphone input signal 8 output signal of a second microphone or reference microphone likewise not equally subjected to a bandpass filtering. A bandpass 14 also cuts out the frequency range between 40 and 60 Hz from the output signal of the microphone and also delivers the filtered signal to a level meter 15.

In a subtraction unit, the levels measured by the level meters 13 and 15 are subtracted from each other and the resulting level difference is made available to an updating unit for updating the Xp variable. However, an update of the Xp value should only take place if the microphone signals have a correspondingly high level. For this purpose, the microphone levels are applied to an input level sensing block 18 which generates an enable Xp signal when both signal levels exceed a certain threshold. This can be prevented that in cases where there are no acoustic input signals but only microphone noise, a microphone adaptation takes place. The enable-Xp signal is therefore looped to the Xp update block.

The value Xp, optionally updated in block 17, is now provided to the compensation filter 3 'to complete the control loop. The determination of the Xp value and thus the adaptation of the microphones to each other in the Xp update block 17 can be carried out by a (N) LMS algorithm (Normalized Leased Meansquare), wherein an "acoustical delay" block is necessary.

In FIG. 4 a circuit diagram of an improved version of a matching circuit is shown. The essential structure corresponds to that of FIG. 3 , wherein the corresponding function blocks perform substantially the same functions. Only the compensation filter, which is also designated by the reference numeral 3 ', has a further signal input, with which also the denominator polynomial can be changed via the variable X _q .

In order to be able to carry out both a change of the numerator and the denominator polynomial, the output signal of the input level query block 18, with which it is determined whether the two microphone signals have a sufficiently high level, is forwarded to a switch 19. This switch 19 alternately generates an enable-X _q signal and an enable-X _p signal in time if it receives an enable-X _p -X _q signal from block 18.

In addition to the X _p -Update block 17 also Xq update block 20 is therefore provided for changing or updating the X _q -value here. Now, if the switch 19 outputs an enable-X _q signal, the X _q value is changed in accordance with the level difference from the subtractor 16. Otherwise, if the switch 19 outputs an enable-signal X _p, X _p value is changed in the X _p -Update block 17 corresponding to the level difference. When the level difference is less than 0, the X _p or X _q value in one direction and when the level difference is greater than 0 are changed in the corresponding other direction.

The compensation filter 3 'receives the modified or updated X _p and X _q values as manipulated variables. As in the previous embodiment according to FIG. 3 The different high-pass frequencies of the microphones in a narrow frequency range around the corner frequencies mean different average output levels of the two microphone signals. This means that the level difference depends directly on the difference between the corner frequencies. To adapt the corner frequencies, therefore, simply the difference of the level is formed (power difference).

The total distance of a directional microphone from the microphone input to the output is often described at low frequencies with other first-order high-passes. In addition to the acoustic high pass, the microphone still has a first order electrical high pass with a corner frequency of about 180 Hz. Another high pass results from a coupling capacitor and input resistance of an IC input stage.

The adaptive methods described above can in principle be applied to all high passes.

Claims (6)

1. Method for reciprocal adaptation of a plurality of microphones (1, 5) of a hearing aid, by

   - measuring (13) a first amplitude of a first output signal from a first one of the plurality of microphones (1) in a predefined frequency range (12),

   - measuring (15) a second amplitude of a second output signal from a second one of the plurality of microphones (5) in the predefined frequency range (14) and

   - filtering (3') the first output signal as a function of the first amplitude and of the second amplitude, so that the difference (16) between the two amplitude signals is reduced,

   characterised in that

   - the filtering (3') is effected by multiplying by a denominator polynomial and a numerator polynomial,

   - only the numerator polynomial is varied using a control system and

   - the predefined frequency range (12, 14) consists of a first frequency band between 40 and 60 Hz and a second frequency band between 80 and 120 Hz.
2. Method according to claim 1, whereby parameters for filtering (3') are adapted in a closed loop such that the first and second amplitudes correspond to one another.
3. Method according to claim 1 or 2, whereby the filtering modifies the value and/or phase of the first output signal.
4. Apparatus for the reciprocal adaptation of a plurality of microphones (1, 5) of a hearing aid, having

   - a first measuring device (13) for measuring a first amplitude of a first output signal from a first one of the plurality of microphones (1) in a predefined frequency range (12),

   - a second measuring device (15) for measuring a second amplitude of a second output signal from a second one of the plurality of microphones (5) in the predefined frequency range (14) and

   - a filter device (3') which is connected to the first and second measuring device (13, 15), for filtering the first output signal as a function of the first amplitude and of the second amplitude, so that the difference (16) between the two amplitude signals can be reduced,

   characterised in that

   - the filter device (3') can be modelled by a denominator polynomial and a numerator polynomial,

   - a control system is connected to the filter device, with which only the numerator polynomial can be varied, and

   - the predefined frequency range (12, 14) consists of a first frequency band between 40 and 60 Hz and a second frequency band between 80 and 120 Hz.
5. Device according to claim 4, whereby the filter device (3') can be adapted in a closed loop such that the first and second amplitudes correspond to one another.
6. Device according to claim 4 or 5, whereby the value and/or phase of the first output signal can be modified using the filter device.

Images