Method of rejecting inherent noise of a microphone arrangement, and hearing device

Abstract

A method for rejecting inherent noise of a microphone arrangement that includes a first microphone and a second microphone. The first microphone generates a first microphone signal from an ambient sound signal and the second microphone generates a second microphone signal from the ambient sound signal. A measure of correlation between the first microphone signal and the second microphone signal is ascertained, and inherent noise of the first microphone and/or of the second microphone in the first or second microphone signal is rejected on the basis of the measure of correlation.

Claims

1. A method of rejecting inherent noise of a microphone arrangement having a first microphone and a second microphone, the method comprising: generating with the first microphone a first microphone signal from a sound signal from the surroundings; generating with the second microphone a second microphone signal from the sound signal from the surroundings; ascertaining a measure of correlation between the first microphone signal and the second microphone signal; rejecting inherent noise of at least one of the first microphone or the second microphone in the first or second microphone signal on a basis of the measure of correlation; and ascertaining at least one parameter selected from the group consisting of a wanted signal level, a power spectral density, a noise level, and a noise power variable for the first microphone signal and/or for the second microphone signal; and additionally controlling a rejection of the inherent noise of the microphone arrangement on a basis of at least one of the signal parameters, and rejecting the inherent noise of the microphone arrangement by way of Wiener filter; using the wanted signal level and/or the power spectral density or using the noise level and the noise power variable as input variables for the Wiener filter; and applying the Wiener filter to at least one of the first microphone signal or the second microphone signal in dependence on the measure of correlation.

2. The method according to claim 1, which comprises rejecting the inherent noise when the measure of correlation undershoots a predefined lower limit value.

3. The method according to claim 1, which comprises setting a degree of a rejection of the inherent noise in gradual dependence on the measure of correlation.

4. The method according to claim 1, which comprises: ascertaining the measure of correlation respectively for each of a plurality of frequency bands; and rejecting the inherent noise of one or both of the first microphone or the second microphone in a signal component of the first microphone in the respective frequency band or in a signal component of the second microphone signal in the respective frequency band based of the measure of correlation ascertained for the respective frequency band.

5. The method according to claim 1, wherein the measure of correlation is at least one measure selected from the group consisting of a covariance, a coherence, and a cross correlation.

6. The method according to claim 1, wherein the microphone arrangement is a component of a hearing device and the method comprises rejecting inherent noise of two microphones of the hearing device.

7. A hearing device, comprising: a microphone arrangement with a first microphone for generating a first microphone signal from a sound signal from surroundings of the hearing device and a second microphone for generating a second microphone signal from the sound signal from the surroundings of the hearing device; and a control unit connected to receive the first and second microphone signals from the microphone arrangement and configured to reject inherent noise of the microphone arrangement by carrying out the method according to claim 1.

Description

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING

(1) FIG. 1 shows a block diagram of a hearing device having two microphones and a Wiener filter for rejecting inherent noise of the microphones; and

(2) FIG. 2 shows a graph of a control function for the Wiener filter shown in FIG. 1 as a function of a measure of correlation.

(3) Mutually corresponding parts and variables are provided with identical reference signs throughout the figures.

DETAILED DESCRIPTION OF THE INVENTION

(4) Referring now to the figures of the drawing in detail and first, in particular, to FIG. 1 thereof, there is shown a block diagram of a hearing device 1 having a microphone arrangement 2. The microphone arrangement 2 comprises a first microphone 4 and a second microphone 6. The first microphone 4 is designed to generate a first microphone signal x1 from a sound signal 8 from surroundings, i.e., an ambient sound signal 8, of the hearing device 1. The second microphone 6 is designed to generate a second microphone signal x2 from the sound signal 8 from the surroundings of the hearing device 1. The first microphone signal x1 and the second microphone signal x2 are supplied to a control unit 10, which has processing and storage means, not depicted in more detail, in the form of one or more signal processors, RAM modules, etc., and in which the two microphone signals x1, x2 are processed by taking into consideration an impaired hearing of a user of the hearing device 1 that needs to be compensated for.

(5) The control unit 10 uses the aforementioned signal processing to generate from the first microphone signal x1 and the second microphone signal x2 an output signal 12 that is converted into an output sound signal 16 by an output transducer of the hearing device 1, in the present case provided by a loudspeaker 14. The output sound signal 16 is supplied to the ear of the wearer of the hearing device 1. The output transducer used in this instance may also be in particular a bone conduction receiver or any other electroacoustic transducer designed to generate a sound signal from the output signal 12.

(6) In the control unit 10, a first secondary signal path 18 is branched off from the first microphone signal x1 and a second secondary signal path 20 is branched off from the second microphone signal x2. The first and second secondary signal paths 18, 20 are supplied to inherent noise rejection 22, which can be implemented in the control unit 10 as an appropriate software module or else by appropriate, hardwired, circuits (for example as an ASIC), for example. A measure of correlation 24 is formed in the inherent noise rejection 22 from the first microphone signal x1, as is present in the first secondary signal path 18, and from the second microphone signal x2, as is present in the second secondary signal path 20.

(7) For example, the measure of correlation 24 can be a cross correlation function of the two microphone signals that is maximized for the temporal argument of the afore-mentioned function, and possibly normalized in a suitable manner. The measure of correlation 24 used can likewise be the cross-power spectrum of the two microphone signals x1, x2, which may need to be normalized in a suitable manner.

(8) The first microphone signal x1 and the second microphone signal x2 are moreover processed in the control unit 10 by directional microphonics 32 to form a preliminary output signal 11. A further secondary signal path 13 is branched off from the preliminary output signal 11, and said further secondary signal path is supplied to the inherent noise rejection 22, which furthermore has a Wiener filter 26. Such a Wiener filter is described in the above-mentioned patent application US 2018/0139546 A1, for example. A wanted signal level 28 and a noise power 30 are then ascertained in the inherent noise rejection 22 from the signal components of the preliminary output signal 11 in the secondary signal path 13 on the basis of frequency band. The splitting into individual frequency bands in this instance can be effected upstream of the directional microphonics 32 already by means of a filter bank (not depicted in more detail). On the basis of the wanted signal level 28 and the noise power 30, a filter function f, the arguments of which are the two aforementioned variables, in the Wiener filter 26 is used to ascertain a gain factor w that is intended to be used to reject inherent noise of the first microphone 4 and/or of the second microphone 6 in the preliminary output signal 11 by means of appropriate multiplication by the preliminary output signal 11.

(9) The application of the gain factor w for rejecting the aforementioned inherent noise is effected in this instance on the basis of a control function s that includes the measure of correlation 24 of the two microphone signals x1, x2 as argument. A gain factor w′ is therefore formed from the gain factor w of the Wiener filter 26 and the control function. The control function s is in this case preferably such that a high level of correlation between the first microphone signal x1 and the second microphone signal x2 results in the gain factor w being applied to the preliminary output signal 11 only a little or not at all, since in this case it is assumed that even substantial noise components in the preliminary output signal 11 and hence also in the aforementioned microphone signals x1, x2 come from noise in the sound signal 8. Accordingly, inherent noise of the microphone arrangement 2 (that is to say from at least one of the two microphones 4, 6), if present in the first place, is masked by the applicable signal components of the sound signal 8. In such a case, the control function s adopts a value of 0 or close to 0. If, however, it is established on the basis of the measure of correlation 24 that there is no significant correlation between the first microphone signal x1 and the second microphone signal x2, then it is assumed that the substantial and mutually uncorrelated signal components in the two microphone signals x1, x2 come from inherent noise of the microphone arrangement 2. Accordingly, a value of the control function s is set such that the gain factor w is applied to the preliminary output signal resulting from the two microphone signals x1, x2 (almost) to the full extent, and that the applicable contribution of the gain factor w is therefore included in the actually applied gain factor w′ (almost) completely. The value of the control function s is therefore 1 or almost 1. By applying the gain factor w to the preliminary output signal 11 in the respective frequency band, the output signal 12 is formed, which can furthermore be subjected to still further signal processing steps, not depicted in more detail, before the loudspeaker 14 effects the conversion into the output sound signal 16.

(10) The characteristic of the control function s as a function of the measure of correlation 24 is depicted schematically in FIG. 2. The normalized measure of correlation 24 assumes values between 0 and 1 as argument for the control function s, with 0 representing completely uncorrelated microphone signals and 1 representing perfectly correlated microphone signals x1, x2. The control function s, plotted on the ordinate, for its part assumes values between 0 and 1, a value of 1 according to the Wiener filter 28 shown in FIG. 1 resulting in the gain factor w produced there being applied to the two microphone signals x1, x2 completely, and a value of 0 for the control function s resulting in such application being omitted completely. The control function s assumes the value 1 for values of the measure of correlation 24 up to a lower limit value GU, or lower threshold GU. The lower limit value GU is therefore the value for the correlation, measured using the measure of correlation 24, below which the two microphone signals x1 and x2 are assumed to be sufficiently uncorrelated to reliably determine the inherent noise. For values of the measure of correlation 24 above an upper limit value GO, or upper threshold GO, the control function s assumes the value 0, as result of which the rejection of the inherent noise using the gain factor w ascertained in the Wiener filter 26 shown in FIG. 1 is therefore stopped completely. Between the lower limit value GU and the upper limit value GO there is continuous interpolation of the control function s, which is linear in the example shown in FIG. 2 but can also have a different characteristic, so long as said characteristic remains antitone (in particular the characteristic of the control function s can also gradually fall from 1 to 0). It will be noted here that the measure of correlation 24 is limited to values between 0 and 1 merely on account of applicable normalization; other definition ranges are conceivable.

(11) The control function s in this instance can additionally have a dependency—not depicted in more detail in the present case—on the signal level and/or on the noise level that is similar in form, based on the characteristic depicted in FIG. 2, to the dependency on the measure of correlation 24, that is to say in particular provides for complete application of the gain factor w for a low signal level and/or noise level and provides for complete stoppage of the rejection of inherent noise for high signal levels and/or noise levels (above a predefined upper limit).

(12) The applicable value of the control function s for the ascertained value of the measure of correlation 24 is now applied to the gain factor w ascertained by the Wiener filter, for example by means of a convex combination in the form
w′=w.Math.s+(1−s),
and the gain factor w′ thus ascertained is applied to the directional signal (the preliminary output signal 11) formed on the basis of the first microphone signal x1 and the second microphone signal x2. The hearing-device specific signal processing 32 for compensating for the impaired hearing of the wearer of the hearing device 1 is preferably effected after the inherent noise rejection 22 so as not, by means of subsequent amplifications, to additionally amplify possible inherent noise of the microphone arrangement 2 as well and thus to minimize the entry of possible inherent noise of the microphone arrangement 2 into the output signal 12 as far as possible.

(13) Although the invention has been illustrated and described more thoroughly in detail by the preferred exemplary embodiment, the invention is not limited by this exemplary embodiment. Other variations can be derived therefrom by a person skilled in the art without departing from the scope of protection of the invention.

(14) The following is a summary list of reference numerals and the corresponding structure used in the above description of the invention: 1 hearing device 2 microphone arrangement 4 first microphone 6 second microphone 8 sound signal 10 control unit 11 preliminary output signal 12 output signal 13 (further) secondary signal path 14 loudspeaker 16 output sound signal 18 first secondary signal path 20 second secondary signal path 22 inherent noise rejection 24 measure of correlation 26 Wiener filter 28 wanted signal level 30 noise power 32 directional microphonics f filter function GU lower limit value GO upper limit value s control function x1 first microphone signal x2 second microphone signal w gain factor w′ gain factor