Method for directional signal processing for a hearing aid

The method enhances hearing aid performance in complex environments by using directional signal processing with adaptive microphones and time-delayed superimpositions to isolate and amplify specific sound sources, addressing the issue of reduced intelligibility caused by multiple sound sources.

EP3726853B1Active Publication Date: 2024-11-27SIVANTOS PTE LTD
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
EP2020155945
Authority / Receiving Office
EP · EP
Patent Type
Patents
Current Assignee / Owner
Priority Date
2019-04-18
Filing Date
2020-02-06
Publication Date
2024-11-27
Estimated Expiration
2040-02-06

AI Technical Summary

Technical Problem

Existing hearing aid technologies with automatic gain control (AGC) and dynamic compression often inadvertently attenuate signals from multiple sound sources, leading to reduced intelligibility in complex listening situations, such as conversations with multiple speakers, as they react to sound events regardless of direction.

Method used

A method for directional signal processing in hearing aids that generates calibration directional signals with maximum attenuation in specific directions, allowing for relative gain parameter determination to optimize signal amplification and compression, thereby isolating and amplifying one useful signal source while minimizing the impact of another, using adaptive directional microphones and intermediate signals with time-delayed superimpositions.

Benefits of technology

This approach effectively reduces level peaks from unwanted sound sources, maintaining comfortable sound levels and improving intelligibility by isolating and amplifying the desired signal sources, even in complex environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure IMGF0001
    Figure IMGF0001
  • Figure IMGF0002
    Figure IMGF0002
  • Figure IMGF0003
    Figure IMGF0003
Patent Text Reader

Abstract

The invention describes a method for directional signal processing for a hearing aid (2), wherein a first input signal (E1) is generated from an ambient sound signal (28) by a first input transducer (24) of the hearing aid (2), wherein a second input transducer (26) of the hearing aid (2) generates a second input signal (E2) from the ambient sound signal (28), wherein a first calibration directional signal (12) is generated from the first input signal (E1) and from the second input signal (E2), which exhibits a relative attenuation in the direction (6) of a first useful signal source (14) of the environment, and wherein a second calibration directional signal (16) is generated from the first input signal (20) and from the second input signal (E2), which exhibits a relative attenuation in the direction (10) of a second useful signal source (18) of the environment.wherein a relative gain parameter (GR) is determined based on the first calibration directional signal (12) and the second calibration directional signal (16), wherein a first processing directional signal (Y1) and a second processing directional signal (Y2) are generated based on both the first input signal (E1) and the second input signal (E2), wherein a source-sensitive directional signal (YQ) is generated based on the first processing directional signal (Y1), the second processing directional signal (Y2) and the relative gain parameter (GR), and wherein an output signal (52) of the hearing aid (2) is generated based on the source-sensitive directional signal (YQ).
Need to check novelty before this filing date? Find Prior Art

Description

[0001] The invention relates to a method for directional signal processing for a hearing aid, wherein a first input signal is generated from an ambient sound signal by a first input transducer of the hearing aid, wherein a second input signal is generated from the ambient sound signal by a second input transducer of the hearing aid, wherein a first directional signal is generated from the first input signal and from the second input signal, which has a relative attenuation in the direction of a first useful signal source in the environment, wherein a second directional signal is generated from the first input signal and from the second input signal, which has a relative attenuation in the direction of a second useful signal source in the environment, and wherein a source-sensitive directional signal is generated from the first directional signal, the second directional signal and the relative gain parameter.

[0002] In a hearing aid, ambient sound is converted into an input signal by at least one input transducer. This signal is then processed and amplified in a frequency-band-specific manner, tailored to the individual hearing loss of the wearer, depending on the specific hearing impairment being corrected. The processed signal is then converted into an output signal via an output transducer, which is transmitted to the wearer's ear. During signal processing, automatic gain control (AGC) and dynamic range compression are often applied to the input signal or to a pre-processed intermediate signal. With AGC, the input signal is typically amplified linearly only up to a certain threshold, and above this threshold, a lower amplification is applied to smooth out peak levels in the input signal.This is intended in particular to prevent sudden, loud sound events from resulting in an output sound signal that is too loud for the wearer due to the additional amplification in the hearing aid.

[0003] However, such a hearing aid with integrated dynamic compression initially reacts to sound events regardless of their direction. If the wearer of a hearing aid is in a complex listening situation, e.g., in a conversation with several people, one person might trigger the compression with a short cry or loud laugh, noticeably reducing the volume of another participant's contributions, which can impair the wearer's intelligibility.

[0004] EP 3 337 188 A1 describes a method for operating a hearing aid comprising at least one first input transducer, one second input transducer, and at least one output transducer, wherein the first input transducer generates a first input signal from an ambient sound signal, and the second input transducer generates a second input signal from the sound signal, wherein a first direction is assigned to a first useful signal source and a second direction is assigned to a second useful signal source, which is spatially separated from the first useful signal source, wherein a first directional signal oriented in the first direction and a second directional signal oriented in the second direction are formed based on the first input signal and the second input signal, and wherein an output signal is formed based on the first directional signal and the second directional signal, which is converted into a sound signal by the output transducer of the hearing aid.

[0005] EP 3 104 627 A1 describes a method for improving a recording signal in a hearing system with at least one hearing aid, wherein the hearing aid is assigned a first directional microphone which has an adjustable first directional characteristic with a preferred direction, wherein the first directional microphone converts sound into a first signal which is included in the recording signal, wherein, depending on a speech activity of a user (1) of the hearing system (2), the preferred direction of the first directional characteristic is adjusted relative to a frontal direction of the user in such a way that the sound sensitivity of the first directional microphone in the frontal direction is attenuated.

[0006] In US Patent 6,865,275 B1, two output signals of a microphone system depend differently on the angle of incidence of the acoustic signals and are divided. A mathematical product of the ratio and a weighting factor is saturated and subtracted from a signal value that can be fed into the system. The remainder of the subtraction is multiplied by the output signal of the microphone system, which also generates the denominator signal for the division. Depending on the weighting factor of the saturation value and the subtraction value, a desired directional characteristic is implemented between the resulting signal of the multiplication and the angle of incidence of acoustic signals acting on the microphone system.

[0007] The invention is based on the objective of providing a method for signal processing in a hearing aid which, in particular in conjunction with AGC and dynamic compression, is also suitable for complex listening situations.

[0008] The aforementioned problem is solved according to the invention by a method for directional signal processing for a hearing aid according to claim 1 or claim 5, and a corresponding hearing system according to claim 10. In this method, a first input signal is generated from an ambient sound signal by a first input transducer of the hearing aid, wherein a second input signal is generated from the ambient sound signal by a second input transducer of the hearing aid, wherein a first calibration directional signal is generated based on the first input signal and the second input signal, which exhibits maximum attenuation in the direction of a first useful signal source in the environment, and wherein a second calibration directional signal is generated based on the first input signal and the second input signal, which exhibits maximum attenuation in the direction of a second useful signal source in the environment.wherein a relative gain parameter is determined based on the first calibration directional signal and the second calibration directional signal, wherein a first instantaneous gain parameter is determined based on the first calibration directional signal, wherein a second instantaneous gain parameter is determined based on the second calibration directional signal, and wherein the relative gain parameter is determined based on the first instantaneous gain parameter and the second instantaneous gain parameter.

[0009] Here, a first processing directional signal and a second processing directional signal are generated based on both the first and second input signals, wherein a source-sensitive directional signal is generated based on the first processing directional signal, the second processing directional signal, and the relative gain parameter, and wherein an output signal of the hearing aid is generated based on the source-sensitive directional signal, which is preferably converted into an output sound signal by an output converter of the hearing aid, wherein a derived signal strength in the direction of the first useful signal source is determined based on the relative gain parameter and the reference signal strength, and wherein a complex superposition parameter for superimposing the first processing directional signal with the second processing directional signal is determined based on the derived signal strength.and the source-sensitive directional signal is generated based on the associated superposition. Advantageous and, in some cases, inventive embodiments are the subject of the dependent claims and the following description.

[0010] An input converter, in this context, refers in particular to an electroacoustic converter designed to generate a corresponding electrical signal from an audio signal. Specifically, the generation of the first or second input signal by the respective input converter may also include preprocessing, e.g., in the form of linear pre-amplification and / or analog-to-digital conversion.

[0011] The generation of the first and second calibration signals based on the first and second input signals preferably includes the direct integration of the signal components of the first and second input signals into the respective calibration signals. This ensures that the first and second input signals are not both used simultaneously for generating control parameters or similar functions that are applied to signal components of other signals. Preferably, at least the signal components of the first input signal, and particularly preferably also the signal components of the second input signal, are linearly integrated into the first calibration signal. A similar approach preferably applies to the second calibration signal.

[0012] The first and second calibration directional signals can be generated, in particular, from intermediate signals, each of which is created from the first and second input signals. For example, a directional first intermediate signal and, furthermore, a directional second intermediate signal can be generated from the first and second input signals, with the directional characteristics of the first and second intermediate signals preferably exhibiting symmetry with each other, e.g., as cardioid and anticardioid. The first calibration directional signal can then be generated from the first and second intermediate signals using adaptive directional microphones such that, as required, the adaptive directional microphones produce a relative attenuation in the direction of the first signal source. A similar process applies to the second calibration directional signal.

[0013] The relative gain parameter is preferably generated from the first and second calibration directional signals in such a way that, with a corresponding superposition of the signal components of the first and second calibration directional signals, weighted by the relative gain parameter, an output signal resulting from such a superposition can be controlled as far as possible via a common AGC and in particular via a common dynamic compression, and for this purpose the signal components arriving from at least two directions are taken into account and weighted as optimally as possible.

[0014] The relative gain parameter can then be used, in particular, for superimposing the first processing directional signal with the second processing directional signal. Preferably, the first and second processing directional signals are generated using the same intermediate signals as the first and second calibration directional signals. However, in particular, an additional, preferably shared, degree of freedom, e.g., in the form of an additional adjustment parameter or similar, can be added to both the first and second intermediate signals during the generation of the processing directional signals, which is not present in the calibration directional signals.Such an additional degree of freedom allows the sensitivity of the source-sensitive directional signal to be adjusted towards both the first and second signal sources by varying the aforementioned adjustment parameter, particularly depending on the relative gain parameter. This enables subsequent processing of the source-sensitive directional signal, or the output signal generated from it, via AGC and corresponding dynamic compression in a manner that takes both signal sources into account. Alternatively, instead of the aforementioned adjustment parameter, the required degree of freedom can also be introduced into a superposition of the two processing directional signals, for example, in the form of a complex-valued superposition parameter during the generation of the source-sensitive directional signal.

[0015] In order to apply AGC and, if necessary, corresponding dynamic compression in a hearing aid in complex listening situations, especially conversations with multiple participants, without attenuating individual signals (e.g., contributions to a conversation) due to the level peaks of another sound signal, the first and second calibration directional signals are preferably determined in such a way that the sound events of one useful signal source (such as the speech contributions of one conversation partner) are emphasized as much as possible, and the sound events of another useful signal source (e.g., speech contributions of another conversation partner) are suppressed as much as possible, whereby the roles of the sound events to be suppressed and emphasized are reversed for the two calibration directional signals.

[0016] Based on these two calibration directional signals, the relative gain parameter is preferably determined in such a way that a corresponding superposition of the two calibration directional signals – i.e., a superposition of the first calibration directional signal with the second calibration directional signal, which is weighted with the relative gain parameter – can be controlled by a common AGC value as a result of the described suppressions and emphasis of the respective sound events, whereby excessive influence of one sound event by the other sound event can be avoided in the superimposed signal.

[0017] The source-sensitive directional signal, from which the output signal may be generated through additional signal processing, is not created by superimposing the two calibration directional signals, but rather by using the two processing directional signals. This is because superimposing the two calibration directional signals for further processing could cause the direction of maximum attenuation of the superimposed signal to fluctuate over a wider angular range of up to 90° depending on the sound events of the desired signal sources, as can occur, for example, with adaptive directional microphones. In particular, this direction no longer coincides with the direction of either of the desired signal sources.

[0018] As described above, an additional degree of freedom can now be introduced via the two processing directional signals, depending on the relative gain parameter determined from the calibration directional signals. This stabilizes the direction of maximum attenuation in the source-sensitive directional signal. A sudden change in the sound of a useful signal source leads, via a corresponding change in the relative gain parameter and a resulting different weighting of the two processing directional signals, to the signal contributions of the other useful signal source in the source-sensitive directional signal being unaffected or only minimally affected by this change in AGC with dynamic compression.

[0019] Preferably, a third input signal can be generated by a third input transducer of the hearing aid, so that a total of three calibration directional signals are generated from the available input signals. Each of these signals exhibits a relative attenuation in the direction of a different one of the three useful signal sources. Two relative gain parameters can then be determined from the three calibration directional signals, allowing for the superposition of three processing directional signals, which are themselves generated from the three input signals, based on these two relative gain parameters. The method can be extended to even higher-order systems in the input transducers.

[0020] According to the invention, a first instantaneous gain parameter is determined from the first calibration directional signal, and a second instantaneous gain parameter is determined from the second calibration directional signal. The relative gain parameter is then determined from the first instantaneous gain parameter and the second instantaneous gain parameter, particularly as their quotient. Preferably, the first instantaneous gain parameter and the second instantaneous gain parameter are determined as "isolated" values ​​of an AGC or dynamic compression for the respective calibration directional signal. Thus, each of the two useful signal sources is "calibrated" independently via a corresponding AGC using the respective calibration directional signal, which attenuates the other useful signal. The relative gain parameter is then determined from these isolated AGC values.

[0021] Advantageously, a first intermediate signal and a second intermediate signal are generated from both the first and second input signals. This makes it possible to generate the calibration directional signals and / or the processing directional signals using adaptive directional microphones, with the intermediate signals being used in the adaptive directional microphone technique.

[0022] According to the invention, the first calibration signal exhibits maximum attenuation in the direction of the first useful signal source, and the second calibration signal exhibits maximum attenuation in the direction of the second useful signal source. This allows the influence of each useful signal on the other calibration signal, and thus on the relative gain parameter, to be minimized particularly effectively.

[0023] Advantageously, the first calibration directional signal is generated using adaptive directional microphones, particularly based on the first and second intermediate signals, and / or the second calibration directional signal is generated using adaptive directional microphones, particularly based on the first and second intermediate signals. This ensures that the calibration directional signal exhibits the lowest possible, preferably minimal, sensitivity in the direction of one of the two signal sources, resulting in high, preferably maximum, attenuation in that direction, and the highest possible, preferably maximum, sensitivity in the direction of the other signal source.

[0024] It is further advantageous if the first intermediate signal is generated by a time-delayed superposition of the first input signal with the second input signal, implemented by means of a first delay parameter, and / or the second intermediate signal is generated by a time-delayed superposition of the second input signal with the first input signal, implemented by means of a second delay parameter. In particular, the first and second delay parameters can be chosen to be identical to each other, and the first intermediate signal can be generated symmetrically to the second intermediate signal with respect to a preferred plane of the hearing aid, the preferred plane being the frontal plane of the wearer when wearing the hearing aid. Aligning the directional signals with the wearer's frontal direction facilitates signal processing, as this takes into account the wearer's natural line of sight.

[0025] Preferably, the first intermediate signal is generated as a forward-facing cardioid directional signal and / or the second intermediate signal as a backward-facing cardioid directional signal. A cardioid directional signal can be generated by superimposing the two input signals relative to each other with an acoustic time delay corresponding to the distance between the input transducers. Depending on the sign of this time delay during superposition, the direction of maximum attenuation is either in the frontal direction (backward-facing cardioid directional signal) or in the opposite direction (forward-facing cardioid directional signal). The direction of maximum sensitivity is opposite to the direction of maximum attenuation. This simplifies further signal processing, as such an intermediate signal is particularly suitable for adaptive directional microphones.

[0026] Preferably, the first calibration directional signal and the second calibration directional signal are both generated using both the first intermediate signal and the second intermediate signal.

[0027] In a further advantageous embodiment, the first processing signal is generated from the first intermediate signal, and in particular "identically" to it, and / or the second processing signal is generated from the second intermediate signal, and in particular "identically" to it. Here, "generating the first or second processing signal from the first or second intermediate signal" (and analogously in the following) means, in particular, that no signal components from other signals are included in the generated signal other than the signal components of said "generating" signal. Signal components from other signals are used, at most, as control signals for parameters during the generation of the respective processing signal. The first processing signal is generated identically to the first intermediate signal by reusing the first intermediate signal as the first processing signal for the subsequent process steps.In this context, it is advantageous for reducing complexity if both the calibration direction signals and the processing direction signals are based on the same intermediate signals.

[0028] It is further advantageous if the first processing signal is generated as a first asymmetric superposition signal by means of a time-delayed superposition of the first input signal with the second input signal, implemented using asymmetric first weighting factors, and / or if the second processing signal is generated as a second asymmetric superposition signal by means of a time-delayed superposition of the second input signal with the first input signal, implemented using asymmetric second weighting factors. This means, in particular, that for the generation of the first processing signal, the first input signal E1 is superimposed with the second input signal E2 according to E1 - w1 E2, whereby the weighting of the two input signals is not identical, but is determined by the first asymmetric weighting factors w1 (w1 ≠ 1).Furthermore, a time delay T1 of the second input signal can be applied, so that the first processing directional signal Y1 (t) can be represented as in the period . Y1 t = E1 t − w1 ⋅ E2 t − T1 .

[0029] The second processing direction signal Y2(t) can be represented (depending on the choice of the global phase) accordingly as Y2 t = E2 t − w2 ⋅ E1 t − T2 bzw . − E2 t + w2 ⋅ E1 t − T2 .

[0030] Using such first and second asymmetric weighting factors w1, w2 (w1, w2 each ≠ 1), an additional degree of freedom can be inserted to fix the direction of maximum attenuation in the source-sensitive directional signal as described above.

[0031] According to the invention, a reference signal strength is defined in the direction of the second useful signal source, in particular based on the first instantaneous gain factor. A derived signal strength in the direction of the first useful signal source is then determined based on the relative gain parameter and the reference signal strength. Based on the derived signal strength, a complex superposition parameter is determined for superimposing the first processing directional signal with the second processing directional signal. The source-sensitive directional signal is then generated based on the associated superposition. In this way, the additional degree of freedom can be introduced by means of the complex superposition parameter to fix the direction of maximum attenuation in the source-sensitive directional signal as described above.

[0032] The aforementioned problem is also solved according to the invention by a method for directional signal processing for a hearing aid, wherein a first input signal is generated from an ambient sound signal by a first input transducer of the hearing aid, wherein a second input transducer of the hearing aid generates a second input signal from the ambient sound signal, wherein a first calibration directional signal is generated on the basis of the first input signal and on the basis of the second input signal, which has a maximum attenuation in the direction of a first useful signal source in the environment, wherein a second calibration directional signal is generated on the basis of the first input signal and on the basis of the second input signal, which has a maximum attenuation in the direction of a second useful signal source in the environment.and wherein a relative gain parameter is determined based on the first calibration directional signal and the second calibration directional signal.

[0033] Here, the first processing directional signal is generated as a first asymmetric superposition signal by means of a time-delayed superposition of the first input signal with the second input signal, implemented using asymmetric first weighting factors, and / or the second processing directional signal is generated as a second asymmetric superposition signal by means of a time-delayed superposition of the second input signal with the first input signal, implemented using asymmetric second weighting factors, whereby a reference signal strength is defined in the direction of the second useful signal source, and a derived signal strength in the direction of the first useful signal source is defined based on the relative gain parameter and the reference signal strength.and wherein the asymmetric first and / or second weighting factors for the first and / or the second asymmetric superposition signal are determined based on the derived signal strength, and wherein the source-sensitive directional signal is generated as the first or second processing directional signal based on the first and / or the second asymmetric superposition signal.

[0034] This means, in particular, that to generate the first processing directional signal, the first input signal E1 is superimposed on the second input signal E2 according to E1 - w1 E2, whereby the weighting of the two input signals is not identical, but is determined by the first asymmetric weighting factors w1 (w1 ≠ 1). Furthermore, a time delay T1, for example of the second input signal, can be introduced, so that the first processing directional signal Y1(t) can be represented within that time period as Y1 t = E1 t − w1 ⋅ E2 t − T1 .

[0035] The second processing direction signal Y2(t) can be represented (depending on the choice of the global phase) accordingly as Y2 t = E2 t − w2 ⋅ E1 t − T2 bzw . − E2 t + w2 ⋅ E1 t − T2 .

[0036] Using such first and second asymmetric weighting factors w1, w2 (w1, w2 each ≠ 1), an additional degree of freedom can be inserted to fix the direction of maximum attenuation in the source-sensitive directional signal as described above.

[0037] In this process, a reference signal strength is established in the direction of the second useful signal source, specifically based on the first instantaneous gain factor. A derived signal strength in the direction of the first useful signal source is then determined based on the relative gain parameter and the reference signal strength. Based on this derived signal strength, the asymmetric first and / or second weighting factors for the first and / or second asymmetric superposition signal are determined. Finally, the source-sensitive directional signal is generated as the first and / or second processing directional signal, respectively, based on the first and / or second asymmetric superposition signal. This provides another way to introduce an additional degree of freedom to fix the direction of maximum attenuation in the source-sensitive directional signal, as described above.If the source-sensitive directional signal is generated based on the first and / or the second asymmetric superposition signal, this can also be achieved by a corresponding superposition using a real-valued superposition parameter. In particular, the reference signal strength and the derived signal strength can be used as equivalent for determining the real superposition parameter and the respective asymmetric weighting factors.

[0038] The invention further describes a hearing system comprising a hearing aid, which has a first input transducer for generating a first input signal from an ambient sound signal and a second input transducer for generating a second input signal from the ambient sound signal, and a control unit configured to perform a method according to one of the preceding claims. In particular, the control unit can be integrated into the hearing aid. In this case, the hearing system is directly provided by the hearing aid. The advantages mentioned for the method and its further developments can be applied analogously to the hearing system.

[0039] An embodiment of the invention is explained in more detail below with reference to a drawing. The drawing schematically depicts: Fig. 1 shows a conversation situation of a hearing aid wearer with two conversation partners, Fig. 2 shows a preferred, directional signal processing for the hearing aid in the conversation situation according to Fig. 1 , Fig. 3 the course of a notch depth of a directional signal at the maximum attenuation angle in the method according to Fig. 2 Depending on an arc parameter for the corresponding superposition parameter, Fig. 4 shows an alternative embodiment of directional signal processing according to Fig. 2 .

[0040] Corresponding parts and sizes are marked with the same reference symbols in all figures.

[0041] In Figure 1The schematic top view shows a wearer 1 of a hearing aid 2 in a conversation with a first conversation partner 4 and a second conversation partner 8. The first conversation partner 4 is positioned relative to wearer 1 in a first direction 6, and the second conversation partner 8 in a second direction 10 relative to wearer 1. The second conversation partner 8 is the main conversation partner of wearer 1; the first conversation partner 4 participates in the conversation only through occasional contributions. The described conversation situation is shown for both the upper and lower images. Figure 1 identical.

[0042] In order to attenuate the peak levels of the speech contributions of the first conversation partner 4 and the second conversation partner 8 for the wearer 1 of the hearing aid 2 in an output sound signal of the hearing aid 2, the following is done, as shown in the image above. Figure 1As shown, a first calibration directional signal 12 is generated using adaptive directional microphones such that it exhibits maximum and preferably complete attenuation in the first direction 6, in which the first interlocutor 4 is positioned. This means that the speech contributions of the first interlocutor 4 are not captured by the first calibration directional signal 12. Consequently, a compression factor, which is calculated based on the first calibration directional signal 12, reacts only to the second interlocutor 8 with respect to the two signal sources 14, 18, which are provided by the first and second interlocutors 4 and 8, respectively. Here, a first instantaneous gain parameter G1 is determined, which, with respect to the signal contributions of the second signal source 18 (i.e., the second interlocutor 8), determines the optimal signal gain and thus implicitly also a corresponding compression ratio at any given moment.

[0043] In the lower image of Figure 1 Analogous to the image above, a second calibration directional signal 16 is shown, which exhibits maximum and preferably complete attenuation in the second direction 10, i.e., the direction of the second interlocutor 8. Since the second direction 10 coincides with the frontal direction of the carrier 1, the second calibration directional signal 16 is designed as a backward-directed cardioid directional signal 20. The second instantaneous gain parameter G2, determined from and assigned to the second calibration directional signal 16, thus represents at any given moment the optimal gain with respect to the first interlocutor 4 and, in particular, a corresponding compression ratio.

[0044] In order to reduce the level peaks in the output sound signal of the hearing aid 2 for its wearer 1 by means of compression, due to the contributions of both the first conversation partner 4 and the second conversation partner 8, such an output sound signal could now be formed from a linear combination of the first and the second calibration directional signal 12, 16, which are each weighted with their corresponding instantaneous gain parameters G1, G2.Since the first calibration directional signal 12 is also generated using adaptive directional microphones based on a forward-facing cardioid directional signal and the backward-facing cardioid directional signal 20, such a linear combination would result in an output sound signal whose directional characteristic is similar in shape to that of the first calibration directional signal 12, but with the notch 22 of maximum attenuation shifted away from the first direction 6. This leads, on the one hand, to a potentially undesirable, completely "deaf" area away from the first useful signal source 14, which, on the other hand, can also fluctuate in its orientation due to the dependence of such a linear combination on the speech contributions of the second interlocutor 8.

[0045] In Figure 2 is schematically shown in a block diagram a method for directional signal processing for the hearing aid 2 according to Figure 1The situation described therein is illustrated, which is intended to attenuate the peak levels of the two signal sources 14, 18 provided by the respective conversation partners 4, 8. The hearing aid 2 contains a first input transducer 24 and a second input transducer 26, which generate a first input signal E1 and a second input signal E2, respectively, from a sound signal 28. The sound signal 28 is the ambient sound, which therefore also includes the contributions of the first conversation partner 4 and the contributions of the second conversation partner 8. Any preprocessing, such as A / D conversion or similar, is already incorporated into the input transducers 24, 26, each of which also includes a microphone.

[0046] The first input signal E1 is superimposed with the second input signal E2, which has been delayed by a first delay parameter T1, and a first intermediate signal 34 is formed. Similarly, the second input signal E2 is superimposed with the first input signal E1, which has been delayed by a second delay parameter T2, and a second intermediate signal 36 is formed. Here, and without loss of generality, the first and second delay parameters T1 and T2 are identical (T1 = T2) and are chosen such that the first intermediate signal 34 is a forward-directed cardioid directional signal 38, and the second intermediate signal 36 is a backward-directed cardioid directional signal 20. Based on the first intermediate signal 34 and the second intermediate signal 36, the first calibration directional signal 12 is then determined. Figure 1The first calibration signal 12 is generated using adaptive directional microphones 40 such that the contributions of the first interlocutor 4 are maximally suppressed. The first instantaneous gain parameter G1 determined for the first calibration signal 12 thus represents the optimal gain and compression of the signal contributions of the second interlocutor 8. Using adaptive directional microphones 42, the second calibration signal 16 is generated from the first intermediate signal 34 and the second intermediate signal 36, which maximally suppresses the contributions of the second interlocutor 8. Since the second interlocutor 8 is positioned frontally to the carrier 1, the second calibration signal 16 is, as already mentioned, provided by the rearward-directed cardioid signal 20. However, this situation can also change, so that a change in the position of the second interlocutor 8 can also be taken into account using the adaptive directional microphones 42.

[0047] The second calibration directional signal 16 is used to determine the second instantaneous gain parameter G2, and a relative gain parameter GR is formed from this and the first instantaneous gain parameter G1, which in this case is given by the quotient G2 / G1. The relative gain parameter GR would result from the aforementioned linear combination of the first calibration directional signal 12 with the second calibration directional signal 16, each weighted by their respective instantaneous gain parameters G1 and G2, if the first instantaneous gain parameter G1 were used as the global gain parameter for the resulting signal of the linear combination, whereby the relative gain parameter GR in the first direction 6 produces precisely the correct signal strength for the speech contributions of the first interlocutor 4.

[0048] To generate the output signal 52, the first intermediate signal 34 is used as a first processing directional signal Y1, and the second intermediate signal 36 as a second processing directional signal Y2, in the following signal processing steps. The first processing directional signal Y1 and the second processing directional signal Y2 are then superimposed with a complex superposition parameter a, which is determined based on the relative gain parameter GR, in the form YQ = Y1 + a·Y2. As a result of this superposition, a source-sensitive directional signal YQ is generated, which may be subjected to further, unspecified signal processing steps 50, such as additional frequency-band-specific amplification, etc. From the source-sensitive directional signal YQ, an output signal 52 is generated, which is converted into an output sound signal 56 by an output converter 54 of the hearing aid 2.The output converter 54 can, for example, include a loudspeaker. The output sound signal 56 is then fed to the ear of the wearer 1.

[0049] For a sound signal that strikes the two input transducers 24, 26 at an angle of α against the frontal direction, the first and second processing directional signals Y1, Y2 can be represented as Y1 ω = A ⋅ 1 − exp − i ω T 1 + cos α ⋅ X ω , Y2 ω = A ⋅ − exp iωT − exp − i ω T ⋅ cos α ⋅ X ω where T = T1 = T2 is the first or second delay parameter, X(w) is the frequency response of the sound signal 28, w is the respective frequency, A is a normalization factor, and i is the imaginary unit. T is such that Y1 and Y2 result as the desired forward and backward cardioid directional signals 38 and 20, respectively. It can now be determined using the formulas YQ = Y1 + a ⋅ Y2 , YQ ω ∧ 2 = H ω ∧ 2 ⋅ X ω ∧ 2 Determine the squared magnitude of the transfer function H(w) for the source-sensitive directional signal YQ as a function of the complex-valued superposition parameter a: H ω 2 = 2 A 2 ⋅ 1 + a 2 + + 2 Re a cos ωT − cos ωT cos α − − a 2 cos ωT 1 − cos α − cos ωT 1 − cos α

[0050] It is preferred that for an angle of incidence α = 0° the magnitude of the transfer function |H(ω)| should be independent of the frequency w, thus achieving a flat frequency spectrum in the frontal direction. This preferred choice, however, does not represent any restriction. This determines the normalization factor A in the two equations (i). It can now be shown that all parameters a for which |H(ω)|^2 leads to a minimum at the same angle α each lie on a circle in the complex plane, and thus a circle position is relevant for the depth of the minimum at α, i.e., for the suppression of YQ at this angle: a φ = sin ωT sin ωT 1 − cos α cos φ + j sin φ + sin ωT cos α sin ωT 1 − cos α

[0051] The position of the circle can be parameterized by an arc parameter ϕ ∈[0, 2π) such that the relative depth D of the minimum of |H(ω)| at a given angle α varies for different ϕ on this circle in the complex plane.

[0052] The relative depth D of the notch in the directional characteristic of the source-sensitive directional signal YQ, created by the aforementioned attenuation, as a function of this arc parameter ϕ, is in Fig. 3 The dependencies D(ϕ) shown there can be tabulated, allowing a relationship between a desired notch and the arc parameter sz for a. Thus, the relative gain parameter GR, which determines the relative attenuation in the directional characteristic of the source-sensitive directional signal YQ (based on the aforementioned tabulated dependencies on the arc parameter ϕ of a), and the first instantaneous gain parameter G1 can be determined from the two degrees of freedom of the magnitude |a| and the arc parameter ϕ of a. Figure 3 It is also evident that for ϕ=0, which, as can be shown, corresponds to a real-valued a, the deepest possible notch is created.

[0053] In Figure 4A schematic representation of an alternative implementation of the method for directional signal processing is shown in a block diagram. Figure 2 The determination of the relative gain parameter GR is completely identical to that shown in [reference]. Figure 2 The method shown is different, except that the generation of the output signal 52 is modified here. The first and second processing direction signals Y1 and Y2, respectively, are now each generated from the first and second input signals E1 and E2, which are superimposed with a delay parameter T (= T1 = T2). However, an additional degree of freedom is introduced via a real-valued matching parameter m during the superposition. According to equations (i) Figure 2 The corresponding equations for the first and second processing directional signals Y1, Y2 as a function of the frequency ω, the frequency response X(w) of the sound signal 28 and the adaptation parameter m are now: Y1 ω = A ⋅ 1 − m ⋅ exp − i ω T 1 + cos α ⋅ X ω , Y2 ω = A ⋅ exp − iωT − m ⋅ exp − i ω T ⋅ cos α ⋅ X ω

[0054] Based on the first and second processing directional signals Y1 and Y2, the source-sensitive directional signal YQ is now generated according to YQ = Y1 + a·Y2, where the superposition parameter a is chosen to be a real value. As in equation (ii) above, the transfer function H(w) can now be determined as a function of the superposition parameter a and the matching parameter m: H ω 2 = A 2 ⋅ 1 + m 2 + a 2 + a 2 m 2 − 4 ma ⋅ cos ωT cos α + 2 a 1 + m 2 cos ωT − − 2 ma 2 ⋅ cos ωT 1 − cos α − 2 m ⋅ cos ωT 1 + cos α

[0055] The normalization factor A is then preferably chosen such that, in the frontal direction, i.e., for α = 0°, the magnitude of the transfer function |H(w)| is independent of the frequency w. This defines the normalization factor A in the two equations (iii) as A = 1 1 + m 2 + a 1 − m 2 a + 2 cos ωT − 2 m cos 2 ωT

[0056] It can now be shown that for all real-valued superposition parameters a, the minimum of |H(ω)|^2 in α, i.e., the angle of maximum attenuation, is independent of the matching parameter m. Thus, the values ​​of the superposition parameter a and the matching parameter m can be chosen as a function of the first instantaneous gain parameter G1 and the relative gain parameter GR such that, on the one hand, the instantaneous overall loudness and any resulting compression ratio have the correct value, and on the other hand, in the first direction 6 – by equating it with α in the corresponding equations – the correct relative attenuation of the speech contributions of the first interlocutor 4 in Figure 1This is done to control the speech signals of both interlocutors 4, 8 via the instantaneous first gain parameter G1. It is very useful that the minimum of the squared magnitude or the magnitude of the transfer function |H(ω)| is independent of the matching parameter m, so that the relative attenuation can be controlled via the latter as a function of the relative gain parameter GR, for example via appropriately tabulated values.

[0057] Based on the source-sensitive directional signal YQ generated as described from the first and second processing directional signals Y1, Y2, the output signal 52 is generated again. As a further, not detailed, alternative embodiment, it is also conceivable to apply the adjustment parameter m, unlike in equation (iii), only to one of the two processing directional signals Y1, Y2 (for example, to Y2), and to use the other processing directional signal (for example, Y1) as the corresponding cardioid directional signal (e.g., the forward-directed 38).

[0058] Although the invention has been illustrated and described in detail by the preferred embodiment, the invention is not limited by the disclosed examples and other variations can be derived by the person skilled in the art without departing from the scope of protection of the invention as defined in the following claims. Reference symbol list

[0059] 1 Carrier 2 Hearing aid 4 First conversation partner 6 First direction 8 Second conversation partner 10 Second direction 12 First calibration directional signal 14 First useful signal source 16 Second calibration directional signal 18 Second useful signal source 20 Backward cardioid directional signal 22 Notch 24 First input transducer 26 Second input transducer 28 Sound signal 34 First intermediate signal 36 Second intermediate signal 38 Forward cardioid directional signal 40 Adaptive directional microphone 42 Adaptive directional microphone 50 Signal processing steps 52 Output signal 54 Output transducer 56 Output sound signal a) Superposition parameter Relative depth E1 First input signal E2 Second input signal G1 First instantaneous gain parameter G2 Second instantaneous gain parameter GR Relative gain parameter T1 First delay parameter T2 Second delay parameter Y1 First processing directional signal Y2 Second processing directional signal YQ Source-sensitive directional signal ϕArc parameter

Claims

1. Method for directional signal processing for a hearing aid (2), wherein a first input signal (E1) is generated by a first input transducer (24) of the hearing aid (2) from a sound signal (28) of the surroundings, wherein a second input signal (E2) is generated by a second input transducer (26) of the hearing aid (2) from the sound signal (28) of the surroundings, wherein a first intermediate signal (34) and a second intermediate signal (36) are each generated on the basis of both the first input signal (E1) and the second input signal (E2), wherein a first calibration directional signal (12), which has a maximum attenuation in the direction (6) of a first useful signal source (14) of the surroundings, is generated on the basis of the first intermediate signal (34) and on the basis of the second intermediate signal (36), wherein a second calibration directional signal (16), which has a maximum attenuation in the direction (10) of a second useful signal source (18) of the surroundings, is generated on the basis of the first intermediate signal (34) and on the basis of the second intermediate signal (36), wherein a relative amplification parameter (GR) is determined on the basis of the first calibration directional signal (12) and the second calibration directional signal (16), in that - a first instantaneous amplification parameter (G1) is ascertained on the basis of the first calibration directional signal (12), - a second instantaneous amplification parameter (G2) is ascertained on the basis of the second calibration directional signal (16), and - the relative amplification parameter (GR) is ascertained on the basis of the first instantaneous amplification parameter (G1) and the second instantaneous amplification parameter (G2), wherein a first processing directional signal (Y1) and a second processing directional signal (Y2) are each generated on the basis of both the first input signal (E1) and the second input signal (E2), wherein in this case the first processing directional signal (Y1) is generated from the first intermediate signal (34) and the second processing directional signal (Y2) is generated from the second intermediate signal (36), wherein a source-sensitive directional signal (YQ) is generated on the basis of the first processing directional signal (Y1), the second processing directional signal (Y2), and the relative amplification parameter (GR), in that - a reference signal strength is established in the direction (10) of the second useful signal source (18), - a derived signal strength is established in the direction (6) of the first useful signal source (14) on the basis of the relative amplification parameter (GR) and on the basis of the reference signal strength, and - a complex superposition parameter (a) for a superposition of the first processing directional signal (Y1) with the second processing directional signal (Y2) is ascertained on the basis of the derived signal strength, and the source-sensitive directional signal (YQ) is generated on the basis of said superposition, and wherein an output signal (52) of the hearing aid (2) is generated on the basis of the source-sensitive directional signal (YQ).

2. Method according to any one of the preceding claims, wherein the first calibration directional signal (12) is generated by means of adaptive directional microphonics (40), and / or wherein the second calibration directional signal (16) is generated by means of adaptive directional microphonics (42).

3. Method according to Claim 1 or 2, wherein the first intermediate signal (34) is generated on the basis of a time-delayed superposition of the first input signal (E1) with the second input signal (E2) implemented by means of a first delay parameter (T1), and / or wherein the second intermediate signal (36) is generated on the basis of a time-delayed superposition of the second input signal (E2) with the first input signal (E1) implemented by means of a second delay parameter (T2).

4. Method according to Claim 3, wherein the first intermediate signal (34) is generated as a forward-directed cardioid directional signal (38), and / or wherein the second intermediate signal (36) is generated as a backward-directed cardioid directional signal (20).

5. Method for directional signal processing for a hearing aid (2), wherein a first input signal (E1) is generated by a first input transducer (24) of the hearing aid (2) from a sound signal (28) of the surroundings, wherein a second input signal (E2) is generated by a second input transducer (26) of the hearing aid (2) from the sound signal (28) of the surroundings, wherein a first calibration directional signal (12), which has a maximum attenuation in the direction (6) of a first useful signal source (14) of the surroundings, is generated on the basis of the first input signal (E1) and on the basis of the second input signal (E2), wherein a second calibration directional signal (16), which has a maximum attenuation in the direction (10) of a second useful signal source (18) of the surroundings, is generated on the basis of the first input signal (20) and on the basis of the second input signal (E2), wherein a relative amplification parameter (GR) is determined on the basis of the first calibration directional signal (12) and the second calibration directional signal (16), in that - a first instantaneous amplification parameter (G1) is ascertained on the basis of the first calibration directional signal (12), - a second instantaneous amplification parameter (G2) is ascertained on the basis of the second calibration directional signal (16), and - the relative amplification parameter (GR) is ascertained on the basis of the first instantaneous amplification parameter (G1) and the second instantaneous amplification parameter (G2), wherein a first processing directional signal (Y1) and a second processing directional signal (Y2) are each generated on the basis of both the first input signal (E1) and the second input signal (E2), wherein in this case the first processing directional signal (Y1) is formed as a first asymmetrical superposition signal on the basis of a time-delayed superposition, implemented by means of asymmetrical first weighting factors (m), of the first input signal (E1) with the second input signal (E2), and / or wherein the second processing directional signal (Y2) is formed as a second asymmetrical superposition signal on the basis of a time-delayed superposition, implemented by means of asymmetrical second weighting factors (m), of the second input signal (E2) with the first input signal (E1), wherein a source-sensitive directional signal (YQ) is generated on the basis of the first processing directional signal (Y1), the second processing directional signal (Y2), and the relative amplification parameter (GR), in that - a reference signal strength is established in the direction (10) of the second useful signal source (18), - a derived signal strength is established in the direction (6) of the first useful signal source (14) on the basis of the relative amplification parameter (GR) and on the basis of the reference signal strength, and - on the basis of the derived signal strength - a complex superposition parameter (a) for a superposition of the first processing directional signal (Y1) with the second processing directional signal (Y2) is ascertained, and / or - the asymmetrical first (m) and / or second weighting factors (m) are ascertained for the first and / or the second asymmetrical superposition signal, - the source-sensitive directional signal (YQ) is generated on the basis of said superposition of the first processing directional signal (Y1) with the second processing directional signal (Y2), and wherein an output signal (52) of the hearing aid (2) is generated on the basis of the source-sensitive directional signal (YQ).

6. Method according to Claim 5, wherein a first intermediate signal (34) and a second intermediate signal (36) are each generated on the basis of both the first input signal (E1) and the second input signal (E2), and wherein the first calibration directional signal (12) and the second calibration directional signal (16) are both generated on the basis of both the first intermediate signal (34) and the second intermediate signal (36).

7. Method according to Claim 6, wherein the first intermediate signal (34) is generated on the basis of a time-delayed superposition, implemented by means of a first delay parameter (T1), of the first input signal (E1) with the second input signal (E2), and / or wherein the second intermediate signal (36) is generated on the basis of a time-delayed superposition, implemented by means of a second delay parameter (T2), of the second input signal (E2) with the first input signal (E1).

8. Method according to Claim 6 or Claim 7, wherein the first intermediate signal (34) is generated as a forward-directed cardioid directional signal (38), and / or wherein the second intermediate signal (36) is generated as a backward-directed cardioid directional signal (20).

9. Method according to any one of Claims 5 to 8, wherein the first calibration directional signal (12) is generated by means of adaptive directional microphonics (40), and / or wherein the second calibration directional signal (16) is generated by means of adaptive directional microphonics (42).

10. Hearing system comprising - a hearing aid (2), which has a first input transducer (24) for generating a first input signal (E1) from a sound signal (28) of the surroundings and a second input transducer (26) for generating a second input signal (E2) from the sound signal (28) of the surroundings, and - a control unit, which is configured to carry out the method according to any one of the preceding claims.

Citation Information

Patent Citations

  • Method for improving a recording signal in a hearing system

    EP3104627A1

  • Method for operating a hearing aid

    EP3337188A1

  • Method to determine the transfer characteristic of a microphone system, and microphone system

    US6865275B1