Hearing device and method for operating a hearing device

Abstract

The invention relates to a method for operating a hearing device (2) which is assigned to a user, in which the hearing device (2) acquires an audio signal (U) and thereby generates an input signal (E), in which an AGC input signal (EAGC) is generated from the input signal (E), which is fed to an AGC unit (12) of the hearing device (2), the AGC unit (12) amplifies the AGC input signal (EAGC) in accordance with a level of the AGC input signal (EAGC) and in accordance with an AGC function (14) and then outputs the AGC input signal as an AGC output signal (AAGC), the AGC function (14) being defined by one or more parameters (P) which are set in accordance with the currently present acoustic scene (S), and the hearing device (2) generating an output signal (A) from the AGC output signal (AAGC) for output to the user. The invention further relates to a corresponding hearing device.

Description

[0001] The invention relates to a hearing device and to a method for operating such a hearing device.

[0002] A hearing device generally comprises an input transducer, a signal processing device and an output transducer. The input transducer is generally a microphone. The output transducer is generally an earpiece, also called loudspeaker or receiver. The hearing device is generally configured to a single user and is used only by this user. The hearing device is used for example to support a hearing-impaired user and to compensate for the hearing loss of this user. The input transducer generates an input signal which is fed to the signal processing device. The signal processing device modifies the input signal and generates therefrom an output signal, so that the output signal is a modified input signal. In order to compensate for the hearing loss, the input signal is amplified for example with frequency-dependent gain factors according to the audiogram (hearing range) of the user. The output signal is finally output to the user via the output transducer. For a hearing device with a microphone and an earpiece, the microphone accordingly generates an input signal from a sound signal in the environment, and the earpiece in turn generates a sound signal from the output signal. The input signal and the output signal are electrical signals, they are therefore also simply called signals. In contrast, the sound signal in the environment and possibly the sound signal output by the earpiece are acoustic signals.

[0003] It is furthermore advantageous to integrate an automatic gain control (AGC) into the hearing device. For this purpose, the hearing device expediently has a corresponding AGC unit by means of which the dynamics of the input signal can be modified in order to obtain a certain dynamics of the output signal. Here, the level of the input signal is amplified to the level of the output signal according to an AGC function. This AGC function for example contains a knee below which the output signal corresponds to the input signal (gain = 1) and above which compression takes place (gain < 1). The AGC function is also called dynamic curve.

[0004] One disadvantage of the AGC is that in some cases the compression can lead to a reduction of the signal-to-noise ratio (SNR) of the input signal. This reduces the speech intelligibility in particular for the case of input signals containing speech. This is particularly problematic in a noisy environment which already has a low signal-to-noise ratio in itself.

[0005] Against this background, the technical problem addressed by the invention is to improve the operation of a hearing device with an AGC unit. For this purpose, a correspondingly improved method and a correspondingly improved hearing device are to be presented.

[0006] This task is solved according to the application by a method having the features of claim 1 and by a hearing device having the features of claim 20. Advantageous design solutions, improvements and variants are the subject matter of the dependent claims. The explanations made in connection with the method apply analogously to the hearing device and vice versa. If steps of the method are implicitly or explicitly described in the following, advantageous design solutions of the hearing device result therefrom, namely that the hearing device has a signal processing device which is designed to perform one or more of these steps.

[0007] A core idea of the application is, inter alia., to determine the AGC function optimally for different acoustic scenarios, i.e. the AGC unit of the hearing device implements the AGC function according to which it performs the AGC, and here also taking into account the individual hearing range of the hearing device user. Thus, in operation of the hearing device, the AGC function is advantageously automatically adapted when the scenario changes, and the hearing range of the user is utilized optimally as far as possible.

[0008] The method according to the application is used for operating a hearing device. The hearing device is correspondingly configured to a user. The hearing device is in particular configured to only one user and is used only by this user. Preferably, the hearing device is used to provide support for a hearing-impaired user and to compensate for the hearing loss of this user, on the basis of which the following is also assumed without limiting generality. In operation, the user wears and uses the hearing device, for example on the head and in or on the ear.

[0009] Hearing devices generally have an input transducer, an output transducer and a signal processing apparatus (also referred to as signal processing unit). In a suitable design, the input transducer is a microphone and the output transducer is a receiver. The following will be explained on the basis of this design, but without restricting generality. In operation, the input transducer generates an input signal which is fed to the signal processing apparatus. Also in operation, the signal processing apparatus modifies the input signal and finally generates an output signal therefrom. For modifying the input signal, the signal processing apparatus has one or more units which are connected in series and / or in parallel as required. In order to compensate for a hearing loss, the input signal or a signal derived therefrom is amplified by frequency-dependent gain factors, for example in accordance with a user's audiogram (hearing range). To this end, the signal processing apparatus suitably has a corresponding gain unit. The output signal is finally output to the user by means of the output transducer in operation. For a hearing device with a microphone and a receiver, the microphone accordingly generates an input signal in accordance with a sound signal in the environment, and the receiver in turn generates a sound signal in accordance with the output signal. The input signal and the output signal are electrical signals and are therefore also simply referred to as signals, respectively. In contrast, the sound signal in the environment and, if necessary, the sound signal output by the receiver are acoustic signals. As an alternative or in addition to the sound signal, the hearing device receives a data signal, for example an audio stream. Any signal (sound signal, data signal) received by the hearing device for modification and output to the user is referred to as an audio signal in general.

[0010] In the method described here, the hearing device acquires an audio signal, in particular by means of an input transducer, and generates an input signal from the audio signal, also by means of the input transducer. From the input signal, an AGC input signal is generated which is fed to an AGC unit of the hearing device. In the simplest design, the AGC input signal is identical to the input signal, i.e. it is simply directly forwarded to the AGC unit. The following will be explained on the basis of this design, but without restricting generality. Alternatively, only a part of the input signal is used as the AGC input signal, or the input signal is modified beforehand by a further unit of the signal processing apparatus and is used as the AGC input signal. The AGC unit is in particular part of the signal processing apparatus and is arranged, for example, before, after or alongside a gain unit.

[0011] The AGC unit amplifies the AGC input signal according to the level of the AGC input signal and according to an AGC function, and then outputs the correspondingly amplified AGC input signal as AGC output signal. To this end, the AGC function correspondingly configures each level value of the AGC input signal to a level value of the AGC output signal. At one point of the AGC function, the ratio of the two level values results in a gain (i.e. a gain factor) which is applied by the AGC unit to the respective AGC input signal level value to amplify the signal. The gain can in principle be smaller than, larger than or equal to 1, i.e. attenuation (<1), mere passing (=1) or true amplification (>1) in the strict sense can take place. In an equivalent design, the AGC function relates the levels of the input signal and the output signal, instead of the levels of the AGC input signal and the AGC output signal. In short: the AGC input signal (and usually also the input signal) has an input dynamic, and the AGC output signal (and usually also the output signal) has an output dynamic, and the AGC function maps the input dynamic to the output dynamic. Input dynamic and output dynamic are usually also referred to as "dynamic" or "dynamic range", respectively.

[0012] The AGC function is defined by one or more parameters. Here, without limiting generality, it is assumed that there are multiple parameters. These parameters define the form or the course of the AGC function, and are also referred to as AGC parameters. These parameters are advantageously adjustable to change the AGC function and, in turn, the modification of the AGC input signal by the AGC unit. Here, these parameters are set in dependence on the currently existing acoustic scenario, i.e. in dependence on the current acoustic scenario. More precisely: the respective appropriate values of the parameters are selected in dependence on the current acoustic scenario. Thus, the AGC function is optimally parameterized in dependence on the current acoustic scenario, respectively. An acoustic scenario (short: "scenario") is generally understood as the acoustic environment in which the user is located. The current acoustic scenario is then the acoustic environment in which the user and thus the hearing device is actually located at the current time. The acoustic environment and thus the scenario is defined, inter alia, by any sound signals, including useful noise and disturbing noise, which can reach the user and the hearing device with different intensities and / or from different directions, and also by any acoustic properties of the environment, such as reverberation time, number of speakers, signal-to-noise ratio (SNR), etc. Examples of scenarios are "conversation", "music", "conversation with background noise", "disturbing noise". Other examples of scenarios are, inter alia, situations with a certain noise level, a certain SNR or a certain dynamic range, as further explained below.

[0013] The currently existing acoustic scene is suitably determined by a scene recognition unit of the hearing device. The scene recognition unit is in particular part of the signal processing arrangement. Suitably, the input signal or a signal derived therefrom is fed to the scene recognition unit, which then determines the scene on the basis thereof. To this end, the scene recognition unit comprises for example a classifier. The AGC unit is in particular connected to and controlled by the scene recognition unit, i.e. the parameters are set in particular automatically depending on the determination of the current acoustic scene. The scene recognition unit outputs in particular a control signal which depends on the respective scene and is received by the AGC unit to set the parameters depending on the scene. Preferably, the speed of setting the parameters corresponds to the speed of the determination of the scene. Optionally, the setting of the parameters is smoothed to optimize, for example, the response to a scene switch. For example, the control signal is smoothed to this end.

[0014] Finally, the hearing device generates an output signal from the AGC output signal (as described above) to be output to the user. The above explanations with respect to the input signal apply correspondingly, i.e. in the simplest design the output signal is identical to the AGC output signal, i.e. it is simply directly forwarded, for example, to the earpiece. Alternatively, only a part of the AGC output signal is used as the output signal, or the AGC output signal is also modified beforehand by a further unit of the signal processing arrangement, for example a gain unit, before being used as the output signal.

[0015] The following explanations are made without limiting generality, assuming that the AGC input signal corresponds to the input signal and the AGC output signal corresponds to the output signal. Irrespective thereof, all explanations with respect to the input signal and the output signal also apply in substance to the AGC input signal or the AGC output signal, and vice versa.

[0016] An important advantage of the present application is in particular that a more precise control of the output dynamics is achieved for different scenes. It is in principle conceivable that the AGC unit uses a fixed preset AGC function in combination with an adjustable time constant, or that a specific predefined scene of the input signal is adapted to a threshold value above which compression takes place. The present application described herein goes a step further by, however, treating different scenes by parameterizing the AGC function differently if necessary depending on the scene, in that the parameters are suitably set depending on the scene, thereby producing a correspondingly optimized AGC function for each scene. Thereby, the user's hearing comfort and speech intelligibility are improved.

[0017] Another advantage consists especially in that, within the same scene, the AGC by the AGC unit is more stable, where "stable" is to be understood as that, within a given scene dynamics, the AGC function is linear only within this dynamics. In principle, the AGC function can be preset, which is stable within a certain value range with a fixed amount. If the scene dynamics exceeds this value range, then simply shift the range to capture the dynamics. If the dynamics is larger than the value range, then this leads disadvantageously to the value range being shifted back and forth repeatedly. The adaptation of the time constant of the AGC unit can also lead to uncontrollable and / or unstable behavior. In contrast, the application described here changes the AGC function and thus the behavior of the AGC unit only when the scene changes, i.e. when a change in the current acoustic scene is recognized by the scene recognition unit. Within the same scene, however, the same AGC function is continuously used, which is optimized for this scene, however, by the parameter settings depending on the scene.

[0018] Another advantage consists especially in that the signal-to-noise ratio of the scene is at least maintained and even increased. Thereby, especially the speech intelligibility is improved. In principle, the compression by the AGC unit, i.e. the slope of the dynamic curve < 1, leads to a reduction of the SNR. This is especially problematic in noisy scenes with a low SNR and can reduce the speech intelligibility. According to the application, the AGC function is adapted to the scene here by setting the parameters depending on the current acoustic scene. This can even lead to an expansion (dynamic curve slope > 1), in which case the SNR of the current acoustic scene is increased accordingly, while the dynamics of the output signal is still optimally adapted to the user, for example to the user's hearing range.

[0019] The terms "dynamic" and "dynamic range" are used synonymously here and denote a measure of the value range covered by the variable level of a signal, for example the input signal, the output signal, the AGC input signal, the AGC output signal. The dynamic is defined especially for a predefined time interval, for example 1 second. The term "lower end" then refers to the minimum value of the dynamic, the term "upper end" similarly to the maximum value. The input dynamic is characterized accordingly by the minimum and maximum values of the AGC input signal or generally by the minimum and maximum values of the input signal, the output dynamic similarly by the minimum and maximum values of the AGC output signal or generally by the minimum and maximum values of the output signal. The input dynamic is preset by the current scene, while the output dynamic can in principle be adapted by the AGC function, which is also used here. Furthermore, it can be advantageous to preset a particularly suitable minimum value or maximum value or minimum and maximum values for the output dynamic, i.e. by a corresponding parameterization of the AGC function. This is used in the preferred design to reduce or avoid comb filtering effects, as explained in further detail below.

[0020] In a particularly simple design, the lower end and the upper end of the dynamic range correspond to the minimum and the maximum of the considered signal level, in particular within a predefined time range. Accordingly, the lower end and the upper end of the dynamic of the current acoustic scene are suitably determined as the minimum and the maximum of the input signal within a predefined time range. In addition to using the minimum and the maximum of the signal level as the upper end and the lower end of the dynamic range, other definitions, in particular based on statistical analysis, are suitable. For example, the upper end and the lower end correspond to a measure of dispersion, e.g. the standard deviation, of the level values, in particular within a predefined time period, with respect to a measure of central tendency, e.g. the median, of the level values. Accordingly, in a suitable design, the lower end and the upper end of the dynamic range of the current acoustic scene are determined based on a statistical analysis of the input signal.

[0021] The dynamic range of the current acoustic scene is advantageously determined, which comprises a lower end and an upper end. In other words: the current acoustic scene has a dynamic range, which is directly determined, e.g. by deriving from the input signal, as the minimum and the maximum of the input signal, or as the minimum and the maximum of an AGC input signal, which itself is derived from the input signal. The dynamic range of the current acoustic scene is also referred to as scene dynamic. The input dynamic, already mentioned, or a dynamic derived therefrom, is particularly suitable as the dynamic range of the current acoustic scene. The determination of the scene dynamic is suitably part of the method described herein. In a suitable design, the scene dynamic is determined by the scene recognition unit or the AGC unit.

[0022] The scene-dependent parameterization of the AGC function is particularly suitable for optimally mapping the scene dynamic to a specific, preset output dynamic, i.e. for keeping the level of the output signal within a respective numerical range. For this purpose, in a suitable design, an output dynamic is preset, having a lower end and an upper end. The parameters of the AGC function are set such that the AGC function maps the dynamic, here the scene dynamic, to the output dynamic. In particular, the lower end of the scene dynamic is mapped to the lower end of the output dynamic, and the upper end of the scene dynamic is mapped to the upper end of the output dynamic. In between, i.e. between the lower end and the upper end, in particular a linear mapping takes place. Thereby, an optimal overlap of the scene dynamic and the output dynamic is achieved.

[0023] In a particularly advantageous design, the output dynamic corresponds to the user's hearing range. The AGC function is adapted to the scene in such a way that the scene dynamic is optimally mapped to the hearing range. The parameters of the AGC function are set in such a way that the output dynamic ultimately given to the user is as consistent as possible with the hearing range, so that the hearing range optimally overlaps the scene dynamic and the user's hearing is optimally fully utilized. It is already advantageous in principle that the scene dynamic is at least mostly, preferably at least 90%, mapped onto the hearing range. The hearing range is defined by those level values that the user can perceive. Signals with levels within the hearing range are audible to the user, while signals with levels outside the hearing range are inaudible. Accordingly, the hearing range is individualized for each user and, if necessary, is different. Signals with a dynamic at least partially outside the hearing range are disturbing for the user and can be difficult to understand. Whether a level can be heard is necessarily subjective. The hearing range is determined, for example, outside the method by suitable hearing tests. The hearing range is suitably stored in or to the hearing device.

[0024] The respective AGC function for mapping the scene dynamic to the hearing range can now in principle be used in this way. In an advantageous design, however, the AGC function is additionally adapted as an additional step after or during the adaptation of the hearing range by shifting the lower end or the upper end of the output dynamic, i.e. by setting a difference between the lower end or the upper end of the hearing range and the lower end or the upper end of the output dynamic, as described in detail below. The output dynamic, although slightly different from the hearing range, can reduce the comb filter effect in this way.

[0025] The lower end and the upper end mentioned here are each a level value (also called a loudness value), for example given in dB. Accordingly, the hearing range and any dynamic range are each a range of levels (also called a range of loudnesses).

[0026] For the hearing range (also called "hearing dynamic"), the design of the dynamic applies analogously, with the difference that the hearing range is not based on the actual signal, but rather gives the user's characteristics. Despite this, the hearing range has a lower end (minimum) and an upper end (maximum) and gives a numerical range of levels. The upper end and the lower end of the hearing range can be determined and defined as described above for the dynamic range.

[0027] In one suitable design, the different scenarios are distinguished "nominally", i.e. by a classifier, for example to be understood as "noisy environment", "speech in quiet environment", etc. Alternatively or additionally, the different scenarios are described by one or more characteristic values, for example a noise level, an SNR or a dynamic range, wherein these characteristic values are preferably used to control the dynamics of the AGC unit. The definition of the different scenarios by a classifier can be suitably implemented in different ways: for example by means of thresholds for the above-mentioned characteristic values (SNR, etc.), or as a statistical classifier which statistically analyzes the characteristic values, the input signal or a transformed form of the input signal and, on the basis thereof, identifies the respective scenario. As a classifier, it is likewise suitable to use a trained classifier, for example a neural network, etc. The speech activity detection unit mentioned below is also suitable as a classifier for classifying the scenarios. Here, the term "scenario" preferably particularly refers to the respective associated dynamic range, which can differ frequency-specifically and is advantageously primarily referenced to the dynamic range of the target speaker and secondarily to the dynamic range of the acoustic scenario (environment), i.e. the dynamic range of the ambient noise or the current noise level.

[0028] The respective AGC function, which is fixedly preset for the different scenarios, can at most only limitedly map the scenario dynamics to the preset output dynamics, in particular to the user's hearing range. In contrast, according to the application, the AGC function is adapted as required. Thereby, the acceptance of using the hearing device is improved.

[0029] Suitably, the hearing range is the user's comfortable hearing range. The comfortable hearing range is defined in particular by a conventional hearing test in which, for example, a hearing acoustician measures the user's hearing threshold frequency-dependently, which represents the lower level limit of perception, and additionally also measures the uncomfortable threshold (UCL) as the upper level limit. The comfortable hearing range lies between these two level limits. This comfortable hearing range can in principle also be further defined by a test for the range of the so-called most comfortable loudness (MCL). Alternatively or additionally, the dynamic range of the current acoustic scenario is the relevant dynamic range of the scenario, which in particular primarily follows the dynamic range of the target speaker and, suitably additionally secondarily (for example in scenarios without speech), the dynamic range of the environment, i.e. the dynamic range of the ambient noise or the current noise level.

[0030] In a preferred design, one of the parameters is a first knee, which is set to lie at the lower end of the dynamic range (scene dynamic) and at the lower end of the output dynamic. The other parameter is a second knee, which is set to lie at the upper end of the dynamic range and at the upper end of the output dynamic. Thus, the dynamic range and the output dynamic are delimited by these two knees. By means of these two knees, the AGC function is divided into three sections: a first section between the two knees, a second section after the second knee and a third section before the first knee. Preferably, all three sections are straight, i.e. linear. The first section is of the highest user relevance, since it preferably determines the gain of the AGC input signal in the audible range. The second and third sections define the gain of the AGC input signal outside the audible range. By setting the two knees to mark the lower and upper ends, the scene dynamic is mapped precisely to the output dynamic, in particular to the audible range.

[0031] In an exemplary design, the lower end of the output dynamic is 40 dB and the upper end of the output dynamic is 60 dB. In a first scene, the lower end of the dynamic range (scene dynamic) is 20 dB and the upper end of the dynamic range is 80 dB. Then the slope of the first section is < 1 and in operation the dynamic range is compressed in this scene, since it is greater than the output dynamic. In a further second scene, the lower end of the dynamic range is 60 dB and the upper end of the dynamic range is 70 dB. Then the slope of the first section is > 1 and in operation of this scene the dynamic range is expanded, since it is now greater than the output dynamic. The dynamic range varies from scene to scene, while the output dynamic usually remains constant.

[0032] In another preferred design, one of the parameters is an offset (gain offset), which is set to a value corresponding to the difference between the lower end of the output dynamic, in particular the audible range, and the lower end of the dynamic range. In other words: Offset g shift The lower end of the output dynamic o low and the lower end of the dynamic range d low is derived according to:

[0033] g shift = o low - d low .

[0034] Preferably, the lower end of the output dynamic is determined such that the difference to the lower end of the dynamic range does not exceed a maximum value. Thus, the offset is limited to a maximum value. In this way, the differences in the gain aspect of different scenes are at least qualitatively preserved, which results in a more natural scene reproduction for the user. Thus, the lower end of the output dynamic is effectively shifted depending on the lower end of the scene dynamic and the maximum value, if necessary, and is then correspondingly offset from the originally preset lower end of the output dynamic.

[0035] In another preferred design, one parameter is a first slope, which is set as the ratio of the difference between the upper and lower limits of the output dynamics (especially the auditory range) to the difference between the upper and lower limits of the dynamic range (scene dynamics). The first slope s hearing The upper end of the output is dynamic. high and lower end o low and the upper limit d of the dynamic range high and lower end d low The following formula is used to derive:

[0036] s hearing =(o high -o low ) / (d high -d low )

[0037] Therefore, the first slope is specifically the slope of the AGC function in the first segment between the two inflection points mentioned above. The first slope also typically represents the gain of the AGC unit for the level values ​​within the scene's dynamic range. Depending on the size of the dynamic range in the given current acoustic scene and the size of the output dynamics (especially the individualized auditory range), a slope <1 (compression), a slope = 1 (simple transmission), or a slope > 1 (expansion) is produced. Optionally, the first slope is limited to a maximum value to avoid possible artifacts. This maximum value is especially > 1.

[0038] Suitably, it is determined whether speech is present in the input signal, and if speech is not present in the input signal, a first slope is limited to 1. For determining whether speech is present in the input signal, a speech activity detection unit is preferably used. This speech activity detection unit is, in particular, part of a signal processing device, and receives the input signal or a signal derived therefrom to determine whether speech is present in the input signal. The limitation to 1 in the absence of speech restricts the expansion of the AGC unit to the presence of speech.

[0039] In another preferred design, one parameter is a second slope, which is set to be equal to the ratio of the difference between the maximum audible level and the upper limit of the output dynamics (specifically, the audible range) to the difference between the maximum scene level and the upper limit of the dynamics. The second slope specifically represents the gain of the AGC unit along the second segment (i.e., above the second inflection point). The second slope s loud From the maximum auditory level h max Output dynamic upper end o high Maximum scene level d max and the upper end of the dynamic range d high The following formula is used to derive:

[0040] s loud =(h max -o high) / (d max -d high )

[0041] The maximum audible level is greater than the upper end of the output dynamic and represents, inter alia, the maximum level allowed for the output signal overall. The maximum audible level is, inter alia, the so-called uncomfortable threshold (UCL), which is measured or approximated. The maximum audible level is, for example, 70 dB. The maximum scene level is, inter alia, greater than the upper end of the dynamic range, which is determined, for example, on the basis of a statistical analysis, while the maximum scene level corresponds to the measured absolute maximum of the level of the current acoustic scene. The maximum scene level is, for example, 90 dB. The maximum scene level is defined, for example, as the highest input level (for a frequency-specific) within a certain time period. If the AGC unit is adapted to this maximum scene level, it means, inter alia, that all levels up to and including this level are transmitted by the AGC unit into the audible range. Levels above the high level, which are statistically less probable, fall appropriately into the working range of additional algorithms, for example, an "output limiter" or a "maximum power output" limiter.

[0042] The second slope is preferably limited to a value of at most 1, i.e. no expansion is allowed on the second section.

[0043] The scene-dependent adaptation of the AGC function according to the application also achieves a reduction in the comb-filter effect. The comb-filter effect is an artefact that arises from the superposition of the output signal of the hearing device with the direct sound signal (short: direct sound). The direct sound bypasses the hearing device, for example through the vent or the ear dome of the hearing device, and reaches the ear canal of the user. Due to the processing within the hearing device, a time offset arises between the output signal and the direct sound, the output signal being usually delayed relative to the direct sound. The comb-filter effect is most pronounced when the background noise of the scene is stationary, i.e. changes little, and the direct sound and the output signal have similar levels. The level of the direct sound signal is also referred to as the direct sound level. The comb-filter effect is therefore strongly dependent on the type and level of the input signal. Since the gain of the hearing device is user-dependent and can be non-linear, the comb-filter effect differs between hearing devices, since the comb-filter effect is configured user-dependently (so-called fitting / adaptation). Accordingly, it is difficult to give a static solution to the comb-filter problem.

[0044] In principle, it is possible to switch off those frequency bands (or frequency ranges) which are particularly sensitive to the comb-filter effect when fitting the hearing device to the user. But these frequency bands would then also be excluded from the beneficial properties, such as noise suppression and others. Furthermore, this approach does not take into account that the comb-filter effect depends on the situation, i.e. on the non-linear relationship between the level of the direct sound and the level of the output signal of the hearing device. Another possible solution would be to predict the comb-filter effect by means of a scene recognition unit and then to switch off the possibly affected frequency bands depending on the situation. However, this prediction is difficult.

[0045] Therefore, it is advantageous to use an adjustable AGC function to avoid the comb-filter effect. In a suitable design, the level of the direct sound of the current acoustic scene is determined accordingly and one or more parameters of the AGC function are set such that the lower or upper end of the output dynamics has a predetermined minimum distance from the level of the direct sound. Thereby, the optimal AGC function is determined and set automatically for each acoustic scene, so that the occurrence of the comb-filter effect is avoided. Here, especially when the scene dynamics is mapped to the hearing range as optimally as possible at the same time, the speech and the environmental noise are preserved to the greatest extent, but now taking into account the possible comb-filter effect. The insight used here is that the comb-filter effect is most pronounced when the level of the direct sound is similar to the level of the output signal and there is a stationary disturbing noise in the current acoustic scene. In this case, the comb-filter effect is particularly disturbing, so that this case is avoided here in particular. It is assumed here that the stationary disturbing noise, for example the stationary background noise, forms the minimum, i.e. the lower end, of the scene dynamics. This is also referred to as the "noise floor". On the one hand, the stationary disturbing noise reaches the user as direct sound, the level of which corresponds to the actual level of the stationary disturbing noise. On the other hand, the disturbing noise reaches the user via the hearing device, i.e. the lower end of the scene dynamics and thus the stationary disturbing noise is mapped by the AGC function to the lower end of the predetermined output dynamics. Therefore, the comb-filter effect occurs in particular if the level of the direct sound is similar to the lower end of the output dynamics. By means of the corresponding minimum distance between the lower end of the output dynamics and the level of the direct sound, the risk of the comb-filter effect is thus reduced.

[0046] The above describes the level of the direct sound as the level of the stationary disturbing noise in the acoustic scene and / or as corresponding to the lower end of the scene dynamics as a preferred application case. But these descriptions are similarly applicable to other direct sound signals and similarly applicable to the upper end of the output dynamics, which is moved analogously if necessary to maintain the minimum distance from the level of the direct sound, so as to avoid the comb-filter effect or other effects.

[0047] The minimum distance is suitably at least 3 dB. Preferably, the minimum distance is at most 6 dB. The direction in which the minimum distance is maintained is secondary. In practice, the corresponding end of the output dynamics can be moved such that the direct sound level is either within or outside the output dynamics; the important thing is to maintain the minimum distance. Suitably, the corresponding end of the output dynamics is moved in the direction requiring the smallest change. For example, if the direct sound level is 59 dB and the lower end of the output dynamics is 60 dB, then the lower end is moved upwards by at least 2 dB, more preferably 5 dB, i.e., to 62 dB or 65 dB. Conversely, if the direct sound level is, for example, 46 dB and the lower end of the output dynamics is 45 dB, then the lower end is moved downwards by at least 2 dB, more preferably 5 dB, i.e., to 43 dB or 40 dB.

[0048] In principle, the direct sound level can be determined in various ways. However, determining the direct sound level is difficult and error-prone. Therefore, it is advantageous to determine the direct sound level simply by measuring the lower end (minimum) of the dynamic range (scene dynamics) and using it as the direct sound level. Suitablely, the direct sound level is the level of stable interference noise in the acoustic scene, in which case the above approximation is particularly reasonable. This method is particularly robust and does not require shutting down individual frequency bands. Instead, the scheme described here is based primarily on scene-dependent adaptations of the AGC function. Additional adaptations to avoid comb filtering effects are also advantageously performed at the lower end of the scene dynamics, i.e., at the interference noise end of the acoustic scene, while the upper end, which is particularly important for speech output, is suitably left unaffected by any corrections made to avoid comb filtering effects. This reduces comb filtering effects while preserving speech intelligibility and noise quality to the greatest extent possible.

[0049] For example, the above scheme is implemented as follows: First, check whether the absolute difference between the lower end of the output dynamics and the direct sound level is less than the minimum distance, as shown below:

[0050] |o low -n low <6dB?

[0051] Where, n low It is the direct sound level, and assumes n low =d low That is, the direct sound level is assumed to be the stable background noise at the lower end of the scene dynamics. If now additionally satisfy

[0052] n low ≥o low ,

[0053] Then set

[0054] o low =n low -6dB

[0055] Conversely, if additionally

[0056] n low <o low ,

[0057] then the following is set:

[0058] o low =n low + 6 dB

[0059] In a suitable design variant, the parameters are set depending on the frequency. In other words, the AGC input signal is divided into frequency bands or the input signal is already divided into frequency bands, for example by means of a filter bank, which is in particular part of the signal processing device. The parameters are then set individually for each frequency band, so that for each individual frequency band its own AGC function is used.

[0060] It is also advantageous to take into account the listener's auditory fatigue (equivalent: tiredness) when setting the parameters. In a suitable design variant, the listener's auditory fatigue is determined accordingly and one or more parameters are additionally (i.e. in addition to the scene dependency) set in dependence on the auditory fatigue. For example, the upper and / or lower end of the output dynamic is modified here by means of a psychoacoustic model, so that they deviate from the initially preset values if necessary. Thus, it is advantageous if the auditory fatigue is great, i.e. exceeds a corresponding threshold, over a certain period of time, for example 15 minutes, in the past, to lower the upper end of the output dynamic, i.e. to move it to a lower level value. Alternatively or additionally, conversely, if the auditory fatigue is small, i.e. below a corresponding threshold, over a certain period of time, for example 15 minutes, in the past, to raise the lower end of the output dynamic, i.e. to move it to a higher level value. The auditory fatigue is determined, for example, simply in dependence on the average level of the scene, it being assumed that there is high auditory fatigue if the average level is above a certain limit value and, conversely, there is low auditory fatigue if the average level is below a certain limit value.

[0061] Embodiments of the application are explained in detail below with reference to the drawings. Therein, in the drawings schematically:

[0062] Figure 1 a hearing device is shown,

[0063] Figure 2 a method for operating a hearing device is shown, Figure 1

[0064] Figure 3 an AGC function parameterized for a first scene is shown,

[0065] Figure 4 an AGC function parameterized for a further, second scene in Figure 3 is shown,​

[0066] Figure 5a The AGC function in Figure 3 is shown for a third scenario parameterization,

[0067] Figure 5b The AGC function in Figure 5a is shown, which is additionally adapted to reduce comb-filtering effects,

[0068] Figure 6a The AGC function in Figure 3 is shown for a fourth scenario parameterization,

[0069] Figure 6b The AGC function in Figure 6a is shown, which is additionally adapted to reduce comb-filtering effects.

[0070] An embodiment for a hearing device 2 is shown in Figure 1 , which is correspondingly configured to one user not explicitly shown. The hearing device 2 is correspondingly configured to the user only and is used by the user only. In the embodiment shown here, the hearing device 2 serves to support a hearing-impaired user and to compensate for the hearing loss of the user. In operation, the user wears and uses the hearing device 2, for example on the head and in or on the ear.

[0071] The hearing device 2 has an input transducer 4, here a microphone, an output transducer 6, here an earpiece, and a signal processing device 8, also called signal processing device. In operation, the input transducer 4 generates an input signal E, which is fed to the signal processing device 8. The signal processing unit modifies the input signal E and thereby finally generates an output signal A. In order to compensate for the hearing loss, the input signal E or a signal derived therefrom is amplified, for example by a gain unit 10, in dependence on a frequency-dependent gain factor according to the audiogram of the user. The output signal A is finally output to the user by the output transducer 6.

[0072] An exemplary method for operating the hearing device 2 is shown in Figure 2 . In step S1, the hearing device 2 acquires an audio signal U by means of the input transducer 4 and thereby generates an input signal E. From the input signal, an AGC input signal E AGC is generated, which is fed to the AGC unit 12 of the hearing device 2. In the simplest design, the AGC input signal E AGC is identical to the input signal E, alternatively only a part of the input signal E is used as AGC input signal E AGC , or the input signal E is modified beforehand by a further unit of the signal processing device 8 and then used as AGC input signal E AGCThe AGC unit 12 is part of the signal processing device 8 and is arranged in front of the gain unit 10 in the Figure 1 alternative, behind the gain unit or in parallel thereto.

[0073] The AGC unit 12 amplifies the AGC input signal E AGC in accordance with the level p E of the AGC input signal E AGC and in accordance with an AGC function 14, and outputs the correspondingly amplified AGC input signal E AGC as an AGC output signal A AGC . Six embodiments of the AGC function 14 are shown in Figures 3 to 5b . The AGC function 14 maps the level p A , more precisely the level value of the output signal A or the AGC output signal A AGC , to each level pE, more precisely to each level value of the input signal E or the AGC input signal E AGC . From the ratio of the two level values p E , p A a gain is generated, which is applied by the AGC unit 12 to the respective AGC input signal E AGC level value in order to amplify the AGC input signal. The gain can in principle be less than, greater than or equal to 1, i.e. attenuation and compression (<1), pure simple transmission (=1) or real amplification and expansion (>1) are possible.

[0074] The AGC function 14 is defined by one or more parameters 18, 20, g shift , s hearing , s loud , wherein, without restricting the generality, it is assumed that there are multiple parameters P. These parameters define the form or the course of the AGC function 14 and are therefore also referred to as AGC parameters. These parameters are adjustable in order to change the AGC function 14 and thus the modification of the AGC input signal E AGC by the AGC unit 12. Here, in step S2, these parameters are set in accordance with the currently existing acoustic scene S, i.e. in accordance with the current acoustic scene S. The acoustic scene S, in short "scene", is generally understood to be the acoustic environment in which the user can be. The current acoustic scene S is then the acoustic environment in which the user and thus the hearing device 2 is actually at the current time. The acoustic environment and thus the scene S is defined by any sound signals, including useful noise and disturbing noise, which can reach the user and the hearing device 2 with different intensities and / or from different directions, and by any acoustic properties of the environment.

[0075] The currently existing acoustic scene S is determined in the embodiment shown here in step S3 by a scene recognition unit 16. The input signal E or a signal derived therefrom is fed to the scene recognition unit 16, which determines the scene S on the basis thereof. The AGC unit 12 is connected to the scene recognition unit 16 and is controlled by the scene recognition unit, i.e. the parameters are set automatically in accordance with the determination of the current acoustic scene S. The scene recognition unit 16 outputs a control signal which depends on the respective scene S and is received by the AGC unit 12 in order to set the parameters in accordance with the scene.

[0076] Finally, the hearing device 2 outputs in step S4 from the AGC output signal A AGC an output signal A for output to the user. In the simplest design, the output signal A is identical to the AGC output signal A AGC . Alternatively, only a part of the AGC output signal A AGC is used as the output signal A, or the AGC output signal A AGC is modified in advance by a further unit of the signal processing device 8, for example a gain unit 10, as shown in Figure 2 , and is then used as the output signal A.

[0077] The AGC input signal E AGC (and generally the input signal E) has an input dynamic range d, and the AGC output signal A AGC (and generally the output signal A) has an output dynamic o, and the AGC function 14 maps the input dynamic range d to the output dynamic o. The input dynamic range d and the output dynamic o are generally also referred to as "dynamic" or "dynamic range", respectively. The terms "dynamic" and "dynamic range" are used synonymously here and are a measure of the range of values covered by the variable level of a signal, for example the input signal E, the output signal A, the AGC input signal E AGC , the AGC output signal A AGC . The term "lower end" then refers to the minimum value of the dynamic range d, and the term "upper end" similarly refers to the maximum value. In a particularly simple design, the lower end and the upper end d low , d high of the dynamic range d correspond to the minimum value and the maximum value of the level of the signal in question, for example the input signal E or the AGC input signal E AGC . But as lower end and upper end d high , d low of the dynamic range d, for example other definitions based on statistical analysis are also suitable, so that the lower end and the upper end d low , d high of the dynamic range d of the input signal E are determined on the basis of a statistical analysis of the input signal E. Similarly, the output dynamic o also has an upper end o high and a lower end olow .

[0078] Here, the AGC function 14 maps the dynamic range d to an output dynamic o. This is understood to map the lower end d low of the dynamic range d to the lower end o low of the output dynamic o high and to map the upper end d high of the dynamic range d to the upper end o low of the output dynamic o low . Between both, i.e. between the lower end and the upper end d high , o high a linear mapping takes place. Thereby an optimal overlap of the scene dynamic d and the output dynamic o is achieved and especially the user's hearing is optimally fully utilized.

[0079] The current acoustic scene S has a dynamic range d which is derived, for example, from the input signal E, for example directly determined as the minimum and maximum of the input signal E, or directly determined as the minimum and maximum of the AGC input signal E AGC which itself is derived from the input signal E. The dynamic range d of the current acoustic scene S is also called scene dynamic. The already mentioned input dynamic or the dynamic derived therefrom is particularly suitable as the dynamic range d of the current acoustic scene S. The determination of the scene S and its dynamic range d is part of the method described here and is carried out in step S2 or step S3.

[0080] The scene-dependent parameterization of the AGC function 14 serves to map the scene dynamic to a specific, preset output dynamic o, i.e. to keep the level p A of the output signal A within the respective value range. For this purpose, the output dynamic o is preset. The parameters of the AGC function 14 are set such that this AGC function maps the dynamic range d, i.e. the scene dynamic here, to the output dynamic o. Here, the lower end d low of the scene dynamic is mapped to the lower end o low of the output dynamic o high and the upper end d high of the scene dynamic is mapped to the upper end o low of the output dynamic o. Between both a linear mapping takes place. Overall, an optimal overlap of the scene dynamic and the output dynamic o is achieved.

[0081] In Figure 3 , 4Embodiments are shown in 5a and 6a, wherein the output dynamic o corresponds precisely to the user's auditory range h. The AGC function 14 is adjusted according to the scene so that the scene dynamics are optimally mapped to this auditory range h. The parameters of the AGC function 14 are set such that the final output dynamic o to the user corresponds as closely as possible to the auditory range h, thereby optimally overlapping the auditory range h with the scene dynamics and making optimal use of the user's hearing. However, it is advantageous in principle that the scene dynamics are at least largely, for example, at least 90%, mapped to the auditory range h.

[0082] The hearing range h is defined by the level values ​​that a user can perceive. Signals with levels within the hearing range h are audible to the user, while signals with levels outside this range are inaudible. Accordingly, the hearing range h is personalized for each user and may vary. Signals whose dynamics lie at least partially outside the hearing range h are intrusive to the user and may be difficult to understand. The hearing range h is determined, for example, by means of a suitable hearing test outside of the method described above. Here, the hearing range h is stored in the hearing device 2. In both embodiments, the output dynamics o correspond to the hearing range h, thereby effectively mapping scene dynamics to the hearing range h. Figure 3 and 4 The two embodiments are described in light of two different scenarios S. Figure 5a and 6a An embodiment is also shown, in which the scene dynamics are effectively mapped to the auditory range h. However, due to the risk of comb filtering effects in both scenes S, an additional adaptation is made to the AGC function 14 to achieve a slightly modified output dynamic o to reduce the comb filtering effects. The results are shown in Figure 5b and 6b As shown in [the diagram]. Therefore, in [the diagram] Figure 5b and 6b In this process, as an additional step after or during adaptation to the auditory range h, the AGC function 14 is further adapted by specifically moving the lower end o of the output dynamic o. low (Move the upper part of the alternative or additional site o) high That is, at the lower end of the hearing range h. low The lower end of the output dynamic o low The difference is set between them. This is described in detail below. In this way, although the output dynamic range o is slightly different from the auditory range h, the comb filtering effect can be reduced.

[0083] The dynamic design approach is similarly applicable to the auditory range h, the difference being that the auditory range h is not based on the actual signal, but rather on user attributes. Nevertheless, the auditory range h has a lower limit h. low (Minimum value) and upper hhigh (Maximum value), and gives the numerical range of the level. When mapping the scene dynamics d to the auditory range h, the parameters of the AGC function 14 are set such that the dynamic range of the output signal A output to the user corresponds to the auditory range h, thereby making the auditory range h optimally overlap with the scene dynamics d.

[0084] Here, the AGC function 14 and therefore the behavior of the AGC unit 12 are only altered when the scene S changes, i.e., when the scene recognition unit 16 detects a change in the current acoustic scene S. Within the same scene S, the same AGC function 14 is continuously used, however, this AGC function is optimized for scene S by setting parameters that depend on the scene. By setting parameters according to the current acoustic scene S, the AGC function 14 is adapted to scene S. This may even lead to expansion (gain > 1), in which case the SNR of the current acoustic scene S increases accordingly, while the dynamics of the output signal A are still limited to the user's hearing range h.

[0085] In the design scheme shown here, one of the parameters is the first inflection point 18, which is set at the lower end of the dynamic range d and the output dynamic o. low h low Another parameter is the second inflection point 20, which is set to the upper end of the dynamic range d and the output dynamic range o. high h high This is in Figure 3 and 4 As can be seen, the dynamic range d and auditory range h are defined by these two inflection points 18 and 20. These two inflection points 18 and 20 divide the AGC function 14 into three parts: the first segment 22 between the two inflection points 18 and 20, the second segment 24 after the second inflection point 20, and the third segment 26 before the first inflection point 18. All three segments 22, 24, and 26 are linear. The first segment 22 is most relevant to the user because it determines the AGC input signal E. AGC The gain within the output dynamic range. The second and third sections, 24 and 26, define the input signal E. AGC The gain of AGC is outside the output dynamic o. This is achieved by setting the two inflection points 18 and 20 as the lower and upper ends d. low h low d high h high ,exist Figure 3 , 4 In 5a and 6a, the scene dynamics d are accurately mapped to the auditory range h.

[0086] exist Figure 3 and 4The lower end h low of the hearing range h is exemplarily 40 dB, and the upper end h high is exemplarily 60 dB. The output dynamic o is the same as the hearing range h. In Figure 3 a first scenario S, the AGC function 14 is exemplarily shown, wherein the lower end d low of the dynamic range d is 20 dB, and the upper end d high is 80 dB. In this case, the slope on the first section 22 is < 1, and in operation, the dynamic range d is compressed in this scenario S, since it is larger than the output dynamic o and the hearing range h. In Figure 4 a further second scenario S, the AGC function 14 is exemplarily shown, wherein the lower end d low of the dynamic range d is 60 dB, and the upper end d high is 70 dB. In this case, the slope on the first section 22 is > 1, and in operation, the dynamic range d is expanded in this scenario S, since it is now larger than the output dynamic o and the hearing range h. The dynamic range d varies with the scenario S, while the output dynamic o and the hearing range h remain constant.

[0087] Here, one parameter is the offset g shift which is set to a value corresponding to the difference between the lower end h low of the output dynamic o and the lower end d low of the dynamic range d. For example, the lower end o low of the output dynamic o is determined such that the difference between the lower end d low of the dynamic range d does not exceed a maximum value m1.

[0088] Further, here, one parameter is the first slope s hearing which is set to a value corresponding to the ratio of the difference between the upper and lower ends o high , o low of the output dynamic o and the difference between the upper and lower ends d high , d low of the dynamic range d. Thus, the first slope s hearing is the slope of the AGC function 14 on the first section 22, which generally gives the gain of the AGC unit 12 for level values within the scenario dynamic d, depending on the size of the dynamic range d in the given current acoustic scenario S and the size of the output dynamic o and in particular the individualized hearing range h. The slope s hearing < 1, i.e. compression, as Figure 3 , slope = 1, i.e. simple transmission, or slope > 1, i.e. expansion, as Figure 4 . Optionally, the first slope s hearing is limited to a maximum value.

[0089] In the embodiment shown here, it is additionally determined whether speech is present in the input signal E, and if speech is not present in the input signal E, the first slope s is... hearing The value is limited to 1. A speech activity detection unit 28 is used to determine whether speech is present in the input signal E.

[0090] In addition, there is another parameter here, which is the second slope s. loud The second slope is set to be equal to the maximum auditory level h. max The upper end of the output dynamic o high The difference between the maximum scene level d and the maximum scene level d max and the upper end of the dynamic range d high The ratio of their differences. The second slope s loud This indicates the gain of AGC unit 12 along the second segment 24. Figure 3 and 4 In the middle, the maximum auditory level h max It is 70dB, and the maximum scene level is d. max It is 90dB.

[0091] Optionally, the second slope s loud It is restricted to a maximum value of 1, meaning that expansion is not allowed on the second segment 24.

[0092] The scene-dependent adaptation of AGC function 14 can also reduce the comb filtering effect, which will be discussed below. Figure 5a , 5b 6a and 6b further elaborate on this. Among them, Figure 5a , 5b A restaurant scene is shown. Figure 6a , 6b A street scene is shown. The comb filtering effect is an artifact generated by the superposition of the output signal A and the direct sound signal (referred to as direct sound). The comb filtering effect is most pronounced when the background noise of scene S is stationary, i.e., varies very little, and the direct sound and output signal A have similar levels. The level n of the direct sound signal... low Also known as direct acoustic level n low .

[0093] Here, the adjustable AGC function 14 is used to avoid the comb filtering effect. To do this, the direct sound level n of the current acoustic scene S is first determined. low Furthermore, one or more parameters of the AGC function 14 are set such that the lower end of the output dynamic o is... low or the upper part o high With direct sound level n lowwith a preset minimum distance x. Thereby, an optimal AGC function 14 is automatically determined and set for each acoustic scene S, so that the occurrence of comb filtering effects is avoided. Here, especially when simultaneously mapping the scene dynamics optimally as far as possible to the hearing range h as described above, speech and ambient noise are preserved to the greatest extent, while now the potential comb filtering effects are taken into account. The insight exploited here is that, when the direct sound level n low is similar to the level p A of the output signal o and a stationary disturbing noise is present in the current acoustic scene S, comb filtering effects are most pronounced. Specifically, it is thus assumed here that the stationary disturbing noise, for example a stationary background noise, constitutes the minimum value of the scene dynamics, i.e. the lower end d low . This is also referred to as the "noise floor". This stationary disturbing noise reaches the user on the one hand as direct sound with the direct sound level n low corresponding to the actual level of the stationary disturbing noise. On the other hand, the disturbing noise reaches the user via the hearing device 2, i.e. the lower end d low of the scene dynamics, and thus the stationary disturbing noise is mapped by the AGC function 14 to the lower end o low of the preset output dynamics o, here i.e. to the hearing range h, as shown in Figure 5a , 6a . Thus, especially when the direct sound level n low is similar to the lower end o low of the output dynamics o, as shown in Figure 5a , 6a , a comb filtering effect occurs. By a respective minimum distance x between the lower end o low of the output dynamics o and the direct sound level n low , as shown in Figure 5b and 6b , the risk of a comb filtering effect is reduced.

[0094] In the embodiment shown here, the direct sound level n low is the level of the stationary disturbing noise in the acoustic scene S and corresponds to the lower end d low of the dynamic range d of the scene S. But these design proposals are similarly applicable to other direct sound signals and similarly to the upper end o high of the output dynamics o, which can similarly be shifted to maintain a minimum distance x to the direct sound level, here n high , not shown in the figure, so that comb filtering effects are avoided.

[0095] The minimum distance x here is 6 dB. It is secondary in which direction the minimum distance x is maintained, two variants are shown in Figure 5b , 6b . In practice, the lower end o lowIt can be moved in this way to achieve a direct sound level n low Within the output dynamic o ( Figure 6b Or outside of the output dynamics ( Figure 5b The important thing is to maintain the minimum distance x. Here, the lower end o of the output dynamic o is... low Move in the direction where the desired change is smaller. Therefore, in Figure 5b In the middle, the direct sound level n low The output is 59dB, and the lower end of the dynamic output o is... low If it is 60dB, then the lower end o low Shifting up by 5dB, that is, moving to 65dB. Conversely, if the direct sound level n... low like Figure 6b The value shown is 46dB, representing the lower end of the dynamic output o. low If it is 45dB, then the lower end o low Shifting downwards by 5dB, i.e., moving to 40dB. This shift is achieved by moving the first inflection point 18. This also results in a different slope on the first segment 22. However, in principle, other adaptations of the AGC function 14 could also be considered to shift the lower end o accordingly. low And establish the minimum distance x. Figures 5a to 6b In this context, the output dynamic o that precisely corresponds to the auditory range h is labeled o1, while the output dynamic o that shifts to reduce the comb filtering effect is labeled o2.

[0096] In principle, the direct sound level n low It can be determined in different ways. Here, it is simply by measuring the lower end d of the dynamic range d. low Use it as the direct sound level n low Determine the direct sound level n low This is feasible because the direct sound level n low It is the level of stable interference noise in acoustic scene S.

[0097] Figure 5b , 6b The additional adaptation shown to avoid the comb filtering effect is applied at the lower end of the scene dynamics d. low The process takes place at the noise-affected end of the acoustic scene S, while the upper end d, which is particularly important for speech output, is not. high It is therefore unaffected by any corrections made to avoid the comb filtering effect. This reduces the comb filtering effect while preserving speech clarity and noise quality to the greatest extent possible.

[0098] Here, the above solution is implemented as follows: First, check the lower end o of the output dynamic o. low With direct sound level n low Check if the absolute difference between them is less than the minimum distance x, i.e., check if the following conditions are met:

[0099] |o low -n low |<6dB?

[0100] Here we assume n low =d low That is, the direct sound level n low Assumed to be the lower end of the scene dynamics d low The background noise is stable. If now additionally satisfied:

[0101] n low ≥o low ,

[0102] like Figure 6a As shown, the settings are as follows:

[0103] o low =n low -6dB,

[0104] like Figure 6b As shown. Conversely, if additionally the following conditions are met:

[0105] n low <o low ,

[0106] like Figure 5a As shown, the settings are as follows:

[0107] o low =n low +6dB,

[0108] like Figure 5b As shown.

[0109] In a design scheme not explicitly shown, parameters are set depending on the frequency. In other words: for example, the AGC input signal E is controlled by a filter bank. AGC The input signal E is divided into multiple frequency bands, and this filter bank is, for example, part of the signal processing device 8. Then, parameters are set individually for each frequency band, thereby using its own AGC function 14 for each individual frequency band.

[0110] Alternatively, user auditory fatigue (equivalent to exhaustion) can be considered when setting parameters. In a suitable design, user auditory fatigue is determined accordingly, and one or more parameters are additionally (in addition to context-dependent settings) based on auditory fatigue.

[0111] List of reference numerals

[0112] 2 Hearing devices

[0113] 4-input converter

[0114] 6 output converter

[0115] 8 signal processing device

[0116] 10 gain unit

[0117] 12 AGC unit

[0118] 14 AGC function

[0119] 16 scene recognition unit

[0120] 18 first knee

[0121] 20 second knee

[0122] 22 first section

[0123] 24 second section

[0124] 26 third section

[0125] 28 voice activity detection unit

[0126] A output signal

[0127] A AGC AGC output signal

[0128] D dynamic range of a scene, scene dynamic

[0129] d low lower end of the scene dynamic

[0130] d high upper end of the scene dynamic

[0131] d max maximum scene level

[0132] E input signal

[0133] E AGC AGC input signal

[0134] g shift offset

[0135] h hearing range

[0136] h low lower end of the hearing range

[0137] h high upper end of the hearing range

[0138] h max maximum hearing level

[0139] m1 maximum value (for offset)

[0140] nlow Direct sound level

[0141] o, o1, o2 output dynamics (e.g. of the output signal or AGC output signal)

[0142] o low Lower end of the output dynamics

[0143] o high Upper end of the output dynamics

[0144] p A Level of the output signal / AGC output signal

[0145] p E Level of the input signal / AGC input signal

[0146] S scenario

[0147] S1-S4 steps

[0148] s hearing First slope

[0149] s loud Second slope

[0150] U audio signal

[0151] x minimum distance

Claims

1. A method for operating a hearing device (2), the hearing device being configured for a user, a. Wherein, the hearing device (2) acquires audio signals (U) and thereby generates an input signal (E), b. Among them, Generate AGC input signal (E) from input signal (E) AGC The AGC input signal is fed to the AGC unit (12) of the hearing device (2). c. Among them, the AGC unit (12) according to the AGC input signal (E) AGC The level of the AGC signal is amplified according to the AGC function (14) and the AGC input signal (E) is amplified. AGC ), and then use it as the AGC output signal (A AGC Output, d. Among them, the AGC function (14) is passed through one or more parameters (18, 20, g). shift s hearing s loud The parameters are defined and set according to the current acoustic scene (S). e. Among them, the hearing device (2) outputs a signal (A) from the AGC. AGC It generates an output signal (A) to be output to the user.

2. The method according to claim 1, in, An output dynamic (o) is preset, wherein the output dynamic has a lower end (o). low ) and the upper end (o high ), Specifically, the dynamic range (d) of the acoustic scene (S) is determined, wherein the dynamic range has a lower end (d). low ) and the upper end (d) high ), The parameter is set such that the AGC function (14) maps the dynamic range (D) to the output dynamic (o).

3. The method according to claim 2, in, The output dynamic (o) corresponds to the user's auditory range (h).

4. The method according to claim 3, in, The auditory range (h) is the user's comfortable auditory range, and the dynamic range (d) is a correlated dynamic range.

5. The method according to any one of claims 2 to 4, in, One of the parameters is the first inflection point (18), which is set to be located at the lower end (d) of the dynamic range (d) and the output dynamic (o). low o low ), Another parameter is the second inflection point (20), which is set to be located at the upper end (d) of the dynamic range (d) and the output dynamic (o). high o high ).

6. The method according to any one of claims 2 to 5, in, One of the parameters is the offset (g) shift ), Wherein, the offset (g) shift The value is set to the following value, which corresponds to the lower end (d) of the output dynamic range (o) and the dynamic range (d). low o low ) difference.

7. The method according to claim 6, in, The lower end of the output dynamic (o) low ) is determined such that it is relative to the lower end (d) of the dynamic range (d). low The difference between (m1 and m2) does not exceed the maximum value (m1).

8. The method according to any one of claims 2 to 7, in, One of the parameters is the first slope (s) hearing The first slope is set to correspond to the upper and lower ends (h) of the output dynamic (o). low h high The difference between the upper and lower ends of the dynamic range (d) and the dynamic range (d) low d high The ratio of the difference between ).

9. The method according to claim 8, in, Determine whether speech is present in the input signal (E). Wherein, if there is no speech in the input signal (E), then the first slope (s) hearing The value is limited to 1.

10. The method according to any one of claims 2 to 9, in, One of the parameters is the second slope (s) loud The second slope is set to correspond to the maximum auditory level (h). max ) and the upper end of the output dynamic (o) (o) high The difference between ) and the maximum scene level (d) max ) and the upper end of the dynamic range (d) high The ratio of the difference between ).

11. The method according to claim 10, in, The second slope (s) loud The value is limited to a maximum of 1.

12. The method according to any one of claims 2 to 11, in, The lower and upper ends (d) of the dynamic range (d) low d high The minimum and maximum values ​​of the input signal (E) within a predefined time range are determined.

13. The method according to any one of claims 2 to 12, in, The lower and upper limits (d0 and d1) of the dynamic range (d) are determined based on statistical analysis of the input signal (E). low d high ).

14. The method according to any one of claims 2 to 13, in, Determine the direct sound level of the acoustic scene (S). Among them, one or more parameters (18, 20, g) shift s hearing s loud ) is set such that the lower end (o) of the output dynamic (o) is... low ) or upper end (o high It has a preset minimum distance from the direct sound level.

15. The method according to claim 14, in, The direct sound level is determined as follows: the lower end of the dynamic range is measured and used as the direct sound level.

16. The method according to claim 14 or 15, in, The minimum distance is at least 3 dB, and preferably at most 6 dB.

17. The method according to any one of claims 14 to 16, in, The hearing device (2) has an ear canal microphone through which the direct sound level is measured.

18. The method according to any one of claims 1 to 17, in, The parameters are set depending on the frequency.

19. The method according to any one of claims 1 to 18, in, Determine if the user is experiencing auditory fatigue. The one or more parameters are additionally set according to the auditory fatigue.

20. A hearing device (2) designed to perform the method according to any one of claims 1 to 19.

Images