Method for suppressing acoustic reverberation in an audio signal

Abstract

The invention relates to a method for suppressing acoustic reverberation in an audio signal (50), wherein an audio signal is provided, wherein a first level measurement (1, 1') of the audio signal is performed, wherein a second level measurement (2, 2') of the audio signal is performed at least sometimes during the first level measurement, wherein the first level measurement is performed with a decay time (16, 16') and / or an adjustment time (6, 6') that differs from a decay time (18, 18') or an adjustment time (12, 12') of the second level measurement, wherein acoustic reverberation of a sound event (4) in the audio signal is suppressed by attenuating the audio signal (50) depending on the contribution of the sound event (4) in the audio signal, and wherein the attenuation is controlled depending on a difference (20, 20') between the first level measurement and the second level measurement.

Description

[0001] This application is a divisional application of a patent application with a filing date of June 21, 2019, Chinese patent application number 201910542189.X, and invention name “Method for suppressing sound reverberation in audio signals”. Technical Field

[0002] The present invention relates to a method for suppressing acoustic reverberation in an audio signal. Background Art

[0003] Sound reverberation is often formed in closed or at least partially closed spaces due to multiple reflections of a generated sound event off the walls and other objects within the space. The decay time of the reverberation varies depending on the geometry of the space and its walls, as well as the type, number, and geometry of the objects within it. The decay time of the reverberation is also influenced by the properties of the surfaces within the space. Unlike echoes, which can be perceived in isolation as a "repetition" of the generated sound event, reverberation forms a substantially continuous "echo" of the sound event.

[0004] Although a minimal amount of reverberation is even desirable for a pleasant sound experience, especially for music, in order to counteract overly "dry" sounds like staccato, reverberation is often disadvantageous for understanding speech contributions, since characteristic sound events, particularly for distinguishing individual consonants, have only a very short duration and, when combined with the corresponding superposition of reverberation, can sometimes significantly distort the spectral information. Depending on the decay time, this can even pose problems for distinguishing formants, which are used to identify vowels.

[0005] In hearing devices, which are often used to compensate for the hearing loss of the wearer, it is particularly important to reproduce the wearer's conversation partner's speech contribution as easily as possible for the wearer to understand. This is because a lack of acoustic understanding of the speech contribution, accompanied by a loss of perceptible information for the wearer, can be particularly noticeable and therefore uncomfortable. Therefore, technologies are often used in hearing devices to improve the intelligibility of speech contributions.

[0006] However, especially when using hearing instruments, acoustic reverberation can significantly affect the crucial aspect of speech intelligibility: in hearing instruments, the wearer-specific signal processing performed there to compensate for the wearer's hearing loss often also involves dynamic compression of the input signal generated by the hearing instrument's microphone. Dynamic compression should particularly help amplify quiet sound events (which the wearer can barely perceive or no longer perceive due to the wearer's hearing loss) until sufficient perceptibility is achieved, while the same amplification is not applied to sufficiently loud sound events that the wearer can perceive without major problems, so that further amplification could result in an unpleasant volume.

[0007] However, this dynamic compression also "compresses" the acoustic reverberation, so that it experiences a correspondingly greater amplification than the sound event that generated it. As a result, on the one hand, the wearer perceives a longer decay time in the current environment and, on the other hand, the intelligibility of the speech contribution is affected due to the interrelationship described above. Summary of the Invention

[0008] The object of the present invention is therefore to provide a method for suppressing acoustic reverberation in an audio signal, which method can be executed in real time with the lowest possible computational complexity and in doing so provides a sonogram that is as realistic as possible.

[0009] According to the invention, the stated technical problem is solved by a method for suppressing acoustic reverberation in an audio signal, wherein an audio signal is provided, wherein a first level measurement is performed on the audio signal, wherein at least sometimes during the first level measurement, a second level measurement is performed on the audio signal, wherein the first level measurement is performed with a decay time that differs from a decay time of the second level measurement, and / or wherein the first level measurement is performed with an adjustment time that differs from an adjustment time of the second level measurement, wherein acoustic reverberation of a sound event in the audio signal is suppressed by attenuating the audio signal depending on the contribution of the sound event in the audio signal, and wherein the attenuation is controlled depending on the difference between the first level measurement and the second level measurement. Advantageous embodiments, which are considered in part to be inventive in themselves, are the subject of the following description.

[0010] In this context, suppressing acoustic reverberation in an audio signal is understood in particular to mean suppressing signal contributions in the audio signal that are formed by acoustic reverberation in a real acoustic situation mapped by the audio signal. In this context, the audio signal is provided in particular by means of one or more electroacoustic converters, which convert the real acoustic situation into one or more, in particular electrical, signals. In order to provide the audio signal, preprocessing can also be performed on the one or more, in particular electrical, signals generated in this way, which preprocessing can include, for example, digitization, amplification, dynamic compression or also noise suppression. In this context, acoustic reverberation is understood in particular to mean reflections of the sound of a generated acoustic event on walls and / or objects, for example of an at least partially enclosed space, wherein the multiple reflections of the propagating sound generated by the acoustic event at fixed positions result in a continuous or almost continuous fading of the acoustic event.

[0011] In this context, a level measurement is understood to mean, in particular, that a mathematical function is formed by the level measurement, or that the level measurement can be represented as a function by which the amplitude of the audio signal and / or the envelope of the amplitude and / or the square of the amplitude value is mapped, preferably in a strictly monotonic manner and particularly preferably without inflection points, to a corresponding level value. In particular, functions are also provided in which the relationship between the function input variable and the mapped level value is not merely logarithmic, and the term level measurement is intended to include more general functions with suitable monotonic properties.

[0012] Performing a second level measurement of the audio signal at least sometimes during a first level measurement is understood to mean, in particular, that only one of the two level measurements is performed at a discrete point in time. This level measurement can be performed during discretization during digitization when providing the audio signal, wherein missing level values ​​between the discrete support points thus formed can be interpolated, thereby enabling a comparison with the other level measurement, in particular a continuous one, over the time period to be examined. However, preferably, both level measurements can be performed continuously (during a discretization predetermined by digitization).

[0013] In this context, the decay time of the level measurement is to be understood as meaning, in particular, the time that elapses after a signal contribution in the audio signal and a corresponding level shift of the level measurement until the level measurement drops to zero or to a predetermined fraction of the level shift in the absence of further signal contributions in the audio signal, i.e., for example, the time constant T60 for a 60 dB reduction. The settling time of the level measurement is to be understood as meaning, in particular, the time that elapses after a spontaneously occurring static signal contribution in the audio signal until the level measurement reaches a predetermined fraction of the asymptotic limit value of the signal level corresponding to the static signal contribution. A shorter settling time in this context means, in particular, a faster reaction of the level measurement to spontaneously occurring signal contributions in the audio signal.

[0014] In particular, each of the two level measurements is therefore associated with a settling time and a decay time, wherein for the method herein, the first level measurement and the second level measurement differ from one another in at least one of the two parameters, the settling time and the decay time, respectively associated therewith, wherein the level measurements may differ not only in the settling time but also in the decay time. In particular, the settling time of the first level measurement may be referred to as the first settling time, and the decay time of the first level measurement as the first decay time, wherein the corresponding parameters associated with the second level measurement may be referred to as the second settling time and the second decay time.

[0015] Therefore, the following three possibilities are obtained, namely, the two adjustment times are the same, but the first decay time is different from the second decay time, or the two decay times are the same, but the first adjustment time is different from the second adjustment time, or the first decay time is different from the second decay time, and the first adjustment time is also different from the second adjustment time.

[0016] A sound event is understood to be, in particular, any event that generates a sound in a real sound situation that is mapped by an audio signal and / or used to provide an audio signal by means of a corresponding converter, wherein the sound-generating event can be associated with a clear end point in time. In this sense, suppressing the sound reverberation of a sound event in an audio signal means, in particular, suppressing the signal contribution that corresponds to the sound reverberation of the sound event in the real sound situation.

[0017] In this case, the audio signal can be attenuated for suppressing acoustic reverberation, in particular according to corresponding amplification factors or attenuation factors, wherein the factors are controlled according to the difference between the first level measurement and the second level measurement.

[0018] The proposed method makes it possible to distinguish the contribution of a sound event in an audio signal from the contribution of a sound reverberation corresponding to the sound event from each other based on the different level curves in the first and second level measurements resulting from different decay times or different adjustment times, and to utilize this to perform corresponding attenuation.

[0019] In this case, the first level measurement differs from the second level measurement at least in terms of the decay time or the adjustment time, or in terms of the two time constants mentioned. Given the different decay times of the two level measurements, the contribution of acoustic reverberation decreases more rapidly in the level measurement with the shorter decay time than in the level measurement with the longer decay time. If an independent sound event is no longer present in the audio signal, but only the contribution of acoustic reverberation is still recorded, the difference between the first and second level measurements increases. This can be used to control the attenuation of the audio signal, so that the audio signal is attenuated more as the difference increases. In the case of different adjustment times, a sound event that starts suddenly causes a sudden increase in the difference between the two measurements, because the level measurement with the slower adjustment time reacts to the onset of the sound event later. Therefore, given the different adjustment times, this sudden increase in the difference can be used to control the attenuation of the audio signal to a certain extent, or not at all.

[0020] In particular, the proposed method can be applied per frequency band, i.e., for the level measurement mentioned, the audio signal is decomposed into the individual frequency bands, and two level measurements are then performed correspondingly per frequency band based on the decomposition, and the respective attenuation of the audio signal in the individual frequency bands can be controlled differently accordingly. This takes into account the fact that the decay of acoustic reverberation in a given environment can have a partially pronounced frequency dependency, and therefore, for certain frequency bands, it may be advantageous to apply no attenuation or only a small attenuation to the audio signal in order to minimize changes or distortions to the audio signal.

[0021] Preferably, a first level measurement is performed with a first decay time, and a second level measurement is performed with a second decay time, the second decay time being greater than the first decay time, wherein the attenuation of the audio signal is controlled by a monotonically increasing function of the difference between the first level measurement and the second level measurement to suppress acoustic reverberation in the audio signal. If only acoustic reverberation of a sound event is still present in the audio signal, without any contribution from the sound event itself being recorded, the signal profile in the audio signal is essentially given by the decay characteristics that characterize the actual acoustic situation mapped by the audio signal. For contributions to the audio signal caused by acoustic reverberation, the difference between the two level measurements increases over time due to the different decay times used for the first and second level measurements.

[0022] This difference can now be used, firstly, as a safety indicator that the relevant signal contribution is actually provided by the acoustic reverberation, and secondly as an indicator of the temporal distance from the sound event that generated the acoustic reverberation. In this regard, in the described case, the acoustic reverberation in the audio signal can now be suppressed in such a way that the audio signal is suppressed more strongly as the difference between the two level measurements increases, i.e., the decay is described by a monotonically increasing function of the difference, since it is assumed that, as the difference increases, no independent new signal contribution is generated by sound events other than the sound event that generated the acoustic reverberation, and therefore, only the decaying acoustic reverberation still provides a substantial contribution to the audio signal. In particular, when the audio signal is further processed by dynamic compression, it can be advantageous to suppress the acoustic reverberation in the described manner as the decay increases, so that the decay is not artificially increased and thus prolonged by the dynamic compression.

[0023] Advantageously, a first level measurement is performed with a first adjustment time, and a second level measurement is performed with a second adjustment time that is identical to the first adjustment time. The second decay time is greater than the first decay time. This has the advantage that sound events in both level measurements each produce a level shift with the same adjustment characteristics, whereby independent sound events can be identified based on the two level measurements, which can be used to control the suppression of sound reverberation. For example, by forming the difference between the two level measurements already mentioned, in the event that the level shift in both level measurements is simultaneously high, an independent sound event can be inferred, and in this case, the attenuation of the audio signal for suppressing the previously initiated sound reverberation can be stopped, if necessary.

[0024] In one advantageous embodiment, a first level measurement is performed with a first settling time and a first decaying time, and a second level measurement is performed with a second settling time that is different from the first settling time and the same as the first decaying time. A sound event is inferred based on an increase in the difference between the first and second level measurements. After the onset of the sound event, attenuation of the audio signal is controlled using a monotonically decreasing function of the difference between the first and second level measurements to suppress acoustic reverberation in the audio signal. In particular, the recognition of a sound event can be triggered by an increase in the difference exceeding a predetermined threshold value.

[0025] If a sound event that may generate a signal contribution to acoustic reverberation in the audio signal is inferred in the described manner, then due to the identical decay times of the two level measurements, the acoustic reverberation exhibits essentially the same decay characteristics in both level measurements when the additional signal contributions disappear. Ideally, the difference between the level measurements in the two level measurements approaches zero at the beginning of the decay characteristics. When the measured levels reach the acoustic reverberation of the actual sound event present in the audio signal, the two level measurements each transition to their decay characteristics and then, according to the common decay time of the two level measurements, decay along with the acoustic reverberation. In this regard, the difference between the two level measurements can be used to control the suppression of acoustic reverberation, such that after a sound event is identified, a decreasing, in particular, disappearing, difference is used as an indication that no additional contributions from the independent sound event are present in the audio signal, but only acoustic reverberation, thereby maximally reducing the audio signal. A variable, in particular, increasing, difference after a sound event is identified can be interpreted as an indication that additional, smaller sound events are still present, and therefore, preferably, the audio signal is still not suppressed to the maximum extent.

[0026] The attenuation can be controlled in this way, on the one hand, such that, after the sound event has been identified based on the increase in the difference, as soon as the difference between the first level measurement and the second level measurement falls below a predetermined limit value again, a constant attenuation is used as a "trigger," since it is assumed that there is a sound event, i.e., in particular, a decaying characteristic of the sound reverberation of the sound event, which causes the difference in the level measurements to disappear. On the other hand, after the sound event has been identified, the attenuation can also be increased in a predetermined manner over time, and if necessary, as a function of the maximum level of the sound event as a reference variable, which further reduces the contribution of the sound reverberation in the audio signal.

[0027] Advantageously, the first level measurement is selected such that it exhibits an exponential decay characteristic. Preferably, the second level measurement is also selected to exhibit an exponential decay characteristic. An exponential decay characteristic is understood to mean that, after a sound event with a clearly defined end point, the level value output by the corresponding level measurement decreases exponentially with increasing time, or decreases linearly with time on a logarithmic scale. This has the particular advantage that the decay characteristic of the first level measurement reflects the actual decay characteristic of the acoustic reverberation.

[0028] In a real sound situation, a sound event may be reflected on walls and objects in a room, wherein, for a given environment, the corresponding reflected sound is related to the sound pressure level of the initial sound event. On the one hand, the reflected sound may be reflected again on further walls and objects, wherein the proportion of the sound reflected again is likewise related to the sound level of the first reflection. The result of this correlation of successive reflections is an exponential decay characteristic of the sound events in a real sound situation. By selecting a first level measurement which also has an exponential decay characteristic, the first level measurement can be adjusted, with a suitable selection of the associated first decay time, so that the first level measurement essentially reproduces the decay characteristic of the real sound situation which is the basis of the audio signal.

[0029] The first level measurement is preferably achieved by a first-order recursive low-pass filter. Such a filter provides the desired exponential decay characteristic. It is also particularly preferred to achieve the second level measurement by a first-order recursive low-pass filter.

[0030] Advantageously, a physical decay time constant is selected for the environment in which the audio signal to be provided is generated, serving as the first decay time for the first level measurement. Generating an audio signal is understood to mean, in particular, a corresponding conversion of the actual sound situation by means of a suitable converter. For example, T60, which is the time for a 60 dB drop in the sound level, can be used as the decay time constant.

[0031] It has further proven advantageous to determine the noise background of the audio signal, wherein the attenuation of the audio signal is controlled as a function of the noise background of the audio signal and particularly preferably the attenuation of the audio signal is modified, applied, stopped or restarted as a function of the noise background. In order to determine the noise background, in particular the knowledge of sound events obtained according to the first level measurement and / or the second level measurement can be used so that, for example, recognized sound events exceeding a minimum level are not interpreted as noise. However, the noise background of the audio signal can also be determined in other ways known to a person skilled in the art. In this case, when suppressing sound reverberation, a modification of the attenuation of the audio signal, in particular in the form of a correspondingly smaller application of the attenuation of the audio signal as a function of the noise background, can prevent the determined sound reverberation from causing the audio signal to drop or be "suppressed" below its actually averagely present noise contribution, which could be perceived as unnatural.

[0032] In this case, it is particularly preferred to perform noise suppression in the audio signal, wherein the attenuation is additionally controlled as a function of the noise suppression. This means, in particular, that the audio signal does not fall below the noise background present in the audio signal after the noise suppression has been applied.

[0033] The present invention further relates to a method for suppressing acoustic reverberation in an audio signal of a hearing device, wherein the audio signal is provided by means of an input converter of the hearing device as a function of an ambient acoustic signal, and wherein acoustic reverberation is suppressed in the audio signal by means of the method described above. The present invention also relates to a hearing device comprising an input converter for generating an audio signal and a signal processing unit configured to perform the method described above. The advantages presented for the method for suppressing acoustic reverberation in an audio signal and its developments are also transferable to the method in the hearing device and the hearing device itself. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[0035] Figure 1 A time diagram schematically illustrates a first level measurement and a second level measurement of the same audio signal with the same adjustment time and correspondingly different decay times,

[0036] Figure 2 schematically illustrates a first level measurement and a second level measurement of the same audio signal with the same fade-out time and correspondingly different adjustment times, and

[0037] Figure 3 A hearing device is schematically shown in a block diagram.

[0038] In all figures, parts and variables that correspond to one another are provided with the same reference symbols. DETAILED DESCRIPTION

[0039] exist Figure 1 , the level values ​​P of a first level measurement 1 (solid line) and a second level measurement 2 (dashed line), respectively, performed on an audio signal (not shown in detail), are schematically shown in a time diagram over time t. At time T0, an isolated sound event 4 (dotted line) is present in the audio signal. This sound event 4, on the one hand, has a clearly defined end point and, on the other hand, produces a contribution of acoustic reverberation in the audio signal due to the physical environment in which the audio signal was recorded. Sound event 4 has only a very short duration ΔT. Therefore, all the acoustic energy of sound event 4 is concentrated within this duration ΔT. This can be the case, for example, with explosions, knocks, applause, or similar noises of very short duration.

[0040] The first level measurement 1 has a first settling time 6, which elapses after time T0 at which the sound event 4 begins, before the first level measurement has a predetermined portion 8 of an asymptotic level 10, wherein the asymptotic level 10 corresponds to the level that the first level measurement would have for a static, continuous sound event with the same signal level as the sound event 4. The second settling time 12 of the second level measurement 2 is identical to the first settling time 6 of the first level measurement.

[0041] For this reason, the first level measure 1 and the second level measure 2 have the same regulation characteristic 13 and therefore have the same maximum value 14 at the time T1 marking the end of the duration ΔT and therefore the end of the sound event 4, which is located a little below the asymptotic level 10. At the time T1, in the real sound situation mapped by the audio signal, the decay characteristic caused by the reverberation now begins, so that the first level measure 1 and the second level measure 2 now also transition to their decay characteristic 15 according to the corresponding decay time.

[0042] Here, the first level measurement 1 has a first decay time 16 which corresponds exactly to the decay time constant T60, after which the sound level decreases by 60 dB from the maximum value 14. The second level measurement 2 has a second decay time 18 which is greater than the first decay time 16. For this reason, for the first level measurement 1, the sound level decreases by 60 dB from the maximum value 14. Figure 1 The logarithmic scale of the diagram describes the exponential decay behavior of the straight line, which is correspondingly steeper than the straight line describing the decay behavior of the second level measurement 2 .

[0043] If a difference 20 is now determined from the first level measurement 1 and the second level measurement 2, this difference 20 can be used to suppress the proportion of acoustic reverberation in the audio signal, so that the audio signal is attenuated in a manner that is monotonically increasing in relation to the difference 20. For this purpose, Figure 1 The sound level of the audio signal 22 (dotted line) resulting from this attenuation is schematically shown in FIG.

[0044] In this case, the attenuation is performed only until the noise floor 24 of the audio signal is reached, so that at time T2, at which the attenuated audio signal 22 falls below the noise floor 24 due to further attenuation according to the difference 20 between the first level measure 1 and the second level measure 2, the attenuation is gradually reversed depending on the noise floor 24. In particular, instead of depending on the noise floor 24, the aforementioned attenuation can also be reversed depending on the residual noise floor 26, which is formed by the residual noise remaining when the noise suppression is performed.

[0045] exist Figure 2 In FIG, the level values ​​P of a first level measurement 1 ' (solid line) and a second level measurement 2' (dashed line) of an audio signal in response to a sound event 4 (dotted line) are schematically shown in a time diagram over the time t, wherein the sound event starts at the time T0 and ends again at the time T1 after a duration ΔT. Figure 2 In the example, the first adjustment time 6' is shorter than the second adjustment time 12'. This means that at time T1, the first level measurement 1' has a maximum value 14 that is significantly above the maximum value 14' of the second level measurement 2'. Furthermore, the second level measurement 2' does not reach its maximum value 14' until later than time T1. This means that the difference 20' has already significantly diverged during the adjustment characteristics 13' of the two level measurements 1', 2'. Therefore, a sound event 4 can now be identified.

[0046] Here, the first decay time 16' of the first horizontal measurement 1' is identical to the second decay time 18' of the second horizontal measurement 2' and is reselected as the decay time constant T60 (see Figure 1 ). After the sound event 4 is identified after the time point T0 based on the difference 20' of the sudden separation of the two level measurements 1', 2', the difference of the two level measurements 1', 2' decreases again, because immediately after the time point T1, the first level measurement 1' has transitioned to its decay characteristic. Here, the second level measurement 2' is first still further adjusted to the decaying sound reverberation of the sound event 4 until the maximum value 14' of the second level measurement 2' reaches the actual instantaneous level of the decaying sound reverberation of the sound event 4. From the decrease of the difference 20' starting from the time point T1, it can be deduced accordingly that the sound event 4 has now ended and only the decay characteristic 15' is present in the audio signal. In this case, if the occurrence of the sound event 4 was previously identified based on the sudden increase of the difference 20', a predetermined attenuation can be applied to the audio signal in the process of the decrease of the difference 20'. As shown according to Figure 1 As described, in order to suppress the acoustic reverberation, the attenuation of the audio signal can be modified or reversed depending on the noise background 24 or the residual noise background 26 .

[0047] In particular, the frequency bands can be used to perform Figure 1 and according to Figure 2 A method for suppressing acoustic reverberation in an audio signal is shown. To this end, the audio signal is divided into frequency bands, in which the first and second level measurements 1, 1', 2, 2' shown are performed. The attenuation of the audio signal can then be controlled for each frequency band individually, depending on the difference 20, 20' determined in this frequency band, according to the corresponding adjustment characteristics 13, 13' and fading characteristics 15, 15'.

[0048] exist Figure 3 In the block diagram, a hearing device 40 is schematically shown, which has an input converter 42, a signal processing unit 44 and an output converter 46. The input converter 42, which is provided by a microphone, generates an audio signal 50 from an ambient sound signal 48, in which a specific sound event 4 also enters. Now, in the signal processing unit 44, the audio signal 50 can be generated according to the sound signal 48 of the environment. Figure 1 Or according to Figure 2 In the manner described, the acoustic reverberation of the sound signal 4 is suppressed in the audio signal 50. The resulting signal is further processed, in particular subjected to dynamic compression and frequency-band-dependent amplification, thereby generating an output signal 52, which the output converter 46 converts into an output sound signal 54.

[0049] Although the present invention has been described in detail through a preferred embodiment, the present invention is not limited to this embodiment, and those skilled in the art may derive other variations therefrom without departing from the scope of protection of the present invention.

[0050] Reference Signs List

[0051] 1. 1' first horizontal measurement

[0052] 2. 2' Second level measurement

[0053] 4 Sound Events

[0054] 6, 6' first adjustment time

[0055] 8 Pre-assigned shares

[0056] 10 Progressive Levels

[0057] 12, 12' Second adjustment time

[0058] 13, 13' Adjustment characteristics

[0059] 14, 14' maximum

[0060] 15, 15' fade characteristics

[0061] 16, 16' first disappearance time

[0062] 18, 18' Second fade time

[0063] 20, 20' difference

[0064] 22 Weakened audio signal

[0065] 24 Noise Background

[0066] 26 Residual noise background

[0067] 40 Hearing devices

[0068] 42 Input Converter

[0069] 44 Signal Processing Unit

[0070] 46 Output Converter

[0071] 48 Sound Signals

[0072] 50 audio signals

[0073] 52 output signal

[0074] 54 Output sound signal

[0075] ΔT duration (time)

[0076] P-level value

[0077] T time

[0078] T0, T1, T2 time points

[0079] T60 decay time constant

Claims

1. A method for suppressing acoustic reverberation in an audio signal (50), in, providing an audio signal (50), Therein, a first level measurement (1') is performed on the audio signal (50), wherein, during the first level measurement (1'), a second level measurement (2') is performed on the audio signal (50), wherein the first level measurement (1') is performed with a first adjustment time (6') that is different from a second adjustment time (12') of the second level measurement (2'), wherein, based on the contribution of the sound event (4) in the audio signal (50), the sound reverberation of the sound event (4) in the audio signal (50) is suppressed by attenuating the audio signal (50); and wherein the attenuation is controlled according to a difference (20') between a first level measurement (1') and a second level measurement (2'), wherein a first level measurement (1') is performed with a first adjustment time (6') and a first decay time (16'), wherein a second level measurement (2') is performed having a second adjustment time (12') and a second decay time (18'), the second adjustment time being different from the first adjustment time (6') and the second decay time being the same as the first decay time (16'), The decay time of the level measurement is the time which elapses after a signal contribution in the audio signal and a corresponding level offset of the level measurement until the level measurement drops to zero or to a predetermined fraction of the level offset in the absence of further signal contributions in the audio signal. The settling time of the level measurement is the time which elapses after a spontaneously starting static signal contribution in the audio signal until the level measurement reaches a predetermined proportion of the asymptotic limit value of the signal level corresponding to the static signal contribution. wherein a sound event (4) is inferred from an increase in the difference (20') between the first level measurement (1') and the second level measurement (2'), and After the sound event (4) starts, the attenuation of the audio signal (50) is controlled by a monotonically decreasing function of the difference (20') between the first level measurement (1') and the second level measurement (2') to suppress the sound reverberation in the audio signal (50).

2. The method according to claim 1, in, The first horizontal measurement (1') is selected such that the first horizontal measurement (1') has an exponential decay characteristic (15').

3. The method according to claim 2, in, The first level measurement (1') is achieved by means of a first order recursive low pass filter.

4. The method according to claim 1 or 2, in, A physical decay time constant (T60) that produces an environment for the provided audio signal (50) is selected as a first decay time (16') of the first level measurement (1').

5. The method according to claim 1 or 2, in, determining a noise floor (24) of an audio signal (50), and The attenuation of the audio signal (50) is controlled according to the noise background (24) of the audio signal (50).

6. The method according to claim 5, in, performing noise suppression on the audio signal (50), and In this case, the attenuation of the audio signal (50) is additionally controlled as a function of noise suppression.

7. The method according to claim 1 or 2, in, decomposing the audio signal (50) into a plurality of frequency bands, wherein the first level measurement (1') and the second level measurement (2') are performed respectively according to the frequency band, wherein the audio signal (50) is attenuated according to frequency bands to suppress sound reverberation, and Therein, in the corresponding frequency band, the attenuation of the audio signal (50) is controlled as a function of the difference (20') between the first level measurement (1') and the second level measurement (2').

8. A method for suppressing acoustic reverberation in an audio signal (50) of a hearing device (40), in, providing an audio signal (50) based on an ambient sound signal (48) by means of an input transducer (42) of the hearing device (40), and Therein, in an audio signal (50), acoustic reverberation is suppressed by a method according to any one of the preceding claims.

9. A hearing device (40) having an input converter (42) for generating an audio signal (50) and a signal processing unit (44), the signal processing unit being configured to perform the method according to any one of claims 1 to 7.

Images