Hearing aid with adaptive noise canceller

Abstract

The invention discloses a hearing aid with an adaptive noise canceller. The hearing aid comprises: an input interface configured to provide a plurality of input audio signals; a first beamformer configured to determine a first beamformed signal; a second beamformer configured to determine a second beamformed signal; a first detector configured to determine and output a first detector output signal; a first noise canceller configured to receive a first noise canceller input signal based on the second beamforming signal and the first detector output signal, determine and output a first anti-noise signal; a second noise canceller configured to receive a second noise canceller input signal based on the second beamforming signal, determine and output a second anti-noise signal; and an output interface configured to determine an output signal based on the first beamformed signal, the first anti-noise signal, and the second anti-noise signal.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of hearing aids. BACKGROUND

[0002] Hearing aids can encounter more than two different types of noise, and it is desirable to provide a hearing aid that can better remove or attenuate different types of noise. SUMMARY

[0003] In one aspect of the application, a hearing aid is provided. The hearing aid comprises an input interface configured to provide a plurality of input audio signals. The hearing aid comprises a first beamformer configured to receive a first beamformer input signal based on the plurality of input audio signals, and determine a first beamformed signal based on the first beamformer input signal. The hearing aid comprises a second beamformer configured to receive a second beamformer input signal based on the plurality of input audio signals, and determine a second beamformed signal based on the second beamformer input signal. The hearing aid comprises a first detector configured to receive a first detector input signal based on the plurality of input audio signals, determine a first audio parameter based on the first detector input signal, determine a first detector output signal based on the determined first audio parameter. The first detector outputs the first detector output signal. The hearing aid comprises a first noise canceller configured to receive a first noise canceller input signal based on the second beamformed signal and the first detector output signal, determine a first anti-noise signal based on the first noise canceller input signal, and output the first anti-noise signal, wherein the first anti-noise signal is an anti-noise signal for a first type of noise. The hearing aid comprises a second noise canceller configured to receive a second noise canceller input signal based on the second beamformed signal, determine a second anti-noise signal based on the second beamformed signal, and output the second anti-noise signal, wherein the second anti-noise signal is an anti-noise signal for a second type of noise. The hearing aid comprises an output interface. The output interface is configured to determine an output signal based on the first beamformed signal, the first anti-noise signal, and the second anti-noise signal, and output the output signal.

[0004] Thus, an improved hearing aid is provided.

[0005] The input interface can comprise a plurality of input transducers. The input transducers can be a plurality of microphones for converting input sound into a plurality of electrical input signals. The input interface can convert the electrical input signals into the input audio signals. The input interface can comprise a wireless interface for receiving the plurality of input audio signals over a wireless link. The wireless interface can comprise a wireless receiver.

[0006] The plurality of input audio signals can represent sound. The plurality of input audio signals can represent near-end sound, i.e. sound in the environment surrounding the hearing aid. The plurality of input audio signals can represent far-end sound, i.e. sound received from another device in communication with the hearing aid. The plurality of input audio signals can be a plurality of time-frequency domain signals. The plurality of input audio signals can be a plurality of time domain signals.

[0007] The first beamformer can comprise a target distortionless beamformer. The target distortionless beamformer can comprise one or more of: a delay-and-sum beamformer, a delay-and-subtract beamformer, a minimum variance distortionless response beamformer, a minimum power distortionless response beamformer, and a linearly constrained minimum variance beamformer. The first beamformer can comprise a plurality of first beamformer weights. The first beamformer can be configured as an adaptive beamformer, such that the first beamformer weights can vary over time. The first beamformer can also be configured as a fixed beamformer, such that the first beamformer weights remain fixed over time.

[0008] The first beamformer can be configured to determine the first beamformed signal by applying the plurality of first beamformer weights to the first beamformer input signals. The first beamformer can be configured to determine the first beamformed signal by beamforming the first beamformer input signals. The first beamformer can be configured to determine the first beamformed signal by applying one or more beamforming algorithms (e.g. delay-and-sum beamformer, MVDR beamformer, GSC beamformer, differential beamformer, etc.).

[0009] The first beamformer input signals can be the plurality of input audio signals. The first beamformer input signals can be the plurality of modified input audio signals. The plurality of input audio signals can be modified by downsampling, upsampling, or other pre-processing.

[0010] The second beamformer can comprise a target cancelling beamformer. The target cancelling beamformer can comprise a delay-and-subtract beamformer configured to cancel a target signal (e.g. a desired speech signal). The second beamformer can comprise a plurality of second beamformer weights. The second beamformer can be configured as an adaptive beamformer, such that the second beamformer weights can vary over time. The second beamformer can also be configured as a fixed beamformer, such that the second beamformer weights remain fixed over time. The second beamformer can comprise a plurality of target cancelling beamformers, each target cancelling beamformer configured to cancel a target signal.

[0011] In the present disclosure, a target can be understood as a desired signal portion. For example, for telecommunication, the target can be the own voice. For a hearing aid, the target can be the speech for the hearing aid user to listen to.

[0012] The second beamformer input signal can be a plurality of input audio signals. Alternatively, the second beamformer input signal can be a plurality of modified input audio signals.

[0013] The second beamformer can be configured to determine the second beamformed signal by applying a plurality of second beamformer weights to the second beamformer input signal. The second beamformer can be configured to determine the second beamformed signal by beamforming the second beamformer input signal. The second beamformer can be configured to determine the second beamformed signal by applying one or more beamforming algorithms (e.g., delay-and-sum beamformer, MVDR beamformer, GSC beamformer, differential beamformer, etc.).

[0014] The first detector can be a voice activity detector configured to detect the presence of speech in the first detector input signal. The first detector can be a self voice detector configured to detect the speech of the hearing aid user in the first detector input signal. The first detector can be a target voice detector configured to detect the speech of a target speaker in the first detector input signal. The first detector can be a pure noise detector configured to detect the presence of pure noise (i.e., no speech) in the first detector input signal. The first detector can be a voice activity detector. The first detector can be a self voice activity detector.

[0015] The first detector input signal can be the first beamformer input signal or the second beamformer input signal. The first detector input signal can be a plurality of input audio signals. The first detector input signal can be based on a plurality of modified input audio signals. An input audio signal of a reference microphone of the hearing aid can be used as the first detector input signal. For example, the reference microphone of the hearing aid can be a forward facing microphone of a behind-the-ear hearing aid. For example, the reference microphone of the hearing aid can be a rearward facing microphone of a behind-the-ear hearing aid.

[0016] The first audio parameter can be a noise level or a signal-to-noise ratio of the first detector input signal. The first audio parameter can be a parameter indicative of the presence or absence of speech in the first detector input signal. The first audio parameter can be a speech presence probability value. The first audio parameter can be a voice activity parameter configured to indicate the presence of speech in the first detector input signal. The first audio parameter can be a pure noise parameter configured to indicate that only noise (i.e., no speech) is present in the detector input beamformer input signal. The first audio parameter can be a pure noise presence probability value.

[0017] The first audio parameter can comprise an energy of the first detector input signal. For example, the first audio parameter can be a power value of the first detector input signal. For example, the first audio parameter can be a plurality of power values of the first detector input signal. The power value can comprise a time-smoothed value of a magnitude of the first detector input signal, wherein the magnitude can comprise an absolute value, an absolute squared value, and a logarithmic value of any of the above. The time-smoothing can comprise a first-order recursive filter with a time constant. Alternatively, the time-smoothing can comprise a first-order recursive filter with an attack time constant and a release time constant.

[0018] The first audio parameter can comprise a plurality of power values of the first detector input signal. Each power value can be determined using a recursive filter comprising an attack time constant and a release time constant.

[0019] The first audio parameter can comprise a likelihood value based on a likelihood function. The likelihood function can be based on a probability model, such as a Gaussian distribution. The likelihood value can be determined based on the first detector input signal and the likelihood function. The first audio parameter can comprise a likelihood ratio or a log-likelihood ratio. The likelihood ratio can be based on a plurality of likelihood values. For example, the likelihood ratio can be based on a first likelihood value under a speech-absent hypothesis and a second likelihood value under a speech-absent hypothesis.

[0020] The first audio parameter can comprise a binary value. The binary value can indicate a presence of speech in the first detector input signal. For example, a value of "1" can indicate a presence of speech, and a value of "0" can indicate an absence of speech. The binary value can indicate a presence of pure noise. For example, a value of "1" can indicate a presence of pure noise, and a value of "0" can indicate an absence of pure noise.

[0021] The first audio parameter can comprise a probability value. The probability value can indicate a likelihood of a presence of speech in the first detector input signal. For example, a probability value closer to "1" can be more indicative of a presence of speech in the first detector input signal than a probability value closer to "0". The binary value can indicate a presence of pure noise. For example, a probability value closer to "1" can be more indicative of a presence of pure noise in the first detector input signal than a probability value closer to "0".

[0022] The first detector output signal can comprise the first audio parameter. The first detector output signal can comprise a processed version of the first audio parameter.

[0023] The first noise canceller can comprise a complex number for modifying an amplitude and a phase of the first noise canceller input signal. The amplitude and the phase of the first noise canceller input signal can be modified by a complex conjugate product of the first noise canceller input signal and the complex number of the first noise canceller. The complex number of the first noise canceller can be determined based on minimizing a squared error between the first beamformed signal and the first noise canceller output signal.

[0024] To minimize the error, the first noise canceller can include an iterative solver, such as a gradient descent algorithm. The gradient descent algorithm can be a least mean square algorithm, a normalized least mean square algorithm, or a sign-sign least mean square algorithm. Alternatively, a closed-form solution can be applied to minimize the error.

[0025] The first noise canceller input signal can be the second beamformer output signal. The first noise canceller input signal can be a modified second beamformer output signal, where the second beamformer output signal has been modified by filtering, amplification, or other pre-processing.

[0026] The first anti-noise signal can be an output of the first noise canceller and determined by a complex conjugate product of the first noise canceller input signal and a complex number of the first noise canceller. The first anti-noise signal can be indicative of a first type of noise in the plurality of input audio signals. The first anti-noise signal can be a noise signal indicative of the first type of noise in the plurality of input audio signals.

[0027] In this specification, an anti-noise signal can be understood as a signal that, when combined with an original signal from which the anti-noise signal is generated, aims to remove noise from the original signal or a first beamformed signal. The anti-noise signal can be a noise signal that is intended to be subtracted from the original signal or the first beamformed signal so as to remove noise.

[0028] The first type of noise can include, but is not limited to, ambient noise, reverberation, microphone self-noise, wind noise, machine noise, colored noise (e.g., sound produced by an electric fan, a vacuum cleaner, noise inside a car cabin). The first type of noise is characterized by its power spectral density (or periodogram) changing slowly over time compared to the power spectral density (or periodogram) of speech.

[0029] The second noise canceller can include a complex number for modifying an amplitude and a phase of the second noise canceller input signal. The amplitude and the phase of the second noise canceller input signal can be modified by a complex conjugate product between the second noise canceller input signal and the complex number of the second noise canceller. The complex number of the second noise canceller can be determined based on minimizing a squared error between the first beamformer output signal and a second noise canceller output signal.

[0030] The second noise canceller can include an iterative solver, such as a gradient descent algorithm.

[0031] The second noise canceller input signal can be the second beamformer output signal. The second noise canceller input signal can be a modified second beamformer output signal, where the second beamformer output signal can be modified by filtering, amplification, or other processing.

[0032] The second anti-noise signal can be an output of the second noise canceller and determined by a complex conjugate multiplication between the second noise canceller input signal and a complex of the second noise canceller. The second anti-noise signal can be indicative of a second type of noise in the plurality of input audio signals. The second anti-noise signal can be a noise signal indicative of the second type of noise in the plurality of input audio signals.

[0033] The second type of noise can include, but is not limited to, (unwanted) speech, transient noise, reverberation. The second type of noise is characterized by a degree of change in its power spectral density (or periodicity) over time that is the same or greater compared to the power spectral density (or periodicity) of speech.

[0034] The output interface can be configured to provide, based on the output signal, a stimulus perceived by the user as an acoustic signal. The output interface unit can be a vibrator of a bone conduction hearing aid. The output interface unit can comprise an output interface transducer. The output interface transducer can comprise a receiver (loudspeaker) for providing the stimulus to the user as a sound signal (e.g. in an (air conduction based) acoustic hearing aid). The output interface transducer can comprise a vibrator for providing the stimulus to the user as a mechanical vibration of the skull (e.g. in a bone-anchored or bone-attached hearing aid). The output interface unit can (additionally or alternatively) comprise a (e.g. wireless) transmitter for transmitting the hearing aid picked-up sound to another device, e.g. a remote communication partner (e.g. over a network, e.g. in a telephony operating mode).

[0035] The output signal can be based on a linear combination of the first beamformed signal, the first anti-noise signal, and the second anti-noise signal. The output signal can be a result of subtracting the first anti-noise signal and the second anti-noise signal from the first beamformed signal.

[0036] A first advantage of the present invention is a better attenuation effect on the first type of noise and the second type of noise. A second advantage is an improved audio quality of the target signal. A third advantage of the present invention is the use of a first noise canceller and a second noise canceller, wherein the first noise canceller determines the first anti-noise signal and the second noise canceller determines the second anti-noise signal. Thus, compared to a hearing aid using a single noise canceller, the present invention provides a hearing aid that is able to better remove or attenuate both types of noise.

[0037] A benefit of using a multi-stage noise cancellation structure is a higher flexibility and a better control of the performance of the noise reduction system. Each noise canceller can set its own parameters. The detector is responsible for ensuring that the noise canceller only updates and adapts to the unwanted noise source based on the detection result of the detector of the presence of the unwanted noise.

[0038] In an embodiment, the interface comprises two or more microphones configured to provide the plurality of input audio signals.

[0039] The advantage of using two or more microphones is a better attenuation effect on the first and second type of noise.

[0040] In an embodiment, the first beamformer is a fixed filter, i.e. the first beamformer weights do not change over time. The second beamformer is a fixed filter, i.e. the second beamformer weights do not change over time. The first noise canceller is an adaptive filter. The second noise canceller is an adaptive filter.

[0041] One advantage of using a fixed filter for the first and second beamformer is a reduced computational complexity and increased robustness in complex sound environments.

[0042] The first type of noise can be a slow time-varying noise. The second type of noise can be a fast time-varying noise.

[0043] The slow time-varying noise can be one or more of the following: environmental noise, reverberation, microphone self-noise, wind noise, machine noise, and colored noise (e.g. sound produced by an electric fan, a vacuum cleaner, noise inside a car cabin). The slow time-varying noise is characterized by a slow change in its power spectral density (or periodogram) over time compared to the power spectral density (or periodogram) of speech.

[0044] The fast time-varying noise can be one or more of the following: unwanted speech, transient noise, acoustic feedback, reverberation, and echo. The fast time-varying noise is characterized by a same or greater degree of change in its power spectral density (or periodogram) over time compared to the power spectral density (or periodogram) of speech.

[0045] Hence, one advantage of the present invention is to propose a hearing aid that is able to better remove or attenuate slow time-varying noise and fast time-varying noise.

[0046] In an embodiment, the first noise canceller comprises a first smoothing factor. The second noise canceller comprises a second smoothing factor. The smoothing factor defines the adaptation rate of the associated noise canceller. The first smoothing factor results in a lower adaptation rate than the second smoothing factor.

[0047] The first smoothing factor can be used to determine the adaptation rate of the first noise canceller. A smaller first smoothing factor can be used to configure the first noise canceller to adapt to slow time-varying noise compared to a larger first smoothing factor. The first smoothing factor can be used in an iterative solver such as a gradient descent algorithm.

[0048] The second smoothing factor can be used to determine the adaptation rate of the second noise canceller. A larger second smoothing factor can be used to configure the second noise canceller to adapt to fast time-varying noise compared to a smaller first smoothing factor. The second smoothing factor can be used in an iterative solver such as a gradient descent algorithm.

[0049] In the present specification, a smoothing factor can be understood as a factor used to determine a weight given to a latest estimate compared to a previous estimate.

[0050] Thus, it is an advantage of the present invention that the hearing aid comprises a first smoothing factor for better attenuating slow time-varying noise and a second smoothing factor for better attenuating fast time-varying noise.

[0051] In an embodiment, determining the first detector output signal comprises comparing the first audio parameter to a first threshold value. If the first audio parameter exceeds the first threshold value, determining the first detector output signal comprises an activation signal. The activation signal allows the first noise canceller to adapt by reducing an error between the first anti-noise signal and a first target signal based on the first beamformed signal. If the first audio parameter does not exceed the first threshold value, determining the first detector output signal comprises a fixation signal. The fixation signal causes the first noise canceller to remain fixed.

[0052] The first threshold value can be a predefined value. The first threshold value can be a predefined value between "0" and "1". For example, the first threshold value can be 0.5. For example, the first threshold value can be selected in the range of 0.1 to 0.9.

[0053] The first threshold value can be an adaptive value. For example, the first threshold value can be determined based on a noise level or a signal-to-noise ratio. The noise level or the signal-to-noise ratio can be determined by the first detector, e.g. as the first audio parameter. A higher noise level can cause the first threshold value to be more biased towards "1" compared to a lower noise level, whereas a lower noise level can cause the first threshold value to be more biased towards "0". A lower signal-to-noise ratio can cause the first threshold value to be more biased towards "1" compared to a higher signal-to-noise ratio, whereas a higher signal-to-noise ratio can cause the first threshold value to be more biased towards "0".

[0054] In the present invention, allowing the noise canceller to adapt can be understood as changing one or more parameters / coefficients comprised by the noise canceller.

[0055] In the present invention, fixing the noise canceller can be understood as keeping one or more parameters / coefficients comprised by the noise canceller fixed, i.e. not changing these parameters / coefficients.

[0056] The activation signal can be a binary value, where a value of "1" allows the first noise canceller to adapt. The activation signal can be based on the first detector output signal. The activation signal can be a value between "0" and "1", excluding "0".

[0057] The first target signal can be a first beamformed signal. The first target signal can be a processed version of the first beamformed signal. The first target signal can comprise speech signals from a talker plus noise. The first target signal can be a linear combination of the first beamformed signal and the first anti-noise signal. The first target signal can be a signal based on sound emitted by a multimedia device such as a television, a telephone, a sound system, etc. The first target signal can be sound from a live entertainment performance.

[0058] The error between the first target signal and the first anti-noise signal can be computed as a difference between the two. The error can be computed as an absolute value (or an absolute squared value) of the difference. The error can be computed as a logarithm of an absolute value (or an absolute squared value) of the difference.

[0059] Reducing the error between the first target signal and the first anti-noise signal can be achieved by reducing an average difference between the first target signal and the first anti-noise signal. Reducing the error between the first target signal and the first anti-noise signal can be achieved by reducing an average absolute value or an average absolute squared value of the difference between the first target signal and the first anti-noise signal. Reducing the error between the first target signal and the first anti-noise signal can comprise using an iterative solver to minimize a mean squared error between the first target signal and the first anti-noise signal, e.g. a gradient algorithm. Reducing the error between the first target signal and the first anti-noise signal can comprise using a closed-form equation based on minimizing a mean squared error between the first target signal and the first anti-noise signal.

[0060] The fix signal can be a binary value, where a value of "0" stops the adaptive adjustment of the first noise canceller, i.e. fixes the first noise canceller. The fix signal can be based on the first detector output signal.

[0061] Thus, the disclosed hearing aid comprising the first noise canceller has the advantage of being controlled by the first detector output signal. In particular, it has the advantage of being able to adapt to the first type of noise, thereby reducing the noise and enhancing the desired speech signal better.

[0062] In an embodiment, the hearing aid can comprise a second detector. The second detector is configured to receive a second detector input signal based on the plurality of input audio signals. The second detector determines a second audio parameter based on the second detector input signal. The second detector determines a second detector output signal based on the determined second audio parameter. The second detector outputs the second detector output signal. The second noise canceller is configured to receive the second detector output signal. The second noise canceller is configured to determine a second anti-noise signal based on the second detector output signal.

[0063] The second detector can be an interference detector configured to detect a presence of interference in the second detector input signal. The interference detector can be based on determining a modulation in the first beamformer input signal. For example, the modulation can be based on determining a change in energy level in the second detector input signal. The interference detector can be based on determining a spatial feature in the second detector input signal. For example, the spatial feature in the second detector input signal can be used to determine a direction or a location of a sound source in the second detector input signal. The spatial feature can be based on determining a correlation between the plurality of input audio signals. The second detector can comprise a trained neural network configured to detect interference in the second detector input signal.

[0064] In the present disclosure, interference can be understood as a highly modulated noise. The interference can be understood as a noise with transient properties. For example, the interference can be speech from an unwanted sound source.

[0065] The second detector input signal can be the first or the second beamformer input signal. The second detector input signal can be the plurality of input audio signals or the plurality of modified input audio signals. The input audio signal of the hearing aid reference microphone can be used as the second detector input signal.

[0066] The second audio parameter can be an interference parameter indicative of interference in the second detector input signal. The second audio parameter can be a noise level or a signal-to-noise ratio of the second detector input signal. The second audio parameter can be a parameter indicative of a presence or absence of interference in the second detector input signal.

[0067] The second audio parameter can comprise an energy of the second detector input signal. For example, the second audio parameter can be a power value of the second detector input signal. For example, the second audio parameter can be a plurality of power values of the second detector input signal. The power value can comprise a time-smoothed value of an amplitude of the second detector input signal, wherein the amplitude can comprise an absolute value, an absolute squared value, and a logarithmic value of any of the above. The time-smoothing can comprise a first order recursive filter with a time constant. Alternatively, the time-smoothing can comprise a first order recursive filter with an attack time constant and a release time constant.

[0068] The second audio parameter can comprise a plurality of power values of the second detector input signal. Each power value can be determined using a recursive filter comprising an attack time constant and a release time constant.

[0069] The second audio parameter can comprise a correlation-based value. For example, the correlation-based value can comprise a cross-correlation between the input audio signals.

[0070] The second audio parameter can comprise a likelihood value based on a likelihood function. The likelihood function can be based on a probability model, e.g. a Gaussian distribution. The likelihood value can be determined based on the second detector input signal and the likelihood function. The second audio parameter can comprise a likelihood ratio or a log-likelihood ratio. The likelihood ratio can be based on a plurality of likelihood values. For example, the likelihood ratio can be based on a first likelihood value under the assumption that no interference is present and a second likelihood value under the assumption that no interference is present. Alternatively, the likelihood ratio can be based on a first likelihood value under the assumption that interference is present at the first spatial position and a second likelihood value under the assumption that no interference is present at the first spatial position.

[0071] The second audio parameter can comprise a binary value. The binary value can be indicative of the presence of interference in the second detector input signal. For example, a value of "1" can be indicative of the presence of interference and a value of "0" can be indicative of the absence of interference.

[0072] The second audio parameter can comprise a probability value. The probability value can be indicative of the likelihood of the presence of interference in the second detector input signal. For example, a probability value close to "1" is more indicative of the presence of interference in the second detector input signal than a probability value close to "0".

[0073] The second detector output signal can comprise the second audio parameter. The first detector output signal can comprise a processed version of the second audio parameter.

[0074] Thus, the disclosed hearing aid comprising a second noise canceller has the advantage of being controlled by the second detector output signal. In particular, it has the advantage of being able to adapt to the second type of noise, thereby better reducing the noise and enhancing the desired speech signal.

[0075] In an embodiment, determining the second detector output signal comprises comparing the second audio parameter to a second threshold value. If the second audio parameter exceeds the second threshold value, determining the second detector output signal comprises an activation signal. The activation signal allows the second noise canceller to adapt by reducing an error between a second target signal based on the second beamformed signal and a second anti-noise signal. If the second audio parameter does not exceed the second threshold value, determining the second detector output signal comprises a fixation signal. The fixation signal causes the second noise canceller to remain in a fixed state.

[0076] The second threshold value can be a predefined value. The second threshold value can be a predefined value between "0" and "1". For example, the second threshold value can be 0.5. For example, the second threshold value can be selected in the range of 0.1 to 0.9.

[0077] The second threshold can be an adaptive value. For example, the second threshold can be determined based on a noise level or a signal-to-noise ratio. The noise level or the signal-to-noise ratio can be determined by the second detector, e.g., as the second audio parameter. A higher noise level can cause the second threshold to be more biased towards "1" than a lower noise level. A lower signal-to-noise ratio can cause the second threshold to be more biased towards "1" than a higher signal-to-noise ratio. A higher signal-to-noise ratio can cause the second threshold to be more biased towards "0" than a lower signal-to-noise ratio.

[0078] The activation signal can be a binary value, where a value of "1" allows the second noise canceller to adapt. The activation signal can be based on the second detector output signal. Alternatively, the activation signal can be a value between "0" and "1", excluding "0".

[0079] The error between the second target signal and the second anti-noise signal can be computed as a difference between the two. The error can be computed as an absolute value (or an absolute squared value) of the difference. The error can be computed as a logarithm of an absolute value (or an absolute squared value) of the difference.

[0080] Reducing the error can be reducing an average (or mean) difference between the second target signal and the second anti-noise signal. Reducing the error can be reducing an average (or mean) absolute value (or absolute squared value) of a difference between the second target signal and the second anti-noise signal. Reducing the error can include using an iterative solver such as a gradient algorithm. Reducing the error can include using a closed-form equation based on minimizing a mean squared error between the second target signal and the second anti-noise signal.

[0081] The second target signal can be a second beamformed signal. The second target signal can include a speech signal from a talker plus noise. The second target signal can be a linear combination of the second beamformed signal and the second anti-noise signal. The second target signal can be a signal based on sound emitted by a multimedia device such as a television, a telephone, a sound system, etc. The first target signal can be sound from a live entertainment performance.

[0082] The fix signal can be a binary value, where a value of "0" stops the adaptation of the second noise canceller, i.e., fixes the second noise canceller. The fix signal can be based on the second detector output signal.

[0083] In an embodiment, the second audio parameter is an interference activity parameter.

[0084] The interference activity parameter can include a correlation-based value. For example, the correlation-based value can include a cross-correlation between the input audio signals.

[0085] The interference activity parameter can comprise a likelihood value based on a likelihood function. The likelihood function can be based on a probability model, e.g. a Gaussian distribution. The likelihood value can be determined based on the second detector input signal and the likelihood function. The interference activity parameter can comprise a likelihood ratio or a log-likelihood ratio. The likelihood ratio can be based on a plurality of likelihood values. For example, the likelihood ratio can be based on a first likelihood value under an assumption of absence of interference and a second likelihood value under an assumption of absence of interference. Alternatively, the likelihood ratio can be based on a first likelihood value under an assumption of interference from a first spatial location and a second likelihood value under an assumption of interference from a second spatial location.

[0086] Interference can be understood as unwanted speech, e.g. speech from a person next to the hearing aid user.

[0087] In an embodiment, the first audio parameter is a voice activity parameter.

[0088] The voice activity parameter can comprise an energy of the first detector input signal. For example, the voice activity parameter can be a power value of the first detector input signal. For example, the voice activity parameter can be a plurality of power values of the first detector input signal. The power value can comprise a time-smoothed value of an amplitude of the first detector input signal, wherein the amplitude can comprise an absolute value, an absolute squared value, and a logarithmic value of any of the above. The time-smoothing can comprise a first order recursive filter with a time constant. The time-smoothing can comprise a first order recursive filter with an attack time constant and a release time constant.

[0089] The voice activity parameter can comprise a plurality of power values of the first detector input signal. Each power value can be determined using a recursive filter comprising an attack time constant and a release time constant.

[0090] The voice activity parameter can comprise a likelihood value based on a likelihood function. The likelihood function can be based on a probability model, e.g. a Gaussian distribution. The likelihood value can be determined based on the first detector input signal and the likelihood function. The voice activity parameter can comprise a likelihood ratio or a log-likelihood ratio. The likelihood ratio can be based on a plurality of likelihood values. For example, the likelihood ratio can be based on a first likelihood value under an assumption of absence of speech and a second likelihood value under an assumption of absence of speech.

[0091] In an embodiment, the hearing aid comprises a first gain controller. The first gain controller is configured to receive the first anti-noise signal. The first gain controller determines a first control gain. The first gain controller applies the first control gain to the first anti-noise signal. The first gain controller outputs a first noise-controlled signal. The output interface is configured to determine the output signal based on the first noise-controlled signal.

[0092] The control gain can be used to control the strength of the associated noise canceller.

[0093] The first control gain may comprise a real value between 0 and 1. The first noise-controlled signal may be a modified first anti-noise signal. The modified first anti-noise signal may be the result of applying the first control gain to the first anti-noise signal. The first control gain may be a global gain configured to apply to the entire signal spectrum. The first control gain may also be a frequency-specific gain configured to apply to a portion of the signal spectrum.

[0094] In an embodiment, the first control gain is a predetermined gain.

[0095] The first control gain may be a value between "0" and "1". The first control gain may be set as a factory setting. The first control gain may be set by an audiologist during a calibration procedure or a fitting procedure of the hearing aid.

[0096] In an embodiment, the first gain controller is configured to receive the first detector output signal.The first gain controller is configured to determine a first control gain based on the first detector output signal.

[0097] The first control gain may be determined as a first audio parameter. That is, if the first audio parameter is a value between 0 and 1, the first control gain may be determined as a value corresponding to the first audio parameter. The first control gain may be determined as a function of the first audio parameter. The function may be determined through empirical testing.

[0098] In an embodiment, the hearing aid includes a second gain controller. The second gain controller is configured to receive a second anti-noise signal, determine a second control gain, and apply the second control gain to the second anti-noise signal. The second gain controller outputs a second noise-controlled signal. The output interface is configured to determine an output signal based on the second noise-controlled signal.

[0099] The second control gain may comprise a real value between 0 and 1. The second noise-controlled signal may be a modified second anti-noise signal. The modified second anti-noise signal may be the result of applying the second control gain to the second anti-noise signal. The second control gain may be a global gain configured to apply to the entire signal spectrum. The second control gain may also be a frequency-specific gain configured to apply to a portion of the signal spectrum.

[0100] In an embodiment, the second control gain is a predetermined gain.

[0101] The second control gain may be a value between "0" and "1." The second control gain may be set to a factory setting. The second control gain may be set by an audiologist during a calibration procedure or a fitting procedure of the hearing aid.

[0102] In an embodiment, the second gain controller is configured to receive a second detector output signal. The second gain controller is configured to determine a second control gain based on the second detector output signal.

[0103] The second control gain can be determined as the second audio parameter, i.e. if the second audio parameter is a value between 0 and 1, the second control gain can be determined as a value corresponding to the second audio parameter. The second control gain can be determined as a function of the second audio parameter. The function can be determined by empirical tests.

[0104] In an embodiment, the hearing aid comprises a third detector. The third detector is configured to receive a third detector input signal based on the plurality of input audio signals. The third detector determines a third audio parameter based on the third detector input signal. The third detector determines a third detector output signal based on the determined third audio parameter. The third detector outputs the third detector output signal. The hearing aid comprises a third noise canceller. The third noise canceller is configured to receive a third noise canceller input signal based on the second beamformed signal and the third detector output signal. The third noise canceller determines a third anti-noise signal based on the third noise canceller input signal, wherein the third anti-noise signal is an anti-noise signal for a third type of noise. The third noise canceller outputs the third anti-noise signal. The output interface is configured to determine the output signal based on the third anti-noise signal.

[0105] The third detector can be a feedback detector (or an echo detector) configured to detect a presence of feedback or echo in the third detector input signal. The feedback detector can be based on determining a presence of howling in the third detector input signal. For example, the presence of howling can be determined based on a spectral analysis, e.g. by determining a power ratio between adjacent sub-bands. The feedback detector can be based on determining a presence of a specific location sound that can generate feedback. For example, the specific location of the feedback can be determined using an acoustic path between a microphone and a receiver of the hearing aid.

[0106] The third detector input signal can be the first or second beamformer input signal. The third detector input signal can be the plurality of input audio signals. An input audio signal of a reference microphone of the hearing aid can be used as the third detector input signal. For example, the reference microphone of the hearing aid can be a forward facing microphone of a behind-the-ear hearing aid.

[0107] The third audio parameter can be a feedback parameter (or an echo parameter) configured to indicate a presence of feedback or echo in the third detector input signal.

[0108] The third audio parameter can comprise a spectral energy of the third detector input signal. The third audio parameter can be a spectral ratio of spectral energies between adjacent subbands. For example, the third audio parameter can be a power value of the third detector input signal. For example, the third audio parameter can be a plurality of power values or spectral ratios of the third detector input signal. The power value can comprise a result of a time-smoothing process on a magnitude of the third detector input signal, wherein the magnitude can comprise an absolute value, an absolute square value, and a logarithm value of any of the above. The time-smoothing process can comprise a first-order recursive filter with a time constant. Alternatively, the time-smoothing process can comprise a first-order recursive filter with an attack time constant and a release time constant.

[0109] The third audio parameter can comprise a plurality of power values of the third detector input signal. Each power value can be determined using a recursive filter comprising an attack time constant and a release time constant.

[0110] The third audio parameter can comprise a correlation-based value. For example, the correlation-based value can comprise a cross-correlation between a plurality of input audio signals.

[0111] The third audio parameter can comprise a likelihood value based on a likelihood function. The likelihood function can be based on a probability model, such as a Gaussian distribution. The likelihood value can be determined based on the third detector input signal and the likelihood function. The third audio parameter can comprise a likelihood ratio or a log-likelihood ratio. The likelihood ratio can be based on a plurality of likelihood values. For example, the likelihood ratio can be based on a first likelihood value under a feedback-absent hypothesis and a second likelihood value under the feedback-absent hypothesis. Alternatively, the likelihood ratio can be based on a first likelihood value under a feedback-present-in-first-spatial-location hypothesis and a second likelihood value under a feedback-absent-in-first-spatial-location hypothesis.

[0112] The third detector output signal can comprise a binary value. The binary value can indicate a presence of feedback in the third detector input signal. For example, a value of "1" can represent a presence of feedback, and a value of "0" can represent an absence of feedback.

[0113] The third detector output signal can comprise a probability value. The probability value can indicate a likelihood of a presence of feedback in the third detector input signal. For example, a probability value closer to "1" can indicate a greater likelihood of a presence of feedback in the third detector input signal than a probability value closer to "0".

[0114] The third noise canceller can comprise a complex number for modifying a magnitude and a phase of the third noise canceller input signal. The magnitude and the phase of the third noise canceller input signal can be modified by a complex conjugate multiplication between the third noise canceller input signal and the complex number of the third noise canceller. The complex number of the third noise canceller can be determined based on minimizing a squared error between the first beamformer output signal and the third noise canceller input signal.

[0115] The third noise canceller can comprise an iterative solver, such as a gradient descent algorithm.

[0116] The third noise canceller input signal can be the second beamformer output signal. Alternatively, the third noise canceller input signal is a modified third beamformer output signal, wherein the third beamformer output signal can be modified by filtering or amplification.

[0117] The third anti-noise signal can be an output of the third noise canceller and determined by a complex conjugate multiplication between the third noise canceller input signal and a complex number of the third noise canceller.

[0118] The third type of noise can comprise acoustic feedback. The third type of noise can be characterized by its howling property. The howling can be detected based on a spectral howling detection algorithm.

[0119] In an embodiment, the third audio parameter is an audio feedback parameter.

[0120] The audio feedback parameter can comprise an energy of the third detector input signal. For example, the audio feedback parameter can be a power value of the third detector input signal. For example, the audio feedback parameter can be a plurality of power values of the third detector input signal. The power value can comprise a time-smoothed result of a magnitude of the third detector input signal, wherein the magnitude can comprise an absolute value, an absolute squared value, and a logarithmic value of any of the above. The time-smoothed processing can comprise a first order recursive filter with a time constant. Alternatively, the time-smoothed processing can comprise a first order recursive filter with an attack time constant and a release time constant.

[0121] The audio feedback parameter can comprise a plurality of power values of the third detector input signal. Each power value can be determined using a recursive filter comprising an attack time constant and a release time constant.

[0122] The audio feedback parameter can comprise a likelihood value based on a likelihood function. The likelihood function can be based on a probability model, such as a Gaussian distribution. The likelihood value can be determined based on the third detector input signal and the likelihood function. The audio feedback parameter can comprise a likelihood ratio or a log-likelihood ratio. The likelihood ratio can be based on a plurality of likelihood values. For example, the likelihood ratio can be based on a first likelihood value under a feedback-absent hypothesis and a second likelihood value under a feedback-absent hypothesis.

[0123] In an embodiment, the hearing aid comprises a third gain controller configured to receive the third anti-noise signal, determine a third control gain, apply the third control gain to the third anti-noise signal, and output a third noise-controlled signal. The output interface is configured to determine the output signal based on the third noise-controlled signal.

[0124] The third control gain can comprise a real value between "0" and "1". The third noise-controlled signal can be a modified third anti-noise signal. The modified third anti-noise signal can be a result of applying the third control gain to the third anti-noise signal. The third control gain can be a global gain configured to be applied to a full frequency spectrum of a signal. The third control gain can also be a frequency-specific gain configured to be applied to a frequency band of a frequency spectrum of a signal.

[0125] In an embodiment, the third gain controller is configured to receive a third detector output signal. The third gain controller is configured to determine the third control gain based on the third detector output signal.

[0126] The adaptive gain can be determined based on the third detector output signal. The third detector output signal can be a value between "0" and "1". The third detector output signal can be a binary value. The third detector output signal can be a probability value indicating a presence or absence of the third type of noise.

[0127] In an embodiment, the third control gain is a predetermined gain. The predetermined gain can be a value between "0" and "1".

[0128] In an embodiment, the first beamformer is a beamformer with a distortionless criterion.

[0129] The distortionless criterion can be a linear constraint. The distortionless criterion can be based on a target sound source. The linear constraint can be a linear combination of relative transfer functions (i.e., steering vectors) of the target sound source using beamformer weights of the first beamformer such that an output of the linear combination equals a predetermined value (e.g., "1"). The target sound source can be a desired speaker located in front of a hearing aid user. The hearing aid user wears a hearing aid, such as a behind-the-ear hearing aid. The target sound source can be a self voice of the hearing aid user.

[0130] The beamformer with the distortionless criterion can be any beamformer satisfying the distortionless criterion.

[0131] In an embodiment, the second beamformer is a target cancelling beamformer.

[0132] The second beamformer can determine the second beamformed signal based on minimizing a power of a target sound signal based on the plurality of input audio signals.

[0133] In an embodiment, the hearing aid can be an air-conduction type hearing aid.

[0134] In an embodiment, the hearing aid can be a bone- conduction type hearing aid.

[0135] In an embodiment, determining the first anti-noise signal can include determining a correlation between the first beamformed signal and the first noise canceller input signal.

[0136] In an embodiment, determining the first noise canceller includes normalizing the correlation using a magnitude of the first noise canceller input signal. The magnitude can include an amplitude value, an absolute squared value, or a logarithm of the above values.

[0137] In an embodiment, determining the first noise canceller includes determining a correlation between the first anti-noise signal and the first noise canceller input signal.

[0138] In an embodiment, determining the first noise canceller includes reducing a magnitude error between the first beamformed signal and the first noise canceller input signal.

[0139] In an embodiment, determining the second noise canceller includes determining a correlation between the first anti-noise signal and the first noise canceller input signal.

[0140] In an embodiment, determining the first noise canceller includes normalizing the correlation using a magnitude of the first noise canceller input signal.

[0141] In an embodiment, determining the second noise canceller includes determining a correlation between the second anti-noise signal and the first noise canceller input signal.

[0142] In an embodiment, determining the first noise canceller includes reducing a magnitude error between the second anti-noise signal and the first noise canceller input signal.

[0143] In an embodiment, determining the second anti-noise signal includes subtracting the second anti-noise signal from the first anti-noise signal.

[0144] In an embodiment, determining the third noise canceller includes determining a correlation between the second anti-noise signal and the first noise canceller input signal.

[0145] In an embodiment, determining the third noise canceller includes normalizing the correlation using a magnitude of the first noise canceller input signal.

[0146] In an embodiment, determining the third noise canceller includes determining a correlation between the third anti-noise signal and the first noise canceller input signal.

[0147] In an embodiment, determining the third noise canceller includes reducing a magnitude error between the third anti-noise signal and the first noise canceller input signal.

[0148] In an embodiment, determining the third anti-noise signal includes subtracting the third anti-noise signal from the second anti-noise signal.

[0149] According to a second aspect of the application, there is provided a method of operating a hearing aid, the method comprising providing a plurality of input audio signals. The method comprises determining a first beamformed signal based on the plurality of input audio signals. The method comprises determining a second beamformed signal based on the plurality of input audio signals. The method comprises determining a first audio parameter based on the plurality of input audio signals. The method comprises determining a first detector output signal based on the determined first audio parameter. The method comprises determining a first anti-noise signal related to a first type of noise based on the second beamformed signal and the first detector signal. The method comprises determining a second anti-noise signal related to a second type of noise based on the second beamformed signal. The method comprises determining an output signal based on the first beamformed signal, the first anti-noise signal and the second anti-noise signal. The method comprises outputting the aforementioned output signal.

[0150] The hearing aid can be adapted to provide frequency-dependent gain and / or level-dependent compression and / or frequency translation of one or more frequency ranges to one or more other frequency ranges (with or without frequency compression) to compensate for the hearing impairment of the user. The hearing aid can comprise a signal processor for enhancing an input signal and providing a processed output signal.

[0151] The wireless receiver and / or transmitter can for example be configured to receive and / or transmit electromagnetic signals in the radio frequency range (3 kHz to 300 GHz). The wireless receiver and / or transmitter can for example be configured to receive and / or transmit electromagnetic signals in the optical frequency range (e.g. infrared light 300 GHz to 430 THz or visible light such as 430 THz to 770 THz).

[0152] The hearing aid can comprise a directional microphone system adapted to spatially filter sound from the environment so as to enhance a target sound source among a plurality of sound sources in the local environment of a user wearing the hearing aid. The directional system can be adapted to detect (e.g. adaptively detect) from which direction a particular portion of the microphone signal originates. This can be achieved in a number of different ways as described in the prior art. In hearing aids, microphone array beamformers are often used to spatially attenuate background noise sources. The beamformer can comprise a linearly constrained minimum variance (LCMV) beamformer. Many beamformer variants can be found in the literature. The minimum variance distortionless response (MVDR) beamformer is widely used in microphone array signal processing. Ideally, the MVDR beamformer leaves the signal from a target direction (also called look direction) unchanged while maximally attenuating sound signals from other directions. The generalized sidelobe canceller (GSC) structure is an equivalent representation of the MVDR beamformer which offers computational and digital representation advantages compared to the direct implementation of the original form.

[0153] Most sound signal sources (except the user's own voice) are located at a distance d compared to the size of the hearing aid, e.g. the distance between the two microphones of a directional system mic far away from the user. The typical microphone distance in a hearing aid is in the order of 10 mm. The minimum distance of a sound source of interest for the user (e.g. the sound from the user's mouth or the sound from an audio transmission device) is in the order of 0.1 m (> 10 d mic ). For such a minimum distance, the hearing aid (microphone) will be in the acoustic near field of this sound source and the level difference of the sound signals incident on the respective microphones can be significant. The typical distance of a communication partner is larger than 1 m (> 100 d mic ). The hearing aid (microphone) will be in the acoustic far field of this sound source and the level difference of the sound signals incident on the respective microphones is not significant. The time difference of arrival of the sound incident in the direction of the microphone axis (e.g. the front or the back of a normal hearing aid) is ΔΤ = d mic / v sound = 0.01 / 343 [s] = 29 μs, where v sound is the speed of sound in air at 20°C (343 m / s).

[0154] The hearing aid can comprise an antenna and a transceiver circuit which enables the establishment of a wireless link to an entertainment device (e.g. a TV set), a communication device (like a telephone), a wireless microphone, a separate (external) processing device, or another hearing aid, etc. The hearing aid can thus be configured to wirelessly receive a direct electrical input signal from another device. Similarly, the hearing aid can be configured to wirelessly transmit a direct electrical output signal to another device. The direct electrical input or output signal can represent or comprise an audio signal and / or a control signal and / or an information signal.

[0155] Generally, the wireless link established by the antenna and the transceiver circuit of the hearing aid can be of any type. The wireless link can be a near field communication based link, e.g. an inductive link based on an inductive coupling between the antenna coils of the transmitter part and the receiver part. The wireless link can be based on far field electromagnetic radiation. Preferably, the frequency used for establishing the communication link between the hearing aid and another device is below 70 GHz, e.g. in the range from 50 MHz to 70 GHz, e.g. above 300 MHz, e.g. in the ISM range above 300 MHz, e.g. in the 900 MHz range or in the 2.4 GHz range or in the 5.8 GHz range or in the 60 GHz range (ISM = Industrial, Scientific and Medical, such standardized ranges are e.g. defined by the International Telecommunication Union, ITU). The wireless link can be based on a standardized or proprietary technology. The wireless link can be based on Bluetooth technology (e.g. Bluetooth Low Energy technology, e.g. LE Audio) or Ultra Wide Band (UWB) technology.

[0156] The hearing aid can be constituted by or form part of a portable, i.e. configured to be wearable, device, e.g. a device comprising a local energy source such as a battery, e.g. a rechargeable battery.

[0157] The hearing aid may, e.g., be a low-weight, easy-to-wear device, e.g. having a total weight of less than 100 g, such as less than 20 g, e.g. less than 5 g.

[0158] The hearing aid can comprise a "forward" or "signal" path between the input and output interfaces of the hearing aid for processing of audio signals. A signal processor can be located in this forward path. The signal processor can be adapted to provide a frequency-dependent gain in accordance with the specific needs of the user, e.g. hearing impaired. The hearing aid can comprise an "analysis" path having functionality for analysing the signals and / or controlling the processing of the forward path. Part or all of the signal processing of the analysis path and / or the forward path can be performed in the frequency domain, in which case the hearing aid comprises appropriate analysis and synthesis filter banks. Part or all of the signal processing of the analysis path and / or the forward path can be performed in the time domain.

[0159] Analog electrical signals representing the acoustic signals can be converted to digital audio signals in an analog-to-digital (AD) conversion process, in which the analog signals are sampled at a predetermined sampling frequency or sampling rate f s , f s , e.g. in the range from 8 kHz to 48 kHz (adapted to the specific needs of the application) to provide digital samples x n (or n) at discrete points in time t n (or x[n]), each audio sample representing the value of the acoustic signal at t b by a predetermined number N n of bits, N b , e.g. in the range from 1 to 48 bits, such as 24 bits. Each audio sample is thus quantized using an amount of N b bits, resulting in 2 Nb different possible values of the audio sample. The digital samples x have a time length of 1 / f s , e.g. 50 μs, for f s = 20 kHz. The audio samples can be arranged in time frames. A time frame can comprise 64 or 128 audio data samples. Other frame lengths can be used depending on the actual application.

[0160] The hearing aid can comprise an analog-to-digital (AD) converter to digitize an analog input, e.g. from an input transducer such as a microphone, at a predetermined sampling rate, e.g. 20 kHz. The hearing aid can comprise a digital-to-analog (DA) converter to convert digital signals to an analog output signal, e.g. for presentation to the user via an output interface transducer.

[0161] The input interface and / or antenna and transceiver circuitry of a hearing aid can comprise a transformation unit for transforming a time domain signal into a signal in a transformed domain, e.g. a frequency domain or a Laplace domain, a Z-transform, a wavelet transform, etc. The transformation unit can be constituted by or comprise a time- frequency (TF) transformation unit for providing a time-frequency representation of the input signal. The time-frequency representation can comprise an array or map of respective complex or real values of the signal of interest at certain time and frequency ranges. The TF transformation unit can comprise a filter bank for filtering the (time- varying) input signal and providing a plurality of (time-varying) output signals, each output signal comprising a distinct range of input signal frequencies. The TF transformation unit can comprise a Fourier transformation unit, e.g. a discrete Fourier transform (DFT) algorithm, a short-time Fourier transform (STFT) algorithm, or similar, for transforming the time-varying input signal into a (time-) frequency domain signal. The frequency range considered by the hearing aid from a minimum frequency f min to a maximum frequency f max may comprise a part of the typical human hearing frequency range from 20 Hz to 20 kHz, e.g. a part of the range from 20 Hz to 12 kHz. Typically, the sampling rate f s is greater than or equal to twice the maximum frequency f max , i.e. f s ≥ 2f max . The signal of the forward path and / or the analysis path of the hearing aid can be split into N1 (e.g. uniformly wide) frequency bands, where N1 is e.g. greater than 5, such as greater than 10, such as greater than 50, such as greater than 100, such as greater than 500, at least a part of which are individually processed. The hearing aid can be adapted to process the signal of the forward and / or analysis path in NP different channels (NP ≤ N1). The channels can be uniform or non-uniform in width (e.g. increasing in width with frequency), overlapping or non-overlapping.

[0162] The hearing aid can be configured to operate in different modes, such as a normal mode and one or more special modes, e.g. selectable by the user or selectable automatically. The operating mode can be optimized for a particular acoustic situation or environment, such as a communication mode, e.g. a telephone mode. The operating mode can comprise a low power mode, in which the functionality of the hearing aid is reduced (e.g. to save power), e.g. wireless communication is disabled and / or certain features of the hearing aid are disabled.

[0163] The hearing aid can comprise a plurality of detectors configured to provide a status signal related to the current network environment of the hearing aid, e.g. the current acoustic environment, and / or related to the current state of the user wearing the hearing aid, and / or related to the current state or operating mode of the hearing aid. One or more of the detectors can form part of an external device in communication with the hearing aid, e.g. wirelessly. The external device can e.g. comprise another hearing aid, a remote control, an audio streaming device, a telephone, e.g. a smartphone, an external sensor, etc.

[0164] One or more of the plurality of detectors can act on the full band signal (time domain). One or more of the plurality of detectors can act on the band split signal ((time-)frequency domain), e.g. in a limited number of frequency bands.

[0165] The plurality of detectors can comprise a level detector for estimating a current level of the signal of the forward path. The detector can be configured to determine whether the current level of the signal of the forward path is above or below a given (L-) threshold. The level detector can act on the full band signal (time domain). The level detector can act on the band split signal ((time-)frequency domain).

[0166] The hearing aid can comprise a voice activity detector (VAD) for estimating whether (or with what probability) the input signal (at a given point in time) comprises a voice signal. In the present specification, a voice signal can be intended to mean a signal comprising speech from a human being. It can also comprise other forms of vocalization (such as singing) produced by a human speech system. The voice activity detector unit can be adapted to classify the user's current acoustic environment as a "voice" or "no voice" environment. This has the advantage that periods of the microphone signal comprising human vocalization (such as speech) in the user's environment can be identified and thus separated from periods comprising only (or mainly) other sound sources (such as artificially produced noise). The voice activity detector can be adapted to also detect the user's own voice as "voice".

[0167] The hearing aid can comprise a self voice detector for estimating whether (or with what probability) a particular input sound (such as a voice, such as speech) originates from the hearing system user's voice. The microphone system of the hearing aid can be adapted to be able to distinguish the user's own voice from another person's voice and possibly from no voice sounds.

[0168] The plurality of detectors can comprise a motion detector, e.g. an acceleration sensor. The motion detector can be configured to detect motion of the user's facial muscles and / or bones, e.g. caused by speech or chewing (such as jaw movements) and provide a detector signal indicative of the motion.

[0169] The hearing aid can comprise a classification unit configured to classify the current situation based on input signals from (at least some of) the detectors and possibly other inputs. In the present specification, the "current situation" can be defined by one or more of the following:

[0170] a) the physical environment (such as comprising the current electromagnetic environment, e.g. the presence of planned or unplanned electromagnetic signals (including audio and / or control signals) received by the hearing aid, or the current environment differs from other properties of the acoustics);

[0171] b) the current acoustical situation (input level, feedback, etc.);

[0172] c) the current mode or state of the user (motion, temperature, cognitive load, etc.);

[0173] d) the current mode or state of the hearing aid and / or another device in communication with the hearing aid (selected program, elapsed time since last user interaction, etc.).

[0174] The classification unit can be based on or can comprise a neural network, e.g. a recurrent neural network, e.g. a trained neural network.

[0175] The hearing aid can comprise a sound (and / or mechanical) feedback control (like suppression) or an echo cancellation system. Adaptive feedback cancellation has the ability to track changes in the feedback path over time. It typically estimates the feedback path based on a linear time-invariant filter, but the filter weights are updated over time. The filter update can be computed using a stochastic gradient algorithm, including some form of least mean square (LMS) or normalized LMS (NLMS) algorithm. They all have the property of minimizing the squared difference signal, the NLMS additionally normalizing the filter update with respect to the squared Euclidean norm of some reference signal.

[0176] The hearing aid can further comprise other suitable functionality for the application at hand, like compression, noise reduction, etc.

[0177] The hearing aid can comprise a hearing instrument, e.g. a hearing instrument adapted to be located at the ear of the user or entirely or partially in the ear canal.

[0178] According to a third aspect of the present invention, there is provided a tangible computer readable medium having stored thereon a computer program comprising program code instructions for causing a data processing system to perform a method according to the second aspect of the present invention.

[0179] By way of example, and not limitation, such computer-readable media can comprise Random Access Memories (RAM), Read-Only Memories (ROM), Electronically Erasable Programmable Read-Only Memories (EEPROMs), Compact Disc Read-Only Memories (CD-ROMs) or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code means in the form of instructions or data structures and that can be accessed by a computer. The term "disk" as used herein includes compact discs (CDs), laser discs, optical discs, digital versatile discs (DVDs), floppy disks and blu-ray discs, which generally carry data magnetically or optically encoded, in some non-limiting embodiments. Other storage media include storage in deoxyribonucleic acid (DNA) (e.g., in a synthetic DNA strand). Combinations of the above should also be included within the scope of computer-readable media. In addition to storage on tangible media, computer programs can be transmitted via transmission media such as a wired or wireless link or network, e.g., the Internet, and loaded into a data processing system for execution by the data processing system.

[0180] According to a fourth aspect of the present application, there is provided a computer program product comprising a program comprising instructions for causing a data processing system to perform the method according to the second aspect of the present application.

[0181] According to a fifth aspect of the present application, there is provided a data processing system comprising a processor and program code for causing the processor to perform the method according to the second aspect.

[0182] In a sixth aspect of the present application, there is provided a hearing system comprising a hearing aid as described above, detailed in the "Detailed description of Invention" and defined in the claims, and a supplementary device.

[0183] The hearing system can be adapted to establish a communication link between the hearing aid and the supplementary device so that information, e.g. control and status signals, possibly audio signals, can be exchanged or forwarded from one device to the other.

[0184] The supplementary device can comprise or be constituted by a remote control, a smartphone, or other portable or wearable electronic device such as a smart watch or the like.

[0185] The supplementary device can be constituted by or comprise a remote control for controlling the functions and operation of the hearing aid. The functions of the remote control are implemented in a smartphone, which possibly runs an APP enabling control of the functions of the audio processing device via the smartphone (the hearing aid comprising a suitable wireless interface to the smartphone, e.g. based on Bluetooth or some other standardized or proprietary scheme).

[0186] The supplementary device can be constituted by or comprise an audio gateway device adapted to receive a plurality of audio signals (e.g. from an entertainment device such as a TV or a music player, from a telephone device such as a mobile phone, or from a computer such as a PC, a wireless microphone, etc.) and to select and / or combine appropriate ones (or combinations) of the received audio signals for transmission to the hearing aid.

[0187] The supplementary device can be constituted by or can comprise another hearing aid. The hearing system can comprise two hearing aids adapted to implement a binaural hearing system, e.g. a binaural hearing aid system.

[0188] In the present description, a hearing aid, e.g. a hearing instrument, refers to a device adapted to improve, enhance and / or protect the hearing ability of a user by receiving acoustic signals from the user's environment, generating corresponding audio signals, possibly modifying the audio signals, and providing the possibly modified audio signals as audible signals to at least one ear of the user. The audible signals can be provided, e.g., in the form of acoustic signals radiating into the outer ear of the user, and / or as mechanical vibrations through the bony structures of the user's head and / or through parts of the middle ear to the inner ear of the user.

[0189] The hearing aid can be configured to be worn in any known manner, e.g. as a unit worn behind the ear (with a tube leading the radiated acoustic signals into the ear canal or with an output interface transducer such as a loudspeaker arranged close to or in the ear canal), as a unit arranged entirely or partially in the pinna and / or ear canal, as a unit connected to a fixed structure implanted in the skull such as a vibrator, etc. The hearing aid can comprise a single unit or several units in communication with each other (e.g. acoustically, electrically or optically). The loudspeaker can be provided in a housing together with other components of the hearing aid, or it can be an external unit per se (possibly in combination with a flexible guiding element such as a dome-like element).

[0190] The hearing aid can be adapted to the needs of a specific user, e.g. a hearing impaired person. Configurable signal processing circuitry of the hearing aid can be adapted to apply frequency- and level-dependent compression amplification of input signals. Customized frequency- and level-dependent gain (amplification or compression) can be determined in a fitting process by a fitting system based on hearing data of the user, e.g. an audiogram, using fitting principles (e.g. adapted to speech) during the fitting process. The frequency- and level-dependent gain can be embodied, e.g., in processing parameters, e.g. transmitted to the hearing aid via an interface to a programming device (fitting system) and used by a processing algorithm executed by the configurable signal processing circuitry of the hearing aid.

[0191] A "hearing system" refers to a system comprising one or two hearing aids. A "binaural hearing system" refers to a system comprising two hearing aids and adapted to cooperatively provide audible signals to both ears of a user. A hearing system or binaural hearing system can further comprise one or more "auxiliary devices" which communicate with the hearing aid(s) and influence and / or benefit from the functionality of the hearing aid(s). The aforementioned auxiliary devices can comprise at least one of the following: a remote control, a remote microphone, an audio gateway device, an entertainment device such as a music player, a wireless communication device such as a mobile phone (e.g. a smartphone) or a tablet or another device, e.g. comprising a graphical interface. The hearing aid, hearing system or binaural hearing system can e.g. be used to compensate for a loss of hearing ability in a hearing impaired person, to enhance or protect the hearing ability of a normally hearing person and / or to transmit electronic audio signals to a person. The hearing aid or hearing system can e.g. form part of or interact with a public address system, an active ear protection system, a hands-free telephone system, a car audio system, an entertainment (e.g. TV, music play or karaoke) system, a teleconference system, a classroom amplification system, etc. BRIEF DESCRIPTION OF DRAWINGS

[0192] Various aspects of the application will be best understood with reference to the following detailed description together with the drawings, in which like reference numerals refer to like elements throughout. The drawings are schematic and simplified for clarity and the most detailed description is given with respect to the preferred embodiments. Like numbers refer to like elements throughout. Each aspect of the application can be combined with any or all other aspects of the application. These and other aspects, features and / or technical effects will be apparent from the following description, the drawings and the appended claims, and will be explained with respect thereto, in which:

[0193] Figure 1 a block diagram illustrating signal processing performed by a hearing aid according to an embodiment of the application;

[0194] Figure 2 a block diagram illustrating signal processing performed by a hearing aid according to another embodiment of the application;

[0195] Figure 3 a block diagram illustrating signal processing performed by a hearing aid according to yet another embodiment of the application.

[0196] Further areas of applicability of the present application will become apparent from the detailed description, the drawings and the claims given below. It should be understood, however, that the detailed description and specific examples, while indicating preferred embodiments of the application, are given by way of illustration only. Various modifications will become apparent to those skilled in the art from the foregoing detailed description, the drawings and the claims. DETAILED DESCRIPTION

[0197] The detailed description set forth below, in connection with the appended drawings, is intended as a description of various configurations and is not intended to limit the scope of the concepts. The detailed description includes specific details for the purpose of providing a thorough understanding of the various concepts. However, it will be apparent to those skilled in the art that these concepts can be practiced without these specific details. Numerous specific details are set forth in order to provide a thorough understanding of the subject matter. Well-known methods, procedures, components, and circuits have been described with sufficient detail for the purposes of providing a thorough understanding of the subject matter.

[0198] Electronic hardware can include micro-electromechanical systems (MEMS), (e.g., application-specific) integrated circuits, microprocessors, microcontrollers, digital signal processors (DSPs), field-programmable gate arrays (FPGAs), programmable logic devices (PLDs), gated logic, discrete hardware circuits, printed circuit boards (PCBs) (e.g., flexible PCBs), and other suitable hardware configured to perform the various functions described in this specification, such as sensors for sensing and / or recording physical properties of the environment, the device, the user, etc. Computer programs shall be construed broadly to include instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executables, threads of execution, procedures, functions, etc., whether referred to as software, firmware, middleware, microcode, hardware description language, or otherwise.

[0199] Figure 1 A block diagram illustrating signal processing performed by a hearing aid according to an embodiment of the application is shown. The hearing aid comprises an input interface IN. The input interface IN is configured to provide a plurality of input audio signals 12. The input interface IN can obtain the plurality of input audio signals 12 via a wireless link with another device. The input interface IN can comprise one or more microphones and be configured to convert acoustic signals 11 into the plurality of input audio signals 12.

[0200] The hearing aid comprises a first beamformer W1. The first beamformer W1 is configured to receive a first beamformer input signal 13 based on the plurality of input audio signals 12. The first beamformer input signal can be the plurality of input audio signals 12 or a processed version of the plurality of input audio signals 12. The first beamformer W1 is configured to determine a first beamformed signal 14 based on the first beamformer input signal 13. The first beamformer W1 can determine the first beamformed signal 14 by beamforming the first beamformer input signal 13. The beamforming performed by the first beamformer W1 can be done based on a distortionless criterion. The first beamformed signal 14 can be seen as a distortionless signal containing both a target signal and a noise signal, where the target signal is a signal that is desired and the noise signal is a part that is desired to be minimized and / or removed, e.g. background noise, transient noise, interfering speech or other unwanted signals.

[0201] The hearing aid comprises a second beamformer W2. The second beamformer W2 is configured to receive a second beamformer input signal 15 based on the plurality of input audio signals 12. The second beamformer input signal 15 can be the plurality of input audio signals 12 or a processed version of the plurality of input audio signals 12. The second beamformer W2 is configured to determine a second beamformed signal 16 based on the second beamformer input signal 15. The second beamformer W2 can determine the second beamformed signal 16 by beamforming the second beamformer input signal 15. The beamforming performed by the second beamformer W2 can be such that the second beamformer W2 acts as a target cancelling beamformer, where the second beamformed signal is determined by minimizing the power of a target in the second beamformer input signal 15. The second beamformed signal 16 can be seen as a targetless signal containing a noise signal, which in turn can be seen as containing different types of noise, e.g. high modulated noise or constant background noise.

[0202] The hearing aid comprises a first noise canceller N1. The first noise canceller N1 is configured to receive a first noise canceller input signal 18 based on the second beamformed signal 16. The first noise canceller input signal 18 can be the second beamformed signal 16 or a processed version of the second beamformed signal 16. The first noise canceller N1 is configured to determine a first anti-noise signal 20 based on the first noise canceller input signal 18. The first anti-noise signal 20 is an anti-noise signal for a first type of noise. The first noise canceller N1 can be a noise canceller that is specialized in handling slowly modulated noise, e.g. constant background noise. The first noise canceller N1 can be configured to filter the second beamformed signal 16 according to a set of first filter coefficients to obtain the first anti-noise signal 20. The first anti-noise signal 20 can be considered to represent a noise signal of the first type of noise in the plurality of input audio signals 12. The first noise canceller N1 is configured to output the first anti-noise signal 20.

[0203] The hearing aid comprises a second noise canceller N2. The second noise canceller N2 is configured to receive a second noise canceller input signal 21 based on the second beamformed signal 16. The second noise canceller input signal 21 can be the second beamformed signal 16 or a processed version of the second beamformed signal 16. The second noise canceller N2 is configured to determine a second anti-noise signal 22 based on the second noise canceller input signal 21. The second anti-noise signal 22 is an anti-noise signal for a second type of noise. The second noise canceller N2 can be a noise canceller that is specialized in handling highly modulated noise, e.g. speech or other transient noise. The second noise canceller N2 can be configured to filter the second beamformed signal 16 according to a set of second filter coefficients to obtain the second anti-noise signal 22. The second anti-noise signal 22 can be considered to represent a noise signal of the second type of noise in the plurality of input audio signals 12.

[0204] The hearing aid comprises an output interface OUT. The output interface OUT is configured to determine an output signal 23 based on the first beamformed signal 14, the first anti-noise signal 20 and the second anti-noise signal 22. The output interface OUT can be configured to determine the output signal 23 as a linear combination of the first beamformed signal 14, the first anti-noise signal 20 and the second anti-noise signal 22. The output interface OUT can be configured to determine the output signal 23 by subtracting the first anti-noise signal 20 and the second anti-noise signal 22 from the first beamformed signal 14. The output interface OUT is configured to output the output signal 23. The output interface OUT can output the output signal 23 for further processing. The output interface OUT can output the output signal 23 to one or more output transducers for conversion to sound to be provided to a user of the hearing aid.

[0205] Figure 2A block diagram illustrating signal processing performed by a hearing aid according to an embodiment of the application is shown. The hearing aid comprises an input interface IN. The input interface IN is configured to provide a plurality of input audio signals 12. The input interface IN can obtain the plurality of input audio signals 12 via a wireless link with another device. The input interface IN can comprise one or more microphones and be configured to convert acoustic signals 11 into the plurality of input audio signals 12.

[0206] The hearing aid comprises a first beamformer W1. The first beamformer W1 is configured to receive a first beamformer input signal 13 based on the plurality of input audio signals 12. The first beamformer input signal can be the plurality of input audio signals 12 or a processed version of the plurality of input audio signals 12. The first beamformer W1 is configured to determine a first beamformed signal 14 based on the first beamformer input signal 13. The first beamformer W1 can determine the first beamformed signal 14 by beamforming the first beamformer input signal 13. The beamforming performed by the first beamformer W1 can be done based on a distortionless criterion. The first beamformed signal 14 can be considered a distortionless signal containing both a target signal and a noise signal, where the target signal is a signal that is desired and the noise signal is a part that is desired to be minimized and / or removed, e.g. background noise, transient noise, interfering speech or other undesired signals.

[0207] The hearing aid comprises a first detector D1. The first detector is configured to receive a first detector input signal 17 based on the plurality of input audio signals 12. In this embodiment, the first detector D1 receives the plurality of input audio signals 12, in other embodiments the first detector D1 can receive a processed version of the plurality of input audio signals 12. The first detector D1 is configured to determine a first audio parameter based on the first detector input signal 17. The first detector D1 can be a voice activity detector and the first audio parameter can be a voice activity parameter, e.g. a speech presence probability or a self voice presence probability. The first detector D1 is configured to determine a first detector output signal 18 based on the determined first audio parameter. The first detector output signal 18 can be determined to contain the first audio parameter or a processed version of the first audio parameter.

[0208] The hearing aid comprises a second beamformer W2. The second beamformer W2 is configured to receive a second beamformer input signal 15 based on the plurality of input audio signals 12. The second beamformer input signal 15 can be the plurality of input audio signals 12 or a processed version of the plurality of input audio signals 12. The second beamformer W2 is configured to determine a second beamformed signal 16 based on the second beamformer input signal 15. The second beamformer W2 can determine the second beamformed signal 16 by beamforming the second beamformer input signal 15. The beamforming performed by the second beamformer W2 can be such that the second beamformer W2 acts as a target cancelling beamformer, where the second beamformed signal is determined by minimizing the power of a target in the second beamformer input signal 15. The second beamformed signal 16 can be seen as a target free signal, which contains a noise signal, which in turn can be seen as containing noise corresponding to different types of noise, e.g. high modulated noise or constant background noise.

[0209] The hearing aid comprises a first noise canceller N1. The first noise canceller N1 is configured to receive a first noise canceller input signal 18, 19 based on the second beamformed signal 16 and the first detector output signal 18. The first noise canceller input signal 18, 19 can be the second beamformed signal 16 and the first detector output signal 18 or a processed version of the second beamformed signal 16 and the first detector output signal 18. The first noise canceller N1 is configured to determine a first anti-noise signal 20 based on the first noise canceller input signal 18, 19. The first anti-noise signal 20 is an anti-noise signal for the first type of noise. The first noise canceller N1 can be a noise canceller that is specifically adapted to handle slowly modulated noise, e.g. constant background noise. The first noise canceller N1 can be activated in dependence of the first detector output signal 18, e.g. if the detector only or mainly detects the second type of noise, then it can be more optimal to fix the first noise canceller N1 as it is adapted to the first type of noise. Or if the detector detects the first type of noise, then it can be more optimal to adapt the first noise canceller to that noise. The first noise canceller N1 can be configured to filter the second beamformed signal 16 according to a set of first filter coefficients, resulting in the first anti-noise signal 20. Thus, the first detector output signal 18 can be seen as a control signal controlling the operation of the first noise canceller, and the second beamformed signal 19 can be seen as a signal that is filtered by the first noise canceller N1 to obtain the first anti-noise signal 20. The first anti-noise signal 20 can be seen as a noise signal representing the first type of noise in the plurality of input audio signals 12. The first noise canceller N1 is configured to output the first anti-noise signal 20.

[0210] The hearing aid comprises a second noise canceller N2. The second noise canceller N2 is configured to receive a second noise canceller input signal 21 based on the second beamformed signal 16. The second noise canceller input signal 21 can be the second beamformed signal 16 or a processed version of the second beamformed signal 16. The second noise canceller N2 is configured to determine a second anti-noise signal 22 based on the second noise canceller input signal 21. The second anti-noise signal 22 is an anti-noise signal for the second type of noise. The second noise canceller N2 can be a noise canceller that is specialized in handling highly modulated noise, such as speech or other transient noise. The second noise canceller N2 can be configured to filter the second beamformed signal 16 according to a set of second filter coefficients, resulting in the second anti-noise signal 22. The second anti-noise signal 22 can be seen as a noise signal representing the second type of noise in the plurality of input audio signals 12.

[0211] The hearing aid comprises an output interface OUT. The output interface OUT is configured to determine an output signal 23 based on the first beamformed signal 14, the first anti-noise signal 20 and the second anti-noise signal 22. The output interface OUT can be configured to determine the output signal 23 as a linear combination of the first beamformed signal 14, the first anti-noise signal 20 and the second anti-noise signal 22. The output interface OUT can be configured to determine the output signal 23 by subtracting the first anti-noise signal 20 and the second anti-noise signal 22 from the first beamformed signal 14. The output interface OUT is configured to output the output signal 23. The output interface OUT can output the output signal 23 for further processing. The output interface OUT can output the output signal 23 to one or more output transducers, which convert it to sound provided to a user of the hearing aid.

[0212] Reference is made to Figure 3 which shows a block diagram of signal processing performed by a hearing aid according to another embodiment of the application. The hearing aid comprises an input interface IN. The input interface IN is configured to provide a plurality of input audio signals 12. The input interface IN can acquire the plurality of input audio signals 12 through a wireless link with another device. In this embodiment, the input interface IN comprises two or more microphones and is configured to convert a sound signal 11 into the plurality of input audio signals 12. The plurality of input audio signals 12 can be modelled as:

[0213] x(k, n) = s(k, n) d(k, n) + v(k, n)

[0214] where s(k, n) is the desired signal at a reference microphone, e.g. a front microphone on the hearing aid, d(k, n) is a relative transfer function (RTF) vector encoding the relative acoustic properties between the microphones and the reference microphone, and v(k, n) is a noise vector.

[0215] In general, a minimum variance distortionless response (MVDR) beamformer is used to suppress v(k, n). The MVDR beamformer is a linear combination of x(k, n) that minimizes the noise power at the output of the beamformer and is constrained to have a distortionless response on the desired signal RTF vector. The output of the MVDR beamformer can be expressed as:

[0216]

[0217] where H denotes the complex conjugate transpose of a vector. w MVDR (k, n) is a vector of MVDR beamformer weights. The MVDR beamformer weights can be solved using a generalized sidelobe canceller (GSC) structure. For the two-microphone case, the beamformer weights can be expressed as:

[0218] w MVDR (k, n) = w DSB (k, n) - w TC (k, n) β(k, n)

[0219] where w DSB (k, n) is the weight of a delay-and-sum beamformer, or the weight of an arbitrary beamformer with a distortionless constraint on the desired signal RTF vector; w TC (k, n) is the weight of a target cancellation beamformer; and β(k, n) is a noise canceller. Thus, the output of the MVDR beamformer can be expressed as:

[0220]

[0221] where * denotes the complex conjugate operator. In the present embodiment, the hearing aid comprises a first beamformer W1. The first beamformer W1 is configured to receive a first beamformer input signal 13 based on the plurality of input audio signals 12. The first beamformer input signal can be the plurality of input audio signals 12, or a processed version of the plurality of input audio signals 12. The first beamformer W1 is configured to determine a first beamformed signal 14 based on the first beamformer input signal 13. In the present embodiment, the first beamformer W1 is a beamformer with a distortionless criterion. In other words, the first beamformed signal can be expressed as:

[0222]

[0223] or at least a processed version thereof.

[0224] The hearing aid further includes a second beamformer W2. The second beamformer W2 is configured to receive a second beamformer input signal 15 based on the plurality of input audio signals 12. The second beamformer input signal 15 may be the plurality of input audio signals 12, or a processed version of the plurality of input audio signals 12. The second beamformer W2 is configured to determine a second beamformed signal 16 based on the second beamformer input signal 15. The second beamformer is a target cancellation beamformer configured to determine the second beamformed signal by minimizing the power of a target based on the plurality of input audio signals. In other words, the second beamformed signal can be expressed as:

[0225]

[0226] Or at least a processed version of it.

[0227] The hearing aid includes a first detector D1. The first detector D1 is configured to receive a first detector input signal 17 based on multiple input audio signals 12. In this embodiment, the first detector D1 receives the multiple input audio signals 12; in other embodiments, the first detector D1 may receive processed versions of the multiple input audio signals 12. The first detector D1 is configured to determine a first audio parameter based on the first detector input signal 17. The first audio parameter is a voice activity parameter. For example, the first detector D1 may be a voice activity detector or an ego voice activity detector. The first detector D1 is configured to determine a first detector output signal 18 based on the determined first audio parameter. The first detector D1 is configured to compare the first audio parameter with a first threshold. If the first audio parameter exceeds the first threshold, the first detector output signal 18 is determined to include an activation signal, which allows the first noise canceller N1 to perform adaptive adjustment by reducing the error between the first anti-noise signal 20 and the first target signal 14 based on the first beamforming signal 14. If the first audio parameter does not exceed the first threshold, the first detector output signal 18 is determined to include a fixed signal, which is used to fix the first noise canceller N1. For example, if the first noise canceller N1 is configured to process slowly time-varying noise such as background noise, the first detector D1 may be configured to activate the first noise canceller N1 when no voice activity or low voice activity is detected, and to freeze the first noise canceller N1 in case of high voice activity.

[0228] The hearing aid comprises a second detector D2. The second detector D2 is configured to receive a second detector input signal 24 based on the plurality of input audio signals 12. In the present embodiment, the second detector D2 receives the plurality of input audio signals 12; in other embodiments, the second detector D2 can receive a processed version of the plurality of input audio signals 12. The second detector D2 is configured to determine a second audio parameter based on the second detector input signal 24. The second audio parameter is an interfering activity parameter. For example, the second detector D2 can be an interfering signal detector configured to detect highly modulated noise. The second detector D2 is configured to determine a second detector output signal 25 based on the determined second audio parameter. The second detector D2 is configured to compare the second audio parameter to a second threshold. If the second audio parameter exceeds the second threshold, the second detector output signal 25 is determined to comprise an activation signal that allows the second noise canceller N2 to adapt by reducing an error between the second anti-noise signal 22 and a second target signal based on the first beamformed signal 14; if the second audio parameter does not exceed the second threshold, the second detector output signal 25 is determined to comprise a fixation signal for fixing the second noise canceller N2. For example, if the second noise canceller N2 is configured to process fast time-varying noise such as speech, the second detector D2 can be configured to activate the second noise canceller N2 when interfering speech is detected; whereas the second noise canceller N2 is fixed in the absence of fast time-varying noise.

[0229] The hearing aid comprises a third detector D3. The third detector is configured to receive a third detector input signal 28 based on the plurality of input audio signals 12. In the present embodiment, the third detector D3 receives the plurality of input audio signals 12; in other embodiments, the third detector D3 can receive a processed version of the plurality of input audio signals 12. The third detector D3 is configured to determine a third audio parameter based on the third detector input signal 28. The third audio parameter is an audio feedback parameter. For example, the third detector D3 can be a feedback detector for detecting feedback. The third detector D3 is configured to determine a third detector output signal 29 based on the determined third audio parameter. The third detector D3 is configured to compare the third audio parameter to a third threshold. If the third audio parameter exceeds the third threshold, the third detector output signal 29 is determined to comprise an activation signal that allows the third noise canceller N3 to adapt by reducing the error between the third anti-noise signal 22 and a third target signal based on the first beamformed signal 14; if the third audio parameter does not exceed the third threshold, the third detector output signal 29 is determined to comprise a fixation signal for fixing the third noise canceller N3. For example, if the third noise canceller N3 is configured to process feedback, the third detector D3 can be configured to activate the third noise canceller N3 upon detection of feedback; whereas the third noise canceller N3 is fixed in case no feedback is detected.

[0230] In the present embodiment, a plurality of dedicated noise cancellers N1, N2 and N3 is used. The first noise canceller N1 is dedicated to attenuate slowly time-varying background noise, the second noise canceller N2 is dedicated to cancel highly modulated interference signals from one or more specific directions, and the third noise canceller N3 is dedicated to cancel feedback. The division of the noise cancellers N1, N2 and D3 into a plurality of parts enables a dedicated control of each individual noise canceller. This is achieved by using different detectors D1, D2 and D3, smoothing factors and gains G1, G2 and G3 applied to the output of each noise canceller N1, N2 and N3. In the shown multi-stage noise cancellation structure, there are I noise reference signals and I target outputs. The noise reference signals can be defined as:

[0231]

[0232] where the index i refers to the i-th noise reference signal. The target outputs are given as:

[0233] y trg, i(k,n) = y trg,i-1 (k,n) - y vref,i (k,n)

[0234]

[0235] where if i = 1, y trg,i-1 (k,n) = y DSB (k,n). In principle, the noise canceller can be represented by a single noise canceller value that is summed and represented by:

[0236]

[0237] To compute each noise canceller, the noise canceller needs to be computed in order starting with β1(k,n) since each noise canceller depends on the previous noise canceller. The computation of each noise canceller can be expressed in a two-microphone system as follows:

[0238]

[0239] where a i is the i-th smoothing factor, and where the operator symbol <·> refers to first order recursive smoothing, i.e.:

[0240]

[0241] Each noise canceller can also be approximated using an iterative solver instead of the closed form equation. One commonly used iterative solver is the gradient descent method, which can be expressed as:

[0242]

[0243] where μ i is the step size, and

[0244]

[0245] The hearing aid comprises a first noise canceller N1. The first noise canceller N1 is configured to receive a first noise canceller input signal 18, 19 based on the second beamformed signal 16 and the first detector output signal 18. The first noise canceller input signal 18, 19 can be the second beamformed signal 16 and the first detector output signal 18, or a processed version of the second beamformed signal 16 and the first detector output signal 18. The first noise canceller N1 is configured to determine a first anti-noise signal 20 based on the first noise canceller input signal 18, 19. The first anti-noise signal 20 is an anti-noise signal for the first type of noise. The first noise canceller N1 can be expressed as

[0246]

[0247] The first noise canceller N1 can be a noise canceller that is specifically adapted to handle slowly modulated noise, such as constant background noise. The first noise canceller N1 can be activated in dependence of the first detector output signal 18, e.g. if the detector detects only or mainly the second type of noise, it can be more appropriate to fix the first noise canceller N1 as it is adapted to the first type of noise. Or if the first detector D1 detects the first type of noise, it can be more appropriate to let the first noise canceller adapt to this noise. The first noise canceller N1 can be configured to filter the second beamformed signal 16 in dependence of a set of first filter coefficients, resulting in a first anti-noise signal 20. Thus, the first detector output signal 18 can be seen as a control signal that controls the operation of the first noise canceller N1, and the second beamformed signal 19 can be seen as the signal to be filtered by the first noise canceller N1 to obtain the first anti-noise signal 20. The first anti-noise signal 20 can be seen as a noise signal that represents the first type of noise in the plurality of input audio signals 12. The first noise canceller N1 comprises a first smoothing factor. The smoothing factor of a noise canceller can define the adaptation rate of the relevant noise canceller. The first noise canceller N1 is configured to output the first anti-noise signal 20.

[0248] The hearing aid comprises a second noise canceller N2. The second noise canceller N2 is configured to receive a second noise canceller input signal 21 based on the second beamformed signal 16. The second noise canceller input signal 21 can be the second beamformed signal 16 or a processed version of the second beamformed signal 16. The second noise canceller N2 is configured to determine a second anti-noise signal 22 based on the second noise canceller input signal 21. The second anti-noise signal 22 is an anti-noise signal for the second type of noise. The second noise canceller N2 can be a noise canceller that is specifically adapted to handle highly modulated noise such as speech or other transient noise. The second noise canceller N2 can be activated in dependence of the second detector output signal 25, e.g. if the detector detects only or mainly the first type of noise, it can be more appropriate to fix the second noise canceller N2 as it is adapted to the second type of noise. Or if the second detector D2 detects the second type of noise, it can be more appropriate to adapt the second noise canceller N2 to this noise. The second noise canceller N2 can be configured to filter the second beamformed signal 16 according to a set of second filter coefficients to obtain the second anti-noise signal 22. Thus, the second detector output signal 25 can be seen as a control signal that controls the operation of the second noise canceller N2 and the second beamformed signal 21 can be seen as the signal to be filtered by the second noise canceller N2 to obtain the second anti-noise signal 22. The second anti-noise signal 22 can be seen as a noise signal that represents the second type of noise in the plurality of input audio signals 12. The second noise canceller N2 is configured to output the second anti-noise signal 22. The second noise canceller N2 comprises a second smoothing factor. The first smoothing factor results in a lower adaptation rate than the second smoothing factor.

[0249] The hearing aid comprises a third noise canceller N3. The third noise canceller N3 is configured to receive a third noise canceller input signal 26 based on the second beamformed signal 16. The third noise canceller input signal 26 can be the second beamformed signal 16 or a processed version of the second beamformed signal 16. The third noise canceller N3 is configured to determine a third anti-noise signal 38 based on the third noise canceller input signal 26. The third anti-noise signal 38 is an anti-noise signal for the third type of noise. The third noise canceller N3 can be a noise canceller dedicated to reduce noise caused by feedback, such as an echo or mechanical feedback noise. The third noise canceller N3 can be activated in dependence of the third detector output signal 29, e.g. if the detector detects only or mainly the second type of noise, it can be more appropriate to fix the third noise canceller N3, since the third noise canceller N3 is adapted to the third type of noise. Alternatively, if the third detector D3 detects the third type of noise, it can be more appropriate to let the third noise canceller N3 adapt to this noise. The third noise canceller N3 can be configured to filter the second beamformed signal 16 according to a set of third filter coefficients, resulting in the third anti-noise signal 38. Thus, the third detector output signal 29 can be seen as a control signal controlling the operation of the third noise canceller N3, and the second beamformed signal 16 can be seen as a signal to be filtered by the third noise canceller N3 to obtain the third anti-noise signal 38. The third anti-noise signal 38 can be seen as a noise signal representing the third type of noise in the plurality of input audio signals 12. The third noise canceller N3 is configured to output the third anti-noise signal 38.

[0250] The hearing aid comprises a first gain controller G1. The first gain controller G1 is configured to receive the first anti-noise signal 20, determine a first control gain, apply the first control gain to the first anti-noise signal 20, and output a first noise controlled signal 30. The first control gain can be a predetermined gain set by an audio engineer, an audiologist or a hearing aid user. In the present embodiment, the first control gain is an adaptive gain determined based on the first detector output signal 18. The first control gain can be determined as a function of the first audio parameter determined by the first detector D1. For example, if the first audio parameter is determined to be a speech presence probability of 0.9, the first control gain can be set to a global gain of 0.9.

[0251] The hearing aid comprises a second gain controller G2. The second gain controller G2 is configured to receive the second anti-noise signal 22, determine a second control gain, apply the second control gain to the second anti-noise signal 22, and output a second noise-controlled signal 24. The second control gain can be a predetermined gain set by an acoustical engineer, an audiologist or a hearing aid user. In the present embodiment, the second control gain is an adaptive gain determined based on the second detector output signal 25. The second control gain can be determined as a function of a second audio parameter determined by the second detector D2. For example, if the second audio parameter is determined to be a probability of interference presence of 0.9, the second control gain can be set to a global gain of 0.9.

[0252] The hearing aid comprises a third gain controller G3. The third gain controller G3 is configured to receive the third anti-noise signal 28, determine a third control gain, apply the third control gain to the third anti-noise signal 28, and output a third noise-controlled signal 27. The third control gain can be a predetermined gain set by an acoustical engineer, an audiologist or a hearing aid user. In the present embodiment, the third control gain is an adaptive gain determined based on the third detector output signal 29. The third control gain can be determined as a function of a third audio parameter determined by the third detector D2. For example, if the third audio parameter is determined to be a probability of feedback of 0.9, the third control gain can be set to a global gain of 0.9.

[0253] In the present embodiment, the first beamformer W1 and the second beamformer W2 are fixed filters, while the first noise canceller N1, the second noise canceller N2 and the third noise canceller N3 are adaptive filters. The first noise canceller N1 is adapted based on a first error signal 32. The first error signal 32 can be determined by an output interface OUT. The output interface OUT can determine the first error signal 32 by subtracting the first noise-controlled signal 30 from the first beamformed signal 14. The second noise canceller N2 is adapted based on a second error signal 35. The second error signal 35 can be determined by the output interface OUT. The output interface OUT can determine the second error signal 35 by subtracting the second noise-controlled signal 24 from the first error signal 32. The third noise canceller N3 is adapted based on a third error signal 40. The third error signal 40 can be determined by the output interface OUT. The output interface OUT can determine the third error signal 40 by subtracting the third noise-controlled signal 27 from the second error signal 35.

[0254] The hearing aid includes an output interface OUT. The output interface OUT is configured to determine an output signal 23 based on the first beamformed signal 14, the first noise-controlled signal 30, the second noise-controlled signal 24, and the third noise-controlled signal 27. The output interface OUT is configured to determine the output signal 23 as a linear combination of the first beamformed signal 14, the first noise-controlled signal 30, the second noise-controlled signal 24, and the third noise-controlled signal 27. The output interface OUT is configured to subtract the first noise-controlled signal 30, the second noise-controlled signal 24, and the third noise-controlled signal 27 from the first beamformed signal 14 to determine the output signal 23. The output interface OUT is configured to output the output signal 23. The output interface OUT can output the output signal 23 for further processing. The output interface OUT can output the output signal 23 to one or more output transducers to convert the output signal 23 into sound for a user of the hearing aid.

[0255] Combine Figure 1 、 Figure 2 and Figure 3 The hearing aids described may be air conduction hearing aids, bone conduction hearing aids or a combination thereof.

[0256] The term "processed version" may, for example, cover features extracted from an original audio signal. The term "processed version" may also, for example, cover an original audio signal that has been subjected to a processing algorithm that applies a gain or attenuation and / or a delay to the original audio signal and results in a modified audio signal (preferably enhanced in some way, e.g., noise reduced relative to a target signal, or simply delayed).

[0257] The structural features of the apparatus described above, described in detail in the "Detailed Description of the Invention" and defined in the claims may be combined with the steps of the method of the present invention when appropriately replaced by corresponding processes.

[0258] Unless expressly stated otherwise, the singular forms "a", "the" and "the" used herein include the plural form (i.e., having the meaning of "at least one"). It should be further understood that the terms "having", "including" and / or "comprising" used in the specification indicate the presence of the described features, integers, steps, operations, elements and / or parts, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, parts and / or combinations thereof. It should be understood that, unless expressly stated otherwise, when an element is referred to as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be an intermediate intervening element. The term "and / or" as used herein includes any and all combinations of one or more listed related items. Unless expressly stated otherwise, the steps of any method disclosed herein do not have to be performed in the exact order disclosed.

[0259] It is to be appreciated that a reference to "one embodiment" or "an embodiment" or "aspect" or "may" in the specification or claims indicates that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one implementation of the application. The appearances of the phrase "in one embodiment" or "an embodiment" or "aspect" in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily referring to one specific embodiment. Furthermore, the particular features, structures, or characteristics can be combined in any suitable manner on one or more implementations. The foregoing description has been presented for purposes of illustration.

[0260] The claims are not limited to the aspects shown in the figures, but include all alternatives consistent with the claims language to the full extent permitted by applicable law, wherein the term "each" in a claim is not intended to mean "one and only one" unless explicitly so defined, and that the terms "a," "an," and "the" mean "one or more" unless explicitly stated otherwise. The term "plurality" refers to "two or more."

[0261] Enumerated list of items

[0262] Examples of hearing aids, systems and methods according to the application are presented in the following items:

[0263] Item 1. A hearing aid comprising:

[0264] an input interface configured to provide a plurality of input audio signals;

[0265] a first beamformer configured to receive a first beamformer input signal based on the plurality of input audio signals, and to determine a first beamformed signal based on the first beamformer input signal;

[0266] a second beamformer configured to receive a second beamformer input signal based on the plurality of input audio signals, and to determine a second beamformed signal based on the second beamformer input signal;

[0267] a first detector configured to receive a first detector input signal based on the plurality of input audio signals, to determine a first audio parameter based on the first detector input signal, to determine a first detector output signal based on the determined first audio parameter, and to output the first detector output signal;

[0268] a first noise canceller configured to receive a first noise canceller input signal based on the second beamformed signal and the first detector output signal, to determine a first anti-noise signal based on the first noise canceller input signal and to output the first anti-noise signal, wherein the first anti-noise signal is an anti-noise signal for a first type of noise;

[0269] a second noise canceller configured to receive a second noise canceller input signal based on the second beamformed signal, determine a second anti-noise signal based on the second beamformed signal and output the second anti-noise signal, wherein the second anti-noise signal is an anti-noise signal for a second type of noise;

[0270] an output interface configured to determine an output signal based on the first beamformed signal, the first anti-noise signal and the second anti-noise signal, and output the output signal.

[0271] Item 2. The hearing aid according to item 1, wherein the input interface comprises two or more microphones configured to provide the plurality of input audio signals.

[0272] Item 3. The hearing aid according to item 1 or 2, wherein the first beamformer and the second beamformer are fixed filters, and the first noise canceller and the second noise canceller are adaptive filters.

[0273] Item 4. The hearing aid according to any of items 1-3, wherein the first type of noise is a slow time-varying noise, and the second type of noise is a fast time-varying noise.

[0274] Item 5. The hearing aid according to any of items 1-4, wherein the first noise canceller comprises a first smoothing factor and the second noise canceller comprises a second smoothing factor, the smoothing factors defining an adaptation rate of the respective noise canceller, wherein the first smoothing factor results in a lower adaptation rate than the second smoothing factor.

[0275] Item 6. The hearing aid according to any of items 1-5, wherein determining the first detector output signal comprises comparing the first audio parameter to a first threshold,

[0276] if the first audio parameter exceeds the first threshold, determining the first detector output signal comprises an activation signal, the activation signal allowing the first noise canceller to adapt by reducing an error between the first anti-noise signal and a first target signal based on the first beamformed signal; and

[0277] if the first audio parameter does not exceed the first threshold, determining the first detector output signal comprises a fixation signal, the fixation signal for fixing the first noise canceller.

[0278] Item 7. The hearing aid according to any one of the items 1-6, comprising a second detector configured to receive a second detector input signal based on the plurality of input audio signals, determine a second audio parameter based on the second detector input signal, determine a second detector output signal based on the determined second audio parameter and output the second detector output signal,

[0279] wherein the second noise canceller is configured to receive the second detector output signal and determine the second anti-noise signal based on the second detector output signal.

[0280] Item 8. The hearing aid according to any one of the items 1-7, wherein determining the second detector output signal comprises comparing the second audio parameter to a second threshold,

[0281] if the second audio parameter exceeds the second threshold, determining the second detector output signal comprises an activation signal allowing the second noise canceller to adapt by reducing an error between a second target signal based on the first beamformed signal and the second anti-noise signal and the second noise canceller input signal; and

[0282] if the second audio parameter does not exceed the second threshold, determining the second detector output signal comprises a fixation signal for fixing the second noise canceller.

[0283] Item 9. The hearing aid according to item 7 or 8, wherein the second audio parameter is an interference activity parameter.

[0284] Item 10. The hearing aid according to any one of the items 1-9, wherein the first audio parameter is a speech activity parameter.

[0285] Item 11. The hearing aid according to any one of the items 1-10, comprising:

[0286] a first gain controller configured to receive the first anti-noise signal, determine a first control gain, apply the first control gain to the first anti-noise signal and output a first noise controlled signal;

[0287] wherein the output interface is configured to determine the output signal based on the first noise controlled signal.

[0288] Item 12. The hearing aid according to item 11, wherein the first control gain is a predetermined gain.

[0289] Item 13. The hearing aid according to item 11, wherein the first control gain is an adaptive gain determined based on the first detector output signal.

[0290] Item 14. The hearing aid according to any of items 1-13, comprising:

[0291] a second gain controller configured to receive the second anti-noise signal, determine a second control gain, apply the second control gain to the second anti-noise signal and output a second noise-controlled signal; and

[0292] wherein the output interface is configured to determine the output signal based on the second noise-controlled signal.

[0293] Item 15. The hearing aid according to item 14, wherein the second control gain is a predetermined gain.

[0294] Item 16. The hearing aid according to item 14, wherein the second control gain is an adaptive gain determined based on the second detector output signal.

[0295] Item 17. The hearing aid according to any of items 1-16, comprising:

[0296] a third detector configured to receive a third detector input signal based on the plurality of input audio signals, determine a third audio parameter based on the third detector input signal, determine a third detector output signal based on the determined third audio parameter and output the third detector output signal;

[0297] a third noise canceller configured to receive a third noise canceller input signal based on the second beamformed signal and the third detector output signal, determine a third anti-noise signal based on the third noise canceller input signal and output the third anti-noise signal, wherein the third anti-noise signal is an anti-noise signal for a third type of noise; and

[0298] wherein the output interface is configured to determine the output signal based on the third anti-noise signal.

[0299] Item 18. The hearing aid according to item 17, wherein the third audio parameter is an audio feedback parameter.

[0300] Item 19. The hearing aid according to item 17 or 18, comprising:

[0301] a third gain controller configured to receive the third anti-noise signal, determine a third control gain, apply the third control gain to the third anti-noise signal and output a third noise-controlled signal;

[0302] wherein the output interface is configured to determine the output signal based on the third noise-controlled signal.

[0303] Item 20. The hearing aid according to item 19, wherein the third control gain is an adaptive gain determined based on the first detector output signal.

[0304] Item 21. The hearing aid according to item 19, wherein the third control gain is a predetermined gain.

[0305] Item 22. The hearing aid according to any of items 1-21, wherein the first beamformer is a beamformer with a distortionless criterion.

[0306] Item 23. The hearing aid according to any of items 1-22, wherein the second beamformer is a target cancellation beamformer configured to determine the second beamformed signal based on the plurality of input audio signals by minimizing a power of a target.

[0307] Item 24. The hearing aid according to any of items 1-23, comprising an air conduction type hearing aid, a bone conduction type hearing aid, or a combination thereof.

[0308] Item 25. The hearing aid according to any of items 1-24, wherein determining the first anti-noise signal comprises determining a first correlation between the first beamformed signal and the first noise canceller input signal.

[0309] Item 26. The hearing aid according to any of items 1-25, wherein determining the first noise canceller comprises normalizing the first correlation using an amplitude of the first noise canceller input signal.

[0310] Item 27. The hearing aid according to any of items 1-26, wherein determining the first noise canceller comprises determining a first correlation between the first anti-noise signal and the first noise canceller input signal.

[0311] Item 28. The hearing aid according to any of items 1-27, wherein determining the first noise canceller comprises reducing an amplitude error between the first beamformed signal and the first noise canceller input signal.

[0312] Item 29. The hearing aid according to any of items 1-28, wherein determining the second noise canceller comprises determining a second correlation between the first anti-noise signal and the first noise canceller input signal.

[0313] Item 30. The hearing aid according to item 29, wherein determining the first noise canceller comprises normalizing the second correlation using an amplitude of the first noise canceller input signal.

[0314] Item 31. The hearing aid according to any of items 1-30, wherein determining the second noise canceller comprises determining a second correlation between the second anti-noise signal and the first noise canceller input signal.

[0315] Item 32. The hearing aid according to any of items 1-27, wherein determining the first noise canceller comprises reducing an amplitude error between the second anti-noise signal and the first noise canceller input signal.

[0316] Item 33. The hearing aid according to any of items 1-32, wherein determining the third noise canceller comprises determining a third correlation between the second anti-noise signal and the first noise canceller input signal.

[0317] Item 34. The hearing aid according to item 32, wherein determining the third noise canceller comprises normalizing the third correlation using an amplitude of the first noise canceller input signal.

[0318] Item 35. A method of operating a hearing aid, comprising:

[0319] providing a plurality of input audio signals;

[0320] determining a first beamformed signal based on the plurality of input audio signals;

[0321] determining a second beamformed signal based on the plurality of input audio signals;

[0322] determining a first audio parameter based on the plurality of input audio signals;

[0323] determining a first detector output signal based on the determined first audio parameter;

[0324] determining a first anti-noise signal related to a first type of noise based on the second beamformed signal and the first detector output signal;

[0325] determining a second anti-noise signal related to a second type of noise based on the second beamformed signal;

[0326] determining an output signal based on the first beamformed signal, the first anti-noise signal and the second anti-noise signal;

[0327] outputting the output signal.

[0328] Item 36. The method of operating a hearing aid according to item 35, wherein the first beamformed signal and the second beamformed signal are determined by fixed filters and the first anti-noise signal and the second anti-noise signal are determined by adaptive filters.

[0329] Item 37. The method of operating a hearing aid according to item 35 or 36, wherein the first type of noise is a slow time varying noise and the second type of noise is a fast time varying noise.

[0330] Item 38. The method of operating a hearing aid according to any of items 35-37, wherein determining the first detector output signal comprises comparing the first audio parameter to a first threshold,

[0331] if the first audio parameter exceeds the first threshold, determining the first detector output signal comprises an activation signal allowing the first noise canceller to adapt to the second beamformed signal; and

[0332] if the first detector output signal does not exceed the first threshold, determining the first detector output signal comprises a fixation signal for fixing the first noise canceller.

[0333] Item 39. The method of operating a hearing aid according to any of items 35-38, comprising:

[0334] determining a second audio parameter based on the plurality of input audio signals;

[0335] determining a second detector output signal based on the determined second audio parameter; and

[0336] determining the second anti-noise signal based on the second detector signal and the second detector output signal.

[0337] Item 40. The method of operating a hearing aid according to any of items 35-39, wherein determining the second detector output signal comprises comparing the second audio parameter to a second threshold,

[0338] if the second audio parameter exceeds the second threshold, determining the second detector output signal comprises an activation signal allowing the second noise canceller to adapt to the second beamformed signal; and

[0339] if the second detector output signal does not exceed the second threshold, determining the second detector output signal comprises a fixation signal for fixing the second noise canceller.

[0340] Item 41. The method of operating a hearing aid according to item 39 or 40, wherein the second audio parameter is an interference source activity parameter.

[0341] Item 42. The method of operating a hearing aid according to any of items 35-41, wherein the first audio parameter is a speech activity parameter.

[0342] Item 43. The method of operating a hearing aid according to any one of items 35-42, comprising:

[0343] determining a first control gain;

[0344] determining a first noise-controlled signal by applying the first control gain to the first anti-noise signal;

[0345] determining an output signal based on the first noise-controlled signal.

[0346] Item 44. The method of operating a hearing aid according to item 43, wherein the first control gain is determined as a predetermined gain.

[0347] Item 45. The method of operating a hearing aid according to item 43, wherein determining the first control gain comprises determining the first control gain as an adaptive gain determined based on the first detector output signal.

[0348] Item 46. The method of operating a hearing aid according to any one of items 35-45, comprising:

[0349] determining a second control gain;

[0350] determining a second noise-controlled signal by applying the second control gain to the second anti-noise signal;

[0351] wherein the output interface is configured to determine the output signal based on the second noise-controlled signal.

[0352] Item 47. The method of operating a hearing aid according to item 46, wherein the second control gain is determined as a predetermined gain.

[0353] Item 48. The method of operating a hearing aid according to item 46, wherein determining the second control gain comprises determining the second control gain as an adaptive gain determined based on the second detector output signal.

[0354] Item 49. The method of operating a hearing aid according to any one of items 35-48, comprising:

[0355] determining a third audio parameter based on the plurality of input audio signals;

[0356] determining a third detector output signal based on the determined third audio parameter;

[0357] determining a third anti-noise signal related to a third type of noise based on the third beamformed signal and the third detector output signal;

[0358] determining the output signal based on the third anti-noise signal.

[0359] Item 50. The method of operating a hearing aid according to item 49, wherein the third audio parameter is an audio feedback parameter.

[0360] Item 51. The method of operating a hearing aid according to item 49 or 50:

[0361] determining a third control gain;

[0362] determining a third noise-controlled signal by applying the third control gain to the third anti-noise signal;

[0363] determining the output signal based on the third noise-controlled signal.

[0364] Item 52. The method of operating a hearing aid according to item 51, wherein the third control gain is determined as a predetermined gain.

[0365] Item 53. The method of operating a hearing aid according to item 51, wherein determining the third control gain comprises determining the third control gain as an adaptive gain determined based on the third detector output signal.

[0366] Item 54. The method of operating a hearing aid according to any of items 35-53, wherein the first beamformed signal is determined according to an undistorted criterion.

[0367] Item 55. The method of operating a hearing aid according to any of items 35-54, wherein the second beamformed signal is determined by minimizing a power of a target based on the plurality of input audio signals.

[0368] Item 56. A data processing system comprising a processor and a program code for causing the processor to perform the method according to any of items 35-55.

[0369] Item 57. A computer program product comprising a computer program containing instructions which, when the program is executed by a computer, cause the computer to carry out the method according to any of items 35-55.

Claims

1. A hearing aid, comprising: an input interface configured to provide a plurality of input audio signals; a first beamformer configured to receive a first beamformer input signal based on the plurality of input audio signals, and to determine a first beamformed signal based on the first beamformer input signal; a second beamformer configured to receive a second beamformer input signal based on the plurality of input audio signals, and to determine a second beamformed signal based on the second beamformer input signal; a first detector configured to receive a first detector input signal based on the plurality of input audio signals, to determine a first audio parameter based on the first detector input signal, to determine a first detector output signal based on the determined first audio parameter, and to output the first detector output signal; a first noise canceller configured to receive a first noise canceller input signal based on the second beamformed signal and the first detector output signal, to determine a first anti-noise signal based on the first noise canceller input signal, and to output the first anti-noise signal, wherein the first anti-noise signal is an anti-noise signal for a first type of noise; a second noise canceller configured to receive a second noise canceller input signal based on the second beamformed signal, to determine a second anti-noise signal based on the second beamformed signal, and to output the second anti-noise signal, wherein the second anti-noise signal is an anti-noise signal for a second type of noise; an output interface configured to determine an output signal based on the first beamformed signal, the first anti-noise signal, and the second anti-noise signal, and to output the output signal.

2. The hearing aid according to claim 1, wherein the first type of noise is a slow time varying noise, and the second type of noise is a fast time varying noise.

3. The hearing aid according to any one of the preceding claims, wherein determining the first detector output signal comprises comparing the first audio parameter to a first threshold, if the first audio parameter exceeds the first threshold, determining the first detector output signal to comprise an activation signal, which activation signal allows the first noise canceller to adapt by reducing an error between the first anti-noise signal and a first target signal based on the first beamformed signal; if the first audio parameter does not exceed the first threshold, determining the first detector output signal to comprise a fixation signal, which fixation signal is used to fix the first noise canceller.

4. The hearing aid according to any one of the preceding claims, wherein the first audio parameter is a speech activity parameter.

5. The hearing aid according to any one of the preceding claims, comprising a second detector configured to receive a second detector input signal based on the plurality of input audio signals, to determine a second audio parameter based on the second detector input signal, to determine a second detector output signal based on the determined second audio parameter, and to output the second detector output signal; wherein the second noise canceller being configured to receive the second detector output signal, and to determine the second anti-noise signal based on the second detector output signal.

6. The hearing aid according to claim 5, wherein the second audio parameter is an interference activity parameter.

7. The hearing aid according to any one of the preceding claims, comprising: a first gain controller configured to receive a first anti-noise signal, determine a first control gain, apply the first control gain to the first anti-noise signal, and output a first noise-controlled signal; wherein the output interface is configured to determine the output signal based on the first noise-controlled signal.

8. The hearing aid according to claim 7, wherein the first gain controller is configured to receive a first detector output signal, and determine the first control gain based on the first detector output signal.

9. The hearing aid according to any of the preceding claims, comprising: a second gain controller configured to receive a second anti-noise signal, determine a second control gain, apply the second control gain to the second anti-noise signal, and output a second noise-controlled signal; wherein the output interface is configured to determine the output signal based on the second noise-controlled signal.

10. The hearing aid according to any of the preceding claims, comprising: a third detector configured to receive a third detector input signal based on the plurality of input audio signals, determine a third audio parameter based on the third detector input signal, determine a third detector output signal based on the determined third audio parameter, and output the third detector output signal; a third noise canceller configured to receive a third noise canceller input signal based on the second beamformed signal and the third detector output signal, determine a third anti-noise signal based on the third noise canceller input signal, and output the third anti-noise signal, wherein the third anti-noise signal is an anti-noise signal for a third type of noise; wherein the output interface is configured to determine the output signal based on the third anti-noise signal.

11. The hearing aid according to claim 9, wherein the third audio parameter is an audio feedback parameter.

12. The hearing aid according to any of the claims 1-4, wherein the first noise canceller comprises a first smoothing factor, the second noise canceller comprises a second smoothing factor, the first and second smoothing factors define an adaptation rate of the respective noise canceller, and the adaptation rate caused by the first smoothing factor is lower than the adaptation rate caused by the second smoothing factor.

13. A method of operating a hearing aid, comprising: providing a plurality of input audio signals; determining a first beamformed signal based on the plurality of input audio signals; determining a second beamformed signal based on the plurality of input audio signals; determining a first audio parameter based on the plurality of input audio signals; determining a first detector output signal based on the determined first audio parameter; determining a first anti-noise signal related to a first type of noise based on the second beamformed signal and the first detector output signal; determining a second anti-noise signal related to a second type of noise based on the second beamformed signal; determining an output signal based on the first beamformed signal, the first anti-noise signal, and the second anti-noise signal; outputting the output signal.

Images