Abstract: Techniques are described for generating parallel data for real-time speech form conversion. In an embodiment, based at least in part on input speech data of an original form, a speech machine learning (ML) model generates parallel speech data. The parallel speech data includes the input speech data of the original form and temporally aligned output speech data of a target form different than the original form. Each frame of the input speech data temporally corresponds to the corresponding output speech frame of the target speech form and contains a same portion of the particular content. The techniques further include training a teacher machine learning model that is offline and is substantially larger than a student machine learning model for converting speech form. Transferring “knowledge” from the trained Teacher model for training the Production Student Model that performs the speech form conversion on an end-user computing device.

Abstract: A method of equalising an audio signal derived from a microphone, the method comprising: receiving the audio signal; applying an order-statistic filter to the audio signal in the frequency domain to generate a statistically filtered audio signal; equalising the received audio signal based on the statistically filtered audio signal to generate an equalised audio signal.

Abstract: An apparatus, method and computer program is disclosed. The apparatus may comprise a means comprising at least one processor and at least one memory including computer program code, the at least one memory and computer program code configured to, with the at least one processor, to receive multimedia data representing a scene, the multimedia data comprising at least audio data representing an audio component of the scene. Another operation may comprise determining a location of unwanted sound in the scene. Another operation may comprise performing first audio processing to remove at least part of the unwanted sound from the determined location. Another operation may comprise performing second audio processing to add artificial sound associated to the unwanted sound at the determined location.

Abstract: A system and method for processing an audio input signal includes a microphone, a controller, and a communication link that may be coupled to a remote speaker. The microphone captures the audio input signal and communicates the audio input signal to the controller, and the controller is coupled to the communication link. The controller includes executable code to generate, via a linear noise reduction filtering algorithm, a first resultant based upon the audio input signal, and generate, via non-linear post filtering algorithm, a second resultant based upon the first resultant. An audio output signal is generated based upon the second resultant employing a feature restoration algorithm. The audio output signal is communicated, via the communication link, to a speaker that may be at a remote location.

Abstract: A system that suppresses a plurality of interferences of different types in a received audio signal. The system comprises one or more microphones and an audio controller. The one or more microphones are configured to detect the audio signal. The audio controller applies an interference estimation algorithm to the audio signal to generate an attenuation coefficient for each of the plurality of interferences of different types. The audio controller applies the attenuation coefficients to the audio signal to generate an interference-suppressed audio signal in which the plurality of interferences of different types is suppressed. The audio controller determines a time domain signal based on the interference-suppressed audio signal to provide to an end user.

Abstract: Examples of the disclosure relate to a method, apparatus and computer program, the method including: obtaining audio signals wherein the audio signals represent spatial sound and can be used to render spatial audio using linear methods; obtaining spatial metadata corresponding to the spatial sound represented by the audio signals; and associating the spatial metadata with the obtained audio signals so that in a first rendering context the obtained audio signals can be rendered without using the spatial metadata and in a second rendering context the obtained audio signals can be rendered with using the spatial metadata.

Abstract: Disclosed are systems and methods for processing an audio signal. In particular, there is provided a method for determining dynamic gain values to be applied on a digital input signal. The digital signal may be arranged in blocks. The dynamic gain values may be used for attenuating input signal values exceeding a clipping threshold. More particularly, the method comprising, for each signal block, passing backwards over the next signal block and the current signal block to produce a preliminary gain contour from the input signal; and passing forwards over the current signal block to produce a final gain contour for the current signal block based on the preliminary gain contour, wherein the gain contours are produced by applying an instant gain ascent and a smooth gain decay to the gain contours.

Abstract: In one implementation, a circuit can include a reference pin and an operational amplifier that can include an output pin, an inverting input pin and a non-inverting input pin. The inverting input pin can be electrically coupled to the output pin via a first impedance and to the reference pin via a second impedance. The non-inverting input pin can be electrically coupled to the reference pin via a third impedance and can be configured to receive a detection signal. The reference pin can be configured to receive a detection reference signal associated with the detection signal.

Abstract: Methods and apparatuses for watermark embedding and extracting are provided. A method for watermark extracting includes obtaining a carrier object embedded with a watermark image; determining at least one encoding region that includes watermark information; determining a plurality of template lattices from the at least one encoding region, the plurality of template lattices comprising a plurality of positioning template lattices and one or more encoding template lattices; and obtaining the watermark information according to the plurality of template lattices.

Abstract: A hearing device includes: at least one microphone for converting sound received by the at least one microphone into an audio signal; a sound impulse detector configured for detecting a presence of an impulse in the audio signal; and a signal processor configured for processing the audio signal into a processed audio signal in response to the presence of the impulse in the audio signal as detected by the sound impulse detector; and a receiver coupled to the signal processor for converting the processed audio signal into an output sound signal for emission towards an eardrum of a user; wherein the sound impulse detector is configured for operation in a frequency domain for detecting presence of the impulse in the audio signal.

Abstract: A method at a wearable electronic device with: a first electro-acoustic input transducer and a second electro-acoustic input transducer arranged to pick up a first acoustic signal and convert the first acoustic signal to a first microphone signal and second microphone signal; and a third electro-acoustic input transducer arranged to pick up a second acoustic signal and convert the second acoustic signal to a third microphone signal; and a processor (140).

Abstract: A noise suppression device includes: a DFT executor that expands a baseband signal into a discrete Fourier series X0(n), the baseband signal being generated by mixing an AM broadcast wave signal including a carrier wave of the angular frequency ?C with a complex sine wave of the same frequency; and an amplitude spectrum calculator that calculates an amplitude spectrum |X0(n)| from X0(n). The noise suppression device also includes: an asymmetric component detector that detects an asymmetric component in |X0(n)|; a suppressor that calculates a discrete Fourier series X1(n) by multiplying the value corresponding to the asymmetric frequency bin by a first factor and multiplying the other values by a second factor in X0(n); and an IDFT executor that performs inverse discrete Fourier transform on X1(n) to obtain a discrete-time signal.

Abstract: An image capture device with dynamic wind noise compression tuning techniques is described. A technique includes detecting of the presence of wind noise by measuring coherence between at least two microphones. For a compressor, adjusting a default compression threshold and default compression parameters based on the coherence measurements. For each microphone, applying by the compressor the adjusted compression parameters when an audio signal is above the adjusted compression threshold and applying the default compression parameters when the audio signal is below the adjusted compression threshold.

Abstract: A method for providing IQ mismatch (IQMM) compensation includes: sending a single tone signal at an original frequency; determining a first response of an impaired signal at the original frequency and a second response of the impaired signal at a corresponding image frequency; determining an estimate of a frequency response of the compensation filter at the original frequency based on the first response and the second response; repeating the steps of sending the single tone signal, determining the first response and the second response, and determining the estimate of the frequency response of the compensation filter by sweeping the single tone signal at a plurality of steps to determine a snapshot of the frequency response of the compensation filter; converting the frequency response of the compensation filter to a plurality of time-domain filter taps of the compensation filter by performing a pseudo-inverse of a time-to-frequency conversion matrix; and determining a time delay that provides a minimal LSE fo

Abstract: Disclosed is a signal processor for headphone off-ear detection. The signal processor includes an audio output to transmit an audio signal toward a headphone speaker in a headphone cup. The signal processor also includes a feedback (FB) microphone input to receive a FB signal from a FB microphone in the headphone cup. The signal processor also includes an off-ear detection (OED) signal processor to determine an audio frequency response of the FB signal over an OED frame as a received frequency response. The OED processor also determines an audio frequency response of the audio signal times an off-ear transfer function between the headphone speaker and the FB microphone as an ideal off-ear response. A difference metric is generated comparing the received frequency response to the ideal off-ear frequency response. The difference metric is employed to detect when the headphone cup is disengaged from an ear.

Abstract: A converter includes a time window cut-out block, a fast Fourier transform (FFT) block, an attenuation amount limitation block, a quantization noise attenuation block, an overtone generation block, an inverse fast Fourier transform (IFFT) block, and a time window resynthesis block. The attenuation amount limitation block determines the maximum attenuation amount of quantization noise to be attenuated in the quantization noise attenuation block based on the magnitude of a signal level of sound data supplied from the time window cut-out block. The quantization noise attenuation block adjusts amplitude in a frequency domain based on the maximum attenuation amount determined by the attenuation amount limitation block, to attenuate the quantization noise.

Abstract: An adaptive beam-forming array uses multiple sensors and noise reference subtraction to reduce noise at an output of the adaptive beam-forming array. A direction of arrival of energy from a desired source is determined and an inter-sensor noise correlation between one or more pairs of sensors is determined. An Adaptive Blocking Matrix (ABM) generates a noise reference from an inter-sensor model representing a relationship between desired signal components received from the desired source and that are present in signals from one or more pairs of sensors. The noise reference is generated with an adaptive filter that filters a first signal from a first sensor in the pairs of sensors and is combined with the second signal from a second sensor in the pairs of sensors to produce the noise reference. The adaptive filter is initialized with an initialization response computed from the direction of arrival and the inter-sensor noise correlation.

Abstract: Systems and methods are disclosed for reducing unwanted noise during image capture. The noise may be airborne or structure-borne. For example, airborne sound may be sound that is emitted from a motor of a motorized gimbal into the air, which is then detected by a microphone of an imaging device along with the desired sound. Structure-borne noise may include vibrations from the motor that reach the microphone. Structure-borne noise may lead to local acoustic pressure variation by the microphone or pure vibration of the microphone.

Abstract: Detection of a blocked microphone involves receiving microphone signals from a plurality of microphones. A plurality of signal feature measures are derived from the microphone signals. The signal feature measures are normalised. The normalised signal feature measures are variably weighted in response to detected environmental conditions in the microphone signals. The variably weighted normalised signal feature measures are combined to produce an output indication of whether a microphone is blocked.

Abstract: A method for improving the robustness of a speech processing system having at least one speech processing module comprises: receiving an input sound signal comprising audio and non-audio frequencies; separating the input sound signal into an audio band component and a non-audio band component; and identifying possible interference within the audio band from the non-audio band component. Based on such an identification, the operation of a downstream speech processing module is adjusted.

Abstract: An audio appliance can include a microphone transducer configured to receive sound from an environment and to convert the received sound into an audio signal and a display. The audio appliance can include an audio analytics module configured to detect an audio-input impairment by analyzing the audio signal and output a detection signal identifying the audio-input impairment in real-time. The audio-input impairment can include, for example, a poor-intelligibility impairment, a microphone-occlusion impairment, a handling-noise impairment, a wind-noise impairment, or a distortion impairment. The audio appliance can also include an impairment module configured to identify and emit a user-perceptible alert corresponding to the identified audio-input impairment in real-time; and an interactive guidance module configured to present a suggested action to address the audio-input impairment in real-time. Related aspects also are described.

Abstract: Under one aspect, a method is provided for processing a received signal, the received signal including a desired signal and an interference signal that spectrally overlaps the desired signal. The method can include obtaining an amplitude of the received signal. The method also can include obtaining an average amplitude of the received signal based on at least one prior amplitude of the received signal. The method also can include subtracting the amplitude from the average amplitude to obtain an amplitude residual. The method also can include, based upon an absolute value of the amplitude residual being less than or equal to a first threshold, inputting the received signal into an interference suppression algorithm so as to generate a first output including the desired signal with reduced contribution from the interference signal.

Abstract: Earpieces and methods for acute sound detection and reproduction are provided. A method can include measuring an external ambient sound level (xASL), monitoring a change in the xASL for detecting an acute sound, estimating a proximity of the acute sound, and upon detecting the acute sound and its proximity, reproducing the acute sound within an ear canal, where the ear canal is at least partially occluded by an earpiece. Other embodiments are disclosed.

Abstract: Disclosed is a signal processing apparatus including a surrounding sound signal acquisition unit, a NC (Noise Canceling) signal generation part, a cooped-up feeling elimination signal generation part, and an addition part. The surrounding sound signal acquisition unit is configured to collect a surrounding sound to generate a surrounding sound signal. The NC signal generation part is configured to generate a noise canceling signal from the surrounding sound signal. The cooped-up feeling elimination signal generation part is configured to generate a cooped-up feeling elimination signal from the surrounding sound signal. The addition part is configured to add together the generated noise canceling signal and the cooped-up feeling elimination signal at a prescribed ratio.

Abstract: A howling detection method is provided. A window separation processing is processed on an audio signal to obtain a plurality of analysis windows. A signal energy indicator value of each preset frequency in at least one analysis window is obtained by using a preset perceptual coefficient corresponding to each frequency, to obtain a perceptual energy indicator value of each frequency, the preset perceptual coefficient corresponding to each frequency indicating a sensitivity of a human ear to a sound of each frequency. It is determined whether howling occurs according to the perceptual energy indicator value of each frequency in the at least one analysis window.

Abstract: A method consistent with certain implementation involves presenting a graphical user interface (GUI) to a user on a display, where the GUI presents a visual representation of a room that is adapted to be adjusted in size and shape by user manipulation of a controller; the GUI has a drop and drag menu adapted to selection of an object from a plurality of objects for placement at any selected position within the room; at least one of the objects comprising a loudspeaker; and where the GUI provides for input of data characterizing the loudspeaker. This abstract is not to be considered limiting, since other embodiments may deviate from the features described in this abstract.

Abstract: A system and method that enhances spoken utterances by capturing one or more microphone signals. The system and method estimates a plurality of echo paths from each of the one or more microphone signals and synthesizes a speech reinforcement signal in response to and corresponding to the one or more microphone signals. The system and method concatenates portions of the synthesized reinforcement signal with the captured microphone signals and processes the captured microphone signals in response to the estimated plurality of echo paths.

Abstract: A system for classifying impulsive noise on a communications signal comprises an impulse signal generator, an integrator, a first comparator, and an impulse peak detector. The impulse signal generator receives a communications signal that includes impulsive noise and is configured to provide an impulse signal that includes just the impulsive noise. The integrator receives the impulse signal and integrates the impulse signal to determine the power of the impulse signal. The first comparator receives the impulse signal and is configured to compare the impulse signal to a first reference signal and indicate the time during which the value of the impulse signal is greater than the value of the first reference signal. The impulse peak detector receives the impulse signal and is configured to process the impulse signal, compare the processed signal to a second reference signal, and detect the peak value of the impulse signal.

Abstract: A user speech interface for interactive media guidance applications, such as television program guides, guides for audio services, guides for video-on-demand (VOD) services, guides for personal video recorders (PVRs), or other suitable guidance applications is provided. Voice commands may be received from a user and guidance activities may be performed in response to the voice commands.

Abstract: A noise suppressor has a band extractor to separate signal by frequency band; and per-band units for each of band including noise estimator and SNR computation units. The per-band unit has a histogrammer to give histograms of current and past SNRs, and a gain-curve updater computes gain curves from the histogram. Gain curves are used to determine raw gains from current SNRs, raw gain is filtered and controls a variable gain unit to provide band-specific gain-adjusted, signals that are recombined into a noise-reduced frequency-domain output. Raw gain filtering may include finite-impulse-response filtering and weighted averaging of intermediate gains of a current and adjacent-band per-band unit. The method includes separating an input into frequency bands, estimating in-band noise, and deriving a band SNR. Then, histogramming the SNR and updating a gain curve from the histogram, and finding a raw gain using the gain curve and current SNR.

Abstract: In some embodiments, a method includes receiving, at a device, digital audio content to be converted by a digital-to-analog converter to produce analog audio content. The digital audio content has at least one audible frequency. The method also includes selecting, at the device, a first signal and a second signal to output with the analog audio content. The first signal has an inaudible carrier frequency and a bandwidth, and the second signal also has an inaudible carrier frequency and a bandwidth. A difference between the first signal and the second signal is an undesired audible signal. The method further includes outputting, from the device, the analog audio content, the first signal and the second signal, such that when the analog audio content is detected by a recording device the undesired audible signal is detected with the analog audio content.

Abstract: A method comprising determining an audio noise spectrum that represents external audio noise, determining an audio reference spectrum that represents audio content to be perceived by a listener, and determining a compensation spectrum based on the audio noise spectrum and based on the audio reference spectrum.

Abstract: A voice control device has a speech command recognizer, a sensor data processor and a decision making circuit. The speech command recognizer is arranged for performing speech command recognition to output a recognized speech command. The sensor data processor is arranged for processing sensor data generated from at least one auxiliary sensor to generate a detection output. The decision making circuit is arranged for deciding a response of the voice control device according to the recognized speech command and the detection output. The same speech command is able to trigger difference responses according to the detection output (e.g., detected motion). Besides, an adaptive training process may be employed to improve the accuracy of the sensor data processor. Hence, the voice control device may have improved performance of the voice control feature due to a reduce occurrence probability of miss errors and false alarm errors.

Abstract: Loudness control systems or methods may normalize audio signals to a predetermined loudness level. If the audio signal includes moderate background noise, then the background noise may also be normalized to the target loudness level. Noise signals may be detected using content-versus-noise classification, and a loudness control system or method may be adjusted based on the detection of noise. Noise signals may be detected by signal analysis in the frequency domain or in the time domain. Loudness control systems may also produce undesirable audio effects when content shifts from a high overall loudness level to a lower overall loudness level. Such loudness drops may be detected, and the loudness control system may be adjusted to minimize the undesirable effects during the transition between loudness levels.

Abstract: The application relates to a hearing device comprising a beamformer of the generalized sidelobe canceler (GSC) type. The application further relates to a method of operating a hearing device. The disclosure addresses a problem which occurs when using a GSC structure in a hearing device application. The problem arises due to a non-ideal target-cancelling beamformer. As a consequence, a target signal impinging from the look direction can—unintentionally—be attenuated by as much as 30 dB. To resolve this problem, it is proposed to monitor the difference between the output signals from the all-pass beamformer and the target-cancelling beamformer to control a time-varying regularization parameter in the GSC update. This has the advantage of providing a computationally simple solution to the non-ideality of the GSC beamformer. The invention may e.g. be used in hearing aids, headsets, ear phones, active ear protection systems, or combinations thereof.

Abstract: A speech recognition apparatus and a speech recognition method are provided. In the invention, whether an original voice sampling signal corresponding to a target voice frame is a consonant signal is determined according to at least one of a ratio of an energy of a low-pass sampling signal to an energy of the original voice sampling signal and a ratio value of an energy of a second consonant frequency band signal.

Abstract: A speech recognition apparatus and a speech recognition method are provided. In the invention, whether an original voice sampling signal corresponding to a target voice frame is a noise signal is determined according to a ratio of an energy of a first consonant frequency band signal to an energy of a second consonant frequency band signal, a ratio of an energy of the first consonant frequency band signal to an energy of the original voice sampling signal and a ratio of an energy of the second consonant frequency band signal to an energy of the original voice sampling signal.

Abstract: In accordance with an embodiment of the present invention, a noise reduction method for speech processing includes detecting if two signals from two microphones are so close to each other in non voice area that the two microphones are equivalent to Single-Microphone for noise/interference reduction processing. Single-Microphone noise/interference reduction processing algorithm is selected if the equivalent Single-Microphone is detected; Multiple-Microphone noise/interference reduction processing algorithm is selected if the equivalent Single-Microphone is not detected.

Abstract: An exemplary system for facilitating binaural hearing by a cochlear implant patient includes 1) a spectral analysis facility configured to divide a first audio signal presented to a first ear of the patient and a second audio signal presented to a second ear of the patient into first and second sets of analysis channels, respectively, and 2) a processing facility configured to process acoustic content contained in a first analysis channel included in the first set of analysis channels and acoustic content contained in a second analysis channel included in the second set of analysis channels, mix the processed acoustic content contained in the first and second analysis channels, and direct a cochlear implant to apply electrical stimulation representative of the mixed acoustic content to the first ear by way of a stimulation channel that corresponds to the first analysis channel.

Abstract: Disclosed is a distortion compensation apparatus that, by appropriately generating a distortion-compensation coefficient, makes it possible to obtain a desired transmission output, and substantially reduce the amount of power leakage to an adjacent channel. Reception section (103) of the distortion compensation apparatus acquires and demodulates a transmission signal to generate a demodulation signal. Delay adjustment section (104) computes the delay amount of the demodulation signal with respect to the baseband signal, eliminates the delay of the demodulation signal with respect to the baseband signal based on the delay amount, and outputs the baseband signal and the demodulation signal in which the delay is eliminated.

Abstract: Disclosed is a video controlled beam steering mechanism for an adaptive filter in a sensor array system that receives input from a target and applies an averaging filter and appropriately steers the beam. An adaptive filter is then used if the SNR of the output of the averaging filter reaches a threshold.

Abstract: An information handling system has a haptic generation module to generate haptic effects including haptic noise and a haptic noise reduction module. The haptic noise reduction module receives characteristics of sound representative of haptic noise generated by a haptic generation module of a device and entering an audio input module of the device, the characteristics including frequencies and timing. It also detects the generation of haptic effects, the generations occurring after the receiving characteristics. It also reduces the effects of haptic noise on digital data representing audio input to the device based upon the received characteristics of the sound. It may reduce the effects by subtracting amplitudes of audio waves representing the haptic noise from amplitudes of audio waves representing the audio input.

Abstract: A system and method for processing close talking differential microphone array (CTDMA) signals in which incoming microphone signals are transformed from time domain signals to frequency domain signals having separable magnitude and phase information. Processing of the frequency domain signals is performed using the magnitude information, following which phase information is reintroduced using phase information of one of the original frequency domain signals. As a result, high pass filtering effects of conventional differential signal processing of CTDMA signals are substantially avoided.

Abstract: An audio processing apparatus includes an acquisition unit configured to acquire an audio signal, and an audio processing unit configured to reduce noise contained in the audio signal, wherein the audio processing unit complements an audio signal in a section containing noise of the audio signal with a signal generated based on an audio signal in a predetermined section before the section containing noise and an audio signal in a predetermined section after the section containing noise, and wherein, in a case where noise is contained in one of the audio signal in the predetermined section before the section containing noise and the audio signal in the predetermined section after the section containing noise, the audio processing unit complements the audio signal in the section containing noise with a signal generated based on the audio signal in a noise-free section.

Abstract: The present technology provides robust, high quality expansion of the speech within a narrow bandwidth acoustic signal which can overcome or substantially alleviate problems associated with expanding the bandwidth of the noise within the acoustic signal. The present technology carries out a multi-faceted analysis to accurately identify noise within the narrow bandwidth acoustic signal. Noise classification information regarding the noise within the narrow bandwidth acoustic signal is used to determine whether to expand the bandwidth of the narrow bandwidth acoustic signal. By expanding the bandwidth based on the noise classification information, the present technology can expand the speech bandwidth of the narrow bandwidth acoustic signal and prevent or limit the bandwidth expansion of the noise.

Abstract: An apparatus for providing one or more adjusted parameters for a provision of an upmix signal representation on the basis of a downmix signal representation and a parametric side information associated with the downmix signal representation has a parameter adjuster. The parameter adjuster is configured to receive one or more parameters and to provide, on the basis thereof, one or more adjusted parameters. The parameter adjuster is configured to provide the one or more adjusted parameters in dependence on an average value of a plurality of parameter values, such that a distortion of the upmix signal representation caused by the use of non-optimal parameters is reduced at least for parameters deviating from optimal parameters by more than a predetermined deviation.

Abstract: A differential microphone array (DMA) is provided that includes a number (M) of microphone sensors for converting a sound to a number of electrical signals and a processor that is configured to apply linearly-constrained minimum variance filters on the electrical signals over a time window to calculate frequency responses of the electrical signals over a plurality of subbands and sum the frequency responses of the electrical signals for each subband to calculate an estimated frequency spectrum of the sound.

Abstract: Provided are an imaging device, program, memory medium, and noise reduction method capable of appropriately reducing noise without causing degradation in a target sound such as voice. The imaging device of the present invention has: a video imaging unit for capturing video; a signal converter for converting a sound generated during video capture to a sound signal; subject determination units that predict or recognize a specific subject; a noise detector for detecting noise included in the sound generated during video capture; a noise reduction unit for reducing the noise signal from the sound signal; a voice detector for detecting non-noise signals in the sound signal; and a noise reduction performance change unit that lowers the noise signal reduction performance of the noise reduction unit when the subject determination units predict or recognize the specific subject.

Abstract: Measurement signals for calculating acoustic characteristics of the acoustic space are reproduced in a plurality of periods at sound reproduction intervals. A picked-up signal is acquired by picking up a reproduced signal. The picked-up signal is divided for each period, and the acoustic characteristics of the acoustic space are calculated from an arithmetic unit of the divided periodic signals and the measurement signal. Before sound reproduction, a background noise signal in the acoustic space is measured, and a peak frequency component of a frequency characteristic of the background noise signal is detected. The number of periods and sound reproduction intervals of measurement signals to be reproduced are controlled so as to cancel out a detected peak frequency component at the time of calculating arithmetic unit.

Abstract: A novel beamforming post-processor technique with enhanced noise suppression capability. The present beamforming post-processor technique is a non-linear post-processing technique for sensor arrays (e.g., microphone arrays) which improves the directivity and signal separation capabilities. The technique works in so-called instantaneous direction of arrival space, estimates the probability for sound coming from a given incident angle or look-up direction and applies a time-varying, gain based, spatio-temporal filter for suppressing sounds coming from directions other than the sound source direction, resulting in minimal artifacts and musical noise.