Abstract: A computer-implemented method for processing audio input data into processed audio data by using an audio device comprising a microphone, a processor device and a memory holding a plurality of neural networks is presented. The plurality of neural networks are associated with different room types, wherein each room type is associated with one or more reference room acoustic metrics. The method comprises obtaining, by the microphone, room response data, wherein the room response data is reflecting room acoustics of a room in which the audio device is placed, determining, by using the processor device, the one or more room acoustic metrics based on the room response data, and selecting, by using the processor device, a matching neural network among the plurality of neural networks by comparing the one or more room acoustic metrics with the one or more reference room acoustic metrics associated with the different room types associated with the plurality of neural networks.

Abstract: A computer-implemented method (600) for training a neural network (100) by using a data processing device (102) is presented.

Abstract: A speakerphone is disclosed. The speakerphone comprises an interface, a speaker, a plurality of microphones, a processor, and a memory.

Abstract: A speakerphone is disclosed. The speakerphone comprises an interface, a speaker, one or more microphones including a first microphone, a processor, and a memory. The speakerphone is configured to obtain an internal output signal for provision of an internal audio output signal in an environment. The speakerphone is configured to output the internal audio output signal in the environment. The speakerphone is configured to obtain a microphone input signal. The speakerphone is configured to determine an impulse response associated with the environment. The speakerphone is configured to determine one or more environment parameters indicative of acoustics of the environment. The speakerphone is configured to determine an environment score indicative of suitability of a conference setup in the environment. The speakerphone is configured to output the environment score.

Abstract: A speakerphone is disclosed. The speakerphone comprises an interface, a speaker, one or more microphones including a first microphone, a processor, and a memory. The speakerphone is configured to obtain an internal output signal for provision of an internal audio output signal in an environment. The speakerphone is configured to output the internal audio output signal in the environment. The speakerphone is configured to obtain a microphone input signal. The speakerphone is configured to determine an impulse response associated with the environment. The speakerphone is configured to determine one or more environment parameters indicative of acoustics of the environment. The speakerphone is configured to transmit the impulse response and/or the first environment parameter to a server device.

Abstract: A method comprising: at an electronic device (100) having an array microphones (101) with a plurality of microphones and a processor (102): receiving a plurality of microphone signals (x1, x2, x3) from the plurality of microphones; generating a processed signal (XP) from the plurality of microphone signals using one or both of beamforming and deconvolution; generating a compensated processed signal (XO) by compensating the processed audio signal (XP) in accordance with compensation coefficients (Z). Generating the compensated processed signal comprises: generating first spectrum values (PXP) from the processed audio signal; generating reference spectrum values (<PX>) from multiple second spectrum values (PX1, PX2, PX3) generated from each of at least two of the microphone signals in the plurality of microphone signals (x1, x2, x3); and generating the compensation coefficients (Z) from the reference spectrum values (<PX>) and the first spectrum values (PXP).

Abstract: A conference device and a computer-implemented method for training a neural network are disclosed, the conference device comprising a conference controller; a microphone array comprising a plurality of microphones for provision of audio signals representing audio from one or more sound sources; a direction estimator connected to the conference controller and the microphone array, the direction estimator configured to obtain, from the microphone array, a plurality of audio signals including a first audio signal and a second audio signal; determine direction data based on the plurality of audio signals, the direction data comprising an indication of an estimated probability of voice activity for one or more directions, wherein to determine direction data comprises to apply an offline-trained neural network; and output audio data based on the direction data to the conference controller.

Abstract: The present disclosure relates to a microphone apparatus and an associated computer implemented method. The microphone apparatus comprising a main microphone array, an adaptive beamformer, a fixed beamformer, and an analyzer. The analyzer is configured to determine a first relative score based on the output of the fixed beamformer and the adaptive beamformer. The first relative score indicating a difference between the adaptive beamformer and the fixed beamformer.

Abstract: An audio device and related methods for speech quality detection are disclosed, the audio device comprising an interface, a processor, a memory and one or more microphones, wherein the audio device is configured to obtain, via the interface, a microphone input signal from one or more microphones including a first microphone; process the microphone input signal for provision of an output signal; determine, using a non-intrusive quality detection model, one or more quality parameters including a first quality parameter indicative of a speech quality associated with the output signal; control processing of the microphone input signal based on the first quality parameter; and transmit, via the interface, the output signal.

Abstract: A method for speech extraction in an audio device is disclosed. The method comprises obtaining a microphone input signal from one or more microphones including a first microphone. The method comprises applying an extraction model to the microphone input signal for provision of an output. The method comprises extracting a near speaker component in the microphone input signal according to the output of the extraction model being a machine-learning model for provision of a speaker output. The method comprises outputting the speaker output.

Abstract: A conference device and a computer-implemented method for training a neural network are disclosed, the conference device comprising a conference controller; a microphone array comprising a plurality of microphones for provision of audio signals representing audio from one or more sound sources; a direction estimator connected to the conference controller and the microphone array, the direction estimator configured to obtain, from the microphone array, a plurality of audio signals including a first audio signal and a second audio signal; determine direction data based on the plurality of audio signals, the direction data comprising an indication of an estimated probability of voice activity for one or more directions, wherein to determine direction data comprises to apply an offline-trained neural network; and output audio data based on the direction data to the conference controller.

Abstract: A signal processing method and a wearable electronic device such as a headphone or an earphone comprising a microphone arranged to pick up an acoustic signal and convert the acoustic signal to a microphone signal (x); a loudspeaker arranged in an earpiece; and a processor configured to control the volume of a masking signal (m); and supply the masking signal (m) to the loudspeaker. Further, the processor is further configured to detect voice activity and generate a voice activity signal (y) which is, concurrently with the microphone signal, sequentially indicative of one or more of: voice activity and voice in-activity; and control the volume of the masking signal (m) in response to the voice activity signal (y) in accordance with supplying the masking signal (m) to the loudspeaker at a first volume at times when the voice activity signal (y) is indicative of voice activity and at a second volume at times when the voice activity signal (y) is indicative of voice in-activity.

Abstract: A method comprising: at an electronic device (100) having an array microphones (101) with a plurality of microphones and a processor (102): receiving a plurality of microphone signals (x1, x2, x3) from the plurality of microphones; generating a processed signal (XP) from the plurality of microphone signals using one or both of beamforming and deconvolution; generating a compensated processed signal (XO) by compensating the processed audio signal (XP) in accordance with compensation coefficients (Z). Generating the compensated processed signal comprises: generating first spectrum values (PXP) from the processed audio signal; generating reference spectrum values (<PX>) from multiple second spectrum values (PX1, PX2, PX3) generated from each of at least two of the microphone signals in the plurality of microphone signals (x1, x2, x3); and generating the compensation coefficients (Z) from the reference spectrum values (<PX>) and the first spectrum values (PXP).

Abstract: A headset with an electro-acoustic input transducer arranged to pick up an acoustic signal and convert the acoustic signal to an electric signal. Based on processing a portion of the electric signal, the voice activity detector is configured to: detect proximal voice activity, distal voice activity and no voice activity, at times when respectively present in the acoustic signal picked up by the electro-acoustic transducer, and to select a respective mode, the selection of which is encoded in the control signal. The first processor is controlled by the voice activity detector to reduce, in the output signal, intelligibility of distal voice activity at least at portions of time periods when the control signal indicates the mode of presence of distal voice activity.

Abstract: Disclosed is a method and a headset for reducing pop-noise in voice communication between a user and a far-end device. The headset has a first, a second and a third electro-acoustic input transducer for reception of audio input signals. The headset also has a first beamformer to provide a voice signal. The first beamformer is configured to optimize the voice-to-background noise ratio. The headset comprises a second beamformer configured for providing a pop-noise signal. The pop-noise signal is based on the first input signal from the first input transducer, the second input signal from the second input transducer, and a third input signal from the third input transducer. The second beamformer is adaptively configured to cancel voice and background noise while not cancelling pop-noise. The headset compares the pop-noise signal to the voice signal to determine time periods and frequency bands having pop-noise.

Abstract: Disclosed is a method for optimizing noise cancellation in a headset, the headset comprising a headphone and a microphone unit comprising at least a first microphone and a second microphone, the method comprising: generating at least a first audio signal from the at least first microphone, where the first audio signal comprises a speech portion from a user of the headset and a noise portion from the surroundings; generating at least a second audio signal from the at least second microphone, where the second audio signal comprises a speech portion from the user of the headset and a noise portion from the surroundings; generating a noise cancelled output by filtering and summing at least a part of the first audio signal and at least a part of the second audio signal, where the filtering is adaptively configured to continually minimize the power of the noise cancelled output, and where the filtering is adaptively configured to continually provide that at least the amplitude spectrum of the speech portion of

Abstract: A headset with an electro-acoustic input transducer arranged to pick up an acoustic signal and convert the acoustic signal to an electric signal. Based on processing a portion of the electric signal, the voice activity detector is configured to: detect proximal voice activity, distal voice activity and no voice activity, at times when respectively present in the acoustic signal picked up by the electro-acoustic transducer, and to select a respective mode, the selection of which is encoded in the control signal. The first processor is controlled by the voice activity detector to reduce, in the output signal, intelligibility of distal voice activity at least at portions of time periods when the control signal indicates the mode of presence of distal voice activity.

Abstract: Disclosed is a method and a headset for reducing pop-noise in voice communication between a user and a far-end device. The headset has a first, a second and a third electro-acoustic input transducer for reception of audio input signals. The headset also has a first beamformer to provide a voice signal. The first beamformer is configured to optimize the voice-to-background noise ratio. The headset comprises a second beamformer configured for providing a pop-noise signal. The pop-noise signal is based on the first input signal from the first input transducer, the second input signal from the second input transducer, and a third input signal from the third input transducer. The second beamformer is adaptively configured to cancel voice and background noise while not cancelling pop-noise. The headset compares the pop-noise signal to the voice signal to determine time periods and frequency bands having pop-noise.

Abstract: A method for attenuating undesired content in an audio signal and to an apparatus adapted to attenuate undesired content in an audio signal. The method has receiving an audio input signal (Si); providing a main audio signal (Sa) in dependence on the audio input signal (Si); determining an input level signal (Ln) indicating a signal level (Ln,i) of the main audio signal (Sa); applying a frequency-dependent gain to the main audio signal (Sa) to provide an audio output signal (So); providing an analysis signal (Sn) in dependence on the audio input signal (Si); determining a classification signal (Sc) indicating the presence in the analysis signal (Sn) determining a threshold control signal (St) indicating a frequency-dependent level threshold (72, 75, 76, Tf) for multiple frequency subbands; and determining the frequency-dependent gain in dependence on signal levels (Ln,i).

Abstract: The present invention relates to a method for attenuating undesired content in an audio signal and to an apparatus adapted to attenuate undesired content in an audio signal. The invention may be used to reduce adverse effects on a user caused by undesired or potentially harmful audio signals received from an audio communication network, such as e.g. a telephone network, and may advantageously be implemented in headsets and other audio communication apparatus that can receive audio signals from an audio communication network and provide corresponding sound signals to one or more users. The present invention provides a method for attenuating undesired content in an audio signal.