Abstract: A method for hardware-based machine learning acceleration is provided. The method may include partitioning, into a first batch of data and a second batch of data, an input data received at a hardware accelerator implementing a machine learning model. The input data may be a continuous stream of data samples. The input data may be partitioned based at least on a resource constraint of the hardware accelerator. An update of a probability density function associated with the machine learning model may be performed in real time. The probability density function may be updated by at least processing, by the hardware accelerator, the first batch of data before the second batch of data. An output may be generated based at least on the updated probability density function. The output may include a probability of encountering a data value. Related systems and articles of manufacture, including computer program products, are also provided.

Abstract: The disclosed subject matter generally relates to a voice assistant system for a vehicle which can distinguish who is issuing voice commands, and thereby also which sub-system in the vehicle to control according to the voice command. This is obtained by allowing the voice assistant system to receive input signals from an image capturing system configured to detect the user issuing the voice command. The vehicle compartment may be divided into several zones each associated with at least one sub-system, and the input signals from the image capturing system provide information to the control unit of the voice assistant system regarding the present zone of the user issuing the voice commands.

Abstract: A computer-implemented method includes obtaining training data including utterances of speakers in acoustic conditions, preparing at least one machine learning model, each machine learning model including a common embedding model for converting an utterance into a feature vector and a classification model for classifying the feature vector, and training, by using the training data, the machine learning model to perform classification by speaker and to perform classification by acoustic condition.

Abstract: A processing load at a receiving side is reduced in a case where a plurality of classes of audio data is transmitted. A predetermined number of audio streams including coded data of a plurality of groups is generated and a container of a predetermined format having this predetermined number of audio streams is transmitted. Command information for creating a command specifying a group to be decoded from among the plurality of groups is inserted into the container and/or the audio stream. For example, a command insertion area for the receiving side to insert a command for specifying a group to be decoded is provided in at least one audio stream among the predetermined number of audio streams.

Abstract: Methods and apparatus assist listeners in distinguishing between electronically generated binaural sound and physical environment sound while the listener wears a wearable electronic device that provides the binaural sound to the listener. The wearable electronic device generates a visual alert or audio alert when the electronically generated binaural sound occurs.

Abstract: A system and method for externalizing sound. The system includes a headphone assembly and a localizer configured to collect information related to a location of the user and of an acoustically reflective surface in the environment. A controller is configured to determine a location of at least one virtual sound source, and generate head related transfer functions that simulate characteristics of sound from the virtual sound source directly to the user and to the user via a reflection by the reflective surface. A signal processing assembly is configured to create one or more output signals by filtering the sound signal respectively with the HRTFs. Each speaker of the headphone assembly is configured to produce sound in accordance with the output signal.

Abstract: An image processing system for generating an image for display to a user at a head-mountable display device includes: a sound input unit configured to receive sound information relating to one or more sounds, an audio processing unit configured to analyse the sound information relating to the one or more sounds, an image generation unit configured to generate one or more image elements that indicate properties of analysed sound information, and an image output unit configured to output display images for display to a user of a head-mountable display device, the images comprising the generated image elements as an image overlay, where the one or more generated image elements are indicative of the direction of the source of the one or more sounds.

Abstract: The description relates to modeling acoustic effects in scenes with dynamic portals. One implementation includes obtaining a representation of a scene having a plurality of portals and simulating sound travel in the scene using a plurality of probes deployed in the scene. The implementation also includes determining acoustic parameters for initial sound traveling between respective probes based at least on the simulating. The implementation also includes identifying one or more intercepted portals in the scene that are intercepted by a particular initial sound path from a particular source location to a particular listener location, using particular acoustic parameters for the particular source location and the particular listener location.

Abstract: The present technology relates to a signal processing device and method, and a program making it possible to reduce the computational complexity of decoding at low cost. A signal processing device includes: a priority information generation unit configured to generate priority information about an audio object on the basis of a plurality of elements expressing a feature of the audio object. The present technology may be applied to an encoding device and a decoding device.

Abstract: A method for playing audio data, includes: detecting a number and position relationships of associated audio playback devices; performing audio signal processing on audio data to be played based on the number and the position relationships to obtain an audio signal matching the number of the audio playback devices; and playing the audio signal through the audio playback devices.

Abstract: Intelligent systems are disclosed that respond to user intent and desires based upon activity that may or may not be expressly directed at the intelligent system. In some embodiments, the intelligent system acquires a depth image of a scene surrounding the system. A scene geometry may be extracted from the depth image and elements of the scene may be monitored. In certain embodiments, user activity in the scene is monitored and analyzed to infer user desires or intent with respect to the system. The interpretation of the user's intent as well as the system's response may be affected by the scene geometry surrounding the user and/or the system. In some embodiments, techniques and systems are disclosed for interpreting express user communication, e.g., expressed through hand gesture movements. In some embodiments, such gesture movements may be interpreted based on real-time depth information obtained from, e.g., optical or non-optical type depth sensors.

Abstract: An information processing device 11 including: a control data generation unit that inputs analysis data X that is to be processed, to a trained model that has learnt a relationship between analysis data X that represents a time series of musical notes, and control data Y for controlling movements of an object that represents a performer, thereby generating control data Y according to the analysis data X.

Abstract: A speaker system uses destructive wave interference to generate “dead spots” with respect to an audio presentation. The signal for the dead spot generating device can be an inverted signal generated using the audio signal. In one embodiment, the inverted signal is generated using the audio signal, an indication of loudness at one or more active speakers, and a determination of the characteristics of the sound path from the one or more active speakers (including delay and attenuation).

Abstract: A sound signal processing method includes: obtaining a sound signal; obtaining impulse response data that was measured in a predetermined space before the sound signal is obtained; generating an early reflected sound control signal not including a reverberant sound by convolving impulse response data of an early reflected sound among the obtained impulse response data into the obtained sound signal.

Abstract: System and method for enhancing audio reproduced by an audio reproduction device is described. A plurality of convolution coefficients are generated for a predefined space. A digital audio signal is modified based on the generated convolved digital audio signal to generate a convolved digital audio signal. The convolved digital audio signal is converted to a convolved analog audio signal. The convolved analog audio signal is fed to the audio reproduction device.

Abstract: An apparatus for generating loudspeaker signals includes an object metadata processor configured to receive metadata, to calculate a second position of the audio object depending on the first position of the audio object and on a size of a screen if the audio object is indicated in the metadata as being screen-related, to feed the first position of the audio object as the position information into the object renderer if the audio object is indicated in the metadata as being not screen-related, and to feed the second position of the audio object as the position information into the object renderer if the audio object is indicated in the metadata as being screen-related. The apparatus further includes an object renderer configured to receive an audio object and to generate the loudspeaker signals depending on the audio object and on position information.

Abstract: In some embodiments, a method comprises receiving audio content comprising left input channel signals and right input channel signals, and generating first and second input signals from the left and right input channel signals. The first input signal is based on a sum of the left and right input channel signals, and the second input signal is based on a difference of the left and right input channel signals. An array transfer function is applied to the first and second input signals to produced audio output signals, which can be provided to a plurality of audio transducers on one or more playback devices.

Abstract: According to an example embodiment, a method for processing a spatial audio signal that represents an audio scene, wherein the spatial audio signal is controllable and associated with at least two viewing directions is provided, the method including receiving a focus direction and a focus amount; processing the spatial audio signal by modifying the audio scene so as to control emphasis in, at least in part, a portion of the spatial audio signal in said focus direction according to said focus amount; and outputting the processed spatial audio signal, wherein the modified audio scene enables the emphasis in, at least in part, said portion of the spatial audio signal in said focus direction according to said focus amount.

Abstract: A method including receiving an audio scene including at least one source captured using at least one near field microphone and at least one far field microphone. The method includes determining at least one room-impulse-response associated with the audio scene based on the at least one near field microphone and the at least one far field microphone, accessing a predetermined scene geometry corresponding to the audio scene, and identifying best match to the predetermined scene geometry in a scene geometry database. The method also includes performing RIR comparison based on the at least one RIR and at least one geometric RIR associated with the best matching geometry and rendering a volumetric audio scene based on a result of the RIR comparison.

Abstract: An audio system generates customized head-related transfer functions (HRTFs) for a user. The audio system receives an initial set of estimated HRTFs. The initial set of HRTFs may have been estimated using a trained machine learning and computer vision system and pictures of the user's ears. The audio system generates a set of test locations using the initial set of HRTFs. The audio system presents test sounds at each of the initial set of test locations using the initial set of HRTFs. The audio system monitors user responses to the test sounds. The audio system uses the monitored responses to generate a new set of estimated HRTFs and a new set of test locations. The process repeats until a threshold accuracy is achieved or until a set period of time expires. The audio system presents audio content to the user using the customized HRTFs.

Abstract: An example playback device is configured to (i) receive, via a network interface, data representing a command to play back audio content, where the audio content is a first type of audio content, (ii) during playback of the first type of audio content via an audio amplifier configured to drive a speaker, apply a first calibration and a second calibration to playback by the playback device, where the first calibration at least partially offsets one or more acoustic characteristics of an environment surrounding the playback device when applied to playback by the playback device, and where the second calibration corresponds to the first type of audio content, and (iii) during playback of a second type of audio content via the audio amplifier configured to drive the speaker, apply a third calibration to playback by the playback device, where the third calibration corresponds to the second type of audio content.

Abstract: An audio headset may receive a plurality of audio signals corresponding to plurality of surround sound channels. The headset may determine, via its audio processing circuitry, context and/or content of the audio signals. The audio processing circuitry may process the audio signals to generate stereo signals carrying one or more virtual surround channels, wherein the processing comprises automatically controlling, based on the context and the content of the audio signals, a simulated acoustic environment of the virtual surround channels.

Abstract: Provided is a sound bar including: a rear sound signal generating unit that generates a rear sound from an input audio signal; and an output unit that outputs the rear sound generated by the rear sound signal generating unit to a rear sound speaker.

Abstract: The present disclosure relates to an apparatus for decoding an encoded Unified Audio and Speech stream. The apparatus comprises a core decoder for decoding the encoded Unified Audio and Speech stream. The core decoder includes an upmixing unit adapted to perform mono to stereo upmixing. The upmixing unit includes a decorrelator unit D adapted to apply a decorrelation filter to an input signal. The decorrelator unit is adapted to determine filter coefficients for the decorrelation filter by referring to pre-computed values. The present disclosure further relates to a an apparatus for encoding a Unified Audio and Speech stream, as well as to corresponding methods and storage media.

Abstract: A mechanism is described for facilitating display-based audio splitting in media environments, according to one embodiment. A method of embodiments, as described herein, includes receiving, at a computing device, a request to split audio streams associated with audio channels in communication with the computing device serving as a host device, wherein the audio streams are generated by one or more applications; selecting a split mode to facilitate splitting of the audio streams; and splitting the audio streams based on the split mode such that the audio streams are continuously mapped to the audio channels to adjust for delays and fine tuning.

Abstract: An apparatus including at least one processor and at least one memory including a computer program code, the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus at least to: determine, for two or more microphone audio signals, at least one spatial audio parameter for providing spatial audio reproduction; determine at least one coherence parameter associated with a sound field based on the two or more microphone audio signals, such that another sound field is configured to be reproduced based on the at least one spatial audio parameter and the at least one coherence parameter.

Abstract: An audio processing system and method which calculates, based on spatial metadata of the audio object, a panning coefficient for each of the audio objects in relation to each of a plurality of predefined channel coverage zones. Converts the audio signal into submixes in relation to the predefined channel coverage zones based on the calculated panning coefficients and the audio objects. Each of the submixes indicating a sum of components of the plurality of the audio objects in relation to one of the predefined channel coverage zones. Generating a submix gain by applying an audio processing to each of the submix and controls an object gain applied to each of the audio objects. The object gain being as a function of the panning coefficients for each of the audio objects and the submix gains in relation to each of the predefined channel coverage zones.

Abstract: Disclosed is an audio signal processing device. The audio signal processing device includes a receiving end configured to receive a 2-channel stereo signal and a processor configured to process the 2-channel stereo signal. The processor is configured to filter the 2-channel stereo signal using a spatial distortion removal filter and output the filtered 2-channel stereo signal to a speaker including two or more channels. The spatial distortion removal filter is a filter for offsetting distortion that occurs when the output signal is transmitted from the speaker to a listener, and includes an ipsilateral filter applied to an ipsilateral signal of the 2-channel audio signal and a contralateral filter applied to a contralateral signal of the 2-channel audio signal.

Abstract: A user listening position is estimated without imposing a burden on a user, and a sound reproduction environment suitable for listening is formed regardless of an actual speaker arrangement state. An estimation part that estimates the user listening position by using position information of N speakers that are three or more speakers, and an arrangement part that sets virtual speaker arrangement by using the user listening position are provided. Therefore, it is possible to estimate the user listening position without imposing a burden on a user and form a sound reproduction environment suitable for listening regardless of an actual speaker arrangement state.

Abstract: In some examples, immersive audio rendering may include determining whether an audio signal includes a first content format including stereo content, or a second content format including multichannel or object-based content. In response to a determination that the audio signal includes the first content format, the audio signal may be routed to a first block that includes a low-frequency extension and a stereo to multichannel upmix to generate a resulting audio signal. Alternatively, the audio signal may be routed to another low-frequency extension to generate the resulting audio signal. The audio signal may be further processed by performing spatial synthesis on the resulting audio signal, and crosstalk cancellation on the spatial synthesized audio signal. Further, multiband-range compression may be performed on the crosstalk cancelled audio signal, and an output stereo signal may be generated based on the multiband-range compressed audio signal.

Abstract: At least one exemplary embodiment is directed to modifying audio content using location specific spatial impulse responses.

Abstract: An audio processing method includes: M audio signals are obtained by processing an audio signal by M virtual speakers; M first HRTFs and M second HRTFs are obtained, where the M first HRTFs corresponding to a left ear position, and the M second HRTFs corresponding to a right ear position; high-band impulse responses of some of the M first HRTFs are modified to obtain modified first target HRTFs, and high-band impulse responses of some of the M second HRTFs are modified to obtain modified second target HRTFs; a first target audio signal corresponding to the left ear position is obtained based on the modified first target HRTFs and un-modified first HRTFs, and the M audio signals; and a second target audio signal corresponding to the right ear position is obtained based on the modified second HRTFs, un-modified second target HRTFs, and the M audio signals.

Abstract: An apparatus and method of rendering audio. A binaural signal is split on an amplitude weighting basis into a front binaural signal and a rear binaural signal, based on perceived position information of the audio. In this manner, the front-back differentiation of the binaural signal is improved.

Abstract: The embodiments of the disclosure provide a method for providing an occluded sound effect and an electronic device. The method includes: providing a virtual environment, wherein the virtual environment comprises a first object, and the first object is approximated as a second object; defining an object detection range of a sound source based on a sound ray originated from the sound source; in response to determining that the first object enters the object detection range, defining a reference plane based on a reference point on the second object and the sound ray, wherein the reference plane has an intersection area with the object detection range; projecting the second object onto the reference plane as a first projection; determining a sound occluding factor based on the intersection area and the first projection; and adjusting a sound signal based on the sound occluding factor.

Abstract: An original noisy signal of each of at least two microphones is acquired by acquiring, using the at least two microphones, an audio signal emitted by each sound source. For each frame in time domain, an estimated frequency-domain signal of each sound source is acquired according to the original noisy signal of each of the at least two microphones. A frequency collection containing a plurality of predetermined static frequencies and dynamic frequencies is determined in a predetermined frequency band range. A weighting coefficient of each frequency contained in the frequency collection is determined according to the estimated frequency-domain signal of the each frequency in the frequency collection. A separation matrix of the each frequency is determined according to the weighting coefficient. The audio signal emitted by each of the at least two sound sources is acquired based on the separation matrix and the original noisy signal.

Abstract: A decoder for generating an audio output signal having one or more audio output channels is provided, having a receiving interface for receiving an audio input signal having a plurality of audio object signals, for receiving loudness information on the audio object signals, and for receiving rendering information indicating whether one or more of the audio object signals shall be amplified or attenuated, further having a signal processor for generating the one or more audio output channels of the audio output signal, configured to determine a loudness compensation value depending on the loudness information and depending on the rendering information, and configured to generate the one or more audio output channels of the audio output signal from the audio input signal depending on the rendering information and depending on the loudness compensation value. One or more by-pass audio object signals are employed for generating the audio output signal. Moreover, an encoder is provided.

Abstract: Embodiments are described for a method of simultaneously localizing a set of speakers and microphones, having only the times of arrival between each of the speakers and microphones. An autodiscovery process uses an external input to set: a global translation (3 continuous parameters), a global rotation (3 continuous parameters), and discrete symmetries, i.e., an exchange of any axis pairs and/or reversal of any axis. Different time of arrival acquisition techniques may be used, such as ultrasonic sweeps or generic multitrack audio content. The autodiscovery algorithm is based in minimizing a certain cost function, and the process allows for latencies in the recordings, possibly linked to the latencies in the emission.

Abstract: A binaural room impulse response (BRIR) can be generated based on a position of a listener's head, and a plurality of head related impulse responses (HRIRs). Each of the plurality of HRIRs are selected for a respective one of a plurality of acoustic reflections which, when taken together, approximate reverberation of a room. Each of the acoustic reflections have a direction and a delay. The BRIR filter is applied to source audio to generate binaural audio output.

Abstract: Apparatus including circuitry configured for: determining, for two or more speaker channel audio signals, at least one spatial audio parameter for providing spatial audio reproduction; determining between the two or more speaker channel audio signals at least one audio signal relationship parameter, the at least one audio signal relationship parameter being associated with at least one coherence parameter, in such a way that the at least one coherence parameter provides at least one interchannel coherence information for at least two frequency bands, to reproduce the two or more speaker channel audio signals based on the at least one spatial audio parameter and the at least one audio signal relationship parameter; and transmitting using at least one determined value.

Abstract: An apparatus for estimating an inter-channel time difference between a first channel signal and a second channel signal, includes: a calculator for calculating a cross-correlation spectrum for a time block from the first channel signal in the time block and the second channel signal in the time block; a spectral characteristic estimator for estimating a characteristic of a spectrum of the first channel signal or the second channel signal for the time block; a smoothing filter for smoothing the cross-correlation spectrum over time using the spectral characteristic to obtain a smoothed cross-correlation spectrum; and a processor for processing the smoothed cross-correlation spectrum to obtain the inter-channel time difference.

Abstract: A method and a device for processing an audio signal in a vehicle are provided. The method includes: obtaining an audio signal by a microphone array; performing echo cancellation on the obtained audio signal, to obtain a first processed signal; and performing beamforming on the first processed signal according to sound zones in which microphones of the microphone array are located, to obtain a second processed signal, wherein the vehicle includes at least two sound zones, and each microphone of the microphone array is located in at least one sound zone. With the beamforming, the requirements for isolation degree between different sound zones is not high, and the sound source of the audio signal can be accurately determined.

Abstract: An audio processing system and a method thereof are provided. The system includes an audio receiver and an audio processor device. The audio receiver includes an audio interface, an absolute value converter unit, a framing unit and a characteristic value detector unit. The audio interface receives an audio signal from an audio transceiver device. The absolute value converter unit takes an absolute value of the audio signal to output an absolute value audio signal. The framing unit divides the absolute value audio signal into a plurality of sub-frame signals. The characteristic value detector unit detects characteristic values of the sub-frame signals. The audio processer processes the audio signal outputted by the audio transceiver device according to the characteristic values of the sub-frame signals.

Abstract: A system and method for tracking a position node for one or more moving objects, the tracking including scanning for one or more moving objects, detecting one or more moving objects, creating a temporary position node around one or more detected moving objects, searching a node pool for an existing position node nearby the temporary position node, setting a timer for the existing position node that is nearby the temporary position node, the timer has a predetermined time limit, and for a temporary position node that is not nearby an existing position node, creating a new position node that is added to the node pool and setting a timer associated with the new position node, the timer associated with the new position node has a predetermined time limit.

Abstract: A method for transmitting audio data performed by an audio data transmission apparatus in accordance with the present invention comprises the steps of: generating playback environment information of three-dimensional audio content; encoding a three-dimensional audio signal of the three-dimensional audio content; and transmitting, to an audio data reception apparatus, the encoded three-dimensional audio signal of the three-dimensional audio content and the generated playback environment information, wherein the playback environment information includes environment information of a room in which the three-dimensional audio content is played.

Abstract: In general, techniques are described by which to support scalable unified audio rendering. A device comprising an audio decoder, a memory, and a processor may be configured to perform various aspects of the techniques. The audio decoder may decode, from a bitstream, first audio data and second audio data. The memory may store the first audio data and the second audio data. The processor may render the first audio data into first spatial domain audio data for playback by virtual speakers at a set of virtual speaker locations, and render the second audio data into second spatial domain audio data for playback by the virtual speakers at the set of virtual speaker locations. The processor may also mix the first spatial domain audio data and the second spatial domain audio data to obtain mixed spatial domain audio data, and convert the mixed spatial domain audio data to scene-based audio data.

Abstract: Various embodiments relate generally to electrical and electronic hardware, computer software, wired and wireless network communications, and audio and speaker systems. More specifically, disclosed are an apparatus and a method for processing signals for optimizing audio, such as 3D audio, by adjusting the filtering for cross-talk cancellation based on listener position and/or orientation. In one embodiment, an apparatus is configured to include a plurality of transducers, a memory, and a processor configured to execute instructions to determine a physical characteristic of a listener relative to the origination of the multiple channels of audio, to cancel crosstalk in a spatial region coincident with the listener at a first location, to detect a change in the physical characteristic of the listener, and to adjust the cancellation of crosstalk responsive to detecting the change in the physical characteristic to establish another spatial region at a second location.

Abstract: Occlusion devices, earpiece devices and methods of forming occlusion devices are provided. An occlusion device is configured to occlude an ear canal. The occlusion device includes an insertion element and at least one expandable element disposed on the insertion element. The expandable element is configured to receive a medium via the insertion element and is configured to expand, responsive to the medium, to contact the ear canal. Physical parameters of the occlusion device are selected to produce a predetermined sound attenuation characteristic over a frequency band, such that sound is attenuated more in a first frequency range of the frequency band than in a second frequency range of the frequency band.

Abstract: An apparatus and a method for acoustic scene classification of a block of audio samples are provided. The block is partitioned into frames in the time domain. For each respective frame of a plurality of frames of the block, a change measure between the respective frame and a preceding frame of the block is calculated. The respective frame is assigned, based on the calculated change measure, to one of a set of short-event frames, a set of long-event frames, and a set of background frames. The feature vector is determined based on a feature computed from one or more of the set of short-event frames, the set of long-event frames, and the set of background frames.

Abstract: Described is a method of processing position information indicative of an object position of an audio object, wherein the object position is usable for rendering of the audio object, that comprises: obtaining listener orientation information indicative of an orientation of a listener's head; obtaining listener displacement information indicative of a displacement of the listener's head; determining the object position from the position information; modifying the object position based on the listener displacement information by applying a translation to the object position; and further modifying the modified object position based on the listener orientation information. Further described is a corresponding apparatus for processing position information indicative of an object position of an audio object, wherein the object position is usable for rendering of the audio object.

Abstract: To enable channel setting to be performed easily and accurately to speakers. Thus, provided is a speaker system including: N number of speakers, the N being three or more; and a signal processing device capable of communicating with each speaker. The signal processing device performs processing of recognizing two arrangement reference speakers with reception of notification that a designation operation has been received from a user, from two speakers among the N number of speakers, and processing of acquiring distance information between each speaker. The signal processing device recognizes a relative-position relationship between the N number of speakers, with the two arrangement reference speakers and the distance information between each speaker. Then, the signal processing device automatically sets a channel to each speaker, on the basis of the relative-position relationship recognized.