Abstract: Systems and methods are provided for detecting wake words. An electronic device detects, from a microphone array, an audio signal in an environment proximate to the audio front end system. The electronic device processes the audio signal using a plurality of wake word detection engines, including dynamically adjusting how many wake word detection engines are available for processing the audio signal. The electronic device independently adjusts respective wake word detection thresholds for the plurality of wake word detection engines used to process the audio signal.

Abstract: Systems and methods for classifying data are disclosed. For example, a system may include at least one memory storing instructions and at least one processor configured to execute the instructions to perform operations. The operations may include receiving training data comprising a class. The operations may include training a data classification model using the training data to generate a trained data classification model. The operations may include receiving additional data comprising labeled samples of an additional class not contained in the training data. The operations may include creating a synthetic data generator. The operations may include training the synthetic data generator to generate synthetic data corresponding to the additional class. The operations may include generating a synthetic classified dataset comprising the additional class. The operations may include retraining the trained data classification model using the synthetic classified dataset.

Abstract: An example first playback device includes programming to perform functions including: (1) storing an active volume state variable in memory, wherein the active volume state variable corresponds to a current playback volume; (2) storing a volume limit state variable in memory, wherein the volume limit state variable corresponds to a playback volume limit of the first playback device; (3) detecting a command to begin playback of media at a proposed playback volume different from the current playback volume; (4) based on comparing (i) the playback volume limit and (ii) the proposed playback volume, selecting a startup playback volume; (5) playing back media at the startup playback volume; and (6) causing at least a second playback device of the media playback system to play back media at the startup playback volume.

Abstract: A speech separation device (12) of a speech separation system includes a feature amount extraction unit (121) configured to extract time-series data of a speech feature amount of mixed speech, a block division unit (122) configured to divide the time-series data of the speech feature amount into blocks having a certain time width, a speech separation neural network (1b) configured to create time-series data of a mask of each of a plurality of speakers from the time-series data of the speech feature amount divided into blocks, and a speech restoration unit (123) configured to restore the speech data of each of the plurality of speakers from the time-series data of the mask and the time-series data of the speech feature amount of the mixed speech.

Abstract: An example method includes receiving, by a computational assistant executing at one or more processors, a representation of an utterance spoken at a computing device; identifying, based on the utterance, a task to be performed by the computational assistant; responsive to determining, by the computational assistant, that complete performance of the task will take more than a threshold amount of time, outputting, for playback by one or more speakers operably connected to the computing device, synthesized voice data that informs a user of the computing device that complete performance of the task will not be immediate; and performing, by the computational assistant, the task.

Abstract: In some embodiments, a first audio signal is received via a first microphone, and a first probability of voice activity is determined based on the first audio signal. A second audio signal is received via a second microphone, and a second probability of voice activity is determined based on the first and second audio signals. Whether a first threshold of voice activity is met is determined based on the first and second probabilities of voice activity. In accordance with a determination that a first threshold of voice activity is met, it is determined that a voice onset has occurred, and an alert is transmitted to a processor based on the determination that the voice onset has occurred. In accordance with a determination that a first threshold of voice activity is not met, it is not determined that a voice onset has occurred.

Abstract: A method for multi-channel voice activity detection includes receiving a sequence of input frames characterizing streaming multi-channel audio captured by an array of microphones. Each channel of the streaming multi-channel audio includes respective audio features captured by a separate dedicated microphone. The method also includes determining, using a location fingerprint model, a location fingerprint indicating a location of a source of the multi-channel audio relative to the user device based on the respective audio features of each channel of the multi-channel audio. The method also includes generating an output from an application-specific classifier. The first score indicates a likelihood that the multi-channel audio corresponds to a particular audio type that the particular application is configured to process.

Abstract: In embodiments of an audio control module (318), audio data (310) is received from an audio data source (314) for output to an audio rendering device (316). An initialization input (326) can be received from a wireless audio headset (320) and, responsive to receiving the initialization input, the audio data (328) is communicated to the audio headset. The audio that would be generated from the audio data (322) at the audio rendering device (316) is also limited, such as by replacing the audio data (322) with null audio data, clearing audio data packets from the audio data (322), or by asserting a mute signal (336) to the audio rendering device.

Abstract: Systems, apparatuses, and methods are described for determining a direction associated with a detected spoken keyword, forming an acoustic beam in the determined direction, and listening for subsequent speech using the acoustic beam in the determined direction.

Abstract: A display apparatus including a flexible vibration module and a method of manufacturing the flexible vibration module are provided. A display apparatus includes: a display panel configured to display an image, and a flexible vibration module on a rear surface of the display panel, the flexible vibration module configured to vibrate the display panel, the flexible vibration module including: a plurality of first portions having a piezoelectric characteristic, and a plurality of second portions respectively between pairs of the plurality of first portions, the plurality of second portions having flexibility.

Abstract: A voice recognition device includes a plurality of mics disposed toward different directions and a processor connected with the plurality of mics, wherein the processor is configured to determine, in a setup mode, a direction of a first sound received through the plurality of mics; set a non-detecting zone, which includes the direction of the first sound; determine, in a normal mode, a direction of a second sound received through the plurality of mics; and skip voice recognition for the second sound or an operation based on the voice recognition depending on whether the direction of the second sound belongs to the non-detecting zone.

Abstract: This disclosure describes transcribing speech using audio, image, and other data. A system is described that includes an audio capture system configured to capture audio data associated with a plurality of speakers, an image capture system configured to capture images of one or more of the plurality of speakers, and a speech processing engine. The speech processing engine may be configured to recognize a plurality of speech segments in the audio data, identify, for each speech segment of the plurality of speech segments and based on the images, a speaker associated with the speech segment, transcribe each of the plurality of speech segments to produce a transcription of the plurality of speech segments including, for each speech segment in the plurality of speech segments, an indication of the speaker associated with the speech segment, and analyze the transcription to produce additional data derived from the transcription.

Abstract: A voice capturing method includes following operations: storing, by a buffer, voice data from a plurality of microphones; determining, by a processor, whether a target speaker exists and whether a direction of the target speaker changes according to the voice data and target speaker information; inserting a voice segment corresponding to a previous tracking direction into a current position in the voice data to generate fusion voice data when the target speaker exists and the direction of the target speaker changes from the previous tracking direction to a current tracking direction; performing, by the processor, a voice enhancement process on the fusion voice data according to the current tracking direction to generate enhanced voice data; performing, by the processor, a voice shortening process on the enhanced voice data to generate voice output data; and playing, by a playing circuit, the voice output data.

Abstract: An electronic device includes a display and one or more processors. The display displays, when a person is at a location, a virtual image at a sound localization point (SLP) in empty space. The one or more processors process sound with a room impulse response (RIR) for the location in order to generate binaural sound that externally localizes at the SLP in empty space in response to the person being at the location.

Abstract: A speech processing device includes a processor. The processor performs operations including: detecting a single-talk state based on a speech signal collected by each of microphones, the single-talk state in which any one of persons speaks; estimating a mixing rate indicating a ratio of a speech signal of the main speaking person to a speech signal of another person based on a sound pressure ratio of the speech signals collected by the microphones in the single-talk state of the main speaking person and a sound pressure ratio of the speech signals collected by the plurality of microphones in the single-talk state of the another person; and determining whether suppression of a crosstalk component due to speaking of the another person contained in the speech signal of the main speaking person is necessary based on an estimation result of the mixing rate.

Abstract: The present invention discloses a dual-microphone adaptive filtering algorithm for collecting body sound signals, characterized in that, using at least two microphones, a primary microphone and a secondary microphone, to collect signals; the primary microphone is used to collect noisy body sound signals, and the secondary microphone is used to collect environmental noise; applying a same high-pass filtering to signals collected by the primary microphone and signals collected by the secondary microphone; using a normalized least mean square algorithm on the primary microphone signals and the secondary microphone signals after the high-pass filtering to calculate a weight of the adaptive filter and to calculate an error signal to filter out environmental noise in the primary microphone signals; processing the error signal for a first time by a low-pass filtering to restore the body sound signals, to obtain the body sound signals output by the adaptive filtering algorithm.

Abstract: A motor vehicle has a component for recognizing a user based on a mobile device carried by the user and a microphone. Once a user has been recognized from the outside, the microphone converts sound waves acting on the motor vehicle into electrical signals. A speech control unit processes the sound waves transmitted by the microphone. The motor vehicle includes a number of microphones oriented in different directions on different sides of the motor vehicle. The number of microphones convert sound waves into electrical signals and transmit them to the speech control unit. The speech control unit is connected to a plurality of loudspeakers directed towards the surroundings of the vehicle for voice communication with the user, with speech output being provided only via the loudspeaker or loudspeakers closest to the user.

Abstract: A vehicle loudspeaker system, including at least two microphones forming a microphone array, at least one loudspeaker configured to emit non-human sound, a processor programmed to receive incoming audio signals from the microphone array. The processor further programmed to apply beamforming to the incoming audio signals, determine whether human generated sound is detected within the audio signal, and instruct the loudspeaker to adjust the non-human sound in response to human generated sound being detected.

Abstract: According to one embodiment, a computer-implemented method for dynamically modifying placeholder text in a conversational interface includes: processing a conversation log reflecting a conversation between a human user and an automated agent; determining, based at least in part on the processing: one or more capabilities of the automated agent; and/or a trajectory of the conversation; and dynamically modifying placeholder text in the conversational interface based at least in part on: the one or more capabilities of the automated agent; the trajectory of the conversation; or both the one or more capabilities of the automated agent and the trajectory of the conversation. Other embodiments in the form of systems and computer program products are also disclosed.

Abstract: Devices, media, and methods are presented for an audio enhanced augmented reality (AR) experience using an eyewear device. The eyewear device has a microphone system, a presentation system, a support structure configured to be head-mounted on a user, and a processor. The support structure supports the microphone system and the presentation system. The eyewear device is configured to capture, with the microphone system, audio information of an environment surrounding the eyewear device, identify an audio signal within the audio information, detect a direction of the audio signal with respect to the eyewear device, classify the audio signal, and present, by the presentation system, an application associated with the classification of the audio signal.