Abstract: A speech recognition system for resolving impaired utterances can have a speech recognition engine configured to receive a plurality of representations of an utterance and concurrently to determine a plurality of highest-likelihood transcription candidates corresponding to each respective representation of the utterance. The recognition system can also have a selector configured to determine a most-likely accurate transcription from among the transcription candidates. As but one example, the plurality of representations of the utterance can be acquired by a microphone array, and beamforming techniques can generate independent streams of the utterance across various look directions using output from the microphone array.

Abstract: In some implementations, a computing device can calibrate media playback channels for presenting media content through a media system by determining the media propagation latency through the media system. For example, the computing device can send calibration content (e.g., audio data, video data, etc.) to various playback devices (e.g., playback channels) of the media system and record a timestamp indicating when the calibration content was sent. When the playback devices present the calibration content, a sensor device (e.g., remote control device, smartphone, etc.) can detect the presentation of the calibration content. The sensor device can send calibration data (e.g., media samples that may include the calibration content and/or a timestamp indicating when the media sample was detected by the sensor device) to the computing device. The computing device can determine the propagation latency (e.g., presentation delay) based on the calibration data received from the sensor device.

Abstract: In some implementations, a computing device can calibrate media playback channels for presenting media content through a media system by determining the media propagation latency through the media system. For example, the computing device can send calibration content (e.g., audio data, video data, etc.) to various playback devices (e.g., playback channels) of the media system and record a timestamp indicating when the calibration content was sent. When the playback devices present the calibration content, a sensor device (e.g., remote control device, smartphone, etc.) can detect the presentation of the calibration content. The sensor device can send calibration data (e.g., media samples that may include the calibration content and/or a timestamp indicating when the media sample was detected by the sensor device) to the computing device. The computing device can determine the propagation latency (e.g., presentation delay) based on the calibration data received from the sensor device.

Abstract: An audio system has a housing in which are integrated a number of microphones. A programmed processor accesses the microphone signals and produces a number of acoustic pick up beams based groups of microphones, an estimation of voice activity and an estimation of noise characteristics on each beam. Two or more beams including a voice beam that is used to pick up a desired voice and a noise beam that is used to provide information to estimate ambient noise are adaptively selected from among the plurality of beams, based on thresholds for voice separation and thresholds for noise-matching. Other embodiments are also described and claimed.

Abstract: A network device in a wireless communication system includes a first communication circuit that receives a first control-plane message from a base station (BS) over a first communication link; a second communication circuit that transmits user-plane data to a user equipment (UE) over a second communication link established responsive to the first control-plane message; and a third communication circuit that receives the user-plane traffic data from a server over a third communication link. The UE communicates user-plane data with the BS and the network device simultaneously at a predetermined time. The first communication circuit receives a second control-plane message from the BS. The second communication circuit transmits a pilot signal to the UE responsive to the second control-plane message to enable the UE to measure a radio link quality of the second communication link using the pilot signal.

Abstract: A speech recognition system for resolving impaired utterances can have a speech recognition engine configured to receive a plurality of representations of an utterance and concurrently to determine a plurality of highest-likelihood transcription candidates corresponding to each respective representation of the utterance. The recognition system can also have a selector configured to determine a most-likely accurate transcription from among the transcription candidates. As but one example, the plurality of representations of the utterance can be acquired by a microphone array, and beamforming techniques can generate independent streams of the utterance across various look directions using output from the microphone array.

Abstract: Several embodiments of a digital speech signal enhancer are described that use an artificial neural network that produces clean speech coding parameters based on noisy speech coding parameters as its input features. A vocoder parameter generator produces the noisy speech coding parameters from a noisy speech signal. A vocoder model generator processes the clean speech coding parameters into estimated clean speech spectral magnitudes. In one embodiment, a magnitude modifier modifies an original frequency spectrum of the noisy speech signal using the estimated clean speech spectral magnitudes, to produce an enhanced frequency spectrum, and a synthesis block converts the enhanced frequency spectrum into time domain, as an output speech sequence. Other embodiments are also described.

Abstract: An audio system includes one or more loudspeaker cabinets, each having loudspeakers. Sensing logic determines an acoustic environment of the loudspeaker cabinets. The sensing logic may include an echo canceller. A low frequency filter corrects an audio program based on the acoustic environment of the loudspeaker cabinets. The system outputs an omnidirectional sound pattern, which may be low frequency sound, to determine the acoustic environment. The system may produce a directional pattern superimposed on an omnidirectional pattern, if the acoustic environment is in free space. The system may aim ambient content toward a wall and direct content away from the wall, if the acoustic environment is not in free space. The sensing logic automatically determines the acoustic environment upon initial power up and when position changes of loudspeaker cabinets are detected. Accelerometers may detect position changes of the loudspeaker cabinets.

Abstract: An electronic device having a device housing includes a loudspeaker and several microphones within the device housing. A control circuit is electrically coupled to the loudspeaker and microphones. The loudspeaker produces speech and/or music. The control circuit determines a statistical measure for a first data set representing individual impulse responses from the plurality of microphones and compares that to a predetermined statistical measure for a second data set representing individual object-free impulse responses from the plurality of microphones to determine if an object is near the device. The statistical measure may be variance and may be computed in the time domain. Variance may be calculated using differences between the individual impulse responses and a mean impulse response that is a linear combination of the impulse responses for the plurality of microphones. The control circuit may include echo cancellers to mitigate common signals and/or other acoustic sources.

Abstract: A speech recognition system for resolving impaired utterances can have a speech recognition engine configured to receive a plurality of representations of an utterance and concurrently to determine a plurality of highest-likelihood transcription candidates corresponding to each respective representation of the utterance. The recognition system can also have a selector configured to determine a most-likely accurate transcription from among the transcription candidates. As but one example, the plurality of representations of the utterance can be acquired by a microphone array, and beamforming techniques can generate independent streams of the utterance across various look directions using output from the microphone array.

Abstract: An audio system includes one or more loudspeaker cabinets, each having loudspeakers. Sensing logic determines an acoustic environment of the loudspeaker cabinets. The sensing logic may include an echo canceller. A low frequency filter corrects an audio program based on the acoustic environment of the loudspeaker cabinets. The system outputs an omnidirectional sound pattern, which may be low frequency sound, to determine the acoustic environment. The system may produce a directional pattern superimposed on an omnidirectional pattern, if the acoustic environment is in free space. The system may aim ambient content toward a wall and direct content away from the wall, if the acoustic environment is not in free space. The sensing logic automatically determines the acoustic environment upon initial power up and when position changes of loudspeaker cabinets are detected. Accelerometers may detect position changes of the loudspeaker cabinets.

Abstract: An audio system includes one or more loudspeaker cabinets, each having loudspeakers. The system outputs an omnidirectional sound pattern to determine the acoustic environment. Sensing logic determines an acoustic environment of the loudspeaker cabinets. The sensing logic may include an echo canceller. A playback mode processor adjusts an audio program according to a playback mode determined from the acoustic environment of the audio system. The system may produce a directional pattern superimposed on an omnidirectional pattern, if the acoustic environment is in free space. The system may aim ambient content toward a wall and direct content away from the wall, if the acoustic environment is not in free space. The sensing logic automatically determines the acoustic environment upon initial power up and when position changes of loudspeaker cabinets are detected. Accelerometers may detect position changes of the loudspeaker cabinets.

Abstract: A device and a corresponding method are provided to tune parameters of an echo control process without re-initializing the echo control process and without interrupting a playback process. A state of the device and environment around the device is computed during use of the device given information from sensors. Such sensors can give information on the position of the device, the orientation of the device, the presence of a proximate object, or handling of the device resulting in occlusion of microphones and loudspeakers, among other things. The computed state of the device is mapped to an associated device state code from among a plurality of device state codes. The parameters of the echo control process are tuned either according to the associated device state code, or a change in such a code, during use of the device.

Abstract: Several embodiments of a digital speech signal enhancer are described that use an artificial neural network that produces clean speech coding parameters based on noisy speech coding parameters as its input features. A vocoder parameter generator produces the noisy speech coding parameters from a noisy speech signal. A vocoder model generator processes the clean speech coding parameters into estimated clean speech spectral magnitudes. In one embodiment, a magnitude modifier modifies an original frequency spectrum of the noisy speech signal using the estimated clean speech spectral magnitudes, to produce an enhanced frequency spectrum, and a synthesis block converts the enhanced frequency spectrum into time domain, as an output speech sequence. Other embodiments are also described.

Abstract: An audio system includes one or more loudspeaker cabinets, each having loudspeakers. Sensing logic determines an acoustic environment of the loudspeaker cabinets. The sensing logic may include an echo canceller. A low frequency filter corrects an audio program based on the acoustic environment of the loudspeaker cabinets. The system outputs an omnidirectional sound pattern, which may be low frequency sound, to determine the acoustic environment. The system may produce a directional pattern superimposed on an omnidirectional pattern, if the acoustic environment is in free space. The system may aim ambient content toward a wall and direct content away from the wall, if the acoustic environment is not in free space. The sensing logic automatically determines the acoustic environment upon initial power up and when position changes of loudspeaker cabinets are detected. Accelerometers may detect position changes of the loudspeaker cabinets.

Abstract: An audio system includes one or more loudspeaker cabinets, each having loudspeakers. The system outputs an omnidirectional sound pattern to determine the acoustic environment. Sensing logic determines an acoustic environment of the loudspeaker cabinets. The sensing logic may include an echo canceller. A playback mode processor adjusts an audio program according to a playback mode determined from the acoustic environment of the audio system. The system may produce a directional pattern superimposed on an omnidirectional pattern, if the acoustic environment is in free space. The system may aim ambient content toward a wall and direct content away from the wall, if the acoustic environment is not in free space. The sensing logic automatically determines the acoustic environment upon initial power up and when position changes of loudspeaker cabinets are detected. Accelerometers may detect position changes of the loudspeaker cabinets.

Abstract: Embodiments of the present disclosure can provide systems, methods, and computer-readable medium for adjusting audio and/or video information of a video clip based at least in part on facial feature and/or voice feature characteristics extracted from hardware components. For example, in response to detecting a request to generate an avatar video clip of a virtual avatar, a video signal associated with a face in a field of view of a camera and an audio signal may be captured. Voice feature characteristics and facial feature characteristics may be extracted from the audio signal and the video signal, respectively. In some examples, in response to detecting a request to preview the avatar video clip, an adjusted audio signal may be generated based at least in part on the facial feature characteristics and the voice feature characteristics, and a preview of the video clip of the virtual avatar using the adjusted audio signal may be displayed.

Abstract: Embodiments of the present disclosure can provide systems, methods, and computer-readable medium for providing audio and/or video effects based at least in part on facial features and/or voice feature characteristics of the user. For example, video and/or an audio signal of the user may be recorded by a device. Voice audio features and facial feature characteristics may be extracted from the voice audio signal and the video, respectively. The facial features of the user may be used to modify features of a virtual avatar to emulate the facial feature characteristics of the user. The extracted voice audio features may modified to generate an adjusted audio signal or an audio signal may be composed from the voice audio features. The adjusted/composed audio signal may simulate the voice of the virtual avatar. A preview of the modified video/audio may be provided at the user's device.

Abstract: A speech recognition system for resolving impaired utterances can have a speech recognition engine configured to receive a plurality of representations of an utterance and concurrently to determine a plurality of highest-likelihood transcription candidates corresponding to each respective representation of the utterance. The recognition system can also have a selector configured to determine a most-likely accurate transcription from among the transcription candidates. As but one example, the plurality of representations of the utterance can be acquired by a microphone array, and beamforming techniques can generate independent streams of the utterance across various look directions using output from the microphone array.

Abstract: A speech recognition system for resolving impaired utterances can have a speech recognition engine configured to receive a plurality of representations of an utterance and concurrently to determine a plurality of highest-likelihood transcription candidates corresponding to each respective representation of the utterance. The recognition system can also have a selector configured to determine a most-likely accurate transcription from among the transcription candidates. As but one example, the plurality of representations of the utterance can be acquired by a microphone array, and beamforming techniques can generate independent streams of the utterance across various look directions using output from the microphone array.