Abstract: In some embodiments, methods and systems for training an acoustic model, where the training includes a training loop (including at least one epoch) following a data preparation phase. During the training loop, training data are augmented to generate augmented training data. During each epoch of the training loop, at least some of the augmented training data is used to train the model. The augmented training data used during each epoch may be generated by differently augmenting (e.g., augmenting using a different set of augmentation parameters) at least some of the training data. In some embodiments, the augmentation is performed in the frequency domain, with the training data organized into frequency bands. The acoustic model may be of a type employed (when trained) to perform speech analytics (e.g., wakeword detection, voice activity detection, speech recognition, or speaker recognition) and/or noise suppression.

Abstract: Computer-implemented methods for training an acoustic model for a far-field utterance processing system are provided. The acoustic model may be configured to map an input audio signal into linguistic or paralinguistic units. The training may involve imparting far-field acoustic characteristics upon near-field training vectors that include a plurality of near-microphone utterance signals. Imparting the far-field acoustic characteristics may involve generating a plurality of simulated room impulse responses, convolving one or more of the simulated room impulse responses with the near-field training vectors, to produce a plurality of simulated far-field utterance signals and saving the results of the training in one or more non-transitory memory devices corresponding with the acoustic model. Generating simulated room impulse responses may involve simulating room reverberation times but not simulating early reflections from room surfaces.

Abstract: A computer implemented system for rendering captured audio soundfields to a listener comprises apparatus to deliver the audio soundfields to the listener. The delivery apparatus delivers the audio soundfields to the listener with first and second audio elements perceived by the listener as emanating from first and second virtual source locations, respectively, and with the first audio element and/or the second audio element delivered to the listener from a third virtual source location. The first virtual source location and the second virtual source location are perceived by the listener as being located to the front of the listener, and the third virtual source location is located to the rear or the side of the listener.

Abstract: Disclosed is an apparatus and method operative to receive packets of media from a network including a receiver unit operative to receive the packets from the network, a jitter buffer data structure for receiving the packets in an ordered queue, the jitter buffer data structure having a tail into which the packets are input; a plurality of heads defining points in the jitter buffer data structure from which the ordered queue of packets are to be played back, the heads comprise an adjustable actual playback head coupled to an actual playback unit and at least one prototype head, each prototype head having associated therewith a target latency a processor having decision logic operable to determine a cost of achieving the associated target latency for each prototype head, wherein the decision logic compares the costs determined for each prototype head to identify a particular target latency and head location for the actual playback head of the buffer and a playback unit coupled to the processor for actual playba

Abstract: Described herein is a method for creating an object-based audio signal from an audio input, the audio input including one or more audio channels that are recorded to collectively define an audio scene. The one or more audio channels are captured from a respective one or more spatially separated microphones disposed in a stable spatial configuration. The method includes the steps of: a) receiving the audio input; b) performing spatial analysis on the one or more audio channels to identify one or more audio objects within the audio scene; c) determining contextual information relating to the one or more audio objects; d) defining respective audio streams including audio data relating to at least one of the identified one or more audio objects; and e) outputting an object-based audio signal including the audio streams and the contextual information.

Abstract: Embodiments are described for a soundfield system that receives a transmitting soundfield, wherein the transmitting soundfield includes a sound source at a location in the transmitting soundfield. The system determines a rotation angle for rotating the transmitting soundfield based on a desired location for the sound source. The transmitting soundfield is rotated by the determined angle and the system obtains a listener's soundfield based on the rotated transmitting soundfield. The listener's soundfield is transmitted for rendering to a listener.

Abstract: A computer implemented system for rendering captured audio soundfields to a listener comprises apparatus to deliver the audio soundfields to the listener. The delivery apparatus delivers the audio soundfields to the listener with first and second audio elements perceived by the listener as emanating from first and second virtual source locations, respectively, and with the first audio element and/or the second audio element delivered to the listener from a third virtual source location. The first virtual source location and the second virtual source location are perceived by the listener as being located to the front of the listener, and the third virtual source location is located to the rear or the side of the listener.

Abstract: Some aspects of the present disclosure involve the recording, processing and playback of audio data corresponding to conferences, such as teleconferences. In some teleconference implementations, the audio experience heard when a recording of the conference is played back may be substantially different from the audio experience of an individual conference participant during the original teleconference. In some implementations, the recorded audio data may include at least some audio data that was not available during the teleconference. In some examples, the spatial characteristics of the played-back audio data may be different from that of the audio heard by participants of the teleconference.

Abstract: Teleconference audio data including a plurality of individual uplink data packet streams, may be received during a teleconference. Each uplink data packet stream may corresponding to a telephone endpoint used by one or more teleconference participants. The teleconference audio data may be analyzed to determine a plurality of suppressive gain coefficients, which may be applied to first instances of the teleconference audio data during the teleconference, to produce first gain-suppressed audio data provided to the telephone endpoints during the teleconference. Second instances of the teleconference audio data, as well as gain coefficient data corresponding to the plurality of suppressive gain coefficients, may be sent to a memory system as individual uplink data packet streams. The second instances of the teleconference audio data may be less gain-suppressed than the first gain-suppressed audio data.

Abstract: Some aspects of the present disclosure involve the recording, processing and playback of audio data corresponding to conferences, such as teleconferences. In some teleconference implementations, the audio experience heard when a recording of the conference is played back may be substantially different from the audio experience of an individual conference participant during the original teleconference. In some implementations, the recorded audio data may include at least some audio data that was not available during the teleconference. In some examples, the spatial characteristics of the played-back audio data may be different from that of the audio heard by participants of the teleconference.

Abstract: A computer implemented system for rendering captured audio soundfields to a listener comprises apparatus to deliver the audio soundfields to the listener. The delivery apparatus delivers the audio soundfields to the listener with first and second audio elements perceived by the listener as emanating from first and second virtual source locations, respectively, and with the first audio element and/or the second audio element delivered to the listener from a third virtual source location. The first virtual source location and the second virtual source location are perceived by the listener as being located to the front of the listener, and the third virtual source location is located to the rear or the side of the listener.

Abstract: Various disclosed implementations involve processing and/or playback of a recording of a conference involving a plurality of conference participants. Some implementations disclosed herein involve analyzing conversational dynamics of the conference recording. Some examples may involve searching the conference recording to determine instances of segment classifications. The segment classifications may be based, at least in part, on conversational dynamics data. Some implementations may involve segmenting the conference recording into a plurality of segments, each of the segments corresponding with a time interval and at least one of the segment classifications. Some implementations allow a listener to scan through a conference recording quickly according to segments, words, topics and/or talkers of interest.

Abstract: Various disclosed implementations involve processing and/or playback of a recording of a conference involving a plurality of conference participants. Some implementations disclosed herein involve receiving audio data corresponding to a recording of at least one conference involving a plurality of conference participants. The audio data may include conference participant speech data from multiple endpoints, recorded separately and/or conference participant speech data from a single endpoint corresponding to multiple conference participants and including spatial information for each conference participant of the multiple conference participants. A search of the audio data may be based on one or more search parameters. The search may be a concurrent search for multiple features of the audio data. Instances of conference participant speech may be rendered to at least two different virtual conference participant positions of a virtual acoustic space.

Abstract: Teleconference audio data including a plurality of individual uplink data packet streams, may be received during a teleconference. Each uplink data packet stream may corresponding to a telephone endpoint used by one or more teleconference participants. The teleconference audio data may be analyzed to determine a plurality of suppressive gain coefficients, which may be applied to first instances of the teleconference audio data during the teleconference, to produce first gain-suppressed audio data provided to the telephone endpoints during the teleconference. Second instances of the teleconference audio data, as well as gain coefficient data corresponding to the plurality of suppressive gain coefficients, may be sent to a memory system as individual uplink data packet streams. The second instances of the teleconference audio data may be less gain-suppressed than the first gain-suppressed audio data.

Abstract: Computer-implemented methods for training an acoustic model for a far-field utterance processing system are provided. The acoustic model may be configured to map an input audio signal into linguistic or paralinguistic units. The training may involve imparting far-field acoustic characteristics upon near-field training vectors that include a plurality of near-microphone utterance signals. Imparting the far-field acoustic characteristics may involve generating a plurality of simulated room impulse responses, convolving one or more of the simulated room impulse responses with the near-field training vectors, to produce a plurality of simulated far-field utterance signals and saving the results of the training in one or more non-transitory memory devices corresponding with the acoustic model. Generating simulated room impulse responses may involve simulating room reverberation times but not simulating early reflections from room surfaces.

Abstract: Disclosed is a method and apparatus operative to process packets of media received from a network including a receiver unit operative, a jitter buffer data structure and a playback head defining a point in the jitter buffer data structure from which the ordered queue of packets are to be played back, and at least one prototype head. Each prototype head having a predetermined latency assigned thereto and defining a point in the jitter buffer data structure from which the ordered queue of packets is being played back containing said latency a processor operable to determine a measure of conversational quality associated with the ordered queue of packets being played back by each prototype head. Also described is a head selector operable to compare the measures of conversational quality associated with the ordered queue of packets being played back by each prototype head to select the prototype head with the highest measure of conversational quality and a playback unit coupled to the playback head.

Abstract: Described herein is a method for creating an object-based audio signal from an audio input, the audio input including one or more audio channels that are recorded to collectively define an audio scene. The one or more audio channels are captured from a respective one or more spatially separated microphones disposed in a stable spatial configuration. The method includes the steps of: a) receiving the audio input; b) performing spatial analysis on the one or more audio channels to identify one or more audio objects within the audio scene; c) determining contextual information relating to the one or more audio objects; d) defining respective audio streams including audio data relating to at least one of the identified one or more audio objects; and e) outputting an object-based audio signal including the audio streams and the contextual information.

Abstract: A computer implemented system for rendering captured audio soundfields to a listener comprises apparatus to deliver the audio soundfields to the listener. The delivery apparatus delivers the audio soundfields to the listener with first and second audio elements perceived by the listener as emanating from first and second virtual source locations, respectively, and with the first audio element and/or the second audio element delivered to the listener from a third virtual source location. The first virtual source location and the second virtual source location are perceived by the listener as being located to the front of the listener, and the third virtual source location is located to the rear or the side of the listener.

Abstract: A method for processing audio data, the method comprising: receiving audio data corresponding to a plurality of instances of audio, including at least one of: (a) audio data from multiple endpoints, recorded separately or (b) audio data from a single endpoint corresponding to multiple talkers and including spatial information for each of the multiple talkers; rendering the audio data in a virtual acoustic space such that each of the instances of audio has a respective different virtual position in the virtual acoustic space; and scheduling the instances of audio to be played back with a playback overlap between at least two of the instances of audio, wherein the scheduling is performed, at least in part, according to a set of perceptually-motivated rules.

Abstract: Systems, methods, and computer program products for creating an object-based audio signal from an audio input are described. The audio input includes one or more audio channels that are recorded to collectively define an audio scene. The one or more audio channels are captured from a respective one or more spatially separated microphones disposed in a stable spatial configuration. A system receives the audio input. The system performs spatial analysis on the one or more audio channels to identify one or more audio objects within the audio scene. The system determines contextual information relating to the one or more audio objects. The system defines respective audio streams including audio data relating to at least one of the identified one or more audio objects. The system then outputs an object-based audio signal including the audio streams and the contextual information.