Amazon Technologies, Inc.

Michael Klingbeil

Siddhartha Shankara Rao

John Joseph Dunne

Adam Marks

Arvindh Krishnaswamy

Michael Mark Goodwin

Christopher LaFata

Carlin Wiegner

Karim Helwani

Sanjeev Kumar

Pratik GuhaRay

Rakshit Karnawat

Tobias DiPasquale

Erfan Soltanmohammadi

Madhuri R Marri

Ankit Kumar Yadav

Hoang Do

William Burnson

Hiten Vasnani

Steven Sabonjohn

Charlotte Barge

John Robert Ciancutti

Puneet K Puri

Sid Shankara Rao

Justin Rowe

An audio assessment machine learning model, for example including one or more neural networks, may be trained for human audio quality assessment estimation, wherein the training comprises performing comparisons of a plurality of training machine learning audio quality assessments of training audio content to a plurality of human audio quality assessments of the training audio content and adjusting the audio assessment machine learning model based on the comparisons. After the training, a first audio analysis of first audio content may be performed by the audio assessment machine learning model. A first machine learning audio quality assessment of the first audio content may be provided, by the audio assessment machine learning model, based on the first audio analysis. The first machine learning audio quality assessment may include a quality score for the first audio content and a quality degradation reason for the first audio content.

Machine learning model that estimates human audio quality assessments

A synchronization system or module controls the transmission of voice or music (or other media) during a media program, consistent with a state of the media program. When a media program is in a voice state, a creator of a media program may speak or sing words that are captured and transmitted to devices of listeners. When the media program is in a media state, the creator may instruct or command that a media entity that has been pre-buffered on the devices of the listeners may be played to such listeners. A synchronization system or module that interprets instructions or commands of a creator or interprets interactions of a creator of the media program may likewise instruct or command devices of the listeners to play voice or media, in a synchronized manner.

Synchronizing live and pre-recorded content of media programs

Implementations for compositing two input signals to form a higher quality signal are described. A first input signal is received from a first input device and a second input signal is received from a second input device. The first input signal and the second input signal are combined into a composite input signal. It is then determined that the composite input signal has a higher quality than either the first input signal or the second input signal individually. Based on that determination, the composite input signal is selected for use by the media conferencing service as part of a media conference.

Automatic signal composition for media conferences

Implementations for selecting an input device based on characteristics of the input signals from those input devices are described. A first input signal is received from a first input device of a participant device participating in a media conference and a second input signal is received from a second input device of the participant device. A first characteristic of the first input signal and a second characteristic of the second input signal are determined. The first characteristic is compared to the second characteristic. It is determined that a quality of the second input signal is greater than a quality of the first input signal based on comparing the first characteristic to the second characteristic. The second input device is selected based on determining that the quality of the second input signal is greater than the quality of the first input signal.

Automatic input device selection for media conferences

A plurality of device-sharing participants may be detected that are participating in a videoconference via a shared computing device. The detecting of the plurality of device-sharing participants may be performed based, at least in part, on at least one of an audio analysis of captured audio from one or more microphones or a video analysis of captured video from one or more cameras. A plurality of participant connections corresponding to the plurality of device-sharing participants may be joined to the videoconference. Each of the plurality of participant connections may be identified within the videoconference using a respective name. A plurality of video streams and a plurality of audio streams corresponding to the plurality of participant connections may be transmitted, and the plurality of video streams and the plurality of audio streams may be presented to at least one other conference participant.

Separate representations of videoconference participants that use a shared device

This disclosure describes selecting the best hosting server(s) in a particular geographical location for the majority of participants in an interactive electronic activity. A request for an electronic interactive activity involving a plurality of nodes is received. At least one factor related to a plurality of potential servers is evaluated with respect to the plurality of nodes, where at least two of the plurality of potential servers are located in different geographical locations. Based at least in part on evaluating the factors, one or more of the potential servers is selected to host the electronic interactive activity, where the one or more of the potential servers selected are located in the same geographical location.

Selecting hosting servers for interactive electronic activities

A cloud-based video/audio conferencing system and method performs locale based aggregation of a full duplex media stream to organize multiple connections to a conference call that originate from the same physical location or a shared locale. The cloud-based video/audio conferencing system performs synchronization of the microphone and speaker audio signals of the same-locale connections. In this manner, a conference call may be held with multiple user devices making connections from the same physical location. User experience is enhanced by allowing each user in the same location to use his/her own individual devices to connect to the same conference call.

Dynamic locale based aggregation of full duplex media streams

A technology for estimating a delay between a far-end audio signal and a near-end audio signal for acoustic echo cancellation is disclosed. A copy of the far-end signal is stored in a speaker buffer and organized in chunks, and a copy of the near-end signal is stored in a microphone buffer and organized in chunks. Cross correlation is performed on each pair of speaker chunks and microphone chunks based on &#x3b2;-PHAse Transform ('PHAT') generalized cross correlation ('GCC'). A peak correlation value can be obtained for each pair of the chunks. Offset values corresponding to the peak correlation values are collected and clustered. A best cluster is selected and the offset value represented by the selected cluster is identified as the estimated delay. Acoustic echo cancellation can be performed on the near-end signal based on the estimated delay.

Delay estimation for acoustic echo cancellation

A latency compensating adaptive jitter buffer method is implemented. In response to a queuing delay event being detected, the size of the jitter buffer is temporarily increased to receive all incoming audio packets up to a certain refill level. The method then produces audio playback using accelerated playback to reduce the jitter buffer size back to the nominal size.

Adaptive audio stream with latency compensation cross reference to other applications

A latency compensating adaptive jitter buffer method is implemented in an audio client running on a user device or in an audio server to adaptively adjust the size of a jitter buffer to optimize latency while minimizing packet loss during audio signal transmission. In some embodiments, the jitter buffer is kept to a nominal size for low latency. In response to a queuing delay event being detected, audio production is temporarily stopped and the size of the jitter buffer is temporarily increased to receive all incoming audio packets up to a certain refill level. The method then resumes audio production using accelerated playback to reduce the jitter buffer size back to the nominal size.

Growing community of inventors

New Haven, CT, United States of America

Michael Klingbeil