Roku, Inc.

Wim Michiels

Jim Funk

David Sharp

Jeff Bush

Dale Luck

David Westerhoff

Andrey Marsavin

Govind Vaidya

Beth Teresa Logan

Vijay Anand Raghavan

Tapan Oza

Zheng Han

Amit Paliwal

Ilya Asnis

Rodrigo Emiliano Tocalini

Robert Fitzgerald Watts

Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for performing adaptive audio bitrate selection. An example embodiment operates by receiving a data streaming request. The embodiment then streams data at a first audio bitrate. The embodiment then determines whether there is a streaming discontinuity associated with the data. The embodiment then determines, in response to the determination that there is a streaming discontinuity, whether a network condition satisfies one or more conditions based on one or more streaming parameters. The embodiment then, in response to the determination that the network condition satisfies the one or more conditions, adjusts the first audio bitrate to a second audio bitrate during the streaming discontinuity. The embodiment then continues to stream the data, subsequent to the streaming discontinuity, at the second audio bitrate to complete the data streaming request.

Adaptive audio bitrate selection

Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for modifying one or more parameters of a data streaming bitrate selection algorithm based on machine learning. An example embodiment operates by training and operating a first machine learning model to predict a sustainable network bandwidth. A second machine learning model is trained to receive the sustainable network bandwidth and predict a likelihood that this network bandwidth will not empty a data buffer of streaming data. A bitrate is selected based on the likelihood being below a threshold percentage, such as 50%.

Machine learning for adaptive bitrate selection

Disclosed herein are various embodiments for server-side scene change content stitching. An example embodiment operates by providing a stream of first content for display on a display device. Second content to be stitched into the stream is received. Metadata indicates a point in time in the stream into which the second content is to be stitched. A scene change is identified in the first content closest to the point in time. A first frame proximate to the scene change is identified. The stream is displayed on the display device, including the second content stitched into the first content.

Server-side streaming content stitching

Disclosed herein are various embodiments for server-side scene change content stitching. An example embodiment operates by providing a stream of first content for display on a display device. Second content to be stitched into the stream is received. Metadata indicates a point in time in the stream into which the second content is to be stitched. A scene change is identified in the first content closest to the point in time. A first frame proximate to the scene change is identified. A distance within the plurality of frames between the identified first frame and the identified scene change is identified. The stream is displayed on the display device, including the second content stitched into the first content.

Server-side scene change content stitching

Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for stitching second content into a stream of first content, where the audio and/or video format of the first content may differ from the second content. An example media device embodiment operates by decoding first content using a first decoder compatible with the first content. The decoded first content is displayed on a display device. The media device then receives second content for stitching into the display of the first content. The media device determines whether an audio format and a video format of the second content match an audio format and a video format of the first content. If the audio format and the video format match, then the media device causes a program time stamp (PTS) of the second content to be consecutive with a PTS of the first content during the stitching, such that a reference clock is also consecutive. This provides continuity during the stitching. The media device causes the first decoder to decode the second content, and then displays the decoded second content on the display device using the PTS of the second content and the reference clock. Also provided herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for stitching second content into a stream of first content, such that the stitching occurs at scene change boundaries of the first content, using single or multiple decoders.

Client side stitching of content into a multimedia stream

Disclosed herein are system, apparatus, article of manufacture, method, and/or computer program product embodiments for providing a representation to a connected device. An embodiment operates by recognizing a new device connected to a display device, collecting device fingerprint information from the new device, and requesting a device class representation information determined based on the device fingerprint information. Another embodiment operates by receiving device fingerprint information from a display device via a network connection, wherein the device fingerprint information is collected from a device connected to a display device, and providing device class representation information to the display device via the network connection, when the device class representation information corresponding to the device fingerprint information is available.

Providing a representation for a device connected to a display device

Disclosed herein system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for using technology in innovative ways to automatically and intelligently predict performance of content delivery networks (CDNs) in a crowdsourced manner in order to deliver the highest resolution content possible with little or no buffering. In some embodiments, a media device downloads and displays primary content on a display device. The media device determines that secondary content should be stitched into the display of the primary content. Accordingly, the media device requests that a crowdsource server determine a bit-rate variant of the secondary content to download. The crowdsource server determines the bit-rate variant in a crowdsourced manner such that subsequent download of the identified bit-rate variant by the media device results in a zero buffer experience. The media device downloads the determined bit-rate variant of the secondary content, and stitches the downloaded bit-rate variant of the secondary content into the display of the primary content.

Growing community of inventors

Los Gatos, CA, United States of America

Wim Michiels