Abstract:
Techniques are provided for optimizing quality in adaptive streaming, where information is obtained relating to encoding bitrates and content quality scores for content segments of a content stream. The content segments of the content stream are provided by a content server and downloaded by a computing device. For each content segment, an encoding bitrate is determined for downloading the content segment from the content server, where determining the encoding bitrate for a current content segment is based upon an estimated current network bandwidth, constraints of a buffer into which downloaded content segments are stored, and also upon quality scores associated with content segments within a selected horizon extending from the current content segment to a future content segment in the content stream at an endpoint of the horizon. Each content segment with the determined bitrate for the content segment is downloaded from the content server.
Abstract:
Content is received within a network by a content receiver based upon fetch requests by the content receiver to a content source, where the fetch requests for content are based upon a first adaptation logic scheme to fill a buffer of the content receiver at a selected rate and to download segments of content at selected encoding rates. A network priority is detected at the content receiver that includes an observed download rate at the content receiver that is greater than a fair share rate for other content receivers in the network. Fetch requests for content are revised in response to detection of the network priority, where the revised fetch requests are based upon a priority adaptation scheme that results in at least one of filling the buffer at a faster rate and downloading segments of content at higher encoding rates compared with the first adaptation logic scheme.
Abstract:
Techniques are provided for optimizing quality in adaptive streaming, where information is obtained relating to encoding bitrates and content quality scores for content segments of a content stream. The content segments of the content stream are provided by a content server and downloaded by a computing device. For each content segment, an encoding bitrate is determined for downloading the content segment from the content server, where determining the encoding bitrate for a current content segment is based upon an estimated current network bandwidth, constraints of a buffer into which downloaded content segments are stored, and also upon quality scores associated with content segments within a selected horizon extending from the current content segment to a future content segment in the content stream at an endpoint of the horizon. Each content segment with the determined bitrate for the content segment is downloaded from the content server.
Abstract:
In one embodiment, a method that receives a manifest for plural encoded representations of a single content stream, each representation fragmented into plural chunks, each representation comprising a different quality level, the manifest listing a plurality of representations, each representation comprising the plural chunks at one of a plurality of quality levels, and requests one of the plural chunks based on selection of one of the plurality of quality levels explicitly indicated in the manifest.
Abstract:
In one method embodiment, receiving a video stream comprising a first compressed picture without associated time stamp information and a second compressed picture having associated first time stamp information, the second compressed picture following the first compressed picture in transmission order; deriving second time stamp information based on the first time stamp information; and processing the first compressed picture based on the second time stamp information.
Abstract:
A method is provided in one example embodiment and includes receiving media from an input; gathering metadata from a plurality of media stream representations that are provided by the input; creating one or more metadata segments corresponding to the plurality of media stream representations in order to generate a metadata track; and making at least a portion of the metadata track available to a client device, where the portion of the metadata track is provided separate from a manifest file.
Abstract:
Content is received within a network by a content receiver based upon fetch requests by the content receiver to a content source, where the fetch requests for content are based upon a first adaptation logic scheme to fill a buffer of the content receiver at a selected rate and to download segments of content at selected encoding rates. A network priority is detected at the content receiver that includes an observed download rate at the content receiver that is greater than a fair share rate for other content receivers in the network. Fetch requests for content are revised in response to detection of the network priority, where the revised fetch requests are based upon a priority adaptation scheme that results in at least one of filling the buffer at a faster rate and downloading segments of content at higher encoding rates compared with the first adaptation logic scheme.
Abstract:
A method is provided in one example embodiment and includes receiving at a first network node a request to obtain data from a second network node, wherein the first and second network nodes are connected via n access networks; partitioning the request into n subrequests proportionally based on relative throughputs of the n access networks; and transmitting each of the n subrequests to the second network node via a respective one of the n access networks.
Abstract:
In one embodiment, a client device includes a processor to send a first GET request to a server for a first data chunk of a data item, a communication interface to receive packets of the first data chunk, a buffer to buffer the packets, wherein the processor is operative to obtain a first measure of round trip time (RTT), the processor is operative to estimate a time t1 when receipt of the last packet of the packets of the first data chunk in the buffer will be completed, the processor is operative to determine a next fetch time, t2, to send a second GET request to the server for a second data chunk of the data item, based on the time t1 adjusted by a portion of the first measure of RTT, the processor is operative to send the second GET request to the server at about time t2.
Abstract:
In one embodiment, a method that receives a manifest for plural encoded representations of a single content stream, each representation fragmented into plural chunks, each representation comprising a different quality level, the manifest listing a plurality of representations, each representation comprising the plural chunks at one of a plurality of quality levels, and requests one of the plural chunks based on selection of one of the plurality of quality levels explicitly indicated in the manifest.