Methods, systems, and devices for improved video throughput using deep learning video coding are described. A device may receive a bitstream including a set of video frames. The device may batch the set of video frames into a first subset of video frames and a second subset of video frames based on a change in a reference scene associated with the set of video frames. The device may select a mode of operation for a neural processing unit of the device based on the batching. The device may generate a set of video packets including the first subset of video frames, the second subset of video frames, or both, based on the neural processing unit and the selected mode of operation

Embodiments provide for a system, comprising a plurality of encoders configured to generate a plurality of variant streams, and a cross-variant Instantaneous Decoder Refresh (IDR) identifier configured to inspect the plurality of variant streams, identify IDR frames in each of the plurality of variant streams, determine which IDR frames correspond to cross-variant boundaries, and demarcate the IDR frames corresponding to cross-variant boundaries.

Embodiments provide for a system, comprising a plurality of encoders configured to generate a plurality of variant streams, and a cross-variant Instantaneous Decoder Refresh (IDR) identifier configured to inspect the plurality of variant streams, identify IDR frames in each of the plurality of variant streams, determine which IDR frames correspond to cross-variant boundaries, and demarcate the IDR frames corresponding to cross-variant boundaries.

A method, apparatus, device and medium for coding a video are provided. An implementation plan includes: determining an initial video frame structure based on a preset threshold for a B-frame number; in response to determining that the initial video frame structure meets a preset condition, shortening the initial video frame structure to obtain a candidate video frame structure set; determining a target video frame structure from the candidate video frame structure set; and performing video coding on video frames in a to-be-coded video frame sequence according to the target video frame structure.

A method, computer program, and computer system is provided for splitting viewport bitstreams. A first coded video bitstream is at least partially encoded using a first Group-of-Pictures (GOP) structure at a first resolution. A second coded video bitstream structured into a plurality of tiles is encoded using a second GOP structure at a second resolution, whereby the second GOP structure includes fewer coded pictures than the first GOP structure. A streaming bitstream for decoding or rendering is created using the first coded video bitstream and the second coded bitstream.

A data coding method includes determining a length of a current first bit-rate-control-unit data frame in to-be-transmitted data, and the length of the first bit-rate-control-unit data frame is less than or equal to a length of a group of pictures (GOP) data frame. The method also includes determining a first target bit rate of the first bit-rate-control-unit data frame according to the length of the first bit-rate-control-unit data frame, a target bit rate of the to-be-transmitted data, and a frame rate of the to-be-transmitted data and, when a first difference value between a real bit rate of a second bit-rate-control-unit data frame and a second target bit rate of the second bit-rate-control-unit data frame is greater than zero, adjusting the first target bit rate according to the first difference value, and coding the first bit-rate-control-unit data frame according to the adjusted first target bit rate.

A data coding method includes determining a length of a current first bit-rate-control-unit data frame in to-be-transmitted data, and the length of the first bit-rate-control-unit data frame is less than or equal to a length of a group of pictures (GOP) data frame. The method also includes determining a first target bit rate of the first bit-rate-control-unit data frame according to the length of the first bit-rate-control-unit data frame, a target bit rate of the to-be-transmitted data, and a frame rate of the to-be-transmitted data and, when a first difference value between a real bit rate of a second bit-rate-control-unit data frame and a second target bit rate of the second bit-rate-control-unit data frame is greater than zero, adjusting the first target bit rate according to the first difference value, and coding the first bit-rate-control-unit data frame according to the adjusted first target bit rate.

Live video streams are produced using a network server system. The network server system initially receives, via a network, one or more captured video streams that are live streams captured from a remotely-located camera, phone or the like. The captured streams are forwarded to a control device via the network to thereby permit a user of the control device to select one of the captured video streams for output to the video production stream. In response to a command received from the control device that indicates the selected capture stream, the selected capture video streams is encoded for output as the video production stream.

Live video streams are produced using a network server system. The network server system initially receives, via a network, one or more captured video streams that are live streams captured from a remotely-located camera, phone or the like. The captured streams are forwarded to a control device via the network to thereby permit a user of the control device to select one of the captured video streams for output to the video production stream. In response to a command received from the control device that indicates the selected capture stream, the selected capture video streams is encoded for output as the video production stream.

A first aspect is a method for coding a group of pictures (GOP) that includes frames of a video. The method includes encoding, at least some of the frames of the GOP, using a first encoding pass to obtain encoding statistics; obtaining, using the encoding statistics, respective temporal dependency likelihoods (TDLs) for the at least some of the frames of the GOP, where the respective TDLs indicate contributions that the at least some of the frames make in reducing prediction errors of the GOP; obtaining a reference frame based on the respective TDLs; and using the reference frame in encoding at least some of the frames of the GOP in a second encoding pass.