Variable coding resolution in video codec

摘要

A video codec provides for encoding and decoding pictures of a video sequence at various coded resolutions, such that pictures can be encoded at lower coded resolutions based on bit rate or other constraints while maintaining a consistent display resolution. The video codec employs a coding syntax where a maximum coded resolution is signaled at the sequence level of the syntax hierarchy, whereas a lower coded resolution is signaled at the entry point level for a segment of one or more intra-coded frames and frames predictively encoded based thereon. This allows the use of a separate out-of-loop resampler after the decoder to up-sample the pictures to the display resolution.

主权项

1. A method of decoding video coded using variable coded resolution, comprising, with a computing device that implements a video decoder:
receiving a video bit stream, where a sequence header is integral to the video bit stream, the video bit stream having a plurality of entry point segments representing separate sub-sequences of pictures of a video sequence; decoding at least one syntax element from the sequence header, the least one syntax element from the sequence header specifying a maximum coded resolution for the video sequence; for a given one of the entry point segments, determining whether to use the maximum coded resolution for the video sequence as coded resolution for pictures of the given entry point segment and, if not, determining the coded resolution for the pictures of the given entry point segment by decoding at least one syntax element of an entry point header for the given entry point segment, the at least one syntax element of the entry point header specifying the coded resolution for the pictures of the given entry point segment; decoding said given entry point segment; and reconstructing at least part of the video sequence with the decoded entry point segment.