NAL Unit Picture Partition Coding for Parallel Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding systems face inefficiencies in encoding and decoding due to the fixed representation of additional syntax elements, which limits the number of offset syntax elements and increases bitstream size, and not all decoders can decode encoded picture partitions in parallel.
Innovation Solution
Entropy encoding of additional syntax elements using unsigned integer 0-th order Exp-Golomb coding, allowing for variable-length sequences that represent more offset syntax elements, and signaling entry points through Supplemental Enhancement Information messages to conserve bandwidth and align encoded picture partitions to byte boundaries for parallel decoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If fixed-length representation is used for syntax elements, then decoding is simple, but bitstream size increases and number of offset syntax elements is limited
Solution Approach 1:
The patent changes the representation parameter from fixed-length to variable-length encoding for syntax elements. Specifically, it uses exponential-Golomb coding for offset syntax elements and context-based adaptive binary arithmetic coding for other syntax elements, allowing the bitstream to use fewer bits when syntax elements have smaller values while maintaining simple parsing through defined termination conditions and length indicators.
2Device complexity
If fixed-length representation is used for syntax elements, then encoding is simple, but the number of offset syntax elements is limited
Solution Approach 1:
The patent changes the encoding parameter from fixed-length to variable-length, specifically employing exponential-Golomb coding for offset syntax elements. This allows the number of representable offset syntax elements to be limited only by the available bit depth rather than by a fixed-length constraint, enabling support for a large number of encoded picture partitions while maintaining manageable encoding complexity through the structured syntax definition.
3Productivity
If more offset syntax elements are added, then parallel decoding capability improves, but bitstream size increases
Solution Approach 1:
The patent applies variable-length encoding parameters to offset syntax elements, which allows the bitstream to efficiently represent a large number of encoded picture partitions without proportionally increasing bitstream size. The exponential-Golomb coding provides compact representations for offset values, enabling the inclusion of multiple offset syntax elements for parallel decoding while minimizing bandwidth consumption.
Solution Approach 2:
The patent segments the syntax elements into different categories with different encoding strategies. Offset syntax elements use exponential-Golomb coding, while other syntax elements use context-based adaptive binary arithmetic coding. This segmentation allows optimized encoding for each type, improving the efficiency of representing multiple encoded picture partitions for parallel decoding.
Data Source
AI summary
A video encoder generates a Network Abstraction Layer (NAL) unit that contains a plurality of encoded picture partitions of the video data. The video encoder generates a bitstream that includes a variable-length value that represents an entropy-encoded first syntax element, a variable-length value that represents an entropy-encoded second syntax element, and fixed-length values that represent offset syntax elements. Lengths of each of the offset syntax elements are determinable based on the first syntax element. A video decoder uses the first syntax element, the second syntax element, and the offset syntax elements when decoding the encoded picture partitions.