In-loop Filtering Signaling Across Video Slice Boundaries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards face challenges in efficiently signaling in-loop filtering operations across tile and slice boundaries, LMCS information, and transform block sizes, leading to unnecessary signaling and potential alignment issues in ALF virtual boundaries.
Innovation Solution
Conditionally signaling loop_filter_across_tiles_enabled_flag and loop_filter_across_slices_enabled_flag based on video unit partitioning, signaling LMCS information in both picture and slice headers, and adjusting maximum transform block sizes relative to coding block sizes to optimize bitstream efficiency and alignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If in-loop filtering operation is always performed across tile boundaries, then filtering quality is improved, but bitstream size increases due to unnecessary signaling
Solution Approach 1:
The patent changes the parameter of flag inclusion based on the partitioning configuration. When tiles are used for video unit partitioning, the loop-filter-across-tiles-enabled-flag is included in the bitstream; when slices are used, the flag is excluded. This parameter change resolves the contradiction by adapting the signaling to the actual processing needs of different partitioning modes.
Solution Approach 2:
The patent segments the video processing into different modes (tile-based and slice-based partitioning) and applies different signaling rules to each segment. This allows the system to optimize bitstream efficiency for each specific partitioning mode while maintaining filtering quality where needed.
2Adaptability or versatility
If LMCS information is signaled at both picture and slice levels, then coding flexibility is improved, but bitstream complexity increases
Solution Approach 1:
The patent makes the picture header structure universal by conditionally including LMCS information at the picture level when slice-based partitioning is used. This allows the same header structure to serve both tile-based and slice-based modes, providing coding flexibility without unnecessarily increasing complexity for all cases.
Solution Approach 2:
The patent applies partial action by including LMCS information at the picture level only when necessary (for slice-based partitioning), rather than always including it. This reduces unnecessary signaling while maintaining the flexibility needed for advanced coding modes.
3Productivity
If maximum transform block size is increased, then coding efficiency is improved, but alignment accuracy with ALF virtual boundaries deteriorates
Solution Approach 1:
The patent changes the parameter of maximum transform block size based on the relationship with coding block size and ALF virtual boundary positions. By dynamically adjusting this parameter according to the specific video region and boundary locations, the system achieves optimal coding efficiency while maintaining accurate alignment with virtual boundaries.
Solution Approach 2:
The patent applies local quality by allowing different maximum transform block sizes for different video regions depending on their proximity to ALF virtual boundaries. This enables optimized coding efficiency in each local region while ensuring proper alignment where boundaries are present.
Data Source
AI summary
A method includes performing a conversion between a video unit of a video and a bitstream of the video. The bitstream conforms to a format rule. The format rule specifies that a syntax element indicating whether an in-loop filtering operation is performed across slice boundaries is selectively included in the bitstream based on whether or how the video unit is partitioned into slices.