Deblocking Block and Sub-Block Boundaries in Video Encoders
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video encoding and decoding technologies face challenges in efficiently deblocking video sequences, particularly when combining longer luma filters with sub-block deblocking, which can result in dependencies between filtering on different boundaries.
Innovation Solution
The proposed solution involves methods for deblocking that restrict the number of samples modified on certain boundaries, allowing for parallel friendly deblocking processes. Specifically, the methods include deblocking block boundaries with restricted sample modification and deblocking sub-block boundaries in a way that minimizes dependencies between different deblocking operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If longer luma filters are applied for block boundaries, then filtering effectiveness is improved, but dependencies between filtering operations increase
Solution Approach 1:
The patent segments the deblocking process into independent operations for different boundary types (block boundaries and sub-block boundaries). By separating these operations, the patent allows longer luma filters to be applied at block boundaries without creating dependencies with sub-block boundary filtering, thus maintaining filtering effectiveness while reducing operational complexity.
Solution Approach 2:
The patent applies different filtering strategies to different boundary types based on local characteristics. Block boundaries receive longer luma filters for enhanced filtering effectiveness, while sub-block boundaries use shorter filters or different processing. This localized approach optimizes filtering where needed without propagating dependencies across the entire image.
2Reliability
If sub-block deblocking is applied, then boundary smoothness is improved, but processing complexity increases
Solution Approach 1:
The patent segments the deblocking process into independent operations for block boundaries and sub-block boundaries. By separating these operations and processing them in parallel, the patent achieves boundary smoothness through sub-block deblocking without increasing overall processing complexity, as the segmented operations can be executed simultaneously rather than sequentially.
Solution Approach 2:
The patent applies sub-block deblocking selectively only where necessary (at sub-block boundaries) rather than uniformly across the entire image. This partial application maintains boundary smoothness at critical locations while avoiding the processing complexity that would result from applying the same operation throughout the entire block structure.
3Productivity
If parallel friendly deblocking is implemented, then processing efficiency is improved, but filtering strength may be reduced
Solution Approach 1:
The patent segments the deblocking operations into independent parallel tasks for block boundaries and sub-block boundaries. This segmentation enables simultaneous processing (improving productivity) while maintaining filtering strength through the use of longer luma filters at block boundaries, which operate independently of sub-block processing and thus do not compromise filter strength for the sake of parallelization.
Data Source
AI summary
A deblocking the method. The method includes deblocking a block boundary between a first block (B1) and a second block (B2) with the restriction that not more than 5 samples are modified on the B2 side of the block boundary as a result of the deblocking. The deblocking method also includes deblocking a sub-block boundary within the B2 block such that not more than 2 samples on the side of the sub-block boundary within the B2 block that is closest to the block boundary between B1 and B2 are modified by the deblocking.