Unified Rounding in Sub-Block Video Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sub-block based prediction technologies face challenges such as higher bandwidth requirements due to small sub-block sizes, misalignment of motion vectors between luma and chroma components, and the need for unified constraints and rounding methods across different modes.
Innovation Solution
The proposed solution involves unified rounding methods for sub-block based prediction, where the size and alignment of sub-blocks are optimized for both luma and chroma components, and unified constraints are applied across merge and non-merge affine modes to improve hardware design and coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If sub-block based prediction is used to improve motion compensation precision, then manufacturing precision is improved, but device complexity increases due to higher bandwidth requirements and multiple processing modes
Solution Approach 1:
The current block is divided into multiple sub-blocks, and motion compensation is performed independently for each sub-block. This segmentation allows for more precise motion estimation at the sub-block level, improving overall motion compensation precision while managing complexity through localized processing
Solution Approach 2:
The patent applies unified rounding methods and constraints across both merge and non-merge affine modes, as well as across luma and chroma components. This universal approach simplifies the processing logic and reduces device complexity by eliminating the need for separate handling of different modes and components
2Manufacturing precision
If separate processing methods are used for luma and chroma components to maintain component-specific optimization, then manufacturing precision is improved, but device complexity increases due to misalignment and separate constraints
Solution Approach 1:
The patent merges the processing approaches for luma and chroma components by applying unified rounding methods and the same constraints to both components. This combining approach eliminates the complexity of maintaining separate processing pipelines while preserving the ability to handle component-specific characteristics through the unified framework
3Manufacturing precision
If different rounding methods are used for merge and non-merge affine modes to optimize each mode, then manufacturing precision is improved, but device complexity increases due to multiple rounding approaches
Solution Approach 1:
The patent implements a universal rounding method that is applied to both merge and non-merge affine modes. This single rounding approach replaces multiple mode-specific rounding methods, reducing device complexity while maintaining optimization through the unified constraint application across all modes
Data Source
AI summary
Devices, systems, and methods for unified rounding in sub-block based prediction are described. In a representative aspect, a method of video processing includes generating, for a processing of a current block of video, a pairwise merge candidate based on a pair of motion candidates, and performing, based on the pairwise merge candidate, a conversion between the current block and a bitstream representation of the video. In another representative aspect, a method of video processing includes generating, for a current block of video coded using a geometry partition mode, a uni-prediction motion candidate based on a scaled motion vector and a List0 motion vector, and performing, based on the uni-prediction motion candidate, a conversion between the current block and a bitstream representation of the video.