Video Coding Mode Selection by Block and VPDU Size
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing Versatile Video Coding (VVC) standard faces inefficiencies due to independent definitions of maximum transform size and CTU size, leading to suboptimal block partition processes and coding tools that do not align with Virtual Pipeline Data Unit (VPDU) sizes, especially when CTU sizes exceed 128x128, and the need for redesigning block partition structures and signaling to accommodate varying resolutions across bitstreams.
Innovation Solution
Proposes conditional enabling of ternary-tree (TT) and binary-tree (BT) splits based on Virtual Pipeline Data Unit (VPDU) dimensions, allowing configurable CTU sizes larger than 128x128, and disabling or modifying coding tools for blocks exceeding certain dimensions to align with maximum transform sizes, thereby optimizing block partitioning and coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If CTU sizes are configured larger than 128x128 to reduce the number of coding units, then the block partition process becomes misaligned with VPDU sizes and transform sizes, but this leads to reduced coding efficiency and suboptimal performance
Solution Approach 1:
The patent introduces dynamic conditional logic that adapts the block partition process based on the relationship between CTU size and VPDU size. When CTU size exceeds VPDU size, the system dynamically adjusts partitioning behavior and coding tool availability, allowing the system to optimize for large-block scenarios while maintaining alignment constraints. This dynamic adaptation resolves the contradiction by making the partition structure flexible rather than fixed.
Solution Approach 2:
The patent changes the state of coding tools and partitioning options based on the parameter relationship between CTU size and VPDU size. Specifically, it enables or disables certain coding tools conditionally when block dimensions exceed VPDU dimensions, and modifies the maximum transform size parameter to align with VPDU boundaries. This parameter-based control mechanism allows efficient large-CTU processing while maintaining system consistency.
2Adaptability or versatility
If independent maximum transform size and CTU size definitions are used to provide flexibility, then coding tools do not align with VPDU sizes, but this causes suboptimal block partition processes
Solution Approach 1:
The patent introduces conditional parameter adjustments that modify the effective maximum transform size and coding tool availability based on the relationship between block dimensions and VPDU size. When a block exceeds VPDU dimensions, the system automatically adjusts parameters to enforce alignment, ensuring that transform operations and coding tools remain consistent with VPDU boundaries while preserving the ability to configure various CTU sizes.
Solution Approach 2:
The system dynamically adapts the behavior of coding tools and partitioning mechanisms based on real-time dimension comparisons. Rather than using fixed independent definitions, the patent creates a dynamic system where the availability and parameters of coding tools are adjusted according to whether block sizes align with or exceed VPDU sizes, optimizing performance for each specific configuration.
3Productivity
If coding tools are enabled for all block sizes to maximize functionality, then processing complexity increases, but this leads to reduced coding efficiency for large blocks
Solution Approach 1:
The patent implements parameter-based conditional enabling of coding tools, where the availability of specific tools is determined by comparing block dimensions against VPDU size thresholds. For blocks exceeding VPDU dimensions, certain coding tools are automatically disabled or modified, reducing the effective complexity of the coding process for large blocks while maintaining full functionality for smaller, alignment-compatible blocks.
Solution Approach 2:
The patent applies different coding tool configurations to different block size categories. Rather than uniformly enabling or disabling tools across all block sizes, it creates local optimizations where coding tools are selectively enabled or disabled based on the specific dimension relationships, allowing efficient processing tailored to each block size scenario.
Data Source
AI summary
Methods, systems, and devices for coding or decoding video wherein the picture partition mode is based on block size are described. An example method for video processing includes using a dimension of a virtual pipeline data unit (VPDU) used for a conversion between a video comprising one or more video regions comprising one or more video blocks and a bitstream representation of the video to perform a determination of whether a ternary-tree (TT) or a binary tree (BT) partitioning of a video block of the one or more video blocks is enabled, and performing, based on the determination, the conversion, wherein the dimension is equal to VSize in luma samples, wherein dimensions of the video block are CtbSizeY in luma samples, wherein VSize=min(M, CtbSizeY), and wherein M is a positive integer.