Variable Transform Size Signaling for VVC Coding Tools
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The interaction between the maximum transform size and other transform coding tools in video encoding and decoding, particularly in the Versatile Video Coding (VVC) standard, is not adequately addressed when the maximum transform size varies between 32 and 64, leading to inefficiencies in processes such as zero-out, multiple transform selection, chroma transform size, transform skip, and block-based delta pulse code modulation (BDCM).
Innovation Solution
Adapting the signaling of zero-out process, multiple transform size, chroma transform size, transform skip, and BDCM tools based on the maximum transform size, allowing these processes to function effectively even when the maximum transform size is 32, rather than the previously fixed 64, by modifying the coding tools to accommodate the variable maximum transform size.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the maximum transform size is reduced to 32 to reduce complexity, then device complexity is reduced, but coding efficiency deteriorates
Solution Approach 1:
The patent applies dynamics by making the maximum transform size configurable rather than fixed. The system can dynamically adjust the maximum transform size between 32 and 64 based on the specific coding scenario, allowing optimization of both complexity and efficiency in different contexts
Solution Approach 2:
The patent changes the parameter of maximum transform size from a fixed value to a variable parameter that can take different values (32 or 64). This parameter change enables the system to adapt to different coding requirements, resolving the contradiction between complexity reduction and efficiency maintenance
2Productivity
If the maximum transform size is fixed at 64 to maintain coding efficiency, then coding efficiency is improved, but device complexity increases
Solution Approach 1:
The system dynamically selects the maximum transform size based on the coding scenario. When high coding efficiency is needed, it uses 64; when complexity reduction is prioritized, it uses 32. This dynamic adaptation resolves the contradiction by choosing the appropriate value contextually
Solution Approach 2:
The patent segments the transform coding process into different modes based on block size and coding requirements. By dividing the processing into different pathways (using 32 or 64 transform size depending on the situation), the system can maintain efficiency where needed while reducing complexity where appropriate
3Adaptability or versatility
If transform coding tools are adapted for variable maximum transform size, then adaptability is improved, but device complexity increases
Solution Approach 1:
The patent creates a universal transform coding framework that can handle both 32 and 64 maximum transform sizes using the same core tools. The coding tools are designed with multi-functionality to accommodate different transform sizes through standardized processing paths, reducing the need for separate specialized tools for each size
Solution Approach 2:
The coding tools incorporate dynamic elements that allow them to adapt their behavior based on the maximum transform size parameter. This dynamic capability enables a single set of tools to handle variable requirements without requiring multiple specialized toolsets, thus improving adaptability while controlling complexity
Data Source
AI summary
Methods and apparatus for implementing Discrete Trigonometric transforms are based on maximum transform size. In one embodiment, matrix-based intra prediction is enabled for coding unit sizes up to a specified size, regardless of the maximum transform size. In another embodiment, low-frequency non-separable transforms are used to improve coding gain. Syntax in a bitstream can be used to indicate a coding tool that is used.