Transform Coefficient Entropy Coding with Dependent Quantization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in achieving high coding efficiency when adapting quantization based on previously quantized and coded data, which affects the interrelationship between data items and the availability of information for context adaptive entropy coding.
Innovation Solution
Implement context-adaptive entropy decoding and encoding of transform coefficients using significance, parity, and greatness flags, along with equiprobability entropy decoding, to efficiently handle dependent quantization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If dependent quantization is used to adapt quantization based on previously quantized and coded data, then coding efficiency is improved, but the interrelationship between data items becomes more complex and information availability for context adaptive entropy coding is reduced
Solution Approach 1:
The patent segments the transform coefficients into multiple groups based on their positions in the scan order, with each group having a specific size (e.g., first group of 4 coefficients, second group of 4 coefficients). This segmentation allows the encoder to process and code different groups of coefficients with different quantization parameters independently, reducing the complexity of managing interrelationships while maintaining the benefits of dependent quantization.
Solution Approach 2:
The patent applies different quantization parameters to different groups of transform coefficients based on their local characteristics and positions. Each group can have its own quantization parameter selected from available quantization parameters, allowing localized optimization of coding efficiency without requiring complex global optimization across all coefficients.
2Productivity
If context adaptive entropy coding is used to code transform coefficients, then coding efficiency is improved, but the availability of information for accurate coding is reduced due to dependency in quantization
Solution Approach 1:
The patent performs preliminary grouping of transform coefficients into distinct groups before applying quantization and entropy coding. This preliminary organization ensures that information about coefficient positions and group memberships is preserved and made available to the entropy coder, maintaining information availability while enabling efficient context adaptive coding.
Solution Approach 2:
The patent introduces group identifiers and quantization parameter indicators as intermediary elements that mediate between the dependent quantization process and the entropy coding process. These intermediaries carry information about the quantization decisions and group structures, ensuring that the entropy coder has access to necessary information for accurate and efficient coding.
Data Source
AI summary
Concepts for transform coefficient block coding are described which enable coding of coefficients of a transform block in a manner suitable for dependent quantization and effectively implementable by entropy coding in terms of coding efficiency.