Transform Block Sub-Block Entropy Coding for Video Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards, such as HEVC and VVC, face inefficiencies in entropy encoding and decoding transform coefficient levels, particularly in handling non-zero coefficients and requiring multiple sub-block coding passes, which can increase complexity and reduce compression efficiency.
Innovation Solution
The proposed method involves dividing transform blocks into sub-blocks and using multiple sub-block coding passes to entropy encode or decode each sub-block, with specific syntax elements being processed in optimized passes to reduce complexity and improve efficiency, including the use of Transform Skip Mode for blocks with no non-zero coefficients and adaptive allocation of context-coded bins.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple sub-block coding passes are used to entropy encode transform coefficient levels, then compression efficiency is improved, but device complexity increases
Solution Approach 1:
The transform block is divided into multiple sub-blocks, and each sub-block is processed separately in different coding passes. This segmentation allows the encoder to handle non-zero coefficients more efficiently by focusing computational resources on relevant sub-blocks, thereby improving compression efficiency while managing complexity through localized processing.
Solution Approach 2:
The coding process uses multiple passes with different levels of detail, where the first pass handles significant coefficients and subsequent passes refine the encoding. This dynamic approach allows the system to adapt the coding complexity to the actual content of each transform block, improving overall compression efficiency without uniformly increasing complexity across all blocks.
2Productivity
If multiple sub-block coding passes are used to entropy decode transform coefficient levels, then compression efficiency is improved, but processing time increases
Solution Approach 1:
The first coding pass performs preliminary encoding of the most significant coefficients, allowing the decoder to reconstruct a coarse version of the transform block early. Subsequent passes add refinement details, enabling progressive decoding that balances compression efficiency with time-constrained applications by providing usable output before all passes complete.
3Measurement precision
If context-coded bins are allocated for each syntax element in multiple coding passes, then encoding precision is improved, but the number of bins required increases
Solution Approach 1:
Different syntax elements are assigned context-coded bins based on their local importance and probability characteristics within each sub-block. High-probability or significant syntax elements receive dedicated context models with higher precision, while less critical elements use simpler encoding, thereby maintaining overall encoding precision while reducing the total number of bins required compared to uniform high-precision encoding of all elements.
Data Source
AI summary
Video processing methods and apparatuses in a video encoding or decoding system for processing residual data of transform blocks. A current transform block in a video picture is divided into multiple sub-blocks, and each sub-block in the current transform block is entropy encoded or decoded using multiple sub-block coding passes. Each transform coefficient level in each sub-block is visited once according to a pre-defined scanning order in each sub-block coding pass. Two or more syntax elements associated with transform coefficient levels in the current transform block are aggregated and coded in one single sub-block coding pass. The video processing method is applied if the current transform block contains at least one non-zero transform coefficient level and the current transform block is coded in Transform Skip Mode (TSM) according to some embodiments.