Residual Coding for Transform Skip and BDPCM Bit-Rate Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in achieving efficient compression and reducing redundancy in video data, particularly in intra prediction and motion vector prediction, which affect bandwidth and storage requirements.
Innovation Solution
The implementation of block differential pulse-code modulation (BDPCM) and context-based adaptive binary arithmetic coding (CABAC) for transform skip mode, which enhances entropy coding and improves the efficiency of intra prediction and motion vector prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If traditional intra coding without transform is used, then device complexity is reduced, but coding efficiency deteriorates due to increased bit rate requirements
Solution Approach 1:
The patent introduces a dynamic mode selection mechanism that adaptively switches between transform skip mode and conventional transform mode based on block characteristics. This allows the system to optimize between complexity and efficiency dynamically, using transform skip for smooth blocks (reducing complexity) and conventional transform for detailed blocks (maintaining efficiency).
Solution Approach 2:
The patent changes the processing parameter by bypassing the transform stage entirely in transform skip mode, directly coding residual samples. This parameter change (from transformed coefficients to raw residuals) reduces computational complexity while maintaining acceptable coding efficiency through optimized residual coding strategies.
2Device complexity
If transform skip mode is implemented, then device complexity is reduced by skipping transform, but manufacturing precision deteriorates due to potential quality loss
Solution Approach 1:
The patent applies transform skip mode selectively based on local block characteristics, such as intra prediction mode and block content analysis. Smooth blocks with simple gradients are processed with transform skip to reduce complexity, while blocks requiring high precision maintain conventional transform processing, ensuring local quality is preserved where needed.
Solution Approach 2:
The patent incorporates rate-distortion optimization feedback to determine whether to apply transform skip mode. The encoder evaluates the impact of skipping transform on reconstruction quality and bit rate, using this feedback to make intelligent decisions that balance complexity reduction with quality preservation.
3Measurement precision
If conventional residual coding is used, then measurement precision is maintained, but loss of information increases due to higher bit rate requirements
Solution Approach 1:
The patent extracts and removes redundant information from the residual signal through optimized coding strategies in transform skip mode. By directly coding residual samples without transform, the patent eliminates transform-related overhead and focuses compression efforts on the most significant residual components, reducing bit rate while maintaining essential precision.
Solution Approach 2:
The patent applies partial precision to transform skip mode residuals, using sufficient precision to maintain quality but not excessive precision that would increase bit rate. This selective precision approach maintains measurement precision where critical while reducing information loss through efficient bit allocation.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of video decoding performed in a video decoder is provided. A bit stream including syntax elements is received. The syntax elements correspond to coefficients of a region of a transform skipped block in a coded picture. The syntax elements include a first flag indicating whether an absolute coefficient level of one of the coefficients is greater than 0, and a second flag indicating a parity of the absolute coefficient level of the one of the coefficients. The second flag is decoded based on an equal probability model.