Video Encoder Quantization Adjustments for DC Shift Contouring Artifacts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression techniques do not effectively address contouring artifacts caused by DC shift artifacts, leading to visible distortions in smooth regions of video frames, which affect perceptual quality despite increased bit rate costs.
Innovation Solution
A video encoder adjusts quantization by identifying DC shift blocks and gradient slope regions, reducing quantization step sizes in these areas to minimize contouring artifacts, using bi-level or multi-level differential quantization and analyzing texture maps to optimize quantization parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If uniform quantization is applied across all regions, then encoding complexity is reduced, but contouring artifacts appear in smooth gradient regions due to DC shift
Solution Approach 1:
The patent applies different quantization strategies to different regions of the video frame. Smooth gradient regions are identified and processed with adjusted quantization parameters to prevent contouring artifacts, while other regions use standard quantization. This local differentiation resolves the contradiction by maintaining low overall complexity while specifically addressing harmful artifacts in vulnerable regions.
Solution Approach 2:
The patent performs preliminary identification of smooth gradient regions before quantization is applied. By detecting gradient slopes and classifying regions in advance, the system can pre-determine which blocks need special quantization treatment, avoiding the need for complex real-time adjustments during encoding while preventing contouring artifacts.
2Productivity
If quantization step size is increased to reduce bit rate, then compression efficiency improves, but perceptual quality deteriorates due to visible contouring artifacts
Solution Approach 1:
The patent applies higher quantization step sizes in textured regions where contouring is less visible, and uses lower or adjusted step sizes in smooth gradient regions where contouring artifacts are most problematic. This local differentiation maintains high compression efficiency overall while preserving perceptual quality in vulnerable regions.
Solution Approach 2:
The patent dynamically adjusts quantization parameters based on local gradient characteristics. By changing the quantization step size parameter according to the detected gradient slope in each region, the system optimizes the balance between compression efficiency and perceptual quality, reducing visible artifacts without sacrificing overall compression performance.
3Quantity of substance
If DC coefficients are heavily quantized to save bits, then bit rate is reduced, but DC shift artifacts cause contouring in smooth regions
Solution Approach 1:
The patent applies differential quantization to DC coefficients based on local gradient characteristics. In smooth gradient regions, DC coefficients are quantized with higher precision to prevent DC shift artifacts, while in textured regions where contouring is less visible, more aggressive quantization is applied to reduce bit rate. This resolves the contradiction by protecting against harmful artifacts only where necessary.
Solution Approach 2:
The patent performs preliminary classification of blocks into smooth gradient regions and other regions before quantization. This preliminary action allows the system to pre-determine which DC coefficients require careful quantization to avoid artifacts, enabling efficient bit allocation that reduces overall bit rate while preventing DC shift artifacts in vulnerable regions.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A video encoder identifies one or more portions of a video picture that contain DC shift blocks and adjusts quantization (e.g., by selecting a smaller quantization step size) to reduce contouring artifacts when the picture is reconstructed. The encoder can identify the portion(s) of the picture that contain DC shift blocks by identifying one or more gradient slope regions in the picture and analyzing quantization effects on DC coefficients in the gradient slope region(s). The encoder can select a coarser quantization step size for a high-texture picture portion.