Switchable Extended-Tap Adaptive Loop Filters for Video Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently managing bandwidth demand and artifact reduction in high-resolution videos, particularly with the increasing number of connected devices and the need for improved compression capabilities beyond current standards like VVC.
Innovation Solution
The implementation of an extended tap in an adaptive loop filter (ALF) and conversion between visual media data and a bitstream, utilizing advanced filtering techniques to minimize mean square errors and adapt filter coefficients for different video regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional filtering techniques are used in video coding, then device complexity is kept low, but manufacturing precision (coding efficiency and artifact reduction) deteriorates
Solution Approach 1:
The filtering process is segmented into multiple tap operations (e.g., 5-tap, 7-tap, 9-tap filters) that can be selectively applied to different video regions. This segmentation allows the system to use simpler filters for uniform regions and more complex filters only where needed, improving coding efficiency without uniformly increasing device complexity across all regions.
Solution Approach 2:
The filter complexity is made dynamic through adaptive selection of tap configurations based on local video characteristics. The system dynamically adjusts the number and arrangement of taps in the adaptive loop filter according to regional variations in video content, enabling higher precision filtering where required while maintaining lower complexity in simpler regions.
2Manufacturing precision
If conventional filtering techniques are used in video coding, then device complexity is kept low, but manufacturing precision (artifact reduction) deteriorates
Solution Approach 1:
The adaptive loop filter applies different filter configurations (different tap numbers and arrangements) to different local regions of the video based on their specific characteristics. This local quality approach ensures that artifact reduction is optimized for each region's needs rather than applying a uniform filter complexity across the entire video, thereby improving artifact reduction without unnecessarily increasing overall device complexity.
3Manufacturing precision
If extended tap configurations are implemented in ALF, then manufacturing precision (coding efficiency) is improved, but device complexity increases
Solution Approach 1:
The filter structure transitions from static to dynamic, allowing the number and configuration of taps to be adjusted based on video content characteristics. This dynamic adaptation enables the system to achieve higher coding efficiency with extended tap configurations only when and where necessary, rather than maintaining fixed high complexity throughout, thus improving manufacturing precision while controlling device complexity through selective application.
Data Source
AI summary
A mechanism for processing video data is disclosed. The mechanism includes determining to apply an extended tap in an adaptive loop filter (ALF). The ALF may also include a spatial tap, and the extended tap may be different from the spatial tap. A conversion is performed between a visual media data and a bitstream based on the ALF.