CCALF Side Information for Video Coding Bandwidth and Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently utilizing side information for in-loop filtering, leading to suboptimal performance in bandwidth utilization and video quality, particularly in high-efficiency video coding standards like VVC.
Innovation Solution
The implementation of a cross-component adaptive loop filter (CCALF) that utilizes side information for reconstruction, including deblocking filter (DBF), sample adaptive offset (SAO), and bilateral filter (BF), to enhance video data processing by applying luma and chroma reconstructions, predictions, and residual values, with adaptive filtering and partitioning techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional in-loop filtering is used without side information, then device complexity is reduced, but video quality and bandwidth utilization are suboptimal
Solution Approach 1:
The patent applies preliminary action by performing deblocking filter (DBF), sample adaptive offset (SAO), and bilateral filter (BF) operations before the cross-component adaptive loop filter (CCALF) processing. These preliminary filtering stages prepare the luma and chroma reconstructions, predictions, and residual values in advance, allowing the CCALF to operate on pre-processed data with reduced complexity while maintaining high video quality.
Solution Approach 2:
The filtering process is segmented into multiple independent stages: DBF, SAO, BF, and CCALF. Each stage processes specific aspects of the video data separately, allowing for optimized computation at each level. The side information from each segmentation stage is independently generated and utilized by subsequent stages, reducing overall system complexity while improving video quality through specialized processing at each level.
2Manufacturing precision
If more side information is utilized in CCALF, then video quality improves, but bandwidth requirements increase
Solution Approach 1:
The patent discards redundant information by selectively using only the most valuable side information (luma and chroma reconstructions, predictions, and residual values) for CCALF input. Less critical data is discarded to reduce bandwidth requirements. The essential side information is recovered and efficiently transmitted to the decoder, achieving high video quality without proportionally increasing bandwidth consumption.
Solution Approach 2:
Instead of transmitting all possible side information, the patent applies partial action by selecting and transmitting only the most critical side information elements needed for effective CCALF operation. This partial transmission approach achieves sufficient video quality improvement without the excessive bandwidth consumption that would result from transmitting complete side information sets.
3Productivity
If adaptive filtering with multiple reconstruction types is applied, then video coding efficiency improves, but processing time increases
Solution Approach 1:
The patent applies periodic action by systematically processing different reconstruction types (luma and chroma) through standardized filtering operations. Each reconstruction type undergoes the same sequence of DBF, SAO, BF, and CCALF steps in a periodic manner, allowing for optimized processing routines that improve video coding efficiency while managing processing time through regular, predictable operation cycles.
Solution Approach 2:
Preliminary processing of luma and chroma reconstructions, predictions, and residual values is performed before the main CCALF operation. This preliminary action prepares the data in advance, organizing it into the required formats and reducing the computational burden during the main filtering pass, thereby improving overall video coding efficiency without excessive processing time increases.
Data Source
AI summary
A mechanism for processing video data is disclosed. The mechanism includes determining to use side information as an input into a cross component adaptive loop filter (CCALF). A conversion is then be performed between a visual media data and a bitstream based on the CCALF. Various types of side information can be used.