Coded Video Processing with Selective LFNST for Lower Bitrates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently managing bandwidth demand due to the increasing number of connected devices receiving and displaying video, particularly in high-resolution formats, and there is a need for improved compression techniques to reduce bitrate while maintaining quality.
Innovation Solution
The implementation of a secondary transform, referred to as Low Frequency Non-Separable Transform (LFNST), is applied between primary transforms and quantization or de-quantization processes, along with methods like intra subblock partitioning, position-dependent intra prediction combination (PDPC), and reduced dimension transforms to optimize video coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional video coding techniques are used, then video transmission is supported, but bandwidth demand increases with the number of connected devices and high-resolution formats
Solution Approach 1:
The video block is divided into multiple sub-blocks, and different transform types (primary transform and secondary transform) are selectively applied to different sub-blocks based on their characteristics. This segmentation allows optimized compression for each region, improving overall compression efficiency while reducing bandwidth demand.
Solution Approach 2:
The patent introduces a secondary transform as an additional processing step that changes the transformation parameters applied to video blocks. By adjusting transform types, block sizes, and application conditions based on video content characteristics, the system achieves better compression ratios and reduced bandwidth requirements.
2Productivity
If a secondary transform is applied between primary transform and quantization, then compression performance improves, but processing complexity increases
Solution Approach 1:
The secondary transform is not applied uniformly to all video blocks but is selectively applied based on local characteristics such as prediction mode, block size, and content type. This localized application improves compression performance where needed while minimizing unnecessary processing complexity in other regions.
Solution Approach 2:
Instead of applying the secondary transform to all blocks, the patent applies it partially to only those blocks that benefit most from it. The transform application is controlled by various conditions and flags, ensuring that the additional processing complexity is incurred only when it provides meaningful compression improvement.
3Quantity of substance
If reduced dimension transforms are applied, then bitrate requirements are reduced, but transform accuracy may be compromised
Solution Approach 1:
The transform process is segmented into primary transform and secondary transform stages. The primary transform handles the main energy compaction, while the secondary transform provides additional refinement. This segmentation allows reduced dimension transforms to be applied effectively while maintaining accuracy through the two-stage approach.
Solution Approach 2:
The primary transform is applied first to perform preliminary energy compaction and identify significant coefficients. This preliminary action prepares the data for the secondary transform, ensuring that the reduced dimension transform operates on already-optimized data, thereby maintaining accuracy while reducing bitrate requirements.
Data Source
AI summary
A video processing method includes performing a conversion between a video block of a video and a bitstream of the video according to a rule. The rule specifies whether or how usage of a secondary transform within a video unit is indicated in the bitstream. The secondary transform is applied before quantization or after de-quantization.