Coefficient Scaling for High Precision Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards face challenges in efficiently managing transform and quantization processes, particularly in handling large block-size transforms and high-frequency coefficients, which impact compression performance and bandwidth usage.
Innovation Solution
The proposed solution involves the use of Low-Frequency Non-Separable Transform (LFNST) and dependent scalar quantization, along with sub-block transform and joint coding of chroma residuals, to optimize transform and quantization processes, reducing computational complexity and improving compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If transform and quantization processes are applied to large block-size transforms and high-frequency coefficients, then compression performance is improved, but computational complexity and memory usage increase
Solution Approach 1:
The transform block is divided into sub-blocks, and only selected sub-blocks undergo transform processing. This segmentation allows the system to process large blocks efficiently by focusing computational resources on specific regions, thereby improving compression performance without proportionally increasing computational complexity across the entire block.
Solution Approach 2:
Different processing strategies are applied to different regions of the transform block based on local characteristics. High-frequency coefficients in certain regions are processed with full transform and quantization to maintain compression performance, while other regions use simplified processing to reduce computational complexity and memory usage.
2Productivity
If transform and quantization processes are applied to large block-size transforms and high-frequency coefficients, then compression performance is improved, but memory usage increases
Solution Approach 1:
By segmenting the transform block into sub-blocks and selectively processing only certain sub-blocks, the patent reduces the total number of coefficients that need to be stored in memory during transform and quantization operations, thereby reducing memory usage while maintaining compression performance through targeted processing of critical regions.
Solution Approach 2:
The patent extracts and processes only the necessary high-frequency coefficients that contribute most to compression performance, rather than processing all coefficients in large blocks. This selective extraction reduces memory requirements by eliminating the need to store and process redundant or less significant coefficient data.
3Productivity
If a scale factor is applied to residual coefficients independent of block size, then compression efficiency is improved, but processing uniformity decreases
Solution Approach 1:
The patent applies a scale factor to residual coefficients independently of block size, changing the scaling parameter to optimize compression efficiency. This uniform scaling approach improves compression efficiency by consistently reducing coefficient magnitude across all block sizes, while the loss of processing adaptability is acceptable given the primary goal of maximizing compression efficiency through simplified uniform processing.
Data Source
AI summary
Methods, systems, and devices for coefficient scaling for high-precision image and video coding are described. An example method of video processing includes performing a conversion between a current block of a video and a bitstream representation of the video according to a rule. The rule specifies that the conversion includes during encoding, skipping applying a forward transform to residual coefficients of the current block prior to including in the bitstream representation, or during decoding, reconstructing residual coefficients of the current block from the bitstream representation without applying an inverse transform. The rule further specifies that a scale factor is applied to the residual coefficients independent of a size of the current block.