LDR Sample Clipping Range Control for Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video and image coding technologies face challenges in efficiently reducing redundancy and signaling overhead while managing bit depth, particularly in low dynamic range (LDR) applications.
Innovation Solution
Implementing sample clipping operations based on determined clipping ranges during encoding and decoding processes, which can be derived from LDR or internal bit depth, and applied at various stages, including loop filtering and reconstruction, to manage sample values within specified limits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If bit depth of video/image samples is reduced to increase coding efficiency, then coding efficiency is improved, but image quality and dynamic range representation deteriorate
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the clipping range parameters (minClip, maxClip) based on the selected LDR curve and block characteristics. This allows the system to adaptively modify the sample value transformation parameters to optimize both compression efficiency and perceived image quality for different content types and display conditions.
Solution Approach 2:
The patent implements dynamics by enabling flexible selection of different LDR curves (e.g., Curve A, Curve B, Curve C) and allowing the clipping operation to be applied conditionally at different encoding stages. The clipping range can be adjusted dynamically based on scene characteristics, preventing over-clipping in certain regions while maintaining efficiency benefits elsewhere.
2Measurement precision
If sample clipping range is restricted to improve LDR representation, then dynamic range accuracy is improved, but flexibility in handling different content types deteriorates
Solution Approach 1:
The patent achieves universality by designing a clipping mechanism that can operate in multiple modes: it can be applied to different color components (luma, chroma) independently, can be configured with different clipping ranges based on LDR curves, and can be selectively applied at different encoding stages. This multi-functional design allows a single mechanism to handle diverse content types including high-contrast scenes, low-contrast scenes, and various color spaces.
Solution Approach 2:
The patent applies segmentation by allowing separate clipping operations for different color components (e.g., luma and chroma can have different clipping ranges). This enables independent optimization of clipping parameters for each component based on its specific characteristics and requirements, improving overall adaptability while maintaining precise LDR control.
3Loss of information
If clipping operation is applied during encoding to reduce redundancy, then signaling overhead is reduced, but complexity of encoding and decoding processes increases
Solution Approach 1:
The patent applies preliminary action by performing the clipping operation during the encoding process before quantization and transform steps. This preliminary clipping of sample values simplifies subsequent processing stages by preventing overflow and reducing the dynamic range early in the pipeline, thereby reducing redundancy in the encoded data while avoiding the need for complex post-processing corrections during decoding.
Data Source
AI summary
This disclosure relates generally to sample clipping in video or image and particularly to methods and systems for sample clipping in video or image based on Low Dynamic Range (LDR) coding. For example, an encoder or decoder may determine a sample clipping range for a bock and apply a sample clipping procedure based on the sample clipping range during one of various stages of encoding or decoding process of the block. The sample clipping range may be derived from an LDR selected for encoding the block.