Decoding Device Flat Scaling List Transform Skip
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In HEVC, the encoding efficiency is reduced when transform skip is applied to transform units (TUs) larger than 4x4 pixels, as a scaling list other than the flat matrix is used during quantization, leading to decreased performance.
Innovation Solution
A decoding device and encoding device that use a flat scaling list for transform blocks larger than 4x4 pixels, ensuring that the transform skip is performed with optimal efficiency by avoiding the use of non-flat scaling lists.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If transform skip is applied to transform units larger than 4x4 pixels, then the encoding complexity is reduced, but the encoding efficiency deteriorates due to use of non-flat scaling lists
Solution Approach 1:
The patent changes the scaling list parameter from non-flat to flat specifically for transform units larger than 4x4 pixels when transform skip is applied. This parameter change resolves the contradiction by maintaining encoding complexity reduction while preventing encoding efficiency deterioration through optimized quantization matrix selection.
2Measurement precision
If a non-flat scaling list is used during quantization of transform blocks larger than 4x4, then the quantization precision is improved, but the encoding efficiency is reduced
Solution Approach 1:
The patent applies different scaling list characteristics to different transform unit sizes. Specifically, flat scaling lists are applied to transform units larger than 4x4 pixels when transform skip is used, while non-flat scaling lists can be used for other cases. This local differentiation resolves the contradiction by optimizing for encoding efficiency in specific contexts while maintaining quantization precision where needed.
Data Source
AI summary
A decoding device includes circuitry configured to decode a bit stream and generate a quantized value and inversely quantize the generated quantized value by using a flat scaling list, in a case where a block size of a transform block to which a transform skip is applied is larger than a 4 by 4 block size.