Adaptive SATD Cost Estimation for Lower-Complexity Video Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern video encoding standards like VVC face significant complexity increases due to extensive rate-distortion cost computations for intra-prediction coding, which hinder efficient encoding while maintaining quality gains from added block partitioning structures and coding tools.
Innovation Solution
The implementation of a DC adjustment ratio table and adaptive SATD (Sum of Absolute Transformed Differences) for video encoding, which reduces complexity by selecting a subset of candidates for later decision stages, using spatial variation and quantization parameter to weight the DC component based on perceptual importance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If extensive rate-distortion cost computations are performed for intra-prediction coding, then encoding quality is improved, but encoding complexity increases significantly
Solution Approach 1:
The patent segments the prediction cost computation by separating DC component analysis from AC component analysis. The DC adjustment ratio table pre-computes and stores DC-related cost values, while AC components are computed separately during encoding. This segmentation allows the encoder to use pre-computed DC values without performing extensive real-time computations, thereby reducing encoding complexity while maintaining quality.
Solution Approach 2:
The patent performs preliminary computation of DC adjustment ratios and stores them in lookup tables before the actual encoding process. By pre-computing these values based on spatial variation and quantization parameter relationships, the encoder avoids performing extensive rate-distortion computations during the encoding stage, thus reducing complexity while preserving quality.
2Manufacturing precision
If more block partitioning structures and coding tools are added, then encoding quality is improved, but encoding complexity increases
Solution Approach 1:
The patent applies partial action by selectively using DC adjustment ratios for specific prediction modes and block types where they provide the most benefit. Rather than applying complex computations uniformly across all coding tools and partitioning structures, the method focuses DC adjustment on intra-prediction coding scenarios where spatial variation is significant, thereby maintaining quality improvements while limiting complexity increases.
3Manufacturing precision
If DC component weighting is adjusted based on spatial variation, then perceptual quality is improved, but computational cost increases
Solution Approach 1:
The patent pre-computes DC adjustment ratios based on relationships between spatial variation metrics and quantization parameters, storing these in lookup tables. During encoding, the system simply retrieves the appropriate DC adjustment ratio from the table based on measured spatial variation and QP values, avoiding the need for complex real-time computations while still achieving perceptual quality improvements through adaptive DC weighting.
Data Source
AI summary
Systems and methods for adaptive prediction cost estimation in video encoding are provided. The techniques improve early cost estimation and reduce the number of candidates for the later decision stages and final RDO stage. In particular, an adaptive sum of absolute transformed differences (SATD) is determined for each candidate, and, based on the adaptive SATD values, a subset of candidates is selected for mode decision search to determine block partitioning, motion vectors, and encoding modes, The adaptive SATD combines a weighted DC component of the SATD and the AC component of the SATD. The weighting factor is selected from a DC adjustment ratio table based on the spatial variation and the QP for a respective coding tree unit. The techniques improve cost estimation accuracy, reduce encoding complexity, and are hardware-friendly for integration into video codecs such as HEVC, AV1, VVC, and AV2.