HEVC SAO Syntax Decoupling to Remove CABAC Bin Interleaving

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current high-efficiency video coding (HEVC) systems face inefficiencies in Sample Adaptive Offset (SAO) signaling due to limitations in syntax and binarization, leading to reduced throughput and suboptimal performance in video compression.

Innovation Solution

The proposed solution involves enhanced SAO signaling mechanisms that decouple SAO type coding, introduce new context modeling, and resolve the interleaving of context-coded and bypass-coded bins, utilizing truncated unary binarization and dedicated contexts to improve CABAC throughput.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If context-coded bins and bypass-coded bins are interleaved in SAO signaling, then coding flexibility is maintained, but CABAC engine throughput is limited

Engineering Contradiction:
ImproveCABAC engine throughputVSAvoidSAO signaling complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the SAO signaling process into distinct phases: first coding all context-coded bins together, then coding all bypass-coded bins together. This segmentation eliminates the interleaving pattern that limits CABAC throughput while maintaining the necessary coding flexibility through proper organization of the two coding types.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If multiple contexts are used for SAO type coding, then coding precision is improved, but processing complexity increases

Engineering Contradiction:
ImproveSAO type coding precisionVSAvoidcontext modeling complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies partial action by using only one context for coding the SAO type parameter, rather than employing multiple contexts. This selective approach maintains adequate coding precision while significantly reducing the complexity of context modeling and CABAC processing, avoiding the excessive complexity that would result from using multiple contexts.

Inventive Principle:
Principle #16Partial or excessive action

3Manufacturing precision

If SAO parameters are signaled with detailed syntax, then decoding accuracy is improved, but bitstream length increases

Engineering Contradiction:
ImproveSAO parameter decoding accuracyVSAvoidbitstream size
Core Design Contradiction:
Manufacturing precisionVSQuantity of substance

Solution Approach 1:

The patent merges the coding of context-coded and bypass-coded bins into an optimized sequence where all context-coded bins are processed first, followed by all bypass-coded bins. This merging approach improves decoding efficiency and accuracy by eliminating interleaving overhead, while the streamlined process actually reduces the effective bitstream length compared to the interleaved approach.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP3917149A1Decoupling enhancements in sample adaptive offset (SAO) for high efficiency video encoder (HEVC)
Publication Date: 2021.12.01 SONY GROUP CORP
  • EP3917149A1 patent drawingFigure 1
  • EP3917149A1 patent drawingFigure 2
  • EP3917149A1 patent drawingFigure 3

AI summary

An encoder/decoder is described using enhanced signaling mechanisms SAO parameters. The various parameters are signaled in various ways according to different embodiments of the invention. In a first embodiment (embodiment A), SAO on / off is decoupled form SAO type coding, with SAO on / off flags being jointly encoded for all color components. The second embodiment (embodiment B), is similar to embodiment A, but modified for application to JCTVC-J0268. In a third embodiment separate signaling is provided for SAO on / off, SAO types BO and EO, and for BO and EO side information (classes or band position). Each of these enhanced SAO signaling mechanisms provide enhanced coding efficiency.