Wrap-Around Motion Compensation for Open GOP Segment Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies struggle to handle motion-vector wrap-around and open GOP resolution switching in HTTP streaming scenarios, leading to non-conformant bitstreams and visual artifacts during segment switches, particularly in 360-degree video streaming.
Innovation Solution
Implement a video coding concept that includes a wrap-around motion compensation tool and constrained RASL encoding indication SEI message to ensure bitstream conformance during resolution switching, using VVC's reference picture resampling (RPR) functionality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If open GOP coding structures are used to improve coding efficiency in HTTP streaming, then coding efficiency is improved, but bitstream conformance and visual quality are degraded during resolution switching
Solution Approach 1:
The patent applies preliminary action by inserting an Intra Random Access Point (IRAP) picture at a predetermined location within the open GOP structure, before resolution switching may occur. This pre-positioned IRAP picture serves as a safety net that ensures bitstream conformance can be maintained during resolution switching events, allowing the system to benefit from open GOP coding efficiency while having a prepared fallback mechanism ready.
Solution Approach 2:
The patent utilizes parameter changes by dynamically adjusting the picture structure and coding mode based on resolution switching events. When resolution switching is detected or anticipated, the system changes parameters such as inserting IRAP pictures, modifying GOP structures, or switching between intra and inter prediction modes to maintain bitstream conformance while preserving visual quality.
2Measurement precision
If motion-vector wrap-around is used to improve prediction accuracy in 360-degree video, then prediction accuracy is improved, but bitstream conformance is degraded during segment switches
Solution Approach 1:
The patent applies local quality by selectively applying wrap-around motion compensation only to specific blocks or regions within a picture where it is beneficial, rather than universally to all blocks. This localized application allows the system to maintain prediction accuracy in regions where wrap-around provides value while avoiding bitstream conformance issues in regions where it may cause problems during segment switches.
Solution Approach 2:
The patent implements dynamics by making the wrap-around motion compensation feature configurable and adaptable based on the coding scenario. The system can dynamically enable or disable wrap-around motion compensation, adjust its parameters, or switch between different motion compensation modes depending on whether segment switching is detected, thus maintaining both prediction accuracy and bitstream conformance.
3Productivity
If RASL pictures reference earlier segments for inter-prediction to improve coding efficiency, then coding efficiency is improved, but visual artifacts occur during resolution switching
Solution Approach 1:
The patent introduces an intermediary mechanism in the form of an IRAP picture positioned within the open GOP structure. This IRAP picture acts as a mediator between RASL pictures that reference earlier segments and the resolution switching event. By providing this intermediate reference point, the system maintains the coding efficiency benefits of inter-prediction from earlier segments while preventing visual artifacts through the availability of a fresh intra-coded reference at the appropriate location.
Data Source
AI summary
A motion-vector wrap-around coding tool may be reliably used, for instance, in the context of segmental video bitstream formation under switching between different video bitstream representations. A video encoder encodes a video and an indication into a data stream, which is valid for a sequence of pictures, and indicates that certain or all Random Access Skipped Leading (RASL) pictures within the sequence of pictures are coded in a manner excluding a predetermined set of one or more coding tools which includes a wrap-around motion compensation tool. A video decoder decodes from a data stream an indication that indicates that certain or all RASL pictures within a sequence of pictures are coded in a manner excluding a predetermined set of one or more coding tools including a wrap-around motion compensation tool.