Scalable Video Coding POC Alignment via Reset Flags
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In scalable video coding (SVC), the picture order count (POC) values are not aligned across different layers when random access pictures are not aligned, leading to inconsistencies in POC values for pictures within the same access unit, which can violate bitstream conformance requirements.
Innovation Solution
A method is introduced to improve the derivation of POC values by using signaling flags, such as POC MSB reset and LSB reset flags, to ensure that all pictures within an access unit have consistent POC values, even when IRAP pictures are not aligned across layers, by resetting the POC values appropriately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If POC values are reset when IRAP pictures appear in the bitstream, then random access capability is improved, but POC alignment across different layers deteriorates when IRAP pictures are not aligned
Solution Approach 1:
The patent introduces two new parameters (flags) to control POC reset behavior: 'pcm_sample_adaptive_offset_enabled_flag' and 'delta_poc_msb_override_flag'. These flags allow selective resetting of POC MSB and LSB components, enabling independent control of POC alignment across layers while maintaining random access capability. The flags are set in the bitstream to indicate whether POC reset should occur, resolving the contradiction by providing conditional control over the reset behavior.
Solution Approach 2:
The patent makes the POC reset behavior dynamic rather than static. Instead of always resetting POC when IRAP pictures appear, the system now dynamically determines whether to reset based on the values of the introduced flags. This dynamic control allows the system to adapt to different scenarios: resetting when alignment is needed, and maintaining current POC values when alignment would cause problems across misaligned IRAP pictures.
2Manufacturing precision
If POC values are kept consistent across all pictures in an access unit, then bitstream conformance is improved, but flexibility in handling misaligned IRAP pictures deteriorates
Solution Approach 1:
The patent segments the POC value into two independent components: POC MSB (most significant bits) and POC LSB (least significant bits). By introducing separate flags to control resetting of each component independently, the system can achieve bitstream conformance for one component while maintaining flexibility in the other. This segmentation allows differential handling of POC alignment requirements across different layers and access unit types.
Solution Approach 2:
The patent applies different POC reset behaviors to different layers and different POC components based on local requirements. The introduced flags allow the encoder to specify that POC MSB should be reset while POC LSB remains unchanged, or vice versa, or both reset, or neither reset. This local quality control enables precise management of POC alignment at the component level rather than forcing a uniform approach across all POC values.
Data Source
Figure 1A
Figure 1B
Figure 2A
AI summary
An apparatus configured to code video information includes a memory unit and a processor in communication with the memory unit. The memory unit is configured to store video information associated with a first video layer having a first picture. The processor is configured to process picture order count (POC) derivation information associated with the first picture, and determine, based on the POC derivation information associated with the first picture, a POC value of at least one other picture in the first video layer that precedes the first picture in decoding order. The processor may encode or decode the video information.