Scalable Video Decoding Buffer Management for Non-Aligned IRAP
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In scalable video coding, complications arise when random access point pictures are not aligned across multiple layers, leading to issues with flushing the decoded picture buffer, particularly in multi-layer bitstreams with non-aligned IRAP pictures, which can result in unnecessary picture removal and buffer overflow.
Innovation Solution
An apparatus and method that determine the value of a flag associated with a base layer picture to decide whether to remove or refrain from removing enhancement layer pictures from the decoded picture buffer, ensuring proper flushing without outputting them before the corresponding enhancement layer picture is coded, thereby maintaining buffer integrity and alignment across layers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If enhancement layer pictures are removed from the decoded picture buffer when random access point pictures are not aligned across multiple layers, then buffer overflow is prevented, but unnecessary picture removal occurs leading to loss of reference information
Solution Approach 1:
The patent changes the parameter being monitored from layer alignment status to individual picture output status. By checking whether each enhancement layer picture has been output before removal, the system dynamically adjusts the removal decision based on actual picture state rather than assuming removal is always safe when layers are misaligned.
Solution Approach 2:
The patent introduces a feedback mechanism where the decoding and output status of enhancement layer pictures is continuously monitored. This feedback information is used to control the picture removal operation, ensuring that pictures are only removed when it is safe to do so, thereby preventing both buffer overflow and unnecessary information loss.
2Loss of information
If enhancement layer pictures are retained in the decoded picture buffer to preserve reference information, then reference picture availability is maintained, but buffer overflow may occur
Solution Approach 1:
The patent changes the control parameter from static layer alignment configuration to dynamic picture output status. This allows the buffer management to adapt to the actual decoding progress of enhancement layer pictures, retaining pictures only as long as necessary for reference purposes.
Solution Approach 2:
The patent makes the picture removal operation dynamic by continuously checking the output status of enhancement layer pictures. The buffer management strategy transitions from a static rule-based approach to a dynamic state-dependent approach, optimizing the balance between preserving reference information and preventing buffer overflow.
3Ease of operation
If pictures are flushed from the decoded picture buffer at random access points, then decoding alignment is improved, but complications arise when random access point pictures are not aligned across multiple layers
Solution Approach 1:
The patent segments the buffer management control into two independent parts: base layer picture flushing based on random access point detection, and enhancement layer picture removal based on individual picture output status. This segmentation allows each layer to be managed independently according to its specific requirements, simplifying the overall control logic despite multi-layer complexity.
Solution Approach 2:
The patent introduces an intermediary control mechanism that mediates between the base layer flushing operation and enhancement layer picture retention. This intermediary checks the output status of enhancement layer pictures and controls their removal timing, preventing direct conflicts between layer alignment requirements and reference picture availability.
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 base layer (BL) and an enhancement layer (EL), the BL having a BL picture in a first access unit, and the EL having an EL picture in the first access unit. The BL picture may be associated with a flag. The processor is configured to determine a value of the flag associated with the BL picture, and perform, based on the value of the flag, one of (1) removing one or more EL pictures in a decoded picture buffer (DPB) without outputting the one or more EL pictures before the EL picture is coded, or (2) refraining from removing the one or more EL pictures in the DPB without outputting the one or more EL pictures. The processor may encode or decode the video information.