Scalable Data Stream Signaling for Lower Parsing Overhead
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The handling of scalable data streams by network entities is complex due to the need for extensive parsing and signaling of operation points and scalability dimensions, which is inefficient and burdensome, especially in the context of HEVC and H.264/AVC standards.
Innovation Solution
A data stream structure is proposed where packets carry operation point identifiers and scalability axes descriptors, with profile and level descriptors signaled efficiently through explicit definitions or references, reducing overhead and complexity for network entities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If extensive parsing and signaling of operation points and scalability dimensions is implemented, then complete scalability information is provided, but complexity and overhead for network entities increases
Solution Approach 1:
The patent extracts scalability information into separate NAL units (VPS, SPS, PPS) that can be independently parsed and processed. This separates the complexity of scalability signaling from the main video decoding process, allowing network entities to handle only the necessary portions without parsing entire bitstreams.
Solution Approach 2:
The scalability information is segmented into multiple hierarchical levels (sequence level, picture level, slice level) with each NAL unit containing specific subsets of parameters. This segmentation allows progressive parsing where network entities can stop at any level based on their capabilities and network conditions.
2Adaptability or versatility
If multiple different NAL unit header structures are used to signal operation points, then scalability is achieved, but parsing effort and complexity increases
Solution Approach 1:
The NAL unit header is designed with a universal structure that can accommodate multiple scalability dimensions (temporal, spatial, quality) using the same field formats. The nal_unit_type, temporal_id, and layer_id fields work together in a consistent manner across all scalability types, eliminating the need for structure-specific parsing logic.
Solution Approach 2:
The patent introduces dynamic flags (e.g., temporal_dependency, spatial_dependency, quality_dependency) that enable the same NAL unit structure to adapt to different scalability configurations. These flags dynamically indicate which dimensions are active, allowing parsers to adjust processing based on actual content rather than predetermined structures.
3Loss of information
If prefix NAL units are used to signal base layer, then operation point information is provided, but additional overhead and processing is required
Solution Approach 1:
The patent merges base layer information with enhancement layer information in a unified NAL unit structure. Instead of separate prefix NAL units, the base layer parameters are integrated into the same syntax structures (VPS, SPS, PPS) that define enhancement layers, allowing single-pass parsing that extracts both base and enhancement information simultaneously.
Data Source
AI summary
The handling of scalable data streams by network entities is rendered less complex by, in the data stream, accompanying packets which actually carry data by packets of a different packet type which have a scalability axes descriptor defining the number of scalability axes and a semantic meaning thereof. In another aspect, the handling of scalable data streams by network entities is rendered more efficient by conveying level and/or profile descriptors using packets other than the data carrying packets, with profile and/or level descriptors being divided up into a first set explicitly signaling the definition of the available coding option set and/or available syntax element value range for a respective operation point, and a second set of profile and/or level descriptors which signal the definition of the available coding option set and/or available syntax element value range for their respective operation points by reference another profile and/or level descriptor.