NAL Unit Header Scalable Video Coding Overhead Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing H.264 video compression standard's Network Adaptation Layer (NAL) unit header design has architectural shortcomings, leading to high overhead and error resilience issues, particularly in scalable and multiview coding, due to its 'bolt-on' nature, which complicates decoding and network element operations.
Innovation Solution
A revised NAL unit header design that includes a layer_id field and a table of layer descriptions, allowing for efficient pruning and decoding by indicating the position of layers in a hierarchy, reducing overhead and improving error resilience through enhanced syntax and parameter sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the existing H.264 NAL unit header design is used, then cross-vendor interoperability is maintained, but overhead is high and error resilience is poor due to the bolt-on nature of scalable and multiview coding extensions
Solution Approach 1:
The NAL unit header is segmented into distinct functional fields: a base header portion containing universal elements (forbidden_zero_bit, nal_ref_idc, nal_unit_type) and an extension header portion that is conditionally present only when scalable or multiview coding is used. This segmentation isolates the complexity of extensions from the base structure, improving error resilience by allowing decoders to ignore extension fields they cannot process while maintaining interoperability.
Solution Approach 2:
The patent introduces a new dimensional organization to the header by adding extension flags (svc_extension_flag, mvc_extension_flag) that act as dimensional switches. These flags create a conditional header structure where extension fields are present only when needed, transforming the header from a flat structure to a multi-dimensional one that adapts to different coding modes without increasing base complexity.
2Adaptability or versatility
If scalable and multiview coding extensions are added to the NAL unit header, then quality and spatial scalability are supported, but the header overhead increases and decoding becomes more complex
Solution Approach 1:
The NAL unit header is made dynamic through conditional presence of extension fields. The svc_extension_flag and mvc_extension_flag control whether extension fields appear in the header, allowing the header structure to adapt its complexity based on the coding mode. This dynamic structure enables scalable and multiview coding support while keeping the base header simple for non-scalable content.
Solution Approach 2:
Extension flags act as intermediaries between the base NAL unit header and the scalable/multiview extension fields. These flags mediate the conditional inclusion of extension fields, allowing the decoder to determine whether to process complex extension data or proceed with standard decoding, thereby managing complexity while supporting versatility.
3Adaptability or versatility
If extension fields are conditionally added based on svc_extension_flag, then scalable coding is supported, but the header structure becomes more complex and parsing becomes harder
Solution Approach 1:
The extension flags (svc_extension_flag, mvc_extension_flag) are evaluated in advance during header parsing to determine whether extension fields are present. This preliminary action allows the decoder to prepare the appropriate parsing path before encountering extension fields, simplifying the overall parsing process by avoiding the need to always parse potential extension fields regardless of their actual presence.
Data Source
AI summary
Disclosed are techniques for scalable, multiview, and multiple descriptive video coding using an improved Network Adaptation Layer (NAL) unit header. A NAL unit header can include a layer-id that can be a reference into a table of layer descriptions, which specify the properties of the layer. The improved NAL unit header can further include fields for reference picture management and to identify temporal layers.