EDRAP Sample Signalling for Clearer Media Track Random Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing media file formats for video coding, such as ISOBMFF, lack clarity in track references and assumptions about the need for specific random access points, leading to inefficiencies in random access mechanisms, particularly with extended dependent random access points (EDRAPs).
Innovation Solution
Clarifies the track reference type 'aest' to specify a single track identifier and defines EDRAP samples as requiring zero or more preceding SAP or EDRAP samples for correct decoding, using the VisualEdrapEntry class to manage sample dependencies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the media file format uses track references without explicit clarity on reference types, then the file format remains simpler, but the random access mechanism becomes ambiguous and less efficient
Solution Approach 1:
The track reference structure is segmented into distinct reference types (e.g., aest for associated external stream track, gest for group track) with specific meanings. This segmentation clarifies the random access mechanism by explicitly defining what each reference type represents, eliminating ambiguity while maintaining manageable complexity through standardized categorization.
Solution Approach 2:
A reference type box is introduced as an intermediary structure between the track reference and the actual track. This intermediary layer provides explicit metadata about the reference type, serving as a mediator that translates ambiguous references into clear, actionable information for random access operations without requiring fundamental changes to the overall file format.
2Adaptability or versatility
If EDRAP samples are defined with flexible preceding sample requirements, then the decoding flexibility improves, but the ambiguity in sample dependency increases
Solution Approach 1:
The VisualEdrapEntry class is introduced to perform preliminary action by explicitly listing and defining all required preceding SAP or EDRAP samples before decoding begins. This preliminary specification of sample dependencies eliminates ambiguity during decoding while maintaining the flexibility to handle various sampling configurations, as the required samples are pre-identified and documented.
Solution Approach 2:
The sample dependency structure incorporates feedback mechanisms where the required preceding samples are explicitly referenced and verified during decoding. This feedback loop ensures that all necessary samples are available before processing, maintaining clarity in sample dependencies while allowing flexible decoding configurations through structured reference to previously decoded samples.
3Productivity
If the track reference assumes a single track identifier, then the random access efficiency improves, but the adaptability to complex track structures decreases
Solution Approach 1:
The reference type box structure provides universality by accommodating multiple reference types (aest, gest, and others) within a single standardized framework. This multi-functional design allows the system to efficiently handle single track identifiers for optimized random access while also being adaptable to complex multi-track structures, as the same reference mechanism can interpret different reference types appropriately.
Solution Approach 2:
The system uses parameter changes in the reference type box to adapt to different track structures. By changing the reference type parameter (e.g., setting reference_type to 'aest' for associated external stream track or 'gest' for group track), the system can optimize performance for simple cases while maintaining adaptability to complex scenarios, achieving both productivity and versatility through parameter-based configuration.
Data Source
AI summary
A mechanism for processing media data is disclosed. In one aspect, a method includes determining, during a conversion between the media data and a media data file, whether a track reference of type associated external stream track (‘aest’) is included in a media track. The method also includes performing the conversion based on the determining. Additionally, when present, a track reference type box with a reference type equal to ‘aest’ shall contain only a track identifier and shall not contain any track group identifier.