CRA and BLA Picture Constraints for Flexible Video Random Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video codec standards limit the flexibility of random access points in video bitstreams, restricting the ability to perform operations like fast forwarding, rewinding, and seamless switching between video streams due to constraints on decodable leading pictures.
Innovation Solution
Introduce 'clean random access' (CRA) and 'broken link access' (BLA) pictures that allow non-decodable leading pictures after CRA pictures, using flags or indicators to manage these pictures, and redefine unit types and constraints to simplify mapping to container formats.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing video codec standards are used to ensure decodability of leading pictures, then decoding reliability is improved, but the flexibility of random access points and ability to perform fast forwarding/rewinding is reduced
Solution Approach 1:
The bitstream is segmented into different access point types (IDR and CRA pictures) with distinct characteristics. IDR pictures provide complete reference resets ensuring decodability, while CRA pictures allow non-decodable leading pictures for greater flexibility. This segmentation allows the system to choose appropriate access points based on operational needs.
Solution Approach 2:
The patent changes the parameter of reference picture availability by introducing CRA pictures that can have non-decodable leading pictures. This parameter change allows random access points to be placed more flexibly in the bitstream without requiring all leading pictures to be decodable, thus improving random access flexibility while maintaining decoding reliability through proper picture ordering and reference management.
2Manufacturing precision
If constraints on leading pictures are enforced to ensure proper decoding, then decoding correctness is improved, but the number of accessible random access points is reduced
Solution Approach 1:
The patent introduces dynamic picture ordering where leading pictures can appear after CRA pictures in bitstream order but precede them in display order. This dynamic arrangement allows the decoder to handle non-decodable leading pictures correctly by using picture order count (POC) values and reference picture management, enabling more random access points while maintaining decoding correctness.
Solution Approach 2:
The patent uses picture order count (POC) values and reference picture set management as intermediaries to handle the relationship between CRA pictures and their leading pictures. These intermediaries allow the decoder to correctly process non-decodable leading pictures by identifying them through POC comparisons and managing reference picture availability, thus enabling more access points without sacrificing decoding correctness.
3Adaptability or versatility
If CRA pictures are allowed in the middle of bitstream with non-decodable leading pictures, then random access flexibility is improved, but complexity of managing picture decoding order is increased
Solution Approach 1:
The patent uses feedback mechanisms through picture order count (POC) values and decoding flags that provide information to the decoder about the decoding status of leading pictures. When a CRA picture is encountered, the decoder uses POC comparisons and flags to determine whether leading pictures are decodable, providing feedback that simplifies the management complexity by automating the decision process.
Solution Approach 2:
The bitstream structure itself provides self-service by embedding sufficient information (POC values, picture types, reference picture sets) within the coded data to allow the decoder to automatically manage the complexity of CRA pictures and their leading pictures without external control. The decoder uses the embedded POC information to self-determine the correct decoding and display order.
Data Source
AI summary
Disclosed herein are innovations for bitstreams having clean random access (CRA) pictures and/or other types of random access point (RAP) pictures. New type definitions and strategic constraints on types of RAP pictures can simplify mapping of units of elementary video stream data to a container format. Such innovations can help improve the ability for video coding systems to more flexibly perform adaptive video delivery, production editing, commercial insertion, and the like.