Video Codec Slice Layout for Parallel CTU Row Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video codecs like HEVC lack efficient parallel processing capabilities for video encoding and decoding, particularly in handling CTU processing and slice organization, which hinders optimal utilization of multi-core processors.
Innovation Solution
The introduction of mechanisms to determine whether to output pictures preceding dependent random access pictures, manage coded picture buffer delays, and control non-scalable nested picture timing supplemental enhancement information, enhancing parallel processing and slice organization in video encoding and decoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If CTUs are processed in raster scan order sequentially, then processing order is simple and easy to implement, but parallel processing capability is limited
Solution Approach 1:
The picture is divided into multiple slices, and each slice is further divided into multiple CTU rows. This hierarchical segmentation allows different CTU rows to be processed in parallel by different processing units while maintaining organized data flow and dependency management within each slice.
Solution Approach 2:
The processing approach transitions from one-dimensional sequential raster scan to two-dimensional parallel processing across multiple CTU rows within slices. This dimensional change enables simultaneous processing of multiple CTU rows while maintaining the raster scan order within each row, thereby improving parallel processing capability without completely abandoning the traditional scanning approach.
2Productivity
If more CTU rows are processed in parallel, then parallel processing efficiency is improved, but processing complexity and resource management increase
Solution Approach 1:
By dividing the picture into slices and further dividing slices into CTU rows, the system creates manageable processing units that can be distributed across multiple processing cores. This segmentation reduces the complexity of managing parallel processing by breaking down the large-scale parallel operation into smaller, more manageable units.
Solution Approach 2:
The system dynamically adjusts the number of CTU rows processed in parallel based on available processing resources and picture characteristics. This dynamic approach allows the system to optimize parallel processing efficiency while adapting to varying computational loads and resource availability, preventing excessive complexity in resource management.
Data Source
AI summary
An apparatus (200) for receiving an input video data stream according to an embodiment is provided. The input video data stream has a video encoded thereinto. The apparatus (200) is configured to generate an output video data stream from the input video data stream.