Video Decoding Parallel Processing Motion Vector Extent
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video decoding technologies face limitations in decoding intermediate pictures derived from preceding frames, especially in fast-forward applications, due to the serial nature of bit streams and dependency on complete previous frame data, leading to jumpy images and inefficiencies in processing.
Innovation Solution
The method involves partially parallel decoding of video frames by commencing decoding of intermediate pictures once the vertical extent of the previous picture exceeds the maximum motion vector extent, allowing decoding to start before the entire previous picture is available, utilizing multiple decoding cores to process frames in parallel.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If parallel decoding is implemented to improve decoding speed, then productivity is improved, but it becomes impossible when pictures are derived using motion vectors from previous pictures
Solution Approach 1:
The decoding process is segmented into multiple independent decoding units that can operate in parallel. Each decoding unit processes a specific picture or portion of pictures independently, allowing parallel execution while maintaining the required data dependencies through proper segmentation of the work
2Reliability
If complete previous picture data is waited for before decoding current picture, then decoding accuracy is improved, but loss of time increases
Solution Approach 1:
The system performs preliminary actions by pre-decoding or pre-preparing picture data in advance using multiple decoding units. This allows the current picture decoding to start earlier without compromising accuracy, as the necessary previous picture data is prepared in parallel beforehand
Solution Approach 2:
The decoding system dynamically adjusts the decoding timing and data availability based on the maximum motion vector extent. By calculating the required data area dynamically, the system can start decoding as soon as sufficient data is available, minimizing delay while ensuring accuracy
3Productivity
If fast forward by skipping intermediate frames is implemented, then productivity is improved, but image quality and smoothness deteriorate
Solution Approach 1:
The video sequence is segmented into multiple independent pictures that can be decoded and displayed during fast forward. Instead of skipping frames, the system decodes and displays intermediate pictures by segmenting the processing into parallel decoding units, maintaining image smoothness while achieving fast forward functionality
Data Source
Figure 1~3
Figure 4
Figure 5
AI summary
A method and apparatus for decoding a compressed motion compensated video sequence is provided. The image data comprising a plurality of reference pictures (R) and motion vectors for deriving intermediate pictures (I,B) from preceding pictures in the sequence. The maximum vertical extent of the motion vector corresponds to a number of lines of image data. A reference picture is received along with a set of motion vectors to derive a next picture in the sequence from the reference picture. This next picture is decoded once the vertical extent of the reference picture received exceeds the maximum vertical extent of a motion vector from a starting position. A further next set of motion vectors for deriving a further next picture in the sequence is then received and for each further next picture the image data is decoded using the respective further next set of motion vectors when the area of a respective previous picture has been decoded to said maximum vertical extent of a motion vector from a starting position.