Scaled Motion Search Section Parallel Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video encoding and decoding methods, such as those based on the H.264 standard, face challenges in achieving efficient and fast processing, particularly in handling high-resolution video signals, which can lead to increased bandwidth and storage requirements.
Innovation Solution
The implementation of a video encoder/decoder system that includes a motion refinement engine with a scaled motion search section and a reduced-scale motion search module, which uses downscaled pictures to generate motion vector candidates and refine motion vectors, allowing for parallel processing and improved efficiency in motion search and refinement operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If conventional video encoding methods are used to handle high-resolution video signals, then processing accuracy is maintained, but processing speed decreases and computational complexity increases
Solution Approach 1:
The motion search process is divided into two independent stages: full-resolution motion search and refined motion search. The full-resolution search operates on original high-resolution blocks to identify coarse motion vectors, while the refined search operates on downsampled lower-resolution blocks to refine these vectors. This segmentation allows parallel processing of multiple blocks at full resolution without increasing overall computational complexity, thereby improving processing speed.
Solution Approach 2:
The patent introduces a resolution dimension by performing motion search at two different resolutions: full resolution for initial search and reduced resolution for refinement. This dimensional approach allows the system to process multiple full-resolution blocks simultaneously by using lower-resolution reference data, effectively increasing processing throughput without sacrificing motion estimation accuracy.
2Measurement precision
If full-resolution motion search is performed on all blocks, then motion estimation accuracy is maintained, but processing time increases
Solution Approach 1:
Motion search is segmented into two phases: a full-resolution phase that provides accurate initial motion vectors for all blocks, and a reduced-resolution refinement phase that quickly adjusts these vectors. By separating the accuracy-critical initial search from the time-consuming refinement process, the patent maintains motion estimation accuracy while reducing overall processing time through parallel execution of multiple blocks in the first phase.
Solution Approach 2:
The full-resolution motion search is performed as a preliminary action to establish accurate initial motion vectors for all blocks before refinement begins. This preliminary full-resolution search enables subsequent reduced-resolution refinement to start from accurate positions, reducing the search space and time required while maintaining final motion estimation accuracy.
3Productivity
If reduced-resolution motion search is used, then processing speed increases, but motion estimation accuracy decreases
Solution Approach 1:
The patent segments the motion search into two functional parts: full-resolution search that ensures accuracy and reduced-resolution refinement that boosts efficiency. The full-resolution search on original blocks provides accurate motion vectors, while the reduced-resolution search on downsampled reference blocks provides efficient refinement. This segmentation allows the system to achieve high processing efficiency without sacrificing accuracy, as each stage compensates for the other's limitations.
Solution Approach 2:
The reduced-resolution motion search acts as an intermediary refinement step between the full-resolution search and final motion compensation. It takes the initial motion vectors from full-resolution search and refines them using downsampled reference data, providing an efficient intermediate processing stage that improves productivity while the full-resolution initial search ensures final accuracy is maintained.
Data Source
AI summary
A scaled motion search section can be used in a video processing device that processes a video input signal that includes a plurality of pictures. The scaled motion search section includes a downscaling module that downscales the plurality of pictures to generate a plurality of downscaled pictures. A reduced-scale motion search module generates a plurality of motion vector candidates at a downscaled resolution, based on the plurality of downscaled pictures. The reduced-scale motion search module includes a column buffer that stores a column of reference data and generates the plurality of motion vector candidates based on a parallel processing of the column of reference data for a group of adjacent macroblock pairs.