Motion Estimation Data Reuse via Stripe Scan Bandwidth Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data reuse methods for motion estimation, such as level C, require excessive bandwidth for external frame memory, especially with high-resolution videos like HDTV 1280×720 at 30 frames per second, leading to increased system loading and power consumption.
Innovation Solution
The data reuse method with level C+ reuses the overlapped searching region completely in the horizontal direction and partly in the vertical direction, loading BH×(SRv+n×Bv−1) pixels from the external frame memory for n consecutive blocks, and adjusts internal memory size according to the video encoder's scheduling, reducing external frame memory bandwidth requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If data reuse method with level C is applied in hardware accelerator, then internal memory state is improved, but external frame memory bandwidth requirement increases to 536 MByte/s for HDTV 1280×720 at 30 fps
Solution Approach 1:
The patent transitions from traditional raster scan (horizontal then vertical) to stripe scan (vertical stripes) for loading reference frame blocks. This dimensional change in data access pattern enables more effective reuse of data in external frame memory, reducing the total bandwidth requirement from 536 MByte/s to 442.4 MByte/s for HDTV 1280×720 at 30 fps while maintaining motion estimation accuracy.
2Measurement precision
If huge bandwidth of external frame memory is introduced for motion estimation, then motion estimation accuracy is maintained, but power consumption of system bus increases
Solution Approach 1:
The patent changes the data access parameter from traditional raster scan to stripe scan pattern, and optimizes the loading of reference frame blocks by considering both horizontal and vertical overlapping regions. This parameter change reduces external frame memory bandwidth requirement by approximately 16% (from 536 to 442.4 MByte/s), directly reducing power consumption while preserving motion estimation accuracy through complete coverage of necessary search regions.
3Ease of repair
If data reuse method with level C is used, then horizontal overlapping region is reused, but vertical overlapping region bandwidth requirement increases for high-resolution video
Solution Approach 1:
The patent merges the data loading strategy for both horizontal and vertical overlapping regions into a unified stripe scan approach. Instead of treating horizontal and vertical reuses separately as in level C, the method combines them by loading vertical stripes that inherently capture both horizontal and vertical overlapping data, enabling more comprehensive data reuse and reducing external memory bandwidth requirements for high-resolution video.
Data Source
AI summary
A data reuse method with level C+ for block matching motion estimation is disclosed. Compared to conventional Level C scheme, this invention can save large external memory bandwidth of motion estimation. The main idea is to reuse the overlapped searching region in the horizontal direction and partially reuse the overlapped searching region in the vertical direction. Several vertical successive current macroblocks are stitched, and the searching region of these current macroblocks is loaded, simultaneously. With the small overhead of internal memory, the reduction of external memory bandwidth is large. By case studies of H.264/AVC, the level C+ scheme can provide a good trade-off between the conventional Level C and D scheme.