Video Decoding Motion Vector Storage Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for higher display quality in video content leads to increased memory and bandwidth requirements for storing motion vector information, which can degrade decoding efficiency and consume excessive storage space and transmission bandwidth.
Innovation Solution
A method and apparatus for storing motion vector information that utilize the direct—8×8_inference flag and reference picture lists to selectively store only necessary motion vector data, reducing storage capacity and bandwidth by discarding unnecessary information based on syntax elements and target resolution levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion vector information of all sub-macroblock partitions is stored, then decoding accuracy is improved, but memory storage capacity and transmission bandwidth are consumed excessively
Solution Approach 1:
The patent extracts and stores only the necessary motion vector information based on the direct-8x8-inference flag. When the flag is set, motion vector information for certain sub-macroblock partitions is discarded rather than stored, thereby reducing memory storage capacity and transmission bandwidth while maintaining decoding accuracy for partitions that require it.
Solution Approach 2:
The patent applies different storage strategies to different sub-macroblock partitions based on the direct-8x8-inference flag. Some partitions have their motion vector information stored while others do not, creating a localized quality approach where storage is optimized per partition rather than uniformly applied across all partitions.
2Measurement precision
If motion vector information of all sub-macroblock partitions is stored, then decoding accuracy is improved, but transmission bandwidth is consumed excessively
Solution Approach 1:
The patent extracts and transmits only the necessary motion vector information based on the direct-8x8-inference flag. When the flag is set, motion vector information for certain sub-macroblock partitions is excluded from transmission, thereby reducing transmission bandwidth consumption while maintaining decoding accuracy for partitions that require it.
3Quantity of substance
If motion vector information is reduced based on direct-8×8_inference flag, then memory storage capacity and transmission bandwidth are reduced, but decoding efficiency may be degraded
Solution Approach 1:
The patent performs preliminary action by setting the direct-8x8-inference flag during encoding to indicate which sub-macroblock partitions require motion vector information. This allows the decoding apparatus to pre-determine which partitions need storage and processing, optimizing both memory usage and decoding efficiency without requiring full motion vector information for all partitions.
Data Source
AI summary
A video decoding apparatus includes a bitstream parser, a calculator and a memory. The bitstream parser is provided to receive a video bitstream and extracting a set of constraints associated with the video bitstream, wherein the set of constraints has information associated with a direct_8×8_inference flag for a macroblock of a picture, wherein the macroblock has N sub-macroblock partitions. The calculator is provided to calculate first motion vector information associated with the macroblock and obtain second motion vector information associated with K of the N sub-macroblock partitions from the first motion vector information according to the information associated with the direct_8×8_inference flag, wherein K is less than N. The memory is provided to store the second motion vector information.