Adaptive Video Compression Motion Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital video encoding methods fail to efficiently address redundancy in video streams, leading to suboptimal compression and increased noise accumulation due to reliance on reconstructed frames, which introduce quantization noise and mismatch noise.
Innovation Solution
The proposed method employs motion compensation using both original and reconstructed frames to determine a blended motion-predicted picture, selecting between them based on quantization noise, mismatch noise, and spatial characteristics to minimize noise accumulation and improve encoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If motion compensation uses reconstructed frames as reference, then encoding efficiency is improved through redundancy removal, but quantization noise and mismatch noise accumulate
Solution Approach 1:
The system dynamically switches between using reconstructed frames and original frames as motion compensation references based on noise conditions. When quantization noise and mismatch noise are detected to be below thresholds, reconstructed frames are used for better compression. When noise exceeds thresholds, the system transitions to using original frames to prevent noise accumulation, thus adaptively optimizing the reference frame selection.
Solution Approach 2:
The invention changes the parameter of reference frame selection by introducing noise level detection and adaptive switching mechanisms. The system monitors quantization noise and mismatch noise levels, and based on these parameter changes, dynamically adjusts whether to use reconstructed or original frames as references, thereby resolving the contradiction between compression efficiency and noise control.
2Object-generated harmful factors
If original frames are used for motion compensation, then quantization noise is reduced, but the number of bits required for encoding increases
Solution Approach 1:
The system employs dynamic reference frame selection, switching between reconstructed frames and original frames based on noise thresholds and encoding conditions. This dynamic approach allows the system to use original frames only when necessary (when noise levels exceed thresholds), thereby reducing quantization noise while minimizing the increase in bit rate, as original frames are not used continuously.
Solution Approach 2:
Instead of always using original frames to eliminate quantization noise, the system applies partial action by using original frames only in specific regions or under specific conditions (when noise thresholds are exceeded). This selective use of original frames reduces quantization noise where needed while avoiding the excessive bit rate increase that would result from using original frames throughout the entire video stream.
3Productivity
If motion compensation uses reconstructed frames, then compression ratio is improved, but mismatch noise increases
Solution Approach 1:
The system implements feedback mechanisms by detecting mismatch noise levels during encoding and using this information to control reference frame selection. When mismatch noise exceeds a threshold, the feedback loop triggers a switch to using original frames instead of reconstructed frames, thereby controlling mismatch noise accumulation while maintaining compression efficiency through selective use of reconstructed frames.
Solution Approach 2:
The invention introduces dynamic switching between reconstructed and original frames based on real-time mismatch noise detection. This dynamic approach allows the system to maintain high compression ratios by using reconstructed frames when mismatch noise is low, while preventing noise accumulation by switching to original frames when mismatch noise exceeds thresholds.
Data Source
AI summary
Methods and systems for video coding are provided. A method includes receiving a digital video stream. The digital video stream includes multiple sequential independent frames. The method further includes storing a first frame of the digital video stream. The method also includes encoding a second frame of the digital video stream using motion compensation with the stored first frame as a reference.