Selective Video Frame Quantization for Quality-Preserving Bit Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data processing methods for video streams struggle with optimizing encoding efficiency, particularly in reducing bit rates without compromising subjective quality, especially when dealing with long videos and varying user equipment and network conditions.
Innovation Solution
A data processing method that analyzes an encoding file of a video stream to identify target image frames and their corresponding quantization parameters based on a quality evaluation index. The method determines new quantization parameters and bit allocations for these frames, which are then used to update the initial parameters during secondary encoding, thereby optimizing bit rate usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing per-title encoding optimization schemes are used, then encoding efficiency is improved, but computational complexity increases and they are not suitable for long videos
Solution Approach 1:
The patent segments the video stream into multiple video chunks based on time duration, and further segments each chunk into groups of picture frames. This hierarchical segmentation allows the complex optimization problem to be divided into smaller, more manageable sub-problems that can be solved independently for each segment, reducing overall computational complexity while maintaining encoding efficiency.
Solution Approach 2:
The patent performs preliminary analysis on the encoding file to identify target picture frames that require optimization before the actual re-encoding process. By pre-identifying which frames need quantization parameter adjustments based on quality evaluation indices, the system avoids unnecessary computations on all frames, thereby reducing computational complexity while preserving encoding efficiency.
2Productivity
If machine learning-based Per-Title implementation is used, then encoding optimization is improved, but computational resources and time are consumed
Solution Approach 1:
Instead of applying machine learning-based optimization to all picture frames in the video, the patent selectively applies optimization only to identified target picture frames that meet specific quality evaluation criteria. This partial action approach reduces processing time and computational resource consumption while still achieving meaningful encoding optimization where it is most needed.
Solution Approach 2:
The patent applies different quantization parameter adjustments to different picture frames based on their local quality characteristics. By evaluating each frame's quality index and applying optimization only where necessary, the system achieves localized quality improvement without the excessive computational burden of global optimization approaches.
3Manufacturing precision
If quantization parameters are adjusted for all frames, then quality is improved, but bit rate increases
Solution Approach 1:
The patent adjusts quantization parameters only for target picture frames that have quality issues, rather than uniformly adjusting all frames. This localized adjustment approach improves quality where needed while avoiding unnecessary bit rate increases in already-acceptable frames, thus resolving the contradiction between quality improvement and bit rate control.
Solution Approach 2:
The patent dynamically changes quantization parameters based on frame-specific quality evaluation results. By modifying parameters only for frames that fail quality thresholds and maintaining original parameters for acceptable frames, the system achieves quality improvement without proportional bit rate increases across the entire video stream.
Data Source
AI summary
A data processing method and apparatus. An implementation solution including: analyzing an encoding file of a video stream to obtain a target image frame set corresponding to the encoding file and target quantization parameters corresponding to the target image frame set, a target quantization parameter being quantization parameter of a target image frame when the target image frame meets a quality evaluation index; one the basis of a comparison result of the target quantization parameters corresponding to the target image frame set, the encoding file and initial quantization parameters of image frames of the video stream, determining new quantization parameters of the target image frames in the target image frame set and new bit numbers of the target image frames corresponding to the new quantization parameters; and updating the initial quantization parameters and the initial bit numbers of the image frames of the video stream.