Second Pass Video Block Mode Changes Without Motion Recalculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-pass video coding, the second pass often requires a full re-encode of video blocks, which is computationally intensive and inefficient, especially when adjusting bit-rate constraints or exploiting unused bandwidth.
Innovation Solution
The second coding pass changes video block modes between skipped and direct modes without altering motion information, and adjusts quantization parameters, allowing for residual information addition or elimination to refine coding, thereby reducing computational complexity and enhancing quality or reducing bit-rate effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a full re-encode of video blocks is performed during the second coding pass, then coding enhancements and bit-rate adjustments can be achieved, but computational complexity increases significantly
Solution Approach 1:
The patent segments the video coding process into two distinct passes: a first pass that performs full encoding with motion estimation and compensation, and a second pass that only performs mode changes and quantization adjustments. This segmentation allows the computationally intensive operations to be performed only once in the first pass, while the second pass focuses on lightweight refinements, thereby reducing overall computational complexity while maintaining coding enhancement capabilities.
Solution Approach 2:
The first coding pass performs all preliminary computationally intensive operations including motion estimation, motion compensation, and initial encoding. By completing these actions beforehand, the second pass only needs to perform simple mode changes and quantization adjustments, significantly reducing the computational burden during the second pass while still achieving coding enhancements.
2Manufacturing precision
If quantization parameters are adjusted in the second pass, then bit-rate constraints can be met, but full re-encoding is required which reduces productivity
Solution Approach 1:
The patent separates the quantization parameter adjustment from the full encoding process. Quantization parameter adjustments are performed only in the second pass after the initial encoding is complete, allowing bit-rate control to be applied selectively without requiring a complete re-encoding of all video blocks, thus maintaining productivity.
Solution Approach 2:
The patent changes quantization parameters in the second pass to meet bit-rate constraints. By adjusting these parameters after the first pass encoding, the system can control the final bit-rate without needing to re-perform the entire encoding process, thereby maintaining coding efficiency while achieving precise bit-rate control.
3Manufacturing precision
If coding modes are changed in the second pass, then quality refinement is achieved, but motion information must be re-calculated increasing computational load
Solution Approach 1:
Motion information is calculated and determined during the first coding pass beforehand. In the second pass, when coding modes are changed, the previously calculated motion information is reused without recalculation. This preliminary action of computing motion data once eliminates the need for redundant calculations in the second pass, maintaining quality refinement while reducing computational load.
Solution Approach 2:
The patent copies and reuses the motion information calculated in the first pass during the second pass. Instead of recalculating motion vectors and compensation data when mode changes occur, the system copies the existing motion information and applies it to the modified coding modes, significantly reducing computational complexity while achieving quality refinement.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
This disclosure describes techniques for second pass video coding in a multi-pass video coding scenario. The coding modes for some video blocks encoded during a second pass may be changed relative to the coding modes used for such video blocks in the first pass. However, motion information does not change for those video blocks that have the changed modes. In particular, mode changes can be made in the second coding pass relative to the modes used in the first coding pass without changing the manner in which motion information will be derived at the decoder, e.g., due to similarities between the original modes of the first pass and changed modes used in the second pass. The second pass coding techniques may also include quantization parameter adjustments, and the mode changes can cause such quantization parameter adjustments to have more profound refinements effects on the second pass coding.