Video Encoder Prediction Update Selection for Faster Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently processing increasing amounts of digital video data, particularly in terms of encoding and decoding efficiency and speed, without increasing processing time or circuit size.
Innovation Solution
The proposed solution involves an encoder and decoder that utilize a bi-directional optical flow (BDOF) process and intra prediction mixing process, which are mutually exclusively applied at a specific stage in the pipeline process, enhancing the selection of appropriate components and operations for encoding and decoding, such as filter, block size, motion vector, and reference picture, to improve efficiency and speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If multiple update processes (BDOF and intra prediction mixing) are applied sequentially to improve prediction accuracy, then encoding quality is improved, but processing time and circuit complexity increase
Solution Approach 1:
The patent applies update processes selectively rather than universally. The encoder determines whether to apply BDOF, intra prediction mixing, or both based on block characteristics and conditions. This partial application maintains prediction accuracy where needed while avoiding unnecessary processing elsewhere, thus reducing overall processing time without sacrificing essential encoding quality.
Solution Approach 2:
The patent introduces dynamic selection of update processes based on varying conditions. The encoder adapts which update process to apply (BDOF, intra prediction mixing, or neither) depending on block type, motion characteristics, and other parameters. This dynamic approach optimizes the balance between prediction accuracy and processing efficiency for different regions of the video data.
2Manufacturing precision
If multiple update processes (BDOF and intra prediction mixing) are applied sequentially to improve prediction accuracy, then encoding quality is improved, but circuit size and device complexity increase
Solution Approach 1:
The patent implements update processes selectively based on determined conditions. The circuitry includes control logic that determines whether to apply BDOF, intra prediction mixing, or neither, based on block characteristics. This selective application reduces the need for always-active circuit paths, thereby reducing overall circuit size and complexity while maintaining prediction accuracy where required.
Solution Approach 2:
The patent divides the update process into separate, independently controllable stages: BDOF update and intra prediction mixing update. Each update process can be independently enabled or disabled based on conditions. This segmentation allows the circuit to implement only the necessary update processes for each block, reducing overall circuit complexity compared to a design that always applies all updates.
3Adaptability or versatility
If more encoding components and operations are selected and applied, then encoding quality and adaptability are improved, but processing efficiency and speed decrease
Solution Approach 1:
The patent implements dynamic selection of encoding components and operations based on block characteristics and conditions. The encoder adapts which update processes to apply (BDOF, intra prediction mixing, or neither) rather than applying all possible operations uniformly. This dynamic approach maintains high encoding quality and adaptability for complex blocks while improving overall encoding speed by skipping unnecessary operations for simpler blocks.
Data Source
AI summary
An encoder: generates, in an inter prediction mode, a first prediction image of a current block to be processed, based on a derived motion vector; and generates a final prediction image of the current block by applying an update process to the first prediction image. Candidates for the update process include a first process and a second process. The first process is a BDOF process. The second process is a process of mixing the first prediction image with a second prediction image generated in intra prediction for the current block. In the applying of the update process, the first process and the second process are mutually exclusively applied.