Motion Estimation Search Memory Layout for Cost-Controlled Video Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding technologies lack a suitable method for reducing costs while maintaining optimal performance, particularly in high-efficiency video coding (HEVC), due to inadequate control over memory access and search range size during video encoding.
Innovation Solution
A method and system-on-chip (SoC) that control and extend the search range size for motion estimation operations by configuring ME search ranges within a search memory, allowing for enhanced ME performance with optimized cost control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the search range size is increased to enhance motion estimation performance, then ME performance is improved, but memory access cost and processing complexity increase
Solution Approach 1:
The patent divides the search range into multiple segments by configuring different ME search ranges (first ME search range and second ME search range) corresponding to different ME regions. This segmentation allows the system to perform motion estimation with limited search range size in specific regions while maintaining the ability to extend search range when needed, thus resolving the contradiction between search range size and memory access control complexity.
Solution Approach 2:
The patent implements dynamic search range configuration where the video processing circuit can flexibly adjust the ME search range based on the current encoding conditions. The system can configure at least one ME search range corresponding to at least one ME region and extend the search range dynamically during motion estimation operations, allowing optimal balance between performance and complexity.
2Device complexity
If the search range size is limited to reduce memory access cost, then processing cost is reduced, but motion estimation accuracy deteriorates
Solution Approach 1:
The patent applies local quality by configuring different ME search ranges for different ME regions within the picture. Instead of using a uniform search range throughout, the system adjusts the search range size and configuration locally based on the specific requirements of each ME region, allowing cost optimization in regions where full search is not needed while maintaining accuracy where necessary.
Solution Approach 2:
The patent implements partial action by performing motion estimation with limited search range size in certain ME regions while having the capability to extend the search range when needed. This partial approach allows the system to reduce memory access cost in typical cases while maintaining the option to perform more extensive search when higher accuracy is required.
3Productivity
If CTU size is increased to improve encoding efficiency for high-resolution images, then encoding efficiency is improved, but backward compatibility and implementation complexity increase
Solution Approach 1:
The patent applies dynamics by making the CTU size adjustable rather than fixed. The video processing circuit can dynamically adapt to different CTU sizes based on the encoding requirements and picture characteristics. This dynamic adjustment capability allows the system to benefit from larger CTUs for high-resolution images while maintaining compatibility with existing implementations through flexible configuration.
Solution Approach 2:
The patent implements parameter changes by allowing adjustable CTU size configuration as a key parameter. The system can modify the CTU size parameter based on the specific encoding task, picture resolution, and performance requirements. This parameter flexibility enables optimization for high-resolution images while maintaining backward compatibility through configurable rather than fixed architecture.
Data Source
AI summary
A method for performing memory access control with limited search range size during video encoding includes: loading partial data of non-encoded video data corresponding to a picture into a search memory of a video processing circuit for performing at least one motion estimation (ME) operation of the video encoding; and configuring at least one ME search range corresponding to at least one ME region within memory space of the search memory for performing the at least one ME operation, wherein the at least one ME search range includes the at least one ME region, and the video processing circuit is arranged to perform the at least one ME operation with an extended search range greater than the at least one ME search range in the memory space for enhancing ME performance.