Video Codec Complexity Control via Dynamic Rate Distortion Threshold
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video coding methods face challenges in efficiently managing resource usage and maintaining consistent encoding times, often requiring fine-tuning and overprovisioning, which can lead to inefficiencies and increased costs, while existing complexity control methods result in high variation of encoding time across Group Of Pictures (GOPs).
Innovation Solution
A fully scalable complexity control method that dynamically adjusts coding units by determining the rate distortion cost for a reduced number of coding modes, recursively splitting coding units into a quad-tree structure, and using a controller to adapt the rate distortion cost error threshold, ensuring optimal resource utilization and minimizing the difference between actual and target complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional video coding methods are used with fine-tuning and overprovisioning, then resource utilization is improved, but system complexity and cost increase
Solution Approach 1:
The patent implements dynamic complexity control by adjusting the number of coding modes evaluated based on available resources. The system dynamically selects between evaluating all coding modes or a reduced set, allowing the encoder to adapt its computational complexity in real-time to match resource availability, thereby improving resource utilization without permanently increasing system complexity
Solution Approach 2:
The patent changes the parameter of coding mode evaluation by introducing a flag that controls whether to evaluate all coding modes or only a reduced set. This parameter change allows the system to switch between high-quality mode and fast mode, optimizing resource utilization without requiring complex hardware overprovisioning
2Stability of the object's composition
If buffering is used to solve resource competition, then processing stability is improved, but encoding delay increases
Solution Approach 1:
The patent dynamically adjusts the encoding complexity based on real-time resource availability rather than using fixed buffering. By monitoring resource conditions and adapting the number of coding modes evaluated, the system maintains processing stability without introducing the delays associated with buffer-based approaches
3Reliability
If over dimensioning the system for worst case is done, then reliability is improved, but resource efficiency deteriorates
Solution Approach 1:
The patent implements dynamic adaptation that allows the system to reliably handle worst-case scenarios when resources are available while efficiently utilizing reduced resources when they are constrained. The system switches between evaluating all coding modes (high reliability) and a reduced set (high efficiency) based on real-time conditions, avoiding the need for permanent over-dimensioning
4Speed
If a reduced number of coding modes is evaluated, then encoding speed is improved, but rate distortion cost increases
Solution Approach 1:
The patent dynamically adjusts the number of coding modes evaluated based on resource availability and importance of the current video block. For important blocks, more modes are evaluated to maintain quality, while for less important blocks, fewer modes are evaluated to improve speed, thus balancing rate distortion cost against encoding speed dynamically
Solution Approach 2:
The patent applies different levels of coding mode evaluation to different regions or blocks within the video frame based on their importance. This allows the system to maintain high quality (low rate distortion cost) for critical regions while using faster, reduced-mode evaluation for less critical regions, optimizing the overall speed-quality tradeoff
Data Source
AI summary
Means for controlling the computational complexity related to video encoding that includes a dynamic rate distortion cost error threshold and constrains the encoding complexity to a given complexity target. The solution is selective and provides fast convergence while incurring only a limited loss in rate distortion performance.