Last-Level Projection Hardware to Reduce VR Motion Artifacts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Head-mounted augmented reality and virtual reality systems suffer from rendering-to-motion miscorrelation, leading to movement artifacts like smearing and strobing, which can induce nausea in users, due to the inefficiencies in existing reprojection solutions that utilize software-based graphics engines and cause significant software overhead and context switching.
Innovation Solution
A hardware-based reprojection method is implemented using a dedicated graphics processor to reprocess frames before display, reducing software overhead and context switching by leveraging hardware acceleration for efficient frame reprocessing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If software-based reprojection is used, then implementation simplicity is maintained, but processing time and software overhead increase significantly
Solution Approach 1:
The patent replaces the software-based reprojection mechanism with a hardware-based solution using a dedicated reprojection unit. This substitution of mechanical (software) system with hardware circuitry enables significantly faster frame repiping without the overhead of software context switching and processing, directly resolving the contradiction between implementation simplicity and processing time.
2Adaptability or versatility
If software-based reprojection is used, then flexibility is maintained, but context switching inefficiency increases
Solution Approach 1:
The patent replaces software-based reprojection with a dedicated hardware reprojection unit that operates independently of the main graphics rendering pipeline. This hardware acceleration eliminates software context switching overhead and improves productivity while maintaining flexibility through configurable reprojection parameters and support for multiple display devices.
3Reliability
If frame repiping is performed in advance of vsync, then rendering-to-motion miscorrelation is reduced, but processing time is consumed
Solution Approach 1:
The patent implements preliminary frame repiping by capturing the rendered frame and immediately piping it to the display through the reprojection unit before the vsync event. This preliminary action corrects rendering-to-motion miscorrelation by accounting for head movement between rendering and display, while the efficient hardware implementation minimizes the time consumed.
Solution Approach 2:
The patent maintains continuity of useful action by keeping the reprojection unit actively processing frames in parallel with the rendering pipeline. The reprojection unit continuously repipes frames without blocking the main rendering process, ensuring both high rendering-to-motion correlation and efficient time utilization through overlapping operations.
Data Source
AI summary
An apparatus and method for efficient image reprojection in a virtual reality system. For example, one embodiment of an apparatus comprises: a sensor interface to collect motion data from one or more sensors during a virtual reality session; graphics circuitry to execute graphics program code to render an image frame during the virtual reality session; a processor to generate motion transform data using the motion data, the motion transform data specifying how the image frame is to be adjusted prior to display; a reprojection engine to perform an in-line reprojection of the frame using the motion transform data to generate a reprojected image frame; and display circuitry to display the reprojected frame.