Improving Real-Time Rendering Quality and Efficiency using Variable Rate Shading on Modern Hardware (DAEV)

The aim of this master thesis is to research the possibilities of variable rate shading, a new hardware feature introduced with Nvidia Turing Graphic-Chips. In order to improve performance and quality of real-time graphics applications, this feature allows to control the sampling rate of tiles in the screen dynamically. For example, it enables algorithms for the variation of the sampling rate based on the scenes current content, called content adaptive shading, and the variation of the sampling rate based on the motion in the scene, called motion adaptive shading. In combination with an eye tracking device, it allows the adaption of the sampling rate based on foveal vision. As such, it enables the development of hardware accelerated foveated rendering. Another goal of this master thesis is to investigate the applicability to existing rendering techniques, like deferred shading. Therefore, the implemented techniques will be integrated into a deferred shading renderer. An evaluation of the performance gain and the perceived visual experience will conclude this master thesis.