Visual and performance comparison: Physically based atmosphere method against Unity Universal Render Pipeline procedural skybox

Sammanfattning

Background. In the pursuit of realism in rendering, a feature that may get overlooked is the sky and atmosphere—especially the consumer who might suspect that something is off without identifying it. A critical aspect of sky rendering is the color that should be dynamic with the time of day to be realistic. With multiple ways of rendering a sky and atmosphere, it is unclear which method best suits the circumstance. Objectives. The objective of this study was to make that decision clearer with both the visual aspect and the realism. It was also crucial that performance was not affected too heavily by the techniques. Methods. As the first objective can be subjective, we decided to ask several people through a survey to get a more objective answer. In the survey, we presented the participants with two images of identical scenes where the only difference between them was the method of rendering the sky. We then asked them which picture they thought was the most aesthetically pleasing and which they thought was the most realistic. The second part of the study was to test the performance of the techniques. We recorded the average frame time through a C\# script over all of the scenes. Results. When using the settings we established, the Physically-based atmosphere scored lower in both realism and aesthetics in the first visual test but scored higher in realism in all of the scenarios for the second visual test. Both the average frame rate and frame time did end up in favor of the Physically based atmosphere. Conclusions. The outcome of this study was surprising, we had thought that the Physically-based atmosphere would score higher in both realism and aesthetics. The performance of the Physically based atmosphere method proved to be much more optimized than the Unity procedural skybox than we would have thought.