Real-Time Shading With Area Light Sources | Two Minute Papers #65

Dear Fellow Scholars, this is Two Minute Papers with Károly Zsolnai-Fehér. In computer graphics, we use the term shading to describe the process of calculating the appearance of a material. This gives the heart and soul of most graphical systems that visualize something on our screen. Let the blue sphere be the object to be shaded and the red patch be the light source illuminating it. The question is, in this configuration, how should the blue sphere look in reality? In order to obtain high-quality images, we need to calculate how much of the red patch is visible from the blue sphere. This describes the object's relation to the light source. It close or is it nearby? Is it facing the object or not? What shape is the light source? These factors determine how much light will arrive to the surface of the blue sphere. This is what mathematicians like to call an integration problem. However, beyond this calculation we also have to take into consideration the reflectance of the material that the blue sphere is made of. Whether we have a white wall surface or an orange makes a great deal of difference and throws a wrench in our already complex calculations. The final shading is the product of this visibility situation and the material properties of the sphere. Needless to say that the mathematical description of many materials can get extremely complex, which makes our calculations really time consuming. In this piece of work, a technique is proposed that can approximate these two factors in real time. The paper contains a very detailed demonstration of the difference between this and the analytical computations that give us the perfect results but take extremely long. In short, this technique is very closely matching the analytic results, but it is doing it in real time. I really don't know what to say. We're used to wait for hours to obtain images like this, and now, 15 milliseconds per frame. What a hefty value proposition for a paper. Absolutely spectacular. Some of the results really remind me of topological calculations. Topology is a subfield of mathematics that studies what properties of different shapes are preserved when these shapes are undergoing deformations. It's super useful because, for instance, if we can prove that light behaves in some way when the light source has the shape of a disk, then if we're interested in other shapes, topology can help us determine whether all these enormous books full of theorems on other shapes are going to apply to this shape or not. It may be that we don't need to invent anything and can just use this vast existing knowledge base. Some of the authors of this paper work at Unity, which means that we can expect these awesome results to appear in the video games of the future. Some code and demos are also available on their website which I've linked in the description box, make sure to check them out! Thanks for watching, and for your generous support, and I'll see you next time!