4 ms·
Pretty awesome visualization. Does anyone know if the authors published a paper on the "shadow computation that is 10 to 15 times faster than older methods"? I
by shrx 9y ago
Pretty awesome visualization. Does anyone know if the authors published a paper on the "shadow computation that is 10 to 15 times faster than older methods"? I couldn't find it on their websites [1-3].
[1] https://vgc.poly.edu/~fmiranda/ https://vgc.poly.edu/~fmiranda/
[2] https://vgc.poly.edu/~csilva/ https://vgc.poly.edu/~csilva/
[3] http://www.harishd.com/home/ http://www.harishd.com/home/
- theoh 9y agoIt sounds like a way to take larger simulation steps by linearising the movement of projected points on shadow outlines. I'm guessing a bit, but the idea of figuring out how to take larger steps is an established trick for speeding up simulations (e.g. the Siggraph paper "Large Steps in Cloth Simulation") "Instead of explicitly tracking the movement of the sun, which is how shadow accumulation is usually measured, the researchers are tracking the movement of the shadows. Using the key insight that shadows move linearly within a short time range, they designed an efficient algorithm that performs an order of magnitude faster than those used in existing approaches. “This also allows users to interactively compute and explore shadows over different time ranges and allows them to quickly test multiple scenarios,” said Doraiswamy."