September 12th, 2020 - working on the beamtracer


No pictures today (who would want a screenshot of pure code anyway?). Now that the backburner simmering is done about how to minimize the required amount of computing power for the (anisotropic) vector beamtracer, I have resumed coding it.
Case in point : how to minimize the number of tests one has to do to determine the facet that a given light ray will intersect, without calculating the (supposed) intersection point of said light ray with each damn facet of the design?

For a long time I had been playing with the idea of a discrete space made up of a collection of elementary angular "segments" (pyramids actually) and use these as a way to filter out the facets that wouldn't be in the vicinity of the direction a given light ray is headed towards. Well I have the first part done (computing the list of elementary angular segments containing a given facet). Next up will be the same for a 3D pointer (a vector and a vertice).


Tom, a.k.a. Ludwig Von Sodabowski