drido

Thanks a lot Yunus! I guess you take a much more elegant approach and shuffled the inputs around a bit compared to mine. Your technique seems to be proven and part of your course, so I'll give it a try understanding, thanks a lot again. Btw, I am am intrigued of how you embedded the video, as the first image before starting shows the code, yet that image is nowhere to be found in the video

Hi, I try to accumulate refracted light through a water surface on the ground with a detail vex wrangle. My code so far is this, with the ground being input 0 and the water surface being input 1: vector sunDir = {0, -1, 0}; float ior = 1.33; int numpt1 = npoints(1); vector hitPoint; float u, v; for(int i = 0; i<numpt1; i++) { vector Psurf = point(1,"P", i); vector Nsurf = normalize(point(1,"N", i)); vector refractRay = refract(sunDir, Nsurf, ior); int hitPrim = intersect(0, Psurf, refractRay, hitPoint, u, v); int nearPoint = nearpoint(0, hitPoint); vector cdTemp = pointattrib(0, "Cd", nearPoint, 0); cdTemp += 0.1; setpointattrib(0, "Cd", nearPoint, cdTemp, "set"); } While I do get the caustic pattern on the ground, it doesn't seem to accumulate the light as I intend to do with the last 3 lines. Instead of adding 0.1 to the current Cd value each time a point gets hit in a loop step, it only ever outputs 0 or 0.1 as final color value. I suspect this being due to some parallel execution when I read from/write to points on the second input. I also tried a foreach loop as well as collecting the hit points in a detail array and iterate over that, with the same result. When I did a nested loop to run over every point on the second input for each hit point it does yield the desired result, but becomes really slow at a grid resolution of about 100x100. Which is understandable as it runs 10.000 times 10.000 = 100 Million times according to my understanding. Would anyone know how I can set this up so I can accumulate the color values efficiently on the hit points of the ground surface? thx+cheers hendrik