Hey Sven, thanks for the info.
I am range testing lowering Irradiance rays against density. I had hoped that I could use a higher density to keep the rays closer to 128-256. (I´m pushing the classroom scene, loads of nasty noise) Irradiance Particles feel very slow, mostly because of the lack of visual feedback, that sucks when tweaking but once that damn preprocessing is done. Am now at 256-350 rays and density 2 My frametimes are all around 01:30:00 - 02:45:00, yes 1 1/2 hours and up at HD720 w/DOF. Clean, presentable frames, using unified sampling (which is really nice) The Automatic FinalGather alternative is similarly fast/slow with 1536-2048 rays. All that is mR 3.9.x, so I lack a few optimizations and yes, it hurts. Cheers, tim On 28.03.2013 00:53, Sven Constable wrote:
Be very careful with the importons density, because the slider range in SI is totally off scale. Value of one means one importon ray per pixel, that’s a *very* high quality setting. I use values down to 0.05 for (ultra)fast previews and ramp it up to 1 for high quality (overnight) renders... so value of 2 or above is like rendering with AA of min5max5 as a comparison.:) Better factory scale for the slider would have been 0.01 to 2 I think. sven -----Original Message----- From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Tim Leydecker Sent: Wednesday, March 27, 2013 23:56 To: softimage@listproc.autodesk.com Subject: Softimage 2013/2014 Irradiance Particles - IP Optimize still single threaded? I´m using xsi2012sap here and find it brutally anoying to see the Irradiance Particle calculation´s IP Optimize use a single CPU thread only, especially for settings with Importons Density dialed above 1. At least I can´t tell a quality increase between a setting of 1 to 2, just the way you´d want things to turn out... Irradiance Particles are really nice otherwise, seem to even be energy conserving actually. If you compare Iparticles to how the Fallof Start Stop option "works" for good old FinalGathering, that is neither desireable nor nice to make use of it... Any info on reworked mR satellite support with Softimage 2014? All that is boring me to the brink of buying a personal V-Ray 2.0 license and even thinking about how much hassle Maya would be anyway... Cheers, tim P.S: I would have thought nVidia Corp. is clever enough to realize they have the fastest GPUs to make mR rendering at least seem nice...