PDA

View Full Version : Mental Ray: Jagged shadows and highlights

schultzi
03-11-2005, 08:14 PM
Well, I'm working on a little personal project, and up until now, I've had no reason to render above 1280x1024. In my scene, renders at and below 1280x1024 look fine, but when I attempted a hi-res panorama shot (3200x1600), I ended up with really nasty white jagged highlights along the bottom edge of the objects and funky looking jagged shadows. Caustics, GI, and Final Gather are all enabled, and the final gather radius is at a max of 15 and a min of .1 with 1000 samples. Under the "Renderer" tab, the min samples per pixel is at 1/64, and the max is at 1024. The main light source is a skylight, and I included a large, white, self-illuminating plane above the objects that is invisible to the camera and doesn't cast or recieve shadows. Here's a small piece of the large panorama view with the problem areas highlighted. Can anyone point me in the right direction?

http://img203.exs.cx/img203/9325/problem8zt.jpg
Crocodilian
03-11-2005, 09:49 PM
when you increase resolution from 1280x1024, did you increase GI and Final Gather numbers? In round terms, you're now calculating 5.1 million pixels, versus 1.3 million previously, almost exactly four times the resolution.
schultzi
03-12-2005, 12:23 AM
No, thank you very much, I completely forgot and left them the same. Wow, didn't even think about that. But I have another question, when doing an ambient occlusion pass, should the reflective output plane above the objects be given the ambient material, or should it remain the same output material originally given to it?
TheWriter
03-12-2005, 11:52 AM
Do not forget to increase your shadow map sizes as well. And perhaps more sampling as artifacts are a lot more noticable in larger renders.
Crocodilian
03-12-2005, 03:26 PM
One more thing:

I'd previously posted a question about mental ray Final Gather settings, and Jeremy Birn made the following observation:

"FG is sort of a "2 in 1" effect - on one hand, it samples everything that's visible from the point of view of a surface, almost like a very diffused version of a raytraced reflection. On the other hand, it also samples any photons in the area, including GI and Caustic Photons, so it tends to smooth out photon mapped GI which otherwise can get very blotchy looking in MR."

I haven't explored these settings in depth, but taking Jeremy's cue (usually a good thing to do. . .) I'd experiment with increasing your FG settings independant of the photon count; this may smooth the jagginess you're seeing.

Re: "Other Question"
[ when doing an ambient occlusion pass, should the reflective output plane above the objects be given the ambient material, or should it remain the same output material originally given to it?] I don't know. Good question.
CGTalk Moderation
03-12-2005, 03:26 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.