View Full Version : interior fg spots
07-11-2007, 04:49 PM
hello, i am rendering an animation in an old house.
i am using mia sun/sky with fg (multibounce)
even with 1200 fg rays (radius control quality), i get thee spots on the wall.
i tried various min/max radius, the spots are still there.
(fg point density is not an option, it takes crazy time to render)
07-11-2007, 05:11 PM
have you tried playing with your fg filtering?
07-11-2007, 08:47 PM
It seems you need to increase the interpolation value.
I wouldn't use radius control, the new mr is very good at automatically adjusting this. Also make sure you have the FG filter set to atleast 1.
I would start at:
point density .1
If you need more detail increase point density to like .3 and decrease interpolation to around 75. If you need more, go to .5, 50. Need more, 1 point density, 10, maybe 20 if it still splotches.
07-12-2007, 04:56 AM
you could achieve something like that with a directional or area light outside of the window, a few fill lights inside, and an ambient occlussion all rendered out as seperate passes, so you wouldnt even need FG.
07-23-2007, 04:36 PM
why is it that every time i rase my FG filter to 2, i get a black interior
07-24-2007, 10:57 PM
FG filter will darken it the higher the value is because it will adjust the FG points based on the values of the surrounding FG points. So in the case of interiors, the farther you go into the corners or darker areas you get fewer FG rays that are hitting bright spots vs the darker spots. So it adds extra darkness into the dark areas because it's darkening the fewer brighter FG points to match their surrounding darker FG points.
Now the question is, since it does sort of artifically darken it in a sense vs following the actual value of the FG ray as it was calculated, is it still physically accurate? I'm not completely sure maybe someone who knows more can offer an explanation. Because technically speaking you would say a FG filter of 0 is the most physically accurate. But I've found an FG filter of 1 to produce an image that reads to the eye as more realistic.
Although I do believe that mental images intended this to work with GI photons for the secondary bounce illumination. Which are not affected by the FG filter and would probably produce a more realistic result because even with an FG filter of 4 it still won't darken the corners as much because the corners overal value would be based on the GI photons.
07-24-2007, 10:57 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.
vBulletin v3.0.5, Copyright ©2000-2013, Jelsoft Enterprises Ltd.