I have a vibration problem with this animated water. I included two frames…putting one on top of the other you can see an area of the pool water go darker. Since it is a large area, it does not seem to be a problem with noise applied to the bump map. For the movement I have animated the phase control of the noise. Below I have posted the material settings and vray settings (vray 1.5 with max 7 on Windows XP 64). In addition, to avoid any flickering, I did the IRMAP calculation on one machine and also each frame is rendered on one machine. Any ideas…have to get this animation out as soon as possible. Thanks!
simple enough. you cant have any animations other than the camera when using saved maps. thats probably why you get such a difference
Though, I do not have an object moving…it is just an effect of the map. For example…I also have a waterfall…animated bitmap…that does not have this flickering problem.
a waterbumpmap is as bad as moving geometry in this case i guess. Due to refraction it changes what part of the pool is seen where. If it changes to much there might be areas that dont have enough Irrmap info. That´s a wild guess tho.
Thorsten
that was my initial gues. reflections and refractions would show different areas if animated. some areas might have no samples
refraction caustics in the GI settings is the culprit, coupled with the big area of refractive surface.
Too much changes direction by the frame as pointed out before.
Lele
So, do I need to turn off the refractive caustics? Or, do I need to exclude the water from the GI calculations?
I tried both of the above, but it did not help the problem. Do not tell me that the only way is to render out each IRMAP for each frame?!?
The problem does not seem to be in the IRMAP calculation…I rendered out a couple frames doing all calculations for each frame…but still had the same problem.
Try using QMC/Lightcache instead of IRRMap
Same thing happens with QMC.
Can you post a sequence of frames to show of the problem more clearly than just 2 images.
How can I post the sequence?
upload to rapidshare.de
Ok…here it is…
http://rapidshare.com/files/463680/piscina.m1v
Hmm ok, that clearly displays the problem, how was that example rendered?
can you also post just the pool section of your scene for us to play with?
The example is rendered with the first settings that I posted. I tried other suggestions, but received the same result…even when excluding the water from the GI calculation. Though, I was always using a saved LC map…I should try without that calculation saved.
Try looking for duplicated geometry… it looks like you might have to copies of the pool surface.
It seems all your normals are out of whack for the pool mesh in that file that you uploaded. There wasnt any duplicate geometry that i could see, but the flipped normals would give that impression. Actually there would be some intersecting geometry near the pool mesh and the surround meshes which might be worth while trying to clean up.
Also it looks as tho your water mesh will have coinciding faces with the pool mesh, I suggest that you use just a plane instead of a closed mesh.
Also you wont be able to use saved GI solutions… you can try but i dont think the results will be that good.. but they might be not noticeable after you fix up the normals and coinciding faces