Did a quick search and didn't find anything...
I turned on GI for a planet render to fill in the shadows on a model. The partial render is below, with a completed render without GI. The area south of the equator is identical to the image without GI.
Detail 2, AA 3, GI 3,4
Bump map, surface texture and cloud blending shader all spherical projected images.
Are you using any coverage adjust in your cloud nodes?
- Oshyan
Awesome render and perspective, by the way.
There is one positive and two negative coverage adjustments in the clouds nodes (forgot to remove the positive one). Their density fractals all have positive coverage adjustments.
It does seem to be cloud related.
- Set all coverage settings (cloud and fractal) to 0: still there
- Disabled all clouds: not there (but not much difference to no GI render, so it may be that it's just not noticeable)
- Disabled existing clouds + default cirrus layer: still there
I have seen similar problems before when using coverage adjustment specifically in a cloud layer. Would it be possible to provide the .tgd for further investigation? You can email it to me privately if you prefer.
- Oshyan
Already attached to the first post ;)
Certainly no secrets in there.
Texture maps are from here:
http://www.oera.net/How2/TextureMaps2.htm (http://www.oera.net/How2/TextureMaps2.htm)
Following in tandem with RArcher's problem... here's a cropped render with GI on and the reflecivity shader removed. The distribution of GI has increased a bit and you can clearly see that it's not a straight line.
I'd added a weak fill light as well but I don't think it made the difference.
Still guessing....
One thing I noted was that the camera position has a large negative Y value. I set it to 0 and it seems to be working although of course this gives a completely different view. I'm running a 10 frame animation between the two positions to see if it's possibly related to this.
Thanks for the continued investigation. I'll try to get a chance to look at you and Ryan's files this weekend.
- Oshyan
Made some progress, but not what I expected.
The first frame displayed the error even though the previous render of the same camera position didn't. The only difference between the two was the detail level (it takes a while to render) so I stopped the sequence and tried some different detail settings. At 0.5 it doesn't appear from this camera angle... it's there at 0.75, and spreads further with detail at 1.0. Added another image with render detail 2.0 showing that the area without GI has spread.
Just giving it a nudge, with the last image added
Hi Ben,
Thanks for the new information. I will look into this problem soon.
Matt
I'm currently hacking on RAarcher's file and will check this one out when I'm through with it. ;D
- Oshyan
No hurry... I was just wondering if there was anything else you wanted me to try rather than you having to check through everything. It may just be coincidence but there did seem to be some sort of link between the two problems. If you find a workaround for RArcher's problem let me know and I'll test it on mine.
I think both are just examples of general GI issues that we're already aware of, although one or both files may be useful for tracking down specific issues and resolving them. RArcher's scene in particular seems to have a very interesting behavior. One thing I can say is cranking GI up to 8/8 got rid of the problem. ;)
- Oshyan
I'll keep that in mind if I get really desperate ;)
I can imagine the render time.
Turns out turning GI Relative Detail down to 1, with any setting of Sample Quality, seems to also fix it on Ryan's scene. I'll run some tests on your scene shortly with what I've learned on his. It would be interesting if it behaved similarly.
- Oshyan
Hi Ben,
It would be interesting to see if you can narrow it down to certain elements of the scene (i.e. not just render settings). For example it would be interesting if it only happens with atmosphere or cloud layers, or with reflective oceans etc.
Matt
OK Matt
I'll pull it apart piece by piece and see if anything changes.