Page 50 of 109

Re: PhotonGI cache

Posted: Fri Feb 22, 2019 11:27 am
by lacilaci
Dade wrote: Fri Feb 22, 2019 11:22 am
lacilaci wrote: Fri Feb 22, 2019 7:55 am seems that some recent changes made light leaking worse... nothing helps
I can not add anything more without a test scene.
here:
glass_interior_test_279.blend
(2.17 MiB) Downloaded 154 times

Re: PhotonGI cache

Posted: Fri Feb 22, 2019 2:07 pm
by marcatore
I've tried your scene and it works in the same way.

I've also tried to change the units scale and scale down the objects according to the "real" scale, apply the scale but nothing change.
I've tried this to check of some units transformation introduce some problem... bu nothing change in the rendering

Re: PhotonGI cache

Posted: Fri Feb 22, 2019 2:14 pm
by lacilaci
This wasnt happening before, something has changed in recent builds... or a bug

Re: PhotonGI cache

Posted: Fri Feb 22, 2019 2:45 pm
by marcatore
if you use a smaller Lookup radius the leaks disappear...try with 5cm for example.

From my point of view, the best solution is to have something not related with the geometry thickness. Modelling the thickness of the wall is not the quickest way to model a room or a building. Often, having just flat surfaces without thickness give more flexibility.

Re: PhotonGI cache

Posted: Fri Feb 22, 2019 3:10 pm
by Dade
lacilaci wrote: Fri Feb 22, 2019 2:14 pm This wasnt happening before, something has changed in recent builds... or a bug
It is caused by the yesterday change: viewtopic.php?f=5&t=840&start=480#p10380

I reverted the change so now your scene works fine (but your radius is really big for this scene has spotted by Marcatore).

@Marcatore, about the original problem (viewtopic.php?f=5&t=840&start=470#p10360), just use the default parameters so the floor will be rendered with normal path tracing in order to avoid the artifact. Or just use the flag to disable PhotonGI on floor material if you want to change the parameters anyway.

Re: PhotonGI cache

Posted: Fri Feb 22, 2019 3:39 pm
by lacilaci
I'm thinking, shouldn't cache filtering radius get larger as cache entry radius gets smaller?

Re: PhotonGI cache

Posted: Fri Feb 22, 2019 4:03 pm
by Dade
lacilaci wrote: Fri Feb 22, 2019 3:39 pm I'm thinking, shouldn't cache filtering radius get larger as cache entry radius gets smaller?
At the moment most parameters a relative to the radius and it is often quite handy. They could be expressed in absolute term but than you would have to change all the parameters if you want the same result with a different radius.

Re: PhotonGI cache

Posted: Fri Feb 22, 2019 5:00 pm
by lacilaci
What is normal angle parameter doing? low values seem to make image darker
higher values seem much more closer to normal pathtracing...

EDIT: I'm testing this on the 2.1 benchmark scene

Also, no matter what I do I cannot make the 2 wall objects gather photons
cacheblack.JPG
(reseted scale, shading, collapsed modifiers and turned ngons to triangles etc...) they just don't gather photons

Re: PhotonGI cache

Posted: Fri Feb 22, 2019 5:34 pm
by Dade
lacilaci wrote: Fri Feb 22, 2019 5:00 pm (reseted scale, shading, collapsed modifiers and turned ngons to triangles etc...) they just don't gather photons
It looks like another NaNs problem, I will check.

Re: PhotonGI cache

Posted: Sat Feb 23, 2019 7:20 am
by lacilaci
So, why does the default normal angle kill so much of the light? 10° seems very low

90deg normal angle:
1min13_90deg.jpg
10deg normal angle:
1min13_10deg.jpg
10degree normal angle kills a lot of GI light and also seems to be darker than pathtracing only..

by the way these are 1min and 13sec renders on gtx1060!