Atmosphere Simulation

Post your tests, experiments and unfinished renderings here.
User avatar
B.Y.O.B.
Developer
Posts: 3753
Joined: Mon Dec 04, 2017 10:08 pm
Location: Germany
Contact:

Re: spectral rendering

Post by B.Y.O.B. » Fri Mar 06, 2020 9:08 pm

With no camera volume instead of auto volume detection:
Attachments
rend3_nocamvol_20min.png

wasd
Posts: 315
Joined: Tue Apr 24, 2018 7:20 pm

Re: spectral rendering

Post by wasd » Fri Mar 06, 2020 10:34 pm

B.Y.O.B. wrote:
Fri Mar 06, 2020 9:08 pm
With no camera volume instead of auto volume detection:
Yes, that's better! I see how different horizon line is. There's no difference if bump is disabled though.
CPU Bidir (sometimes GPU Path) + Metropolis | Log Power | Core i5-4570

User avatar
B.Y.O.B.
Developer
Posts: 3753
Joined: Mon Dec 04, 2017 10:08 pm
Location: Germany
Contact:

Re: spectral rendering

Post by B.Y.O.B. » Fri Mar 06, 2020 10:57 pm

To me it looks like the horizon line is wrong without volume, because the camera rays start in the incorrect volume, so they do not scatter regardless at which distance they hit the ocean surface.
There might or might not be bugs in the volume system, but from what I can see this scene is not affected.

wasd
Posts: 315
Joined: Tue Apr 24, 2018 7:20 pm

Re: spectral rendering

Post by wasd » Fri Mar 06, 2020 11:05 pm

B.Y.O.B. wrote:
Fri Mar 06, 2020 10:57 pm
To me it looks like the horizon line is wrong without volume, because the camera rays start in the incorrect volume, so they do not scatter regardless at which distance they hit the ocean surface.
There might or might not be bugs in the volume system, but from what I can see this scene is not affected.
This scene, in my opinion is not affected, at least at 1300 samples. Did you have bump on sea enabled while render with disabled camera volume? Because I try to disable both bump and camera volume and didn't see any difference with picture with camera volume and no bump.
CPU Bidir (sometimes GPU Path) + Metropolis | Log Power | Core i5-4570

User avatar
B.Y.O.B.
Developer
Posts: 3753
Joined: Mon Dec 04, 2017 10:08 pm
Location: Germany
Contact:

Re: spectral rendering

Post by B.Y.O.B. » Fri Mar 06, 2020 11:23 pm

Good point, I forgot about the bump. It was not present in the first render, but enabled in the second, so it might be responsible for the change. Will render with auto-detected volume and bump.

User avatar
B.Y.O.B.
Developer
Posts: 3753
Joined: Mon Dec 04, 2017 10:08 pm
Location: Germany
Contact:

Re: spectral rendering

Post by B.Y.O.B. » Sat Mar 07, 2020 12:03 am

With auto-detected volume and bump.
So the darkening of the ocean was caused by the bump mapping, and wether the camera volume is detected automatically or forced to no volume doesn't seem to make a noticeable difference.
It would be interesting to know which volume was chosen by the auto-detection, it could be printed to the console.
Attachments
rend4_autocamvol_23min.png

User avatar
B.Y.O.B.
Developer
Posts: 3753
Joined: Mon Dec 04, 2017 10:08 pm
Location: Germany
Contact:

Re: Atmosphere Simulation

Post by B.Y.O.B. » Sun Mar 08, 2020 6:31 pm

I have disabled scattering on all atmosphere layers, then re-enabled it one by one to find the one responsible for the fireflies, and it looks like the "Glare" layer is the sole culprit.

It seems to be caused by the high Asymmetry: Values higher than about (0.7, 0.7, 0.7) seem to increasingly create fireflies, and the glare layer has asymmetry (0.93, 0.92, 0.92). Can the volume scatter code be optimized to better sample such strong directed scattering?

(4 minute renders on RTX 2080 below)
Attachments
Scattering in glare layer enabled
Scattering in glare layer enabled
Scattering in glare layer disabled
Scattering in glare layer disabled

Fox
Posts: 437
Joined: Sat Mar 31, 2018 11:17 am

Re: Atmosphere Simulation

Post by Fox » Mon Mar 09, 2020 2:40 pm

Maybe glass material with 0 0 0 reflection and shadow color 1 1 1 would be better where null material is used?

wasd
Posts: 315
Joined: Tue Apr 24, 2018 7:20 pm

Re: Atmosphere Simulation

Post by wasd » Tue Mar 10, 2020 9:21 pm

With sun lamp set to "distant", light source no longer visible. Is it possible to make it visible, like "sun"?
CPU Bidir (sometimes GPU Path) + Metropolis | Log Power | Core i5-4570

User avatar
Dade
Developer
Posts: 4667
Joined: Mon Dec 04, 2017 8:36 pm
Location: Italy

Re: Atmosphere Simulation

Post by Dade » Tue Mar 10, 2020 11:18 pm

wasd wrote:
Tue Mar 10, 2020 9:21 pm
With sun lamp set to "distant", light source no longer visible. Is it possible to make it visible, like "sun"?
Not really, with a "distant" light source, all the rays are parallel (so there isn't an emitting spot).
Support LuxCoreRender project with salts and bounties

Post Reply