Test scences GPU Path vs CPU BiDir

Post your tests, experiments and unfinished renderings here.
wasd
Posts: 85
Joined: Tue Apr 24, 2018 7:20 pm

Re: Re:Strange effect using volume, CPU BiDir

Post by wasd » Sat Jul 07, 2018 10:54 pm

FarbigeWelt wrote:
Sat Jul 07, 2018 10:25 pm
Yes, and as you can see, beam above spot of second spot light is not visible.
I think there is a beam. Your dust is just too thin. The reason why you have brighter beam above glass is that there's another beam going up, reflected from glass and wine.
I tried it with laser, it gives more evident beam.
11.png
But suddenly! I replaced dusty air with clear one and now there's some spectacular beam over the glass. The beam that should not be:
12.png
CPU Bidir + Metropolis | Log Power | Core i5-4570

User avatar
B.Y.O.B.
Developer
Posts: 1466
Joined: Mon Dec 04, 2017 10:08 pm
Location: Germany
Contact:

Re: Test scences GPU Path vs CPU BiDir

Post by B.Y.O.B. » Sat Jul 07, 2018 11:25 pm

wasd wrote:
Sat Jul 07, 2018 10:54 pm
The beam that should not be
Did you also change the camera volume?
Support LuxCoreRender project with salts and bounties

wasd
Posts: 85
Joined: Tue Apr 24, 2018 7:20 pm

Re: Test scences GPU Path vs CPU BiDir

Post by wasd » Sat Jul 07, 2018 11:53 pm

B.Y.O.B. wrote:
Sat Jul 07, 2018 11:25 pm
wasd wrote:
Sat Jul 07, 2018 10:54 pm
The beam that should not be
Did you also change the camera volume?
Nope. It was Air and I changed Air volume itself.
CPU Bidir + Metropolis | Log Power | Core i5-4570

wasd
Posts: 85
Joined: Tue Apr 24, 2018 7:20 pm

Re: Test scences GPU Path vs CPU BiDir

Post by wasd » Sun Jul 08, 2018 12:00 am

Maybe I forgot dust somewhere though.
CPU Bidir + Metropolis | Log Power | Core i5-4570

wasd
Posts: 85
Joined: Tue Apr 24, 2018 7:20 pm

Re: Test scences GPU Path vs CPU BiDir

Post by wasd » Sun Jul 08, 2018 12:03 am

wasd wrote:
Sat Jul 07, 2018 11:53 pm
B.Y.O.B. wrote:
Sat Jul 07, 2018 11:25 pm
wasd wrote:
Sat Jul 07, 2018 10:54 pm
The beam that should not be
Did you also change the camera volume?
Nope. It was Air and I changed Air volume itself.
Nevermind that. I experimented with beam disappearence and submerged everything in box of dust, changed external volume to that dust, then removed dust and forgot abot external volume. It is clearly my fault.

UPD: Nope. It was reversed normals on Wine.
CPU Bidir + Metropolis | Log Power | Core i5-4570

wasd
Posts: 85
Joined: Tue Apr 24, 2018 7:20 pm

Re: Re:Strange effect using volume, CPU BiDir

Post by wasd » Sun Jul 08, 2018 3:50 am

FarbigeWelt wrote:
Sat Jul 07, 2018 10:25 pm
Yes, and as you can see, beam above spot of second spot light is not visible.
It seems for me, light passing homogenous, scattering volume keeps scattering outside the volume.
I guess I'm done with my research on that matter. Air in your scene is not so dusty, that's why beam from second spot is barely visible. Beam from spot over wine is apparent because normals of your wine are reversed and (this is my guess) scattering is done with wine volume settings.

With simpler setup (even thickness glass tube) I've got following results:
1) glossy sticks (Sky for lighting)
sticks.png
2) laser beams (world lighting is None)
lasershow2.png
3) glass of wine — it takes a while though for beam to become apparent; 790 samples image
wineglass.png
Would be nice to see B.Y.O.B. comment on that.
Some real life scattering as a bonus:
scat.jpg
Attachments
laser-10.blend
(711.31 KiB) Downloaded 24 times
Glass of Wine-5.blend
(868.11 KiB) Downloaded 30 times
CPU Bidir + Metropolis | Log Power | Core i5-4570

User avatar
Dade
Developer
Posts: 1232
Joined: Mon Dec 04, 2017 8:36 pm

Re: Test scences GPU Path vs CPU BiDir

Post by Dade » Sun Jul 08, 2018 8:54 am

I haven't yet read the discussion but are you aware that SDS (Specular-Diffuse-Specular) paths can not be rendered by a normal BiDir but they require BiDirVM (or something equivalent).

Note: a volume scattering point is considered a Diffuse path vertex (i.e. "D") while a glass intersection point is considered a Specular path vertex (i.e. "S").
Support LuxCoreRender project with salts and bounties

User avatar
B.Y.O.B.
Developer
Posts: 1466
Joined: Mon Dec 04, 2017 10:08 pm
Location: Germany
Contact:

Re: Test scences GPU Path vs CPU BiDir

Post by B.Y.O.B. » Sun Jul 08, 2018 9:23 am

Dade wrote:
Sun Jul 08, 2018 8:54 am
are you aware that SDS (Specular-Diffuse-Specular) paths can not be rendered by a normal BiDir but they require BiDirVM (or something equivalent).
But aren't the laser beams behind glass SDS paths?
In the past I always had the impression that LuxCore Bidir can in fact render SDS paths, however they converge very slowly.

wasd, the normals of the wine were indeed reversed which I did not check, nice find.
With correct normals (pointing outwards) this is what I get:
wine3.jpg
I think you're right that the still weakly visible beam of light we see over the glass coud be the reflection of the spotlight.

Nice render with the green lasers.
Attachments
Glass of Wine 3.blend
(340.47 KiB) Downloaded 11 times
Support LuxCoreRender project with salts and bounties

User avatar
Dade
Developer
Posts: 1232
Joined: Mon Dec 04, 2017 8:36 pm

Re: Test scences GPU Path vs CPU BiDir

Post by Dade » Sun Jul 08, 2018 9:38 am

B.Y.O.B. wrote:
Sun Jul 08, 2018 9:23 am
In the past I always had the impression that LuxCore Bidir can in fact render SDS paths, however they converge very slowly.
Nope, it is like rendering caustics of a point light for a path tracer, it is just not possible (because the point light is a not intersectable light source so it can not be hit by a ray).

BiDir can not connect a light path and eye path when one of the end vertices is a specular event.
Support LuxCoreRender project with salts and bounties

User avatar
FarbigeWelt
Posts: 179
Joined: Sun Jul 01, 2018 12:07 pm
Location: Switzerland
Contact:

Re: Strange effect using volume, CPU BiDir

Post by FarbigeWelt » Sun Jul 08, 2018 10:02 am

Thank you for the analysis and answers wasd, Dade and B.Y.O.B.

Although I checked the normals of the two objects, I made obvisouly a mistake for wine. It is not perfecly clear for me why Blender struggles sometimes with automatic normal setting for objects made by mesh spin. I see, I have to be more careful with direction of normals.

Sometimes it is not clear to me what is meant to exterior or interior, e.g. for a plane a 2D object with no volume.

Using laser is a good input. I should use laser too in 'dubious' cases. Their intense beam is good for tracing effects. At least lasers generate nice picutres.

By the way, congratulations for all coders and testers for the V2.0 speed, stability and very smooth UI integration in Blender.
Laser and Volumes.jpg
Laser, Bragg, Dispersion and Volume
Laser and Prism.jpg
Laser, Prism, Volume Scattering
Microsoft Windows 10 Professional, intel i7 4770K, 32 GB, AMD R9 290x 4 GB, AMD R9 390x 8 GB
Instagrammer, please join #luxcorerender for your renderings.

Post Reply