Smoke ring machine

I made smoke ring machine to test smoke interaction with geometry.

Basically it’s just a tube with a smoke emitter inside and animated cylinder.
Cylinder pushes the smoke out of tube in 5 frames and slowly moves
back to start position - repeating.

Some minor shadow rendering errors. :roll:

http://vimeo.com/18043755

funny :slight_smile:
really there are artifacts near to the tube.

Truly awesome

btw, what kind of advection is used here?

It’s using classic advection and symmetric conservation.

try with slow moving advection, it is designed for such sims.

Yes, I tried that, but didn’t see that much difference for some reason.
I could try it tomorrow anyway just to see the difference.

Today I need to simulate steam train. :slight_smile:
btw, question related to that simulation. I’m getting shadow errors there.
Shadow in the ground is rendered differently in some buckets.
It’s vray sun causing this. Changing the shadow bias doesn’t have any effect to it.
Any ideas?

the slow moving advection is just the classic one with small modification. the difference is clearly visible when the movement is really slow.
about the shadows - usually the bucket related errors are caused by different versions of the software on the render hosts. why you think that the sun is involved?

Hmm. I’m not using dr, only single machine, so there can’t be any different versions.

Ah, I meant that the shadow on a ground is from vray sun, if that matters.

could you post a image with these broken shadows?

There you go.

Buckets are more visible on a right side.

damn! can you do some tests?
i mean to try what happens when rendering with different phoenix/vray versions.

I’ll see what I can do. I could install earlier Phoenix, but I can’t mess our renders with older VRay.
I’m still using 1.5 sp5 on Max 2011 x64

I tried that scene using envFog and it renders just fine.

Same frame using Phoenix 1.1 looks good.

well, there is a bug in 1.2 then

Damn! You need my scene for testing?
It now simulates in 4 minutes to frame 100 where errors are this time really visible.

Well, attaching it anyway.

Good luck!
train_Smoke.zip (29.4 KB)

i think i found it. it affects only the analytic scattering (it is on by default), so if you switch it off the problem should disappear.

yes, it works!

Thanks.