Another cornellbox

Olis · Post by **Olis** » Tue May 01, 2007 8:24 am

My first ever cornellbox test. The fluid is baked with resolution 100, subsurfed one time and normal smoothed. It is rendered for a little more than 12 hours.

CTZn · Post by **CTZn** » Tue May 01, 2007 9:03 am

Hi Olis,

Not sure I'm correct but that looks like a simple surface, if that's the case extrude the borber edges vertically through the floor and close the bottom. You may need to separate the sim surface from the downside box (of watter) in order to preserve normals, smooth for the top and hard for the "body" of the volume. Get me ? Or I'm wrong, you already did a volume out of it...

Olis · Post by **Olis** » Tue May 01, 2007 10:09 am

It is actually a volume. But I agree with you that the caustics doesn't look right... Maybe it is too much IOR. It is set to 1.33 at the moment.

alex22 · Post by **alex22** » Tue May 01, 2007 10:24 am

It would be nice with a slight sss, so that you have underwater volumetric caustics.

Olis · Post by **Olis** » Tue May 01, 2007 10:32 am

alex22 wrote:It would be nice with a slight sss, so that you have underwater volumetric caustics.

Good idea!

Kram1032 · Post by **Kram1032** » Tue May 01, 2007 10:35 am

1.33 is ok...
how's the scale?
this drops aapear quite small, but it's 2*2*2m big!

look in your liquid-simulation-settings: somewhere, there is a "scale" part... try to scale the whole scene, so that it fits

manitwo · Post by **manitwo** » Tue May 01, 2007 11:01 am

quoting myself:

the smaller the lightsource(s) the sharper the caustics.

CTZn · Post by **CTZn** » Tue May 01, 2007 3:31 pm

It is actually a volume.

Oh, ok. However in the image the front part looks clipped, maybe it's an optical illusion... In fact I have not the sensation that refraction is occuring beneath the perturbed surface, maybe you could do a test by adding an oblical stick through the water down to the floor...

Dunno, I'm not convinced for now, I need to see that stick

But still, the sim is good and the image quite clean

Ever tried Cauchy b coeff (0.003 physically accepted value) ?

Kram1032 · Post by **Kram1032** » Tue May 01, 2007 9:58 pm

yeah, sun-caustics just are such sharp, 'cause the light nearly comes stright down.

Olis · Post by **Olis** » Tue May 01, 2007 10:16 pm

After changing the simulation scale in the domain options, everything is much slower. The simulation itself does not take longer to compute it is just that the fluid is heavier. The image in the first post had the simulation scale set to 0.03 which is the default. The frame I've rendered is the second one in the simulation. When the simulation scale set to 2.00, it takes much longer for the icosphere of liquid to fall down. I baked this one with the resolution 150 and this is what it looks like after 30 frames:

I think I have to bake it at least 60 frames so that it has enough time to get to the splash, because it is what I want to render.

Kram1032 · Post by **Kram1032** » Tue May 01, 2007 10:49 pm

you shouldn't reduce THAT size, but you should look, how big it is, and then SCALE DOWN THE SCENE (or just use the resizer: set it to 0.03

)

Olis · Post by **Olis** » Wed May 02, 2007 12:29 am

Kram1032 wrote:
you shouldn't reduce THAT size, but you should look, how big it is, and then SCALE DOWN THE SCENE (or just use the resizer: set it to 0.03 )

What is the advantage of that?

Kram1032 · Post by **Kram1032** » Wed May 02, 2007 12:52 am

the thickness of the water has the correct scale....
also, the lightsource is MUCH smaller...