BaVolumeCloud intersecting volume problems

[chaney]

I am getting some cubicle artifacts when two or more BaVolumeClouds intersect in Maya .

The density is being driven by particles from Maya and the cubes are much larger than the cell size . Part of the problem seems to be shadow artifacts but most of it seems to come from an additive comp .

Check out these images for details :

http://www.superchrishaney.com/images/volumeCloudIntersectProblem.jpg
http://www.superchrishaney.com/images/volumeCloudIntersectProblem2.jpg
http://www.superchrishaney.com/images/volumeCloudIntersectProblem3.jpg

Any advice would be greatly appreciated .

[schoenberger]

If multiple volumes intersect, then Mental Ray calls first one volume, then the other. For each image tile mental Ray renders it decides differentely which volume is rendered first.

After both volumes are rendered, it is only possible to choose a compositing mode.
Just to check: are you using the an older version? Then you have to enable the checkbox "support for intersecting volumes" to enable the compositing.

But this is not real 3d volume intersection. The only way to get real 3d volume intersection is to use one single volume shader.
_________________
Holger Schönberger
Binary Alchemy - digital materialization

[chaney]

We are using the BA_particle_density3 and BA_volume_cloud from the 3.6 package for 64bit Linux and I do not believe there is a "support for intersecting volumes" checkbox on either of those nodes . There is that option with the BA_volume_solid.

I will install the 3.6.2 package as that seems to be the latest .

Thank You ,
Chris

[goshone]

Is there any support for this now? Trying to render multiple overlapping volumes (3) in the same scene and they are flickering.
Using Maya 2011 on Red Hat 64bit Linux and BA_Volume_Cloud v3.7.1.

[schoenberger]

No, there is no fix for this issue.
The only possibility would be to combine multiple particle density nodes into one volume shader.
I have not tested it in Maya.
You need two pt_loader and a slef-made BBox object that sourrounds both systems.
Then apply to the BBox the volume shader as usual. In the particle density, pick the pt_loader geo shader.
Set visibility attributes the geo shader object to no prim rays, no sec rays, no trans, no refl, no shadows.
So it gets exported, but it not rendered.
_________________
Holger Schönberger
Binary Alchemy - digital materialization

[schoenberger]

An other idea:
I remember someone who has joined two Maya cache files into one.
Perhaps you can set different userA variables for both clouds.
Then you can apply the volume as usual and drive the "Use particle" switch in the particle_density node.
Take a look at XSI, sample scene rendering/ICE_fireworks.
_________________
Holger Schönberger
Binary Alchemy - digital materialization

[goshone]

[schoenberger]

If you use the same cache file, why not using two particle_density nodes, add color and density together and use one volume.
_________________
Holger Schönberger
Binary Alchemy - digital materialization

[goshone]

I am now, but that method doesn't allow me to set contrast or cell size separately. Basically, I would like to modify attrs found on the BA_volume_cloud node, but using the same sim cache. As far as I know, that isn't possible unless using multiple volume_cloud nodes.

[goshone]

This is what I imagine with multiple volumes



Red = mainly fire looking volume - this may be a different sim altogether
Green = relatively high density and contrast. Small step size to retain detail
Blue = larger step size for efficiency, low density and contrast

Is this possible with one volume (i.e. one BA_volume_cloud)?

Also, is this (intersecting volumes) a limitation of mental ray or the shader?

[schoenberger]

Hi

The cell size can not vary, but the marching steps can. If you enable the manual mode for the marching.
Then you can set adaptive subdivisions. (You can view the level of subdiv them in preview mode).

Multiple intersecting volumes is not easy to implement. The renderer does not know enough about the volume shader to support it and the shader does not know enough about other shaders in the section.
It would probably be possible with an additional shader that combines multiple volumes.
But then the cell size could differ, but the marching steps will be the smaller ones.


Workaround for now could be a comp solution.
e.g. particle life is 100 frames.
You need high detail until 40.

Layer1:
Particle life 55, high detail

Layer2:
Particle life 100, low detail

Layer 3:
Particle life 100, low detail
BW-Gradient as a mask from frame 40-50.
_________________
Holger Schönberger
Binary Alchemy - digital materialization

[goshone]

Hi Holger,
Thanks for the reply. This is how I would do it in the past if needed, which works most of the time. I am still trying to wrangle it all in one volume for the time being, and will try to post if I achieve a breakthru.
cheers
Josh

[schoenberger]

I have also logged this issue into the feature requests for the next version.
_________________
Holger Schönberger
Binary Alchemy - digital materialization

[danyargici]

Hi,

Are overlapping volume shaders still an issue for MR / BA_Volume?

I could really do with separating my volume settings for different elements of my cloud (that I've merged in order to avoid this old problem). I already have it split up with Multiple Particle Densities and I'm very familiar with this workflow, however I'm forced to use really inappropriate settings for whole cloud because I need them for a small part of it.

I did a quick test using XSI's built-in Volume shader and didn't see anything too nasty, but due to it's sporadic nature, it's hard to be sure and I've been burnt badly in the past when trying this on a job (which is when I first became aware of the issue!)

Failing all that, I don't suppose there is a way to 'blackhole' or holdout certain parts of a volume? At least that way I could try and render in two approximate layers and fudge something back together... Having said that, I definitely still need any holdout portions of the cloud to still cast shadows...

Argh, man do I hate MR!

Thanks for any tips,

DAN
btw I'm on 2012SP1 - in case the version of Mental Ray might have any influence on you answers...

[danyargici]

Softimage 2012SP1 I should say. In case it's not clear.

[schoenberger]

Yes, the Problem still exists in Mental Ray.
There is always one cloud rendered in front of an other cloud.

You can render a black/white mask for a cloud. But mising intersecting clouds with an alpha channel is not possible. You can place an Volume/Object with Alpha channel in front of a second one, but for mixing volumes, you would need a deep image.
_________________
Holger Schönberger
Binary Alchemy - digital materialization

[danyargici]

OK, thanks for the clarification Holger.

It would be awesome to be able to render this particle-density based stuff in Arnold. Being back in MR land is making me squirm to say the least...

[schoenberger]

After the fluid shader is done I will create a particle shader.
But the volume shader in Arnold is not the baVolume shader used in MRay. So there are not speedups.
_________________
Holger Schönberger
Binary Alchemy - digital materialization

[danyargici]

No speedups I'm sure, but I'd take noise over flicker every time. Presuming that would be the case of course...