Filling cup with flip using vdb collision update..

[nigelgardiner]

Hi There

 

Sorry last post didn't include the glass in the hip file. now included I hope.

Reducing the voxel size for the vdb seems to work but I still get some fluid popping out of the original glass geometry.

So I guess I need some advice on how to get this working correctly.

 

Again thanks.

Nigel.

glass liquid v5 with glass.hipnc

[paulh]

Reduce the voxel size on your glass vdb and turn on Collision Separation on your FLIP object. Set the Collision Separation to 0.5x your Particle Separation and you should be good to go.

 

Also, turn on the Collision Guide in your Flip Object to get a better sense of how your collision geo is working.

glass liquid v5.1 with glass.hipnc

[nigelgardiner]

Hi Paul

 

Thanks, that keeps the particles inside the collision object but when I turn on the fluid surface that is breaking thru the collision geometry?

Any idea how to handle that? 

 

thanks

Nigel.

[paulh]

Hi Nigel,

 

That can be fixed by adjusting values on your Particle Fluid Surface node. Take down the Voxel Scale and Droplet Scale values and you should see the surface more closely align to your fluid particles. Hope that helps.

[nigelgardiner]

Hi Paul

 

Thanks, that makes sense.

Tho that probably won't guarantee the fluid surface is sitting on the inside surface of the glass perfectly.

Not that I know much about it, but there must be a relatively simple way of saying if the fluid is outside of the inside wall of the glass get rid of it which would then leave an accurate fluid surface against the inner of the glass.

 

cheers

Nigel.

[Hamp]

Try a VDB boolean before meshing maybe?

Subtract whatever penetrates your glass could be an idea..

[fathom]

how will you render this?  even if you have a perfectly aligned fill, you're unlikely to have enough precision to render using only two different materials (glass and fluid).  you'll probably need a third (the glass-to-fluid surface).  i would suggest you come up with a means to utilize the interior of your glass that is contact with your fluid as a separate surface.  throw out the glass and fluid in these areas so you have a single boundary (glass-to-fluid) instead of two (glass-to-air and air-to-fluid).

 

this might be doable using refraction masks and a couple extra phantoms of your glass and fluid...

[nigelgardiner]

rendering was/is my next issue, the glass to fluid boundary is something I've been thinking about a lot.

i photograph a lot of liquid in glass type scenarios so it's all very intriguing.

suspect i won't get it working in a hurry given my current houdini skills.

[Juraj]

how will you render this?  even if you have a perfectly aligned fill, you're unlikely to have enough precision to render using only two different materials (glass and fluid).  you'll probably need a third (the glass-to-fluid surface).  i would suggest you come up with a means to utilize the interior of your glass that is contact with your fluid as a separate surface.  throw out the glass and fluid in these areas so you have a single boundary (glass-to-fluid) instead of two (glass-to-air and air-to-fluid).

 

this might be doable using refraction masks and a couple extra phantoms of your glass and fluid...

 

Good solution for this problem is using nested materials. It enables to set priority to different shaders and it will nicely fix problem with overlapping areas. However I haven't seen this feature in Mantra and I am not skilled enough to do that.

But there is jf_nested_dielectric shader for Arnold which can do it. And also Maxwell renderer has this feature.

It could be nice if somebody tried to create nested shader for Mantra.

 

EDIT: Just found something for Mantra.

Edited January 11, 2016 by Juraj Tomori

[nigelgardiner]

Hi Juraj

 

Thanks very much for the advice much appreciated, will look into nested materials now. 

 

cheers

Nigel

[nigelgardiner]

Hi There

 

I'm a long way from thinking about the rendering aspect of this as I'm still trying to get my head around houdini and vdb. Tho I have looked at the physhader for mantra and that looked promising.

I am still after playing around for ages having problems with particles not colliding with my vdb collision field and if i turn on the collision guide in for the static object Houdini crashes so I guess I've messed up something fundamental.

 

I've looked at the example in this thread    https://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&t=37760&view=previous&sid=afb6f5e148c82c67ce6807adba179ed1    

 

and can't really see what I'm doing differently, would someone please be able to enlighten me?

 

Thanks very much

Nigel.

glass liquid v13.hipnc

[Juraj]

Hi,

I checked your file and it crashed my computer twice before I could inspect it. You had generated there some huge resolution VDB volume. I recreated DOP setup again with shelf tools and now it works like expected.

 

glass liquid v13_fix.hipnc

[nigelgardiner]

Hi Juraj

 

Thanks. I tried to create the setup as manually as I could to try and figure out how it all fits together. Guess I had something not quite right.

The huge vdb resolution volume, yeah I couldn't understand why there was no collision so tried increasing resolution.

So much to learn.

[nigelgardiner]

Hi Juraj

 

I looked at your file setup, I see you or the shelf tools reverted to the default fluid surface setup. I've been trying to make that interact nicely with a contained surface for a while and I don't think it will work well enough.

I replaced it with a convert fluid particle to vdb volume  node and a convert volume to polygon soup node and it seems to work better (which is what I've been toying with). I've got to do some paid work now so attached is just a rough patch but to me it looks more promising tho I could be wrong.  To me it seems there's more chance of getting the collision surface and the fluid surface to sit together accurately, perhaps not exactly like this is set up but with some sort of vdb node arrangement.

 

 

thanks

Nigel.

glass liquid v14_fix.hipnc