Chocolate + Milk Splash

[Mokiki]

Heyo everyone,

here is a recent vfx shot I did.
Simulated with Houdini H16, Meshed with Frost and rendered with Vray.

https://vimeo.com/226574607

[prateekvishwa]

very nice!

[vlado]

It looks great, but for some reason I'm not sure that it will look like that in real life; I'm not exactly sure why, the liquids seem to behave oddly at the time of the collision. Did you use any references?

Best regards,
Vlado

[Mokiki]

prateekvishwa
Thanks!

vlado
At the time of the project we were shooting a lot of water splashes with a high speed camera (2k fps) but for this cg shot I went the "what it could look like"-route.

Here is a stockfootage clip of something alike: https://www.videoblocks.com/video/co...otion-53s0isg/
There are parts that behave very similar to the simulated fluid but getting anything closer was outside of the time frame and budget.

[vlado]

Wow that reference is beautiful, thanks for posting! Yeah, it looks close enough.

Best regards,
Vlado

[kagemaru]

Very nice!
You got a very decent smooth mesh that's always the issue for this kind of stuff. Can i ask some specs like how many particles, frost algorithm you used, etc? I'm asking because more and more i started dropping frost and using directly houdini where i can use vdb, attributes, etc.

[Mokiki]

kagemaru
Thank you.
In all honesty I had/have problems with frost. There is another really old chocolate test I did with RF on my channel where I didn't manage to get rid of weird reflection issue on the mesh. (Jagged reflection). Luckily this wasn't the case here, at least no real obvious parts.
Also I am pretty new to houdini so I didn't have the time to really dive deep into the vdb mesh creation.

As for the particle count. It was around 15 to 20 million particles. (In total) Zhu/Bridson as meshing method, Radius 0,1cm - Blend radius 2.3 (low density trimming checked) 1.0 / 50,0. Also I had a relax modifier (value: 0,5 / iterations: 10 Keep boundary pts fixed) on top.

Hope I didn't forget anything

[kagemaru]

kagemaru
Also I am pretty new to houdini so I didn't have the time to really dive deep into the vdb mesh creation.
As for the particle count. It was around 15 to 20 million particles. (In total) Zhu/Bridson as meshing method, Radius 0,1cm - Blend radius 2.3 (low density trimming checked) 1.0 / 50,0. Also I had a relax modifier (value: 0,5 / iterations: 10 Keep boundary pts fixed) on top.
Hope I didn't forget anything

Thanks. Sounds pretty standard workflow, but it worked very nicely. I guess the massive amount of particles really helped a lot.
Congrats again!

[jstrob]

Nice! The ref has that unfamiliar cause it's shot with a high fps. It's beautiful! And I understand getting more closer to the ref would be never ending.

[flino2004]

I agreed with Vlado... in my mind, the chocolate would be more dense and heavier than milk but still like the simulation

[AJ Jefferies]

I think this looks great! I agree that the 'tendril' effect when they meet feels a little off for liquids as dense as milk/chocolate but this sh*t is so difficult to get right, I think you've done a brilliant job.

[Mokiki]

Thank you all very much for the nice words.

[Paul Oblomov]

The only problem is to add volumetric blending. For now to be able to match reference with two refractive fluids - you have to fake it

If vlado can point to a right direction of making use of vdb volumetric masks as a blend amount - would be super cool.