Techniques for creating an internal 3d structure of a fluid simulation

[Neilg]

I have been trying to create a semi transparent fluid that contains an internal 3d structure and i'm starting to run out of ideas. I'd like to run over everything i've tried and if anyone knows any workarounds or solutions to getting these methods working (or can offer up a new method) I would appreciate the discussion.
The fluid stays pretty much the same volume outside of being squashed and stretched, no more is added or removed during the sim, and the viscosity is pretty high so it all sticks together.

First I tried particle flow & birthing particles on the surface (planning later to use those to draw splines in between them) but the particles dont read the phoenix sim as an animated object and stay frozen on the first frame.

Next up I tried Joker martinis Labyrinth script - works great, reads the geometry - but hangs on transforming vertices when applied to a phoenix sim. works fine on collapsed meshes...

To try and get the Labyrinth script working I tried making a cloth simulation that wrapped around and encased the fluid sim tightly - then i'd sim that and swap out my fluid sim for that animated cloth mesh. (I figured the consistent vertex count would help) Neither mass fx or standard cloth can recognize that a phoenix simulation is animated, they hold the first frame.

I then tried using environment fog - setting up a procedural network of 3d maps to give me a series of random internal points connected by tendrils. Looked great on my dummy object when setting it up... turns out vray environment fog cant use the phoenix object as a mesh gizmo, so it wont render.

So now I am all out of ideas. I felt like they were all pretty promising but now I have ran out of steam.
Some degree of flickering and instability is fine - I just want to be able to give the fluid some depth and internal detail.

[LarsSonparsson]

What about having Phoenix smoke inside Phoenix liquid?

I mean that you could use that same simulation two times with two overlapping simulators. The other one would be rendered as geometry and the other one renderer as volumetric.


Just a first idea came to my mind. Haven't tested it.

[Svetlin.Nikolov]

I also think this is the way to go - I have a scene where I duplicated the simulator, redirected the second one to use the first one's Input path by using "$(same_as_output) PhoenixFD001" in the second one's Input path, and rendered one in mesh mode, and the second one in Volumetric mode.

[squintnic]

do you have any references of what the structure should look like - could you do a low res phoenix viscous liquid sim then export the particles as PRTs and script splines between each point or mesh them with frost, messing around with zhu bridson thresholds and the amount of particles you mesh? then duplicate that sim an use one as a transperant outer volume?

[Neilg]

do you have any references of what the structure should look like - could you do a low res phoenix viscous liquid sim then export the particles as PRTs and script splines between each point or mesh them with frost, messing around with zhu bridson thresholds and the amount of particles you mesh? then duplicate that sim an use one as a transperant outer volume?

Originally I saw it as being a sack of floating neurons, but given the complexity of making that work i'm flexible on the actual look just to be able to get something kind of otherworldly in. If it doesn't move with the shape and they pop on and off I could make that work too. It's not set in stone - It's for a personal project i've been chipping away at for the last few months - got the storyboard down, a few shots near enough finished, but this particular thing is totally new territory for me. I think the original idea would need houdini to make soft bodies float inside a sack in zero-g.

We dont have frost unfortunately - I was thinking early on it seems like the right kind of tool for this job.
I will try exporting the particles as a prt - that'll get me pretty close to what I was hoping to work with when I started. I also completely forgot I could switch it to render in volumetric mode and regain the ability to use procedural density maps.
I have also considered modeling it by hand and seeing if there's a way to link vertices of the mesh to an ffd to pull the structure along with it.
I'm basically trying to cover every possible method that gets me a result then i'll pick the best looking one to run with.

[Raven]

You Could use a PhoenixFD Force in a pflow system to generate particles in the fluid that are influenced by the fluid.
You could then use Labyrinth on those.