I came across this [80.lv] article on lvl80 regarding Ben Watts' foam simulation. I tried to piece together his approach from what little Information is provided by the article.



I have no trouble setting up a decent flip-sim, but I'm a little stuck creating the bubbles.

I use a group expression node to select about 5% of all particles and randomize their pscale. Doing this however leads to the issue of bubbles intersecting the collision geo.

Watts writes: "The foam bubbles are just a subset of the original particles. I had to make sure the bubbles didn’t protrude too far out of the foam, or else they’d end up inside the collider walls, so some compression parameters were added based on the main fluid’s gradient."

Does this maybe mean, he only selects particles to be possible bubbles, that are far enough from the collision geo, so as not to intersect the collision geo?

The phrase "...some compression parameters were added based on the main fluid’s gradient.", confuses me, and I'm not sure what he means.

Has anyone suggestions on how this might be achieved or has anyone recreated this kind of setup before?
I'd appreciate any help.
Cheers!

I would love to know more about this as well.