On a sort of sad note, I lost half my data on this render, hence the length of the actual thing. But I still wanted to post it because there is still valuable info.
The purpose of this was to show how much you can do with the Thinkbox plugins using limited data. I started with a fairly low-res Realflow Hybrido solve. It's final particle count was close to 20k particles. Because currently Realflow's PRT export is wonky I chose to just import the .bin sequence into Max using a PRT Loader. From here I used Stoke to upres the initial 20k particles to 500k particles. This took less than 10 seconds to run. I wanted to keep everything as close to out of the box as possible, so I selected every available channel from the .bin sequence to influence the Stoke particles.
The next thing was to create foam. I'm not an avid Realflow user, and in the past I remember foam/splash creation to be kind of convoluted and if you weren't careful you could overwrite your data. So because I knew I was going to mesh my particles, I chose to use the exported Stoke particles to be my foam. Once I imported them I used Magma to cull out particles under a certain velocity. Then I rendered them as voxels. It was as simple as that. I ended up just using the alpha channel on a white solid so they would show up better.
Lastly! was meshing with Frost. I chose the Anisotropic method because I love the look. The downside is the increase in mesh creation and rendering. However, meshing per frame was about 10 seconds so it wasn't that bad. Just to be safe I used Xmesh to export the mesh JUST so I have a backup. I used the imported Xmesh as my final rendering geo. Each frame took about an hour, but this was because of Vray, I wanted refraction, so I paid for it in render time.
The 1 thing I like about Frost and Xmesh(Although not revolutionary) is that you don't have to have it visible in the viewport to render a frame. This came in handy in the past for me handing assets to a lighter. I would lock the Xmesh position and then give them a brief overview of Xmesh and Frost so they wouldn't touch sliders.
Again, I lost data on this, there was also a second part planned based off of this example but I'm going to save it for another video. The good part is that if you're reading this far, I'll inform you that the wall of text would have been twice as long.