Advected Texture Coordinates from RF + Stoke Spawn + more

After looking through Bobo’s posted Stoke tutorials, I thought that this one could be kind of neat and useful. I posted this image with exaggerated features so you could see what is happening. So the process goes as follows:

  1. I did a generic Realflow solve that were exported directly as a PRT. Just to note, I had a ton of problems when I tried using the PRT exported directly from Realflow. Whenever I did distance tests based on surface distances, the grid would flip every which way and scale randomly. So to get around that I just scaled the flow and position and then cached out a PRT so I could deal with it problem free.

  2. The next think I did was follow Bobo’s tutorial on advecting the texture coordinates. It took me a few look overs to understand what was going on, but I’m like 75-80% there. I multiplied the velocity by 2 in the Stoke in the Field Sim flow so that you could really see what was happening.

  3. After that, this is where it got fun. I took the RF PRT and in Magma did a Logic test based on distance ~> Basically, any particles that collided with the painting were not selected and the rest deleted. The left over particles were used in the Stoke sim to be “birth particles” for the next step.

  4. Wherever those stoke particles were born, I pumped them into a very standard pflow(Krakatoa Birth) with an age of 1 frame and spawn + kill parent. This did a typical “droplet with a streak” flow that had trailing particles. I added a bit of wind to give it some path variation…etc.

  5. I then exported out the pflow setup, making sure to export an age channel. After importing that cache back in, I added a Magma flow on top that would read the color of the nearest surface point of the painting.

  6. I used Frost to mesh the paint particles. I did a “Shrink over Age” of 15. Lastly, for the color I used a vertex color in the diffuse so that I would get the individual particle colors read into the Frost mesh.

Lastly, I gave the paint mesh a shinier material so that it would look more “wet”. I know a lot of the steps I took could be done right in pflow, but honestly, I’d rather just cache it all out and not deal with it. In the attached image I didn’t at the Realflow splash that I did. When my machine is free’d up I’m going to render the 200 frames out so you could better see what is happening. I think this is really neat for A LOT of purposes. You could get accurate things that rinse off, especially because you can produce wet maps now. I wanted to do that with this, but I didn’t want to waste all of the fun.

I hope you guys find this interesting.

thanks! cant wait to see the animation…

btw, i think realflow has a PRT export bug…bobo? have we told them? i can follow up

cb

Cool Mr. Stark :slight_smile:.

Yes, we told them.
On the one hand, the channels are named incorrectly, e.g. “ID” is called “id” and thus does not work in our tools.
And RF still writes v1.0 particles, so scaling and transforms don’t come over. Once v1.1 support is added to RF, it will get easier.

i followed up with NextLimit again today

cb

Yep, these were the problems I ran into immediately. I thought I could get around it by switching the transform axis’. That worked until I tried doing anything with the particles; then everything went haywire.

Just use the BIN files for now :slight_smile:
You can resave them to PRTs with a PRT Loader and Krakatoa if you want to save space.

That’s exactly what I did. I positioned it and scaled it then exported to a PRT. It took me 30 minutes until I thought “duh”.