Copying velocity particle attributes

[phong]

Hi!

In the Fluid Technical Evening video with Jeff (http://www.sidefx.com/index.php?option=com_content&task=view&id=1490&Itemid=310) he demonstrated a particle fluid sim rock splash at the beginning. He used two-passes to generate the highres splash and mentioned, that he transfered the velocity from the lowres fluid sim to the splash sim. How can I transfer the velocity attribute without overwriting the particles own velocity then? Or in other words, how did he do it?

Thanks a lot,

Phong

[graham]

It's been a long time since I did those files for Jeff but I'm pretty sure the velocity for the new points was just the average velocity for all the points that got grouped out, though I can't be certain. If you use something like an attribute promote to get the average, scale the values up a bunch to get a bigger splash, then apply them as the V attr to the high res points before creating the new particle fluid simulation. Basically the key is to just gives the points a large velocity towards your splash object and let it sim from there. Sorry I can't be of much more help since it was so long ago.

[phong]

Ah, so you generated a kind of iso volume out of the separated lowres collision particles, on which you applied a scatter with the points having the same velocity attribute and emitted particles from those?

[Netvudu]

Phong, I don´t know if Graham did that, but I have done this myself before and it worked pretty well. You should try that method.

You can always add some more velocity or variance for artistic purposes if the original velocity doesn´t perform dramatically enough.

Edited February 16, 2010 by Netvudu

[graham]

We created a volume and then created points inside, not on the surface of. These points most likely got a scaled up average velocity so they had roughly the same motion but exaggerated. These points were then fed to a new simulation that used them as the initial state. There was no emitting done. The large number of new particles with the increased velocity was what would be used for the splash.

[phong]

interesting. But what is if it takes a while until the wave reaches the collision object? Then the points of the new simulation would dissipate in the lowres particle sim or not? Or are you using a cached version of the lowres sim and do an independent splash simulation? if so, how do you get the splash interact with the lowres sim then (like hitting the surface and dissipate)?

@Netvudu: I tried that method and it did work pretty well, but I had to manually animate when particle emission begins and when it ends. It would be cooler, if that is all done procedurally (i.e. if you have tons of rocks and stuff in the water with which the wave can collide!). How cool would that be?!