Vortex woes
#1
hey Tyson,

this is a rather hard to explain issue, so I´m posting it here instead of the facebook group...

I´m working on a blizzard shot, where a storm is raging around a character. As the Camera is orbiting around the character as well, I figured I could use a vortex sim to cover the whole camera path and not just have the wind come from one direction.

But I´m having troubles getting what I want and I´m wondering how you would tackle this, or if I´m overseeing something obvious.

As I also need a phoenix smoke sim, I thought the easiest way would be to use that as base with fluid force and birth fluid.
Unfortunately I couldn´t use tyVortex as a force in Phoenix, so I had to resort to max standard vortex force. I had to constantly birth new smoke, as the smoke would leave the grid after a while and I used open walls.

Maybe I should rethink that approach, the idea was to not have the smoke bunch up on the walls, but the smoke leaving the grid and the constant feeding of smoke are probably the cause of the issues described further down...

Anyway, I got a good enough results, that looked good in the viewport, but I can´t get an exact representation with tyFlow.
In the simplest version, adding Birth Fluid and Fluid Force, I get two issues:

1. As the phoenix grid has open walls, the particles expand to infinity, when they leave the grid.
2. I get "dead" particles, mostly in the middle of the vortex.

As I´m also having a slowmotion retimed part in the middle, in order to avoid any "popups" of newly birthed particles in that part, I wanted to birth all particles at once, instead of birthing them over time, as the smoke sim does.
But that lead to even more "dead" particles.

I´ve started playing with various workarounds, like adding an object bind to keep the particles inside the grid area, deleting them after they leave the area, adding curl force and slow op to the sim etc.

I also tried without any fluid sim, by mimicing the forces with tyvortex/max vortex force, but the vortex either kept expanding or shrinking too much, leading to holes in the middle or clumped particles in the middle.

I´m gonna keep exploring how to tweak it, but I´d love some input, maybe I´m just missing a step here...
  Reply
#2
Quote:Unfortunately I couldn´t use tyVortex as a force in Phoenix

I ran into the same issue myself recently, and implemented a Spacewarp wrapper for tyFlow helper forces, that allow you to input them into PhoenixFD.

I'm not sure when the next build will be ready, but if you're working a tight timeline shoot me an email and I'll send you a new build early so you can use that spacewarp feature.

As for dead particles, why not delete or scale down when particles reach a certain low velocity, so you cleanly clear them out without worrying about them. Once exported to tyCache, it also has options to scale particles up from zero after they are born that will respect retiming options on the tyCache (so they'll scale slower if you slow the cache down)...that may help the birth transitions.

Beyond that, gonna need to see a video example of your effect to really know what the issues are.
  Reply
#3
(02-26-2020, 03:49 PM)tyFlow Wrote:
Quote:Unfortunately I couldn´t use tyVortex as a force in Phoenix

I ran into the same issue myself recently, and implemented a Spacewarp wrapper for tyFlow helper forces, that allow you to input them into PhoenixFD.

haha, thats why we love you. You keep fixing things that bug all of us...

Its for my personal short though, so I´m not in a hurry. Just getting knots in my brain again...
Actually I figured out how to solve part of my problem: I had tried to break up the uniformity of the tyVortex by changing the Multiplier of the velocity Affect in the force op. Once I dialed that back to zero and put vortex and pull force at the same value, it stopped randomly expanding.

I´m still wondering where those "dead" particles come from though, if its missing velocities in the phoenix grid or the way tyFlow interpolates those velocities between grid voxels.
The resolution of the phoenix sim I used to influence the tyFlow particle was relatively low (just around 10 mill voxels), so maybe there is the culprit?

Deleting the "dead" particles is probably a better idea for now than what I´m doing.
Now I´m sending them to a new event with extra forces, but then I have to retime or slow them down them separately as well, as they don´t follow the phoenix retimed sim anymore in that event...

I´ll see that I can step back to my previous attempt and make a preview of the issue.
  Reply
#4
Ok, I think I jumped the gun on this.
I´m pretty sure now that the issue is within the phoenix Sim.
Looks like I have "dead voxels" in there, they were just harder to see than in the tyFlow sim.
So...is the spacewarp wrapper in the next build already...?
  Reply


Forum Jump: