Tweet
Hi there,
I´m planning on improving my underwater-particles-workflow for a short I´m working on.
There will be several underwater shots. I´ve done quite a few medical animations and I´ve often used a very simple setup with just spawning a couple of thousands of small particles in a defined area, with those particles being a group instance of various modeled little objects resembling sand particles or plant fibres, and then just adding some random movement with random walk or a wind force with turbulence.
This works ok, but is far from the look I´m going for (For reference, this is what it SHOULD look like: https://youtu.be/cjV2WP0wkvo?t=53s):
A) There is no reaction with animated objects in the scene with my setup.
B) The amount and look just doesn´t feel dense enough with only 10-20.000 Particles of roughly the same size.
C) I need a setup that works for close up shots with little camera movement as well as for shots where the camera travels some distance, or where the camera isn´t so close, that the BG is hidden in DoF anyways.
For Part A I´m thinking about using Phoenix FD with the Phoenix Force op, to get a more fluid like look to the particles and not just some random movement, as well as reaction to other objects in the shot.
For Part B I´m going to try and use several Particle Flows or Events (depending on the final particle count and responsiveness in the viewport): One for the "bigger" particles that actually look like "stuff" (small sand particles, plantfibres etc.), one for smaller particles that are either just very small lowpoly spheres or boxes to catch highlights, and maybe in addition some drag particles from Phoenix FD, rendered with the pointshader.
I´m not sure about the setup I need to make this work with phoenix though.
Here is what I was thinking, please comment on anything that might not work or could be improved:
1. For the background I´ll be using the same setup I´ve used before, with the instanced particles and some random movement.
2. For the foreground, I´ll create a phoenix grid covering my hero objects and fill it with liquid (can´t remember right now, if there is an initial fillup option or if I have to use an emitter object and "imprint" it with an animated discharge).
3. Simulate the whole length of the shot. Make sure to export all necessary channels (please fill in the blank I´m drawing right now....
4. Create a particle flow covering the same area and birth smaller instanced particles like in Step 1.
5. Add a Phoenix Force and voila: I should have particles that realistically move around as if they were submerged under water and react to moving objects...
6. Maybe also emit drag particles and render them using the point shader to further fill up the water with movement...not sure If that would create particles through the whole liquid container or just on its surface...?
7. For the shots where the camera travels a longer distance I´ll probably be fine with my original setup and maybe just modifiy it a little, so I don´t have to fill up the whole camera path with thousands of particles, but rather spawn them only a couple of frames (maybe 50 or so) before the camera moves through and delete them after the camera has moved past them.
I´m not quite sure how I´ll be able to do that, but thats not really a Phoenix issue...I was thinking of cloning the animated camera, linking an emitter object to it and then just moving the keyframes forward.
I could probably also link a phoenix grid to the original camera, check inertial forces and then repeat the same process as in the previous steps 2-7, I´m not sure if that would work though.
There are some more questions regarding this setup:
For step 3: Would it be enough to just have the animated objects moving through the grid to get some nice turbulence going on? Or would I also have to creat some basic movement in the container, and if so: how would I best do that?
Should I use a Plainforce? Or slowly emit liquid into the cotainer from one or more sources?
What difference would my cell count make on the simulation? I´m guessing I´ll need a certain resolution for the simulation to correctly react to the animated objects?
Or coula I just simulate the grid with a pretty low resolution and still get a usable phoenix force op for my pflow setup?
The boundaries of my grid should be open, right? How does that affect my phoenix force op? Would particles in my pflow setup bounce of the grid boundaries, simply dissapear or just continue with there movement once they reach the simulated boundaries?
As soon as I´ll have some time to do some R&D, those questions might just answer themselves, but I´d like to get a head start and start some simulations in my head until then...
I´m planning on improving my underwater-particles-workflow for a short I´m working on.
There will be several underwater shots. I´ve done quite a few medical animations and I´ve often used a very simple setup with just spawning a couple of thousands of small particles in a defined area, with those particles being a group instance of various modeled little objects resembling sand particles or plant fibres, and then just adding some random movement with random walk or a wind force with turbulence.
This works ok, but is far from the look I´m going for (For reference, this is what it SHOULD look like: https://youtu.be/cjV2WP0wkvo?t=53s):
A) There is no reaction with animated objects in the scene with my setup.
B) The amount and look just doesn´t feel dense enough with only 10-20.000 Particles of roughly the same size.
C) I need a setup that works for close up shots with little camera movement as well as for shots where the camera travels some distance, or where the camera isn´t so close, that the BG is hidden in DoF anyways.
For Part A I´m thinking about using Phoenix FD with the Phoenix Force op, to get a more fluid like look to the particles and not just some random movement, as well as reaction to other objects in the shot.
For Part B I´m going to try and use several Particle Flows or Events (depending on the final particle count and responsiveness in the viewport): One for the "bigger" particles that actually look like "stuff" (small sand particles, plantfibres etc.), one for smaller particles that are either just very small lowpoly spheres or boxes to catch highlights, and maybe in addition some drag particles from Phoenix FD, rendered with the pointshader.
I´m not sure about the setup I need to make this work with phoenix though.
Here is what I was thinking, please comment on anything that might not work or could be improved:
1. For the background I´ll be using the same setup I´ve used before, with the instanced particles and some random movement.
2. For the foreground, I´ll create a phoenix grid covering my hero objects and fill it with liquid (can´t remember right now, if there is an initial fillup option or if I have to use an emitter object and "imprint" it with an animated discharge).
3. Simulate the whole length of the shot. Make sure to export all necessary channels (please fill in the blank I´m drawing right now....
4. Create a particle flow covering the same area and birth smaller instanced particles like in Step 1.
5. Add a Phoenix Force and voila: I should have particles that realistically move around as if they were submerged under water and react to moving objects...
6. Maybe also emit drag particles and render them using the point shader to further fill up the water with movement...not sure If that would create particles through the whole liquid container or just on its surface...?
7. For the shots where the camera travels a longer distance I´ll probably be fine with my original setup and maybe just modifiy it a little, so I don´t have to fill up the whole camera path with thousands of particles, but rather spawn them only a couple of frames (maybe 50 or so) before the camera moves through and delete them after the camera has moved past them.
I´m not quite sure how I´ll be able to do that, but thats not really a Phoenix issue...I was thinking of cloning the animated camera, linking an emitter object to it and then just moving the keyframes forward.
I could probably also link a phoenix grid to the original camera, check inertial forces and then repeat the same process as in the previous steps 2-7, I´m not sure if that would work though.
There are some more questions regarding this setup:
For step 3: Would it be enough to just have the animated objects moving through the grid to get some nice turbulence going on? Or would I also have to creat some basic movement in the container, and if so: how would I best do that?
Should I use a Plainforce? Or slowly emit liquid into the cotainer from one or more sources?
What difference would my cell count make on the simulation? I´m guessing I´ll need a certain resolution for the simulation to correctly react to the animated objects?
Or coula I just simulate the grid with a pretty low resolution and still get a usable phoenix force op for my pflow setup?
The boundaries of my grid should be open, right? How does that affect my phoenix force op? Would particles in my pflow setup bounce of the grid boundaries, simply dissapear or just continue with there movement once they reach the simulated boundaries?
As soon as I´ll have some time to do some R&D, those questions might just answer themselves, but I´d like to get a head start and start some simulations in my head until then...
).
Comment