[Fisio SB addOn released] Fisio: Physics Simulator
Alberto
Posts: 1,436
Fisio: Physics Simulator is a fast and easy-to-use physics simulator for Daz Studio based on the Project Chrono library.
Fisio simulates the motion of solid objects, referred to as bodies. The main features of Fisio are:
- Any close-geometry prop or figure (including instances) can participate in the simulation as a body.
- There is no limit to the number of bodies in simulation, other than the computer's memory.
- The bodies can be subjected to animatable gravity, damping, point, directional, torque, oscillatory, or wind-type forces.
- Some bodies can be manually animated, and nonetheless, participate in the simulation, by colliding or linking to other bodies.
- The motion of the bodies can be constrained by using links between them.
- Like in the natural world, a body can be defined by its material and its shape.
- A material is defined, like natural materials, by its physical properties such as density, bounciness, etc., and it can be shared by many bodies.
- The shape is the specific geometry of the body. The user can select a simplified shape or the actual mesh of the body, or let the plugin determine one.
- Fisio is extensible with future add-ons.
This is the core Fisio plugin, and it is dedicated only to rigid-body physics simulation. An add-on will deal with soft bodies (flexible or deformable bodies).
popup_01.jpg
1300 x 1000 - 188K
popup_03.jpg
1300 x 1000 - 355K
popup_04.jpg
1000 x 1300 - 190K
popup_06.jpg
1000 x 1300 - 140K
Post edited by Alberto on
Comments
This will bring DAZ into a whole new realm of 3D and CGI creation. Up there with the professional animation tools. I am looking forward to being able to do in DAZ what (up till now) I needed software costing several thousand dollars to do.
Nice video! Skimmed quick this time. Simulations for rigid bodies looked promising.
In OP you mention add-on for soft body physics. Will this be your own creation? Or is there a product it is based on?
Just wanted to say Happy to see this initiative.
Am also curious if an option to save all properties will be possible.
Looking forward to this!
Thank you!
Fisio and its future add-ons are based in this library: http://projectchrono.org/.
All properties can be saved.
Will it be able to interact with dForce (e.g. drop an object onto a trampoline and have the trampoline deform and then bounce the object back up)?
Yes!
Yes, but as the calculations of dForce and of Fisio cannot be done simultaneously, the process has to be iterative: run Fisio, then dForce, then Fisio, then dForce, and so on, until the solution get stabilized.
The future Fisio's add-on for soft bodies, however, will do the process in a simple run.
Thanks, I was expecting a simple no so that is still good even if it is a bit fiddly.
Great!
Do you implement the fluid model, too?
WOW...Instabuy
Sounds intriguing, something we've veen waiting for!
No. Let the fluid to FLUIDOS
You're welcome!
Your answers to Richard and tfis really highlight why these shouldn't be PA products, but native functions of DS.
can we have this NOW or do we have to wait until DAZ soon ?
I think it's a regular soon (not DAZ soon )
This is a video of tests by Imago, the beta-tester.
Notice the earrings at about 0:32 sec. Note: As the figure itself isn't taking part in the simulation in this quick test, you can see some overlapping, It can easily be fixed by including the figure or a proxy of it.
So nice to see fragile boxes drop and earrings do a physics dance. Thanks to you & Imago.
Some movements look close enough for me, some look good, some more stylistic and not quite close enough (house rocks).
How much would it slow down the earring movement calculations when including a collision fig or proxy mesh, which should be denser and more complex in shape? Mention cos DS is so snappy with less nodes/active morphs in scene. So wonder when this impacts physics calcs.
Also am curious. The way your plans are shaping up, will the soft-body add-on Only work with dForce? Or is working with dF just another option for vertex deforms? For whatever it is worth, myself, am not a fan of idea of a 3rd party physics product being dependant on dF and also its future dev unknowns .
You're welcome!
Some of them (like the rocks) are quick tests. They can be refined by adjusting the parameters (e.g. the density of the objects) and scaling of simulation.
Yes, the number of polygons in a mesh is the main variable in slowing the calculations. But there are alternatives. You can select the base resolution or use a decimated mesh in the calculations, or force the plugin to use simple predefined meshes. In any case, the mesh used for calculations doesn't need to be the same for rendering (that is, for the rigid bodies). The plugin generates keyframes of positions and rotations of the bodies. This can be copied from a low-poly body to a high-poly equivalent body.
Like the core plugin, the Fisio add-ons will be completely independent of dForce, as they will be based on the Project Chrono engine (https://projectchrono.org) as well.
I second this!
As Alberto said, those are quick tests, the results are way better with more precise settings. If you got questions, just ask.
Can I define an simulation range or do i have to simulate the whole timeline, like in fluidos?
Yes, you can define the range of simulation.
Finally!!!!!!!!!!!!!!!!!
Thanks Alberto & Imago for info.
Will look forward to more videos. Same time know how much time/work they can be. So won't expect too many.
So totally get video sections were variable as a result. Just nice to hear confirmed.
Thank you especially for clarifying that this physics product is designed to be stand-alone.
Truthfully was a bit worried when I saw your reply about Trampoline & soft body deforms; worried it was maybe built as an interplay with dForce only or heavily.
Have a few more questions if you have time?
re: mesh physics calcs
Can any calculations be parallelized (avoid animations issue of limited in-large to single core)?
Can proxy physics meshes have another layer of weight painting and factored in for variable strength responses?
BIG question: will soft body physics work fully & properly with geografts (allowing denser mesh and more detailed deforms)? Initial thought is should work, having worked with geografts alot. But don't want to assume.
Sounds dumb maybe. But was so nice to see damping again in 3d.
Will there be an option for x, y, z spin/twist deforms as well? (sorry didn't look at Chrono engine yet & scanned video-1 a bit quicker - less time atm for proper deep-dive).
Your're welcome!
I have more time than solid answers for the time being because you're asking about the soft-bodies add-on, which isn't finished yet.
The collision calculations step (the slower one) is multithreaded for the present Fiso core (hard bodies physics). The soft-bodies add-on's calculations will be even further multithreaded.
This isn't defined yet.
It should work.
Damping is already available in the Fisio core.
It isn't defined by now, but the deforms will follow the physics laws and the chosen material properties.
When I have more advanced the soft-bodies add-on, I could tell you more.
It's out, but currently of no use, as DAZ has not generated the serial number (again).