Hello people. I imported this FBX character and simulated him falling into this
hammock, however, the animation is not working. How do I make animation work together with dynamics?
Thanks.
Noob's question about animation clip on dop network
2384 5 0- magaodaslanes
- Member
- 3 posts
- Joined: Aug. 2021
- Offline
- kriegmesser74
- Member
- 41 posts
- Joined: May 2021
- Offline
I've never done something like this in Houdini, but what do you mean by "simulated him"? Simulated how? Hand keyed animation? Mocap? Some sort of rigid body simulation? In either case I think (based on my experience in other packages) that you need to do this in layers... i.e. first simulate/animate the object (in this case a man), maybe bake animation on vertex level, and then do the simulation/cloth part with your model as a collision object...
- MilanB
- Member
- 159 posts
- Joined: Nov. 2018
- Offline
kriegmesser74
you need to do this in layers... i.e. first simulate/animate the object....
Hello! Could you please explain a bit about doing it in ''layers''?
Also, what do you mean by baking the animation on vertex level, and what would be the use for that? (is baking process anything similar to this?) It might be just the terminology, but i never came across that.
As for this situation, for example, the way i would do this is, simulate the (man) falling as a collision object in vellum, and let it interact (with gravity maybe) with the blanket thing. I assume this is the same thing you mentioned @kriegmesser74 ?
Edited by MilanB - Aug. 23, 2021 19:53:10
- kriegmesser74
- Member
- 41 posts
- Joined: May 2021
- Offline
By baking animation I was thinking on saving point cache (in 3DS Max) or alembic cache (Maya) or "whatever is called" cache in Houdini... "Baking" of animation to the vertex level of the geometry, i.e. writing positions of vertices per frame or tick, to a file on HDD or SSD... enables you to delete the rig,(optimize scene), or to stop simulating the first object as well as rewinding the animation or rigid body simulation back and forth so you can choose the exact frame when your character touches the canvas... of the trampoline's semi/dynamic cloth rig (no idea how to rig/make that in Houdini, maybe something like this https://www.youtube.com/watch?v=1nSR4LIsPAs), so the character stops being just a bunch of animated vertices, and becomes full fledged collision object with a mass etc. I.E you can choose when one animation/simulation ends - a character... and second one begins... -the trampoline rig animation/simulation... This is what I meant by doing it in layers... maybe events would be a better term... Although the Houdini is really powerful I,(imho), think it is very VERY unlikely you can simulate all at once... maybe you can but it would most likely last for days... I repeat... never done this in Houdini, I'm focused on rigging inside of it... but this is some general direction...
- tamte
- Member
- 8766 posts
- Joined: July 2007
- Offline
you can simulate this at once either using Vellum or Bullet
- if you want to keep your character rigid ragdoll, I'd use bullet
instead of using just a pointcache which is essentially a deforming geo I'd export rigidly transformed objects for each part to that they come in as transformed packed prims already
or you can also extract that using Extract Transform SOP and Transform Pieces
then you can feed them as passive animated packed rbd's and activate when you need it to become fully dynamic
for this you will need to create your hammock cloth rig fully in bullet
- if your character is more like a softbody I'd use Vellum, you can use Tetrahedral Softbodies, pin to target to original anim and also update rest shape every frame unil you want to release it
or you can also treat it as individual pieces with additional constraints
you can also add shapematch constraints to try to keep it rigid if you still want rigid behavior, but want to use Vellum instead of Bullet
and hammock rig in Vellum may be more straightforward to create, even though bullet may run faster for such type of sim
- if you want to keep your character rigid ragdoll, I'd use bullet
instead of using just a pointcache which is essentially a deforming geo I'd export rigidly transformed objects for each part to that they come in as transformed packed prims already
or you can also extract that using Extract Transform SOP and Transform Pieces
then you can feed them as passive animated packed rbd's and activate when you need it to become fully dynamic
for this you will need to create your hammock cloth rig fully in bullet
- if your character is more like a softbody I'd use Vellum, you can use Tetrahedral Softbodies, pin to target to original anim and also update rest shape every frame unil you want to release it
or you can also treat it as individual pieces with additional constraints
you can also add shapematch constraints to try to keep it rigid if you still want rigid behavior, but want to use Vellum instead of Bullet
and hammock rig in Vellum may be more straightforward to create, even though bullet may run faster for such type of sim
Tomas Slancik
FX Supervisor
Method Studios, NY
FX Supervisor
Method Studios, NY
- kriegmesser74
- Member
- 41 posts
- Joined: May 2021
- Offline
-
- Quick Links