Is it possible to use rigid body and dynamics but control exactly where an object stops and how?

[Adam Trachtenberg]

It’s possible to mix mograph animation with dynamics, but it’s hard to control. I think your best bet would be to cache the dynamics and then animate the visibility of the pieces that you want to animate manually, substituting in non-dynamic objects.

[Jon Simpson]

This helps, thanks.

I’m thinking I should show you what I am trying to do. My client wants me to duplicate this scene, but with various colors and images on the puzzle face:

https://www.pond5.com/stock-footage/708258/puzzle.html

So according to you description (if I am correct), I would start by throwing each piece into the scene with various bounces, then replace them with non-dynamic objects up until the point they land in their final spot – back to dynamic objects for the bounce upon landing in position. Is this correct?

[Jon Simpson]

Or, what if I were to keep the dynamics on all of the pieces individually by mixing 25 scenes to where each piece ends up where it is supposed to from each angle (using the static puzzle as a universal target)? I hope that makes sense, but looking at this over and over (and over and over, lol), It is the only way I can make sense of it working. Additionally, none of the pieces bounce off of each other, a few of them go through eachother in their path, and even the last piece that falls, is sitting on the floor while it is still vertical, NOT on the puzzle piece it is standing above – indicating some scenes mixed, unless there is a way to control rigid body and dynamic attributes individually, object to object?

[Adam Trachtenberg]

I don’t know how the sample video was done, but there’s really no way to tell a simulation where the pieces should end up. I think what you’ll have to do is start with the pieces assembled, bounce them out, and then run it backwards.

[Jon Simpson]

Okay, after looking at this again and again, is it possible to animate each piece separately with a different camera view for each, and then mix them all together to where they end up landing naturally as a whole?

By this I mean letting each piece do what it does naturally when dropped, but then set the camera properly to where that piece sits exactly as it should at a universal angle with the puzzle in tact – then play all of them at their own time.

[Eurika Le roux]

Hi there… my name is Eurika
Did you get an answer on you question for the falling puzzle pieces?
I’ve got something simular with soccer balls falling and I want them to start later on in scene and rolling and stop at certain places and positions?
Please help if you can.
Thank you
Eurika

[Brian Jones]

can’t help really except to say Adam said it. Dynamics is not exactly reality but it’s a pretty good simulation and that means objects go where ‘physics’ says they should. You can blend keyframes and dynamics but you can’t keyframe the final positions and then blend without the objects ‘gliding’ to the keyframes (it doesn’t look like natural dynamics). Same goes for using the Transition setting in the dynamics tag.
Starting in your final position then play it backwards by attracting the objects can maybe work but not if they bounce a bunch (hard to get that to go backwards).
Like the trick shot in pool/billiards you can plan the shot by carefully using other objects to force the bounce in the correct direction (and in 3D these ‘helper objects’ don’t have to be visible).
Or you can cheat and put shots together that don’t actually show the full shot of the ball’s whole path but just enough to imply they are moving the right direction then some end shots to show them ending up where you want.

[Eurika Le roux]

Thx Brian .. ur a star .. tnx for your reply.