The ability to disable the physics on unions

Feature Requests Studio Features
#1

So, as some of you may know, the models I make are mostly unions.
and although the new physics are great for certain parts, other parts don’t need the specialized physics, but would instead benefit by being calculated like a part with a mesh on it.

For example:
My M4A4 Sherman Firefly

Union count:66
Part count: 61

Now, It runs okay. I could probably run 4-6 before stuff started to slow down (on my machine)

But some of these parts don’t need the specialized physics (which I have marked by changing their materials to neon)

[spoiler]

3cef0d46a536cf4a99b61cab43d349c1.png695×351

(tracks are of course would be mirrored, I just forgot to change their material)[/spoiler]
after going over the model, I have determined that 30 parts don’t need these specialized physics.

Quite simply, A way to either Disable the physics of some parts entirely, or a way to disable the special physics of union parts is a rather needed feature, that would allow me and other users to do more withing the limits of ROBLOX.

True Custom Decomposition Geometry
#2

Yeah, unions are intensive on performance. We really need to be able to disable collisions completely, so the .Touched event doesn’t even fire.

#3

An option like this would help me a lot with the big game I’m currently working on

#4

Past staff responses have pretty much been “prove it”, so you’ll have to provide an example of a game where union physics are severely hurting the performance and CanCollide = false and Invisible+CanCollide=true bounding boxes around the unions doesn’t work.

#5

It’s ridiculous the effort we have to go through to get this. Wingman8’s thank-you should be enough.

#6

I wonder why the staff needs us to prove that we need something that everyone is asking for, like wtf.

#7

Then just all make a repro place and get it over with? There are tons of requests for it, but barely anyone has delivered a repro where unnecessary CSG physics are significantly slowing down the game.