Feedback Request: Deprecation of Unity Cloth

Glitching through stuff? Did you adjust the physics in the inspector? Not saying ot didn’t happen for you but it sounds like an inertia & mass thing. I’m curious how detailed of a mesh you were using. This sounds like one of my first tests going way more high poly than needed. "0.05 distance between nodes glitched like crazy, but when I adjusted closer to .25 it was more natural. It’s been a few months since I did that project, but that’s my recollection. I hinestly haven’t experienced what you describ3d outside of no ideal settings. There’s no glitching in my example. Also, I provide variations without out any scripts So I don’t really understand what you’re saying for sure.

EDIT: Do you want a voucher or what? Sounds like you maybe don’t have good settings. Just to be clear, this user has not asked for, nor tried my example to my knowledge. Just so there is no misinformation here. Fancy? Not really, more like boilerplate Unity.

Edit 2: You’re doing it wrong. Check my example to learn how to avoid the problems you mentioned. Don’t create a bunch of work for other people when you have nothing to gain. I don’t understand why anybody would go put of their to say they want something deprecated…like just don’t use it ityou don’t like but why you gotta help screw me and other people over? Kinda cruel tbh especially if it doesn’t affect you. Deprecating cloth cloth especially before there is a replacement is absolutely terrible. @Baste

It must be an amazing feature, maybe not he best in the industry, but something that is pretty good for all users

If they can manage to make cloth in realtime like the hand moving through the hair in the Time Ghost demo I’d be sold. That was a flex.

  • No
  • Yes, MagicaCloth 2
  • No
    Is there any replacement planned?
    I don’t use default cloth in my project because MagicaCloth2 is much better.
    But I could see that completely removing it could affect the accessibility of some education courses and tutorials where getting things to work quickly is the main point.

No, because Unity Cloth is inferior.

Yes, because Unity Cloth is inferior.

No, because Unity Cloth is inferior.

In general: If something isn’t good enough, drop it. With characters becoming super realistic, the realism is impacted by Unity’s inferior Cloth simulation. One has to revert to assets.

I have developed 2 Assets with Unity Cloth: Basketball and Football.

Your breaking changes in Unity 2022 forced me to recreate Cloth Mechanics, and now you said that it will be deprecated… It’s very sad.

1 Like

Love Unity Cloth. Used many times for dance stuff. Just requires patience and understanding :slight_smile: So sorry to se it go. It’s free and no different in outcome as paid versions of similar.

3 Likes

I think I have used Unity cloth once in my 8 year career. I would always use a custom solution instead.

Unity cloth doesn’t really perform outside of basic contexts and wouldn’t have a problem with it being deprecated.

While I do think it’s bad practise to expect 3rd party plugins to fill holes in the engine, unless it’s done well enough that it’s state of the art, it shouldn’t be worked on.

I don’t know if that was expressed directly earlier but let me add one question:
Are you planning to remove Unity Cloth or replace it?

If it is the latter - then it is ok. If it is the former, then it might feel weird. Wouldn’t it create a precedence - how many features will you be able to remove from the engine and make us pay for the 3rd party solutions?

1 Like

We’re using Magica Cloth 2 since Unity Cloth is too simplistic and Obi Cloth is too bloated. We would like it if Unity invests more in physically-based animation in the future and we’re looking forward to the Sentis cloth example from the Time Ghost demo.

2 Likes

I’m pretty sure that that wasn’t cloth physics per se - as in it wasn’t an in-engine physics engine that was moving that cloth.

Instead, they had a high quality, but slow cloth sim that was running in… some other software than Unity, can’t quite remember. Then they used that to simulate cloth for a bunch of animations, and ended up with like 80 gigs of data. Then they used the cloth positions and bone positions from that to train the AI model. In other words, they essentially used Sentis to create a highly compressed lookup table of bone position => cloth position, which they could use to get close to original fidelity cloth without having to run any models.

The clear downside is that that approach won’t work if you want arbitrary other objects to interact with the cloth. It only works for the scenario kinds you had in your training input. So while it created very impressive results, and is for sure a valid approach, it’s not going to replace a realtime physics based cloth simulator.

2 Likes

As said in other thread I would recommend to offer a new cloth system as a separate package, build on the great foundation of the Hair package with all the great features like SDF collision, automatic finding of primitive colliders in scene, LOD, ….
There would be a huge overlap with the existing code of Hair package, too.

Though I understand it’s a different approach than pure physics, I didn’t think about how much can it be used in different scenarios. Anyways, will see.

I’m using Cloth for the Cape of my Character and it’s working quite well.

Yes

No

Well obviously, I wouldn’t have a cloth solution then. And writing that stuff with Verlet, Mesh deformation would be way to much work. And assets … some of them are very low quality or are designer tools which have 4000 lines of Editor Code which I wouldn’t need, which impact my reload time / iteration speed.
Having a cloth solution provided by Unity is a good thing. This one isn’t perfect but for simple cases it just does the job and I am glad it exists.

I think this is valid, if not, then sorry but:
I good feature for cloth would be: https://youtu.be/w_GJrE-6HtI?t=937
I don’t know if Unity can do that without scripts or 3rd party assets

After several tests with Bi Cloths and other assets, we use Magica Cloth, it’s simple and efficient. We yet have to port on V2

Were you aware that one of the leading asset store developers has a package dedicated to sails and rigging? I haven’t used it myself, but some of his other assets are incredible and highly performant.

1 Like

I think most people just repeat what they heard from others and never bothered to even test how it works for their use case.

1 Like

Thanks for your suggestion screaton :slight_smile: yes, I know this asset and I will consider it, depending what happens with the Unity Cloth. In fact the look is similar to the solution that I use for the ships in the background of the simulation. At the moment when ships are far the sails transition smoothly from cloth simulation to a skinned mesh renderer with blendshapes (using Cloth.SetEnabledFading() ) It’s more performant that cloth simulation but less than a shader solution.

For the player’s “hero” ship, which is always in the foreground, I still hope to have a good (Unity?) cloth simulation. For flags, sails being furled/set or with lot’s of movement, the behavior of the cloth is too complex (too many degrees of freedom) to get it right with shaders or blendshapes, I think. You can see it in my last video about sailing in bad weather (below). The movement of the rear flag or the extra sails being set at the end is too dynamic and complex. Or perhaps it’s only my problem: I already sailed in some tallships, so I’ve seen the real thing too much and expect more realistic behaviour

2 Likes

If you want to deprecate it, then make it open source. That way, people who find Unity Cloth helpful can also contribute to maintaining it.

4 Likes