Animation Status Update (Q3 2023)

12 Likes

Mecanim has been de facto abandoned for many years now with all hands on deck trying to figure out the DOTS mess thatā€™s supposed to be 1.0 since last year, but it still takes a 30 minute tutorial to get a damn hello cube spinning in place so I wouldnā€™t hold my breath expecting a usable animation system any time soon

2 Likes

Hi @UnityChinny . Is the experimental package release date has been finalized internally? Really hope to see it available at Unity 2023.3 cycle next year asap like January 2024.

Thanks to everyone for the feedback, as always.

Mecanim started around 2009
Mecanim has served production needs for over 10 years already, however we all know the architecture simply will not scale for production demands for the next 10 years, which is the main reason we started working on a new system.

Which bugs are getting fixed?
As I have highlighted here - we will continue to maintain Mecanim and continue to fix bugs, mainly crashes, blockers, regressions etc. However it is really unrealistic to fix every issue that is logged for such a mature system which serves so many varied demands. As part of this, we recently closed a bunch, which is why some of you got multiple updates in one go.
Does it mean we are abandoning Mecanim? No, it means we are going to keep it productive, while trying to deliver the new system as fast as possible.

Quarterly status updates
Since joining Unity I have attempted to be as transparent as I can, in order to keep you all informed so you may plan accordingly, instead of saying nothing at all, which invites wild speculation.

Answers to some other questions in this thread
I canā€™t answer all specific feature/workflow questions asked, but will answer what I can.

Yeah we are looking into this ECS overhead

Humanoid is very popular, but also very limited which is why we have generic. So we are looking at somehow combining the two.

It is not possible to remove Humanoid system from Unity core. But we do want any new system to be flexible.
You also asked about releasing an experimental package already. As I mentioned previously, the plan is that as soon as one is properly ready to be tested, it will be released.

Not sure if you are referring to me personally or Unity. Either way, I think your statement is overly harsh.

I donā€™t have that visibility, but will say I do not expect our announcing animation next month.

This is a very fair comment, and I do talk about this at the top as our over-promising and under-delivering, so we are keenly aware of it. Maybe under-promise and over-deliver? But that might be swinging too far in the opposite direction, it is a tricky balance, and has already taken much longer than we expected - which we are very aware of. There is no quick fix to this problem unfortunately.

Similar to Basteā€™s post above, this is definitely something we do NOT want to be the case for the new animation system

Chinny

8 Likes

:hushed: I think official start working on new animation system too late. The work should start during Unity 5.

Any plan to have announcement at Unite 2023?

Ah, that is what I meant when I said next month, I should have said ā€œnext month at Uniteā€

1 Like

Thank you for your response here

But first thing first, your answer was not touch the main concern

What I concern is, would the new animation you talking about be a new system? A new package? A new paradigm?

Would it integrated into the core of unity? As an extension? Or being the completely new way to do animation?

Would it be a breaking change? Will it work with humanoid system? Is it expand the functionality of humanoid system or rework it? (I have a same concern to mecanim but be as it seem it was being replacement of mecanim, which I think many people donā€™t have much to say on that)

And will it apply to current tech stream or is it will got into next year version of unity only? And how the old system will be maintain for the project in older version?

One thing I have in mind for such time. Could you (unity) please opensource some functionality, at least for the old version of unity and especially the deprecated functionality? As you can see the old animation system (and many system) was marked as wonā€™t fix even people still rely on those feature. If you opensource some of them we then could try to fix it where it possibly could

Your decision on crashes, blockers, regressions are not always align with us as you can see many people are always complain about it

@UnityChinny Also you are not even touch anything about Animation.Rigging package I have been asking. I am currently doing something related to it and I donā€™t want to put my effort on something that soon would be deprecated. I beg you please be more clarified on this

From what I know this new animation system support both game object and dots. For game object user, is that only entities core package is required? If yes, that means the rendering part is still using game object mesh renderer.

Official has the issue of keep shipping half finish product that always not working really well and buggy. Basically unusable for serious production. One of it I think is new Input system that until now itā€™s still buggy that canā€™t support mobile input properly like Android yet. But now official seems like become scare to deliver any new product and become never ship any new product which become even worst like dots utility AI solution I asked for years but never see any experimental package available to public yet

The Animation Rigging package will be around for the same amount of time as Mecanim, and Mecanim will be around for many years to come.

A fair point. But as I have said there are so many different ways Mecanim is currently being used, it is practically impossible to address them all, so if there are workarounds etc, then we are very unlikely to address them.

All fair questions - however I canā€™t answer them all at this time, sorry. But I will say that while the system will be new, it will not be a completely new way to do animation. It will be entrenched in traditional methods, built on a new architecture. An architecture which will rapidly accommodate any traditional or non-traditional way of working.

This is certainly above my pay grade, so canā€™t comment on its feasibility.

Donā€™t expect any positive changes made to it for the following years and expect a few negative ones. If it does what you need it to do and more, use it, but keep in mind as various regressions happen over time, they will not get fixed.

Being around like Mecanim probably means it will be neglected, bug will not fixed, and not compromise with new system, that it the meaning of being deprecated. And we will not know any critical use case will be blocking issue in the future so we cannot trust to use it just because it being around. Thatā€™s the point

As I have said, people are complaining right here, right now, that the bug in animation system was marked as wonā€™t fixed even if itā€™s their blocking issue

What about my question regarding low level access to the system?

1 Like

Do you have any plans to restructure the team?

Iā€™m looking forward many years in the future when the new system fails to deliver on all fronts, weā€™ll get a Unite talk that Mecanim should be the secret weapon in our toolbelt or something.

3 Likes

So to use this new animation system, the ECS package will need to be installed even if nothing else in our project uses ECS? Iā€™m already forced to require the Burst package which drastically increases domain reload times.

I pray CoreCLR comes soon and saves us all.

1 Like

Is there a reason you canā€™t disable Domain Reloading? From what Iā€™ve read, the upgrade to CoreCLR will [require everyone to do this]( Unity Future .NET Development Status page-34#post-8918231) anyway.

Are there any update since at this time of 2 weeks ago? Are there any new information about the new animation system as of yet?