Someone at Unity, do you know what happened to Kinematica?

Hi, Unity team,

I don't know where to ask this question as there is no forum or any post about Kinematica.

Kinematica, by far the best highlight from United Berlin and I was really excited to see it. It can be a game changer if it works as advertised.

I know Michael Buttener is the main guy and I have lots of respect for him that he can pull it off.

He promised some updates by the end of summer and trying to get Kinematica into our hand before the end of this year.

I was hoping to hear some update on it from United LA, yet there is no mention of it and not even a blog post.

I'm dying to know what's going on. First of all, is Michael still with Unity? If so, what's the reason for the such a dead-silence. I think we deserve some answers for waiting quitely for a few months. If someone at Unity sees this message, please pass the words to Michael.

Thanks in advance.

Cheers!

 

yeah, I am dying to hear more about kinematica!
Please someone inform us.

 

Hi, Michael. With all of my respect, I know you will blow our minds again when it releases. And I'm also very glad that you are busy working on it. Sure, a couple of months delays is not a problem for a project can change the animation workflow forever. Yeah, please let us know once in a while so that you are well. Cheers! ^^

 

Thank you Michael for the update. It's amazing to read that this is not some intern red herring tech demo. While i fully support the idea of making the tools first and make sure it's sort of usable for the end user - it would be a nice move if not all and everything happens under the hood. When i first heard of it, we stopped working on our current character-animation pipeline - spending time working on other topics. Yes, it sounds that crazy but we decided to go that way because our current game-pitch approach is to make use of current and upcoming technologies that may redefine how way smaller game studios are creating their games. Kinematica, from what we heard and see could be a tool that gives us a new way how we let characters interact with the environment.

What we're looking for now are some informations on prequisites. Maybe it's possible to "leak" the one or another information (if they exist) regarding the following topics:

1. - What is Kinematica being capable of in its current state compared to an regular character controller solution?

2. - Is the final goal giving the end-user an alternative solution to current character controllers (like the new FPS Sample Project or the upcoming Third Person Character Controller) or even replacing these in the future?

2. - What does the end user like us need on both hard- and software side when we're going to get our hands on the first public tool sets?

2a. - Does is come with predefined motions libaries? It looks like most of the input comes from mocap-animations you're doing. With that in mind, will there be some sort of library with animations for subscription / per animation-take payment like seen on https://assetstore.unity.com/motion-library ?

2c. - Is this going to be an solution a smaller game studio without much ressources can actually use for a game pitch / game production? I really seek for an realistic answer here because at the moment, we don't have the budget for any mocap-studio / equipment during the pitch.

3. - Is it worth waiting a few more months and keep waiting instead of using an regular character controller solution?
Yes, stupid question but i don't want to spend worktime on something that we probably won't need anymore then.

Thank you for all the great work your team is doing.
We keep our fingers crossed and can't wait to read and see more from it.

Tobias

 

Anyone come across any updates on this? This tech is super exciting.

 

Yeah, I very antsy also about the progress. I will appreciate if we can get some updates. I think it's about time since we waited from the Summer.

I'm too holding off some aspects of character animations until I know more about it. I believe Kinamatica will give us very nice locomotion and I think all we have to do is to layer upper-body animations on top of it.
I think it will give us the best-looking/easiest-setup character system.

I may be wrong but I won't invest too much developing our own locomotion system but keep it a very basic until then.
More info would be appreciated for the future planning.

Thanks.

 

As always with experimental/promised/scheduled unity features: It's generally not a good idea to plan your current production schedule around them.

That being said, I'd also be very happy about any updates.

 

From what michaelbu said on 31 Oct 2018 in this thread I wouldn't expect anything you could use before Jun 2019 realistically. And if things don't go good then maybe not til end of 2019 or later!

 

On one hand users have asked unity for more transparency since the company began and of course on the other hand now people know what things are planned years in advance, they want it now

 

My takes are, if they are not going to share any information, why make such a big announcement so early?

Not only they make the big announcement, but they also promised for more info by the Summer and something to play with the end of this year.

The last thing I heard is that it's pushed to the end of 2019. We don't even know if that's going to be a release? or just an experimental,

Just like many other "experimental" and it doesn't mean much(I learned the hard way); it will be an uphill battle to become "preview" and quite another battle ahead of it.

In Unity universe, time goes about 2 times slower to say it nicely. Right now, it's the end of the year season, so we all have to wait for the next Unite for more info.

In 2019, I wish there is no more Unite so that they share the info as they are available instead of saving for the event.

To be honest, there aren't much new except rehashing what's been said and just a few new announcements that are like this. It happens too often and it's such a big waste for them to prepare for the event just for the sake of the event.

I got so excited about this(and the new network stack) when I first heard it but now my resentment deepens.

Anyway, I wish the very best and Merry Christmas to everyone! and I hope things will change in 2019.

Cheers!

 

Any new info about this?

 

No updates since Halloween, eh?

I bet this goes the way of so many things. One by one, individual developers will make their own version of this, sell them in the store, and about a year after I've already bought one or two, this will come out.

 

even no mention on roadmap. i guess that was inspirational presentation. hey we made it, so you can!

https://unity3d.com/unity/roadmap

 

This is not quite the same thing as Kinematica, and it hasn't been released yet, but at least we're getting WIP updates on it ...

https://forum.unity.com/threads/motion-magic-mxm-motion-matching-animation-system-for-unity.620713/

 

Unity never fails to disappoint users.
I was so excited when I first learned about Kinematica and it turns into big resentments.

MxM is really interesting and I hope it out-do Kinematica, if Kinematica ever releases...
Cheers!

 

saw this in unite videos,

2019.3 kinematic (preview)

 

Don't forget the bit about GI that they unfortunately forgot at the end of vid since it was clipped by mistake. That's the only logical explanation *stares at Unity*

 

Is Kinematica(preview) still scheduled for 2019.3? I really hope so since it's delayed about a year.

 

Based on a roadmap presentation held today by @cirocontinisio - it appears so! However, it kinda seems to be shrouded in mystery like most of the Unity labs projects, apparently.

But hey, in the meantime @CptKen's motion matching asset has been released.

 

This means it'll likely drop as a package at some point quite far off so don't set your watch considering it can run to next January.

 

No news :thinking_emoji:

 

This quadruped motion solution has just hit the store. You might find it interesting.

https://assetstore.unity.com/packages/tools/animation/bio-animation-147937

The author says it will soon be followed with a biped motion solution.

So, for those of you keeping track, there will shortly be two solutions for humanoid motion in the asset store. Which means we are probably only two years away from Unity's Kinematica.

In case anyone needs a quick link, the first humanoid motion solution to the store is this one: https://assetstore.unity.com/packages/tools/animation/motion-matching-for-unity-beta-145624

 

FWIW, it looks like the biped was added to the quadruped in the Bio Animation solution mentioned above. Also, a small price increase from $40 to $50.

 

Interesting! Looking forward to some videos (and if/how they're different from previous work of his research group.)

 

There are some restrictions, like forcing the game to run at 30 or 60 fps, the models must be based on the same skeleton, and you must allow 2-4 ms per creature for the calculations.

It sounds like MxM is a little more flexible and a little more performant, with an estimate of about 2.2 ms per character.

 

You're completely mistirecting people. Bio Animation is hardly suitable for commercial product. It is based on this tech and more like a research.
https://github.com/sebastianstarke/AI4Animation

You cant change skeleton.
Only biped and quadruped animations.
You cant add animations (Deal breaker).
It runs very very slow.
You have to run 30 or 60 fps (with 4ms per frame yeah you will).

MxM is proper way to do this(Since kinematica isnt released yet). Highly jobified and runs under 0.5-1ms. %100 flexible, do what you want. You can watch performance and tutorial videos on forum page posts (Store page doesnt have them all)
https://forum.unity.com/threads/bet...ation-with-no-animation-state-machine.685489/

 

There's plenty of room for you to give your perspective. I don't need to be on the receiving end of an attack. I'm getting pretty tired of the "punch first and often" approach so many people use on the internet.

I have mentioned MxM in several posts in this thread. I haven't said which is better, of the two packages. I did happen to mention the restrictions on Bio Animation, which you can see in the post just above yours.

People can research them and decide which better fits their needs. I'm not advocating for anyone to buy either.

 

That is not an attack. Thats why we're writing and reading on this forum. I dont even own any of these(yet) and researching like anybody else. I saw Bio Animation as soon as it released and got exited. But its far from motion matching system as you want to put in your game.

 

Myea the hype for motion matching got so hot there for a while it almost had me convinced it was the next great leap in game animation... Not sure why it seems to have died down.. Or did Unity just drop it in order to make room for some sexier feature such as.. I duno.. an API for drone technology that enables mobile AR-gamers to have RTS-like gaming experiences. Allowing the elevation needed to effectively enable a top-down like AR experience..

I mean, yeah, no lets not do THAT because of some sort of 'innovation to animation in game development' for nerds.

 

MxM is having trouble adapting the system to use cut clips, which is the most common type of animation sold in the asset store. It could be that Unity is trying to get their system to work with cut clips, and is finding it difficult as well.

 

Well, from experience I know making a system which simply "works" is much easier than making a system that works AND it's easy to use in both the editor and code/API. From the video, I gather the system is already working, maybe they are just dealing with edge cases, user experience and documentation, etc. The "little" things that take forever, haha...

 

It could also be a matter of getting things to look as good as the developer wants them to look, something like that. He's said to me that the cut clips just don't have as much detail / information as mocap does, and that's the issue.

People come up with solutions for problems all the time. There could be a solution in the works somewhere for this too.

 

True, though I should clarify, just in case, I was talking about the Kinematica video posted above, not MxM. Buttner does mention Kinematica works with cut clips too, since it'd be useless if it worked only with perfect animation/mocap. He does explain a lot of his work getting there (can't say I understand all of it though, my math is not that great, haha).

 

Any new info ? I see that we are now in 2019.3 beta, when should we expect it to be available ?

 

roadmap pdf https://drive.google.com/file/d/1h4u_vHmm6ISxwn6-NAlcwOsuvVqb4iyK/view

 

WTF, another slip... it better be good when it comes out!

 

Just to keep the dream alive really.

NECRO!!!!!!!!!!!!!!!!!!!!!!!!!

So flame me

 

Yeah, sadly it's still not available as a preview package in 2020.1.0a8. Hope it does not get delayed any further.

 

I hear meditation is a fantastic remedy for this sort of thing.

There's no rushing stuff like this though obviously, it's a very tricky subject to tackle without A+B = C.

 

no rush. no rush.

 

FWIW, MxM is about to exit beta. I think the plan is for the price to double, or something like that. If anyone is interested, check the MxM thread and/or the Discord for MxM.

 

Maybe Unity could hire Sebastian Starke et al.?

 

Yeah, I agree, that project looks awesome. Having Kinematika do all these stuff like this would give Unity games an advantage in animation quality, AAA feel even for the most noob user. That would be fantastic.

 

He's already working for EA as a AI Scientist Intern and he's still working on finishing his PhD so...

 

It was more of a joke and to highlight a new cool video about the progressing state of the technology. I don't expect a random forum post to have much of an impact on recruiting decisions, even if I definitely wish unity managed to get some people from that research group on board (rather than Ubisoft and EA.)

 

expect the unexpected. Think positive!

 

Well, regardless, the project is public, so people in charge of Kinematika could take a look at the code to see if there's anything worth mimicking.

 

Any news about kineamtica, it is incredible that this is set aside by a common rigging system, there are not enough in 3D editor programs?