Hello guys I have a couple avaters all setup to go.
But with Mecanim I want facial animation. I have my character setup but my character has eyes, jaws, eyebrows, eyelids and all that.
Mecanim does not play facial animation properly whether assigned or unassigned. The layers don’t do anything at all regarding my facial animation even with masking.
Also when I have the jaws selected and go to muscle setup the jaws are always open on that setup for some reason whether there is no animation applied or not. My problem here is that I want to use mechanim with legacy animations as there is no way mechanim will ever be able to work with facial animation alone since that already is buggy and most facial rigs have a complex setup.
I know I could have the face as a seperate gameobject using legacy animations but I want to have 1 complete model that can have legacy animations for face and mecanim for the body / retargetting.
Seriously, No one knows the answer…well then I will count this as a bug as I met another person who experienced the same problem.
For now a way around this would be to separate the face as a separate max file and make sure the positioning is correct or use vertex snapping in Unity to place in position. This way the body and face are separate and so the body can be Mecanim while the face is Legacy. BUT this still doesn’t solve re-targetting because the face is connected to the neck bone. So a real solution is needed.
Wow, slow down SilverStorm, 5 hours of bumping a specialized question on a public forum? Please don’t bump more frequently than once per day!
The feature is called Additional Bones… and if you check this feature both when you import the Humanoid model and the Humanoid animations this should work out of the box.
Thanks I will give it a go. But last I checked that never worked either but I will double check. And as for the bumping, noted :).
I will let you and the other person struggling know how things went.
Hello yes I saw both features and the first thing I have to say Is to please only have 1 keep additional bones button because this is really confusing since together with my retargetted character that = 4. Also it still does not work. I have created 2 Masks and 1 New Layer. I also removed all the facial bones from the mecanim setup and this fixed her facial expression from open mouthed to closed.
But in the new layer I followed the instructions but still no show. The character just won’t Facially animate AND do the retargetted animations.
I am really not happy here how could something which was so simple in 3.5 become near impossible in 4. I urge you to make Legacy Animations compatible with Mecanim.
This way I can have an Animator Component and a Animation Component attached. 1 for the retargetted animations and 1 for the face…so simple.
Anyway regardless I still need a solution so if you don’t mind please let me know what else I could do.
And sorry to bother you but this issue has been driving me crazy for the last few days, the layering seems to do absolutely nothing at all.
You should not use gameplay characters to play facial animations thus you don’t need Mecanim for this.
Unless you are aiming for a AAA super realistic game I don’t see why you would want to waste extra resources for facial animation for gameplay characters.
Facial rigged chars are most used for cinematics or close up cut scenes, so legacy is just fine for that.
That face rig on Mecanim is certainly for future features, not useful for anything right now but movement of eyes.
Bruno I don’t think by adding facial animation to a gameplay character is going to eat up so many resources. Maybe two or three years ago I would be thinking like that. Also it’s not about trying to make AAA super realistic games that you would only use facial animation.
Now is a good time for game developers to go to the next level, we’ve seen so much of the same thing. Games need more humour and charisma and utilising facial animation is a great way to do that.
If you look at the facial rig in unity you will see that there’s a reasonable amount of bones, so I don’t think it’s going to use up so much resources.
I have confidence in the developers of mecanim. They are motivated and take pride in their work. I just think more info on the future of mecanim would be nice. And to reply to post’s about facial animation.
When do they plan to release more features. And what do they recommend when using the current rig for facial animation.
I’ve watched a video made by iClone devs showing what they can do with iclones facial animation and unity. They want to release this version but are waiting for unity to support blendshapes. I assume unity is working with them on this which is great as I’m definitely interested in buying iclone if unity supports this feature.
Bruno check it out and give me your opinion please
It is the lack of blendshapes operability that causes the issue. A functional hack is to hand rig a bone implementation from your standard rig. Mecanim handles additional bones just fine. Rigify and MakeHuman and most auto-rigging software use some type of deform key implementation which is where Mecanim falls down.
I’ve experimented with this in a limited way(just enough to prove the concept). Hand rigging from the neck bone and building a custom face rig will perform like you think it should. It’s just not in step with any of the authoring programs I’ve found and requires some really tough hand weighting. Not ideal but not impossible.
I’m not sure if I’m in the same boat as OP, but it sounds familiar. When I import my facial animation(bone based) for layering on top of Mecanim body animation, the only animation type that lets the fbx play back is Legacy. If generic or humanoid is chosen, even with “keep additional bones” selected, the fbx will not animate (even when dropped into the scene standalone, without mucking about in mecanim).
As I understand it, legacy types are not usable in the animator.
I’m sorry that we have become so closed lipped, but everytime we’ve said something like feature X is in development, people have their own opinions on how long that can take and when we should release. The Mecanim animation system will support everything the old system did and then some, but I cannot say when. As usual the timeline is we ship when its ready.
But I can say that I sat down with the Mecanim guys during our last hack week. Among other things we played with facial animation using blend shapes, and that worked out very well. I think that blend shapes is the best way to do facial animation at the moment. In my opinion it worked better than what I could possibly rig with bones. There are several solutions on the asset store that offer support for blend shapes and it wouldn’t take long to implement your own. One day Mecanim will support blend shapes too but not right now.
I’m not sure, Bruno! We are on a bit of a limited budget(both financial and manpower) in the studio, So I’ve been tasked to find the most cost effective solution for face animation + character body movement. Does anyone have experience with Speecher? http://www.achest.pro/index.php/2012-03-15-19-15-06/mspeecher/video
The player demo is in rough shape… but it looks similar to FaceRobot as far as workflow goes, and it seems to offer the ability to bake down the animations.
We may end up bypassing Mecanim for our current project, but it’s a very tempting platform for animator collaboration in engine!
At the moment I have found a solid workaround for animating faces. It does not need Mecanim though. Mecanim being used for the body and then adding a “Animation” component to the Character via miscellaneous component.
Together with face bones removed from Mecanims rigging page and a legacy face animation it works, this can also be extended with MixingTransforms to animate ONLY the face while the body is using re-targeted or just simply Mecanim animations.
Its a little tricky via my explanation and I could provide a tutorial but I think the point is NOT having a workaround because it should be built in. I am sure they will fix this if its something they care about. I myself have had a lot of issues with Macanim so I have been sticking to legacy except for retargetting.
Though in the end I chose bones, in the long run and time saving it is the better choice and is industry standard, most companies avoid blendshapes as it is not cost effective, for example you cannot use blendshapes on several characters but you can with bones as you can skin as many characters as you like to that bone and each blendshape has to be manually created PER character, PER expression so I don’t think its worth it.
With bones you can retarget the facial animations like I have done and continue to re-use it over unlimited amount of characters even if theirs extra bones or a little less.
If you combine face mocap (apparently you can track white dots on videos of yourself talking and it can be applied a face mocap! It works) with it you can create some very convincing Re-usable face mocap since the markers get translated onto facial bones).
I have done a little bit of testing and its very flexible.
I think you guys should try this method out its dirt cheap and provides for effective mocap…I might just take it more seriously for my female paradise projects!
“I’m sorry that we have become so closed lipped” Don’t you mean disconnected? that’s what it is right.
“people have their own opinions on how long that can take and when we should release.” Is that such a bad thing? if their opinions are unrealistic then educate them. isn’t that what makes forums interesting “opinions”
"As usual the timeline is we ship when its ready " pfff it’s sad that you guys have become so comfortable with saying that… it’s rude and sounds smug
why can’t you say for example x feature 40% complete or 10% or 99% and what is the target date of release. And if you don’t make it on time then explain a little what’s the hold up and beg for forgiveness… don’t underestimate the power of forgiveness and the joy it brings… although the GUI thing is hard to forgive many of us have, haven’t we??? hello anyone there?? ok maybe just me.
my point is this, we mess up you mess everyone does but it’s a good thing that even though unity team spent a lot of time and money invested in the new GUI you still decided not to go through with it because it lacked awesomeness so instead you decided to kidnap the NGUI genius and now he’s going to inject some of his awesomeness with yours = super awesome GUI … no pressure… really… it better be super awesome.
“There are several solutions on the asset store that offer support for blend shapes and it wouldn’t take long to implement your own.” No thanks… bugs on top of bugs… And I’m thinking Mecanim is the reason I upgraded.
BrUnO XaVIeR “Don’t FaceFX play facial animations on top of Mecanim?” $899.00 is too expensive but good question does it play facial animations on top of Mecanim?
SilverStorm thanks for info "companies avoid blendshapes as it is not cost effective, for example you cannot use blendshapes on several characters but you can with bones as you can skin as many characters as you like to that bone and each blendshape has to be manually created PER character, PER expression so I don’t think its worth it. "
hmm interesting indeed. it would be nice to know if and when unity does release the mecanim with blendshapes would we be able to reuse them on different characters?
I’ve looked at this little youtube clip and it’s quite impressive what’s been achieved with little amount of bones. have a look
Anyway i hope unity team communicates with us a little more, it would help us to plan out our projects. This is important guys, less of the "As usual the timeline is we ship when its ready "
That’s a bit harsh. I think the point is that you should plan with what features are in Unity. The second you start holding back and wait for a particular feature you’re essentially putting your project on hold and I wouldn’t recommend it. We will always strive at creating something new and better! And once that has been achieved I’d be surprised if the continual development of the industry doesn’t send us ten new concepts to add.
So following that logic to its ultimate conclusion you might as well sit down and wait for Unity 10 with the voice activated create virtual reality world tool that will give you full power to unleash your creativity.
But seriously, even if I told you everything was 90% ready, we all know that its the last 10% that takes 90% of the time. Then we push to alpha group and they add their changes, then we push beta and they add some changes, before you know it a feature that’s 90% can be a year before its released. Are you really better of waiting for a year?
Eh, no pressure… sure…
The Mecanim team is using the word blendshapes, so this will be added at some point hopefully not too far into the future.
Well Blendshapes are almost per definition per character. But I think the guys from Faceshift is on to something with their list of basic expressions and then blending between those. That way you can define a facial animation as a blend between a number of basic expressions. If your modeler knows he has to create X blend shapes per character then it becomes a finite amount of work. Giving you infinite possibilities.
That is pretty impressive! I can’t help but wonder if he is using more than four bone-weights though. Still I bet it would look acceptable with 4.
I’m sorry you interpret this as disconnected and smug, that is certainly not how we feel about it. We are merely trying to remove the temptation to wait for a particular feature to be released. When we released Mecanim to the alpha group we released only Mecanim Humanoid, and the response we got from Alpha was quite simple: “Not good enough, we want generic or we can’t use it”.
Given such clear feedback that a feature is incomplete or too cumbersome to work with, we simply have to react to that such ruining any ideas we had about a timescale. While such feedback is painful for us, for our customers who have been delaying projects waiting for a particular feature this is truly devastating. And that’s why we are trying to educate you to use what features are released right now, rather than waiting for something that may need another iteration or ten.
That being said, I think we could be a bit more open about the future. Hopefully we’ll find a good balance, but right now we’re leaning back in the silent direction.
MoHoe thanks for the reply, I do agree and I find many bugs with Unity. I don’t bother submitting them because I just don’t think they care at times…
The other day after several memory leaks with Mecanims re-targeting the engine finally crashes (for the 20th time in the last two months) BUT this time it deleted the whole scene file and I couldn’t load the file back up. How this occurred I don’t know…I couldn’t find a search for it.
Meta-data errors, Referencing previous file errors etc etc (mostly import crash related - from a result of importing of native 3ds max files that have a key per frame and can span for 5000 frame animations).
Apparently with Mecanim each time you import the engines memory footprint is bigger and bigger to the point of going around 1 GB.
So I gave up on it rebuilt the project from scratch into a clean project excluding many complicated’s such as breast physics which I had in my Open Island game…but In return I struck gold with new lighting tricks that make the game look wonderful that I missed in my last project!
But yes the engine being optimized as it is has an unusually high amount of bugs and lacks many simple and standard tools such as a rotate 90 and mirror tools as well as many other basic features. Its Shadows in Pro are broken and I encountered bugs with the light-mapping system (the console actually said it was a bug and many other things). I think I have gone over the tangent here but yeah I have stuck by this engine even at times when I have hated it because I truly think this engine and its company are good ++.
I think we should have a monthly feedback forum where everyone goes on a spree of feedback on Unity things they don’t like and want changed and chances of it being accepted are based on replies to the request etc.
What do you guys think or do we have something like that already>?
Perhaps a big Poll system too where “Insert Feature or Change here” and people can thumb it up and you can put like 100 features per page each with a +1 from you or not.
I think if its anything worth taking its that before jumping to the big features that could complicate things (humanoid didn’t play legacy animations for me and re-targeted animations = it was VERY painstaking and I still haven’t got some of the features of Macanim working yet…and probably won’t.
I am only using Mecanim for legacy animations, it should have been designed from the get go to work with the legacy system rather than trying to create a whole new This OR That system and Not Both system and brought with it a whole lot of problems. And I still have not heard or seen anything much from the public or even heard positive feedback from Mecanim. I will not be using Macanim at all in my new clean project.
Its ok though because we all learn and grow.
Oh a feature request please for the next version of Unity input Create able/Savable Import Presets-Unity should ask that you have a new file incoming and you should activate a preset OR edit all fields manually BEFORE the file loads so importing has to be done ONE TIME not 4 as it currently is : That is one of the many features that should have been! I can expand upon this if anyone wants to know.
UnLogick “So following that logic to its ultimate conclusion you might as well sit down and wait for Unity 10 with the voice activated create virtual reality world tool that will give you full power to unleash your creativity.”
ok well you proved my point about smug replies. that is the 2013 new culture at Unity Team, drop little digs at unity users when they ask questions. yep there’s plenty of videos online showing this sort of thing. You the smart guys and we the clients are childish and unrealistic and therefore don’t deserve respect. if the guys at iclone have proven that facial animation is possible in unity, and are waiting for blendshapes support then I don’t think me asking when is unity expecting to have this feature deserves a reply like “might as well sit down and wait for Unity 10 with the voice activated create virtual reality world tool that will give you full power to unleash your creativity.”[/B] If you tell me that you are expecting to support blenshapes say after 12 months then I would spend more money on a alternative solution. If you tell me less than 6 months then I will delay the facial animation and focus on other elements in my game, like modelling, ai ect. “Well Blendshapes are almost per definition per character. But I think the guys from Faceshift is on to something with their list of basic expressions and then blending between those. That way you can define a facial animation as a blend between a number of basic expressions. If your modeler knows he has to create X blend shapes per character then it becomes a finite amount of work. Giving you infinite possibilities.” Thank you for being helpful. When we released Mecanim to the alpha group we released only Mecanim Humanoid, and the response we got from Alpha was quite simple: “Not good enough, we want generic or we can’t use it”. So you left humanoid unfinished broken and unstable and started working on generic. hmm I think you went with generic because you thought if you added this feature it would help you push mecanim as a good reason for people to upgrade. Why would you channel your energy in that direction if most people who are going to be making games are going to use humanoid characters, maybe less than 10% will use generic. So the 90% users are left with a broken humanoid system. wow how much fun the users are going to have with this. “Given such clear feedback that a feature is incomplete or too cumbersome to work with, we simply have to react to that such ruining any ideas we had about a timescale. While such feedback is painful for us, for our customers who have been delaying projects waiting for a particular feature this is truly devastating. And that’s why we are trying to educate you to use what features are released right now, rather than waiting for something that may need another iteration or ten.” That’s better “Hopefully we’ll find a good balance” I hope so, maybe it’s not too late for unity to fix their reputation. You don’t need a good reputation to get free users but you do if you want paying customers.
