UMA - Unity Multipurpose Avatar on the Asset Store!

Glad to hear it, welcome to the community.

You need to add the Global bone with the distinct 90 degree rotation similar to how the current content is set up.

To create a new race you need to set up a dna converter that converts from slider values and into transform modifications. The bones you modify in your dna converter should be adjustment bones that are not animated. For facial animations like lip-sync you should modify the normal bones and in case the normal bone is animated, you might wanna add additional bones just for the lip sync.

What the dna values means are up to your dna converter.

The easiest way to do soft bodies are by adding a non animated bone with a simple "spring" script. We were working on cloth support in a branch but it's currently bugged.

I just checked the commit log on Unity53 branch, and throughout January there is several commits a day.
https://github.com/huika/UMA/commits/unity53

Overhaul
About the overhaul... the reason we invented our own naming scheme is that what we're doing is radically different, and while it adds a burden on the learning curve you are unlikely to confuse it with alternate uses of the words. So while Slots can be considered meshes, there is a distinct difference to meshes. And while Overlays can be considered textures, there is a distinct difference to textures.

Finally we do have a fairly large user base, and we cannot change the names without forcing all of them to refactor their code. We did this once in the 1.x series and the reply was a very loud and very clear: don't do this ever again.

If I ever get time to start talking about an uma 3.x the only thing I really want to do is make "Textures/Atlasses" a first class citizen, rather than something nested on the slots. Again I would take care not to break the existing code, but rather keep the Slot.AddOverlay and make it call the new methods so that old code still works even if we changed the underlying logic.

UMA buildings!
See this is what happens when something is open source, people take it to the next level. This is perhaps the best reason to pick uma!

Pro's of UMA:

• It's lightning fast
  It is faster than the morph based approach.
  
• Versatile
  • It's not limited to characters, (though at the moment it doesn't output static meshes, that is likely to change when needed)
    
  • Supports both Humanoid and Generic
• Built in texture LOD
  
• Open Source
• In active development
• Very Stable
  • Extremely few bugs (Currently 2 bugs and 2 feature requests according to https://github.com/huika/UMA/issues)

Con's of UMA:

• Poor documentation
• No built in morph support
  Morphs are still the industry standard for customization
• Few artists familiar with the workflow of content creation

 

There is an expression system built in for this exact purpose. The human characters have the standard phonemes and expressions created by default, but your custom mesh will need to define these.

Dynamic bone adjustment would work just fine (that's what UMA is doing). You can add bones at run time. Since the models are generated at runtime, you would have to have a way to identify them (for example, name them a specific way), or you could just let UMA do it all by creating Dynamic DNA (which is just a way to manipulate bones).

For soft bodies, you would have to create or add a solution. I'm not aware of any components in the asset store for this, but I don't know them all. This would go into the physics system, etc. I would recommend perhaps segregating those portions of the models into their own slots (with their own material) so they don't get combined. They would then become a separate submesh, which you could then access the vertexes and deform.

All of this takes time, and we all have day jobs, etc. Fortunately, SecretAnorak has created a set of videos documenting much of this, and illustrated a "Jet" and "horse" UMA just a few messages above. The vast majority of people want human players, so those are provided. However, you can add whatever you like.

We have created a simple system for using UMA in the new version, and documented it in an easy to use sample scene with help text. However, if you want to use advanced concepts like introducing soft bodies, you'll need to get into the code.

I don't agree with this.

Again, I'm not sure what technology and advancements you're talking about. If you're talking about using PBR, and being updated for Unity 5, we have done all this.

UMA is extensible also - you can add stuff on the fly. Unlike Adventure Creator or Playmaker, it's not a game making framework. It's a component. But you can create packages and automatically add them if you wish. For example, this has been done with FinalIK (on the wiki).

This is usually referred to as "blend shapes". These are more limited as they don't combine well, the setup takes longer, and you don't have the ability to propagate the morphs to other geometry. You can however with bones. All the clothing, hair, etc morph along with the underlying body changes in UMA. So you can put on a helmet and it will still fit regardless as to how you morph the head, as it morphs along with it. UMA has "adjustment" bones in addition to animating bones. You aren't stuck with just scaling them. You can scale, rotate, transform them (it's all setup in the DNA). This gives you a lot of flexibility, and doesn't disturb the animations.

They're called "slots" because the idea was that you would have a base set of meshes, and each would fit in a "slot". Then these would combine into an end mesh.
The overlays are called that because you can layer them, and recolor each part of the layer. They are way more than just textures being added to the meshes. You can even have a solid shirt converted into a strip one simply by adding a striped overlay, and recolor the different stripes to different colors - all on thefly.

UMA doesn't belong to me, or to anyone actually. It's all open source. I'm just one of the guys working on it.

We agreed that it needed an update. So we have made a massive update in quality and features. I've noted them a few messages ago. But you can see on the Unity53 branch that we have done a giant overhaul. I recommend downloading it and checking it out. We have 40 commits with hundreds of file change just since the new year, it's a massive change.

 

TeagansDad is correct. The Github is being updated all the time. The Slack channel is very busy.

We use IK and Opsive's controller with UMA but we don't go to the UMA people to solve problems with those. UMA is a character system. If you want to add other things, they you need to get the developer of those assets to help you.

Makes perfect sense to me! We have been using UMA in our game for sometime now. While I have lamented the lack of documentation, I never thought the terms were bad. Slots are meshes, Overlays are the textures on those meshes. The system allows for so much creativity. I personally love it.

If you prefer the updated Morph3d and how it is done, then you should buy it.

While this may not include you, I think people should realize that the developers do this because they want to do it. They don't get paid, as I said above. Because the system is open source, they cannot go to Patreon and ask for money. It has an open source license. And they respect that.

Morph3d has a for profit company behind it. So if you like it, then use it. Again, not necessarily directed at you, but I wonder if sometimes the really over the top attacks on UMA are people who are upset that a free system is not up to par with a paid system.


Personally, I looked at Morph3d. I did not like the models at all and they did not suit the style we want. I enjoy making my own clothes and Morph3d has a finally after no news on the subject released a tool that people say does not work. Willb makes much nicer clothes than the ones available on Morph3d. The options and creativity that can be used with UMA is amazing, mostly because it is not locked down. We can do all sorts of things with it, as I have seen in the Slack channel.

There are many options out there, lots of other character systems. Fuse, Morph3d, IMUV, and some rather expensive ones. You can use those if you prefer. You need to stop comparing UMA to other systems. Unlike Fuse, which is also free, you can customize easily in game. Unlike Morph3d, you have total control over your system, through code and through the way you build UMA into your own character creation system.

However, I don't want UMA to change to a point where I have to depend on others to make my content....unless it is paying for my choice of custom content. I like being able to see the development and get involved in the project.

But it is not for everyone. Docs are almost non-existent. In the past, UMA was difficult to understand but it is getting easier. Lack of 3rd party content and the cost of the tools necessary to make clothing is a factor...although that is a factor with any model you want to customize yourself.

Your comments were unfair to the people who have worked so hard on this. They are a FABULOUS group and very hard working. This is an OPEN SOURCE PROJECT. If you want to get involved and offer your own expertise and skills, please do.

I will be doing some more Twitch TV's on UMA and plan to do one on how to set up your recipes with UMA. However, waiting for the latest updates which are coming soon.

 

And Joen answers at the same time.

 

I think we were all typing at the same time!

 

Some extra clarity never hurts : )

Can just throw in that i use UMA + Lip Sync Pro and it works very well and took me 5min to set up, since they already have support for UMA.

 

Well, as promised, here's a short series explaining how to create a new humanoid race for UMA. It's shockingly easy once you know how. And not even a sniff of code in this one. Hope you find this useful.

 

This is really, really good.

 

I the words of our immortal Secret Anorak, "So easy, it almost hurts."

Thanks, SA, for your hard work! Great videos!

 

I apologize if I came off as rude and arrogant, it appears I had outdated information.

When I began researching UMA, I started from the basics and kept learning about it, but I hadn't seen any links or updates to it, considering the main UMA Asset Store page had people complaining it was heavily outdated.

Some people say to check out the GitHub page, which I did, but I found the /huika one, not the Unity53 as specified above.
I'm not sure how I missed it, but I did.

With new information, I am seeing a new light and realize I was acting on behalf of old information.

Again, my apologies.

I will begin more research for UMA with this new information.
Also, disregard my issues with the layout and such, it appears updated.

However, I still have to say... the slots, race and dna naming conventions still sound awful in my opinion.

I understand this is open source and I have no expectations for anyone to add features or take my word for it, I didn't intend to come off that way.

At this point, I am looking forward to what SecretAnorak is doing as I don't have a lot of time in my full-time job to simply toy around with UMA.

Regarding Soft Bodies, me and my team have come up with a workable, temporary solution.

Thank you for your time, things are looking great :3

 

You did find the correct UMA Github page, but active development is currently in the Unit53 branch, not in master. Also, you are welcome to join our slack group if you have questions or suggestions . )

 

No worries. We all get frustrated and need to vent. Thank you.

Really, it is not that bad. Once you get used to it, any change is confusing. UMA has been around for a couple of years now and most of us have survived the learning curve.

 

I've been using UMA for a while now and I like the system alot, thought I might change to morph3d but UMA so far has outlasted and remained the best option for my project. The only downside to UMA in my opinion right now is customization through morphs...

Do you think that might be something possible added or possible to add in the future? That is the only thing that morph3d has that attracts me, the ease of use with UMA is there.. Just not blendshapes..

 

Can you tell me what you like about blendshapes, and what is the use case? I'm wondering if we already have it covered with another method (expression sets), or if there is something you want to do that we haven't considered. The problem with blendshapes is that you can't apply them to arbitrary meshes (like a 3rd party shirt that you purchase). But maybe we can come up with something if you can tell me what you are needing?

 

Bones is also in general a faster solution than morphs

 

Just mainly higher level of customization, so far so editing things with the DNA sliders presented in UMA there are some things I can adjust to my liking, but I don't have control over finer things like customization the face to a higher degree or more realistic adjustments of body shape like fatness or muscle that blendshapes provide. Adjusting the bones in UMA work in a generic sense but can produce some wonky effects. Unless there is potential that I simply overlooked, adjusting the DNA slider for things like height or size quickly threw my characters out of proportion by scale on the xyz axis instead of being able to do simply adjust a height and have a taller character without the whole character becoming larger in all dimensions than someone that is shorter.

 

This sounds like something we could do with Dynamic DNA - basically it lets you define what the DNA means, so you could specify that you are adjusting a specific set of bones (like adjusting one of the spine bones and pelvis bones, and then inverse adjusting the next bones down the hierarchy).
It wouldn't be simple, but it wouldn't be hard either. From what you are asking, it really sounds like you want more and better DNA?
The main problem with morphs is that they don't affect the things that you add after the fact, so every piece of clothes would have to support every morph, and that's a ton of work to get right (which I think is why Morph3D is having a hard time getting stuff working perfectly -- no bleed through, etc)

 

Sounds good, if that is the best option. I prefer to use UMA since it is very easy to work with, and I don't mind having to swap heads out for customization. Just being able to fine tune more than what's currently available would be a plus. I don't need Black Desert type customization but something that will allow a player to feel like they made something themselves and not just another copy of everyone else.

Edit: How hard would it be to be able to add slots for like different noses, different head shapes, different eye shapes (models)? Different lips or mouths etc. I am wondering if editing the human dna or slots would allow for more fine tuning like that, so I can have a head04 but match it with ear07 and mouth02?

 

Its exactly the same process as adding new slots for clothing´, so its pretty easy. Just make sure it has the same rigging and yu can model whatever you like. I think there are some videos out there about content creation, and the same process is true for body slots.

 

You can also provide the player with different heads they can customize.

 

I don't have a problem making customized faces and bodies with UMA. One thing to consider, and a former colleague in the indie game industry told me this....In most games, unless you are right on top of someone, everything looks pretty similar. SWG had a great system for customization but still, everyone looked very much alike.

It might be better to give your players other ways to customize. Overlays are so easy with UMA that you could provide freckles, moles, scars, tattoos as well as a myriad of skin, hair and eye colors.

As for slots, the nose and ears are already separate so you could change those out if you want to make the meshes. Lips are an overlay. Maybe you could make meshes for those, but I don't know. Heads are separate and Will has a good example of a combined head you might want too look at it in the current Github release.

Just remember, a game has to be balanced. While modern games can handle a lot of poly's, you may want to save them for other parts of your game rather than thousands of unique UMA pieces.

 

Just want to show you a small new feature i just pushed to the unity53 branch, placeholders!
Thanx to @SecretAnorak for his code contribution i have integrated this into the new DCS system, so you can see where you have characters placed in the editor during editor time.
At the moment you can choose from Male, Female or add any custom GameObject with a mesh which can be good if you have custom races. Color is customizable in the inspector.
NOTE: This is just placeholder UMAs, they will not reflect the UMAs actually in game, but can be good for visualizing placement.

.

 

Actually, it's already setup for you to do that - there is a second head supplied that has the eye area/mouth/nose/ears cutout (and it has a set of each as separate slots). Actually it has 3 ears (round, elf, anime-elf). So you can even just load the "swapped head" mesh and modify the various pieces and/or create new ones.

 

It's notable to see that the limitation of UMA's slider aren't from UMA, they are from the designed basic rig. I shared once a video of someone going through the rig and finding solution to get better fine deformation. So if you need better deformation it's up to you to make your own advanced rig.

Now there is still small thing that blendshape can do better, but they are only a few of them. It's still possible to get a system that automate offline the fitting of all mesh to blend shapes. Basically you use the source vertex as "bones" and use a distance check to the nearest target vertex using weighted influence, but with the starting blendshape. Then you iterate through all blendshape of the source to bake the target's modified vertex position into new blendshape.

Blendshape advantage is that they allow crazy deformation difficult to obtain with bone, but given that naughty dog themselves bake blendshape back to bones for performance, we only need some of the unity wizard to replicate it for us lol.

Also in this thread someone had success using blendshape with UMA anyway, they aren't mutually exclusive.

 

This is a very good point. UMA is a base package but since it is open source, anything can be done to it. You are limited only by your own skills or lack of skills.

We find it great for our project. Unlike avatars that layer clothing and then hide what you don't want, UMA's hide mesh slots, helping with performance. That is important to us as we plan to have player characters and a decent NPC population. UMA's are fast and they can be optimized further. There is no problem with skin poking through the clothing, which was a huge problem with the package we used with another engine.

I really look forward to seeing what others do with the system. It is not locked down so if you want to add morphs or extra bones or whatever, you can do it. If you don't have the skills, find someone who does and pay them for it. At least you are saving money since UMA is free.

 

You need blendshapes for things that you just can't do with bones like increasing muscularity, fatness, skinniness, or shifting an UMA from male to female. Bones are fantastic for adjusting gross proportions, but they don't really provide the fine control you need to convincingly do those kinds of transitions--the "upper muscle" slider is a good example of a very unconvincing transition. You can modify bone weights to get some more convincing deformations--I've seen those videos too--but the workflow for adding lots of adjustment bones and properly weighting them is simply too difficult which is why Naughty Dog developed a tool to bake blendshapes automatically to a set of adjustment bones. The bones are more efficient, but building them by hand is just not feasible--you would need an adjustment bone properly placed and weighted for dozens of muscles all over the body just to get a convincing "muscularity" morph. Sculpting a blendshape is much easier. But I don't expect UMA will get a production quality blendshape-to-bones tool any time soon.

One problem is the existing UMA base mesh is is not really an average person. He is at the 99th percentile for height and weight. He's great for making mafia goons or action heroes but there isn't enough expressive power with just bone adjustments to squeeze him into a convincing regular-sized person. I saw someone trying to make him into a child on Slack with vaguely terrifying results. Blendshapes let you use a more neutral base model so you can convincingly scale back the muscles and get a greater range of characters (including unifying males and females into one race). Bones are very useful, but just not enough.

Adding blendshapes is not too complex; they are just per-vertex linear transformations after all. The only difficult part was getting access to the DNA values when the mesh is being constructed because UMA wasn't designed with that in mind. I had to do some jiggery pokery to pass through the DNA values to get the weights for the blendshapes, but I did get it to work ultimately.

The biggest disadvantage is that you need a base mesh for each blendshape, and you need to fit each item of clothing to each one. I am working off a set of five base meshes right now: a base (vaguely androgynous) male, femaleness, muscularity, fatness, and skinniness. Which I think is about the minimum set of base meshes I need to get a reasonable set of transitions, but I am leaning a bit toward ditching skinniness just to lighten the modeling burden a bit for clothing.

I've been playing around with them for a while, and I'm pretty convinced that I want blendshapes for UMA and I am working on them for my own purposes. Once I get something I'm happy with I have no problem releasing the base meshes and code, but I give no guarantees on any sort of time line. And it will certainly break most of the existing UMA resources without someone manually fitting them to the blendshape base meshes.

I could imagine a system of automatic blendshape fitting if you calculated something like a signed distance field for each blendshape, and then use that to morph each article of clothing. I suspect that wouldn't always produce satisfactory results though because there is a bit of artistic license when fitting clothes to different frames. Clothing just doesn't necessarily scale linearly with the body; it may fit differently on a very muscular or fat person.

So I am very pro-blendshape, but right now I think it is more important to focus on getting the next version of UMA stable and released before worrying too much about adding drastic new features. But I will be working on it once I get some time and get up to speed on the newer UMA features. Going to watch Secret Anorak's new videos today haha. Apparently, adding a new race is easy now? Last time I was playing with UMA it seemed like it was excruciatingly arcane.

 

Sounds to me that blendshapes are a nightmare for content creators.

 

I tend to agree, and performance has been an issue as well.

OTOH, if anyone ever gets blendshapes working in a more friendly manner, I'm sure we'll all welcome that addition to UMA.

 

Blend shapes are trivial to implement, support could be added in a few hours, they're just extremely demanding on content creators. Having better DNA converters in UMA would go a long way to answering the worst of the limitations of bones compared to blend shapes.

 

Hey, I'm new here, Morph3D has just been pissing me off to no end.. no customization at all in the long run =[

I'm trying to import UMA 2 into a fresh project, Unity 5.5.0f3, and getting 5 errors right out of the box.
Profiler non-existent, and transformModified obsolete. Are we aware of this?
Also 21 obsolete warnings.

edit: thanks Jaimi. Going through the tutorials now

 

Yeah 5.5 broke a few things. We have a new version coming, and it's almost ready (just redoing the textures, moving things around - which sounds simple, but there are actually hundreds of files that need to be looked at).

You can get the latest version here that works with Unity 5.5:

https://github.com/huika/UMA/tree/unity53

Select the "clone or download" button from the top right, and select "download zip"

Things will move around some, so be sure to keep your modifications outside the base folders so it's easier to merge the release version when it's ready.

 

Morphs would be fine if you are creating your characters and then freezing them, such as in a single player linear story game or a fps game.

But if you want a variety of clothing and armor, want your players to be able to change clothing, and you want to be able to make your own clothing, you will prefer the way UMA does it. Clothing has to be rigged and skinned to the bones in UMA and this allows you to change the size of the bones and the clothing along with the changes. Any clothing you put on a UMA will simply work. You don't have to rig and skin clothes to every body size/style. You would have to pay big bucks to get someone else to do it for you.

UMA may not be as good as a AAA morph solution you see in Naughty Dogs' games, but it works for us. There are other 3rd party solutions that could probably be used in Unity if you want to pay for them, some very very expensive. Or better yet, you can hire someone to make a system for you.

But UMA is amazing for the price. And it is amazing for its flexibility, and the ability to make your own content. I know UMA will stick with bones, but if they were to switch, it would greatly reduce the ability of the average person to make content and that would really ruin the package for me.

As for the size and appearance of the UMA's, I have only gotten compliments on ours. My team gave a talk at our local Unity User group and everyone loved the looks of the UMAs. Everyone has a different idea of what they want for their game. Style is subjective.

On the other hand, more bone adjustments would be a great idea. I like the idea of expanding the DNA.

 

This is fabulous! It would make a lot of people happy and add to an open source project. It would be nice as an alternative version as I have no desire to fit all my clothing to blendshapes. lol

This is what makes open source projects viable. Adding new code and resources to the project only makes it better.

 

I was thinking...dangerous, I know. I wonder if one of the problems with the inability to build the types of UMA's people want is the lack of good skins. Right now, it is really impossible out of the box to have a UMA without muscles. There is a smooth skin in there but it doesn't match the faces.

So, you can only create muscular UMA's at the moment. I am going to see what we can do about that in the future.

 

Yeah the number one real complaints I hear about UMA is that teh base mesh isn't pretty/generic/whatever enough, it's an aesthetics thing. The truth is that is UMA is a script solutin, I consider the mesh a placeholder to prove the concept. But people want good mesh too ...

 

I am "okay" with the mesh but would love more choice in textures.

With Secret Anorak's videos showing how to create new races, you guys could use any mesh you want as a base for UMA. You just have to find it or buy it or make it. Can't wait to see what you do with it all.

I plan to make a giant baby race!

 

We have a higher resolution head mesh now, and higher resolution textures (all thanks to Will B!). Hopefully this bump in quality will better show off the base UMA.

 

To be honest without Will B's assets I probably wouldn't have invested as much time with UMA as I have, I have roughly 1K lines dedicated to an Avatar system that essentially wraps around UMA that uses a slightly custom version of it. Right now I just wrapped up a 1 line equipping and unequipping system that handles overlays slots and such automatically. Hope to have my public test soon for everyone

 

Yeah, Will's clothing really does help make UMA much better as do his textures, some of with will be included in the next release. Your UMA's look great. You might try using the SSS skin shader that is on the asset store. It looks fabulous on the UMA's.

I have a few other videos using the shaders as well.

 

Mind linking the one you are referring in particular?

 

https://www.assetstore.unity3d.com/en/#!/content/42535

If you read through Will's forum post, he does talk about using these. His face backs are already set up for SSS and he includes the slots for SSS. However, if you don't have those, you will have to change the overlays.

Anything that overlays on skin, like underwear, tattoos and others will need a change. It is easy though.

 

Everytime I find something cool, Teila is there >_> I might have to start forum stalking you.
I had never really got serious about my armor equipping until now, and Morph3D was just ugly ugly. I don't need nor really want AAA super rendered quality, so I'm pretty happy with what little I've seen of UMA so far. There first review on the asset store is really negative... If someone else can vote on Gregorik's review it will move up... I just posted my own too. Anywho, I'll be chiming in from now on. Caught up in some other stuff again as usual.

 

Actually, the SSS shader was Will's idea, not mine. But thanks for the compliment. Followers welcome, stalkers..not so much. I have already had to get rid of a few of those. lol

Will vote and make a review of my own. Thanks for that information.

 

That was actually me, writing that review in early 2015.
UMA is obviously harder to get into than Fuse or Makehuman. But it's a gift that keeps giving once you figured it out.
I love the relatively new UMA stuff such as SecretAnorak's output and the Github developments.
What UMA needs is a strong community and actually released games that showcase everything that it is capable of. Keep it up!

 

Working on some very WIP in-Unity updating for UMA to allow us to easier push smaller updates faster without having to go through the store, looking good so far : )
Of course, we would also push updates to the asset store, but maybe at a slower pace than to the auto updater.

 

Yep! That is my plan! I did post a video of a prototype city with some UMA's and AI. Give you a feel of what a real game would be like with better art and a bit more room. lol

Until our game is done or at least enough to really show off, that will have to do. I have seen a few games here with UMA's. It is like seeing a friend in a game.

 

Can someone explain what DK UMA is? sort of a visual integration plugin for UMA?

Do we have an ETA for when UMA2 will work with 5.5?

Time to plug this into my start a new character screen....

 

They are working very hard on getting it out asap. I imagine this week or next but I can't promise as I am not the one doing all the work.

You really don't need DK UMA for the new UMA. Honestly, it is very easy once you figure it out. And not that hard to figure out. There are also example scenes with UIs that I used to make my own prototype. Even if you want DK, wait until the new version comes out and they have a chance to update it. I see people struggling with it right now with the Github version.

 

UMA 2 has been working with U5.5 for quite a while on github, so if you want to get started now, you can get it here:
https://github.com/huika/UMA/archive/unity53.zip
There are still some changes going in, so things might change a little bit, just so you know.

 

I have animations that I purchased to use with the UMA models.
But the UMA models are not aligned with the model the animations were made for.
So when using the more complex animations(2 handed staff or polearm) both hands don't align.
Because of this the staff will not always appear to be in both hands.

I was told by the creator of the animations the model avatars would need to be in sync at the T-Pose level.
But because the UMA avatar is created dynamically I was unsure how to make this happen.
How would I adjust the T-Pose of the avatars?