Adobe Fuse CC Character in Unity 2018.2

Found a neat tool from Adobe called Fuse (it's been around, but it's new to me). Wanted to test a character export to Unity and check out how it renders. What I like is that there are toon characters as well for stylized character exports. With a little polish and some work on the materials and textures I think this could be used for The Answer is Never! character exports (saves me a ton of modeling and UVing time).

Here is the link to the tool for anyone interested:
https://www.adobe.com/products/fuse.html

 

Adobe loves their acquisitions. I remember when Mixamo first released it but I don't remember if I've ever tried it. I can't think of a case of anyone talking about it either but then most people just referred to Mixamo's products as "Mixamo".

 

I was very late to finding out about Mixamo, I do believe Adobe had already owned it. I was only aware of the online animation packs. I had no idea until very recently they had a character generator.

 

Good thing you found it though. It's a bit late but can be really useful you know

 

Yes I'm seeing that! I used to rely on MakeHuman for quick character exports that I could prototype and test, but this makes such nicer characters that can be made to look unique (not that they can't in MakeHuman). My only critique is that I had to break out the eyeball texture to it's own material and the eyelashes as well. In fact the eyelashes were the only transparent part of the entire body of the character, and yet the entire body has to be set to be a transparent material.

 

If you're still investigating what tools are out there, I'd also recommend Reallusion Character Creator 3

https://www.reallusion.com/character-creator/ (General info)
https://www.reallusion.com/character-creator/game.html (Game pipeline info)

 

Oh forgot one fun fact. It's on Steam lol
https://store.steampowered.com/app/257400/Fuse/

 

Last updated 2015, yeah another pointless aquisition at the hands of Adobe a company absolutely useless at doing anything involving 3d.

 

I have some Adobe Fuse Tutorials. (my avatar is a Fuse figure). It's a great base mesh. A lot of people use it as a base figure for UMA.

 

Just curious, what is the poly count and number of materials in that example?

 

FYI the mixamo team is on record in the adobe forums that adobe has abandoned this project and nobody is working on it or fixing any bugs. They say they may come back to it someday, but don't hold your breath.

 

@kdgalla I think the character is between 20-30k triangles if I'm not mistaken. It's a desktop model as it comes out of Fuse. However with easy-to-use retopology tools nothing stops these from being converted to mobile assets with minimal work.

@sinzer0 Yeah I agree this wouldn't be something one would expect any new features or expansions from. However it may live a while as users can upload their own content into it and modify that as well, so maybe the communities can keep it going. I simply like the fact this software spits out a decent topology base mesh, with UVs and a basic texture with the click of a button. After that you can use that base to build something truly unique out of it.

@wetcircuit Ha ha your avatar is really nice looking too, I especially like the unique hairstyle. I'll have to look into UMA and how folks are leveraging these softwares together.

 

Fuse basemesh is around 17k… Clothing and hair are extra, but the basemesh is designed to be cut away underneath to save poly (you don't have to cut away the body). Textures can atlas or remain separate.

 

Your tutorials are very well done, thank you for that link! I'd like to add that the topology is good on these models, so edgeloop decimations are not usually that problematic. You can bring polygon counts down pretty decently.

 

I found it odd that Adobe bought a program that was reliant on another companies texturing. I thought at the time that they may have their sites on Allegorithmic. Maybe now they have bought Allegorithmic we will see updates. Probably not.

 

well, the characters have color keys and are rendered and textured based on the same tech, I used Quixel Suite to render better materials, using the same color key, easy peasy...this was in 2017.1

"The Bad Guys"

quite...usable?

p-

 

When you refer to "color key" are you talking about something like an ID map similar to Substance Painter? I'm trying to figure out a good workflow for editing these Fuse texture maps.

I also have another question. The maps Fuse spits out don't seem to do the specular/gloss properly, also the normal maps looks strange. Are there good settings or "gotchas" to be aware of when using Fuse texture maps in Unity?

 

Yes, I needed better textures, some clothes I hand drew the maps based on the UV's, others were somewhere in the program directory. my main character's face was hand done from hi-res photo's. Time for a Tutorial!!

as far the textures fuse spits out, try this:
https://forum.unity.com/threads/scr...cter-model-into-unity-fixes-materials.482093/

From the same batch: Unity 2017, Character: Fuse

and the Upgraded version of the same Fuse Character..same model, just improved textures..yep...better.

p-

 

The skin substance in Fuse will need tweaking, the default settings are poor but the substance can do better. I export some helper textures by reseting the skin to white and manually creating a hair mask for a custom skin shader I made with Amplify Shader Editor.... I replace the spec/gloss for the skin, it's not procedural, it just outputs the same noise map over and over (it does some invert filtering on that map, but it's not realistic it seems to just be noise). I think Unity might have even inverted the gloss map since Fuse was designed, so you'll need to play around.

You can export just the textures. Maybe try un-atlased textures while learning what looks good because the Standard shader isn't always great about having a full-range of options (different transparencies for glass and hair).

 

It's a nice tool. I like the older version that can be found on Steam. Has a better variety and outputs better optimized models Just beware because the two are not compatible.

 

Oh believe me I'm aware. Discovered this very early on as I was testing both. I've found I liked the older Steam version as well. Seems to me some of the scanned models look far more natural in the older version. In the newer Adobe version it also appears harder to give a standard character a more unique look.

 

I'm not sure characters polygons are optimized for real time when your game has more than two characters on screen. This can still be lot of work if you need tweak polygons to get lower poly versions.

Can you use it for commercial purpose ?
I'm surprised there any no commercial limitations or royalties.

 

They are pretty decently optimized, they are not quite the resolution you'd see on characters in a next-gen game. They have decent topology too, so mesh reduction on these meshes has gone pretty smoothly for me.

I don't know the specifics on the license limitations, but considering it's an open source character building tool intended for developers to create characters for their projects...well I'd assume the license will allow for usage in your commercially shipped games. As for me, I retopologize these meshes and end up bringing them into Substance Painter, so I don't end up using the mesh they provide or even 50% of the texture maps, so I don't concern myself with that aspect.

 

UE4 is also free and open source but there is royalties when using it for something commercial.

Where did you found it open source ? Perhaps you are using the Steam limited version ?
This is what i read in Adobe site
https://www.adobe.com/products/fuse.html#x
"For a limited time, Fuse is available as a free beta to Creative Cloud members. "

I still didn't found the license usage description, but i don't want to download it.

 

The Fuse license is here:https://wwwimages2.adobe.com/conten...etou/Fuse_Additional_Terms_en_US_20180605.pdf

There are no royalties. Free to use in your game, but you cannot redistribute or sell the figures or animations as "content packs".

 

Fuse isnt open source, you may be thinking of MakeHuman.

 

No I just used the wrong terminology, it's free through Steam. Proper term would be free, not open source, though that would be really cool as the community could probably do amazing things to improve it!

 

I'm not sure open source without fundings would change something for Fuse.
MakeHuman is available from long time, it's great, it's free and open source, but i never seen major improvements or major additions.

 

I'm not entirely sure about that. I used to use MakeHuman for years, maybe in 2007 or so I started using it. When I look at what it can do today and how the UI looks and such it's a vastly improved product today.



It also appears within the last 5 months they've made some pretty big improvements as well.

 

2007, this is in some way what i said, it's still the same software without major improvements.
Big new features people need with a funding campaign or kickstarter would help i think.
Anyway, Fuse is better too much limited about the Steam version if you can't do models clothes and accessories.

 

Open Source is death by leaches. Manuel Bastioni LAB just died from an acute case of open source. But hey, now it's on github so maybe open source will save it….

This thread may need a reader caution for some mis-information.
There aren't that many figure systems out there, so do actual side-by-side comparisons. Use the demos. Download the models. Look at their basemesh and UVs. Even Poser can rig and export game figures – the problem is not the software it is their content, but once you know what you are doing the tools are in any major 3D app.

You want a system that is stable that you understand, for when you eventually resculpt the basemesh and create your own clothing/customization system. There are workflow issues that effect how you work with figures. like everything else in Unity there isn't so much a learning curve as an awareness curve.

There's never going to be a profitable "universal figure system" because within a few years you will level up with professional tools which are figure agnostic, or you will need to control aspects of the figure within your game so you will learn to create runtime systems which make the universal figure system obsolete. You end up creating the same figure over and over, using a custom workflow that you've adapted for your needs.

The more you create your own, the less "universal" you want it to be. At some point you are just picking a basemesh because the workflow is so customized to whatever runtime systems you use.

 

Unlike MakeHuman, Bastioni main goal was to reach some status he would start to monetize and make money.
He stopped because it was not making enough money.

This is very different from free software created without any goal to generate money.
For example Blender or MakeHuman goal has never been money ( while money funds they receive helps ).


Also many was not aware about Bastionni funding program , this was an issue in communication.
Another mistake, he did not used popular funding platforms like Kickstarter or Patreon, because this could succeed lot better.
I guess he lost interest and motivation to continue working on it, because he could try anytime a new campaign with Patreon or Kickstarter.

There is also characters generator plugins allowing to create characters in the fly using morph targets.
This kind of tool would help 90% of projects using realistic characters that does not need very unique character sculpting.
An game character generator in Unity also in game could be the next generation tool Unity will work on.

 

He was the original creator of MakeHuman, too.

Sure, but Poser does that, and has all along. PoserPro is half the price and 3x the tool. Morphs are not that revolutionary.

Adobe Fuse and UMA seem like a popular match. Fuse to generate endless humanoid seeds, and UMA to handle runtime.

Morph3D tried to control too much and now they have vanished with everyone's money.

Don't believe the (standalone) hype. Work with base meshes you like. The NextNewAmazing™ tool(s) will be inside Unity.

 

He was aspiring to make a monetized version of MakeHuman with Bastioni, perhaps to make some living of it.
I don't know if he lost motivation and interest, but he could still try it with new features proposal and starting a Patreon or Kickstarter.
(Some game developers reach 4000$ per month with patreon making tutorials,3D models and code).

I was meaning in engine character tool creation, not external software.

 

Wanted to share a couple quick tips.

To take your Fuse(any) characters to the next level, alway unwrap the head.

Shape your head/ face, close to the features of the pre-made photo, in Fuse. export for unity.
separate the mesh. head above the neck, merge the inner eye flesh, inner mouth flesh.(all the little parts).

Best keep in separate UV shells; Tongue, teeth, eyelashes, (any other mesh need to have separate shells.)

Why?..the face textures in fuse are bad..and creating a universal UV standard is quite handy.(see pic below)

Pin the UV Edges: top , left and right edges, edges around the mouth/ears.
Cut a seam in that edge and a little up the the center back.

Make sure your alpha has black around all the UV shells of any hair transparency.

now unwrwap the uvs..take a snapshot of those UV'a and export the image to photoshop. use liquify to warp the new hi-res photo to those uv shells. use diffuse to create new normals/spec/gloss, for me..Quixel NDO and DDO.

Use the photo for the facial hair, using these frees alphas!! (some work)
https://opengameart.org/content/hair-alphas-for-days

And just to learn more, about custom hair / alpha cards:
https://80.lv/articles/creating-hair-for-real-time-projects/

I'll probably make a video, after I finish.





now I can have...a lot.. of hi res, quality characters....


(some are free, but most cost a little..)@.3d.sk
more soon.

Patrick

 

I think a video would be great actually!

 

No, the project died because the developer didn't care enough to make more than a token effort at getting funds. If you read the comments and the link to the forum thread discussing it he made little to no effort to announce funding and he gave up after only one week.

 

The project had been listed as open source since at least 2016 when i first found it.

Maybe you are thinking he should have done a kickstarter? No, sorry but it was always open source. Nice try with the 1 week thing though. LOL.

 

No. I think he should have made more than a token effort at getting funding. That he didn't bother with it says to me that it wasn't crucial or that he didn't really want to continue development. Open source has nothing to do with either of those.

 

LOL! How many leaches used it and gave nothing back…?

(all of them!)

That's open source.

You seem to blame him for not being for profit. How does this work again?

 

If you wanted to get funding for a project from the community which would be the smarter approach to you?

Making an announcement to your core audience and as well as multiple smaller mentions during the funding process and keeping your funding process open for at least one month?

Or informing no one about the funding process and only keeping it open for one week?

 

LMAO! Do you need to be explained everytime you download a project that open source has to be supported by the community that is using it or the one person who actually does all the work will get bored, burn out, stop hosting a leaches-only project? I'm sure this isn't your first day online….

If you REALLY want to lecture someone, go tell M Bastioni that he just doesn't understand how OPEN SOURCE works – that he should have been clearer for the last 3 YEARS when he said "open source" that it meant USERS – and boy do I mean USERS, will scream and cry and pretend they had no idea and LIE to say he only made it "open source for 1 week". lolololol

Bro, you are making my point.
You can keep moving the goal posts all you want. The guy with the ball went home.

 

I haven't been moving the goal posts in the slightest. My original post was about the fact that he put in only a "token effort at getting funds". With your grasp of English this shouldn't be impossible for you to figure out so the only conclusion is that you're trolling.

 

One of us is still on topic.

Here is a visible, real world example why I am glad that Fuse is not open source. Open Source is not magic. Open Source is leaches who kick you after you stop giving them free things.

Adobe makes more money than some nations. They get the same insults hurled at them by leaches, but Adobe will not fold up to go live on a mountain. Adobe leaves the server Auto-rigger connected so Fuse still works. They do bare minimum maintenance but Fuse still works. It's half-arsed noblesse oblige but ok, I have a free Adobe membership and I understand there are no promises, but I also understand Adobe is not a person and doesn't need funding.

You are saying he did Open Source wrong. LOL. Sure, agree. What a dope. Had he managed it better everyone could still be leaches and he would still be the happy slave improving his plugin. What a fool he is. Poor him.

When this conversation becomes about you winning, and stops being about a long-term human figure strategy, I'll grab the popcorn. I'm just here for the open source cupcakes and weenies. When they are gone I'm out of here.

 

Like I said you're just trolling.

 

LOL

I think his decision to make a funding program came from he was already very bored, and he already wanted to abandon making it , he had enough of Bationni plugin and already wanted to let that and do something else

He though money could keep him motivated, but it's clear he has no motivation at all to continue working on it; because people told him and he could write new features presentation and start a patreon right away anytime.

I also don't care about Bastionni plugin could it be open source or not because it's not something i need.
Open source software is not always a great thing, most software i use is not open source and is commercial, because it is a factor of great quality and big features( Substance, Mari, Zbrush etc ...).

Anyway we're out of subject, the discussion is not about Bastionni or MakeHuman but Fuse.
Make a new thread if you wish talk more about other things than Fuse.

 

It doesn't matter people picked up the project (manuel bastioni) where it was abandoned while fuse is no longer updated by adobe, just like many other stuff like sculptris. I thrust open source more than corp by a long mile, I have been backstab too many time and I can't even reverse engineer BY LAWS, HEIL OPEN SOURCES.

 

Sculptrishas been bought by Pixologic and has been included in Zbrush as a new tool.
Zbrush has never been open source and it's the best sculpting software that exists.

So you should be using UE4 and not Unity LOL

 

Decoding: They bought a potential competitor to stiffle the growth before it became better that their main software, I was there when "drpepper" made sculptris and diffuse the news on tigsource and there was a huge hype around it from indies, it deflate as soon as it was bought because we already knew the legacy, the soft won't grow and the state of the art will remain stagnant. I don't see value in corp being big enough to kill market shift before they happen.
To be frank I'll probably shift away from unity anyway, i'm kind just there because habit and lack of time to invest elsewhere (yet). I'll just finish the current set of challenge I have to do then reconsider.

But we are out of topic anyway, fuse isn't having the dynamism I expected and it used to be, which is kinda the point. It's a space that interest me, it's part of the challenge I have to deal with (character generator) But I'll see that when I finish dabbling with the math of afro textured hair (rt and offline), which is a feature all these fancy character generator lack.

 

The author did not want to expand Sculptris with new functionality before it has been baught by Pixologic.
Sculptris was already stated as staying a small tool, itwas not a real challenger to Zbrush.
Perhaps Pixologic did not want Sculptris to gain some small market Zbrush could own, or they did not want the technology to be used by another company or some developers team that could make it grow, who knows ?

But Sculptris would never been a challenger to Zbrush, having used Sculptris a lot, it's not the same technology as Zbrush, it has the same issues as 3D Coat or Blender sculpting , it crashes a lot once you reach some limit polygons, Sculptris is good for concepting or rough sculpting only.
Zbrush is unmatched about highly detailed sculpting without crashing and memory issues, it's the fastest even on old hardware.

I don't find Sculptris tool addition in Zbrush interesting, because i need uniform and constant topology , i prefer concepting with insert mesh or Zspheres and remesh , it's faster.

Anyway we're out of subject for sure lol