Its 'bye Unity'? (Character Creator - MetaHuman)

My appeal on the Unreal forums got results, the fight was huge, but and, it seems, it worked, at least to give (Epic) the beginning of the year 2022. They created what they call 'MetaHuman'.

Here's one of the hardest things to craft in games - and I look at games as a 'media' and not merely a 'waste of time' or 'slots' - characters. Far from getting rid of modelers (I'm one, I work crafting characters in Zbrush and Maya) but it's time for the game industry to take bigger leaps and increase the game market that, with each passing day, becomes more and more monopolized by rare companies, rare workers (it is one of the most layoffs in the world, ruining the lives and families of people across the globe).

My call is clear and objective, for Unity developers to facilitate the creation and implementation of tools that make 'making games' and something 'not only fun' but effective. It is time to take 'bigger steps' and with a view to the massification of this means of communication (games/virtual reality).

Well, indeed, it's not a topic that everyone likes, but a necessary one.

 

Metahuman, as far as I'm aware does not currently offer ability to load custom clothes.

Also when you're using a "Character kit" you'll risk all your characters looking "samey". The closest example to it is the fairly recognized "Daz Look".

Frankly, you're free to take the first steps yourself. In VR Alone ther are few widely open niches that nobody took. Also, "massification of the means of communication" is not necessarily a good thing. We have Facebook, for example.

 

Yeah feel free to go ahead. I would probably never touch it as it's going to be way too far away from properly prepared content for a game.

Put it like this: I reject pretty much everything on asset store as being not good enough or actually costing me more time than it saves. Anything that's community led is pretty much an instant nope from me.

This is just because this stuff is pulled in so many different directions by people of all kinds of skill levels, experience, needs and so on as to be effectively worthless in a real scenario.

To find out what's worth it in a real scenario requires a) money b) a live game - Unity has a but not b, so even Unity's own efforts fall short for years sometimes, before they're finally robust and practical enough to use, since feedback takes years if you're not doing it inhouse.

If it's not an actual game but a dirty internal project, then it doesn't count, and never will since it doesn't have skin in the game at all = almost no budget, and no meaningful results.

If the community is going to do this, then I think it would be useful for hobbyists.

 

If you look at Unity as an editor and how it progressed in the last... lets say ~5 years, do you really want Unity Technologies to make the content creation tools that you use to make your games? Really? From the bottom of your heart? Is this really the thing holding Unity back as an engine?

 

I disagree. I have no idea what you're talking about.

 

I don't think MetaHuman is that relevant for most forum goers here. That fidelity also has to be matched in your environments and your custom animations at which point you're likely dealing with photogrammetry and motion capture all of which are expensive and time consuming endeavors.

Could be nice for some archviz virtual assistants where a certain character creator look doesn't really matter or for large studios that have the resources to feed the system their own unique assets and data. Not really relevant for the regular indie imo.

 

I also agree. It would be better if people did great amazing and good things and gave them to me for free. I sign the petition.

 

Metahuman is far from perfect to be subject of drool.
Free Quixel Megascans, Hair Strands, Lumens, and Nanite however..

Emphasize on Lumens. I was just blown away with how much creative freedom I can have when everything works as it physically should without limitations. Emissive materials being able to light up environment and cast shadows in real-time. Bounce light and global illumination lighting up indoor scene as it should.

Nanite saving countless hours and trouble of retopo and LOD creation workflow for static mesh.

Meanwhile URP is limited to having 8 lights in view. (It was just recently raised to 16)

 

This isn't clear at all.

But also, reality check, there are basic things in Unity that are broken, its workflow is a mess split 5 ways, half the things are abandoned in favor of DOTS and DOTS is nowhere near ready. Like if Unity released its own Metahuman like system right now I think it would be the biggest joke ever.

What's the saying, walk before you run? Something like that.

Give it 10 years. If I haven't fulfilled my dream of buying out Unity and shutting them down by then, maybe they'll have a base set of features mature and stable enough that you all will be able to able to ask about Metahuman like features for your grandkids to enjoy one day.

 

Why does game development need to be so goddam dramatic.

 

It's just what passes for entertainment here before the next arowx post.

 

Being mostly a lurker here, I just wanted to say I loved this post. Sorry, a like wasn't enough to express this. Carry on.

 

Hi, how are you?

There is a difference between someone who loses a huge developer base and someone who wins: those who lose are always skeptical of change - they prefer, out of fear, not to change, not to innovate. Nor do I expect - apart from the traditional disrespect of this moderation - something like 'arguments to the contrary'. For now we will stay, me and my team, we are 8, with Unreal exactly and also for that reason.

Cheer

 

Hi how are you? I'm just a volunteer moderator, and not Unity staff. So please don't think that this is Unity's view point. Also, why are you replying to my reply that is for someone else? Good day.

 

I disagree. Creating a character by itself is not a difficult process. Creating a character and then setting up a proper character customization system that runs within the game is the real difficulty. MetaHuman only solves the former not the latter and it only achieves quality on the level of a non-hero character.

If it were a complete character system that could run within Unreal I'd be far more impressed. Instead it's just one of several existing character generator solutions like Daz 3D or Real Illusion 3 which are not only in a much more complete state but much more powerful too. If MetaHuman were paid it wouldn't be receiving the attention it is.

https://www.daz3d.com/
https://www.reallusion.com/character-creator/default.html

 

The most expensive part of characters like this is the fact you need a few hundred thousand dollars at minimum to support those photorealistic characters. Logically, anything like this will need a matching world like, this, voice acting, the entire works, that far exceeds the cost that such a feature like this saves. So much so, that this feature is only of worth to people making arch vis, needing stand ins, or crowds for films and things like that. But even then you can find plenty of libraries for it and it's never an expensive part of the process.

Hiring someone to even work with these systems and prepare them for a game (if you care about quality), is the same expense as hiring a talented generalist making bespoke characters to order.

It's a great thing to have, IMHO - serious now - but it isn't saving anything nor being revolutionary. It's going to be a narrow band of use cases where it gets actually used and succeeds in creating profit.

 

Making a character for me is absurdly easy and simple, I do one every 2 hours in Zbrush, Maya/Wrap/Substance (only the character - no hair, no cloth). I've been working with 3d MAX since version 2.5, most here weren't even born. I'm an old guy, for the last 30 years, I worked on TV in computer graphics, I graduated from two Universities, one of them, Bachelor of Piano (music is also something simple and easy for me).

I'm not interested in what's in the 'store' (I wonder when a moderator puts it as a 'principle' when it comes to, theme, it's the engine tool and not 'what I'm going to buy'. I don't understand this mercenary logic and I don't even want to understand- there.

My idea was to continue my 3d point and click, with the help of engine tools and not purchased... I also think of those who don't use the dollar as a parameter... well, I won't even enter this debate either.

I know something, who doesn't innovate, loses. Unity is an excellent engine, games like 'Geshing', 'Disco Elisyum' and so many others seem to have a unique stability if made by Unity, I just challenged the community to question, to pressure developers to make more tools to automate the work of small businesses, makes them competitive. But, from what I saw, the 'absolute absence of will, of desire' rules here... I like games, do you?

 

Vroid is made with Unity. Please.

 

Early acess now (UE5)* its planned to 2022. MethaHuman, maybe 2022/23.

 

Well, this will be my last post here, after that, nothing more to say - if you were any member of the community, you would understand, but coming from a moderator... no.

Good luck to everyone, I wish you all the best.

 

Complete character includes rig, painted weights, retopo and ideally character specific animation set. To make it a game character you'll also need animation controller, sounds, collider and physics configuration.

This takes far more time than just 2 hours.

Also this an interesting distinction you're making here. Buying in store is a no, but using an automation tool, which produces a result that is not made by you is somehow okay. Why is that?

Community isn't the one making decisions, and even if unity were to make an automation tool they'd probably spend at least a year to make it half-usable. We also already have UMA on store, for purposes of customization.

Regarding competition ... things come and go. One day unity may cease to exist, or their interests might drift to a new direction, where they will be no longer compatible with mine. Then it will be the time to move on, and this is fine. As it is just one more tool, and there's not much point in getting attached to it, especially when I'm not the one calling the shots and making decisions.

Regarding MetaHuman, the problem with that tool is that it is nowhere close to being done. Its only interesting feature is ability to lipsync with data from iphone. Sculpting isn't great, and there are no clothes and no body shapes.

 

"Complete character includes rig, painted weights, retopo and ideally character specific animation set" - wrong, no animation is needed to who work with MODELS (in general, have a team to animations). About RETOPO: wrong again, try use Wrap 3d, a Russian tool, very good. In this environment there is no secret, there is research, study, a lot of practice. Sometimes I spend an entire year just building tools or a set of tools for something specific. Over the past two years, I've found ways to make characters - rig models - in less than two hours. But watch out, I didn't mention what kind of character, obviously characters that demand more complexity, would take more time, it depends a lot on 'what you want'. An 'Alien' style character, for me, would take more than two days... that for me, my facility is in humanoid models, exclusively. Another thing I found concerns textures... a stylistic type which I will use in a long-term personal project, something I don't want to comment on here and it took me a long time to find the right tools, customize them, change them, so to speak , to serve my purpose.

 

There is a guy who exhibits his work at Devianrt, he works with 'armors' and found a way to make them in his own style using a little known tool, I tried this tool, looks like Zbrush, but works more with ' hard surfaces'. I forgot the name for now, I could see it and post it here, but this tool is something 'indie'... it's called 'Teya Conceptor'. Apparently, little investment in it and I find it fantastic, intuitive.

This is the way, intuitiveness to break the barriers of time and production techniques. This will happen regardless of 'wants'. I just wish it existed for everyone's good and not to be used negatively (fire people from their jobs - and technology unfortunately is used that way these days). But, that's just my wish.

EDIT - * Just to point out, he puts on or exposes a new 'armor', static model, once a day, fast, no?

 

What is your point, really?
First you speak about efficiency, then spending a year to make a tool.

I can make a character in 4 hours. That's from nothing, fully rigged, with animation loops and I'll setup the rig by hand, every bone. It is this slow, because I'm not sufficiently skilled.

However it will be low poly reduced to a bare minimum to make the point and minimal animation set and cutting every possible corner.

Spoiler

So, going back to your claims, if you're producing highly detailed models at speed of hamburgers, then you have a prototype mesh that you slightly distort and the models you produce will be a slight variation of thereof and won't be on the level of, for example, Witcher 3 Geralt.

Spoiler

Because when you need high detail, but keep cutting corners, at some point there will be no more corners to cut and still a ton of work to do.

Which reminds me of a werewolf story. A guy I worked with in the past said that everyone figured out how to make quick human and have them a base mesh, but give them a werewolf, and they screw up and can't do a thing. Which makes me wonder why people aren't studying more fundamental techniques that allow them to model ANYTHING instead of just a variation of a human.

Now, going back to the original post.

What were you trying to say?

 

Do you think big companies make 'things from scratch'? Who is being naive here?! They used to, that doesn't exist anymore in big companies, they developed their own tools by deconstructing others - simple as that, although they swear or claim that they are 'original' things (this doesn't exist in programming - even it is the result of mathematical logic, something which has existed since antiquity). But I'm not going to debate about it here on this forum either.

Unity needs automated tools - or more automated ones, like 'metahuman', 'nanite', 'lumen', etc... as simple as that.

(Interpreting what a third party says is at your own risk)

 

How is that a reply to anything that I said?

 

Do you claim to have ability to read minds, or something?

Why would anyone care what big companies do?

Big company means big budget. With sufficiently high budget you can simply buy a company with technology you want or order things to be made.

Spoiler

Actually LA Noire comes to mind where they made a new tech from scratch.

Now, to use big company strategy, you'll need a big company budget. And if you have a spare billion in your pocket, what are you doing here, arguing?

Unity, in my opinion, needs to get their S*** together instead. They've invested into many technologies since version 5, and all of them are half-broken, unfinished or poorly supported. Adding one more piece half-broken tech won't improve anything.

 

He got a point, this is the reason why Unity exist back then. Providing industry quality tools to close the gap between indie and big companies in terms of quality, accessibility, and production efficiency. This is why they made tools like NavMesh, Shader Graph, VFX Graph, etc.

Though, Unity's business model have changed since then. Their whole purpose used to be in creating tools to create successful titles to their users. Now they no longer have to starve if studios fail because they've opened up various other source of income such as ArtEngine, Multiplay, Vivox, SpeedTree, etc.

I theorize this might also be a strategy from Unity to stay competitive in terms of resources with the competing engines.
The recent acquisitions and diversification strategy is directly related to the recent Unity IPO. With the goal of focusing on creating a good income statement so they can gain a lot more investors and appeal more to various industries.

So far stocks have been strongly separated by their categorization, Mining, Agriculture, Property, Goods, etc. I argue this is Unity's true hidden edge. A very robust and flexible tool that appeals to almost all other industries such as Film Entertainment, Construction & Property, Military, Gaming, Goods, etc.

Even if their interests have to now be divided among many other interests, its a good expansion strategy with the cost of slowing down their game-side of the engine for a year or two. Once they've gained enough resources, Unity might strike back with full force against the competing engines.

I can only hope Unity's true end-game goal is to use all this capital gain for the game-side of the engine.

This is just a theory though, a game theory

 

But modertators here are literally just "members of the community". The forum is heavily targeted by spam bots and the official Unity community managers don't have the time to monitor the forum around the clock and delete that trash. Having a couple volunteer moderators living in different timezones is a good solution for this problem.

I'm sorry, but I'm having trouble following your train of thought. Feels like I'm missing some information to even understand why you want Unity to make a character creator like metahuman. Especially since you are so fast making your own characters and tools. Wouldn't you want to keep that competitive advantage, since clearly most people take a lot longer to make characters?

But I would be super curious to see how far a pro like you can push a new character in 2 hours! Please show us some screeshots.

Correct me if I'm wrong, but didn't NavMesh lag far behind A* Pathfinding on the assetstore and wasn't Shader Graph only developed after Shaderforge and Amplify Shader Editor had been in use for a long time? And didn't forcing/urging people to use shader graph cause lots of problems for advanced custom shader solutions like microsplat? I thought back then there were very convincing arguments made that Unity should focus on opening up those blackbox systems and allow more solid access to internals, or adding low level tech features like texture arrays, so that assetstore devs with more hands on game-making experience can develop solutions that are better fits for specific usecases, instead of Unity making generic solutions that don't fit any usecase really well?

 

I would expect Unity to expect, that as a moderator, you show some respect and act as an official staff moderator would, but maybe not...

 

He hasn't got a point, because he's talking about big studios.

Big studios often use inhouse engines. Frostbite, anyone?

That doesn't seem to match reality, though.
Visual shader editor appeared in Unreal/Substance first, and was only available in unity as an addon.
NavMesh isn't great.

What unity had was an incredibly clean well documented api. The quality of documentation has been steadily going down recently, though.

"Start implementing a bazillion tools, but stop halfway and make sure that none of them ever get finished" doesn't seem like a sound strategy to me. It does sound like a good way to frustrate users and make them leave, though.

 

Unity's goal was to democratize game development not provide industry quality tools to everyone. Making game development accessible to everyone does not automatically imply a certain standard of quality and in fact nothing about Unity for the first decade plus of it's life was anywhere close to being on par with the industry standard.

 

Stop me if im wong, but...i do understand your point (well taken). But that doesnt mean that MetaHuman is going to be a "breaker" for anyone.

Ive seen tech like this arrive, and disapear. Is it good? yes... its a first of its kind, and others will develop similar software, and will be avalible/integrated into Unity AND UE. Only thing you can do is sit back...and see what we have to work with...simple...

 

Metahuman still the best high quality source for varied non white rigged animated character head, which is sad in itself when you make a game not based in northern country, worse the quality is higher than in high production game that can spend time sculpting individual characters.

 

While I think creation tools like meta human and speed tree are inevitable, there are 2 big things that stop me from drinking the kool-aid.

1) Unity is still king when it comes to bringing games to market.

And 2) I absolutely think there’s going to be a huge watershed moment towards the ever elusive “make game button”, but that is going to flash out in less than a couple years.

People will want new experiences.

A “make game button” can never do that and you will have a lot of pretty games, that are all the same and people will get sick of it.

 

A new version of RPGMaker effect. Or Daz effect. You see the first game like that, it is cool. Second one, the third, the fourth, and then you start seeing similarities between them.

 

Never thought about that, but YUP!

RPG maker games are all DOA because of his this.

 

Still coming out at a moderate clip it seems. And apparently they're less likely to get Metacritic scores, but are among the highest-scored (grouped by engine) when they do.

 

Well, doesn’t that kind of makes sense? When you’re using a tool that makes all the games the same you really have to stand out.

 

Suppose that makes sense, as the RPGMaker/Gamemakers games that are good generally tend to be VERY GOOD (Undertale is an example), also catering to a niche that clamours for quality in their genre.

 

Game maker is a little more open though.

Hotline Miami is another jewel in their crown, but I can’t see that being made with something as rigid as RPGmaker.

 

The game that put rpg maker on the map, critically and commercially, was "to the moon".

A make my game button will just shift the struggle elsewhere, not in "execution" but in "vision". I have noticed in my research, that resources didn't matter much on the length of making "deep" works, it always seem to take 5years to 10years, "vision" being the main indicator of progress. A lot of works are derivative, because it's a whole lot faster to use a template (example genre) where most decision are made, than coming with a structure from scratch, then things can be churn up from 1years to 3 years cycle for complex works where you still have have a minimal amount of differentiating elements from the model. Most of the time, dev time of the original template are deceptive, they ted to be thought out and worked way before execution actually happen, and sometime mandate to create something news just fail under the weight of decisions. Movies for example, use the cheapest form of execution as "scripts" revision before committing to anymore costly decision. I don't think a make my game button solve that, if anything it will allow more exploration of structures to get a vision down. A lot of genre has been invented or diversified since the indie revolution, mostly due to the advent of cheap game engine allowing experimentation, they lowered the execution ceiling, but making a noticeable game is still hard. The cycle of trashing what don't works just get longer everytime you lower the cost.

 

Also by the way ...

Make my code button is outdated, just talk to it to explain what you want

 

Again, being a moderator has no effect on people's opinion. Being a moderator is simply cleaning up spam, helping people find where to post, answering questions. It is not a higher rank. And it does not shut me up from having my own opinion

You mean are still in the process of making while tiny indies outshine them on their own asset store. Frankly, VFX is the only one in this group that is ahead.

Sadly haven't been using that phrase for several years now as far as I can see.

 

Yeah, exactly. I have a few RPGMaker games in my Steam library. Looking at a game's store page, you can quickly tell the difference between a game like To The Moon that has something interesting to offer vs. the generic default-asset type of shovelware that people crank out.

 

Perhaps because they already succeeded at that? Anyone can get access to the tools now if they have a PC and a 'net connection, and not just from Unity but also from the competitors who followed suit.

Personally, though, better onboarding materials would help on that front. Access is a start. Next, people need to know how to use stuff.

 

I remember a time where we could in good conscience answer all the beginner threads about where to start with "go to https://unity.com/learn to get started", but I'm seeing that answer less and less.

 

At some point I just started pointing people towards a reputable YouTuber like Brackeys.

 

By the way it is possible to design characters in Daz and bring them to Unity.
Isn't that like MetaHuman on steroids?

 

Daz models are subject to rpgmaker/daz effect. Meaning they're recognizable.

Additionally, I do not recall Daz having any ability to create any clothes. Meaning this cool mech/goblin/orc is a shop item, and another person who bought the character on daz would have exact same thing in their game.

One option worth investigating is MakeHuman, since they seems to have dealt with their licensing insanity. That's partly because MakeHuman has means to import user clothes and refit them to different models, though the mechanism itself is so arcane that every time you do that, you have to offer players Yog Sototh first.