Discussion Unity's another shiny new demo

The Lion
Featured Speakers | SIGGRAPH 2022 (scroll down to the end)

 

Wow. Cool. Another demo. Yay.

At this point I've seen so many demos from Unity that never amount to anything I'm wondering why I should care.

 

Here is the new physics simulation tools previewed https://dl.acm.org/doi/abs/10.1145/3528223.3530058

I'm eager to see new realtime GI solution. Maybe they'll show it at Siggraph.

 

I am not going to join the dissers gallery and instead say..cool.. Looking forward to being able to leverage these technologies.

 

When there is no demo, they will cry
When there is demo, they will cry

 

Nearly all the "big" demos so far haven't been representative of the core engine capabilities and were highly modified to look good for marketing purposes, but didn't really benefit the engine or its users much. There was some cool tech there, but it didn't really go past the demo projects and wasn't kept up with new Unity releases.

Then they announced GIGAYA with a big fanfare at GDC, the first demo that would go 100% native without source or other deep modifications and would benefit the regular engine users like us and they cancelled that.

Now they have a new marketing demo, clearly aimed at movies and TV, likely with minimal benefit to game developers. Good for Unity I guess? But I'm not excited either. It remains to be seen if any of the Weta library of assets they acquired will be game ready when all of it was crafted for movies rendered in GPU farms.

 

I didn't care when they trotted out the subject matter. I am so fricking over the morbid and morose mindset and subject matters. What ever happened to beautiful? I am more impressed by the rayleigh scattering of a sunset than by an army of photorealistic undead transhumans or some battle scarred antique street punk in dragoncon drag. It reeks of puerile hail satan wannabe vibes. I hope to be impressed with the muscle tech that will give running big cats the muscle flex and body mass shifts of a real world big cat. That is beautiful and truly brutal power.

 

....

Lions are cool.

 

Literally nobody worth listening to gives a S*** about the God of War rip-off. It's just an attempt to relitigate boring garbage like the "asset flip" discussion.

 

I honestly don't care.

Lions are still cool.

 

I still don't care.

If unity dies, so be it. If unity doesn't die, so be it. It doesn't matter, I'll roll with either outcome.

The attitude on those forums, however, in past month frequently becomes hysterical to an unhealthy degree.

And that's why I think that "lions are cool" is a better approach. I'm just tried of reading yet another text where somebody is freaking out over something.

 

So then we'll see some fruits of Ziva acquisition, Ziva VFX (2.0) to be precise. I guess we have seen Ziva for face contruction previously but that wasn't Ziva VFX.

 

People make ripoff games with all sorts of engines. This is a non-issue.

 

Except the God of War ripoff you brought up yourself has literally nothing to do with this and is also literally nothing new. Why are you so intent on derailing any discussion of what's actually being discussed here? You only registered your account today.

 

"We have a nice human face construction tech"
"This is a lion"
"Things happened"

 

Mmmm déjà vu, they "stole" the Weta from unreal, in which they already demoed the meerkat there. Weta using its lion king back asset I see...

Unreal is really the future of unity, lool back a few years there to see what's coming to unity.

Not impressed by copycat.

How did you leverage Unity's the heretic demo unsafe code to put beard on mesh?

 

What does any of this have to do with this:

 

I could not possibly be developing what I currently am and have been over the last half year in Unreal. With Unity I was able to tell them truthfully it was possible. Unreal is not the big schtick you seem to think it is. I don't care if Unity doesn't add features that are mostly used by the FPS/RPG crowd. I have not had a contract for one of those in my career of freelancing with Unity since 2010. MakeHuman was MetaHuman before MetaHuman was Unreal. Add shaders and better hair/ skin maps. I could not have controlled a whole stage of animatronics ..5 humanoids, four jig booms, a dolly and pulley crane and DMX lights with Unreal..on my own..in three months time without Mecanim Retargeting and Unity's massive knowledge base. Not impressed by samurai sword. Doesn't do whittling or pop a beer cap well without the distinct possibility of cutting yourself. I would rather use a swiss army knife and my brains and talent to win bacon for my table.

 

Take some time and watch the NatGeo series Savage Kingdom. Just the narrators voice is worth it. The strategy of each animal group is brutal and necessary.

 

Nice screenshot! Id be curious to see what the video looks like.

 

Oh heck no! I'm just in the middle of reevaluating current solution, makehuman is an overall agender ageless character creation, with very bad expression morph, metahuman skip the whole deep character creator and offer a full workflow based on mostly the hardest part of doing character, good facial expression, and if you wanted more precise and define character, mesh to metahuman is a very neat way to retopo, transfer blendshape and animate, that insert itself in any pre existing workflow. It works even with sloppy phone scan, to build very convincing.

Technically they don't use blendshape, unlike unity for the heretic, enemy's situation pending, because that's not optimal for performance but I'm a nerd I had to point that.

That thing is a marvel of razor focus tools that doesn't solve what didn't need, while removing pain points. It's shame it's locked out by license only, it's engine agnostique so that's a low blow.

But I'm not gonna lie, mecanim retargetting was the metahuman of its time, sharing animation across character's varying humanoid skeleton was unheard and remove a big pain on animation. Though now even goober do the same with the democratization of the concept of ik rig, that generalize beyond humanoids, that's good that unity also offer that.

Anyway, when you make something accessible and affordable, here it's unmatched quality for the ridiculous price of nothing, it's bound to create new opportunities never seen before as people start experimenting. Like this.

Who will win the real time rendering race? Unity with complex inaccessible rig? Or a goober in his room with meta human and other free software? I mean I have seen people using dalle 2 to composite matte painting stuff were contact shadows and reflections are added faster and better than any path tracer in existence, artificial voice acting are becoming a thing that's getting better and cheaper, like mod using xvatrainer, time just ticking down before a whole industry upset.

I hope unity is catching this train with their ML team because the whole concept of enemy like demo might become obsolete soon.

 

i dunno why isn't unity making dozens of generic templates people can use to get their ad-riddled phone games made faster. Thats what everybody uses the engine for, don't they?

could be tons of sample projects that just have basic framework set up for all the common phone game types so that any 10 year old could just plug in art and a few scripts and be on to earning that ad revenue.

Is anybody using unity for film stuff?

Heck, I'd still have unity hub installed if there was a sample project available for sudoku, match 3, tower defense, tetris-styles, etc. That would be an awesome resource for any game devs starting out.

 

I think they actually do, their actual origonal game projects are all small mobile game clone, like the infinite runner cat, and they have a whole service dedicated to implement common gameplay used in these games.

 

No. I am building a content creation application based on a tech stack of machine and deep learning, voice and natural motion synthesis for a range of industries. A tricorder kind of intelligent agent that may or may nor have an avatar as it's human interface.

 

Do you have a portfolio of works you have completed? Are any of your lectures at SIGGRAPH available? Where are the games you have published linked at?

I see you sneering at WETA technology. To me that is funny as hell.

 

I'm not sneering at weta at all, they produced both the unreal and the future unity demo.

Though your criticism still hold as I'm targeting unity (current) direction, i have nowhere unity achievement. You will be free to roast me when I released my low tech version of whatever they are doing anyway, this time that will be good sport. Doesn't change the fact that unity is both well placed and seems misguided.

If you read carefully you will have see I pointed at how they have the resources, talent and feet firmly set to take over (me mentioning their ml initiative) but they can still be xerox out of the market by not being pro active and wasting those talents on 5 years old implementations passing for breakthrough. I mean you are still free to see that as a sneer, you are in a free country, and that mean everyone can make their own little truth.

 

https://github.com/Unity-Technologies/com.unity.demoteam.hair
The hair technology is also out now

 

Oh that was the highlight of the demi, pretty sure the lion will make great use of it. Seems like an evolution of Boston.

I want to precise, I never put in doubt the quality of the work made in the demo. It's just not a product for the game engine and the positioningis weird, that's the criticism.

 

Not sure about the movies, but commercial (as in business to business/non-gaming) Unity application is a thing. It's out there, it's used but you don't hear about it much because it isn't that exciting, however Unity tried to bring a bit more attention to that in the past (https://create.unity3d.com/unity-forma-contest).

Asset Store offers all of that already and that market is not as uniform as you might think.

 

Fair enough. Your good rep with me will remain untarnished I work on very cutting edge stuff often and know that usually there is one guiding vision and alot of helping hands. Folks can't just sit down at their workstation and blaze through. This often takes staring at the wall..so to speak..for a long time to produce a burst of code that does the trick within the framework. There is often a build up of code functionality that when the goal has been achieved needs refactoring into a more compact form and hooks made accessible that makes sense to give developers access o those tools. I am gonna give the guys at Unity a break. They got rid of alot of deadwood recently..much to the woes of those who do not understand business and the bell curve..

 

If by "out now" you mean dumped on github and not part of the core Unity tech stack, supported or officially on package manager... then sure. But I think we've done that rodeo a few times. I don't see any guarantee of sustained development as a feature.

Probably go the way of all their other github demos. Some work. Most don't. A shame, since the work to make it looks fantastic. The author deserves to be proud. The movement and rendering is superb.

But I definitely would not use it in a game, only movies or such, because of the lack of definable support.

 

Dude this looks sick! Hyped!

*goes back to seeing if Godot is a viable fit for me*

 

So there is not even a thread on forum anywhere to discuss how to use it? Just like for the heretic? Not even a technical blogpost? I feel like some talented people hijacked unity to make their dream real time realistic character happen, and just do the bare minimum to continue get the finance for it. And nobody's really looking at the code heavy repo, because artist are the one interested and are allergic to that, and programmer goid enough to get it just go to the source paper to do their own implementation. And it's not really documented. And there is zillion alternative already that have real workflow and the benefit of whatever improvement they did isn't relevant because obfuscated.

 

Yeah I'll pass. Give us a game and we'll talk.

 

The difference between Unreal demos vs Unity demos, is that Unreal actually demo tools of the engine that are accessible and you will use (think of Metahuman), where Unity demo things that won't be part of the engine and you can only build in a AAA environment. Think of Heretic, Adam and absolute every Unity demo in the past, only assets can be recovered now, not the tools or systems built for those demos...
Enemies hair tools dumped into github that means it will get forgotten in limbo in a year or so. Sure HDRP team is working on a high end hair shader but will the hair tools be part of it? I pretty much doubt it.
At this point, Unity should just focus on DOTS everything and then build upon that.

 

I'm still going to study them, we will see if I can make something out if it. I have been looking into hair solution for sometimes already, and I haven't checked simulation yet, so that will be useful, but I'm on low end so this actual implementation is a definite no go. But maybe the principle will help realize the idea I had about texture space simulation to complement bone physics on hair. At least the reference paper will help too. Assuming the idea produce good results.

 

https://unity.com/demos/enemies they just edited one line from the previous blog
I think that the technical talk would be at SIGGRAPH

 

That's maybe true for the graphics side of things but on the coding side, the thing that was most beneficial for my projects was their ECS Mutliplayer-Demo Project.
For this, they created
- GUID_Component, a unique ID component that works for any scene object and can be used for SaveGame, Cross-Scene-References as well as references to Checkpoints and more.
- Material Property Override, let's you alter materials, based on PropertyOverrides written to the renderer. You add the component and then you can change the texture, material or any value from the shader for this one GameObject without creating a new material.

Unfortunately most demos since then were only(?) graphic demos. We need them to make more games, then we would get better tools.

I consider GUID_Component and Material Property Override essential tools, which greatly improve the editor as a whole.

 

I'm not sure if the demo will be shiny (how shiny is a lion, really?)

That said, from the outside this looks more like a "Wētā made a demo", but by extension now that is Unity's demo. If the demo actually shows how Wētā tools integrate with the rest of Unity tools, that would be great! Since right now it's a bit of a mystery.

 

Please, just make realtime, no baking, light probes based GI and improve SSGI in 2023.1.

 

This thread is not the place, make your own thread instead of jumping on everyone elses asking for this each time.

Also, aras does not work for unity anymore, so asking him is completely pointless anyway.

 

https://twitter.com/the_f_key/status/1557286750191591425?s=20&t=qs4PyOjAQ1y7KGFaHXmMfw

 

It's kinda weird that Unity obviously paid the devs a lot to develop that (it surprisingly even supports the built-in pipeline!) but yet they do not firmly decide to bring it to an end. In other companies that's decided right after a proof of concept..

Well, on to the voting booths then xP

 

I know opinions are tinted by ones point of view, but are people really that excited for a hair and fur system?
Of all the things I’d like to see, this is far…. faaar in the back.

 

I would be excited if Unity had a plan of supporting fur and hair with card and mesh generation, and that plan was supported as a baseline engine feature.

Otherwise, not in the slightest, no.

Edit: THEY HAVE A PLAN 2023.x

 

Because they built it for a marketing demo. Very few of marketing demo technologies go back to the core of the engine. This system has the same fate as most other Unity's GitHub's dumps - use at your own risk until it inevitably becomes out of date in a year or two.

https://github.com/Unity-Technologies/com.unity.demoteam.hair

And arguably one of the best parts are only available in 2023 alpha which is where Enemies was likely rendered. So you'll maybe be able to use it in a year if you're comfortable using tech stream for production?

 

Partially but not purely. For demo it wouldn't need this many modes and support of all 3 pipelines (whichnonly top notch variant being newest hdrp).

Guess it will depend on whether people demand it on their roadmap feedback feature.

 

Hair simulation tools are on roadmap for both HDRP and URP

https://portal.productboard.com/8uf...ion?utm_medium=social&utm_source=portal_share

 

Really this a demo for specialists not users. Grab it, learn, then make an asset on the store.

I'm going to look at it to see if I can adapt to make afro type of hair, which has almost no real research done on them, I don't expect it to fully support that (curl variations, locking strand effect, stress effect and spiral perversion aren't seemingly part of their model at a quick glance), but I can build on it, it's still a very good basis on many things I'm not as verse as this is.