I know a lot of people will come and say “you know nothing” etc. And I say in advance that I may well be wrong… but today reading some stuff, different links and thinking about, this came to my mind, based on some clues.
Roadmap has almost no more significant new tech left to be released. Some being finished, but it’s overall almost empty for the rest of this year. For the future, actually.
The skookumscript guys were hired long ago to work on a “new paradigm” scripting language for UE4 but there’s no mention of it yet.
There are each time more tools being worked on that allow the user to model, sculpt, rig, animate etc all inside of Unreal, which I personally always thought is the natural move - if you’ll end up putting everything together there, why not make all these tasks in final context, in realtime? Final lighting from the environment and lights, all appendages simulated, final shading etc etc? This also allows much greater interactions between assets like rigs from the start.
Unreal has been expanding a lot to supply new industries/areas other than game dev. Like product design, arch viz etc.
So, if this is not the case… what happened to the roadmap?
By unreal 5 do you mean a whole new code base? The codebase is so big now I cant see them building a whole new engine because they have to support all this functionality in a new version due to all the industry adoption. I think the roadmap has been like this for a long time now.
If you mean a new code base but with some similiarity to the current one (like Unreal Engine 3 to Unreal Engine 4), I mean, yes. But maybe a very different branch.
Yeah, I agree with you, that may be unlikely due to the burden to support both etc.
About the roadmap… yes, it’s been like that for a while, I agree, but it’s getting closer to completing all main features in the plans. So… what will happen after? Only improving on current features? Maybe, as there’s a lot of features already to be worked on. But it’s a little strange.
You see everywhere in the industry content creation tools planning or announcing features that use machine learning to aid in tasks, specially repetitive ones, asset fixing etc. But nothing here yet, aside from interpolation for raytracing samples, denoising. Correct me if I’m missing something.
I know many in this forum would welcome just that… a tighter focus on the current features and get back to game dev roots. I would like to see them split out all this enterprise / arch vis stuff into a different product and have unreal just be about game dev and get scripting in. Then off course split the stuff that makes sense in both products into plugins that can be shared.
I think the ML stuff for game dev is all experiments(gimmick) at this point but maybe makes sense for some enterprise automation or something.
I think if they were going to jump to a new version like that they would have done it already at this point, since there’s the new consoles coming. But since they’ve added support for the new consoles to UE4 it doesn’t look like that’s something that would be necessary. The main reason they made UE4 was that they needed to build something that wasn’t reliant on third party code so that they could freely distribute the source the way they wanted. They can continue iterating on UE4 so I don’t really see them jumping to a new version.
I don’t see the reason to move to UE5 while UE4 is so strong.
Also, unlike game consoles, game engines can be updated over time, like how UE4 has been with version 4.25.
It would make more sense to add to the current engine (and properly test it) rather than starting from scratch.
The thing I wished they’d do is cut out of the fat off the engine.
But I’m still curious: where are some new desired features in the roadmap that they could have added? For instance: for creating digital humans, as they acquired 3Lateral and CubicMotion and supposedly now can work together in much more integrated tools for this, integrated from the ground up, including thinking together about new tools, new workflow paradigms?
I don’t mean the engine is lagging behind. It’s actually very in the edge regarding tech. But things like those expertise + tools acquisitions leaves this question mark in the air.
Wow, we have a prophet in our forum. ^^ I’m already impressed with UE4 rendering. I would like to see what they come up with the new scripting language to fill the gap between BP and C++. That is when I’ll move to UE4 from Unity.
Again, I don’t see an emoji here for laughing intensely, as for Xodroc’s comment.
If this humble forum prophet can give 2 cents: my guess is that the language is less typed (close to python in syntax, structure or even python, but I believe they may create a custom tailored language, close to python so that people coming from VFX pipelines, Virtual Production, feel at home quickly) and used while in play (PIE, standalone). Modify the game while it is running. Maybe even some or all stuff in this new engine will be be done while in play… who knows? It’s a natural tendency for the future.
Maybe is even AI assisted (spelling, auto-correction), for a more natural language, but I think this is betting too far.
As for digital humans… I believe more stuff is coming. It’s a + b, if you think about the facts, moves.
Hopefully more procedural stuff for terrain (but with erosion style realistic results). Would be great.
The integration of cloth, hair, rigid body, destruction… all physics, looks like a great and smart move to me.
Now I’m wondering: what’s the specs of the machine the demo is running in? I mean… With lots more GB of data in the meshes being read from disk, with this new Directx tech (finally fixing those cumbersome and heavy, nonsense tessellation implementations), PS5 at least (I don’t about Xbox yet) is working a lot on fast and big SSDs, custom tailored.
Which kind of SSD would we need for PC? Like… recommended.
oh ffs I didnt know about that, there was nothing about there scripting language I even liked… all they had to do was support c#, well thats killed off that hope.
I see.
I actually liked it. As someone who learnt programming first with python and MEL, I liked the fact that the syntax is very different from languages I know, because it’s easier to remember the commands. I find it harder, for instance, to speak spanish, as someone who speaks portuguese, than to speak say english or norwegian. Or french. Because speaking spanish a lot of times I instinctively resort to portuguese words, because they a lot of them look alike, sound alike.
Some more whispers from the oracle:
as I said, an interpreted language, where you change code with the game running, and is heavily geared towards/structured to use with networking and multi-collaboration. Atomic transactions, pieces of code/objects that don’t break other functionality other “online”/collabo users are working on…
Digital humans: automatic rigging tools, in-engine, including automatic facial rigging, FACs compliant, with face customization tools, maybe. Auto body rigging for very integrated procedural animation.
In-engine modeling, sculpting tools.
Still regarding geometry, maybe arbitrary data stored/read from the the vertices, as now Directx Ultimate possibly allows. Good for Houdini users, for instance. Possibly they’re already working with Sidefx guys on this.
There’s speculation about UE5 releasing “when PS5 is released” floating the internets since before Fortnite launched.
Even *Paragon *forums already had plenty discussions about it, but the release date was earlier than expected.
Oh, man. I’m not going to get into this. Forums duels. Sorry. By the way, I’ve checked your products before. They’re all great. Incredibly clean and organized. Seriously.
One more prediction:
machine learning (+ facial recognition) + facial rigs, to automatically match rigged faces to photos of people.
