Unreal Engine 4

[Murdoc]

I don't think MegaTextures is critical for having smoothly blending surfaces and stuff like that, the only real visible difference it makes from a player perspective is the lack of repetition when it comes to textures. The actual blending is more down to the artists and what the engine's texturing facilities are like in general — for example in CryEngine it's always been possible to paint textures and smooth things together beautifully.

In fact CryEngine can already do global illumination and live changes to the code too (We all remember the same code running simultaneously on 360 and PS3, right?). I'm really looking forward to seeing how Crytek up the game with their next major release.

It isn't the only way, but it certainly provided artists tools, freedom and general performance gains to do so. CryEngine, even Farcry 1 can do this, so can Unreal, but the performance to do this on many surfaces hasn't been possible because its basically a brute force method... in both CryEngine and Unreal.

We'll get more ram, but to continue to use a brute force method is just going to eat up a lot of it that could be applied to other things, I'd suspect that is the reason why the tech demo didn't distinctively show this off.

Cool to hear about Carmack, hope his pull is still strong enough within the company to keep that (I know it sounds stupid, but they were going to license the engine out and are no longer doing so since Bethseda, so, I wonder with zero knowledge of what goes on)

[Frenetic Pony]

To answer: No, Carmack didn't patent anything, I don't think he ever does. Texture blending is fairly basic and has been around for a while, it's just an oversight on the part of the artists, or maybe wasn't totally ready for the demo or something.

The idea is that Megatextures could have solved the low res texture stuff, though it does bring to question other technical details to do with their lighting.

Speaking of which, Cryengine's global illumination is far, far inferior to what you've seen demonstrated in UE4. Of course Crytek's stuff came out two years ago now, and whatever they've done too it since then will probably be a big advance.

[Thrik]

Yeah, don't get me wrong MegaTextures is wonderful but there's no excuse for the UE4 demo looking quite as ropey as it does. It is after all just a tech demo though, so I'm sure we'll see some absolutely splendid stuff coming from actual games.

As for CryEngine, one of their guys said it's limited to one bounce right now for performance reasons but I see no reason that couldn't be ramped up significantly for the next generation — plus as you say, Crytek are unlikely to have been twiddling their thumbs since the last demos.

Yay, graphical advancement is getting exciting again. Longer shelf lives on consoles and console-orientated development has slowed things down so much.

[Roderick]

What concerns me is that as graphics get so incredibly detailed and the tech supporting these worlds expands into insane possibilities, it is becoming less and less accessible to newcomers. If I think about learning the basics of how all of these 3D worlds are built up, it's a daunting prospect. It has become a highly specialized job, where in the past the average person could at least get the fundamentals right on account of everything being a lot simpler (Duke3D editor simple).

Or am I a hopelessly incompetent non-programmer of a person?

You decide.

[brkl]

From what I hear, UE4 should definitely be more accessible than previous versions.

[toblix]

Yeah, I don't think accessibility is a problem at all, and it definitely doesn't have to be. Even though the game code and maths behind the scenes is super-advanced, it's always possible to encapsulate these into more user-friendly building blocks. Though I don't have any experience with these new game engines, I'm pretty sure it's trivial to make a sweet game today compared to what it was like five, ten or twenty years ago. Especially with the increasing competition on these sorts of development tools (Unity, CryENgeIn3e, etc.) the focus on user-friendliness is only going to increase. It won't be long before you can just voice-recog your cyber-unit to dev a sweet game.

[Roderick]

I think I didn't phrase my point properly, in fact, I know I didn't. The editors may be more accessible, but what I'm trying to say is that the underlying mechanics get so complex as to be incomprehensible. It's like computers themselves, I know how to operate them because they're user friendly, but I wouldn't know how to construct them or even what it is that makes them work.

[toblix]

Ah, ok.

[Thrik]

It's certainly a lot more complicated than it used to be, but it's dealt with by people only working on relatively small parts of a game rather than building an entire level by themselves like they did in the 90s. You'd be amazed by what budding artists and designers can come up with by themselves though, even if it does take them a huge amount of time — spending endless evenings figuring these things out is just what you do when you're passionate about something.

That said, the core way of producing things isn't that different. Artists mostly go about things in the same way they always did (albeit with more user-friendly tools speeding many things up and allowing for their vision to be more accurately realised), and programmers have always had it pretty hard as they've forever trying to push hardware to its limits while balancing highly-sophisticated systems.

I don't think there's anything to worry about, more people are getting into game development than ever from what I can see.

[Roderick]

My real question is: what is game?

[Murdoc]

I think I didn't phrase my point properly, in fact, I know I didn't. The editors may be more accessible, but what I'm trying to say is that the underlying mechanics get so complex as to be incomprehensible. It's like computers themselves, I know how to operate them because they're user friendly, but I wouldn't know how to construct them or even what it is that makes them work.

It also concerns me that it will bottom out the "big" industry. I mean look at what happened last time around, half the companies went under or consolodated under 3-4 larger ones and a lot of the jobs that would have been created were just sent over seas as outsourcing contracts. When it takes 1000 people to make one assassins creed on the 360, where does it go from there?

People need to start figuring out better tools for content creation, not just compiling it an editor.

*** When I said CryEngine and Farcry 1 can do "this". This was refering to texture blending. And by blunt force, I meant it's all housed in unique shaders(more or less) and is a pain in the ass to set up for every combination of intersections in most games and is expensive in terms of memory, gpu, and artists time. Hence the mega texture system alievates a lot of set up time and creates a standard memory overheard that is easily workable from the get go. Which is exactly the reason RAGE was gorgeous in its blends and every other game ever only has splatterings of it.

[elmuerte]

You're right Rodi. You used to have just artists and level designers. Then it became texture artists, modelers, and level designers. But right now it's texture artists, modelers, level designers, lighting designers, ... I'm probably forgetting some roles.

But then again, the number of LDs hasn't increased much. They just work on more levels simultaneously together with lighting designers, etc.

Back in the day I designed quite some levels for Duke3D (and other build engine games), it was really easy to create awesome stuff. I experimented a bit with Unreal1, but that was already too complicated for me (read: not within my interest to learn to use).

So UE5 will probably feature concurrent editing (i.e. multiple people working on the same level at the same time). iirc there is already a editor that has that feature.