Just a thought

[London Geezer]

What worries me more than anything else in the next generation is how on earth will the artists keep a continuous flow of *good stuff* to be included in the inevitably high number of next gen games.

At the end of the day there WILL be coders that more or less take advantage of the architectures. What will not be there all the time will be someone who can make all that power into something that doesnt look like some cheap CG TV series.

If the next generation of hardware will be technically capable of producing visuals to rival FFX FMV in real time (and i think it will be close), then very few studios will have the resources and the imagination to come up with art like that. I think we will just end up with many titles with cheap "bumped up" graphics and the number of truly great games will be even smaller than it is now.

As much as Sony or Microsoft or whoever else make it easy to develop on next generation hardware (with all this talk about CG and middleware and all that), what they cannot make easy is getting the artists to come up with decent art (FFX FMV level, without even getting into the racing games part and the sheer size of next generation games) for their games. they can make it so easy that a 5 year old kid can develop on PS3 or Xbox2, but that doesn't mean that the results will be acceptable.... (still, a 5 year old kid might come up with something more interesting than a lot of stuff that came out this generation)

all i'm asking for is true VGA support... then i'll wait for Square and the other big names to do the work and i'll be happy.

 

[rabidrabbit]

Maybe, when the credits roll in the end of Final Fantasy XIV, there will read:

Programming: Square Enix
blah
blah
blah
Visual Effects: Industrial Light and Magic
Character modelling and animation: Weta Digital

 

[Almasy]

Perhaps an entirely new industry can appear, one that specializes in textures and such and can be contracted by publishers on a game basis or something of the sort...

 

[Grall]

Content creation is going to take a LOT more time for a system that can do upwards of a billion polys/sec compared to one that struggles to manage a hundred thousand. In some ways, higher-order surfaces will let modellers "cheat" because they don't have to design each and every poly of course, but a lot of stuff can't be done that way. I am personally totally amazed by many of the rooms in Metroid Prime, they're so detailed it's hard to understand how they did it all. I have scoured some rooms for over ten minutes and still noticed tiny new details, it's totally amazing.

With even more powerful hardware, the load on the 3D modellers will only increase. Texture artists won't be off quite as bad I think, doubling a texture's res in each direction would result in little or no extra work under most circumstances. Actually I would think most artists already do their textures at a fairly high res and then scale it all down to a target size.

Programmers have to do more too of course, but certain stuff like physics, AI and even a whole engine itself can be bought from outside parties. So far most software houses have done their own engines on consoles, possibly because hardware resources have been relatively scarce and you can't waste RAM and processing power heedlessly like a madman in good 'ol PC-style. With this coming gen of consoles, I would not be surprised to see id/Epic-style engine licensing on a large scale even on consoles. Of course the big companies will probably make custom jobbies themselves, but chances are the smaller might not have the resources needed.

*G*

 

[zurich]

I'm just worried that the 'super high production value RPG' genre (read: FF, KH, etc) will die out completely due to insane costs.

 

[Blade]

Almasy: Not that bad of an idea, if it works well on a financial level.

 

[Berserk]

What I'm worried of from a gamer's point of view is whether tge games will cost more or not. I think the former will be the case

 

[Almasy]

What I'm worried of from a gamer's point of view is whether tge games will cost more or not. I think the former will be the case

If the market continues to expand as quickly as it currently has, I think we´re safe from a price increase.

Almasy: Not that bad of an idea, if it works well on a financial level.

Isn´t it?

I think it cold work like this:

For example you´d have a couple of companies that focused on RPG kind of textures, adventures realistic textures (and perhaps some large ones that could be involved in several genres). Or perhaps some that specialized in certain themes.

For example, S E could contract them and say "Hey, I need this kind of textures for FFXIV" and then S E would devote most of their resources to the programming side, while textures would be sent to them through a broadband network. Development times could decrease or at least kept on par thanks to this (or perhaps things are already handled similarly?).

Or perhaps such a thing is unfeasable and unrealistic?

 

[Blade]

I guess there could be a few downsides, like how the texture company involved might want in on game profits and whatnot. Also, where do you get the talent? A lot of texture artists are also responsible for art direction.. characters and such..

I'm sure that your average art guy/team from.. say.. Rare.. probably wouldn't want to become a member of th' "Texture Artists Dream Team Sweatshop" and I'm sure that Rare wouldn't want to lose him/her/them.

It's a good concept, though. Just work out the kinks (like getting talent and working out royalties) and you're set. It might open a can of worms, though. Next thing you know, they'll have contractors doing polygonal models and making music!

..Or are they already doing those things?

 

[Jabjabs]

I don't think it will be that big an issue next generation, considering the modeller have to down sample quite a bit to get things running on even high end Pc's I think it won't be much a problem.

Modeller my just make full quality models instead, with loads of textures (and texture passes) with 100,000 polygons and simply wack them in and let the brute force of the hardware do all the work. Example Doom 3, the pink model was modeled at 200,000 polygons and then down sampled to 2500 with normal maps. They can avoid the normal maps and just have true geometry instead.

 

[notAFanB]

I'm just worried that the 'super high production value RPG' genre (read: FF, KH, etc) will die out completely due to insane costs.

nah that just means less dross and more polish.

we'll probably see more sweeping epics a la Xenosaga/shenmue/FFX-2 as we get to see more assest reuse.

 

[zurich]

There's already companies that sell libraries of (or individual) models.. you can buy them in XSI, 3DS, or Maya format (or .obj), at a variety of polygon counts.

 

[Josiah]

check this out:
http://www.marlinstudios.com/products/products.htm

Like zurich said this stuff is already out there in droves. For the most part I think developers will adapt to advancing technology, they always have.

 

[Deepak]

LB! I think you should be happy than anything else, everyone should be happy....

more complex games = more manpower = more employment opportunities!

It is good for mankind!

 

[Grall]

I don't think it will be that big an issue next generation, considering the modeller have to down sample quite a bit to get things running on even high end Pc's I think it won't be much a problem.

It's not just in-game models that are affected, but the actual levels themselves as well. It's all well if you can build a city environment with cut and paste jobbies to create highly repetitive surfaces like the front of a highrise building or place objects from a library in the scene, stuff like streetlights, trashcans, mailboxes etc etc. However, you can't build nature scenes like that. If you have fifty trees in a level all looking the same or fifty rocks etc, it starts looking fake.

*G*

 

[London Geezer]

HEY AGAIN,
well what i originally meant is that some things like the sheer size of the levels will be an issue. IF models can be "bought" from external team, levels cannot. Take a next generation The Getaway... it took them like 5 years to make it for PS2, i can't imagine how long it would take them to increase the detail enough to make it look like a good PS3 game.....
The Getaway featured a rather simplistic version of London, i guess that a game like that for PS3 will need so much in temrs of resources it's not even funny.....
i guess the guys at Studio Soho have higher resolution textures of London stored for "future games", so that should lower the pressure...

 

[Gollum]

I wouldn't say that the work for modelers will increase way more than for texture artists next generation. For offline rendering modelers have been using SDS for quote some time now, if they now become available in real time it might even reduce the work, as they don't have to produce a low and high detail version of a model and just concentrate on one version instead. Texture artists on the other hand won't have to worry about one single big texture anymore (which all the shading is baked or painted into), but they will have to create a whole array of different textures (diffuse, specular, colour, reflection, bump, normal, whatever) for the complex shaders that next generation consoles will have to support. Either way, artwork/content creation will become an even bigger problem, as will the production costs of AAA titles.

 

[Grall]

Gollum,

Most of the different textures used for procedural shaders will only have to be created once for each material type. There will probably be libraries of such textures available to give for example a brushed steel look, or a rough wood look etc etc, if there isn't already. Look at the MadFX 3D modelling program for example that does more or less realtime rendering of complex images using a library of shaders.

Also, a texture artist wouldn't create a normal map by hand, that would be done by the 3D modelling program, or some other program and would thus be pretty much automatic.

Texture artists will not be overly bogged down is my guess. Modellers WILL, however. There's a reason they have like a hundred guys working on CGI sequences for blockbuster movies, and that's just a couple minutes worth of screen time (more in some and less in others).


*G*

 

[BoddoZerg]

IMHO, as complexity increases, things will get a lot more modular.

Take for example DOOM3. Instead of the traditional way of building FPS levels, they have a map editor that works almost like Starcraft/Warcraft map editors - everything is built using prefabs and rendered in realtime as you place it.

In the future, it is likely that there will be an increasing number of content-production companies, or intracompany teams, that churn out prefab objects for use in games. This has already happened to a great extent with game code; most FPS games these days are assembled using a game engine bought from a 3rd party company and a physics engine bought from a 3rd party company. Sooner or later, the same evolution will apply to game art, although art tends to be more game-specific than physics and rendering code, so game-art houses will do much more custom work.

Note that making a single game entity is just as easy to do no matter how high-poly you are talking about. A high-poly 3d model is no harder to make than a good looking low-poly one. High-res textures are also no problem, since they are always drawn at ultrahigh res and scaled down anyways.

What is rapidly becoming more complex is the number of game objects in the world. In the Half-Life/Quake2 days, no one cared that gameworlds were sparse, devoid of any small objects that would clutter up a real scene. Nowadays, people demand more complex game environments. Indoor scenes have mazes of wires, pipes, and tiles on the ceiling, and tons of common little objects lying around the place. Placing all these by hand will eventually become prohibitively tedious. Outdoor scenes have lush vegetation, every tree has to be slightly different. IMHO, this is where a Starcraft-style map editor will come in handy. Anyone who's made SC/WC maps knows that the editor automatically makes random variations and places random "doodads" whenever you draw tiles of terrain. The same process (albeit more sophisticated) could be used when creating surfaces or rooms in a FPS.

 

[PeterT]

However, you can't build nature scenes like that. If you have fifty trees in a level all looking the same or fifty rocks etc, it starts looking fake.

Well, you could just generate plants and terrain procedurally, or at least add detail that way.