Home General Discussion

non linear gameplay vs art production

polycounter lvl 17
Offline / Send Message
SHEPEIRO polycounter lvl 17
was reading this article:
http://kotaku.com/we-are-still-shooting-the-hinges-and-it-still-sucks-1445267097

and it made me wonder about the difficulty/cost in producing the art(forget about design for the mo) for a non linear game compared to a linear game.

my basic question is:
Are there savings to be made art production wise by reusing areas more completely and implementing a less linear approach to level design or does the effect of having multiple routes/set-peices/puzzles using the same space negate any saving made or even increase it?

in my experience of making games, I know a little about it, but I was wondering if anyone with alot of FPS experience designing/building both of these thinks (as these can possibly give a good comparison without brining in too many other differences)

ku-xlarge.jpg

for me personally I would like to get to know an area, we spend so much money on these worlds that we flash through wouldnt it be nicer to slow it down BUT make that journey interesting by making it denser. the arts dense, but when you compare the interactivity it doesnt seem that much more dense... maybe this is more of a discussion about pacing than necessarily about linearity

Replies

  • adam
    Offline / Send Message
    adam polycounter lvl 19
    I've worked on linear and non-linear games and find that the linear games are more involved, and you work on far less of a total space as an artist.

    In my experience

    Linear:
    • Tasked with a specific area of a specific map OR you're tasked with a single map. You'll work on that map for the entire product of the game. Only once has this not been the case for me, and this was on PS2/old titles.
    • Much more one-off, specific set pieces and other non-reusable pieces. This is primarily due to the amount of scripted events happening in the map and the art required for those.
    • More time spent telling micro stories ("If this is a rose garden, then what is the deal with the destroyed feathers, that shoe over there, and these few blood drops? What happened?")

    Non-linear:
    • A far greater space, when compared to a linear title, is your responsibility.
    • Art is reused much more aggressively and creatively to cover far greater surface area.
    • More time is spent on telling macro stories ("this is a rose garden").
    • Far more planning in the beginning before hitting full production than a linear title.

    In the end, I wouldn't say one is more involved than the other. Linear is the Apple, Non-linear is the Orange when it comes to that sort of development question. Both are fruit.

    EDIT: For what its worth, I prefer non-linear. I love having to think about the impact I'll be having on a players emergent gameplay and giving them a sandbox to play in much more than telling an "epic", scripted story.
  • Crimson
    Offline / Send Message
    Crimson polycounter lvl 18
    I've never worked on a linear-narrative driven game professionally before, but I have spent the last seven years working on MMO's.

    With these sandbox games, I agree with Adam, the focus becomes about efficiency. Especially with a limited art team. The live team has to fight a constant battle of keeping ahead of the player base. We get to be the healer class of the MMO trinity. Constantly trying to keep the tank's HP above the mob's damage output.

    So, every asset has to be reused to provide as much content as possible. It can be really taxing, creatively. You want to give every item, and every prop it's own spotlight to shine in. Time just isn't on your side for that though, and art must be recycled. Picking battles is key, and if you can spend a little extra time making one diamond of the bunch it can make the rest shine as well.

    Though there is one thing I'll say about working under such constraints. You become really familiar with every asset available to you. Every time a new problem arises you can quickly run through your mental catalog of available weaponry to throw at the issue. You also become extremely creative at how to make a model remixable from the get go, future proofing for potential challenges.


    I hope to work on something a bit more confined at some point; Be able to poor my energies into a refined experience where I don't have to worry about as many potential problems and focus on a core experience.

    [Edited for formatting]
  • McGreed
    Offline / Send Message
    McGreed polycounter lvl 15
    But with nonlinaer games, is the preasure more on the level designer, then the props/model designer? I personally prefer nonlinare games and would love to create/design one, but I also understand all the pitholes there is when trying to make everything work all the time.

    Like a western street in a movie, linare the houses would just be facades, except for the one bar, where nonlinare would have all rooms made with furnitures, because you got the option to enter it. It will require more work, but I would really prefer it. But if you have triggered sequences, you can easy break them because of the freedom.

    Good example is Morrowind, you could actually break the storyline since you can go everywhere and kill everyone, howerever they reconize that and gives the player a message that you will break story progresion if you continue, however something important: they dont take the option away from the player.
  • slipsius
    in terms of asset creation, im not entirely sure it would matter if its linear or non-linear. It would depend on how each was designed. You can easily take a linear level and flesh it out into a mass sandbox style level using all the same assets. Maybe a few more tile sets for rooms and such, but for the most part, you could do it. The problem comes when you want each area / room to look and feel unique. If the designers want all new furniture, or different style rocks or anything like that, that creates more work for the artists.

    I think the real difference between the two though isn't so much how much work there is, but who that work is assigned to. Linear games tend to have more in game cut scenes / cinematics, which require more of the animators. They cant just use common cycles for it all if they are to be interacting with a very specific location. Nonlinear games tend to have less of that, and the ones that want more variety in locations require more work from the enviro guys.

    Also, I've seen that picture thrown around before. And although it may be true, i dont think it necessarily means its bad. the reason games back in the day were so non linear was because they lacked realism. Enemies had the mindless roaming, placed in a spot and never to move. But now, AI interacts with the environments, and they react to how the user is playing. You shoot a guy, other guys may run over and see what you`re doing. In more nonlinear games, you still see the mindless roaming with enemies just thrown in random locations.

    In terms of development, it's all the same pie. How each slice is divided is all that changes. At least, from my understanding of it all.
  • Snader
    Offline / Send Message
    Snader polycounter lvl 15
    I'd like to say this: what matters isn't so much the linearity of a game, but more the scope and the style of the environments.

    If you're building a factory with walkways and such, make them modular. If you're building a forest, make a bunch of plants and instance them. If you're making the sphinx of gizeh, you better make that a unique model. (though you can use tilable materials and vert shading)
  • pior
    Offline / Send Message
    pior grand marshal polycounter
    The very interesting part is that games like Doom, Dark Forces and Duke Nukem (all good representative of the non-linear fps genre) were all designed as 2D blueprints, seen from top down.

    This is how their engine worked, how their editor worked, and in great part how the games played too. If you bring up the Doom map while playing, you are basically playing a game of Asteroids, down to the movement and friction model.

    [ame="http://www.youtube.com/watch?v=psaM7kK5Toc"]1979 Atari Asteroids Arcade Game - YouTube[/ame]

    Now it seems like this legacy actually carried over well to Quake, Quake 2 and Unreal. I suppose that this is simply because the designers who worked on the older, 2D based FPSs were still around and applied the same logic as before. But with the transition to 3D, the space of the game world gradually opened up (with levels actually built inside 3D editors instead or Qradiant) and in ended up in a much more cinematic experience, gradually moving away from the original sense of exploration that the limitations of 2D map design forced onto the designers. Movement models changed too, in order to accommodate a richer aiming and shooting experience. It made the shooting itself more fun, but also marked the end of super fast FPS character movement allowing the player to quickly go back to the beginning of a level in order to find a recently revealed key or hidden door.

    I suppose that with the added fidelity of 3D space it also became more and more tempting to design game events and spaces from the point of view of the player, like a shooting gallery with successive "panels" to be experienced linearly, one after the other.

    In short : I think that games like non-linear level design can totally be done today, with the current visual standards to be expected from modern games ; However it would require a strong change of attitude and approach to game design compared to what is often being done with linear FPSs.

    The irony is that, the great liberty of movement and the visual fidelity in modern FPS games goes a bit against that. The 3D space can be so rich and detailed that basing gameplay on exploration inside such rich spaces would be "too much".

    Portal is a great example of how necessary it is to "trim down the 3D" in order to make 3D exploration enjoyable. With its simple tiled walls and strong graphical clues, Portal is delightful ; if it was coated with the unique, non-repeating detail and realism of BF4, CoD or The Last Of Us, it would be a really annoying experience.
  • GarageBay9
    Offline / Send Message
    GarageBay9 polycounter lvl 13
    perna wrote: »
    slightly off-topic, but I've always been waiting for the point that a non-linear world becomes so solid that it's re-used not just for DLC, but for another title entirely. It seems we're not too far from that point, looking at GTA5.

    The idea is that you would continually refine and expand one game space, instead of re-making everything from scratch every time and because of that, never hitting a very high level of interactivity.

    Chunks of STALKER: Shadow of Chernobyl were reused multiple times across Clear Sky and Call of Pripyat.

    Not sure that's what you were driving at, but it did immediately come to mind.
  • eld
    Offline / Send Message
    eld polycounter lvl 18
    perna wrote: »
    slightly off-topic, but I've always been waiting for the point that a non-linear world becomes so solid that it's re-used not just for DLC, but for another title entirely. It seems we're not too far from that point, looking at GTA5.

    The idea is that you would continually refine and expand one game space, instead of re-making everything from scratch every time and because of that, never hitting a very high level of interactivity.

    Wouldn't the problem with that be that technology is racing us faster than we can make the content, even in the same generation.

    Every time we wish to reuse all the content for a new game we find ourselves having the ability to make it look so much better.




    And on topic: I find the budgetary choices between linear and non-linear has far greater connections to the more predictable nature of a linear game, as we can always adjust here and there to accommodate for the differences in need between linear and non-linear.
  • Snader
    Offline / Send Message
    Snader polycounter lvl 15
    Does Far Cry: Blood Dragon count? =P

    Really though, you don't even have to re-use the entire world to re-use assets. I'm betting GTA VI will still be in a city with trashcans, swingsets and highways, for example.
  • cman2k
    Offline / Send Message
    cman2k polycounter lvl 17
    Carmack touched on that a lot with how they made Rage. Their toolchain was set so that they built at a ridiculously high fidelity, and downsampled for the actual game. Meaning later on he could use that raw, super-high-fidelity data, and try different ways of compressing or baking the data - to get better results.

    I'm sure you can all imagine a time when we just zbrush and polypaint our trashcans - and the tools are what make it game ready. And for the next game...we tweak the algorithms and that trashcan is a little higher fidelity than before.

    Clearly there are complications with textures/material models/physics and all types of stuff. I'm not saying there's a clear-cut solution already. I'm just saying, I think we could all imagine a future where there are clear-cut solutions and we've hit more or less hit the ceiling as far as asset creation fidelity goes, and we can stop rebuilding the same things over and over for a series of games, or even games within a company that are all the same style.

    I tend to agree heavily with Perna, in that I'd freaking love a GTA 5 1/2 - where there are way more buildings I can enter and systems I can interact with. Enriching a world to a ridiculous, as-yet-unseen point instead of starting from scratch again.

    While it may not be true for art - this HAS been true for several game series from a design perspective. There are definitely devs that have taken their previous designs and iterate/refine/add/polished it over a course of several sequels, over decades, and it's clearly a benefit. It'd be nice to get there with art one day.
  • ExcessiveZero
    Offline / Send Message
    ExcessiveZero polycounter lvl 6
    I don't see what's wrong with having both

    Bioshock is a great linear experience, fallout 3 is a great non linear experience.

    I don't think you could have such a strong single narrative like bioshock in a non linear environment, without the possibility for linear games you simply won't get the same experience, sure COD or whatever can do it wrong, I personally hate stuff like gears of war where I see the cover and know oh this is where the enemies attack me before it even happens, but theres the right way and wrong way to do everything.

    and hell I expect some people like that gears of war stuff, they must sell enough copies, so more power to them, id support this more if there were no non linear games but there are plenty.
  • slipsius
    The Gears of war cover example just goes back to what I said about design. You could make a nonlinear game with the same cover system, and to avoid your hatred for knowing the baddies are coming, you put cover everywhere. will you need it? will you not? or have spots that has lots of cover, no enemies. then spots with no cover, but lots of guys. balance and design.
  • Snader
    Offline / Send Message
    Snader polycounter lvl 15
    Honestly, Rage was graphically kind of a letdown. Not because of the low resolution textures (okay, not just) but because the benefits were not very noticeable. The decal stamping was still visible in many places, there were near-identically textured assets in the same room, and so on.

    In the end, the problem wasn't as much the hardware, as it was the artists having to paint every square foot of the world. Which cost a lot of time, and still wasn't perfectly 'unique'. Which is something I'd usually not complain about - but uniqueness was the whole point of the engine.

    I think that we should look more and more at complex, combined materials. Throw 5 sandy rock materials together, vertexpaint, and you've got a near infinite desert. High resolution detail textures (1024px per m) combined with rough shape maps (1024 per 10 and per 100 meter), and you've got a giant sandbox game. Without costing too much ram, because all that stuff tiles. Plus, it can be much more dynamic. You leave the door open? Sand can blow in. You can make rocks with snow on them, dependent on wind angle:
    [ame="http://www.youtube.com/watch?v=sXl3z_Zcyck"]Rock and Snow Up Vector Blending Demo Part 2 (UDK) - YouTube[/ame]

    We need to step away from handcrafting entire worlds, and more towards building rulesets to grow a world. Simulating the (fractal-ish) nature of materials rather than editing a specific texture on a polygon. Rendering techniques are headed slowly towards simulation, I.E. raytracing, and that will play much nicer with a substance-based approach.

    At least, that's roughly my vision on the future.
Sign In or Register to comment.