News | Forum | People | FAQ | Links | Search | Register | Log in
Style And Level Design
Some time back Gilt started a rather interesting discussion about style. The discussion was cut short when the QMap db got fragged (some say 'twas the reason), and I believe it's well worth reviving.

Let's go...
First | Previous | Next | Last
refresh my memory. 
i can remember the tumor layout.
the GTA3 discussions
one guy doing the architecture and one the texturing (i LIKE that since i loathe texturing and only ever have time to do blocks and now my blocks are gonna be called suspension parts for formula student car: yep not enough time for mapping so far :-( 
testing stuff 
lunaran and rpg are doing that -- lunaran has already designed a layout, rpg will do all the brushwork, and then lunaran will "skin" the resulting geometry. 
i think it's the best way of doing it. if you know each other enough to have trust and also if you take the job seriously. does anyone else do it regularly? (dunno if daz and killaz did it like that or if they just merged various parts of maps...)

i'd like to ask rpg how he works. does he choose the rough textures and then lun realigns them or does rpg "clothe" all brushes in any random tex and lun chooses the tex set and makes the wad? 
I do the architecture in 3 or so basic textures, and then Lun will make appropriate textures for all the surfaces and retexture the map. 
I like texturing and textures often inspire a lot of my brush work, but I really hate when you need a certain texture to fit a theme and you just can't find one. It makes me wish I knew enough about photoshop to do all my own textures. 
It Makes Me Wish... 
...i had the patience to do them. how long does it take to make a dozen or so textures, anyway? 
I find it easier than mapping anyway.
If you just want a particular texture - you can always modify exesting one. I do this all the time, rather than struggling with the set that doesn suit my brushwork.
Creating new textures that would look good in quake pal is not easy, but making average truecolor stuff is just a matter of knowing how to use the software right.

Tutorials at
And texture forums at 
About Design... 
BWDM1 = design of the century. 
About Time... 
BWDM1 = vaporware.

Please prove me wrong. :/ 
Haha Beat This RPG 
Title : King Of My Castle
Date : 18/6/99
Filename : bwdm1.bsp
Author : Bram "geuss who?" Weterings
Email Address :
Description : A medium sized DM leven for 2-6 players for Quake 2 
bwdm1 was for Q1 f00

I did have a really old version around once, but it was on an old PC. 
i have a feeling if i tried it would be exciting untill i finally used the tex in my map(s) and it looked crap. like when i'm sitting for an hour with some music, adding stuff, go down for a cuppa and listen when i come up and it sounds muted (if you add too much and it all collides), boring, bollox (rather sums it up).

practice makes perfect of course... 
Texture Time... 
The textures I made for Q1 were a lot easier than those for Q3, mainly because of the double scale of the latter. The color palette for Q1 is not actually as much of a problem as you may think - I find that some of the best creativity is when people have limitations, but those limitations are clearly defined. It's really a case of understanding what the palette is made of and therefore what you can get away with once you reduce colors down to 256.
Another thing that makes Q1 textures easier is the almost complete freedom to make what you want; crumbling blue brickwork laced with stainless steel gigeresque pipework? No problem! Such things don't always work in other games, even Q3 :(
The time I spent making textures for my first Q3A map, The Color Of Anger, was vastly greater than the time spent learning Gtk, building brushwork, meshes etc.
Complex, intricate textures like trim or feature-specifics such as doors and jump-pads, can take many hours. I think my kothic jump-pad took about two weeks, a few hours each day. But when making textures specifically for your map, you also have to remember all the textures that are made but end up not being used.

Anyway...mapping styles:
...or somethin' like that. 
Hiya Kell... 
I used your Knave textures when building a section for Q1-Turtle Map 1. Now I know how much work went into them, I must thank you again. I can sling you the .bsp if you want to check out the usage.

re: Mapping Styles
I grasp the "tumor" style and the "monumental" but I'm wondering if you can clarify "cavernous" with an example. By concentric do you mean "chaotic", where the major design feature is iterated at every level from layout down to detail? 
I'd like to see the map, distrans :)

cavernous: maps that are based on expending brushwork for size rather than detail. Insomnia was like that IIR.

concentric: where the map is built around a main area/feature which the player visits and revisits multiple times throughout completion. Peripheral areas are all smaller and carry diminutive versions of main motifs.
I could probably cite Libris Vertiginis as an example, but there are others.

re: chaotic; what you describe sounds like another approach worth mentioning, though I think the term 'fractal' is more accurate terminology. Can't think of an example off-hand, though I am slightly drunk and I've just discovered my website has been reverted to an older backup :( 
Bad luck about the site :(

Insomnia struck me more as monolithic rather than cavernous, although in truth both are appropriate.

The concentric notion makes sense now. Perhaps some of [Kona]'s maps fall into this category as well, although the revisit in those instances tend to be on a new level vertically. The "e3" map from ProdigySE would seem to be a good example of what you are talking about.

Fractal, yep that's the notion. Biff_debris would've made a fair fist of this sort of level but I can't think of anyone who's actually pulled it off (apsp1 maybe?).

/me waves at biff

Level fragment for the Q1 Turtle1 on it's way :) 
Satanic Concentricity 
apsp1 was more of a mutliple-concentric map (being centered in certain areas around a certain colum, ie the beginning area or the generator) but these are rather chaotically placed so I can see where you're coming from. 
I take your point about apsp1 as a multiple-concentric level, but the notion of chaos I was alluding to was wrt chaos theory. So to distinguish from the randomness of chaotic (in your sense) level design we could do worse than adopt Kell's notion of "fractal" level design as a descriptor. 
Evolution On The Visual Front: Part 1 
In terms of style and design, the possibilities are approaching the unlimited. It is very near the point where engine strain is no longer a factor as they can now handle almost anything someone can conceivably build (with the exception of stupid things like rooms the size of the grid). So then, the limiting factor then becomes, the design skills, work ethic and imagination of each individual. This, in turn, further raises the standard of level design, and provides for more variation of form, color and composition; more detail, interaction and dynamism. Now, not only is gameplay more flexible, but the quality of the environment which defines it must be more cohesive and believable. And when I say believable I do not mean that it must reconstruct anything that exists in the world, past or present, but only that it must present a "yes" to the question "Could this reality exist as it is?". This is where the imagination may run to the horizon in a kind of "controlled" directional creativity.

Point two. Such a degree of design freedom is changing the way level designers must approach projects. The days of wedges, cylinders, blocks and spikes are drawing to a close. Now, these forms must be considered merely the starting point of construction in exactly the same way that they are regarded in ultra-high-number polygonal modeling, and they must not be considered acceptable as the details of the final environment (with some exceptions). Such a mundane acceptance would allow a balloon to represent a human head, which is absurd considering the power and capability of modern rendering. Thus, such objects are, appropriately named, primitives.

Moving on, there's textures. Textures have, up until now, been used primarily to assist three dimensional constructs with detail as well as the visual illusions of depth. Hence, one sees many textures with highlights and shadows very 'loudly' pronounced. This is obsolete. Again, the cutting edge is being defined by film-quality modeling. Textures' new primary purpose must be to give forms a material. That is, to make a cylinder metal, or a rectangular prism brick. Details such as dents and rivets on the metal cylinder or cracks and mortar on the brick should no longer be grouped with the texture but 'physically' modeled into the cylinder or rectangular prism, respectively. It can be summarized quite easily: More 3d, less 2d. You might be thinking "Well hey, 2d is important too and it adds so much!" but you must remember that this is, ultimately, three dimensional (or four if you include time, which you should) design. Stop and study the real world about you; look at how much variation there is in form; look at what juts out or dimples in; look inside your computer with all the small three dimensional details! Most everything is a material and a form. Thus, form should not where at all possible be dictated by two dimensions exclusively. And yes, that means more work for you, the level designer. Now you must create multiple textures for smaller areas of architecture and further develop the details. However, this problem is somewhat alleviated with the advent of displacement mapped textures (in Doom 3). But, this merely allows for the creation of prefabricated details, which still had to at some point be modeled (unless I'm misunderstanding how these new 'textures' work).

Anyway, this innovation in texture functionality (displacement mapping) provides a suitable transitional, easily manipulative and highly practical methodology for refined detail construction. Trying to build a circuit board as detailed as possible in 3D is very tedious to say the least. Now what if you needed 10 variations? OI! The headaches and long hours foreseeable there! A better solution might be to create many small details in low level 3D, sometimes referred to as 2.5D, (meaning 1 dimension of the 3 is considerably reduced in size in relation to the other 2) and use them interchangeably like modules to create a larger 3D construct. Thus, the smaller details may be moved and rotated, and even given different materials to provide for much more variation with minimal hassle in miniature design. It would be possible to have a computer program generate such variations for you as well. But ultimately, level designers must provide more variation at a smaller scale or risk bland, bare design (where realism is concerned).

Now, since I can only post 5000 characters in this post, I must continue in the next one. 
Evolution On The Visual Front: Part 2 
Now, if you were to stand out on a city street and try to observe all there is to observe, you would be there all day, and would still not recount everything. The amount of information presented to the senses is staggering. There is so much there that selective attention is necessary. This is what the level designer must understand. They must provide enough detail to imply ALL the information that selective concentration ignores. This is no easy task since there are seemingly limitless areas that the attention of an individual might wander. The cutoff, in the real world, is the point at which the senses become saturated or the variation is so minute as to be indistinguishable. It is the point at which the eyes can see no smaller or farther or the ears can hear no softer or word-that-means-distinguishing-multiple-sounds-mixing-together-in-a-wash-of-noise-er. Now, obviously, virtual worlds are getting closer, with each new engine, to providing the ability to present such massive amounts of information, but there is still a long way to go (if it ever gets there). So, the 'professional' mapper must keep all of this in mind to create an environment and further imply more then what is there. Remember, it's a place, not a map.

Also, I do not mean to imply that more is always better or that more is good everywhere. Only that detail and design must be in balance with probable, plausible realities and imply whatever they cannot practically represent.

Finally, I have forgotten the full reason for having started this post as well as the ideas and points I wished to convey. And since I'm quite susceptible to narrow thinking (tunnel vision) I have probably made some blatant errors in logic or fact. But since I won't find any of them until after I click 'Submit', this will have to do.

A final, final note, the gameplay dictates EVERYTHING else. Design that first. Then build everything else around it. Also, perhaps if I get a second wind I'll dive into theory on gameplay design (which could be totally wrong too).

And so they all lived happily ever after.

I Won't Get Into Details, 
but the main point i disagree with you on is the claim that the burden is shifting from textures to geometry. I believe that both textures AND geometry should do as much work as possible, and that means more work this year than last year. I can understand stylistic choices that lead to either flat textures, or simplistic geometry, but these aside, you should get the full value out of both geometry and textures. 
How Much Can One Man Do? 
A set designer doesn't create every facet of a set, a lot of it just happens. If we wants to create a room he builds some walls, paints them, buys a used table and some drapes. Sure he chose the paint, drapes and table, but the subtletly of the textures and shapes came for free.

Right now a mapper doesn't get that, if you want you table to look a little worn you have to think about where to put that nick and the scrap marks, and then create them very slowly and carefully. I think that mapping tools are going to have to start taking more of the work.

Like, (if I can take my ill thought out example further) a tool for scuffing up the table. Tell the tool where and a few paramters and the AI does the rest. It's like taking the terrain generating tool to the next logical step.

It's late and my example sucks but I think you can see what I'm getting at. I've taken all the work off mappers and put it squarely on the programmers. :) 
First | Previous | Next | Last
You must be logged in to post in this thread.
Website copyright © 2002-2023 John Fitzgibbons. All posts are copyright their respective authors.