Regardless of toolset and engine, how do different people go from brainstorming an idea to having a designed level that's ready enough to be assembled for a game? I'm not talking about actual in game implementation. For me, "Action/Adventure" can include FPS, platformers, or anything with abilities, obstacles, items, or enemies. Do you start with a story that's contionually refined, a random concept sketch, a random game dynamic, or just tinkering around in an editor? How do you decide when and where to place and how to pace obstacles, enemies, and items? Do you create challenges to suit the players abilities or the other way around? I've read of pacing being handled like sheet music, organizing a tempo of activity to set the mood for participants. I'm just looking to see how different people work. [Edited by - BeastX on May 30, 2006 4:19:29 PM]

The overall setting of a level is dictated by story/plot. Then I do sketches and concepts of the setting (such as a castle, a farm, a factory, etc). I also do a stream-of-consciousness flow chart of events for script-heavy sequences. I then would do a floor plan, build it out of blocks in 3D, refine the floor plan, build again, refine. The floor plan is where I do the pacing and the fundamentals of the level (enemy/item location). Then do concepts of some more specific individual objects/pieces of the environment. Finally, I start building the level, and creating needed props/buildings similtaneously.

There's several approaches to designing levels, depending on the game and the level-designer himself.

I usually take a gameplay-first approach, laying down rough architecture and testing puzzles/combat/exploring/other before doing the visual layer, but often I also create small sub-maps that are highly detailed so I won't loose a feeling for the theme and the mood.
I also plan some things beforehand, while other idea's come up halfway down the process. Some people, like me, prefer sketching layouts and settings, others prefer starting in the editor itself.

Of course, when I'm working on one level that doesn't mean I can't think up on interesting puzzles or situations for other levels... sometimes you just build a library of ideas and take from them when you can use one. I also always create a library of reference art very early on.

As for storylines, that depends on the game. Some games depend more on them than others, for those that depend a lot on them you're theme and setting are probably going to be dictated by it, as are your enemies and puzzle types. A story that's all about high speed chasing isn't going to offer much time for vast, exploring-oriented levels, for example, neither does a high-tech game offer ancient temple levels. That aren't infested with missile-launching tech. :P

Of course, at some point level-design ties together with game-design pretty tight. As a level-designer I would probably create challenges fitting the players ability, I think the story-line and game-design are responsible for the abilities and when these are introduced. Good communication is important there... :)

Keeping in mind that I haven't actually done any complete level design yet, I'd view the process as similar to writing an essay. You start out with some ideas (or, perhaps, not even that, in which case you spend time brainstorming). Say, for example, the phrase "futuristic melee combat adventure" pops into your head. From there, you come up with a basic environment that makes sense for your idea. What kind of world would favor melee combat over ranged combat? What sorts of abilities will your protagonist have? You rough out a universe, and in slightly more detail, a specific area where the game will take place. That done, you make a few passes at drawing maps. The first one might be just blocks with names, attached to other blocks with names, much like a UML diagram. The next one might have major landmarks (a fortress, a waterfall, a crashed spaceship, et cetera) and enemy ideas (remember, enemies are part of the map!). The next one might actually sketch out the terrain for each of the areas (okay, we'll have a cavern here, which goes down into a small mazelike area, the exit to which is over here and leads to the next region...), and then the one after that actually gets built into the game using a map editor or something. From then on, it's mainly a process of testing, tweaking, and polishing.

At least, that's how I see it happening.

The game-type and story definitely have a huge impact on how I go about designing a level, or an entire world for that matter. Assuming that you already have a well established story with particular player objectives I would begin organizing a logic or flowchart to make sure that all the ideas can be implemented seemlessly into one environment. The story itself should set basic pacing for the game play through the level, even if you don't yet know how the environment will be visually interpreted.

The next phase is not a very rigid step, but an organic process of brainstorming intended to determine the artistic style, mood, and scale of the world. At this stage I also try to come up with any key features that will need to be included in the level such as a massive landmark tree, a wrecked ship, a power station, a spaceship docking bay, an impact crater etc. All these ideas are simply jotted down in a list or outline form.

Once I begin to have an understanding of what content and events that the level requires then I start addressing the task of composition and design. I've used a number of approaches to this step in the past including sketching orthographics on graph paper, simple perspective sketches/drawings, and have even used clay sculptures as an initial starting point. As my levels have become more complex I've actually begun creating many of my design layouts in Photoshop with layers so I can keep terrain, enemies, event triggers, items, clutter objects, and text all seperate from each other. I've discovered that you don't actually need to get all that detailed though at this stage. You're just looking for a sensible compostion with good aesthetics. The reason for limiting the detail at this stage is because I've recently added a new step to my level design process, which is typically referred to as the blockout.

The objective of the blockout phase is to create extremely basic static meshes that you can actually begin putting into a game engine to begin test playing. For me, the editor I'm most experienced with using is Unreal, but the process can be applied to any game editor. What this allows you to do is actually play through your level in a very primitive form to discover if it has good pacing, a readable layout that is easy to follow, decent variation, etc. Level blockouts should be as basic as possible with extremely simple untextured geometry. This is also a good stage to test collisions, dynamics, and bot pathfinding if it is going to be used.

Once you have confirmed that the level plays as you want and you've made all the changes necessary then you can actually begin creating the static meshes and textures that will be used in the final version and begin populating your level. If you did your blockout correctly you can often simply replace the "stand-in" meshes of your blockout with the final static mesh. I was fairly proud of myself for developing the blockout process until I discovered that other game developers were already implementing very similar practices.

Anyway, this approach works well for large-scale FPS, RPG, Action, and Adventure games but all of the steps may not be necessary for smaller-scale level projects.