Sponsored By

GDC 2001: The Architecture of Level Design

The bar for the visual quality of environments in games keeps getting higher. To meet that bar and push it even further will require the integration of real world design skills as well as a keen sense of gameplay and game design. Attending an architectural design or any sort of design school is of course not absolutely necessary to be a good level designer, but it can help widen your scope of experience. Architectural Design and Level Design are two very different pursuits, the point is really to investigate how looking beyond traditional and often overused game design references and approaches can help to bring an added dimension to the experience of making and playing games.

Steve Chen, Blogger

July 16, 2001

23 Min Read

The bar for the visual quality of environments in games keeps getting higher with each new generation of hardware and software. To meet that bar and push it even further will require the integration of real world design skills as well as a keen sense of gameplay and game design. Attending an architectural design or any sort of design school is of course not absolutely necessary to be a good level designer, but it can help widen your scope of experience. Architectural Design and Level Design are two very different pursuits, the point is really to investigate how looking beyond traditional and often overused game design references and approaches can help to bring an added dimension to the experience of making and playing games. What we would like to do is give you an overview of what you might learn in the first year of an architectural design studio and given our own experiences, how that ended up tying into wide variety of games.

Going to an architecture school these days is much less about the technology of how buildings are built and much more about looking at how human culture continues to express itself through shelter, one of our most basic requirements. Technology and the expression of design ideas and gameplay also seem to be becoming less closely tied together. We in the industry are witnesses to the explosion of licensable engine technologies and subsequent number of games with the same core technology but very different game ideas. Without the technology games like the kind we've all played and worked on are just crazy game ideas and without those crazy game ideas games become just slick technology demos.

The most important thing you learn in architecture school in the first few years is; what's the difference between an architectural idea and an architectural device. Basically, an architectural idea is an idea that says something meaningful about how people experience a given piece of architecture; an architectural device is what specific means you use to express that idea. For example, a window is an architectural device to get light, air and visibility into a building and can be as simple as a hole cut in a wall. An architectural idea about that window might be something like how that window is made to separate the pedestrian public view from a space that's supposed to be private. To express this idea you might frost a piece of glass or place the window very high up; suddenly that window becomes much more than just a hole cut in the wall. The same approach goes for level design. It's important to have a clear idea for what a level is about. If you say, "I'm going to do a deathmatch level", that isn't a level design idea, it's just a vehicle for expressing an idea. You could think of the term "deathmatch level" as a vehicle or device for expressing some idea about people getting together in a space for the purpose of killing each other until someone reaches a certain score.

Having an architectural idea or level design gameplay idea is the most important thing in either pursuit, and beyond that finding an appropriate means to express that idea is the real trick. You will often discover in the process of finding that means of expression whether your idea was any good in the first place, or in the case of games whether or not your idea was any fun. Let's assume that you have a really great idea for a game and its levels, how do you go about coming up with ways to express that idea?

We can start by defining what the qualities of a level are. Level design is very architectural in nature. Levels are spatial experiences of environments that are usually inhabited by you as some sort of avatar and a host of other vaguely humanoid creatures. They have a program in the architectural sense of the word, meaning basically a list of functions like living room, conference room, mad-scientist experiment room, temporal anomaly generator room, etc. They require expenditures of time, labor, funding, and to some extent natural resources, and they are constructed by human hands. So, applying an architectural design methodology makes a certain degree of sense. In order to apply such a methodology to games, it is useful to describe some of the basic elements of spatial design. These elements are really architectural devices in the form of basic design principles for organizing and developing space. They are applicable principles instead of physical objects like the window. It's kind of like the principle in games that puzzles should be hard to do, not hard to figure out what to do. A-lot of this presentation will be a description of those basic principles as they apply to level design, in the hope that they will help you to more clearly and more inventively express your ideas.

Games are ultimately about having a fun, entertaining, and meaningful experience. What we're interested in here is in how you get from a great starting idea to a final finished game. Designing that experience is not unlike the process of designing architecture. Level design, like architectural design, is about finding the appropriate means of expressing your ideas about gameplay. What follows is an investigation of how using the approaches and methodologies of a very closely related field can serve to enhance the process of designing games. Some of the key concepts to design and evaluate architecture are introduced. Path or circulation, how you move through a building, will be outlined. Tools to organize a scheme will be described, and then we will discuss event and character as it relates to level design.

The first book many architecture students are required to read is, "Architecture; Form, Space and Order", by Francis Ching. There are many introductory architectural texts but this one is the probably the easiest to understand and best illustrated texts for the uninitiated. It gives a-lot of examples of built work and has some really great diagrams. A-lot of what we refer to in this lecture is also described in more detail in this book. It's a great way to get an idea of how architects all over the world have been exploring the same basic principles of design since people have been constructing shelter.

The Path

The path or circulation is the key to organizing good design. How you approach and move through a building largely governs the quality of your experience. You can think of circulation as a mechanism that supports the design. The spaces, functions and volumes that make up an architectural experience take place and are guided by the armature of circulation. You can apply this idea to lots of experiences, for example your route on a trip as a sequence of events, or your movement through a museum, and of course the critical path through a game, and not just a real-time 3D game but any game experienced as sequential events. Platformers, RPGs, isometric games, can all be thought of as organized around some idea of path.

Circulation Patterns
There are of course many different types of circulation patterns and they vary according to building function and game type. In a real building, often the circulation patterns are kept as tight and economical as possible; there's always a budget for materials and limit on the amount of space finding the right flow of 'tight' circulation and 'looser' more meandering circulation is an integral part of good design. Games are the same way, only the budgets are in polygons not dollars and the type of gameplay intended governs the right mix of circulation patterns more. In multiplayer games, the routes are usually closed and interlocking. In net games particularly, there is almost rigidity to the circulation pattern. With RPG and adventure games, there is the opportunity for more exploration, so the routes can meander and there is not the need to interlock.

Circulation Components
The basic components and principles of circulation are applicable to both games and buildings, and can be broken down into basic concepts for analysis.

Entrances. In games, everyday elements may actually not be used in the way they were intended - you are probably not going to go in through the front door, but you need to know how to set up the scenario so that it functions effectively as a background to the gameplay. You can announce a lot about the character of the building by its entrance - almost like the load-screen for a level. Is it grand, ceremonial or a purely functional entry/exit? It is important to have a good transition. The approach should be designed in a landscape or city space so that there are gateways or markers that reinforce the sequence. Once you are actually in the building you should have an entrance room, so that you are not fed directly onto the circulation system

Corridors. Corridors are the bane of architecture and level design - you can always try and design them out of your schemes altogether. There is less of an excuse for having them in games, because you can always create an opening or introduce a courtyard when you want to. You can almost introduce a set of game regulations where you don't have corridors that are longer than 50m long. However, if you absolutely have to have them, consider varying their height and width, introducing a syncopated rhythm of columns and openings, or ntroducing exterior views to courtyards or even skylights.

sfhyatt.jpg

Real-life elevators are a pretty dull experience, except for dramatic elevators like at the Hyatt in San Francisco.

Stairs. While we are encouraged to exhibit our scripting skills by using elevators, like real-life they are a pretty dull experience. Unless, that is, you trying to defuse a time bomb as it plunges down a multi-story shaft. Or it is one of those dramatic elevators like at the Hyatt in San Francisco. Stairs are an opportunity to create some great set pieces. Think about all the grand-staircases in movies. You can really open up a volume of space. Landings can become the staging areas for mini-events. Is there room to do battle on the steps. It is not just a matter of getting from A to B. Is it a good sniping location? Or are there views ahead to an unexplored part of the level.

Organization

In an architecture studio you are asked what the concept for your design is. They are asking what the big organizing idea or 'parti'/ spatial idea behind your project is. Organizing ideas and types are often tied to specific programs like schools or hospitals or structural types like brick arches or gothic vaults. In games the player isn't going to get a chance to ask you that question so you better make it clear what you concept or idea is. When designing game levels you might want to be think about some these types to reinforce the idea of feeling of space that game narrative might involve. If you have a story idea about some huge corrupt corporation trying to take over the world with cyborgs (an often used and hackneyed story line) then it might help to go look at what some of the great corporate headquarter buildings look like; how does a corporation express it's power in the world by building architecture?

Type
The term 'type' or 'typology' in architecture refers to the idea that there are fundamental groups of building layouts that can be characterized by a particular functional use. More specifically, if you were shown a plan or a drawing of a building and you could identify it as a church, hospital or an office building without knowing much more about it, that's the basic idea behind the notion of 'type'. Granted, there are different types for each program, but you can generally understand what a building's use is. This idea can be a pretty powerful one when you are designing a level. If you start with and idea about a game and then pick an appropriate type to express that idea you avoid having to re-invent the wheel when it comes to organizing and developing your thoughts and you can make use of this information to build up a scheme. Type is a framework for you to elaborate your game events around. Because a level often combines a number of types, that is why we put circulation first as it ties them all together. For example let's say you have an idea about a level that takes place in a hospital sort of like the final scenes in the movie Hard Boiled. All hospitals share certain characteristics that identify them. These characteristics make up the hospital 'type', lots of corridors with small rooms, lots of elevators near the center of the building, each floor usually has a central hub, there are usually some very specific rooms for very specific functions like the O.R. or emergency room. Using the idea of the type, you might start developing all sorts of gameplay ideas about how a character gets around the level and what sorts of gameplay might happen in each situation or room.

chrysler.jpg

How does a corporation express it's power in the world by building architecture?

Organizing Systems
Architects and level designers love to organize stuff into manageable chunks of architecture and gameplay. According to Ching, in architecture, the five most basic devices used to organize any scheme are Centralized, Linear, Radial, Cluster, and Grid.

Centralized organizations are schemes oriented around a major hierarchically important space. That space is often but doesn't have to be very large. Churches and sports arenas for example tend to be centrally organized buildings. The most important thing happening occurs in the central space. This sort of organization tends to work well with ideas about multiplayer deathmatch levels.

Linear organizations describe schemes that tend to collect themselves around or on some sort of a major axis. For example, a major street in a city or pedestrian paths in a park are types of linear organizations. The most successful example I can think of in games is the train level, AS-HiSpeed, in Unreal Tournament. The entire level is a set of train cabins and the gameplay is all oriented along this line. It provides a great deal of focus and tension to the gameplay, alternatively I think this level could also have worked as a tower for instance. The point is not to think of these as just rigid plan fixated devices but also as three dimensional principles.

ut_hispeed.jpg

AS-HiSpeed in Unreal Tournament is a set of train cabins, and the gameplay is all oriented along this line.


Radial organizations are often similar to centralized ones except that the most important things happening occur in the spokes or arms that radiate out from a central hub. The center is more of an organizing node than it is a major space. Lots of housing subdivisions tend to be radially organized, as well as some prisons. Some centralized pattern multiplayer and CTF levels work well as radial organizations.

Clustered organizations occur when spaces are grouped together close to each other. They can be clustered by function, size or perhaps even materials. In clustered organizations there may be any number of equally important things going on simultaneously or perhaps any number of individually important functions cluster up to describe a single much more important function. You can think of neighborhoods in cities as a type of cluster organization. Buildings and functions of similar types are gathered together and mutually support one another. Levels in games are quite often organized as clusters of spaces. When you look at plans of game spaces, this is the type you most often see. It's also the most difficult to achieve successfully.

Grid organizations are when identical units are distributed evenly, and can be used effectively for mazes. Grid organizations tend to work well with ideas about making efficient systems of infrastructure, a street grid for getting cars around, or a structural grid for supporting floors, or a cubicle grid that can expand and contract with different functional needs. In games, the grid is most prevalent, interestingly enough, in our creation tools. All creation tools from Radiant to Maya use some form of grid to divide Cartesian space into an addressable point or reference. 3D real-time games In a way are all the result of grid organizations.

These descriptions might seem sort of obvious or simplistic and they are, but you really can boil any spatial design project down into some combination of these devices. Most of the time, you probably organize your design using these devices without even thinking about it, but it helps to be aware of these as discrete ideas when you're analyzing and editing what you do. They can help to clarify an idea you have and expose ideas that aren't really working. There is a lot of published work on the subject of how organizational systems affect design ideas perhaps the best known of these is The Ten Books of Architecture written by a Leon Battista Alberti in the 14th century. Who went far beyond these basic systems in the belief that these systems of organization were the keys to harmony and balance in the world. Much of the Renaissance owes it theoretical background to him. More recent investigations into this have been centered around organizational ideas and cities; Colin Rowe's Collage City is an excellent reference for the combination of different types of ordering system.

Spatial Relationships
In his study for a district in Milan, Steven Holl developed a table of links and correlations of four conditions of architecture; on, over, in, and under the ground. Part of evaluating a design would be to consider whether you had introduced a variety of these conditions. Rather than have all the spaces relate directly on a single level, elevate one part of the design and sink another below the ground. Then work to create connections between the two, try to generate spaces that overlap and interlock rather than sit discreetly adjacent to one another.

Events

Having selected an appropriate path and organizational type to act as a framework for your level, you can now focus on the smaller events that bring a design to life. Usually in the real world type or organization takes precedence over path, in the game world this relationship is inverted, and the design is often developed around a number of game play events.

In his book A Pattern Language, Christopher Alexander and his colleagues identify 253 patterns, "spatial bits," that through combination can go towards creating meaningful environments. The list of patterns begins at the city-scale and many of the items are little too small-scale or domestic. However, a cursory examination of the list can be useful for generating ideas. In moving from large to small scale, they follow the process of developing a design.

On a streetscape for example, you could introduce an arcade to create the focus for an event and enliven an otherwise flat façade. Although frontier towns and gunfights spring most immediately to mind, an arcade is an opportunity to create a transition space between inside and outside.

As already mentioned, on a building interior think about developing a staircase as its own separate space. The stairs should not just be a slot between floors, but treat it as "a stage." Give the stair its own volume, open up the surrounding walls and ceiling, make the landings large enough to do battle on.

At the room scale, Alexander suggests considering detail like interior windows and half-height partitions. For gameplay, they create locations to hide and snipe from, and they also break up the feel of a level being a series of discreet cells.


Character

Materials, light and scale are the key areas to focus on developing the character of your levels.

Building materials, or texture in the case of games, lend a great deal of character to a space. Brick adds warmth, stone is cold, and white walls clinical. The first rule is to keep it simple. Try limiting yourself to three materials. One for the floor, ceiling and walls or even simpler. The architecture should be background and not draw attention away from gameplay with dazzling textures. Add detail with trim like wainscoting and mullions, but keep them in a consistent material. Avoid mismatching construction types, the weaker, lighter materials should always go on top. And pay attention to how you turn corners and match edges.

Light is the most effective way to dramatize and enliven a space. You can improve a corridor simply by adding skylights or backlighting. Architecture has a long history of illuminating spaces naturally. The works of Louis Kahn, Richard Meier and Steven Holl are modern examples of inventive lighting techniques. They are experts at manipulating indirect light and playing with the transparency of glass.

It is important to achieve the correct scale and proportion of elements for a convincing sense of space. There are numerous historic precedents for scale and proportion that can be referenced to achieve the appropriate character for your game. Each period of history has its own sets of laws for the size and distribution of elements. Greek is different from Roman and Gothic, as they are different from modern day proportions.

You can also modify the scale in a level for dramatic effect. For example, you can distort and stretch details in a racing game to increase the sense of speed. In Disneyland, many of the attractions are built at three-quarter scale for emphasis, and in first person shooters many of the motifs are compressed and repeated to intensify the experience.


Conclusion

It's impossible within the scope of this paper to cover all the aspects of an introduction to architecture and it's potential uses in game design. It should be clear however that going outside the standard references of the short history of computer game design can serve to widen the range of experiences possible in games. The importance of expanding that experience can't be underscored enough. Games and architecture have always been an important part of the way in which people express themselves. Architects continue to evolve this process over time and game designers could stand to learn a great deal by looking at how other design professionals have dealt with similar problems. As computer games develop with the ever-advancing technology there will be opportunities to express with much more subtlety the full range of human experience both real and imagined. Designing that experience is not the static application of a set of rules and principles. No set of rules or principles will make a boring game design idea a better one, and by the same token great game design ideas can only be made better with the smart application of the right principles.


Bibliography

Leon Battista Alberti, The Ten Books of Architecture, New York: Dover, 1987.

Christopher Alexander et al., A Pattern Language, New York: Oxford University Press, 1977.

Francis D. Ching, Architecture: Form, Space, & Order, New York: John Wiley & Sons, 1996.

Banister Fletcher, Dan Cruickshank (Ed.), Andrew Saint (Ed.), Sir Banister Fletcher's a History of Architecture, Butterworth-Heinemann, 1996.

Kenneth Frampton, Modern Architecture: A Critical History, New York: Thames & Hudson, 1992.

Steven Holl, Parallax, New York: Princeton Architectural Press, 2000.

John Lobell, Between Silence and Light: Spirit in the Architecture of Louis I. Kahn, New York: Random House, 2000.

Richard Meier et al., Richard Meier: Architect, vol. 3, New York: Rizzoli, 1999.

Vitruvius Pollo, Ten Books on Architecture, New York: Dover, 1960.

Colin Rowe, Fred Koetter, Collage City, New York: MIT Press, 1984.

________________________________________________________

 

 

 

Read more about:

Featuresevent gdc

About the Author(s)

Steve Chen

Blogger

Prior to entering the game industry, Steve Chen worked as a practicing architect for five years in Los Angeles, San Francisco, and New York. While at Lucas Arts, Chen was a level designer for Jedi Knight, designing three single-player levels and the Canyon Oasis Jedi multiplayer level. Chen also worked as a background 3D artist for Grim Fandango and as lead level designer for Indiana Jones and the Infernal Machine, where he oversaw the level design and scripting efforts, as well as completing three of the single-player levels. A long-time gamer, Chen entered the gaming industry after studying architecture and computer graphics at Columbia University. He is currently doing research for an as-yet-unannounced new title.

Daily news, dev blogs, and stories from Game Developer straight to your inbox

You May Also Like