McGraw-Hill - 2003 - Ultimate Game Design. Building Game Worlds - DDU03
McGraw-Hill - 2003 - Ultimate Game Design. Building Game Worlds - DDU 03
20 U L T I M A T E G A M E D E S I G N
Building Game Worlds
drawings, and then provide some rough lighting and texturing and try to play in the area as soon as
possible to see if it “works.”
Are the areas too big, too small, too boring, or too confusing? Then it’s just a question of feedback,
refinement, and more feedback and refinement. Until, hopefully, a great environment is produced.
TM: Do “real” floor plans translate into exciting game geometry? Do you consider them a point of
reference?
AH: They can, but I think that most often while one can start with a realistic house or town, one is
going to want to tweak it. Exaggerate it for effect. Except if your game requires verisimilitude such as
a baseball field or a car racing game.
TM: When it comes to basic environmental design, what would you say about room entry and exit,
room flow (room to room), and transitioning from interior to exterior spaces? This happens plenty in
most games.
AH: Well I think that the flow of spaces has to be somewhat grounded in reality. In the study of
architecture, flow is an important aspect of designing any space and is equally important in the game
world.
That being said, sometimes one needs to tweak the reality to get a more fulfilling game
experience … I am thinking of Half-Life, for instance, where the spaces are almost caricatures
of actual physical spaces.
TM: Do you build or plan on paper before geometry construction begins?
AH: Definitely. Yes! I sketch a floor plan and some elevations and then do some rough axon metrics
(3-D views) of the space. I go back and forth … plan, elevations, and 3-D and back to plan again until I
get something polished.
TM: What is the learning process like for 3-D environmental work? How does one begin?
AH: A lot of it is trial and error. I travel to unusual spaces or environments. I draw a lot and
photograph as much as possible and then do research for the particular project at hand. Then it comes
down to modeling the environment and experiencing it first hand to see if it works. Trial and error, and
hopefully each time it gets a little easier because each time I am adding to my knowledge of what makes
environments work.
TM: Should game geometry be built with more modularity?
AH: It depends on the nature of the beast. A lot of shooters can get by on modularity, where it’s
almost more the pursuit that’s important.
But if your game required, say, rooms in a mansion, there would be a downside to having it all
modular, as the aspect of uniqueness in the spaces contributes to the feel of the game and forces one
to confront each room as a different puzzle to solve. So it really depends on the type of game you are
making I think.
TM: What would modularity mean for game production speed? Could designers build environments
faster that way?
AH: I think that is definitely a good thing to consider. When doing Mos Eisley we did it in a
modular fashion because of the twin requirements of storage capacity, which was low, and the need
for visual complexity, which was high.
By varying the appearance of the same object by using scale and texture, we were able to create
the appearance of complexity while only drawing on the resources required for a few models. Thus,
modularity was the way to go. And it is faster to build a level this way, since you are using a kit of
C H A P T E R 1 21
Previsualization
parts approach to the design. It all becomes drag and drop … like Legos … complexity from
simplicity.
TM: For game environments, what do you use as a source of visual inspiration?
AH: I hunt the Web for a general theme, and I also am religious about checking out
bookstores—especially those with a large amount of art and architecture books and magazines. You
never know when something will just slap you in the iris.
TM: What tips do you have for someone interested in learning more about building 3-D
environments?
AH: I think just traveling to see cool spaces first hand, and then drawing and photographing these
spaces, helps tremendously. And then, just hit the computer and have fun … that is what it should all
be about.
TM: What do you enjoy most about working in 3-D? About building game (or other) environments?
AH: Well, I think there is the aspect of being God without any of the downside. I get to create real
or imagined characters and environments and then watch as other people interact with them so that
they become something greater than what they were originally.
I love to create in a medium that is still flexing its wings, and I love that by creating things on the
computer I am able to experience the world around me in a more profound way. It has changed how I
look at the world, and how I react to the world around me in a tremendous fashion.
TM: With large-scale MMOG-style games like Everquest out there, how will developers continue to
deliver large environmental experiences? What will be some of the challenges?
AH: Well, I think to a certain extent there will be a modular component to this like we said, and I
see more algorithmic practices developing where an environment becomes more complex by the more
time you spend in it.
Level of detail on a massive scale. Terrain generation will become more important as will display of
organic shapes likes plants and trees, etc. With computers getting more and more powerful, it should
be no problem.
TM: With all of your experience across mediums, what three tips would you provide to someone
interested in 3-D environmental work?
AH: 1. Draw like a banshee. 2. Photograph environments and shapes in the real world that interest
you. 3. Research art, architecture, and sculpture all the time.
TM: Thanks for taking the time to talk! Offer up any tips, suggestions, warnings, or reprisals you
would like!
AH: I think traveling, drawing, and reading are constants. It is possible to get burned out because of
the constant trade-offs between what you want to create and what you can create based on technical
constraints. Get away from the computer and experience the world so that what you do in 3-D is
expressive of the world, and not just a rehashing of some computer game’s look … and have fun …
have a vision and have fun with it!
TM: Seriously, how does bullfighting relate to working in games?
AH: In the course of your career, you will see a lot of bullcrap … it’s important not to take your
eye off the bull, but you also have to watch your step.
TM: Reminds me of Rodney Dangerfield in the movie Back to School—“Look out for number one.
And don’t step in any number two.” Try and look out for more than number one, try and look out for
your team, your ideas, and your vision. Stay committed to your path. Andrew, thanks so much for the
wise words!
22 U L T I M A T E G A M E D E S I G N
Building Game Worlds
M EGA TIPS
1. As resources allow, always complete a previsualization sequence and
exercise for every project.
2. Organize and assemble abundant reference material to support the
previsualization process.
3. Commit to saving development time later by doing work up front to
establish all major visual reference points, a style bible (which is a formal
document sometimes provided by licensors and sometimes requiring its own
invention), and an understanding of how visuals will be used in support of
gameplay. This process will evolve, but having locked reference points is
meaningful and valuable.
CHAPTER 2
Level
Planning and
Building
23
you’ve completed your previsualization phase, it’s time to
AFTER gather up all of that information into a clean and
ready-to-use design document. If you’re interested in finding out how to build a com-
plete design document, you can find a growing number of design document templates at
www.gamedev.net. The specific type of information you’ll include in your design doc-
ument is dependent on your game’s genre details and platform specifics. Platform
specifics shape your design. Your platform hardware will determine in detail the largest
aspects of your design. A basic shooter for a PDA or cell phone will have a vastly dif-
ferent design orientation than a third-person action title for the PlayStation 2.
When I’m writing the “Level or Mission Description” sections of a design docu-
ment (for example, the game content specifications), I like to try to include detailed
information about the following:
Location What is the action setting? A graveyard? A tunnel system?
Time of day Is it fixed time or dynamic? Does time change visibly during play?
Weather conditions Rain? Lightning storm? Hail? Wind gusts? Does
weather change over time?
Story setup What main story points set up the player for the level or
mission?
Story continuity What main story points need to be reinforced in the
mission or level?
Mission or level summary Provide a complete general “walk-through”
description of the mission or level.
Topographic map Provide a bird’s-eye view of your level or mission.
Specific objectives or goals What will the player be asked to accomplish in
the game setting?
Subobjectives or subgoals What peripheral challenges are required of the
player, if any?
Terrain features Describe and specify. How do terrain features connect
with one another?
24
C H A P T E R 2 25
Level Planning and Building
Props What props are present? Are they static or can they take damage?
How to behave?
Hazards Define all hazards. Open graves? Collapsing trees? Electric fences?
Puzzles What puzzle efforts will be required of the player, if any?
Enemies Which enemies will populate this level? How do their abilities
mesh with the environment?
Nonplayer character (NPC) involvement What NPCs are present? What is
their function and purpose?
Power-ups and collectibles Which power-ups and collectibles are present?
What is their function and purpose?
Effects considerations What kinds of visual effects will be present? Disease
clouds? Altar glows?
Sound and music How will sounds, ambiance, and music be used to
support play?
Script lists What scripts will be required for the level or mission?
Artificial intelligence (AI) considerations What tricky “behaviors” will the
player encounter in single-player mode?
Network or multiplay aspects What must the design account for to support
network or multiplayer modes?
Team guidelines Establish grid size and workflow procedures.
Gameplay notes Collect notes and ideas from team brainstorm sessions.
This list seems like quite a bit of information to settle, and it is. In the end, success
depends on having a roadmap that helps you know where you’re going with a game.
Once you’ve tested and refined your level ideas on paper and have done your de-
sign document work, you are ready to begin the level planning and layout process.
The planning phase will help you lock down design and technical details, while the
stubbing or layout phase will help you to bring out the rough and basic form for your
level or mission. This chapter examines a number of workflow considerations for
taking your ideas from concept toward execution.
With adequate planning, the idea is to begin to translate your topographic map
and reference drawings into basic mesh or geometry form. This allows you to begin
to test, at an early stage, most of your level’s basic properties, like scaling, sizing, and
jumping requirements for an action title.
Again, as discussed in Chapter 1, it’s important to have a starting point and to have
tested your ideas on paper. Don’t begin by building blind geometry. Utilize reference
26 U L T I M A T E G A M E D E S I G N
Building Game Worlds
drawings or reference points for all they’re worth. Utilize your design document—it
helps immensely to know where you’re going. It seems simple, but it’s a challenge for
many game developers.
Building a game requires consistent forward momentum. If you don’t know where
you’re going, you probably aren’t going anywhere. For most level builders, it’s far
easier to start from sketches than to try to build up a level by holding it all in their
mind and “free-forming” it. Now, let’s take a look at what we should keep in mind as
we begin to build levels.
P LANNING YOUR LEVEL WORK
These days, the game world stakes are pretty high, and the bar of measure is seem-
ingly raised every day. Game worlds have to be built quickly and efficiently, and they
have to deliver maximum impact. With the advent of massively multiplayer online
games (MMOGs), entire new “world sections” or regions and updates must be created
and delivered to the player to continue, support, and enhance their gaming experi-
ence. You’ll learn much more about these kinds of games in Chapter 8.
In the recent past, many games were constructed using tile-based systems. Tiles are
basically square 2-D “postage stamps” that link up or connect with each other along
one, two, three, or all four of their edges. The pieces fit together like a jigsaw puzzle.
By connecting many of these pieces in the intended way, you can build a game level.
Oddly enough, even though most titles are done in 3-D today, some of the same con-
struction principles apply. You might even say that, from a design standpoint, there
is a small renaissance of old ideas applied to new technology. If you think about it,
3-D environments often can be reduced to an assembly of various simple components
to form more complex ones.
The time that you commit up front to planning out your level and environmental
work will generally save you time at many points along the way in executing it.
Diving in to build “something” is usually both costly and inefficient.
Game development teams have several decisions to make in choosing software
tools for level building. Here are some of the choices:
Build custom level tool functionality into a 3-D package like Maya
(www.aliaswavefront.com) or 3ds Max (www.discreet.com) using their
own script systems (Maya Embedded Language [MEL] or MAXScript,
respectively). You can build small- or large-scope level assembly and
editing parameters right into Max or Maya.
Use a game development package solution like RenderWare
(www.renderware.com) and its Studio toolset.
C H A P T E R 2 27
Level Planning and Building
Use the tools that accompany a licensed engine technology like the Unreal,
Quake, LithTech (www.lithtech.com), Serious Sam (www.croteam.com), or
NDL/NetImmerse (www.netimmerse.com) systems.
Write your own stand-alone editor to build content for your proprietary engine.
With the time demands placed on developers to provide new and exciting content,
many developers are searching for game construction systems that rely on prefabri-
cation, modularity, and maximum reuseability. There are also several scaling and
world grid issues to consider, and a wide array of mistakes to try and avoid. Let’s take
a look at each of these concepts in detail.
Prefabricated Geometry and Modularity
Prefabricated geometry is simply geometry built to be used and reused in certain
ways. A spiral staircase that you can use in a scene to transition from floor to floor is
an example of a piece of prefabricated geometry. Its use is obvious—it’s a stairway.
You might see five to ten of these used in different locations throughout a level. A
player might not even recognize it as the same piece used later, once it is retextured
and scaled in size.
Piping system components and door moldings are two examples of prefabricated ge-
ometry that can be reused throughout a game environment. These days, for speed and ef-
ficiency, we have to think about building up environments out of components. Maybe
the door molding can be flipped or mirrored in a scene to become something else. These
prefabricated units become something akin to the 2-D tiles mentioned earlier. They
become pieces in our toy construction set aimed at building up game environments.
Why modularity? Because we need “digital Legos.” Game environments have to
be built quickly and efficiently while offering maximum visual impact and maximum
gameplay support. Development teams are under great pressure to create stunning
worlds in short order. Although it seems obvious, only recently have game develop-
ers started to build tools that maximize efficiency with a “snap-together” Lego-like
functionality. These tools now are essential to deliver large-scale environments.
Although some people may think that modular design is a recipe for repetition,
that is not necessarily true. You can be very creative through crafty use of your con-
struction pieces. All of your favorite 2-D titles were built with modular tile sets. 3-D
offers new game vistas, but must share some of the same mindset.
Modular world design is a solid solution to several problems. If we didn’t think
modularly, and all environmental features and all prop features were entirely custom
or unique, we would
Suffer many machine performance issues.
Constantly be building geometry and texturing, extending schedules, and
burning valuable resources along the way.
28 U L T I M A T E G A M E D E S I G N
Building Game Worlds
Not be able to build, refine, and test critical game play as efficiently.
Not be able to make fairly large changes rapidly.
Possibly never complete the game successfully.
In our planning stage, we need to think in terms of modularity and reuse. Commer-
cial development schedules really don’t allow for anything else. How can we get the
most from the least? It’s a design engineering issue, really. It always has been. Games
are built with limited manpower and resources. Development teams regularly face the
challenge of figuring out how to get the most complexity out of minimal resources.
From our topographic maps and basic asset breakdowns, demonstrated in Chapter
1’s cathedral example, we now know some of the key environment components that
define this part of a level section. This is how we start to go from concept to execution.
Scale and Grid Sizing Considerations
Scale, when thinking modularly, is usually dependent on game genre. Ask yourself
and your team these questions: What are you trying to accomplish game-wise? What
about genre-wise? You’ll know the answers to these questions because you’ll have
done the planning. Yet, how you focus your modularity scale (how big or small you
section up your pieces of geometry) depends on answers to these kinds of questions.
Are maps built from entire village sections of terrain? Is a single hallway and room in-
terior built up from many smaller individual pieces? These are examples of large
modular scale and small modular scale, respectively.
If you’re building a space-vehicle racing game through futuristic cityscapes, you
can probably have larger-scale prefabricated pieces, like entire blocks of buildings in
different combinations. On the other hand, if you are taking a player through inti-
mate interior physical spaces, like abandoned mining tunnels, you need detailed
components or building pieces that can be assembled to form complexity and detail
from apparent simplicity.
This approach, in concept, works fine for first- and third-person action titles. It
works for racing and sports titles, with track or arena layout, as well as for role-play-
ing game (RPG) and real-time strategy (RTS) titles, depending on map detailing and
population. It will even work for simulation-oriented games like Zoo Tycoon and
Rollercoaster Tycoon. All of these can feature maps or levels made from geometry
construction sets. This approach also can work with some elements of flight simula-
tor design, like laying out airports and hangars. The “in flight” experience is handled
somewhat differently by the game engine requiring the creation of custom scrolling
and camera moves.
Grid systems are very important in game development. Programmers use grids to
determine location and to handle other programmatic issues. Grids establish world
scale, too. If a character is 128 units tall, then a fence that is 32 units tall may not have
C H A P T E R 2 29
Level Planning and Building
FIGURE 2-1
Global grid settings
the intended effect—Stonehenge being crushed by a dwarf. Figure 2-1 demonstrates
the settings for a world or global grid scale. This setting means that every piece of ge-
ometry snaps to every 16 units on the grid. Figure 2-2 shows an example of the top
view for your grid in an editor.
Also, keep in mind that character actions and interactions are grid-related. Char-
acters can usually jump a certain number of units. They can reach a certain number of
units. They can grab onto a ladder at a certain number of units from the ground. They
can step up a certain number of units. The grid is always game-aware, so we have to
FIGURE 2-2
Top view of grid
