This page was automatically generated from a pdf file. Download the original here: leveldesignprocess.pdf
GREYBOX PROCESS
greyboxing and concepting for level design
foreword
1
preparation
2
establishing parameters
3
about picking a concept
3
the concept
4
breaking up the experience
5
principal level layout
7
designing the hero space
8
first null space
10
second null space
13
first play pass
14
playtesting
16
tools used: Trenchbroom, QuakeSpasm
fast forward
18
engine: Quake
screenshots
26
final thoughts
29
jasper oprel, 2019
foreword
This document is a record of my thoughts and actions when starting to design a
level. Hopefully this will be helpful for anyone starting to get into level design. I
hope that it shows you that some things are universal to every design process (start
designing from the end-experience, thorough testing, break problems in smaller
chunks etc.), most of the process evolves from the problem you intend to solve. In
the process of making, you will find out what works best for you.
This project was a way to get myself acquianted with the Quake engine, so you might
see some dirty brushwork here and there. I would like to extend my thanks to Andrew
Yoder (@Mclogenog) for pointing me to the right resources and dumptruck_ds
(@david_spell) for his illuminating youtube tutorials.
1
BEFORE GETTING STARTED
preparation
It's time to undergo some preperation before you start designing anything. Learn
the basics of the engine you are going to work in and what the limitations are of
that engine. For quake this meant that I watched a tutorial series for the workflow in
trenchbroom and played through the first half of the game. After that I opened the
maps in trenchbroom to analyze the structure and the tricks they used to achieve
different play experiences. This was very helpful to get a general sense of what goes
into making a Quake map.
After that I looked at maps that are being produced today such as the products from
recent Quake mapping jams and the mod Arcane Dimensions. This is an explora-
tion of the context that this map will be released in. From it you get a sense of which
expectations you are going to adhere to and which you are going to reject.
If you are working in a larger project a very important step is to
study the rest of the project along with the mechanics of the game.
How does your level fit into the overall structure and what skills have
the players already learned at this point? What themes is the game
going for and what kind of artstyle will it be? (I actually overlooked
the artstyle for quite a bit in my process. more on that later)
2
establishing parameters
The context that surrounds your project is very important to make explicit note of.
This will give you clear goals to not only inform your design decisions, but also give
you a measure to test your design against. Without a clear vision of the design prob-
lem, there really isn't much to design.
Example questions:
Who are your players? In what context will they be playing? How does your level fit in
with the rest of the game? What do you want to communicate to them?
Here's the context for my level in a nutshell:
• I'm making a single standalone level for people to download off the internet.
• I'm designing for a game that I have little history with and all the playtesters I have
physical access to have never played the game.
• Coming into the Quake community with an outside perspective gives me the
chance to provoke.
• I have set out about 2 weeks to complete the greyboxing of the level. (detailing
and texturing will come later)
From this I set the following goals for myself:
• Singleplayer level that takes about 5-10 minutes to complete.
• A fresh, new experience that aims to surprise.
• Must be playable for novices without failing to entertain hardcore Quake players.
Very simple stuff, but it's easy to lose sight of it in the middle of designing.
about picking a concept
To give the level a clear sense of theme, I center my map around a single mechanic
or experience. Mechanics are clearly defined by the game design, teaching the
player about the use of new verbs in their toolset or showing a new combination of
existing verbs. An experience, however, is a more ephemeral concept to build a level
around. It can be a specific feeling that a space gives (e.g. walking into a large open
hall after a long tunnel) or it can be a specific concept that talks directly to the fanta-
sy of the player (e.g. the player is shrunk down to the size of an ant). While all levels
contain sections that focus on the mechanical and experiental, I have found that it
helps to pick one of them to function as the backbone of the level.
3
the concept
This map was going to be a standalone game experience. So it made sense to me
focus the concept on an experience to offer the player rather than a progression of
mechanics.
works by Krijn de Koning that shift the context of a space
For the experience concept I was inspired by a recent talk I had with the artist Krijn
de Koning. In his work he transforms the perception of a space through the addition
of minimalist geometry. More specifically he showed me the main hall of the Musée
des Beaux-arts in Nantes and how he went about changing it. The original hall in
its minimalist glory was very well known to the visitors of the museum. By extruding
existing geometry and overlaying different scales of architectural grids Krijn was able
to completely change the the feeling of the room; hiding the original space.
Musée des Beaux-arts, Nantes
I want to recreate this experience for players of my level. I want them to get to know
a space and then deconstruct what makes that space special by transforming it into
something completely different.
4
breaking up the experience
Now it’s time to break up the experience to see what kind of player/customer jour-
ney we want to achieve. From here on I’ll call the hall that we’ll be deconstructing the
hero space, since it’ll be at the core of the experience.
The dots are the highlights on the roadmap and are all related to the hero space. The
space between these dots will be normal gameplay that doesn’t involve this space.
We’re using the rules of threes and playing a bit by feel to get the pacing right and
make sure the player doesn’t get bored.
With these ideas we can finish the player journey and with it sketch out an intended
tension curve.
5
These goals will give us lots of concise points to test against later in the process. But
it's also just a guideline, not a strict rule that the map must adhere to. Looking at the
graph also gives some insight in the optimal timing of these highlights. To bend the
curve into a desired shape I move the highlight points.
There are lots of thoughts and ideas that run through your head while in the
concepting phase. Be sure to write these down in the margins of your prepa-
ration. As an example here are two thoughts I wrote down in my margins:
What’s interesting about having this clear divide between hero and null spaces is that we
can use it in the deconstruction. When we start deconstructing the space we can overlay
the grid from the null spaces to give new meaning to the same space. That way the envi-
ronment starts to intrude upon the territory of the hero space.
I thought about having the gameplay take place in clones of the hero space exclusively,
but that would be hitting the player with the concept a little bit too early. We want there
to be a bit of exploration and the pattern recognition in the player’s mind to do the job.
There might be room for this as a finale, though.
6
SKETCHING
principal level layout
Phew, that's all the preparation done and now get to open our map editor to play
around with some blocks. I like to start sketching in engine to get a sense of scale.
I like the idea of keeping the hero space as a courtyard as a reference to the origi-
nal inspiration and as a rigid space with a lot of strong lines to break. After trying
out a couple of different size rooms, I ended up with a courtyard that 512x512 units
surrounded by walkways with a width of 96 units. This 96 unit width then transfers
over to be the dimension for general hallways.
Now let's start by sketching out the first couple of rooms.
changed
original
hero
hero
null space
changed
hero
original
hero
null space
null space
null space
original
hero
original
hero
White lines indicate the general path that the player will take and the magenta lines
are sightlines that promise future play. This gives players an objective when moving
forward. We start with a very short-term one at the beginning and a hidden long-term
one when the player is working towards achieving that goal. Distances between hero
spaces are based on the graph we put together earlier.
7
With the setup of the copied hero spaces we want
to adhere to as much invisible rules as possible.
We are keeping them all on the same level, on the
same grid and connected by the same entrances.
This is so we have more rules to break when we
are deconstructing these spaces. We don't want
the player to get bored with this limited space, so
variety of traversing is key.
At the end we start deconstructing these spaces.
The first is slightly off-grid and the second has the
player entering on the top, giving access to a new
level.
designing the hero space
We still have a couple of deconstruction rooms
to go, but before that I want to take a look at the
actual hero space that we'll be deconstructing.
scale test
For this I started out with a cloister shape, similar
to the space that Krijn deconstructed in Nantes.
All of the sightlines combine into the open center,
which focuses the attention of the visitors into a
single point. It's been used in the past to bring
attention to objects posited in the center or the lack
thereof. The clearness of the focus will be interest-
ing to play with when we deconstruct the space.
8
I want the large open space of the cloister to give a moment of peace to the player.
When they experience a brief elation after intense gameplay it will help cement the
hero spaces as chapter headings for each little side adventure (null space).
To achieve this sense of the space, though, the space needed more detail. In video
games it's very hard to get away with simple geometric shapes that speak mostly
through their material. And the Quake engine especially has problems with this.
That's why we are adding trimming and arches to further pronounce the intentions of
our existing geometry.
This might already be too much detail. We should make sure not to get too attached
to it. The detail has also given our space a definite character that we can choose to
accept or reject. The associations with mediterranean architecture might be a prob-
lem in building out the null space.
This also presents us with a problem that we should have thought of before. We
didn’t settle on a theme for the space in our concepting phase. There’s no real idea
for the physical location or narrative for this deconstruction to take place in. During
the detailing phase I subconsciously modeled it after those found in French monas-
teries and churches. I think this idea of a large, medieval, mediterranean building for
semi-public use is one I want to continue in the building of this project. I imagine the
building to have been a monastery that has been continuously remodeled through
the ages to fit new groups of people.
9
DEFINING
first null space
Here we have a space we need to traverse and an estab-
lished timeframe and tension curve. We'll have a little
increase in tension and then mellow out for a reintroduc-
tion to the hero space. I think it's easier to work back from
our destination in this case.
The final room in this space is some
sort of giant open space. It relieves the
tension of going through smaller halls
of for a while, and builds up a sort of
deciding confrontation. The large open
space directly next to the hero space
also makes for a nice non-symmetrical
path (you don’t start with narrow paths
from the hero space and then end with
them again), this adds to the sense of
realism of the place.
In the narrative this could be some sort
of large mess hall, since they are clas-
sically joined to open courtyards. In this
line of thinking it immediately makes
sense for the narrower hallways to be
kitchens and food storage to provide
First mockup of a dining
the mess hall with meals.
hall that was way too large.
10
But back to the large open space. We want to give players access to the upper levels
at the end of the first null space. Let's make the staircase rather obvious and show
the second floor through the use of a mezzanine. Aiming to keep the upper stairs
rather simple at this point of the level, we just directly connect them to the openings
of the hero space. This leads to a figure 8 shape on the first floor that has no clear
resolve. We are building up to something that doesn't have any pay-off. We'll have to
make some changes to the hero space later to make sure the flow can continue.
So we'll have a kitchen in the bend and storage space on the left. Entering the null
space we give the player the choice to continue onto the main path to the kitchen,
or take a sidepath to fully explore the storage space. This diligence is rewarded with
weapons and a bit of high ground to fight enemies from.
overview from high ground
entering from main path
11
The kitchen then funnels the player to the dining hall. We tell the story of a renovated
and expanded building through the shape language of the ceiling. Basic bend with
chest-heigh covers.
We add an extra connection to the top floor to give players the option to clear out the
dining hall from the mezzanine and give the first floor a more interesting shape.
Next we change the hero space a bit. After entering from the first floor, the player
is hit too suddenly with a wall blocking their way. The solution is expanding the first
floor. This means that neighbouring hero spaces will encroach onto each other. This
coincidentally provides us with a solution to the figure 8 problem we were having
before. We join hero spaces at any level higher than the first. This also gives the
player two different paths to continue onto the next area.
We add some detailing to sell the experience more, fill some leaks here and there
and that's the geometry for the first section done.
12
second null space
The player is now introduced to the first hero space again and then immediately to a
mirrored version of this room. That gives us the repetition we wanted in the tension
graph. The original plan was to have this lead to another standalone hero space that
would then slowly lead to the first deconstructed area. Some walking around shows,
however, that fully exploring the map already makes players quite familiar with the
hero space (we do have 8 possible entrances). That's why I'm pushing the decon-
struction a bit forward with the next space already being a first step to deconstruc-
tion. We'll repeat the space again, but next to it place another hero space on a lower
level.
But let's talk about the null space that connects our first hero space set to our
second. We want the player to follow the narrative between the two and see how
they are connected. However, we have quite a bit of physical space to cover between
the two. That's why I chose for a large open space in the outdoors. There's a direct
sightline between the two and the entire space that separates them is instantly
comprehesible by the player. It also is nice breath of fresh air after all the indoor
spaces.
13
The blocking started out with covering all the exits of hero space 2. A simple stair-
case to the first floor in case players took the lower path. And also introducing
outside to the player before we allow them to experience it. If the player just contin-
ues on their course they end up in a small room that serves a vantage point over the
courtyard. Here they can activate a button to open up the gate.
In this vista the player first sees a large clocktower to indicate where they need to
go. The entire courtyard is designed to feed into this clocktower, while still leaving a
couple of key interest points for the player to interact with on their way there.
first play pass
Now we have a lot of space with a designed interaction in mind. It's time to fill it out
with enemies, health and weaponry to nudge players along this path. This allows us
to properly playtest the space and see if the assumptions we made about the experi-
ence were correct.
In the building of the spaces I already explained what kind of experiences I wanted to
build. I'm mostly following those decisions and the tension graph to place interacta-
bles, so I won't be going over every room again. Here are a couple of interactions
that I thought were most interesting.
14
Level entrance has extra ammo and health
for the player to retreat to. Rewarding their
memory.
First hero space is a purely spatial experi-
ence with only a single soldier slowly wander-
ing in.
You can see (and foolishly engage with)
the shambler that faces away from you in
the second hero space. Foreshadowing an
increase in enemy strength.
Enemies in storage room are all faced
towards the center and a very easy to sneak
up on via the back entrance.
Back entrance has a you stocking up on
supplies while your eye is drawn to the
window (movement from a caged spawn).
There you see a dog walking to the door to
attack you.
Shalraths are exclusive to the outside area.
The first one has an area that's designed so
that the player can easily counter the homing
bullets. While fighting this one, the second
one engages from a less favourable position.
Nailgun right before the player's encounter
with lots of squishies in the dining hall. Elat-
ing relief of tension.
15
TESTING
playtesting
At this point there is enough of the level finished to sit myself down for an actual
playtest. This consists of clearing my mind and playing through the level from start to
finish, writing down issues as I run into them. I categorize these problems and ideas
following the MoSCoW principle. Each issue gets one of the following labels:
• Must fix
right now
• Should fix before testing with others
• Could fix if I have enough time
Would fix if I had unlimited time and budget
This helps to give me a clear priority of what I should change first and if things that
I notice during gameplay can be fixed within a reasonable timeframe. Before I want
to involve outside playtesters I need all my M's to be resolved and ideally my S's as
well.
my testing notes
at this point
16
One of the first things I addressed was a Could fix, but it took so little time that I just
wanted to quickly get it over with. Basically I set the vantage point over the garden to
be at an angle. This gives a better overview of the space and the tower. It much more
clearly sets the tower as a goal for the player. It also is the first time I break up the
grid, giving the space more meaning in the experience of the player.
The upper gallery also needed a rework. It allowed for players to move past the the
hero space much too quickly. I liked the idea of giving this option to advanced play-
ers, so in place of the boardwalk there is now a trapdoor that closes behind the play-
er. Forcing them to fight the Shalraths with less weapons at their disposal.
In hero space 2 it was very difficult to get the right interaction. At first I wanted the
player to see a shambler at the start of the stage, which they could later explode
from above. I found out, however, that the shambler is strong against explosives
and it would wander around too much after engaging. I tried to constrain its move-
ment and give the player a super nailgun to fire more accurately, but the magic of the
moment was lost. I ended up replacing the shambler with two hell knights that need
to work their way down from a pedestal. We still get foreboding and comeuppance
this way.
17
REPEAT
fast forward
We're now a week into the project and about a quarter way through. I don't have
enough time free time left to keep up writing down a concise play-by-play of every
decision, so from here on out I will describe my process more globally. Fortunately
for you, we have already gone over every individual aspect that composes my
process.
I also took a little summer break at this point to participate in a game jam in South
Korea. Getting a chance to come back to your work with a fresh mind is always a
delight.
day 1
I’m alternating between playtesting and improving existing spaces and
working on adding new spaces. It helps me find a flow in the ways players move
through the space and find the flow of the level in general.
Added more cover opportunities
to the kitchen and seperated the
enemy layout in clearer waves.
Player is encouraged to take the
upper path through enemies and
is rewarded with extra weaponry.
Drops down after this "bonus."
18
With the previous path rearranged,
I had to look critically at every-
thing that was connected to it. The
upper floor of hero space 2 was
now a dead end. It made sense to
move the button for the large gate
to the focus point of this section.
Hero space 3 is put together. A
simplified room path. I want the
player to be able to move through
it quickly, since they are so familiar
with it right now. Broke one of the
facades to indicate the critical path.
After this the player plummets into
deconstruction. They fall down a
large chute of repeated hero space
floors.
day 2
I wanted to move forward a bit, so I ended up blocking out the last
remaining section of the level. Going for a large drop and a danker dungeon atmos-
phere.
Large closed of gate with side
chambers.
19
One of the side chambers opens
up to the atrium.
Large staircase spiralling down
to the ultimate destination. Small
hidden path gives player access to
area between floors.
Shrine pops out in the center
after the objective here has been
achieved.
More clear divide for the two
sections in the garden. This area
isn't flowing nicely. Removed the
enemy encounters here for the
time being.
Playtested with some other
people for a bit to test out different
encounters.
20
day 4
The fourth day, in which a miraculous breakthrough is achieved. And I
feel so proud that I want to explain it in detail.
I was struggling with connecting hero space 2 and 3. It was just too much of an
unexplained dead zone. A long straight walk without anything interesting to do for
the player. There's friction here in the way tension is heightened and released. The
section directly before the garden has variable tension with different paths that lead
out into the open area. This means that if we release the tension on the path to a
new area we might give a breather to one player and horribly bore another.
I liked having the moment of pause, though, so I went forward with accepting the
peace as a destination instead of a transit path. That's when solution clicked for me
in the most satisfying way. I flipped the third hero space and all the of the boss cham-
ber on its axis and moved it all the way to the left to directly connect to the second
hero space. Now the player has to pick up a key in a dead-end garden to continue
to the the third hero space. It's up to them to decide when to continue. Making it less
utilitary changes the entire context of the garden to a more pleasurable space.
The connection of the hero spaces also makes for a nice square layout. I was afraid
of doing this during the initial blockout, because the route might become too predict-
able. Having ran the route a couple of times now, I can say there are enough digres-
sions from this path to engage with to keep the player guessing.
21
Overview of the newly made
square layout.
Quite a bit of time went into prop-
erly moving and welding all the
geometry. The chute underneath
the level start was a very tight fit.
Designed a new encounter for
the third hero space that consists
exclusively of zombies and explo-
sives.
day 5
Fleshing out the last areas. Also ran through the level a couple of times
with a new tester, added some pickups everywhere and moved some enemies to
better foreshadow their presence.
Hidden space on top of the grate
and a couple of little fixes to make
the garden more interesting to
explore.
Scripted the final teleporter to
move up through the floor at the
end of the level.
22
Designed the "final boss room".
Players walk through four differ-
ent gates to summon the final gate
(and enemies in the process).
The place is now more heav-
ily guarded to build up the final
encounter. Large heaps of "some-
thing or other" to indicate the end
of the path and a used space.
day 6
Mostly a day of maintenance and preparing for texturing. I started doing
large light volumes first to give a broad impression of every area.
Did lighting for most of
the areas to indicate
where to go and where
to look.
23
Replaced the large bounding
skybox with smaller local skyboxes
for optimzation.
In the process I'm finding leaks in
the level and temporary brushes
that I intended to delete earlier.
Started experimenting with texture
wads to use. (Ended up going with
Hexen2 and hipnotic.)
day 7
Texturing! It has a very slight connection to the topic of greyboxing, but it
just makes that much nicer to play for people outside of the project. That's why I took
out a day of texturing before getting some online consultancy.
24
At the end of the day I sent a beta version of the map to some online friends and
shared it on a Quake Mapping discord server and the func_msgboard. They sent me
back some demos of them playing the level and I got some valuable Quake-specific
feedback about lighting and indicating key use.
In their opinion I could also up the difficulty a bit. Since this was one of the design
goals I hesitated to bring about dramatic changes on this front, but I decided to add
some enemies in places where it was optional to fight them.
notes from
the demos
25
The feedback I'm getting from the playtests are now more and more
focused on the polishing side of things (detailing, texturing, attuning
lighting). I'll keep polishing the map for publication, but for the sake
of this document I'm calling the map finished right here. Otherwise
we would stray way too far from the subject of greyboxing.
screenshots
nantes.map, 31st October 2019
26
27
28
CONCLUSION
final thoughts
The map came together in a total of two weeks with a couple of breaks in between
and I'm fairly happy with it. I would have preferred to have settled on a setting/theme
beforehand and there are some sections that could use some iteration to let them
really shine (the section between the chute and the boss room is very linear and not
very engaging). All in all it was very fun to work in the Quake engine and Trenchb-
room is a delight. I'd heartilly recommend it to anyone starting out with mapping and
level design in general.
I remember looking for detailed documentation of design decisions when I was just
starting out with creating levels. When starting to learn something new I first like to
emulate someone else's workflow before making it my own. However, I found it very
hard to find accounts of the day-to-day of designing. I hope that this write-up proves
useful for someone that is looking to make a start designing maps.
I would also like to state that this is how I personally like to design levels. My back-
ground in architecture makes me want to design the space first and then find interest-
ing interactions within it. Others might like to think of engaging encounters first and
then puzzle them together. As I said in the beginning, there's not one way to go about
it; you have to find a style that works for you.
Thank you very much for reading!
If you have any questions or comments, feel
free to email me at: jasper@oprel.work
Or you can find me on twitter: @jasperoprel
29