Advanced level design: borrowing cinematic techniques for gameplay
In the final article of this series we look at some location design techniques borrowed from film. We discuss setting up metrics, opportunity maps, core loops, color codes, and more.
Hi, I’m Vasiliy Skobelev, a Lead Level Designer at Pixonic, MY.GAMES, and it’s time for the third (and final article) in our article series on level design. We’ve already talked about the general terms and principles of location production, but many asked for a more substantial deep-dive — and no sooner said than done. So, get your notebooks ready because you’re going to want to jot some notes down!
Today we’re going to outline a few level design rules that originate from movies. We’ll also talk about calculating metrics, working with core loops, as well as such seemingly simple things as silhouette and color. Let’s roll!
Level design techniques borrowed from movies
Of the advanced features of level design, I’d like to talk about those that came from movies. And to cover several things at once, let’s look at the example given below, which I prepared specifically for this article.
What do you notice here? There are at least five interesting objects: sunshades like in old shops and stalls, a fence, a car, a window showing the inside of a cafe, and a person in the distance, too small compared to the one in the foreground. But why do they catch the eye?
Let’s start with the person. We see other objects of the same metrics (sizes) repeat at different distances. And in relation to them, we can determine the dimensions of the environment as though with a ruler. Probably the most vivid example from a live release is from The Legend of Zelda: Breath of the Wild. Here we have a tower, we understand its dimensions because we climbed on it and learned its size, and then we see other towers in the distance and understand how big the game map is.
And now, thanks to these “mushrooms” on the sidewalk in the first screenshot, we realize that the character standing in the distance is really shorter than the one in the foreground.
What makes the fence stand out? Why does it strike the eye? It’s because of the color. After all, yellow stands out among orange and gray. In addition, it points out a perfect path to the right. Moreover, it’s also emphasized by the line of the sidewalk, which in this case is quite an obvious guide. If both the sidewalk line and the fence were removed, the image would be much less clear. They also chose the material in such a way that the turn is obvious, but if, together with the railing and the sidewalk line, we remove the light source around the corner, then the turn won’t be distinguishable, no matter how you turn the camera.
Moving on to the cafe entrance, we see it thanks to the light source, contrasting color, and shape of the silhouette. That is, not only do we have color and light highlighting — we also see a hint of a cinematic technique called framing.
If anyone has played BioShock Infinite, there is a mission where we meet Elizabeth and watch her through glass windows. The player walks along the corridor, and there appears a smaller frame, so whether you like it or not, this is the area attracting our full attention.
In our first screenshot, due to the far distance and the small screen space, we don’t immediately look at this stall. But if we had a narrower space and more framing, then it would definitely catch the eye. But it works just fine as a point of interest in this case, that is, passing by, the player is likely to look inside.
Another example of framing is from The Last of Us 2.
Here everything is more artsy and craftsy, and the machines are buried in the wall. This is literally a “frame in a frame”, inside of which the noise level is also much higher. It also stands out because of the vertical half-doors that break the overall pattern. These are the only ones that have horizontal lines, which also don’t repeat with this frequency in this location. In addition, they are presented in contrasting colors. Seemingly prosaic things can be used everywhere, even at the stage of the earliest blockouts.
And now the least obvious thing — what’s with the car on the screenshot below?
There is a good way to get a player around the corner — the principle of an incomplete silhouette. With this, you partially show something already semantically recognizable, something very familiar to the player (especially if it’s an image from real life). The player understands what it is, but doesn’t see the whole object. In this situation, we, as a rule, tend to eye this silhouette entirely, which is a kind of “low-level validation” of the object, characteristic of any moving living creature.
Both parts of The Last of Us are built on this principle, at least their street levels. When you keep on walking, and see the car’s back sticking out so obviously from around the corner, you cannot help but follow it with your eyes to see what is in front of it, because you know it’s within the playing area, which means that this area stretches farther. Not only that, Naughty Dog still uses this technique with incomplete silhouettes from time to time — from opponents suddenly jumping out in front of players to triggering an alarm and attracting opponents in the area.
Below is an example of an incomplete silhouette — a carpet. But there is another technique I love — an open door, which works much more efficiently than many other things in design.
If Alfred Hitchcock said that there is nothing scarier than a closed door, then I would add that there is nothing more inviting than an open door.
Metrics: how to measure spaces and obstacles
When we, as level designers, are faced with the task of defining game metrics, that is, measuring the game space around our character, all you have to start with is the character– how they work and function in the project. If this is some kind of early preprod, and you know that it will change,defining game metrics is not the most rewarding thing, because they’ll have to be updated. How do you do it? What technical data about the character do you have?
At the very least, you know how tall the character is. Even if this is a first-person game, we can still roughly imagine a one-story building. Accordingly, we look at the size of our character, and thanks to life experience, we understand where the objects should be located.
In most 3D games, the character’s body is usually mechanically represented as a capsule. This capsule has only two parameters: the length, or the height of the character, and radius. Roughly speaking, this is the minimum width that you must comply with in order for the character to be able to squeeze into a particular space. In first-person games, you have to consider how high and how deep the cameras are relative to this character, because there are different techniques. If you’ve played, for example, Halo Infinite multiplayer and played with the Field of View slider, then you’ve probably noticed that when you increase it, the legs simply turn off, so that you don’t inadvertently notice they aren’t attached to a torso. Yes, Halo’s first person player controller does have hands and legs but nothing more.
In 3rd person games, usually, the only thing you have to rely on is the metrics of your character. That is, the capsule and the cameras.
Here’s the trick with the camera angle. In particular, the developers of the first 3D Mario games cut their teeth on this — kudos to them. They fell into many traps, saving others from falling, too. Well, except perhaps Naughty Dog, when they made the first Uncharted, they also fell into a bunch of all sorts of traps, which they “cured” with a fixed camera: in the screenshot below, the camera is positioned far away to understand where the character is moving to.
When you walk inside some building, you need to keep track of how much space you have for the camera, because if the character turns and the camera bumps into the wall, that’s not good. Why is it called “spring arm” in Unreal? Because, like a spring, it begins to shrink and relax back when you move away from the wall.
To avoid this “drunk” effect when the camera moves back and forth relative to the character, you need to make the passage wide enough. The Last of Us comes to mind, when you enter the hospital at the end of the first part, you have a super-wide corridor. In this space, on the sides, there are often some low objects such as wheelchairs and other things, and on the left, there are railings. Why was this done? No matter how you twist the camera, you won’t touch the wall with it. Moreover, these low objects won’t let you approach these walls. The developers focus the players’ attention on the audio and the objects at the end of the corridor; the designers cut off the other, unnecessary, things.
In most cases, the width of a passage can be calculated using the hero capsule and two camera arms (to the left and right of the capsule), plus a small margin, because the player won’t walk strictly in the middle. If the visual aspect allows you to design it in a natural way, it’s better to make the passages even wider. For example, in Star Wars: The Force Unleashed II, Kamino has super-wide passages, but it looks authentic within the framework of the setting. In The Last of Us, with a setting more grounded in realism, such a wide corridor would look unnatural.
If our game is, for example, a co-op like Army of Two or Dead Space 3, in most cases, you should make sure that at least two characters are able to squeeze through the passage. Then you should add — at least — one more capsule to this formula of two arms and one capsule.
Most of the time they will follow each other, because psychology works that way. But if they want to pass together, then they should still be able to turn the camera around and, again, not have this “drunk” effect.
If you have a PvP project, this presents a special pain. Let me point out that developers try not to make PvP experiences where characters can pass through each other, this is a technical atavism. There can be passages where several characters from the same team must pass or characters from different teams collide in a battle, let’s consider the last case first. Even if this is a first-person shooter, you will have to add more space to the arms and capsules to make the gameplay comfortable at any gameplay speed. And if it’s a third-person game, even more so.
As for more complex metrics, let’s look at two different examples, which, nevertheless, have something in common.
What should a level designer describe in the metrics in advance in order to make the job of artists, technical designers, or anyone else easier in production?
The height of half-height and full-height covers — let’s call them that. Cover depth — how long and how deep an object we can jump over. And how would we do it? And is this really all that we should consider at the pre-production stage?
There is a universal rule to answer these questions. It is quite harsh, but it helps not to forget what needs to be described:
Placement, rotation, and scale along all available axes matter for any object in the game area.
In other words, you have three transformations in the engine: placement, rotation, and scale. A level designer has to consider all three. Moreover, sometimes by “object” we don’t mean the whole object, but each of its sides. Let’s take a bedside table as an example. It’s a half-height cover item, and we need to define the acceptable height, width and depth. With allowances, if any.
This is how we account for placement and scale because in this particular case the measured sizes will influence how we move and place the object in the available space. What about the rotation? Here you need to understand whether you can tilt this or that side of the table in the mesh, which the artist draws for it. Every side. That’s how we take into account the rotation on all axes.
Among other things, when you work out game situations, you will have to look at how these entities fit together, for example, at what angle corridors connect, what door frames are there. Or, for example, if you have a half-height cover, say, a bar counter, which bends, you need to look at what angles it bends, does it have any gaps or not. You also need to consider if there are ways to overcome these gaps, and so on. This is not rocket science, but this work is still quite meticulous.
Finally, let me give some advice based on personal experience: when you develop these metrics, during pre-production, check the character’s level of readiness. This is very important. If the character is finalized, then good, you have the metrics good to go, if it’s not ready — push your colleagues to finalize it. (Developing locations for a changing character is a very thankless job.)
Don’t try to push your “finalized version”. On the contrary, discuss it with the team. Someone might notice something and ask you to correct it right away. After all, game development is a team effort.
Possibility Map
A possibility map is the number and quality of paths available to the player at key points in the level. Let’s look at an example from Dishonored 2.
The beginning of the road is the key point in this location, and there are three paths available to us (marked with numbers for clarity). Each of them also has a special design that makes the player stop to examine them. But the main thing that I would note is that we have a more prominent path, an open wide inviting ramp, and a less attractive path, a narrow, but still open alley to the left of the ramp. Finally, there is a closed path along the stairs, which we recognize by the railing on the right.
It’s also useful to consider that different players may read the possibility map differently: if I have already played the game and have, for example, a teleport ability, I will also have the opportunity to teleport on a balcony in the upper right. (But these are extreme cases. After all, not every game has a fully thought-out New Game+ replay mode.)
You can also adjust the difficulty with a possibility map. If your game features frequent forks, it’s better not to paralyze the player with this choice from the very beginning, but let them slowly adjust to this.
Again, if anyone remembers the beginning of Dishonored 2, where you walk inside the building before going outside, inside there are only straight corridors. There are no forks there, at most there are some super-short appendixes with the loot. This was done for a reason: first, the player gets used to the mechanics, and then learns full-fledged traversal skills.
It’s also important to determine how powerful an opportunity map can be, that is, the maximum paths for one fork. If it’s, for example, a PvP shooter like the first Titanfall, then the possibility map is usually reduced to four, maximum five paths, if there are many layers. If not, there will be a maximum of four paths, that is, a city intersection.
You’ll have to gradually up to this number of paths from the beginning of the game. In PvE, of course, this is much easier to do than in PvP. Now let’s see how they deal with forks in Dark Souls.
This is the first location of the third part, and here all the forks are divided into a maximum of two paths. Moreover, at the number one in the center after Gundyr, one could argue that there are three paths, but in fact, semantically, for the player are two of them: to pass along the hill or under it. The game trains the player to walk a few meters to face a fork: to the right and to the left. You get to the first ruins, and again a fork: to the right and to the left. From the second fork you already understand that the choice at these forks can be fatal. On one hand, in Dark Souls III there is a standard increase in the complexity of forks after the first location, on the other hand, we aren’t overwhelmed by a large number of paths at the very beginning, but they hint that they have forks, many of them, and their complexity is different. Among the icons in the screenshot there is a lizard on the right, and whoever stumbles upon it will know that this is a very formidable opponent.
Level rhythm or core loop in level design
With linear locations, rhythm is achieved through the use of game core loops, that is, by altering main types of mechanics.
Let’s take Tomb Raider as an example. The screenshots below show us three different types of activities that the character goes through: combat, acrobatics, and solving puzzles. And if you watch the gameplay (or play it yourself), you’ll notice that you usually don’t have three fights in a row or three puzzles in a row. They alternate, quite rhythmically.
All you have to do as a level designer to make a level rhythmic is to understand the core loop you have, how you really want to use it, how often you’d like to do it — and try to avoid repetition to make the experience as attractive and varied as possible.
By the way, in the old Tomb Raider there was a rather serious split in the target audience due to the fact that someone wanted to acrobatically whack opponents and didn’t much like puzzles, while some players solved puzzles perfectly, but were sick to death of dances with tigers and pistols. As a result, in the 2013 reboot of the series, the critical path began to consist of only combat and acrobatics. If there were any puzzles, they were only rudimentary, and did not impede player progress. Further, everything hardcore was put within side paths, side quests and DLC.
If you take a top-down look at the location from DOOM — here’s one from DOOM 2016, and a second from DOOM Eternal — you can see the rhythm right from the morphology of the screenshot. The narrow sections are for the traverse skill-check; how the player moves and explores.
As a rule, round, rectangular and square areas indicate arenas for battle. There is a small detail here also worth noting. In the first screenshot, this rule is followed 100% of the time, and in the second, the location is actually without opponents. This is a short location that takes just a couple of minutes to complete and precedes the boss fight, but it still follows the rhythm. By this time, the player has already gotten used to an alternation of narrow oblong spaces and wide ones, so the developers decided to build a location without opponents with the same rhythm. Apparently, this was done for consistency and in order to maintain consistency with the art direction, because modules (a modular constructor made of models and materials) in a game like this can be created in advance according to general rules, and then the developers might build the location without opponents with the same modules. Production costs.
Let me emphasize again:
A level designer achieves the rhythm in linear levels by alternating mechanics within the core loop.
For example, if you have combat and movement, you alternate them. If you have something else, you also include it.
Core loop of open worlds, or “POIs Diversity Rule”
From the level design point of view, we consider an “open world” as a location with a matrix structure. For example, in terms of the landscape, in Unreal Engine (excluding the fifth version, which has some new methods), this will almost always be a lined grid along two axes. Yes, there is no third axis. And this is really such a mosaic, or a matrix, in which you need to take into account the frequency of these events and what mechanics to use.
In order for the open world to be consistent and interesting (and not cause frustration), it’s worth considering what I personally call the “POIs Diversity Rule”. What is it? In an open area, you cannot guarantee which way the player will go, or rather, you can somehow influence this with the help of landmarks and other similar tools, but nothing more. Therefore, observe the following: when you stand at a random point on the map, only 2–3 objects (points of interest) should attract your attention in the visible horizon. It doesn’t have to be some high point, on the contrary, it can be a gap — a forest area can be “broken” in the middle by a river, for example, passing through the forest, and there also can be something on the river. Only 2–3 elements should be visible in one area. And ideally, at each point of interest, you should use different mechanics.
We’re back to discussing core loop again. For example, let’s look at the Metro Exodus: Sam’s Story DLC. You’re on a boat in the middle of the lake, and you see three abandoned houses around the lake. You enter the first house, there are no opponents, you loot and read notes. In the second with a horror setting, there’re only monsters with whom you have an aggressive shooting. In the third, there are no monsters, only people, and you use stealth mechanics to pass this location.
Let’s hammer home the point about the “POIs Diversity Rule”: In order to have a good core loop in an open area, at most points on the map, the horizon line must be broken by at least three points of interest with different mechanics.
Next, let’s look at some panoramic shots from the fourth Fallout and The Legend of Zelda: Breath of the Wild. The first one isn’t the most striking example, but nonetheless.
On the left we see a city silhouette, which has narrow streets and a bunch of all sorts of approaches, so even if you see a conflict, you still can fit into it in a not very obvious way. For example, from a long distance with a good gun, or using stealth. There is a separate point of interest in the form of a red rocket — a gas station, where you can collect loot. And there is a bold landmark on the horizon to the right, which usually means open space. If there is a combat clash, it will happen right away.
In Zelda, we see a tower on the left — when you see towers like this, you realize that most of the time you’ll be climbing while keeping an eye on stamina. In the middle is a foggy forest. A small spoiler: this is one of the shrines. Shrines in this case are puzzle skill checks. On the right is a volcano that suggests a different climate and some kind of hazard in the game environment. All three things use different mechanics, provide different skill checks, but it’s fun to go in any direction no matter which one we choose. Therefore, this map is good.
Intensity of shade and shape of silhouettes
Can you guess what we’ll talk about next, based on the screenshot below? Obviously not about wallpapering, although it’s indirectly related to this. We’ll talk about player perception.
We have many different ways to influence how the player perceives a location. The basic principles that everyone talks about are leading lines, breadcrumbs, framing, an incomplete silhouette, a landmark, and all sorts of other things that make up the morphology of the picture, its structure, its content from the silhouette point of view. But we can also influence player perception with the help of the color and, moreover, the forms of these silhouettes. And if I’ll just give a small remark on the forms, then I’ll dwell on the color in more detail.
Next, let’s talk about forms. You’ve probably heard about the difference between round and sharp shapes. Many people consider sharp shapes to be repulsive, while round ones seem to look nicer. But that’s not the end of the story.
Prickly shapes with sharp corners may indeed be repulsive in general, but they attract attention, too. Most often, we use sharp forms for negative objects: those that cause damage, limit the playing area, and so on. Torn, incomplete forms also work this way.
Round shapes often denote useful objects. You don’t have to look far: we’ve got round shapes of red and blue “fortune balls” in Diablo and in Horizon Forbidden West locks.
At the same time, if you want to highlight some place in a location and you make it round (and all lines don’t lead to it), then the player is unlikely to pay attention to it. Therefore, for example, in Horizon Forbidden West there are some interactive elements that attract attention precisely due to the fact that they break the overall composition with their torn and sharp silhouette. So be careful: you need to understand why you use certain things as a designer.
As for the color, we’re now approaching what was shown in the first picture of this section.
There are studies on how players react to color: red, blue, green, and so on, and the emotions it raises in them. There are some simple fundamentals that games have utilized for as long as they have existed, and one of those is the banal “traffic light”: red denotes alarm, yellow means attention, green is calm. One of the latest releases where devs overused this is Control. Dangerous zones are highlighted in red and safe zones are indicated with green lights. This is a rather basic approach, but it works.
In addition, there are more complex uses of colors to consider. For instance, because the color used in the traffic light is more orange than yellow, it causes more tension. Gaming studies have revealed that yellow can evoke a feeling of happiness. Green indicates calmness, of course, but also it attracts less attention, so if you are making a safehouse where you have a lot of interactions, the last thing you need to do is paint it green. Blue adds mystery, and science fiction actively uses it. For example, the first Mass Effect is very difficult to imagine without this bluish-cosmic haze in all possible forms.
There are a few studies on color perception but all of them indicate quite a specific reaction when it comes to something interactive like video games. If you make a game with identical rooms of different colors you will see the following. Players will run out of the red rooms very quickly, as if someone’s pushing them out of it. They wouldn’t be very relaxed in purple rooms, either. They will study the blue room, to see what could be hidden there. In the green, they can relax, and in the yellow one, they can achieve a sense of happiness. This can be useful to know.
We could talk about how patterns or shades of different intensities can be used in a room and what can be achieved with this with a million cases, but that doesn’t explain why it works this way. So, the rule for using shades, that is, how light or dark a color is, is quite simple:
The darker the shade, the deeper, but heavier it feels.
That, in fact, is all there is to it. At least when applied to space. For instance, a room painted entirely in white, as in the first picture, will feel lighter and smaller than the same room painted entirely in a dark shade, as the one next to the white.
However, this is not the best example: the point is that if we’re talking about intensity, we must still use shades that would not merge into one in monochrome. That is, if I take this image and make it monochrome (not black and white), but monochrome, or shades of gray, then we get pretty even gray pictures.
Ideally, in order to consider such things, it’s better to replace light blue with black or dark gray in the example. That would be relevant then. If you make any color image in monochrome, you will have a scale from 0 to 256. And if you want noticeable contrast, then the shade values should differ by at least half a scale, that is, by 128 points. It can be shade 64 paired with 128 plus 64. Here they will lie pretty close on the scale, if you turn it all in monochrome. But if you still play with contrast and brightness, you can get a good example. See for yourself: a completely black room seems to be “cut out” from space, but at the same time it will feel like a “little hole”. And light objects seem closer, and have more air in them, as if they are “lighter”.
A little more on this subject. You can zone the room in different ways, in terms of the last room in the third row, you could set a direction if you paint the room dark, and the two opposite walls light. If you paint an entire room light and then paint one entire wall dark, you give the feeling that this wall is either further away, or absent.
If you have a bright room, and you paint the ceiling dark, this will convey the feeling that there’s no roof. Museums often use this trick, as well as airports. That is, when you need to show all navigation to visitors/passengers at their eye level, or at least on walls perpendicular to surfaces relative to their eye level, then the ceiling, as a rule, is rendered in dark shades. Take a look at typical floor/ceiling contrast from the airport below. As you can see displays with a flight schedule really stand out.
One more thing: if we consider gradients from white to black, the darker the environment, the warmer it will seem to a person, because black is considered a warm color, and white is cold.
Finally, another example from Dishonored. In this game, we frequently come across buildings. There are very rarely any bridges between them, and the buildings have the same color almost along the entire height. This is done in order to maintain clear navigation at all height levels. A very elegant solution, and a very non-obvious one until you start analyzing the whole thing.
What else might be interesting here? The boundary of the playing area quite obviously ends with the channel. The rest of the city isn’t playable. This is often done with the help of one technique and works all the time: the color level and the saturation of the non-game area are reduced. On the one hand, this is how the developers indicate the vast world, its scale. On the other hand, we move around the level and don’t experience any dissonance, automatically realizing that there is nothing to the right, because we have a gray faded backdrop compared to the colorful location in the game area.
Read previous articles:
