Games have their own visual rules which are often contrary to other kinds of camera. Camera design dictates how players see into the game world, and ultimately how they play, so without good camera design your whole game may end up unplayable.
It’s very difficult (read: expensive) to change a poorly designed camera without rewriting a game entirely. This is why I consider gamatography (like photography, but for games) to be the first design task on any project, the first spec to be written and the first code that needs to be prototyped.
There need to be clear rules that will govern the camera throughout the whole of the game, sooner rather than later. Do it right and camera acts as a foundation on which you can build. Do it wrong and you’ll create a user experience nightmare.
A camera is a player’s vantage point in a game, her eye into the world. The vast majority of games use one or more of eight basic types.
1. Fixed Point: A fixed point camera neither moves, scrolls nor rotates. Many simple games like Tetris or Bejewelled use one fixed point. Larger games like Jet Set Willy use a series of fixed point cameras, one for each room. Adventure games like Blade Runner also use some fixed point cameras, but the perspective differs from scene to scene.
2. Rotating: A rotating camera is unable to move or scroll but it can turn. Adventure, platform and survival horror games sometimes use rotating cameras for especially grand rooms. An alternative use is for tight spaces in some otherwise tracking-based games (see below), or for special modes.
3. Scrolling: The game world is a flat two dimensional grid and the scrolling camera moves on a plane parallel to it. Scrolling cameras can move horizontally, vertically, or both depending on the game. Many 8-bit and 16-bit platform games use scrolling cameras, as do shooters.
4. Movable: Sim and strategy games often use a movable camera. The player looks down from overhead and moves the camera around the world using arrow keys or the mouse. Some games, like World of Goo or Angry Birds, use the same kind of camera but look from the side rather than above.
5. Floating: Floating cameras are movable cameras with no fixed orientation. The floating camera adds rotational control allowing full 3D navigation. The Total War games, Black and White and the Homeworld games use floating cameras.
6. Tracking: This camera tracks the doll along a pre-defined line in 3D space. It may rotate, speed up or slow down, fall behind or even move ahead of the doll as required but the player has little or no control over its movements. Tracking cameras are often used in linear action or platform games like God of War or the Kim Possible DS games.
7. Pushable: Pushable cameras occupy a default position (usually behind the doll) when not controlled, but the player can push them using a second thumb stick or mouse. The camera then rotates around the doll. This kind of camera is very common in modern games. Pushable and tracking cameras are often casually grouped as ‘third person’ perspective.
8. First Person: First person perspective combines the camera and the doll, which means the player sees through the doll’s eyes. He uses one set of controls to move the doll’s body and another to turn its head. One limited variant of first person perspective is the bonnet camera in some racing games. Another is sniper mode, where the player can aim but not move (or not move well).
Rule #1: The Golden Rule
Cinematography is about composition of lights and camera to convey plot or emotion. Though many games try bring the two together, the requirements of game cameras are often the exact opposite.
Nobody really wants to see the buttocks of a space marine blocking their view, nor 100 jump cuts to bits of action just because an animation looks cool. Good gamatography has to start with what players need to see rather than what story developers want to tell. How the player sees the game, and how she interprets what she sees is a creative constant (called lensing) that affects everything.
Cameras are a key part of that constant, so the golden rule of good gamatography is: If your camera doesn’t help the player play then design one that does.
Rule #2: Surrounding Space
When playing a Mario game the player is not really looking at Mario. He is looking at swinging platforms, star coins, turtles and mushrooms. In a racing game he is not really admiring his own car. He is looking at other cars and the track, anticipating the next turns. In a strategy game he is not really looking at his own troops. He is looking at opponent’s troops, the landscape of battle and the resources on the map.
During play, a player is like a ghost possessing a machine, in this case a doll. The doll is just an instrument in the world, and conduit of information, so the player is not looking at the doll. He’s looking at the space surrounding the doll.
The game world is what gets most of the play brain’s attention because it’s where allies and enemies, switches and objects, puzzles and possible solutions present themselves. The surrounding space of an action game is usually the immediate area surrounding the doll plus a cone extending out from its front. In a strategy game the surrounding space might be a doughnut shape around your units, or the battlefield between two armies. In FarmVille it might be your crop fields rather than your farmer.
If you try to force players look at their doll in order to make an emotional connection, you just block their view of the surrounding space. They see the back of their car, the butt of their doll, or have 25% of their viewing area blocked by the doll’s shoulder. The only emotion that that breeds is irritation.
So the second rule of camera design is to make sure that the camera can always see the surrounding space. Wherever the primary action of the game is likely to occur (relative to the doll) must be visible, and if this means that your level or track ideas must be reshaped in order to work with the camera then that’s what you have to do.
Rule #3: How Much Agency?
Cameras often give a degree of control to the player. Moving perspective becomes a part of her play experience, which makes the world to seem like more of a flowing environment than a series of linear challenges.
However if you’ve never played a first person shooter, the prospect can be daunting. The player needs to get used to the idea that she can control her upper and lower body independently of one another. It takes a lot of practise to be able to perform manoeuvres like strafing confidently, which many interested gamers have acquired but novices lack.
Similarly, a sizeable part of the population finds some kinds of game disorienting. Some strategy gamers much prefer movable cameras to floating ones because they otherwise end up getting lost in their map. Some people naturally sense which way is north and can easily find their way through even the most byzantine cities. Others get lost easily, even in grid-like cities like Chicago, without points of reference.
An easily disoriented player needs grounding. He needs to know which way is up or down, north or south, left or right at all times. However for every disoriented player, there is another that finds a lack of camera control frustrating. He wants to be able to rotate in 360 degrees in multiple axes.
The tension caused by lack of expertise or general disorientation versus the desires of expert players leads to the third rule of camera design: Know your audience’s tolerance for camera control.
Don’t try to foist an alien system onto an audience that has no patience for it because you’ll make them feel stupid. But don’t dumb down camera for a crowd of ninja players because you’ll make them feel patronised. Figure out whether your market wants to be empowered or held by the hand, and then design a camera that speaks to their needs.
Whatever you do, don’t try to do both at the same time. That way lies madness.
Rule #4: Distance to Action
While a videogame might be a 3D world, both of your eyes see the same image on screen. So it may seem to be in stereo but you are actually looking at it in mono. You can easily judge distances between objects that are equally near, but not if they are moving toward or away from you.
A 2D game often feels comfortable because it all operates on the same plane relative to your eyes. However for a 3D game, not so much. Judging a jump in Tomb Raider is much harder than in Sonic the Hedgehog because the jump is in three dimensions. You can’t really tell how far away that ledge is for Lara, so you have to jump slowly and hope for the best. Whereas in Sonic you can complete ten confident jumps in a few seconds.
Depth is less of an issue at distance, where the difference between stereo and mono is negligible. First person shooters work really well because enemies are often far away. The only accuracy that’s required is effectively two dimensional (place your reticule on target, then fire). However when enemies move in close in a first person shooter, the player becomes acutely aware of how little information he actually has.
So the fourth rule of camera design is to compensate for distance to action. If the main action is far away (as in a first person shooter) then aim is more important. If it’s close-up (as in a fighting game) then the ability to judge close distances matters more. If you need to know a lot of information from all over a battlefield then the ability to see over the whole field is vital.
Don’t try to force close or faraway perspectives simply because they look cool or cinematic. Players will not care if they can’t actually play.
Rule #5: Light The Way
Humans normally have 170 degrees of horizontal vision and 100 degrees of vertical, but a screen is lucky to occupy 100 degrees of horizontal and 50 degrees of vertical. As she plays your game, the player is actually looking at a world framed by the borders of her screen, like horse blinkers.
For most cameras this means that the player is less aware of the game world than you might realise. Even when she rotates a floating camera or turns her head in first person, she’s still seeing a windowed version of the world in front of her. So it’s very easy to blindside her.
Blinkered vision has many implications for game design. Mainly it means that the camera and lighting need to lead rather than follow. That’s the fifth rule of gamatography. What does it mean?
One of the most common techniques in cinema is when the camera lets itself be pushed by the actor. As he walks from left to right, the camera will only start to follow him when he reaches the right-hand third of the screen, appearing to allow itself to be pushed.
In games it’s usually the opposite. As I push my little flying saucer rightward in Insanely Twisted Shadow Planet, the camera moves immediately and more quickly than I do, effectively pushing my little ship over to the left of the screen and lighting the path ahead. The game also structures most of its new content in the direction of new camera, so I’m rarely attacked from behind. Attacking me from the direction of movement feels fair because I can see it coming. Attacking from behind (where I can’t) does not.
Floating, pushable and first person cameras all have weaknesses when it comes to lighting the way. Instead, games which use those cameras lighting, signs or indicators within the surrounding space to nudge the player’s attention. Valve, for example, are experts at using lighting to continuously hint to the player. A cool poster on a wall, a sign saying exit, a suspicious blinking light and so on all lead the player down the intended path without ever grabbing control away from him.
Other games use arrows or markers. Halo uses the convention of the Cortana character guiding the Master Chief to mark goals on his HUD and keeps its single player level designs linear. Grand Theft Auto uses a system of indicator arrows and points that overlay the 3D world of the city to tell the player where to go.
Pushable cameras are also frequently constrained by walls or other obstacles, which can make leading with the camera difficult. So those walls need to be removed or extended. Otherwise this leads to behaviours like the player moving her doll into non-optimal positions just to get the camera to pivot. If your game is a third person action adventure with tight corridors, use a tracking camera instead.
Rule #6: Transitions
Modern games commonly use more than one type of camera. Halo uses two: first person when you are the Master Chief but pushable when you are in vehicles. Ico mostly uses tracking cameras (and the player has limited rotational control) but uses scrolling cameras on wall sections.
There’s nothing wrong with using multiple cameras in this way. The complex part is managing the transitions. Cinema often uses cuts, however games are usually better served with movement rather than cutting because cuts during play are disorienting.
Halo uses a one-second animation sequence of the Master Chief getting into or out of a vehicle as an opportunity for transition. The camera moves out from the back of his head to a vantage point behind the vehicle, and the reverse happens when he gets out. That second is enough for the player to comprehend the change of mode. Eventually he barely notices it. Perfect.
Sometimes movement transitions won’t work. When passing through a doorway in third person action games, it is often impractical to move the tracking camera under the door and back up again. The game would have to pause for more than a second or move the camera through a wall. Neither is good, so the camera just cuts instead. It might fade in and out for a second, as in Ico, however a fading cut is not always ideal.
One way to reduce the disorientation from a faster cut is to preserve the direction of movement from the previous camera. It’s also a good idea to define a rule that no action happens immediately on the other side of a doorway. It is unfair on the player to expect him to reorient and adjust to a challenge all in the same second. Give him a moment to figure out which way is up or down before attacking.
Overall, the sixth rule of camera design is this: Always ground your transitions. Give the player a second or two to get his bearings before moving onto the next task.
The goal of good gamatography is for the camera to feel so natural that it is a part of what the game is rather than a system to be fought against. Nailing it early informs many decisions, such as what detail level models need to have and whether they need to be fully 3D. It also plays an integral part of level design.
Ideally, you should reach a point where the game camera becomes robust and works consistently throughout the whole game. Rather than being exciting, the game camera should be predictable, even boring. Once you have that, camera is your friend rather than your enemy, and great things will emerge.
There are exceptions to all rules, but the above six are a good start. If you can visualise what types of camera you want to use and how issues like transitions or surrounding space will be handled before you have built a lot of your game then it will be time well spent.
Above all, don’t get trapped into thinking that the camera’s job is to make players feel. That may be be true in other media, but not in games. Gamatography’s job is to let players play and establish emotional connections on their own. So get your inner film director out of the way and focus on what players need first.