« Picture Imperfect – Introduction | Home | Picture Imperfect – Floating Head Has Gun, Will Game »
Picture Imperfect – Chasing Third Person Cinema
By Corvus | September 9, 2008
Until Wolfenstein 3D (W3D) used pseudo-3D rendering tricks to put the first person shooter on the map, all 2D games used a third person perspective to show us the game world. W3D might have launched an endless parade of first person shooter successors, but it wasn’t long before they followed in the impressive footsteps of Ultima Underworld: The Stygian Abyss and truly rendered their environments in 3D.
But not every 3D game opted for the first person camera. A great many of them decided to translate what they were used to (that is, seeing the avatar on screen), and brought the third person camera to 3D. I’m not sure which 3D game was first to the third-person-perspective table, but the first game I remember playing with a cinematic 3rd person camera was Alone in the Dark. The first game I played that used a chase cam (it sounds odd to refer to it as a chase camera) was Fade to Black (the sequel to Flashback). The first 3rd person 3D game I remember actually liking was Tombraider.
Alone in the Dark, Resident Evil, and Grim Fandango all use cinematic cameras. If you’re not familiar with the term–the cinematic camera is moved from one fixed point in the environment to another, based upon the avatar’s location. This is often used to provide maximum suspense in horror survival titles. In Grim Fandango (if I recall correctly) it was often used to provide interesting or artistic views of the world. Regardless, it puts the game designer in complete control of how the world is displayed, and therein lies my problem with the cinematic camera.
From a designer and programmer’s perspective, I understand how useful the cinematic camera is. You don’t need to worry about level geometry coming between the player and their avatar. Because you completely control what the camera is going to see, you don’t need to build complete environments, or populate every wall of a room with digital bric-a-brac. The cinematic camera also allows you to fully utilize the powerful tricks film cinematographers use to maximize a shot’s emotional impact on the audience. Although to be honest, I have yet to see a cinematic camera in a video game used with such care.
To put it another way–the cinematic camera represents authoritative control and I believe the power of storytelling in games comes from reducing that very thing. As jarring as it is to have my control abruptly removed during gameplay by a non-interactive cut scene, I find it even more jarring when a camera suddenly jumps across a room as I walk, or worse, when I’m in the middle of combat. When the movement controls are mapped in relation to the camera, it’s even more upsetting. I gave up on Alone in the Dark after several attempts to use one of my precious shotgun shells on a zombie, only to find myself turning and walking away from him when the camera jumped to the opposite side of the room.
Clearly, I’m a supporter of the chase cam. I am not, however, claiming that the chase cam is a solution that’s been perfectly implemented. As Jules noted in a recent post on game cameras, we most frequently see the phrase “bad camera controls” in reference to chase cams. No one thinks to mention it if it doesn’t get in the way, but the moment the camera gets stuck on a piece of the landscape, pops to the other side of a wall from an avatar (I’ve had that happen), or refused to allow the player to focus in on some important environmental detail, it’s a serious problem.
The issue here is one of sharing control. Allowing the player to control the camera is ideal, but requiring them to control it at all times is not really acceptable either (although more so when the camera is controlled with a mouse, rather than a thumbstick). No two games implement the chase cam in exactly the same way, and so it’s difficult (for me, anyway), to lump them all together into a single paragraph of praise or condemnation.
Consider the number of variables involved in implementing default chase cam behavior. How does the camera follow the character? Is it rigidly attached to the character, as if by an iron rod, fixed in place directly behind the avatar at all times? Or, like the original Tombraider, does it follow a flexible path that corresponds to the movements of the avatar? The former, more commonly implemented, approach attempts to reveal everything to the player as soon as the avatar would see it. The latter can be used to subtly heighten tension as the player steers the avatar around a corner into a new environment. The former can be a problem when the player backs into a corner. The latter can pose an entirely different problem in the same situation.
The second decision then becomes how to handle it when the camera collides with level geometry. Rune solved the problem by moving the chase cam closer to Ragnar and turning him transparent. Tombraider would swing the camera around so that it was facing Lara Croft. The first didn’t do much for immersion, the latter was frustrating when Lara was backed into a corner because she was being attacked by a pack of wolves that you couldn’t successfully target because of the camera.
Then there’s a myriad of decisions regarding where the default camera position should be and where it should be pointing. Should it be slightly above the head or over the shoulder? Which shoulder? Should it be pointing forward or slightly downward? Should it be a fixed orientation or should its focal point be locked on the avatar, or some point in space in front of the avatar? Does the camera remain at a constant height or follow the terrain? And we haven’t even begun to address how to deal with player control over the camera yet!
When the player wants to move the camera, how far around the avatar do we let them move it? 360 degrees? 180? 90? How about vertical movement? 45 degrees? 30? Do we let them zoom the camera? If they move the camera to view the front of their avatar, does is move closer to the avatar on its own?
Speaking of moving the camera, is the player moving the camera itself–pushing left to move the camera to the left? Or is the player moving the avatar’s focal point as if the avatar were a pivot–pushing right to move the camera left? Or do you give the player a choice to invert the camera control. If so, can they invert the vertical control separately from the horizontal control? And if the the controls move the focal point, how far in front of the avatar is that point? Is moving it a one-to-one ratio, or does it speed up the further you move? Can the focal point’s distance from the avatar be changed? If the player is moving the focal point, is it represented with a targeting reticule? [1]
The balance of all these decisions either results in a successful camera system,which likely means you’ll never really have to think about it, or an unsuccessful one, which likely means you’ll find the game frustrating or even unplayable. I find that I tweak my control sensitivity and directional mapping differently for different games. This, I suspect, is because the choices a designer makes ultimately lends itself best to a particular style of camera use.
The unfortunate thing is that, with a few possible exceptions, many of these decisions are being made out of pure technical necessity and not out of a desire to use the camera in a particularly compelling fashion. Certainly, an aesthetic is being striven for (one would hope) and usability is a concern (one would also hope), but I cannot think of a game that uses the chase cam in a way that… um… increases the narrative fidelity (sorry, that’s the only phrase that would come out of my brain). Perhaps if the designer/programmer in charge of designing the chase cam were called a cinematographer, we would begin to see more artistic and meaningful implementation.
I will give special note to the latest Tombraider game, where the camera pulls around for nice cinematic shots of certain wall climbing sections. This provides an impressive sense of scale, as well as insight into the path you need to take. Fallout 3 also seems to be taking an interesting approach to the chase cam, taking advantage of the wide screen aspect ratio and placing the avatar to the side so that the landscape still dominates your view. It looks great in gameplay videos,but how well this will work in practice remains to be seen.
Whether the imbalance of text in favor of discussing chase cams is an indication of the difficulty of implementation, or merely a reflection of my own bias, is up to you to decide. Likely, it’s a bit of both. You’ll also notice that I have spent most of this post talking about what has been and not about what could be, or should be. I do have thoughts in that direction, but I’m busy assembling some supporting images and making sure my research ducks are in a row. Likely, we won’t see that post until I’m done with this week of retrospection.
What’s your preference for third person cameras–cinematic, chase cam, or some combination thereof? What games have felt natural and what games have not? What games have I neglected that used a cinematic camera to good effect? What game implemented a third person camera so terribly that it was unplayable? And I’m sure at least one of you are wondering how I could go an entire post about third person cameras and not mention Super Mario 64, a game I’ve not played. Feel free to rave about that in the comments too.
[1]More discussion of targeting reticules will either be included in tomorrow’s post, or Thursday’s post will be an amalgamation of topics.[return]
Tagged:camera, third person, video games. | 22 Comments »
September 9th, 2008 at 11:15 am
I don’t think I mentioned it on Twitter but great post!
I couldn’t help but think again of Assassins Creed – I think that game does a great job of providing the level (and type) of control you need as a player. Because you’re constantly looking for paths through the environment, you want to be able to see from as wide an angle as possible.
In comparison, I have a really hard time watching the preview videos for Mirrors Edge because the camera control is basically tied to your head! And actually on second thoughts, it’s not even tied to your head so much as your whole torso! You can’t move your head to the side to look at anything other than what you are running towards.
So, to return to Assassins Creed, placing the camera up and above Altair lets me be able to A) get the information about the world I’d miss otherwise and B) allow for a faux-separation of body and view-point much like our heads (or even eyes!) are not rigidly connected to the direction of our torso.
Sorry if this comes over as a bit ramble-y, it’s rather late y’see.
September 9th, 2008 at 11:25 am
I’ll be interested to see how well ME handles looking in directions that you aren’t moving. When your primary interaction with the world is centered around shooting targets…
Oh, but this is tomorrow’s conversation.
But, yes. I do feel it is much easier to portray physical presence with a 3rd person camera. And a game that has a primary game mechanic revolving around complex environmental navigation greatly benefits from a solid portrayal of physical presence.
September 9th, 2008 at 11:31 am
I think most folks agree that a cinematic camera is rubbish. I think most instances of its use were due to technical limitations, as you mentioned above. They’re increasingly rare to the point where even the most grievous offender (Resident Evil) has used a chase cam since RE4.
“Perhaps if the designer/programmer in charge of designing the chase cam were called a cinematographer, we would begin to see more artistic and meaningful implementation.”
But this goes right back to the problem with cinematic cameras, doesn’t it? To create more interesting “shots,” control of the camera has to be (however briefly) taken away from the player. Or is the idea that this would only occur occasionally for an especially impressive shot, as you mention with Fallout 3?
September 9th, 2008 at 11:36 am
@Nelsormensch I think we’ll find it’s quite possible, with a bit of work and experimentation, to create meaningful camera dynamics so that the storyteller can communicate important information, while leaving the camera under the control of the audience. There are, after all, great camera tricks that don’t involve moving the camera.
I guess it’s a matter of building intuitive cinematic tools for the audience, rather than the storyteller.
And, yeah! I hope the F3 3rd person camera is a viable and artistic means of viewing the Wasteland. It’ll likely be a while before I can get my hands on it (unless someone want to buy it for my birthday), but I look forward to finding out.
September 9th, 2008 at 12:30 pm
My favorite third person cameras are
1. The second splinter cell game.
2. KOTOR
3. God of War, because the camera never really got in the way except in the hateful jumping puzzles.
I tend to go for cameras that I can push around but are also smart enough to follow me.
I sort of like the RE4 style over the shoulder camera too.
September 9th, 2008 at 12:33 pm
@psu Can you articulate what differentiated Splinter Cell 2’s camera from the first game?
September 9th, 2008 at 1:36 pm
Oni had an interest solution to the camera\level geometry problem. When the camera moved through a surface that surface turned transparent so you could freely move the camera and still witness all the action going on within the confines of the level.
When it comes to chase cameras, it’s difficult to pin down games where it’s really well implemented because if the camera works you tend to ignore it.
Though I think Prince of Persia: The Sands of Time was a game that made good use of the camera, though it was still far from perfect. It used a combination of a cinematic camera in certain locations and a controllable chase camera in others. This allow for some player control of where they were looking but also allowed the designers to adjust the camera to highlight particular elements.
September 9th, 2008 at 1:40 pm
several attempts to use one of my precious shotgun shells on a zombie, only to find myself turning and walking away from him when the camera jumped to the opposite side of the room.
i know this doesn’t add much to the convo, but this is my ultimate pet peeve, and really the only thing/bug i can’t handle in videogames. with devil may cry or AC i would often just hold down a direction if it happened once and see how many times i could go back and forth between camera angles before dying (in game or IRL).
clipping – ok
framerate issues – ok
poor level design – ok
nonresponsive controls – ok
cinematic camera angles that don’t adjust for your movement when switching – the burning hot antipathy of a thousand suns
September 9th, 2008 at 2:27 pm
Great post, Corvus. I’m personally fascinated by the notion of a camera in video games, especially as it relates to a customary way of seeing we have been conditioned to accept and understand.
You may not recall this, but when instant replays first began to appear on telecasts of sports, a reverse angle shot was always noted on screen with a visual disclaimer “Reverse Angle” so as not to disorient the viewers. This, I think, stemmed from our conditioning to understand visuals that conformed to the old “180-degree rule” that was standard practice in Hollywood until the 1960s, and on television even later.
Today, we’re all accustomed to this rule being broken routinely, and with video games we’re able to break it ourselves. I’m not sure if I’m contributing anything significant to your essay here, but I do think it’s worth thinking about how media trains us to accept certain conventions, not only as techniques, but as ways of seeing that we don’t often stop to consider or analyze.
September 9th, 2008 at 2:32 pm
@Michael Absolutely we become trained by our media. And, our feedback helps train the producers of the media too. It’s part of developing a meaningful vocabulary within any medium.
But sit down to play Assassin’s Creed with the default camera settings for a bit and switch over to Fable, also with the default camera settings, and see how different an experience it is. Unless we, as an industry, actually sit down and quantitatively talk about the emotional impact of our technical decisions, we won’t ever be able to earn the benefit of developing that common vocabulary and game design will continue to suffer because of it.
September 9th, 2008 at 2:33 pm
I didn’t play all the way through the first game. It might have been the same.
In the *third* and later games, they pulled the camera in a bit tighter and it seemed to bounce more as you moved around.
The second game seemed to me to have the right balance between the camera sitting where you want it and allowing you to move it around without undue consequences.
September 9th, 2008 at 6:00 pm
I’m *shocked* that nobody mentioned echochrome nor Indigo Prophecy/Farenheit.
Alright, so echochrome is not *that* interesting but it’s only mechanic rests on camera controls and nothing else. The player is able to move the camera around the level but the focal point remains the same at all times. The goal is to obscure or create new paths for little manequins to walk on so they can reach the exit while avoiding hazards and other things.
It’s really interesting as a concept but I found it quite boring and devoid of almost any personality. Maybe it’s just me.
As for Indigo Prophecy, that game used cinematic cameras on purpose to make the whole game feel like a movie. And I have to say it worked, once you start walking around, talking to people and interacting with objects the game designer’s intent seems rather obvious. Plus the whole game is broken down in scenes… so, yeah, it was made to feel like a movie.
Also, the player has the ability to nudge the angle of the camera to look around although it’s never required (I think). Still, it’s nice to be able to look around.
September 9th, 2008 at 6:08 pm
Yeah, Echochrome probably deserves a mention, but not really in the context of cinematic or chase cam 3rd person. And while that it uses the camera as a game mechanic is interesting and innovative, it isn’t really something that applies to games where manipulating the camera isn’t the point.
I was not a fan of Indigo Prophecy’s camera. It served well enough, but didn’t really break the mold as far as I’m concerned. I still found the jump cuts to be jarring.
September 9th, 2008 at 8:06 pm
At the risk of derailing the conversation, I feel compelled to insert a brief comment about the connection between the camera and the character controls. That is, are the controls relative to the player, or relative to the camera? The later causes no end of problems, especially when the camera is abruptly moved to a “cinematic vantage point” halfway through a tricky jumping sequence.
September 9th, 2008 at 8:14 pm
@Alan, it doesn’t derail the conversation at all. I touch on the latter in my post (admittedly without explicitly referencing controls).
September 10th, 2008 at 5:01 am
It’s hard to recommend you bother with Super Mario 64 at this point, but it’s a vital title in the history of videogame cameras. It was the first to conceive of the notion of a camera object, and to give the player camera controls to rectify the problems that resulted.
Interesting examination of the subject of cameras, although I’m sceptical of your stance here, only in so much as I feel that it requires huge budgets to tailor the camera to such a great degree.
But I would be delighted to be wrong about this!
Best wishes!
September 10th, 2008 at 6:20 am
@Chris It’s exactly why I think I should play it, despite how much I’m likely to not care for the game itself. And you know me, I like my theory to be free of constraints such as budgets and hardware limitations!
September 10th, 2008 at 10:58 pm
Despite it being a ridiculously easy game, the camera style for RE4 was my favorite chase cam. It used an “over the shoulder” third person that, while not putting you directly into the action, conveyed enough of the action that on the first run through the suspense and tension had a chance to sink in… the camera never once got into a frustrating place (for me at least, but I’ve played it at least 7 times each in both normal and professional) and seemed almost natural in its implementation. Again, the only problem was the ease of beating it (although this was rectified by limiting myself to 3 saves per game… and later, no saves per game. Or attempting to play through with only one type of weapon. Though the ease of the use of the camera seemed to decrease the difficulty of the game by always allowing for what you would do in any given situation.
October 12th, 2008 at 4:08 pm
Entering this discussion a little late (alright, over a month late) but what of a game such as Skate, which utilises a fixed chase cam that focusses on the skaters board, not so much the skater — the camera sits low to the ground, with the skater and board placed either to the right or left of field as per your Fallout 3 example.
While this provided a greater focus on the feet and the board, which as a skater, you do when riding, it also made the game more difficult to play as the focal point was pulled too tight to the character, you also lost sense of direction — often approaching ledges you plan on grinding from a too sharp or too shallow angle.
Yet, the camera was far more interesting than the fixed 3rd person chase cam used by Tony Hawk’s.
After legitimate complaints were made regarding the inflexibility of the camera, the developer of Skate 2, Black Box, said they’d provide users with a range of camera options and some tweaks to alleviate the problems. However, if you were to tweak the camera too much, I believe you’d lose that focus, feel and vision that the developers originally intended, resulting in a game that might be easier to play but felt less like real skating.
October 13th, 2008 at 7:04 am
[...] platformer: de in-game camera. Man Bytes Blog heeft nu al twee delen over dit fenomeen geschreven: één over de bekende ‘chase cam’ en één over de ‘first-person’ camera. Niet al te lang en zeker leuk om [...]
October 14th, 2008 at 10:59 am
Just a bit of a word in support of cinematic cameras. Whilst some people dont like it, I found that the cinematic camera was essential in the Devil May Cry games (3 was the best). I guess it’s a usuability thing – there’s simply no fingers left to control the camera when fighting. The later games added some small degree of camera control, mostly used in the exploration/traveling parts of the game, which was much needed. But the cinematic camera is very important to the fighting parts. (the game design of DMC doesnt require aiming, so attacking things that are offscreen is not as big a problem.)
We see some serious crimes in DMC4 tho – several areas (some with JUMPS! >_< ) have cameras that flip 180 in the middle of the jump, which completely messes with you. I think those were sloppy mistakes, its a shame they made it into the game.
December 1st, 2008 at 5:45 pm
[...] There has been some interesting discussion on storytelling versus authorial control of the camera 3rd person cinema plus also the ability to either create your own trails in video -interactive video object [...]