Lose control

[Michaël Samyn]

Recording some video from Sunset intended to be submitted to a film festival made me realize how awkward a game looks when controlled in the conventional fashion.

Sunset uses the WASD-mouselook convention for navigation. And while that feels alright to play, like moving a sort of cursor through a 3D space, it's damn horrible to look at. The camera jerks all over the place, there's no aesthetic logic in the motions, many movements happen sort of by accident, or as a result of moving the gaze from one point to another. It's awkward and terrible and really makes the 3D world look much worse that it actually is.

Third person controls tend to feature a much better camera. And even if the consistency of the screen composition is kind of boring, the constant gaze on the avatar also gives something to hold on too, aesthetically. The problem here arises with the animation of the avatar. Combing the motions of its body with the control that the player has over it again results in awkward movement. Not of the camera this time but of the body. So by controlling the avatar you make it look less natural.

Second person navigation, as in point and click to tell an avatar where to go, can solve the camera problem completely by allowing cuts, if possibly adding some disorientation. But it doesn't solve the awkwardness of avatar motion entirely. The player can still make the avatar do stupid things like walk into a wall. And the algorithms that take care of collisions, avoidance and relating to the rest of the world never look quite natural.

I would like to figure out a way of playing with a character in a space that removes the awkwardness.

 - We may need to give up the notion of direct control, of camera or avatar, and to consider the character as another person. This might harm the feeling of presence, though. But maybe there's ways to compensate for that.

 - We can't rely on the computer controlling the character because that always looks awkward. So all animation needs to be baked. Unless the character is controlled by a computer in the fiction (or is awkward, or cartoony perhaps). Which I consider a serious candidate but it does limit the stories we can tell.

 - Maybe we should abandon the always-on realtime nature of the medium. A pause functionality might alleviate lots. Just stop the animation before the character starts doing something awkward. And ponder the still screen before continuing.

Any other ideas? Or examples?

[Albin Bernhardsson]

I think the slugcat in Rain World animates wonderfully for being directly controlled by the player. It's a combination of hand-made animation and physics. They've got it easier since it's non-humanoid and 2D, though.

[Mick P.]

WASD+mouse is the absolute worst. Unfortunately there's no magic bullet.

How I've approached this is to develop a kind of robot with variables on top of variables as a kind of shrine to the magnificence of walking, moving through space. I hold my own work to standards at least an order of magnitudes above commercial video games. I don't think it's practical for every game developer to undergo the same journey and this sort of thing needs to be standardized and packaged for ease of integration ASAP.

It's a hard problem, that cries out for 3D media players made first and foremost for people.

- Definitely, there's only so many fingers (edited: two thumbs+3 good fingers to be precise.) We have to use every piece of the controls in every permutation, but that can't begin to cover every kind of movement, it can only drive heuristics. The game can ask is this character clumsy or adept? That kind of thing. I get the impression you are anti-controls, but at some point that is a screensaver. I think it makes more sense to learn to use a controller, like riding a bicycle, and barring that there needs to be assistive technology. To make this easier lets assume all bicycles are essentially the same everywhere.

- Ideally the computer can control the character, but that's a technological feat. Ideally you train it or let it learn from other players. Baked animations are the bane of natural controls. Games use them because they are stupid easy to arrange for. This robot I speak of, its movements are the products of the interplay between all of the variables at once. It's combinative. You can't animate most of it by hand, but it can drive/synthesize canned animations. It's not romantic, but the results are.

- There definitely needs to be a standard framework for making games that can constantly record themselves, and play themselves back. Games that feature time-twisting dynamics can usually do something like this. Something is really lost when a game with a replay system removes the replay system in future iterations. This has always made me long for a replay/playback system for everything I do. Add some editing features and you have a useful way to generate cut-scenes, and trailers. Pause, rewind, fastforward (play back what I did or play the game for me) should be standard features for every game or notgame with a story to show and tell.


What I want to stress in this reply is the necessity for conscious collaboration towards better tools and resources for the job so that ultimately we reach a point where it's child's play to make not only non-linear digital media, but linear digital media as well. It's too much to shoulder alone. Cameras are complicated but they're still basically point and shoot. Virtual worlds are every bit as complicated as reality itself, made even more so by the shear amount of possibilities.


EDITED: I wanted to say that controlling the trajectory of the eye might require an absolute pro to make it look good enough for a trailer, barring editing technology. But that could change with the VR headsets around the corner, since they let you do a kind of motion capture performance with your own neck. Controlling movement is much simpler. Our legs are more like controllers... our eyes/necks less so (I do think we are stuck with controllers, even if we control them with our mind one day, I don't think anyone wants to play a game that you control every little detail of an avatar's movement with our mind. Even if that can be done, it would lose all of its liberating qualities.)

[Michaël Samyn]

It all sounds exciting and I hope you succeed.

But personally I'm not looking for "the same but better." I want a wholly different paradigm. I'm not entirely against requiring some learning, though I prefer my audience learn something nice (like read a story that will help them enjoy the game). But I feel design is the art of creating with what exists, working with how people are and not demand that they change.

I will abandon any interface that stands in the way of my content. If that means making screensavers then I'll make screensavers.

[Michaël Samyn]

I think the slugcat in Rain World animates wonderfully for being directly controlled by the player. It's a combination of hand-made animation and physics. They've got it easier since it's non-humanoid and 2D, though.

This is not what I meant. As far as I can tell, the avatar in Rain World is designed for the mechanics. I'm looking for the opposite. The avatar, the story, the content come first. Everything else follows.

[Mick P.]

It all sounds exciting and I hope you succeed.

But personally I'm not looking for "the same but better." I want a wholly different paradigm. I'm not entirely against requiring some learning, though I prefer my audience learn something nice (like read a story that will help them enjoy the game). But I feel design is the art of creating with what exists, working with how people are and not demand that they change.

I will abandon any interface that stands in the way of my content. If that means making screensavers then I'll make screensavers.

I didn't mean to disparage screensavers. But I guess that's the extreme of "losing control". Movies are screensavers, but I'm not sure how this relates to a game like Sunset. It sounds like you don't like what Sunset looks like, how it moves. I just want to remind you that it looks the same whether it's being demoed or played. What makes the player less valuable than the audience you'd like to impress now?

You can see how little people really care about videogames if you look at the controllers. I've never found a comfortable controller, and I'm not sure how to use the ones that exist. They do not even come in basic sizes like Small, Medium, Large. I'd gladly spend more money on something that will have a long term impact on my hands than the game player contraption itself. I'd hazard to say I cannot really play videogames as such as things are for the time being. Only elevating the medium can remedy this neglect. I think in reality you have to grapple with your body. To the extent videogames imitate reality or are their own reality you still always must grapple with a body. It's a truism.

This is not what I meant. As far as I can tell, the avatar in Rain World is designed for the mechanics. I'm looking for the opposite. The avatar, the story, the content come first. Everything else follows.

But you say you don't want story. It's probably ideal to let the character act on their own volition. But how that manifests itself is tricky. For example in scenarios with keys and doors--this is unavoidable unless you do your damnedest to avoid it and just happen to come out unscathed--having the player fumble in an inventory for a key is not ideal in my opinion. The character should act on their own volition, recall that they have a key (not necessarily a literal key of course) and make use of it... and it also follows that the character may believe they have a key, and attempt to use it, only to discover that it doesn't fit.

The description Ariea gives of your games that require you to not input anything to let the avatar do their own thing IS incredibly attractive. Still often very attractive things fall just out of reach and for this niggling reason or that cannot be had in practice.


EDITED: Cut-scenes are an always effective way to Lose Control. I guess you have to ask what is the point of being in control? For me that's the edge videogames have. It's not a dominating control, but a sense of going through the motions that has the potential to enhance the experience. It's a really complicated problem sometimes, because usually you have to deal with not only the player character doing their thing, but a whole cast of characters and extras to boot.

[Mick P.]

- We may need to give up the notion of direct control, of camera or avatar, and to consider the character as another person. This might harm the feeling of presence, though. But maybe there's ways to compensate for that.

 - We can't rely on the computer controlling the character because that always looks awkward. So all animation needs to be baked. Unless the character is controlled by a computer in the fiction (or is awkward, or cartoony perhaps). Which I consider a serious candidate but it does limit the stories we can tell.

 - Maybe we should abandon the always-on realtime nature of the medium. A pause functionality might alleviate lots. Just stop the animation before the character starts doing something awkward. And ponder the still screen before continuing.

Any other ideas? Or examples?

I was pondering this in the shower, taking the idea seriously not as how could "Sunset" be improved, but how to make a completely different kind of game.

The only thing that springs to my mind is a scenario where the player is like a Greek god seated at a chess-like table and gets to move mortals around on a board and see what that does, kind of like the Sims. It sounds like a big production, but if you want to animate everything? (you could map out all of the possible moves in advance and use other gods seated at the table to keep the number of possibilities manageable project-wise.)

I'm not sure it would be a format with broad appeal, but it might be an interesting next project since it kind of fits with the thematic direction of the games you've done before.


UPDATE: When I wrote this I think I was misreading the original post. I think I misread it a few different ways. What I was describing here is a game where the player has no control over the camera or protagonists that would not devolve into something like Dragon's Lair or David Cage's post Omikron games, although I'm not sure this concept would be a game as such, but it would probably look good on/in keeping with Tale of Tales resume

[Michaël Samyn]

This has nothing to do with Sunset.
The thought was mainly triggered by ubiquitous conventions that are also used in Sunset.

[Mick P.]

This has nothing to do with Sunset.
The thought was mainly triggered by ubiquitous conventions that are also used in Sunset.

Yes, I worked out as much ... or my subconscious did after a day or two. But I wonder is that a good example? Even though not an existing product.

[Michaël Samyn]

I'm just trying to imagine a different, more aesthetically convincing way of using this medium, specifically in terms of pictorial representation. I feel I'm still thinking too close to games, at least to videogames. I want to think closer to reality instead, or at least to figurative painting. But it's difficult to avoid thinking in terms of film, which I'm convinced is a dead end.

[Kjell]

Any other ideas?

- Automate / limit navigation.
- Disconnect navigation from the character.
- Cut out navigation ( sequences ) entirely.

[Mick P.]

(I clearly can't wrap my head around this. Too much into game/cinema, not sure if there can be anything outside that box or not.)

I deleted three posts because it dawned on me your three points seemed to be towards addressing a hypothetical game project instead of a general design principle. A stop-gap project to sidestep lingering problems you've identified? They don't seem like permanent problems as you put them. The awkward computer could always be upgraded could it not?

Typically games solve this by cutting whenever you do something. So if you open a door, cut to scene of opening the door. So it doesn't matter what the configuration was when you did the door "activation".

You can solve these problems with programmatic solutions also. You might want to look into "inverse kinematics" and animations can be blended. I'm not sure what the obsession is with super convincing games myself. The medium hardly has an identity yet and everyone is so eager to jump to doing super experimental work intensive productions.

I find realism bankrupt mainly because shadows. The current state of the art for shadows is so primitive that if Super Mario Bros. is an "8-bit" game, then shadows today are 12-bit shadows (as much as I resist calling them two-bit. Only large untessellated vertex shaded shadows have qualities suggestive of shadows. I think the future is in shadows that will themselves be models, meshes.)

Maybe it is more modernist, but there is nothing wrong with an abstract visual style, much like the Dragon Cancer game employs. And in this style, maybe a little awkwardness will go unnoticed. I don't feel like we've earned the right to progress beyond this level of presentation, and I feel like artists who do so are a little bit ignorant of history and a little too gung-ho for their own good.

Generally anything is acceptable, it can just be hard to communicate to the player how to interact.

[Mick P.]

I would like to add that games tend to tell stories inefficiently.

I definitely enjoy the first person out of body experience, but that necessarily adds time to the clock. I think how you play should ideally be up to player, which would not be difficult if games are made with preexisting platforms that bring these options to the games for no extra work.

NOW. Good movies work by suggestion. They are cut down to the bone. So you'd never waste screen time walking through a doorway just for navigation sake. Often in movies you'll notice that if its clear what a scene is about, the scene will not address the thing that it's about, not give it a moments dialogue or anything, because that's just telling the audience what they already likely know, and it would feel like a beating to just regurgitate what the audience already knows. Games can be extremely guilty of this.

In cutscenes it is practically unacceptable how literal games can be. Beating their audience up as it were. But it's a tougher question when asking, when you press X to walk through the door, what then? Do we waste the time playing the open/walk-through-door animation? Or can we just cut to the next scene? As if the door is an exit to the next scene. It depends on the focus of the game I think. And the focus can be changed if it will make the game easier to make, nothing wrong with that (a lot of the time doors represent loading screens, but ideally games should be seamless, and cuts should be instantaneous. Load screens are a failing that we can't criticize enough)