TIGSource Forums

So I've got this narrative game, right?  I also fancy I have a certain amount of talent as a writer, both in the sense of being able to put a story together and being able to convey it through language.

Who the hell wants to read in a game, though?

But it's not that, so much, it's this -- I've got a specific sort of feel I'm after in the gameplay experience.  Quiet.  Subtle.  Visceral.  Unbroken.  I want to demonstrate ideas visually.  I want body language to do most of the talking for my characters.  But that's hard to do.  I'm a good programmer -- a damned good programmer, but I'm having pipe dreams about using IK systems and custom body rendering to do all this and maybe it's a little far-fetched.

I've got this vision of my game as using no text, or cutting its use of language down to tiny morsels -- punctuating the most important moments in the game with a single sentence.  "Luka sees more with one eye than others do with two."  I'm in love with the idea of speaking a universal language.  But I'm throwing away a skill I have, and a very powerful method of communicating ideas.

This isn't so much a question as a regurgitation of internal conflict, but bear with me.  Recently I used Fuck This Jam as an opportunity to break away from the stagnant indecision I'd been wallowing in, and prototyped one of the game's mechanics as a sort of text adventure, which allowed me to develop its structure and gave me some hope about the project.  A 360-line script parser and a 160-line runtime are enough for me to tell a story, and for once my content exceeds my code.

At this point the most reasonable course of action seems to be prototyping the game in text and moving towards my more ambitious visions after.  I feel this may have its pitfalls if I grow too dependent on writing, though, and that the conversion may be painful.


So I ask you, writers of TIGSource:  If I think I can tell a story without words, is it worth forgoing my skills as a writer?  Doesn't speaking the language of action resonate with what a game is?  Or am I simply crippling myself with unnecessary restrictions and technical goals?

I don't consider myself as a writer, but I will say that the language of action is maybe not so different from the written language, it's all a question of rythm and of choosing what elements to show and what elements to suggest.

Suggestion -- yes.  I have a million ideas about that.  My characters largely have a sort of placid nature, and I like the idea of making their emotions more implied than express in many situations.  While that's not quite expression through action, I feel it works in much the same way.

My game has no text and a lot of story. I'm planning for most of it to be communicated through character behaviour, sometimes meaning just animations.

IK systems isn't even in the right ballpark. They wont solve your problems for you. They are just tools. What you need is to build a vocabulary.

For example: FF6 chars can fall over, sit-down (like wounded, feinted, tired), wave a finger, raise their arms (in exclamation/celebration), and lower their heads. They can also blush, walk, walk slowly, and run. These are a set of things used to re-enforce narrative elements. In a sense they are their own vocab. They have a range of expression.

You can teach the player to read very complex mannerisms. One example from my design is like this.... The NPCs communicate complex ideas through very subtle maneuvers, so when one char points his finger at you he does so in a way that communicates elements of a wide range of his feelings at that moment. There is no strong system of interpretation in reality for this kind of thing so I'm going to build one and teach the player it.

The design is complicated, but I think I can give the gist with an example. The whole game will consist of challenges during which the player will have to master the meaning of his companions'/enemies' behaviour. Simple system: NPC acts, player responds, NPC/the-game responds in a way that indicates the level of success achieved by the player i.e. did he interpret correctly or not. This stuff happens non-stop for the entire game. That way I can teach the player to be aware of subtleties in NPC behaviour.

Simple example for you.... Say your chars have 2 "axes" of expression: love/hate and anger/sadness. You could create 3 walking anims for each axis: high, med, low. So you'd get this:
  . high love                      | high anger
  . medium low/hate         | med anger/sad
  . low love (high hate)      | low anger (high sad)
So 6 anims total.

Then you can blend the anims using some skills - this is your IK equivalent. So you can blend in two ways: you can create a gradient for each axis, creating something like "32% love," or "95% sadness;" and you can blend each axis together, creating something like, "high love _and_ medium anger/sadness."

Make sense?

Like: high love mixed with medium anger/sadness - shows a char with a lot of love and no particular anger/sadness.
Of course you can mix the gradients too: 72% love, 10% sad.

And you can repeat for other kinds of anims... attacking, gesturing etc.

Then you can add stuff for concepts - like sign language (!) - like a circle motion with the head to represent hunger.

--

Speaking through action _does_ resonate with what a game is. The single most powerful method of communication for an NPC is through their mechanics - the ones the players are always engaged in, the reason they're playing.

Do not write a story _then_ try to translate it. The story you create in an action-expressed way will be limited by the vocabulary you create. Mario jumps and runs, then his entire game is about jumping and running. You can create a diverse story with action-only expression, but you must know what those expressions are before creating a story out of them. Imagine designing Mario levels assuming Mario can do anything then deciding he can only run and jump, and in only a very particular way.


  

(http://sphotos-c.ak.fbcdn.net/hphotos-ak-snc6/5412_99558307479_6418602_n.jpg)
(http://sphotos-e.ak.fbcdn.net/hphotos-ak-snc6/5412_99566032479_6025857_n.jpg)

Sorry if it's in french but that's the basics, then you would need the book "animation survival kit" and making comics from scott mccloud and you will have all the basic bricks

I think there are two issues going on here, and it's what you mean by action.  Communicating primarily with visuals and movement is something that resonates with what a movie is.  It can be used in games, just as words can be used in movies, but I think the equivalent for games would be communicating story through interactions and mechanics, which is another sort of action.

That's not to say that communicating through visuals, action and sound rather than relying on words isn't a noble goal.  Here's a movie clip for inspiration:  http://www.youtube.com/watch?v=fx15s_HAkgU

Yeah, they are different. Mechanics are just so much more complicated....

Graham:  This topic was meant to discuss action as it plays into interactive storytelling rather than the technical implementation of that action, and that was the nature of my question.  Regarding the specifics, I plan on doing things subtler than intentful gestures and explicit expressions.  I want logic for eye contact between characters (or the lack thereof), body posture and the like.  People are always doing things with their hands, as you'll notice if you go to a public place and people-watch.  Always doing something that gives them tactile feedback -- I want to model that sort of thing too.  I don't think a set of blended animations will suffice, so I'm interested in investigating a more semi-procedural approach.  Hence my mention of FK/IK systems for modeling all of it in addition to intentful motions.  But again, this all gets a bit far fetched.


Gimmy:  I'd definitely do well to look into some books on animation and body language.


Alec...

I've got some ideas on the interactions in this game which I've been thinking over for a long while.  The body language I mention isn't something I envision as a mechanical part of the game, but would be ever-present.  You wouldn't control it directly, but you'd certainly have an influence on it.  The game itself is very much about interpersonal interactions and self-exploration on the part of the protagonist, and these things are manifest in the mechanics I've designed.  The player would be capable of reminding the character of something which makes him frustrated or wistful, but wouldn't be able to make him act that way directly.

I think the best writers already tell the story through action.  For a book, they're limited to words.  For a game, you'd actually show the actions instead.  Isn't the advice usually, "Show, don't tell."?  It can be taken literally, but it is meant figuratively.  You don't say "Jack is cool."  Instead, you write about things Jack does and let the audience decide he's cool.  Likewise, you don't tell a game story in text when you can show the action on-screen. 

I think your overly-elaborate IK schemes are actually a whiff of the future.  LA Noire was lauded for their facial changes.  That's just the start.  Having full-body body language will be important in future realistic games.   (There will always be room for stylized games, though.)

@ OP

The reason I brought up that stuff is b/c it seems like you're slightly underestimating the problem. It's like you're asking what a level design will look like before you even get the gist of your mechanics.

Whatever vocab you create out of action will be a very small subset of an actual language. When you understand what the subset is, say by exploring how you might build it, then you can ask real questions about what sort of story you can tell (and how you might do it).

If you do have a very general question about story through action I don't know what it is. The general answer is yes, any story can be told. What matters is how rich your "action language" is, how easy it is to "read," and how well you teach the player it.

Body langage isn't hard to do at all, it's mostly parametrics. The real problem is parsing and generating "sentence" with meaning. But even that you could still use "template" and match them to situation "pattern" by filling the splots, basically making a chatterbot but with body language. As there is no such a things in game, it would be a major progress if you pull it even in a crude way, kinda like eliza.

I've always fixated on how a game's player-avatar is the one thing that is onscreen at all times.  Why, then, is the character so much simpler than the world around it?


Gimmy:

I plan to supplement the body language with some symbolic stuff, though I'd like to minimize its use where possible.  I acknowledge that motions alone aren't going to be enough to convey everything I want to convey.


Graham:

It's a less-than-ideal arrangement in that my story is decided in the broad strokes and I'm in the process of evaluating whether a system of my creation will be able to convey it adequately.  Text is the conservative decision -- the fallback, perhaps -- in this regard.  My goals with the "vocabulary" of this language are creating things that are intuitive to read, but ideally can be very subtle.  Culture-independent mannerisms:  Shifting weight from one leg to the other when uneasy.  Avoiding eye contact with another character when a lie is told.  Exchanging glances wordlessly.

As a programmer I want to approach these technologically.  As a storyteller I know that human actors (made interactive) could convey the story elements I've designed.  So my goal is to simulate their role, or at worst make an awkward effort in that direction.  In that sense I have some understanding of how my "level design" and "mechanics" fit together.  At least insofar as I understand the limitations of pantomime and visual symbols, which I will concede is presently inadequate.  It's understanding and working within those limitations which is the biggest challenge for me.

Assuming (wildly) that the technical aspects of those systems are in hand, it becomes a problem of game design to find a means of implementing something like the desired story in the format I describe.

So I suppose the best direction to go from here would be to switch to design-mode and think about what kinds of mannerisms would play into a specific interchange between characters, or the behavior of the player character when left alone with his thoughts.  These are the fine strokes of my design and storytelling, into which I've yet to delve.  The moments which will comprise the broader story I imagine.

Pardon if I seemed dismissive before; you've brought up a very good point.

What I meant is that body language is already expressive (especially combine with facing and positioning) but very little has been done how different body gesture work in a sequence, in parallel, how focus sort between meaningful gesture and noise, how turn taking take place in multiple character interaction and how context modify meaning ... There is only 4 game so far that tackles that in their own different ways: Façade, The act, sweaty palm, spy party.

That's all right.


This is the smaller of your two problems. The tech is tied to the design. You can't assume one is complete and build the other, then return to the first to create it.

Building Mario so that running and jumping felt good took a lot of iteration. Creating levels that took advantage of that was a secondary step. You could say that the two were developed in tandem, though really the control of Mario probably usually took precedence.

You won't be able to find a general solution for expressing all of English through action, and especially not one that you can build yourself, that is interesting to interact with, and suits your game. You will be able to build something extremely small, very limited; and that limited system will reflect your strengths and weaknesses in creating it, as the control of Mario - and every other interesting character - reflected those of his creators.

I recommend being very clear with yourself about how you might express even a handful of ideas through action, then sketching out how much of your story you can tell with that, then adding more to your "action language," then seeing how much you can express with that, and so on. See the parallel to the prototype cycle? Develop some mechanics. Nail them. Then see what kind of game you can make with that. Then repeat.

Aha.  But this isn't my goal at all.  I want to express relationships and emotions.  These characters won't be speaking with some kind of sign language, just letting on the broad strokes of how they feel.  I'll deal with more specific ideas in other ways, though I'd like to minimize their use.



I'm inclined to avoid treating mannerisms too much like islands of meaning.  They're more interesting when taken together.  Again, I'm going for a sort of subtlety that's not well suited to the mentality of "clearly conveying specific ideas" -- this isn't my goal most of the time.  Anyway, that said I could certainly stand to start putting a list of mannerisms together as a designer and getting them operative once my tech is up to speed.

Yeah I understand. The point is still valid though.

Want to know what kind of levels you can create with a "jumping" mechanic? Make a jumping mechanic, then try out a level.

I'm working on a very subtle context-sensitive system too. That system demands prototyping even more than a non-subtle one. So your argument only furthers my point.

ps. Don't wait for the tech. Prototype on paper right away.

Yeah.

My concern was that I wouldn't be able to know how it'll work/feel until the tech was in place.  But if I do more conceptual prototyping work that'll help a lot.

Well I can understand that too. I like code a lot. Programming is my background.

Just sketch like 5 things your chars can do. Implement 3. Then create like a handful of "story pieces" that arrange those 3 in interesting ways.

Then return to the general question.

Somewhat relevant:
http://computationalcreativity.net/iccc2012/wp-content/uploads/2012/05/103-PerezyPerez.pdf

Graham:  Following your advice I've implemented the "simplest" facet of the body language I want to create -- the eyes -- though they ended up being a bit deeper than expected.  (Which is good!)  I referred to some helpful materials on anatomy and animation, which have kept me aware that what I'm doing is for all its cleverness still a very crude illusion.  But it's already enough to express quite a lot, with the help of some context.

Check it out:

https://vimeo.com/56082473 (https://vimeo.com/56082473)

EDIT: scroll down for a GIF.


I might start a devlog at this point.

Try reading some of Robert E. Howard's books/short stories (especially Conan). I love how the fast-paced, actiony writing of it. Unlike most writers, he skips trivial details and focuses on the important stuff, which gives his writing a very fast paced look.

I think the problem with a lot of writing is that it tries to cover everything. It shouldn't; nobody cares about a lot of descriptions. It's often info dumping, though you can include detailed descs to simulate a slower moving 'camera'. It's perfectly possible to describe motions in writing.

This would be very useful advice if I were using text as my predominant means of conveyance.

In other news I found out how to make those silly GIFs.

(http://evanbalster.com/crap/creepy-eyes.gif)

That's pretty good.

This is what I would do:
  . collect some resources: animation, movies, photographs
  . write a short dialog between 2 chars

  . for each sentence, phrase, or piece of a phrase, write down a
  description of the "emotion" the eyes show at that point - ideally
  the most important emotion - just choose 1 simple one to start
  . for each "emotion" collect the resources that most closely
  approximate what you have in mind
  . script your eyes for each emotion
  . play out the scene. dialog appears on the screen and the eyes
  behave in time with it.

Next step would to generate some bodily movements, ideally that suit
the mechanics you're thinking of, paste eyes on them, then have them
react to the player.

This might be interesting as an inspiration:
http://www.gdcvault.com/play/1012460

I Would advise you to go 3D because it will be easier to the long terms as movement would be clumsily constrain if 2D:
http://cmpmedia.vo.llnwd.net/o1/vault/gdc09/slides/Fortier_Universal%20Saints%20Row%202.ppt


This one is only tangentially related, I just put it there:
http://www.gdcvault.com/play/1014362/Cinematic-Character-Lighting-in-STAR

Graham:  Mos def.  Since I'm interested in getting these to a point where they're semi-autonomously expressive I plan on messing with algorithms for blink rate, dilation, focal point, and squint level being controlled by more abstract variables or state representing emotion.


Gimmy:

The characters are going to be in a sort of 3D perspective rendered to 2D.  It's a weird imitation-animation system I'm trying to homebrew and not necessarily a good idea but I'm trying it.  The thinking was things like eyes would be slapped on like decals in this system, which I've learned isn't ideal...

Thanks for the references.

One tip. Nail a few big emotions you know play well before tackling the general formula. Easy to misstep there.

"Decals." I don't know what you mean but it doesn't sound ideal.

Do what you want, It's your project, experiment :)

I have loosely plan something like you but I need to take care about some step before reaching that point, I'm not as good with programming.

I just want to point that despite those reference use realistic character rendition, they might work with simpler (even 2D) representation in spirit, especially "wind wakerish".

Eyes in wind waker are a sort of decals too, not sure.

Edit:
also
(http://www.dhimantvyas.com/blog/wp-content/uploads//eye-direction-meaning_fin0.jpg)
(http://img.photobucket.com/albums/v388/kyleGarrett/headEyeDirection.jpg)

http://skeptics.stackexchange.com/questions/9792/do-eye-movements-reveal-how-a-person-is-thinking

It's crap but it useful for communicating intent in a fictional setting, aka it is a representation ;)
It's the same for facial feature, it's crap, but work well in character representation.

At the very least conveying introspection through non-focused eye movement is a very good idea.  My characters (or at least the protagonist) will frequently be stopping to think.

Fernando Ramallo reminded me of "body noise" today, as a response to my eye video.  Small, pseudorandom movements as the body adjusts and rebalances itself.  Like the motions of one's hand held at arm's length, or the twitch of an eyelid after a quick movement.  That's something worth considering as well, though I'm contemplating perhaps a more animated look.

The thing about the NLP one is that it's made-up unscientific bullshit, and no one can read those subtleties intuitively (because they're not true). The second one, on the other hand, does correspond with the way we communicate non-verbally, so it could be used anywhere :)

In a game about NLP (nevermind its validity ;) ), you could teach the eye cues and then it would be consistent in that game's universe, while you could use the poses in the second image in any game because even a child would get the meanings.

Anything in game is made up bullshit, look at sword fighting, and combo system! I m not sure if you are saying it is bad, but any layer of consistent and intuitive information is good for a design. At least, while the nuance of direction might be lost to the beginner, the overall meaning (pausing to think) get through. Better is that there is a logical consistent explanation for the nuance that can help the beginner pick them up.

I was referring to the specific meanings of each direction. It's not intuitive if the game has to explain it, and it would have to, since it can't be intuitive if it's not true.

Looking to a side when thinking is intuitive, a lot of gestures carry meaning, that's intuitive. Top-right meaning "I'm making images up in my head" is not.

You can still go the naruto way, up is "hopeful", down is "fearful", left is memory, right is thinking, LA noir would have it and would be a better game

My point is, if you want to use that in a game, you have to tell the player what's what first. On the other hand, if you use common body language, the player will just intuitively know. Nothing wrong with the former, it's just different.

Since the body language in my game is less a mechanic and more an aesthetic (though arguably is a bit of both) I'd rather avoid things that require explanation.

A sort of tenet I'm following for the design of this game is "speak little, tell much".

Why not make it both mechanic/aesthetic? Or at least link them.

For example... in a lot of action games monster tells are critical. What if battle tells thematically linked to behaviour in story sequences, etc?

Nice tenet.

to me this thread reads like:

hi i'm a visual artist so i'm going to make a game without graphics, how do i do that?

where did you get that from?

there's a religious myth among game developer and game design theorists that text in games is somehow an immoral thing to do, when players of videogames are under no such impression; games with text are loved and enjoyed by everyone except for a small subsection of weird mentats

example: from the first post: "who wants to read in a game?" the answer of course is: pretty much everyone loves to read in videogames (or else games would not use text at all!). rpgs and visual novels sell wildly. the trend in modern games has been towards more text, not less (compare even fps games of 10 years ago to today, there's a lot more text in them). text sells, it's what people want, and it's fun to read, text adds to the fun and gameplay of a game. what people *don't* like are walls of text, which aren't a good thing even in novels. but the solution to boring walls of text isn't no text, it's less and more precise and carefully chosen/edited text

yeah I see what you're saying.

I love text too. I think he was just being hyperbolic.

My game is going for no text too... but it's not b/c I hate text. I like procedural stuff because it lets me tell a story more dynamically. Procedural behaviour is easier than procedural writing. You can also link behaviour to mechanics more easily. So I empathize.

Also I understand systems much better than I do the written word. I think it's a good idea to make a game that suits your strengths. I imagine Balster is in a similar boat.

Also, language barriers disappear.

I was in love with the idea of no language barriers, maybe more than anything.

I did a little experiment last night, though, and I'm realizing total avoidance of language probably isn't forward thinking with what I want to do.  I'd need to use symbolic representations a lot.  It would be extremely difficult to animate it all, and a bit forced.  A slightly more realistic goal would be extremely minimal use of language -- simple vocabulary applied with extreme terseness.  This would permit the conveyance of simple concepts which would act as a sort of "scaffold" on which to build more complex, unspoken ideas.

The time I've spent thinking about the game having no language has been insightful though -- Robin Arnott uses the term "temporary belief system" to describe this sort of thing.  I've found myself paying much greater attention to body language and facial expressions in film and animation and thinking about systems of mannerisms I could construct and convey.

The little script I've written details the mental/emotional state and consequent body language of two characters as they exchange five lines of dialogue.  There's a lot going on behind the words said -- white lies, feelings, motives -- which motivates eye movements, postures, et cetera.  A lot to work with, when I pair a great deal of subtext with a bare minimum of speech.

At this point I'm toying with the idea of having the game's text read like a crude translation of a foreign film, or a children's book.  That's a reasonably "low" language barrier, and more than enough for me to convey complex ideas.  I also feel as though words will have greater weight if few and far between.


As an example:

The protagonist stands at the base of a great tree, wearing a satchel on his back.  Another like it sits against the trunk.  Him arms are crossed, and he is looking upwards into its high branches.  He shifts his weight from one leg to the other.

A leaf flutters into view from above and drifts to the ground beside him.  Giving it no notice, he breathes deeply, and sighs.

Onscreen narration:

"To those who do not know her,

perhaps she does not seem frail."


The unspoken ideas here (which I expect I am entirely redundant in recounting) are that the protagonist is traveling with someone he is protective toward, she has elected to climb into the tree, and he is quietly worried and impatient for her to come back down.

Yeah, true expression is very hard.

Huh, somehow I saw this thread in the writing section and assumed it was about using text to describe body language. Misleading context :P

My motivation, articulated more carefully:

Action, not words, is the native language of visual games.
I wish mine to speak articulately.

We're having an impromptu Iowa game jam with the theme of "work on your crap".  I'm taking the opportunity to do some research and development on kinematics systems.  If I had a giant pile of money I'd license IKinema, but alas...