“Moods as state always fails”. Mood as static state fails. The basic emotions represented by these systems are not basic or primitive in a human. Angry and Happy are complex states, or really, the system exhibiting simple signs resulting from complex processes. Typically, these are not worth modeling for most games, can be worth modeling for virtual worlds and are a subject of intense and well-financed studies and projects in circles where diplomacy and guile play a significant role.

The key is developing the input system. Facial recognition is good, but has the problems listed above. (do you really want to emote “embarrassed”?) Keyboard shortcuts seem to be the easiest, but the system has to work well and be intuitive. Sliders might work, but I think they would have to be “generalized” in some way.

People take for granted the fact that their characters in SWG “laugh” or perform other animations based off of their chat. It comes across as so natural that it makes you want to chat. People don’t sit there and think “I’m going to chat now because I like how the animations play.”

So if you asked people “what should we spend development time on” it’s fairly likely they won’t mention chat/emotes/moods, even though those things play a very important role in their experience. I’m sure those kinds of systems are even harder to sell to the “higher-ups” at companies like Blizzard and SOE. That is too bad.

Anyway, I think all this applies to text-MUDs as well, but I believe text players are inclined to emote more because that is the only way to stand out and give your character “pathos”. So, if you want the same effect in a graphical MUD/MMO/whatever you have to give everybody the same dull avatar forcing them to emote their identity… Good luck!

You didn’t play SWG (Star Wars Galaxies), did you? When I first started playing it, I really found the system magical. You typed “No” and your avatar would wag his/her finger at the same time. You typed “lol” and the avatar would laugh. “Thank you” and the avatar would join an animation to the typed sentiment. It really really enhanced interaction, IMO.

And this was great as the early SWG had this most social and promising of classes: The entertainer. But even my more gear oriented doctor was more alive when dealing with “patients” simply because the avatar reacted to what was said. And all that without having to trigger a canned emote (although those were there too, of course).

I really miss avatars being naturally expressive and would like to see that aspect developed more in games. I understand that the cost / return on investment ration might not be that enticing though. Pity. anything that enhances interaction and avatar feedback without needing an Animation Override hud (see Second Life), a much less natural process, improves the immersion into the virtual world.

Wendelius

You didn’t play SWG (Star Wars Galaxies), did you? When I first started playing it, I really found the system magical. You typed “no2 and your avatar would wag his/her at the same time. you typed “lol” and the avatar would laugh. Thank you and the avatar would join an animation to the typed sentiment. It really really enhanced interaction, IMO.

And this was great as the early SWG had this most social and promising of classes: The entertainer. But even my more gear oriented doctor was more alive when dealing with “patients” simply because the avatar reacted to what was said. And all that without having to trigger a canned emote (although those were there too, of course).

I really miss avatars being naturally expressive and would like to see that aspect developed more in games. I understand that the cost / return on investment ration might not be that enticing though. Pity. anything that enhances interaction and avatar feedback without needing an Animation Override hud (see Second Life) improves the immersion into the virtual world.

Wendelius

If you want to have players willingly participate in this kind of aspect of game play, then you first have to make the game conducive to social play.
Take away the desire to level and replace it with the desire to do other things, or at least as much so. Otherwise, you end up with the RPers as a separate entity, outcasts in a way. For this emote body language, you need to make it part of the game play through NPC interactions and reactions. NPCs need to recognize and react to players moods and actions, and slot players accordingly for present and future interactions. And it needs to be advantageous to the player.

I am sure big developers have thought a lot about facial expressions and then ruled them out as frivolous — throw in another suit of armour, as Raph says. But they’re not frivolous, they just need a few other complementary technologies to make them work.

1. Intensity as a routable value into and out of the emote engine to a library of animations (per avatar so a standard for avatars such as h-anim is critical) is a means.

2. Proximity and individual histories are also key (say persistence of a dynamic data set).

3. If there isn’t a base set of emotions that can compound, the results are very primitive and subtlety is vital to realism.

The problem going forward was having the spooks show up too early and too heavy handedly, and the fact that as soon as they could, a small group started filing patents. My disgust for the end of what was a promising effort could not have been more pronounced.

Advanced stagecraft is where it leads so one might as well get on with the technical reality of building an assisted puppeterring system.

Facial animations are limited to a small number of presets. I’m in the process of building lollipops for the Hair Fair event, and using the “tongue out” face for the licking animation. Unfortunately, the face is a “nyah” face, so along with the tongue sticking out, the eyes narrow and the brows draw together. I’ve started referring to the result as “anger pops”.

One of the things I would hope to see out of initiatives like Project Natal would be cheap, effortless motion capture. Using a tool like that, we could build a vocabulary of body language, and then trade/sell/swap the lexicons. Thus the cam might capture my exhausted, after-work schlump, but then translate that into a bold, heroic stride or sinuous ninjutsu stalk.

But will we ever have a peripheral ubiquitous enough in the PC world to be worth developing for?