Animating for Videogames

This week it's been a very busy one at the studio, I will be honest by saying Videogames are one of the HARDEST things to do. It's very difficult to get the whole thing right. It's like a Tango dance between art and programming at the rhythm encompassed by the design elements of the game. One could not exist without the other.

What makes it really difficult on the animation side is the use of constraints. Of course you will always have certain constrains in a production environment. While in movies these are the limited amount of frames you will use for a shot, or certain camera angles, in games are way more complex. That might also explain why animation doesn't look THAT good as in movies.

In feature film animation you have a much more controlled environment since the director knows already what's about to happen, where the characters will go next, where they will stand for a while and where they will turn to at an specific time. Not so like in games, where the action is dictated by the decisions YOU, the player, will take.

In games you need to have an animation ready for almost every basic action the player might take. Like walking forward, backward and to the sides. Usually there's some algorithm for blending between 2 animations, but even with it, you need to consider which would be the best part of the animation for a change in direction to happen. Usually where the character's pose is more compressed so it doesn´t "jump" that much while blending. Otherwise it will get choppy, affecting the immersion of the player (also called suspension of disbelief). And because the computer doesn't understand what a walk action is, your character will never be alone when you animate it. You might also need to animate some additional "controllers" to help the computer understand that there's a displacement to be considered while your character walks or to tell it in which direction the character is walking to or looking at. Others might include the position where a prop (a gun, a sword) is being held (the hand?, the hip?).


But the player won't see them while playing. He will just push forward and Voilà! the magic will happen.

There's also the issue about "responsiveness", which is utterly important for the game play. When you press a button on the controller that supposedly makes your character jump, you don't want to watch it taking its time to do so. Chances are you might want it to just jump right away, otherwise you will be losing that moving platform where your character is supposed to get. So you might be limited to just a few frames to animate this action. And with few frames I mean something between 3 and 6!!!, just miliseconds (and hard to get it right).


Another constrain is the amount of memory you have available, resulting in actions that will rely heavily on repetition. This is where the idles will come in handy. An idle is a cycling loop where your character "repeats" an action over and over, so it never stops and look like a statue in the game. Loops are usually around 10 seconds, and they get divided in segments both starting and ending in the same pose so you can shuffle them and make 2 characters look different.


To keep the reads short and interesting, I will talk about other aspects of animation in games on further posts; like behavior systems, navigation, scripted scenes, cineserts and cinematics.

Peace.