Saturday, March 05, 2005

Language Barrier

When I discuss planning a scene with many CG animators I find that there is a kind of communication breakdown happening between those who have a understanding of the traditional background of the craft and those who don't. Folks will use the same words, but often these words won't mean the same thing to all parties in the conversation. Animators who have come from a purely CG background (through no fault of their own) hold different meanings to certain words and phrases than have been in use in animation for 80+ years. I'll admit, I used to be one of these people. Earlier in my animation career when I had conversations with traditional guys I didn't understand what they were on about sometimes. They seemed frustrated with me and my lack of understanding of the basic language of the trade. It's almost like a person from the Bronx was talking to a person from Edinborough. They're supposedly speaking the same language, but I'd be willing to bet good money that they'd have a hard time understanding each other. Such is often the case between animators with only a CG background and those with some measure of experience in traditional animation.

One of the earliest and most significant language casualties of the CG animation era has been the term "key frame". When you're just plugging along as an animator on the box, "setting keys" as often as you blink, the meaning of this term "key frame" gets washed away. It's not the animator's fault. Some programmer some many years back needed a way to describe the function of telling the computer to remember a given value for a given object at a given frame in time- a vital foundation of the entire computer animation activity. Being a programmer (and a very bright fellow to be sure) this person must have looked across the animation lexicon and settled on the term "key frame" to describe this function. A good try, but an unfortunate misuse of the original term. Still, it stuck and here 15-20 years later we have a legion of CG animators who have adopted this new computerized meaning for the term "key frame". So ask any CG animator who hasn't had any experience in traditional animation what the meaning of the term "key frame" is and they'll tell you this: it's when you tell the computer to remember an object's values at a place in time. Ask them how many keyframes are in a scene and they'll likely say "Well, hundreds, perhaps thousands." In the language of CG, all of this is correct.

But what is the original meaning of "key frame"? If you ask an old timer traditional animator what the meaning of a "key frame" is you might hear this: It's a defining moment in a scene that is the foundation of the performance. Ask them how many are in a scene and they'll likely say: Sometimes as few as 1, maybe as many as 4 or 5 for a longer more complex scene.
Now wait... if there's only 1 key frame, then how does the character move? The key to the term is the word "key". See? There are doubtless many frames and poses in any given scene. But only a very few of them are KEY frames. They are the foundation of the performance. Change one key frame (the drawing is called the key pose, where it occurs in your scene- based on your scene break out- is the key frame) and the entire flavor of the scene changes. A KEY frame is just that- it is the key to unlock the entire thing. And so yes, sometimes there is only 1 key frame in a scene. There are plenty of pose drawings and breakdown drawings and inbetween drawings, but very few key drawings.

So why go on about this? Isn't this just some blathering about semantics? I mean, CG is the new paradigm, it rules the feature film landscape. Stop clinging to the past like some old fart, you may say. Well, I think it is important because if we are going to move forward as CG animators we had better come to understand our past animation heritage. I started this post off by saying this language barrier comes up during scene planning discussions. At work as a supervisor I'll often get together and help an animator break down their scene, check each the thumbnials and pose ideas, etc. One of the first questions I ask when looking at the thumbnails or pose test is "Which one of these are key frames?" I used to get the head tilted curious dog look. "Keyframes? There's gonna be tons of 'em by the time I'm done." No, no. I mean, which of these poses, which of these thumbs- which idea am I looking at here is the one or two or maybe 3 that define the entire peformance? Which pose- if changed or taken out- causes the entire scene to change down at the core level? If you as an animator cannot answer that question when you're planning your scene, you're lost before you even leave your house. You have no map to guide you. If nothing is key, then everything is just bleh. Down the road you may end up noodling the snot out of your ideas and the end result will be a performance that is a complete mystery to understand. It's locked up in a mystery because you never found the keys to unlock it in your planning stage.



David Andrade said...

Semantics you say? Definitely! I, being brought up as a student during the CG age, am essentially fed those words. In fact it wasn't until (I knew somewhat of key frames original meaning) I read Williams' book that I truly understood what key frames were! I wonder if there have been similar terms, but none come to mind quickly. Perhaps new animators have it all backwards ^_^ O well, I guess that’s what mirrors are for... heh. Brilliant post Keith!

Anonymous said...


Here is what I came up with for my students. I should maybe add in "poses" and distinguish them from key poses":

A brief digression on terminology

I have made some decisions regarding the 3D animation lexicon. Hopefully this will eliminate any confusion and mesh well with existing 3D and 2D terminology. The terms we will be using are as follows:

Keys or Keyframes: This is what we will use to refer to the actual Keyframes saved in Maya. These are the points on your animation curve, the ticks on the time slider and the boxes in the dopesheet. A bunch of these on the same frame will make up a key-pose.

Key-poses: These are the “extremes”: the poses that define the action. Without them, our animation will not be readable. Preferably in the beginning stages of your animation, your key-poses will have keys saved on all the controllers on the same frames.

Breakdowns: These are the important poses that get you from one key-pose to another and tell the computer how you want that to happen. Remember the computer is dumb and lazy and will take the shortest possible route between two points. Again, to keep things organized, try and place these for various controllers on the same frames at the start of animating your shot. (Not to be confused with the “breakdown” in Maya which is a form of key that maintains its proportional time relationships with adjacent keys. I NEVER use these.)

Inbetweens: These are the frames you let the computer handle. Usually these are crap. You use the key-poses, breakdowns and graph editor to wrangle the best possible inbetweens out of the computer. If it’s not doing the right thing, get in there and fix it. Add a breakdown, tweak your animation curve or fix your key-pose. (Not to be confused with the “inbetween” in Maya which allows you to add or remove frames without affecting the keys or breakdowns.)