Here’s animator Jeff Gabor doing some of the best and most extensive video reference for animation I’ve ever seen [Addendum from 3/19 – apparently the folks at Fox asked that this be taken down. Hopefully this screen-grab gives you a little taste.]:
Jeff Gabor at work
When I first started working in animation in the late 1990’s, video reference was looked down upon, like some kind of unworthy crutch. I remember during Quest for Camelot another assistant whispering to me that he’d seen one of the animators enlarging frames of live-action reference at the photocopy machine. The animator looked embarrassed and quickly went back to his office. At DreamWorks every now and then I’d get a glimpse of an animator doing a bit of reference photography or videography, and I noticed it was usually done at a time and place where it wouldn’t get too much attention. Even using self-produced reference as a general tool, without taking individual images and registering and pegging them up, was seen as cheating. Plus it was a big hassle for a variety of reasons, even after video camcorders became commonplace.
Somehow the shift to CG animation has taken the curse off using video reference. Maybe, with everything digital, it’s just much easier to shoot and use your own reference at your desk. Maybe it’s that CG animation tends to require more realistic movement to not look wonky. I think a big part of it is that CG animation is less spontaneous than hand-drawn – there’s no CG equivalent of a scribble test (at least not yet), and few CG animators (even those who started in 2D) do extensive thumbnails. Shooting a few passes of video reference can go a long way towards organizing one’s thoughts about basic timing and posing, while also revealing bits of nuance that CG animation is particularly good at capturing.
That said, many animators still never use video reference. Jeff Gabor clearly enjoys being in front of the camera and performing. Most animators are a lot more shy. One of my old office-mates, Kevin MacLean, is one of the best animators I know, but he’s shy enough that I had to take his place in a DreamWorks improv class. The class would have been torture to him. I know he never shot reference for his scenes, yet he turned in some of the most sensitive and expressive acting shots in Over the Hedge. On the same film, my pal Sean McLaughlin did some beautiful, dead-on video reference of himself, sometimes playing multiple parts, with great results.
Myself, I’m somewhere in the middle. I know I can’t act worth a damn, and I’m moderately camera shy, but I’ve enjoyed having a web-cam to shoot quick bits of reference for particular pieces of acting or performance. It can be a huge time saver, and sometimes it’s the only way to really figure out how something should move. On the other hand, when I really try to act, or when I’m thinking about trying to hit certain acting beats or do certain gestures, it’s almost always stilted, unnatural, and a waste of time.
In a sense, video reference isn’t any different than using reference like Muybridge, or finding a great clip of animal action on a nature show. On the other hand, if you can’t work out of your own head, then you’re going to be limited to what you personally can act out. Then a tool becomes a straight jacket.
What I’m talking about is distinct from rotoscoping. Strictly speaking, rotoscoping refers using the machine invented and patented by Max Fleischer, as in the Out of the Inkwell series and Gulliver’s Travels. When Don Bluth or Ralph Bakshi are said to rotoscope some of their films, they’re really using registered and pegged live-action photostats, which is a mouthful, so we now tend to refer to any technique of transferring filmed live-action poses to animation as ‘rotoscoping.’ And rotoscoping, like motion capture (a different beast, but in the same family tree), really can be a crutch. I think both rotscoping and mo-cap have their place, but there’s no doubt they both have inherent limitations.
Here are the few tips I have for effectively doing video reference. First, if you’re too shy, you should probably not bother. I used to be so inhibited that I could only shoot reference if no one else was around, and even then it was crummy reference. It took animating at a 10-20 seconds/week pace, in a studio with nothing but open cubicles, to get me over most of my inhibitions.
Second, make it super simple. If your production allows a webcam and appropriate software on your computer, great. If not, consider a notebook computer with a webcam or a digital camcorder. If the process is simple enough, you won’t get precious with the process. I found when I had to go to a production coordinator to get the key to get into the dedicated video room, and when I was done filming then had to spend several minutes encoding the video and transfering it to my computer, I tended just skip the whole process. It needs to be as easy as standing up from your chair and acting things out, as we do all the time, whether we’re shooting reference or not.
Third, grab a friend or two if they’re likely to be better actors or better at a certain behavior. Know your limitations. If you really, really can’t act, then you’re risking a garbage-in/garbage-out scenario.
Forth, I find it necessary to trick myself, by planning to do 4 or 5 consecutive takes in rapid succession. I figure that by the third or fourth take, I’ll be relaxed and natural. Ironically, I usually find that my first take is the best, but if I try to do it in one take I always over-think, or tighten up. And if the first passes aren’t any good, delete the file and start over. There’s also nothing to stop you from editing parts of different takes together.
Fifth, use props. Jeff does a nice job of that. I find I need to set up ‘prop heads’ if I want to appear to look at other characters. If I don’t set up a fake head or two, I have to consciously think about where to direct my gaze, which kills my meager acting. Set things up to be as natural as possible.
Finally, I actually prefer the low resolution and low frame rate of a cheap webcam. For me video reference works best when I’m using it to get the basic gesture and timing of the movement. I know I’ll need to exaggerate, and I find it easier to exaggerate from a fuzzy image where everything isn’t perfectly clear (just remember to convert everything to 24 fps!). The exception is closeup facial shots, where I make sure the lighting is decent to get as good an image as possible.
In those close-up scenes, I try to be ‘in the moment’ (which is frankly something you can’t really try to do, but you know what I mean) and let the chips fall where they may. If I focus on saying and feeling the dialog, I tend to get a fairly natural performance in the eyes and brows. If I think about facial expressions, it all becomes contrived. It’s important to know the dialog cold, and to know just where the accents are and what kind of emotional tone the voice actor used. It’s also crucial to actually say the line out loud, and not try to lip-sync it, or your facial performance will be hesitant and restrained.