By Curtis Hoffmann
Converted to HTML by Rob Kelk
From: fjohnson@police.rutgers.edu (Floyd Johnson)
conty@cbnewsl.cb.att.com (E. Kontei) writes:Because it was bad rotoscoping. The Fleischer brothers invented the technique waay back in the 40's, and it can be done right (like the way Disney does it) or wrong (like the way Bashki and Filmation used to do it). Rotoscoping per se is not a bad tool, what matters is using it right! Several anime have used rotoscoping to good effect, mostly for modeling mechanical devices.Given the definition of rotoscopy, "tracing over" live action with pencil sketches of a character, I thought the only was you could screw it up was be either using too few frames or by imprecision in the drawings.
Well...
It's real easy to screw up rotoscoping...
(Ignoring the possibility of the animator not tracing a person's outline exactly in every single frame.)
First, think about shading -- that is changing every frame, while most directors prefer to have the animators draw one picture for every two frames of film. And the lighting sources used inside the animated scene usually don't match up with that from the original live film, which makes matters much more difficult for the animators (if they even think about this problem at all.) Because the shade lines aren't solid and consistent in real life, when the animator tries to nail them down in pencil and ink, the shaded areas of a person's face or clothes are going to jump around a lot.
Second, the entire issue of perspective comes in. In real life, if a person extends their hands away, or towards the camera, you'll see an apparent difference in size relative to the rest of the body. Most animators ignore this situation under normal conditions, but can't ignore this when rotoscoping. Unfortunately, the standard way to show that one part of the body is closer to the camera than the others is to paint it with gradiated shading. This requires more time and money (and a larger variety of paints that have to be mixed exactly the same when replenished.) So, most directors will ignore the details of perspective, causing the body parts to look funny when colored in.
And, when two people are on the screen at one time, and one walks behind the other, if both characters have been filled in with the same colored paints, they look like they're in line with each other and should be bumping noses.
Third, the fleshy parts of the face are very difficult to pin down with a pencil, so you'll get a jumpy effect in things like jaw lines, curves in the base of the nose, and cheek bones, when a character turns his head around. Further, if a director chooses to transfer the pencil drawings to cel via xerox (rather than hand-inking the lines with colored paint,) you'll get a hard, solid line in places where you'd normally have a simple, soft change in color. An excellent example of this is in Omoide Poroporo -- the main character's cheek bones look very puffy and badly drawn. They also jump around a little when she turns her face towards the camera.
Really, the things that suffer the most when rotoscoping is used: Clothing movement, shading in the face, solidity of the jaw line, hands and shading of the wrists, and mouth movements.
To look right, rotoscoping should not be used to replace one character with another (ie -- Tony the Tiger in the cereal commercials) but rather as a guide for complicated movements. You watch how some action is made, and then draw the character independent of the actual appearance of the original actor/model. The old Bugs Bunny cartoons, with influences from Tex Avery are great examples of this. And the stuff done by Blair Preston.
Third, the fleshy parts of the face are very difficult to pin down with a pencil, so you'll get a jumpy effect in things like jaw lines, curves in the base of the nose, and cheek bones, when a character turns his head around. Further, if a director chooses to transfer the pencil drawings to cel via xerox (rather than hand-inking the lines with colored paint,) you'll get a hard, solid line in places where you'd normally have a simple, soft change in color. An excellent example of this is in Omoide Poroporo -- the main character's cheek bones look very puffy and badly drawn. They also jump around a little when she turns her face towards the camera.
To look right, rotoscoping should not be used to replace one character with another (ie -- Tony the Tiger in the cereal commercials) but rather as a guide for complicated movements.
Well, live action + animation combos have been around for ages. It can be done right (see Who Framed Roger Rabbit for a good example).
I agree. The thing about the Tony the Tiger commercials is that it is pure rotoscoping. A live actor does the work, and then the animators paint the tiger character over him -- which makes Tony look bulky and awkward when compared to the kid actor working with him in the ad. This is not the same as the early Gene Kelly movie where Kelly's dancing by himself on the set, and then Tom and Jerry are painted in later. Of course, the best example of live action + animation is the Out of the Inkwell series by the Fleischer Brothers.
While most people would think of Who Framed Roger Rabbit as a rotoscoping-heavy film, the places where a live actor is replaced by paint are rather rare, and most people have missed those scenes (for the record, the majority of the "actors" were mechanical props that were later painted over -- like the penguin waiters in the Ink and Paint Club. A live actor was used, though, during Jessica's song routine when she grabs Eddie's tie.) Technically, most of the animation in WFRR could be considered to be "half-rotoscoping," I guess.
You watch how some action is made, and then draw the character independent of the actual appearance of the original actor/model.
This is what I call half-rotoscoping. You don't draw the model, only your impression of it. Gives a much "truer" (if you will) look to the animation than rotoscoping.
Ok. But if you watch Omoide Poroporo, you'll see the artifacts of rotoscoping that I described above, which would have been smoothed out or rendered invisible if the animators had used "half-rotoscoping." Since I don't have a line on any inside information that no one else does, I can't state "without a doubt, Omoide Poroporo used pure rotoscoping," though. (Well, actually, yes I can...)
-- Curtis H. Hoffmann
Nov. 16, 1992