Rotoscoping and Paint
Review constantly at full speed. Don’t waste time heading down the wrong road. Carefully assess your
footage via a RAM Preview every few frames, and determine if your approach is working. If it’s not, be
prepared to switch tactics.
• Combine strategies. Effective rotoscoping may employ a mixture of keying (whether color keying or a
hi-con matte), tracking, animated masks, and paint. Start with the one that protects the most crucial edges,
and work your way down to the cruder stuff.
“Keyframing” is so-called because originally (at Disney, in the 1930s) it was the practice of drawing
in the key frames—the top of a leap in the air, the moment of impact, the kiss—that would be done by
the top animators, leaving lower-level artists to add the in-between frames thereafter. This is the right
way to think of keyframing animated masks: Look for the pivotal moments and try to let the computer
in-between the others as much as possible.
With these guidelines in mind and a solid grasp of the fundamentals of creating selections (explored in Chapter
3, “Selections: The Key to Compositing”), you’re ready to consider some rotoscoping specifics.
A hi-con, or high-contrast matte, is created by taking luminance data from an image—typically one
or all of its color channels—and raising the contrast to create a luminance matte, typically to hold out
areas of the same image. Chapter 6 explores this process in greater depth.
I’ve said it before, and I’m saying it again: When animating, or articulating, mattes, keep it simple. All of the
strategies offered here are for setting up and animating your mattes with as few steps as possible. Not only
will this save you time in setup, you’ll have fewer elements to fix if something goes wrong. Each new edit you
keyframe introduces the possibility for more error, and each new point you add to a mask has to be accounted
for on every other frame.
The first question to decide when starting a roto job is whether you prefer to mask in the Composition panel or
in the Layer panel. The advantages of the Composition panel are that you can see the layer in the composition
and that it updates live as you work. This approach, however, does have several potential pitfalls:
• On a large scene you may have to wait for all elements to render and the screen to redraw each time you
draw a point.
• If the layer you are masking has been transformed (and particularly, rotated in 3D, scaled down, or made
transparent), you may find it difficult to edit a mask cleanly.
You may need to select and lock other layers to be certain that you’re editing the correct layer, particularly
if you’re creating more than one mask.
• The mask you draw in the composition viewer is live by default, causing masked areas to disappear as you
work. This can make it impossible to draw complex shapes.
• You lose the use of selection tools that exist only in the Layer panel.
You may have guessed by now that the Layer panel is generally a preferable place to create masks. Tweaking
mask points in the context of the entire comp is often useful, but I recommend this only as a secondary approach.