Goal: get a sense of what it takes to make a system for doing animation, and some key ideas
Prelude: (pre-Computer) Historical Cartoons
- McKay (1911) – make thousands of drawings
- Hurd&Brary (1914-1916) – cells to cut down the work
- Sometime later (can’t find)
- Keyframing and Tweening
- definitely by Disney in 1930s
- Onion Skin, pegs, 5 fingers (ability to hold multiple pages and flip is valued)
Traditional Animation “Pipeline”
- Design / Story
- Master Animators
- Tweeners
- not just interpolation – really filling in what happens
- Catmull ‘74
- Ink and Paint
- Effects
Some things to think about
- Each drawing is independent
- Shape can change any way between drawings
- The “model” is in the artists head
- Or the “Model Sheet”
There are (and still are) systems to help with 2D cartoon production. They have quite a different flavor, since they focus on managing drawings.
Big point of the historical papers:
- the problems are the same – and you need to solve them all
- these paper really identify the problems (and the kinds of solutions)
- these papers weren’t first: but they did show the emergence of the key ideas early
So you want to make a movie with a Computer
Utah, 1970, Dave Evans’ lab…
Ed Catmull’s Hand. Fred Parke’s face (yes, facial animation paper 1972)
Ed Catmull wants to make an animated movie (of 3D objects) – what problems does he need to solve?
- everything is hard – some of the stuff is being invented down the hall, others he has to invent himself
- this set of problems doesn’t change
relate this set to the modern pipeline
- How do I represent objects?
- How do I represent objects in a way that I can make them move?
- How do I describe the motion?
- How do I draw the pictures? (and make them look decent)
- How do I get this into a film so I can watch it?
Ed’s answers at the time?
- Polygon meshes (measured from real objects)
- Hierarchical modeling, sharing points are the edges (single skin)
- Parallel animation language
- Watkins box with Gouraud’s new shading tricks (Phong was figuring out lighting)
- Worries about the movie camera
Foresight:
- “a great deal of research still needs to be done on the creation and manipulation of three dimensional objects”
- wants higher level description “walk charlie to table”
- flexible surfaces are hard to model
- curved surfaces (Catmull-Clark), texture mapping, Z-Buffer === his PhD thesis
So you want to make a movie with a Computer
It’s 1972, National Film Board of Canada, There are these “regular” animation artists around, …
Burtnyk and Wein
Key ideas:
- How do I represent objects?
- want arbitrary curves (can, and will, do anything)
- How do I represent objects in a way that I can make them move?
- embed curves on skeletons (controls)
- idea of “control skeleton”
- How do I describe the motion?
- keyframing, mathematical interpolation
- skeletal deformations
- drawing and filming – not such a big deal
Catmull’s Tween (1975) does similar 2D animation stuff
What are the big ideas here?
- in-betweening (keyframe animation)
- skeletons (?)
- control rigs
- idea that what you do to make things easy to control may be separate from how you draw
MENV – A very specific kind of solution
Same problems:
- How to model
- How to articulate
- How to animate
- How to render (OK, Renderman – or its pre-cursor)
Key ideas (none really novel, its just a famous system that had all of them):
- Avars
- just something that you might want to change
- Model as a box with knobs (avars)
- Animation curves as a first class object
- Procedural AND Interactive
- Language-based modeling
- compact
- avoid precision problems
- PARAMETRIC
- Cue sheets
- curve based control – or computed – or …
What do you see in systems today?
