If you listened to the Pixar Post's
most recent podcast, you heard me talking to T.J. and Julie about Pixar's formation and their first year as a standalone company. I love Pixar's early history. I mean, before they made
Toy Story, Pixar was a computer manufacturer. And before that, they were part of Lucasfilm revolutionizing digital film-making (actually, creating it). And before that, folks like Ed Catmull, Alvy Ray Smith and David DiFrancesco were at the forefront of computer graphics and animation at the New York Institute of Technology (NYIT). When you look back at that time frame (from the mid 70s to mid 80s), before they could act on their dream of creating a feature-length computer animated film, they had to create the tools. Many of the techniques and algorithms found in today's animation tools (motion blur, texture mapping, rendering) were discovered and created by these pioneers.
I thought it would be fun to highlight some of these people and the techniques they pioneered, plus provide links to further information for those of you interested in learning more.
Ed Catmull
Most people know that Ed Catmull is one of the founders of Pixar (along with Alvy Ray Smith). But more than a decade before spinning Pixar out of Lucasfilm, Catmull was studying computer graphics at the University of Utah. As a graduate student in 1972, one of Catmull's projects (along with classmate Fred Parke) was to animate his left hand. He made a plaster cast of his hand, drew a few hundred polygons on the cast, then measured, digitized and entered the coordinates of each polygon into his computer. He then developed a 3D animation program to animate his reproduced hand. Finally, he used a 35mm movie camera to capture the animation from the screen and record it to film. About a minute long, this is one of the first ever computer rendered movies and an amazing piece of work. Not only does the hand rotate and curl up, the camera goes inside the hand. Portions of the film (along with a film that Parke made of animating his wife's face) ended up in the 1976 film
Futureworld. To see the video, along with some "making of" and background footage, check out
Robby Ingebretsen's blog. Ingebretsen's dad helped out with the making of the video. I strongly recommend reading his post regarding this video - not only does he provide interesting details on it, he tells a nice story of how he was invited to Pixar and met Catmull. And make sure to read through the comments, which include entries from Fred Parke and children of other graphic experts of the time like Dave Evans and Ivan Sutherland (creators of the first frame buffer).
As part of Catmull's
PhD thesis, he developed a solution for tracking distances between objects in a 3D space and determining which portions of an object are hidden from the viewer. His solution is called the Z-buffer, which stores the distance between the viewer and objects at every point in a scene. At the same time, Catmull discovered a way to project an image onto the surface of an object. This allowed objects in a computer to take on the look of concrete, wood or any object and is known as texture mapping.
The Pixar Touch by
David A. Price provides a thorough history of not only Catmull's work at the University of Utah but of the key people and developments of Pixar's history, from the earliest days at the New York Institute of Technology and Lucasfilm, through the forming of
Pixar as a standalone corporation and its sale to the Walt Disney Company in 2006.
Alvy Ray Smith
Smith was a pioneer in the area of digital paint systems. His first experience with these systems was at Xerox PARC using Dick Shoup's SuperPaint application. Smith developed his own paint application at NYIT, and followed it up with Paint3, which utilized 3 frame buffers simultaneously to provide a palette of over 16 million colors (each frame buffer supported 256 shades of a single color like red, green or blue. 3 frame buffers therefore could support 256^3 different colors).
In 1978, Smith also made the discovery of the alpha channel, or opacity, which allowed images to be superimposed on top of each other.
These advancements won Smith a Scientific and Engineering Academy Award, along with Dick Shoup and Pixarian Tom Porter, for their "pioneering work in the development of digital paint systems used in motion picture production."
Smith did a
fabulous interview (with a detailed
companion piece) with
Mike Seymour on the
fxguide podcast. Seymour and Smith discuss paint systems, the alpha channel and Smith's work at Lucasfilm and Pixar's early days, and is definitely worth a listen. In addition, Smith's website is full of great information on the
history of digital paint systems.
Rob Cook
Early on, Catmull realized that for computer generated animation to look believable, it would need to include
motion blur.
Motion blur is an artifact of using a film camera to record an image
with moving object(s). While it is normal to try to reduce or eliminate
motion blur when taking a photograph (to stop the motion of a wave
crashing on rocks one would use a very fast shutter speed), motion blur
is necessary to provide a realistic movie experience. Without it, a film
would look similar to stop-motion and have a stutter feeling.
|
Tom Porter's 1984 pool ball image |
droidMaker by
Michael Rubin
gives a good, in-depth discussion of how Rob Cook solved the motion
blur problem while at Lucasfilm. The solution was used in Tom Porter's
"1984" pool ball image and in
The Adventures of André and Wally B.
The book includes interesting bits of the struggles Cook had in finding
the solution, and includes the story of the friendly competition
between Cook and Catmull to see who could find the solution first.
I am greatly enjoying
droidMaker. It is not solely about Pixar like
The Pixar Touch,
as it is follows George Lucas' career and his impact on the digital
film revolution. But Pixar has a significant history with Lucas as Pixar was spun out of Lucasfilm, and the book does an
excellent job covering the period of time up to where Pixar was spun
out. It also covers a wide range of technical topics such as the
beginnings of digital audio, computer graphic workstations and Sun
Microsystems (much of my professional career has been done on Sun
workstations), which I really enjoyed.
Loren Carpenter
In 1980, Loren Carpenter, then an engineer at Boeing, wanted to get into Lucasfilm. The problem was that
everybody wanted to get into Lucasfilm at that time. He knew just sending a resume wouldn't be sufficient, he needed to impress Ed Catmull and Alvy Ray Smith. Carpenter had been applying fractal mathematics in his job and realized he could use fractals to create a whole range of objects from mountains to lightning bolts. He decided he would create a short film for the upcoming SIGGRAPH conference demonstrating his new algorithms. It took him over 4 months to create the 2 minute film, and when it was shown at SIGGRAPH it completely blew away the audience. Catmull and Smith, sitting in the front row, immediately offered Carpenter a job. Carpenter's algorithms were used extensively in creating the Genesis effect for
Star Trek II: The Wrath of Khan.
Carpenter has posted his SIGGRAPH video on
vimeo, and even today it is a stunning piece of work. It's hard to imagine what he must have gone through to complete that video on the equipment of the day, all in his spare time!
Loren Carpenter and Rob Cook
Carpenter and Cook also were key contributors to Reyes (which stands for Render Everything You Ever Saw), the rendering
system at Lucasfilm that later became the basis of RenderMan. Carpenter and Cook won an
Academy Award in 2001 (along with Catmull) for their
"significant advances in motion picture rendering".
Cook did a very
interesting interview with
fxguide's
Mike Seymour a few years ago. The interview covers a large span of Cook's involvement with computer graphics and rendering, and includes an in-depth discussion of solving the motion blur problem, in addition to the history and advancements of RenderMan. It also covers Pixar's migration from being technology-focused to art-focused. Seymour and Cook also discuss the implications of John Lasseter's phrase, "Art challenges the technology, and the technology inspires the art."
Bill Reeves
|
Bill Reeves (center) with myself and
my son Sam at the CAM Benefit in 2011 |
In late 1981, Ed Catmull, Alvy Ray Smith and others in the computer graphics group of Lucasfilm were busy working on the Genesis effect for
Star Trek II. Their original concept included exploding volcanoes and mountains growing from the flat landscape of the planet. But then Bill Reeves showed them his revolutionary new
particle system. A particle system can be used to model objects that do not have well-defined borders, such as clouds, smoke and water. Reeves' new system gave a much more natural and realistic look to these objects, and Catmull and Smith agreed it should be incorporated into the effect. The exploding volcanoes were replaced with fire that spread across the planet, and the final effect was stunning with a high amount of realism. It was the first use of this technique in computer graphics, and won Reeves a Scientific and Engineering Academy
Award® in 1997.