Computer Graphics World - 1984

Interview

LUCASFILMS ENTERS THE FRAY

Ed Catmull is Vice President and managing director of the Computer Division at Lucasfilm. He has a Ph. D. in Computer Science from the University of Utah, and was director of the computer graphics lab at the New York Institute of Technology.

CGW: What led to your involvement in computer graphics?

CATMULL: I grew up wanting to be an animator— until I realized I couldn't draw very well. Later, when I was attending the University of Utah in the early 70's, I was influenced by the long succession of people there who had advanced the state of the art of image making.

I began the graphics lab at N.Y.I.T., working on animation in a converted garage. Later, I was offered a position with Lucasfilm to bring high technology to the film industry in the areas of video editing, digital audio and computer graphics.

CGW: What do you feel is the significance of your work at Lucasfilm?

CATMULL: I believe we have the strongest group anywhere in the area of image synthesis. At SIGGRAPH'84, we demonstrated the first film with articulated motion blur. This was a very significant development because, although image quality in computer graphics has improved dramatically over the years, animation has not. One of the things we tried to do was to make animation more pleasing to the eye, and motion blur is a very important element in that process; without it, special effects don't look natural. Star Wars started a revolution in special effects with motion control cameras that gave inanimate models motion blur.
This same concept extends to computer graphics; however, to get motion blur in graphics, you have to have the proper computations. This year, for the first time, Lucasfilm presented papers on the subject and also used the concept in the short film we showed at SIGGRAPH.

In short, we're trying to move away from the "pretty picture" business and towards serious applications for those pictures.

CGW: What is unique about the work your team is doing?

CATMULL: We're fortunate in that we've got the financial support to fulfill our research and development goals. To continue to be successful, we have to keep on pushing ahead.

A lot of companies don't publish papers, for fear of giving ideas away. We take a different view, that publishing will help establish the fact that we are on the leading edge of what's going on.

CGW: What's the most significant issue Lucasfilm must address in order to successfully implement its long-term goals?

CATMULL: One of the main issues in computer graphics is the cost of making pictures — it's incredibly expensive to produce a high-quality image, especially with motion blur. If we're to succeed as a business, we'll have to get the costs down.

To that end, we have a hardware group that's building a marketable series of devices specialized for computer imagery. It took a while to pull in the right people, but now we've got an exceptionally strong team.

The first device is called the Pixar 2D, which has a frame buffer with a 4000 x 4000-resolution address space and a high-performance processor. It has 48 bits per pixel, with the possibility of going up to 64 bits per pixel. It's a very fast, general-purpose machine that can be used for 2-D image processing, mapping, medical imaging, and prepress work. We've also developed the Editdroid film editor and the ASP Audio Sound Processor.

We plan to market these systems — that's how we're justifying R&D costs. From the very beginning, our intent has been to produce systems that can address a general need in computer graphics.

These devices will definitely change things in the film industry, though costs will prevent that from happening overnight. One problem that needs to be addressed is that a lot of digital special effects still look like computer graphics. However, it's just a question of time before computer imagery becomes cost-effective and commonplace.

CGW: Have you encountered any problems within the computer industry because of your base in filmmaking?

CATMULL: For a long time, whenever we visited a manufacturer and made suggestions about increasing the performance of a piece of equipment because we had a need for it, the manufacturer would insinuate that we were "out on the fringe" of computer graphics, doing exotic work that didn't represent the "real world:' Later, the things we originally wanted became industry standards. After this happened several times, we realized that the kind of leading-edge performance we were seeking would eventually be required by everyone, so we decided to fill the need ourselves.

CGW: What would you like to change about Lucasfilm's image within the computer graphics industry?

CATMULL: One of our fears when we exhibited Pixar at SIGGRAPH '84 was that people would perceive us only as a film company, not examining how our system could be used in general business areas. We were there to sort of test the water, inviting a select group to view a demonstration of the Pixar. Fortunately, the response turned out to be extremely positive. People are realizing that, even though our particular application may not apply to them, the technology could be enormously useful.

As a result of that response, we will be out in force next year at computer graphics conferences and exhibitions. We intend to be a major contender in high-end graphics. The Pixar 2D is only the first in a family of high-performance graphics systems; we plan to extend our efforts into 3-D processing. We will be conducting additional software development in such areas as image generation, modeling, and control of images.