INTERVIEWS

Marc Levoy
Marc Levoy is an Associate Professor of Computer Science and Electrical Engineering at Stanford University. Levoy is a pioneer of computer-assisted cartoon animation, creating tools instrumental to the shows such as The Flintstones. His later interests in volume rendering and technology and algorithms for digitizing three-dimensional objects won him a 1996 SIGGRAPH Computer Graphics Achievement Award. He is currently researching sensing and display technologies, image-based modeling and rendering, and applications of computer graphics in art history, preservation, restoration, and archaeology.

What first drew you to computer graphics?

In the first week of my freshman year in Cornell University's School of Architecture, I was assigned to draw a cube in perspective and to remove the hidden lines. This seems easy, but the cube had several holes cut into it and several protrusions, so the work was tedious. Since the work was also methodical, it seemed to me that a computer could be programmed to do it - both drawing the perspective and removing the hidden lines. Having learned to program a computer in high school, I thought I would give it a try.

Later that day, the prof wandered by my drafting table and found me scribbling on a Fortran coding sheet instead of completing the assignment. He sternly told me to put it away, take out my T-square and triangle, and learn how to draw like a real architect. He also told me to go visit a Professor Donald Greenberg, who knew something about computers. I was terrified; suppose this Professor Greenberg laughed at my silly idea?

Well, I did as I was told - I resumed the tedious task of drawing the cube, and I made an appointment to see Professor Greenberg. When I showed him my Fortran coding sheet, he burst into laughter - my worst fear! But the next instant he was showing me an article by Ivan Sutherland, which described how a computer *could* make perspective drawings, and could remove hidden lines. I was entranced. The date was 1971. I spent the next 10 years under Don Greenberg's wise mentorship.

Do you have any favorite computer graphics mentors?

If Don Greenberg was my mentor in the early days, Ed Catmull and Alvy Ray Smith were my teachers. Although I never worked at the New York Institute of Technology, I visited frequently during the late '70's, learning more about computer animation from Ed and Alvy on each visit than the NYIT administration would have liked. Their creativity astonished me, their enthusiasm infected me, and their openness inspired me. I can never thank them enough. During my time at the University of North Carolina ten years later, my mentors were Turner Whitted, Henry Fuchs, and Fred Brooks. Fred taught me how to think, Turner taught me to think out of the box, and Henry taught me how to act on my ideas. On a more practical level, they also encouraged and supported my work on volume rendering. Finally, strange as it may sound, I consider Pat Hanarahan - who shares my laboratory at Stanford - a mentor as much as a colleague. But so does everyone who meets him.

What was the first time you contributed to SIGGRAPH?

I contributed here and there to publications by Don Greenberg throughout the 1970's, but my first authored SIGGRAPH paper was in 1977, "A Computer Animation System Based on the Multiplane Technique". It described my (somewhat lame) attempt to simulate using computer graphics the unique multi-layer camera Walt Disney used on their early feature animations (Snow White, Pinocchio, and Fantasia). Most modern computer-assisted cartoon animation systems include this capability, although it's doubtful they owe much to my fumbling first attempt.

What year/city was your first SIGGRAPH? Which was most intense? Why?

I attended the first official Siggraph, in Bowling Green, Ohio in 1975. There were hardly a hundred or so attendees. One evening during the conference, a screening was given of Peter Foldes' pioneering film Hunger, which used key-frame interpolation techniques developed by Marceli Wein. With this film, computer graphics became a storytelling tool. I was captivated. I returned to Cornell and worked for the next 10 years on computer animation, leading to a Bachelor's and Master's thesis on the topic, and to the Hanna-Barbera computer animation system.

What contributions to SIGGRAPH are you most proud of? My acheivement award was for work in volume rendering, but I didn't invent volume rendering. Judging from the online citation indices, I guess my 1996 paper on Light Field Rendering, with Pat Hanrahan, has been the most influential. If you measure pride in terms of blood, sweat, and tears, my 2000 systems paper (with many co-authors) about the Digital Michelangelo Project tops the list.
What's your favorite thing at this year or last year's SIGGRAPH? My favorite thing at *all* SIGGRAPHs is meeting with colleagues - old and new. The SIGGRAPH proceedings is about ideas, but the conference is about people.
What near/intermediate developments in CG do you look forward to?

I am enjoying watching graphics scale down in price and size. Last summer, I saw a 3D model of Michelangelo's David rotating on a cell phone. New killer apps for graphics will certainly come out of this astonishing downsizing. Also, like many researchers I am watching with interest as digital cameras become smaller and cheaper than wristwatches. This trend has already brought vision and graphics researchers together. It will undoubtedly spawn new markets, and these will create new opportunities for research. If the 1970's were the pioneer days of graphics, and the 1980's put graphics on a scientific footing, and the 1990's saw graphics break into the billion dollar entertainment and gaming industries, then the 2000's will be the decade of ubiquitous graphics and imaging. I've lost count of how many computer chips are in my car. Soon I'll lose count of how many cameras and displays are in my car.

Do you still follow computer animation? I only follow computer animation in so much as I have friends in the industry. 2D and 3D animation is really merging, so I follow it in that it's almost the same world these days.
Are you a fan?

Who isn't? It helps if the story's good, which it isn't always. I think the original Toy Story was quite an accomplishment. I like anything that John Lasseter has worked on.

Do you watch computer graphics movies, or play video games?

No, I don't play video games at all. At the end of my computer graphics course, I always have a video game competition. The students will say, "Oh, this is just like 'blah', and I don't know what 'blah' is."

It was kind of embarrassing, a few quarters ago, some students did a recreation of the elevator scene from the Matrix and I didn't recognize it. I hadn't seen it. I was a computer graphics professor who hadn't seen the Matrix yet. I've seen the Matrix now.

You won the Computer Graphics Achievement award in 1996. Do you think this sort of recongition is important? It was probably useful during my tenure year, I won't deny that! I think it's useful for SIGGRAPH to have these awards. It gives people something to strive for. Any community should have awards, and most communities do and for good reason. It's an important people thing.
You've worked on a very diverse set of topics in computer graphics and vision over the years. Is there some sort of unifying theme that ties it together? I tend to switch research areas every 7 or 8 years. I find it keeps me fresh. Some people accuse me of just picking low hanging fruit, but I've found it a fruitful way to conduct my career. So it probably doesn't all tie together.

I guess, finely sampled representation is a unifying character of what I've done. Finely sampled image rendering, finely sampled volume rendering, finely sampled display technologies. All the work I've done works on fine sampling. But that's a very thin thread, a loose connection.

What changes have you noticed in SIGGRPAH over the years?

It's gotten a lot bigger. It's now a billion industry. In fact, it's several different biillion dollar industries, and that's changed its character. There seem to be two phases, one before it's a billion dolloar industry and one afterwards, although there was no sudden change visible.

SIGGRAPH used to be an academic conference. The computer vision conferences are like what siggraph used to be. They are not not surrounded by all this buzz, there aren't reporters running around, there are not these giant exhibitions. You'd have to go back 20 years, but SIGGRAPH used to be like those other conferences are now.

Do you have any memorable anecdotes about hijinx from the early years of SIGGRAPH? I'm not a good person to ask about that. I'm a square.

 

ADDITIONAL INFORMATION

people

terms

related reports/links links

 


 

This site is maintained by ACM SIGGRAPH Reports.
Last updated 8/11/03.

The ACM SIGGRAPH Reporter program is sponsored by ACM SIGGRAPH. Photos courtesy of Cybershot digital cameras generously loaned by SONY.