DETAILS


COLUMNS


CONTRIBUTIONS



ABOUT THE COVER


Vol.33 No.3 Aug. 1999
ACM SIGGRAPH


Cover Images Expand Issue Topics

Karen Sullivan
Computer Graphics Cover Editor



August Columns
From the Editor Entertaining the Future


About the Cover
Previous About the Cover Next About the Cover


More on Fiat Lux and The Story of Computer Graphics

For more information, contact:

Paul Debevec
University of California at Berkeley
387 Soda Hall #1776
Berkeley, CA 94720-1776

Front Cover

Fiat Lux by Paul Debevec

The front cover shows how image-based modeling, rendering and lighting techniques were used to create the animation Fiat Lux both for the SIGGRAPH 99 Electronic Theater and inclusion in the ACM SIGGRAPH HDTV 90-minute video documentary The Story of Computer Graphics, premiering at SIGGRAPH 99 (see sidebar). The film features a variety of dynamic objects realistically rendered into real environments, including St. Peter’s Basilica in Rome. The geometry, appearance and illumination of the environments were acquired through digital photography and augmented with the synthetic objects to create the animation. The film builds on the techniques of The Campanile Movie and Rendering with Natural Light from SIGGRAPH 97 and 98.

The problem with most image-based modeling techniques is that they are only able to render the models as they originally appeared in the photographs. This makes it difficult to add new objects to the scenes, since adding such objects requires illuminating them properly and having them cast all the appropriate shadows and reflections in the scene.

After finishing The Campanile Movie, we set out to try to be able to make changes and modifications to image-based models. The key to doing this was to accurately record the illumination present in the environments that we model. Using a special high dynamic photography method (without it, real cameras don’t have nearly the range to accurately record real-world illumination conditions), we found that it was possible to use measurements of real illumination in the world to illuminate synthetic objects - a technique we called “image-based lighting.” We demonstrated this basic technique at last year’s Electronic Theater in Rendering with Natural Light (RNL), a short animation of a collection of ornamental spheres on a pedestal, illuminated by the light in UC Berkeley’s Eucalyptus grove. The film also demonstrated how we try to use the measurements of real illumination to produce a variety of realistic effects of cinematography, such as soft focus, vignetting and flare effects.

For Fiat Lux, there were many new things we wanted to try to do with the technology. First, we wanted to show shadows and reflections based on real illumination, calculated without any synthetic light sources. Second, we wanted to render dynamic, moving objects into the scenes, which would cause us to reinvent and refine our rendering techniques to increase their efficiency and in order to achieve acceptable consistency from frame to frame. Third, we wanted to combine image-based lighting with the techniques we used in The Campanile Movie to create virtual camera moves through luminous space. Fourth, we wanted to show majestic environments with beautiful complex lighting, with sets of objects one could clearly only imagine being there. And finally, we wanted to use these techniques to tell a story.

The result was Fiat Lux. Set in a recovered model of St. Peter’s Basilica, featuring hundreds of gigantic steel spheres and monolithic falling dominos, it presents an abstract interpretation of the conflict between Galileo and the church, between science and religion and between truth and superstition. Fiat Lux — Latin for “Let there be light” — seemed to resonate with both the scientific and religious themes of the piece, as well as to directly reference the new technology used to create it. And “Fiat Lux” is the motto of the University of California at Berkeley to boot.

Cover Image

Cover Image

Cover Image

Images

Top: A section of a panoramic rendering of the main environments used in the film, with synthetic objects in dynamic motion.

Bottom left: Two images of a two-inch mirrored sphere placed in front of Bernini’s Baldacchino inside St. Peter’s.

Bottom right: An example of one of three panoramic enviroments created by taking multiple radiance images and assembling them into panoramas. These were used to create the background plates for the film. Three-dimensionality was added to these backgrounds by projecting them onto a model of the corresponding environments.

Contributors

Christine Cheng, H.P. Duiker, Tal Garfinkel, Tim Hawkins, Jenny Huang and Westley Sarokin. Supported by Interval Research Corporation, the Digital Media Innovation program and the Berkeley Millennium project.

See also Paul's website.

References

1. Debevec, Paul. “Rendering Synthetic Objects into Real Scenes: Bridging Traditional and Image-Based Graphics with Global Illumination and Dynamic Range Photography,” in SIGGRAPH 98, July 1998.

2. Debevec, Paul and Jitendra Malik. “Recovering High Dynamic Range Radiance Maps from Photographs,” in SIGGRAPH 97, August 1997.

3. Debevec, Paul, Camillo J. Taylor and Jitendra Malik. “Modeling and Rendering Architecture from Photographs,” in SIGGRAPH 96, August 1996.

Back Cover

Illustrating our focus topic, “Presenting Visual Information Responsibly” guest edited by Nahum Gershon, these three images produced by Pierre-Yves Bertholet of the MITRE Corporation are Web accessible representations of uncertainty in a battleship visualization. The top image depicts error bars around the history trails of the red force and different future paths based on probability. The image also graphically represents velocity vectors, units of forces and time stopped.

The middle image depicts the spread of two units. The program varies the opacity of the spread polygon as a function of the density of constituent units, so that operators are not overwhelmed by large spreads with low unit density.

The bottom image depicts imperfection in time of the history information of two battleships. It is taken from a sequence in which time passes, with the associated history trails beginning in white, turning yellow, then red and finally disappearing as the information becomes older and less useful in the present. The most current information is thus always displayed in white, using a small but opaque symbol to grab the users’ attention. Successively older pieces of information are represented with progressively more transparent symbols and colors to add a representation of the age of the information to the display.

This program was developed by Pierre-Yves Bertholet of the MITRE Corporation, and aims to give users a sampling of ways in which visualization displays can be improved. It also demonstrates that a complex military visualization can be served over the Web. It uses a Java applet to control a VRML scene that in turn is served to the client’s Web browser. The user can then use a Java control panel to turn on and off 3D error clouds, velocity vectors, history and future trails, unit spread and flight corridor representations while “flying” through the VRML scene.



Karen Sullivan is a faculty member at Ringling School of Art and Design in the Department of Computer Animation and Foundation Studies. Her major focus of research is in concept, narrative and literacy for media and animation. Her video installations and single channel pieces have been shown nationwide. Karen received her M.F.A. from the University of Massachusetts, Amherst, and her B.F.A. from Indiana University, Bloomington.

Cover Image

Cover Image

Cover Image

For more information, contact:

Pierre-Yves Bertholet
The MITRE Corporation
1820 Dolley Madison Blvd.
McLean, VA 22102-3481

Karen Sullivan
Ringling School of Art and Design
2700 North Tamiami Trail
Sarasota, FL 34234

Tel: +1-941-351-5100


The copyright of articles and images printed remains with the author unless otherwise indicated.