Communications of the ACM
Home/Magazine Archive/May 2001 (Vol. 44, No. 5)/Building Large-Format Displays For Digital Libraries/Full Text
Communications of the ACM
Building Large-Format Displays For Digital Libraries
Imagine visiting a museum and being asked to view the latest exhibit through a 20-inch window. Museum curators will gladly maneuver items to any position you ask, but visualization is always limited to the window. Such is the current state of digital libraries.
Digital libraries house media of ever-increasing complexity, from models of statues to virtual archeological landscapes. Future digital archives will only increase in resolution and complexity as acquisition hardware improves and data becomes more intricate. Although digital content is constantly improving, the typical display device (the desktop monitor) remains relatively unchanged.
The problem is the spatial resolution and color quality of digital media surpasses the display real estate of typical monitors. It is common for objects such as artwork and manuscripts to be digitized at a much higher resolution than can be displayed on even the best monitors. This forces the data to be down-sampled, or restricts viewing to selected regions of interest at a time. Even with improved resolution, small display areas are often less than compelling. For example, the high-resolution facsimile of Michelangelo's 17-foot statue of David [4] is somewhat less spectacular when displayed on a high-resolution 20-inch monitor. The simple fact is monitors have become under-powered technologies for the visualization capabilities now needed to fully appreciate digital archives.
The immediate reaction is to enlist the help of large-format displays for visualization available as video walls, domes, and immersive environments such as the CAVE. These environments provide a compelling sense of presence and help break the "window on worlds" paradigm.
Unfortunately, these technologies also present serious challenges that limit their accessibility and usefulness to a small number of institutions. Of primary concern is the cost. Initial costs of projectors, rendering hardware, and subsequent recurring operational expenses are fiscally prohibitive. A second concern is the renovation necessary for installing a large-format display system. Because most systems rely on mechanically aligned projectors and rigidly constructed display surfaces, room modification and projection surface infrastructure are necessary. An additional difficulty is the complexity of continuous operation. These systems are not for novice users and can quickly become unusable without expert maintenance and tuning.
Our goal is to bring large-format displays to the digital library community by breaking down the barriers that make them expensive and difficult to set up and run. We are attacking the price/performance barrier by engineering displays built entirely from commodity components and assembled in a scalable configuration. Specifically, commodity light projectors, unlike monitors, can be positioned collectively to form a single logical desktop (see Figure 1). Inexpensive video cards can perform at the level required to drive a projector. Groups of PCs, each driving a projector and communicating via a local area network, are inexpensive and powerful [2, 6].
While the physical alignment of projectors creates a very compelling display, it is tedious and requires major effort. Within the digital library community, where libraries must operate and maintain equipment, it is important to enable novice technicians to build up and maintain a display by a more casual placement of projectors. The overlap and relative arrangement of projectors could be arbitrary, as long as they illuminate a continuous area on the display surface. With the functionality of the PC and graphics card rendering the image, the overlapped projectors can be blended into a single, logical display.
Recent research efforts [3, 5, 7] into less restricted front-projected visualization environments are making this a viable solution for large-format displays. These techniques use cameras to register the geometry of the display area. The cameras capture images of what is being displayed, and these images guide the process of feathering the projectors into a seamless display. Projectors can be placed in a room and positioned so they illuminate a desired region on a wall or even on arbitrary surfaces (see Figure 2). One or more cameras can be used to configure and blend the projected images together to form a single logical projector. Groups of computers, each driving a projector, communicate to synchronize the display, share models, and react when cameras detect changes in the display quality. This research is promising to make wide field-of-view, large-format display much more accessible in everyday computing.
By lowering the cost and automating all but the most basic set-up and operational functions, it is reasonable to envision libraries that own and operate wide-area displays [1]. University and public libraries could convert existing rooms into immersive displays, where patrons load and visualize digital media. Our hope is the digital library community will find ever broader support and appeal by providing compelling, affordable, easily configurable, high-resolution displays, which do justice to the amazing digital collections being created around the world.
References
1. Brown, M.S. Digital Special Collections: Acquisition, Restoration, and Display of Scholarly Texts. (Ph.D. proposal Oct. 6, 1999), University of Kentucky Technical Report TR31401.
2. Humphreys, G., and Hanrahan, P. A distributed graphics system for large tiled displays. In Proceedings of IEEE Visualization. (San Francisco, Oct. 1999).
3. Jaynes, C. et al. Radiometric Correction in the CoRE Display Environment. University of Kentucky Technical Report TR31301, Jan. 2001.
4. Levoy, M. et al. The digital Michelangelo Project: 3D scanning of large statues. In Proceedings of SIGGRAPH '00. (New Orleans, July 2000), 131144.
5. Raskar, R. et al. Multi-projector displays using camera-based registrations. In Proceedings of the IEEE Visualization Conference. (San Francisco, Oct. 1999), 161168.
6. Samanta, R. et al. Load balancing for multi-projector rendering systems. In Proceedings of the 1999 Eurographics/Siggraph Workshop on Graphics Hardware. (Los Angeles, Aug. 1999).
7. Welch, G. et al. Projected imagery in your office in the future. IEEE Computer Graphics and Applications (July/Aug. 2000), 62-67.
Figures
Figure 1. A display created by two abutted light projectors. The desktop is generated by a PC graphics card that supports multiple video outputs. The resulting imagery is 8 x 4 ft, with an effective resolution of 2048 x 768.
Figure 2. (a) Arbitrary placement of projectors; (b) A display of casually aligned projectors. The imagery, created from projectors situated in the corner of a room, looks correct to the viewer's perspective.
©2001 ACM 0002-0782/01/0500 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2001 ACM, Inc.
No entries found