Viewer explodes virtual buildings
Ted Smalley Bowen,
Technology Research News
Fully immersive virtual reality programs
are an impressive way to experience digital architectural models, but
they don't always provide the best view, especially where action is involved.
Researchers at the University of California at Santa Barbara,
Stanford University, Microsoft Corporation and the University of Virginia
have devised a way to add expanded, or exploded, views to the three-dimensional
architectural graphics used by real-time programs like computer games
and training simulations.
An exploded view renders components of multipart objects like
buildings and machines separately, opening up a building, for example,
to make it possible to see the interiors of all floors at once. The view
preserves relative positioning among all the model's details, including
vertical supports, doors, and furniture.
The researchers' prototype software works with applications that
portray action within three-dimensional spaces and can make them easier
to follow, according to Christopher Niederauer, a researcher at the University
The software provides a bird's-eye view, which is useful for following
team interactions, said Niederauer. "Military strategists and even police
could potentially use our software... to aide in planning group interactions
within a building," he said. "It is a really cool way to... see everything
that is going on."
The software renders architectural models from an external vantage
point rather than generating a more compute-intensive immersive, or first
person, perspective, which allows the user to view the model as if he
were inside. The external viewpoint also allows the software to show more
of an interior model, because immersive viewpoints selectively omit structural
details and objects to simulate a person's perceptions of space.
The software automates portions of the design-intensive process
of locating each of a model's stories and generating the exploded view.
It uses Chromium, which is software that can modify, delete or replace
graphics commands on-the-fly from programs written in the OpenGL programming
language. This allows the exploded view visualizer to alter three-dimensional
graphics programs as they run.
The software first determines the location of the model's layers
-- for example each story of a multistory building -- by analyzing the
architectural model. It then modifies the application's graphics output
to separate each story and render the exploded view.
The program splits stories off from the original architectural
model just below the ceiling, affording a less obstructed view of the
interior. And it allows the user to choose the viewpoint and the spacing
between stories in the exploded view.
The exploded view is axonometric, meaning it represents three-dimensional
objects with vertical and horizontal dimensions drawn to scale, but distorts
curves and diagonals.
The on-the-fly rendering of each frame of an interactive program
takes a lot of computer power, but is within the capability of a desktop
PC, according to Niederauer. "Games these days are optimized to only show
one room at a time, but in order to show everything, we need to draw everything,"
he said. "We still get interactive rates when using a single computer,"
The researchers are currently working on improving the interface,
including simplifying the characters' appearance, possibly representing
them as icons in order to make them easier to track in the exploded view.
The idea of converting computer models of multistory buildings
to better show both exterior and interior details is not new, but the
researchers have automated the process and found new applications for
the technique in computer graphics, said Michael Ashikhmin, assistant
professor of computer science at the State University of New York at Stony
Brook. "The low level techniques are extremely simple and well known but
the [researchers] combine them in a novel and interesting way."
The software could be used to better visualize interactive multiplayer
games such as Doom without any modification to the existing code, he said.
"This by itself is an interesting development, especially in environments
where a third person is allowed to follow the game action."
The software can be used with existing three-dimensional programs
now, according to Niederauer.
Niederauer's research colleagues were Mike Houston of Stanford
University, Manesh Agrawala of Microsoft, and Greg Humphreys of the University
of Virginia. They presented the work at the Association of Computing Machinery's
Symposium on Interactive 3-D Graphics (ACM's I3D '03) in Monterey, California,
April 27 to 30, 2003. The research was funded by the Department of Energy
TRN Categories: Data Representation and Simulation; Graphics;
Story Type: News
Related Elements: Technical paper, "Non-Invasive Interactive
Visualization of Dynamic Architectural Elements," presented at the ACM
Symposium on Interactive 3-D Graphics (ACM's I3D '03), April 27-30, Monterey,
California, and posted at graphics.stanford.edu/papers/archsplit/archsplit_lowq.pdf
August 13/20, 2003
Skulls gain virtual faces
Tool blazes virtual trails
keeps it simple
Video keys off human
Device simulates food
nears quantum limit
Molecule makes ring
expand nano toolkit
Research News Roundup
Research Watch blog
View from the High Ground Q&A
How It Works
News | Blog
Buy an ad link