标题: Exploring the Effect of Tiling on Large Displays [打印本页] 作者: 彬彬 时间: 2012-1-3 11:16 标题: Exploring the Effect of Tiling on Large Displays 1 Introduction
Over the past several years tiled displays have slowly evolved from
expensive blended-projection systems to Liquid Crystal Display
(LCD) tiled displays. The LCD systems are typically cheaper to
cons***ct and maintain, but the bezels introduced at screen edges
are often a common source of complaints. Our long term research
goal is to settle the debate between the two technologies to see if
there really is any perceivable benefit to one versus the other. The
work outlined here presents the first in a series of comparative experiments.
Figure 1: The Immersive Visualization Center at Texas A&M University
consists of a single continuous curved screen. In this image
software seams are introduced to investigate the effect of tiling on
performance in a navigation task.
2 Experiment
We investigated performance on a simple navigation and wayfinding
task in each type of system, one a low-cost, tiled, multiscreen
immersive visualization system and the other a more expensive,
continuous screen, immersive visualization facility. The low
cost system (LCD) is designed using off-the-shelf components and
cons***cted by arranging LCD displays in a curved tiled layout.
The expensive system (IVC) is a Rockwell-Collins semi-rigid, rear
projected, continuous curved screen. With the low cost paradigm,
physical seams are introduced into the image where the displays
are tiled. Our hypothesis is that the tiled system presents an equivalent
visual experience for navigation and way-finding tasks, despite
the separating seams introduced by connecting the screens. Each
immersive system was evaluated by measuring timely task performance
in a psychophysical experiment. There were 20 participants,
ten of which viewed the LCD system first and ten viewed the IVC
system first. There were five conditions which manipulated the independent
variable,seam size (please see supplemental material for
further details). We introduced software seams in the IVC to mimic
the physical bezels present in the LCD system. In each condition
participants were asked to navigate five separate routes, labeled A,
B, C, D and E. Each route was indicated by a starting point and a
target end point. A 2D map was clearly marked with start and end
points was provided. Participants could view the map for as long
as they wished, but typically only studied it for a minute or two
before declaring themselves ready to navigate the Virtual Environment
(VE). After studying the map participants were placed in the
e-mail: ann@viz.tamu.edu
VE located at the start point. The task was to navigate as quickly
as possible through the 3D environment, beginning from their start
position and finishing at the indicated end position. This is known
as a goal-directed navigation, or primed navigation task [Darken
et al. 2001].
3 Results
Time to target per participant across system was recorded and compared
in an effort to gauge performance. As previously mentioned,
our premise is that faster times to reach the end position demonstrate
higher spatial understanding and hence performance in the
system. Average “end time” across participant in each environment
was compared. A two-way ANOVA showed no significant differences
across seam conditions:F(4) = 2:54; p < 0:04 or across
route map F(4) = 31:54; p < :005. This suggests that both seam
size and route influence performance time, but there is no evidence
of an interaction between the two F(16) = 0:63; p > 0:5. These
results suggest that, for this navigation task, the presence of seams
did not adversely impact the performance of simple navigation in
the tiled display environments.
4 Conclusion
The goal of this work was to show that navigation performance is
equal (on average) in systems with or without seams, and, therefore,
not impacted by the size or presence of display seams, be
they physical or virtual (introduced by the software). We compared
performance on a simple goal-directed navigation task in both immersive
systems and discovered that no significant differences exist
across systems, which validates our hypothesis that there is no di***ption
to the visual experience of the user when navigating a VE
populated with physical or software seams. By comparing navigation
to a target across seamed VE systems, it appears that seams
have no significant impact on the time taken to complete a simple
guided navigation and way-finding task. In fact, in some cases participants
actually performed navigation tasks better with the presence
of seams. Informal comments indicated that some users actually
preferred the presence of seams as they helped to partition the
scene and made navigation more manageable. This also may be due
to an extension the phenomenon discovered by several researchers
in which the separation of displays helps users to streamline tasks
in 2D desktop systems [Mackinlay 2004]. Clearly, much work remains
to be done in this area. The experiment represents the beginning
of a larger study in which we hope to fully investigate the
impact of seams on performance across immersive systems.
References
DARKEN, R. P., PETERSON, B., AND ORIENTATION, B. S. 2001.
Spatial orientation, wayfinding, and representation. In In K. M.
Stanney (Ed.), Handbook of Virtual Environments: Design, Implementation,
and Applications, Erlbaum, 493–518.
MACKINLAY, J. D. 2004. Wideband displays: Mitigating multiple
monitor seams. In CHI 2004 Extended Abstracts, ACM Press,
