Spatial Cues in 3D Visualization - Ambient Intelligence for Scientific Discovery - page 120

In the positioning task, subjects viewing the objects in stereo were more ac-
curate, faster, and used less rotational distance than when positioning objects
viewed in mono. With objects casting a shadow, subjects were more accurate,
took more time, and used less rotational distance than when positioning ob-
jects viewed with no cast shadows. While positioning objects against a (more
complex) zig-zag background surface shape, subjects were equally accurate, but
took more time and used less rotational distance than when positioning objects
against a (more simple) flat background surface shape. Finally, while position-
ing objects against a (more complex) checker-board background texture, subjects
were equally accurate, but again took more time and used less rotational dis-
tance than when positioning objects against a (more simple) solid background
texture.
In the resizing task, subjects viewing the objects in stereo were more accu-
rate, faster, and used less rotational distance than when resizing objects viewed
in mono. Objects casting shadows had no significant effect on the resizing perfor-
mance measures. While resizing objects against a (more complex) zig-zag back-
ground surface shape, subjects were less accurate, took more time, and used less
rotational distance than when resizing objects against a (more simple) flat back-
ground surface shape. Finally, while resizing objects against a (more complex)
checkerboard background texture, subjects were equally accurate, but again took
more time and used less rotational distance than when resizing objects against
a (more simple) solid background texture.

4 Results Summary

Instead of summarizing the results across the three experiments together, we
summarize the results for each of the depth cues of interest separately in the
sections that follow. In general, stereo viewing was the most effective cue for
enhancing performance in each task. However, motion was 'a close second' in this
regard. Finally, we discuss a limited number of salient implications for the design
of effective 2D and 3D user interfaces and for depth cue theory in Section 5:
Conclusions and Discussion.

4.1 Motion

Motion is a powerful depth cue that rivals stereo viewing in providing information
about the relative distances of viewed objects. Consequently, motion, in general,
can enhance spatial task performances and can provide useful information about
the spatial properties (such as size, orientation, distance) of objects viewed in
space. However, the utility of motion in promoting spatial task performances is
also a function of the task type, and is tied to individual characteristics, as well
(for example, see the discussion of gender that follows).
The incorporation of motion cues in visual activities can take many forms,
from moving or rotating a viewed object, to moving or changing the perspec-
tive of the viewer, to moving or rotating the entire field of view. In general,