Fish Tank Virtual Reality

1
?? ?? ??
Fish Tank Virtual Reality Colin Ware and Kevin
Arthur Faculty of Computer Science University of
New Brunswice
2
Contents

• What is Fish Tank VR ?
• Advantages over immersion VR.
• Design properties of Fish Tank VR.
• Experiment 1 Subjective Impressions of 3D.
• Experiment 2 Tracing Tree Paths.
• Conclusion.

3
What is Fish Tank VR ?

• The head coupled system
• including a monitor, a device for measuring head
  position, and stereo goggles Fig 1.
• Variables that experiments investigate
• whether or not the perspective view is coupled to
  the actual viewpoint of the observer.
• whether stereopsis is employed.

4
Advantages over immersion VR(1)

• Resolution
• In immersion VR, given the typical resolution of
  current displays each pixel will subtend 12
  minutes of arc.
• Depth-of-Field
• Without directly measuring the focal length of
  the observers lens, cannot take the
  depth-of-field directly into accounting in
  creating images for VR.

5
Advantages over immersion VR(2)

• Stability in the Presence of Eye Movements
• A 40 deg eye movement result in a 3.8 mm
  translation of the center of the lens.
• Integration of the VR Workspace with the Everyday
  Workspace

6
Design properties of Fish Tank VR(1)

• Factors leading to the subjective impression of
  3D space
• Objects
• Fig 2. and Fig 3.
• Depth-of-Field
• made the background out of fuzzy discs, to give
  the illusion of depth-of-field
• Shading

7
Design properties of Fish Tank VR(2)

• Spatio-Temporal Accuracy in Head Tracking
• 3Space Isotrak which may be as much as 80 msec
  behind the current position
• ADL-1 which is 2 ms behind the current position
• Stereo Display
• StereoGraphics CristalEyes stereo system
  ability to display 120 frames/sccone
• High Frame Rate
• maximum of 60HZ

8
Experiment 1

• Subjective Impressions of 3D
• Experimental Conditions
• 1. Picture
• 2. Stereo only
• 3. Head Coupled Monocular
• 4. Head Coupled Binocular
• 5. Head Coupled Stereo

9
Results From Experiment 1
All Subjects 1 2 3 4 5 All 1 Picture
43 1 0 7 12 2 Stereo only
57 7 11 0 19 3 HC monocular
96 93 29 61 70 4 HC binocular
100 89 71 68 82 5 HC stereo
93 100 39 32 66
10
Experiment 2

• Tracing Tree Paths
• Recurrence relation for Tree consturction
• HorizontalSpacing(root) 8.0cm
• VerticalSpacing(root) 8.0cm
• HorizontalSpacing(child) 0.7HorizontalSpacing(r
  oot)
• VerticalSpacing(child) 0.7VerticalSpacing(root
  )
• Y(child) Y(parent) VerticalSpacing(1.0Rand(0
  0.25)
• X(child) X(parent) HorizontalSpacingRand()
• Z(child) Z(parent) HorizontalSpacingRand()

11
Task domain of Experiment 2

• Medical imaging where doctors may wish to trace
  blood vessels in brain scan data
• 3D software visualization where software
  engineers may wish to trace object dependencies
  between software modules represented as networks
  in 3D space.

12
Result From Experiment 2

• Timing data times(sec)
• 1. Picture 7.50
• 2. Stereo only 8.09
• 3. HC monocular 8.66
• 4. HC binocular 9.12
• 5. HC stereo 6.83
• Error data (errors)
• 1. Picture 21.8
• 2. Stereo only 14.7
• 3. HC monocular 3.7
• 4. HC binocular 2.7
• 5. HC stereo 1.3