1. VR equipment

2. Virtual reality High-end user-computer interface
Real-time simulations and interactions
Multiple sensorial channels:
visual
auditory
tactile
smell
taste
Simulation in which computer graphics is used to create realistic-looking world. Not static, but responds to user’s input. Key feature: interactivity
Interactivity and captivating power of graphics contribute to feeling of immersion.

3. Three I’s
Extent to which simulation solves problem is dependent on imagination. Imagination also refers to mind’s capacity to perceive non-existent things, such as this triangle (after Kanizsa).

4. History 1960: Heilig patents head-mounted display (for slides)
1962: Heilig patents Sensorama (VR video arcade)
1966: Sutherland builds CRT-based HMD
1973: Evans & Sutherland design graphics accelerator
1981: NASA develops LCD-based HMD
1985: Zimmerman & Lanier develop sensing glove
1993: IEEE organizes first VR conference
Sensorama: 3D video, motion, color, stereo sound, aromas, wind effects (using fans), vibrating seat. Possible to simulate motorcycle ride sensing wind, feeling pot-holes, smell food passing by store.
Evans&Sutherland: precursor of modern graphics accelerators; framerate 20/s; polys.
NASA: ripping apart Sony Watchman TVs; special optics to allow focussing

5. VR survival: PC vs. high-end SGI
Drawback of early VR: expensive (SGI Reality Engine > $ in ’93). Expectations to high due to media hype; VR could not deliver overnight and funding dried up. Mid-90s critical because focus was on Internet and Web apps. Rebirth late 90s: CPU speed, PC-based graphics accelerators. Also due to large-volume displays (instead of HMD).

6. Components

7. Input devices 3-D position tracking devices
Navigation and manipulation devices
Gesture interfaces

8. Tracker performance: Accuracy
Difference between actual position and measured position
Separately given for translation and rotation. Not constant, and degrades with distance from the origin of coordinate system.
Resolution: granularity of minimum change in position that can be detected
Repeatability: relates to jitter (next slide)

9. Tracker performance: Jitter
Change in tracker output when stationary (sensor noise)
No jitter -> constant value for stationary object. Needs to be minimal to avoid tremor, jumpy objects, etc. Also not constant and influenced by environmental conditions near tracker.

10. Tracker performance: Drift
Increase in tracker error with time
Inaccuracy grows over time (stationary object shown). Controlled by zeroing, using a second tracker that has no drift. Typical of hybrid trackers (not discussed).

11. Tracker performance: Latency
Time delay between action and result
Can induce motion sickness.Elements: measure object change, communication time between tracker electronics and host computer, time to render scene.
Sync tracker and communication with display loop (genlock).

12. Tracking devices
Mechanical
Ultrasonic
Optical
Magnetic

13. Mechanical trackers high accuracy low jitter no drift low latency
Drawbacks:
limited range
weight
limited freedom of motion
few mm, 0.2 microseconds latency
Larger: inertia and weight increase resulting in mechanical oscillations

14. Ultrasonic trackers high accuracy acceptable jitter no drift
high latency
Drawbacks:
very limited range
direct line of sight required
background noise corrupts signal
Based on triangulation; 3 microphones and 3 speakers; speakers activated in cycles, 3x3 datasets, derive pos and orientation.
Update rate low, wait ms to allow echoes to die out.
1,5 m range (air absorpts signal)

15. Optical Trackers high accuracy low jitter no drift low latency
Drawbacks:
direct line of sight required
environmental reflections
IR sources corrupt signal
0.5 mm, 0.03 deg.

16. Magnetic trackers high accuracy low jitter no drift low latency
Drawbacks:
limited range
electromagnetic devices corrupt signal
metallic objects corrupt signal
Three antennas formed of three orthogonal coils wound on ferromagnetic cube. Excited sequentially.
Intensity of signal falls with cube of distance.
2.5 mm, 0.5 deg

17. Navigation and Manipulation Devices
Wanda
Cubic mouse
5DT Data glove

18. Wanda

19. Cubic mouse
9x9x9 cm with tracker. translating rods can be pushed in and out, 6 buttons. Designed for manipulation of single large model.

20. 5DT Data Glove
Gesture interface measures position of fingers (sometimes also wrist) to allow gesture-recognition based interaction.
Here: one sensor per finger and tilt sensor to measure wrist orientation. Finger bending measured based on intensity of returned light emitted from LED. Fiber walls change index of refraction upon finger flexion.
Thumb not used: 2^4 = 16 combinations for gestures.

21. Output devices
Stereo issues
Graphics displays
Haptic displays

22. Stereo vision Field of view in humans: 180 (H) by 120 (V) degrees
Stereopsis (binocular overlap): 120 degrees
Interpupillary distance: 53 – 73 mm

23. Active stereo

24. Passive stereo

25. Graphics displays Head-mounted Floor-supported Desk-supported
Large-volume
Projector-based

26. Fishtank VR
3D Display
2D Input
1D Input
3D Input

27. Projection-based displays
Single display
Theatre
TiledWall
CAVE (4- and 6-sided)
FlexWall

28. Workbench

29. Dual Workbench

30. Theatre (curved screen)

31. TiledWall

32. CAVE (Cave Automatic Virtual Environment Uni. Illinois 1993)

33. Reality Cube (Cave)
Zernikeborg
Theater

34. CAVE computer graphics issues

35. FlexWall

36. Next lecture
OpenGL pipeline
Graphics hardware
Scenegraphs (Performer)