Effects of Field of View on Performance with Head-Mounted Displays

Doctoral Dissertation at University of North Carolina at Chapel Hill (2000),
Experiments
Tasks
Findings

What are the effects of FOV on task completion times, cybersickess and other factors in navigation and search tasks in a VE?

Variables and Constants

    Visual Fidelity
  • field of view (fov) - 48, 112, 176 degrees horizontal
  • form factor - the weights between the different hmd's differed, as the "real-world" hmd was lighter than the others.
  • resolution (number of pixels) - please see the display specs section for more details.
  • latency - this factor differs a little between the two hmd's and the real-world version.
  • jitter - since one of the conditions compared a real world version, latency & jitter differed between the computerized and real versions of the environment.
Independent Variables
  • different hmds - two brands of hmds were compared against as well. please see the display specs section for more details.
  • visual realism - this differed between the different display conditions, for both between the two hmd's and the real-world version.
Constants

FOV vertical remained constant at 47/48 degrees. Head tracking, stereo (or perception of depth in the real-world condition) and latency, among others, were held constant.

System Info

Displays
  • head-mounted display - while the analysis of this study was an independent components analysis, another controlled variable were two different hmd's with different specs, along with a real-world scenario with blinders to control fov. this variable did not turn out to be signifi
Input Hardware
  • hiball head tracker sensor - head tracking sensor
  • mouse - for a task that required recalling spatial memory, the participant used a monitor & mouse to navigate through the scene.
Software

The machine was a SGI Onyx2 Reality Monster with InfiniteReality2 graphics and a R12000 CPU.

Participant Info

All were screened for color-blindness and corrected vision.

Total # Age Range Gender Balance
5 21 - 30

The participant had to go through a maze in a quick and comfortable pace, without hitting any of the walls. Collisions with the walls are the same as error rates in this task.

Environment

The environment was similar to the first task, except there were tall maze-like walls.

Dimensionality Scale Density Visual Realism
3D Small Medium Medium
Metrics

  • time - completion times
  • errors - errors really were collisions of the participants hitting the maze walls.
  1. There was a significant inverse effect of field of view (fov) on time for a travel - maneuvering task.

    The linear model made for this correlation seems especially strong (R^2 = .73, P < .0001). Though, with 5 participants and other controlled variables that did not give significant results (particularly, different HMDs & a real-world version), This result

    Specificity: Neither
    The linear model made for this correlation seems especially strong (R^2 = .73, P < .0001). Though, with 5 participants and other controlled variables that did not give significant results (particularly, different HMDs & a real-world version), This result was combined the results from all the different display conditions all-together. If 176 degrees FOV gave 100% performance, 112 degrees FOV would result in 88%, and 48 degrees would result in 76%.

  2. There was a significant inverse effect of field of view (fov) on time for a travel - maneuvering task.

    The linear model made for this correlation seems especially strong (R^2 = .73, P < .0001). Though, with 5 participants and other controlled variables that did not give significant results (particularly, different HMDs & a real-world version), This result

    Specificity: Neither
    The linear model made for this correlation seems especially strong (R^2 = .73, P < .0001). Though, with 5 participants and other controlled variables that did not give significant results (particularly, different HMDs & a real-world version), This result was combined the results from all the different display conditions all-together. If 176 degrees FOV gave 100% performance, 112 degrees FOV would result in 88%, and 48 degrees would result in 76%.

The participant had to turn their head to find a target that is initially outside their FOV. The participant could not translate or move from their position. All the participant had to do was rotate their heads.

Interaction and Environment

Interface

N/a, though it is assumed that there must have been some push-button mechanism to indicate when teh target was found.

The environment was an empty room with the walls textured to be like a lab. It appears the wall textures were taken from real-life photos of a lab.

Dimensionality Scale Density Visual Realism
3D Small Medium Medium
Metrics

  • time - completion time
  1. There was a significant inverse effect of field of view (fov) on time for a interactive experience of the ve task.

    Wider FOV led to shorter walking time in a VE maze.

    Specificity: Neither
    The linear model made for this correlation seems especially strong (R^2 = .66, P < .05). With 5 participants and other controlled variables that did not give significant results (particularly, different HMDs & a real-world version), This result was combined the results from all the different display conditions all-together. If 176 degrees gave 100% performance, 112 degrees would be at 77%, and at 48 degrees would yield 69% performance.

  2. There was a significant inverse effect of field of view (fov) on time for a interactive experience of the ve task.

    Wider FOV led to shorter walking time in a VE maze.

    Specificity: Neither
    The linear model made for this correlation seems especially strong (R^2 = .66, P < .05). With 5 participants and other controlled variables that did not give significant results (particularly, different HMDs & a real-world version), This result was combined the results from all the different display conditions all-together. If 176 degrees gave 100% performance, 112 degrees would be at 77%, and at 48 degrees would yield 69% performance.