A user study comparing head-mounted and stationary displays
How do head-tracked or hand-tracked camera controls (specifically, head motion to move a camera in an HMD, versus stationary display with viewpoint controlled by orientation-tracked hands) affect search tasks?
Variables and Constants
- head tracking - the hand's orientation controlled the viewpoint, with the hmd mounted in a fixed position on the ceiling, like a fixed periscope. the orientation tracker could be held by either hands. so, the experiment has 2 main conditions: head tracking controlling th
To keep all conditions constant, the same display and tracker were used in all conditions. The hand-tracked condition had a mount from the ceiling to keep the same HMD in a stationary, fixed position. So the following were held constant: Same tracker with lag and jitter Same physical display in all conditions with the same brightness resolution, density, FOV, etc. Constant environment.
- head-mounted display - the hmd was a vpl eyephone, with a resolution of 185 x 139 color pixels with 80 horizontal x 60 vertical fov.
The computer is s Silicon Graphics Crimson VGX. I am assuming that this computer came with its default monitor, PowerVision Graphics, that supports stereoscopic viewing, hardware enabled texture mapping, 1280 x 1024 px, 24-bit color, (source: http://www.sgistuff.net/hardware/graphics/powervision.html) Frame rate: around 10 per second Note: Differences in FOV, resolution were fixed to be the constant. Listed above are the specs for the hardware, unchanged.
All participants were from an introductory undergrad CS class.
|Total #||Age Range||Gender Balance|
The participant had to identify 20 targets in the room. A target had a number. The participant had to shout out the number, and then that target would disappear.
Interaction and Environment
The only interaction supported is moving the camera, and that changed depending on the condition.
Subjects were in an environment where they were in a 6 meters long, 6 meters wide, and 3.4 meters tall room. There were 20 targets, with each containing a number around .3m tall. Each corner of the environment had a piece of furniture, as an orientation cue.
There was a significant inverse effect of head tracking on time for a visual search task.
Participants with head tracking performed 42% faster than participants with hand tracking to move their viewpoint for a generic search task.
Specificity: Somewhat specific
The graph looks like it would be significant, but there was no wordage nor indication that the results were statistically significant. No mention of type of test made, nor the p-values.
There was a significant direct effect of head tracking on training transfer for a visual search task.
If participants started with head-tracking for viewpoint movement, they were faster on the second condition (hand-tracked) than those who did not, for search tasks.
Specificity: Somewhat specific
They found that those starting with head-tracking were 23% faster on their second condition than those that started with the hand-tracked condition. No mention of statistical significance, however, nor were they explicitly evaluating order effects.