Reevaluating stereo and motion cues for visualizing graphs in three dimensions

Proceedings of the 2nd symposium on Appied perception in graphics and visualization - APGV '05 (2005), doi: 10.1145/1080402.1080411
Experiments
Tasks
Findings

The goal was to examine the effect of stereo, kinetic or motion depth and the use of 3D tubes versus lines, on understanding of graphs.

Variables and Constants

Independent Variables
  • motion - a second condition was motion. the graph was either static or rotated smoothly, taking 36 seconds to complete a rotation.
  • rendering style - a second condition was motion. the graph was either static or rotated smoothly, taking 36 seconds to complete a rotation.
  • graph size - 33, 100, 333 and 1000 nodes
Constants

Since the display remained the same, FOV, display size and form factor, resolution, latency, pixel density, refresh rate, frame rate, and other display factors remained constant.

System Info

Displays
  • - the author implemented a wheatstone mirror stereoscope where who displays are on the side of the user, and mirrors in the front of the user project the side display on the user's eyes. this was done to prevent ghosting and image brightness problems. the
Input Hardware
  • mouse - the mouse was used for the participant to answer task-related questions.
Software
None

4 PCs, each with an NVidia Quadro FX 3000G and an AMD Anthalon PX51 3400 processor with speeds of 2400 MHz, powered both displays. It appears from the images that the mirror projection is covering the maximum FOV of human perception.

Participant Info

15 participants were undergraduate students. 2 were the authors to get an estimate of what experienced observers would find.

Total # Age Range Gender Balance
17 -

Two nodes in the graph were highlighted in red, and participants were asked whether they were linked by a path of 2 or 3.

Environment

The density varied depending on the graph size, which had anywhere from 33 to 1000 nodes.

Dimensionality Scale Density Visual Realism
3D Not reported Controlled variable Low
Metrics

  • time - task completion time
  • accuracy - correct or incorrect
  1. There was a significant interaction between stereoscopy and motion on accuracy for a comprehension of spatial information task.

    Stereo & motion cues led to less error rates than having no cues or one of the other for 3D graph understanding tasks.

    Specificity: Neither

  2. There was a significant interaction between stereoscopy and motion on time for a comprehension of spatial information task.

    Stereo & motion cues led to faster response times than having no cues or one of the other for 3D graph understanding tasks. In addition, Both stereo and motion cues and stereo led to faster response times than non-stereo conditions for graph understandi

    Specificity: Neither
    The second result was stronger than the rest, since it was a result of a post hoc Tukey test in addition to an ANOVA. Teh authors found 15% speedup.