Establishing the Range of Perceptually Natural Visual Walking Speeds for Virtual Walking-In-Place Locomotion
The objective of this study was to uncover how the upper and lower bounds of perceptually natural visual speeds might vary across different FOVs during walking-in-place (WIP) and treadmill locomotion.
- field of view (fov) - vertical fovs of 25°, 50° and 75° (aspect ratio=1.25) and the unconstrained fov of the oculus rift developer edition
- movement type - the user either walked in place or walked on the treadmill
- head-mounted display - nvisor sx60
- optitrack motion capture system - a 16 camera camera motion capture system was used to track a marker placed on the head-mounted display
ProForm 520 XLT treadmill
The participants were recruited via a mailing list comprising volunteers from Aalborg University Copenhagen and readers of the Danish periodical Ingeniøren (The Engineer). All participants had prior experience with virtual WIP locomotion due to participation in previous studies and reported having normal or corrected-to-normal vision and hearing.
|Total #||Age Range||Gender Balance|
|21||18 - 44|
For each condition (2 movement types x 4 display FOVs), the participant was presented with 11 visual speeds repeated twice. Thus the user performed 22 walks. The 11 visual speeds were defined as the normal walking speed of the user multiplied by gains ranging from 1.0 to 3.0. A gain of 1.0 would correspond to the normal walking speed of the user. This normal walking speed was established prior to the first trial and corresponded to the speed of the treadmill which the user found comfortable while walking at 1.8 steps per second.The series of visual speeds presented during each condition were either commenced with the slowest visual gain (1.0) or the highest (3.0). After each walk, the gain would change in increments of 0.2. If the series started with the lowest gain, the gains would gradually increase until the highest gain was reached and then decrease until returning to the lowest gain again. The same logic applied if the first gain in the series was 3.0, albeit in this case the gains would gradually descend before ascending. For each series it was randomly decided whether the first gain should be 1.0 or 3.0. For each walk the participants were asked to report whether they had found the virtual speed ‘too slow’, ‘natural’, or ‘too fast’. These verbal judgments were given once the participants felt confident enough to do so or when the walk was over. The participants did at all times walk at a fixed step frequency of 1.8 steps per second.
Interaction and Environment
The visual stimulus was comprised of a 14m long hallway. The participants were instructed to fix their gaze at a painting on the back wall. During all walks the participants were located on the treadmill and asked to hold on to the handlebars.
- user task perception - for each walk the user was asked to rate whether the visual speed was 'too slow', 'natural' or 'too fast'.
There was a significant inverse effect of field of view (fov) on user task perception for a comprehension of spatial information task.
The results revealed significant main effects for both movement type and viewing condition, but no significant interaction between the two variables was found. The post-hoc analyses did not uncover any significant differences in regards to movement type,
It would be beneficial to investigate how properties of the visual stimuli (e.g. the geometric FOV, latency, resolution, weight, environmental motion cues and external peripheral stimulation) influence the range of perceptually natural visual gains. Research emanating from fields such as biomechanics has led to many insights regarding the relationship between gait parameters and walking speed. However, considering that the biomechanics of stepping in place differ from those of real walking, it is necessary for future studies to investigate how variations in gait parameters, such as step frequency and height, influence the perceived naturalness of walking speeds during WIP locomotion. Finally, the current studies were limited to forward translational motion and it remains to be seen whether the results are applicable to walks involving rotational motion.