Valve Index VR System
Methods Summary
Anthropometry, A/B testing, Usability Testing, OOBE testing, Manufacturing tolerance threshold
I was with Valve Corporation from February 2018 to March 2029 on a series of two 6-month contracts. While there I was initially handed a research question about how severe tracking errors need to be before they cause motion sickness in a VR user.
This was an exciting new space for me to investigate, as I knew nothing about VR at the time. I jumped right in and became an expert in the subject by trying out all of the top VR experiences in the Steam VR store. I read up on the differences between Steam VR headset’s lighthouse tracking system, as compared to the inside out tracking of the Windows Mixed Reality headsets.
Tracking Error Induced Motion Sickness
How I learned to dig on a research question
The research question that was handed to me when I began was:
How severe can the tracking errors become before inducing motion sickness in a user?
Abstract
After becoming an expert in the VR space, I determined that the best route to tackling this problem was to benchmark the VR experience. Using our VR headset, I enlisted the help of 30 employees of the most diverse representation I could find. The participants tested three different types of content in VR in an attempt to induce tracking errors and therefore motion sickness. From this study I was able to identify that users are very quick to overlook tracking errors. Many not even noticing, or forgetting they occurred by the time they reached the survey 15 minutes later. However motion sickness was more of a factor of the content with which a user interacts. Only the most severe tracking errors would induce motion sickness, but in those instances all participants took the VR headset off or closed their eyes until tracking was reestablished. Once this was understood, it felt like I could have reframed the initial research question as a question of purely understanding “What causes motion sickness in VR users?” Either way the recommendation that ended up coming from the study was to add a category on Steam that provided a rating for how easily the content could induce motion sickness to the general population, something that the Oculus store was already doing.
Participants
30 employees 50/50 gender split Ethnic diversity Mix of previous VR experience (None, Some, A lot)Method
60 minute session, participants play for 15 minutes and then answer survey questions. They repeat this process three times total for three categories of VR content (roomscale movement, Fast paced movement, Slow and detailed 360 degree movements). At the end they would answer questions on a closing survey.
Each post-task survey included the Simulator Sickness Questionnaire (SSQ) to measure the level of sickness someone is feeling. As well as Likert measures of comfort on a 7-point scale.
Results
Work In Progress
Rear Cradle Adapter
Discovering Additional Insights and Changing Priorities
While conducting the study on tracking errors and motion sickness, I identified an issue that I saw as incredibly important.
A number of women were having trouble getting a secure fit on the headset.
Identify the Problem
I went and found the employees that had participated in the study that were having an issue achieving a secure fit of the headset and asked if I could take some measurements of their head. Once I had these measurements I could figure out where they fell in the distribution of human size and whether they were part of our supported target market.
Sure enough they came in at 40th percentile female head circumference, which is well within our purported target size of 5th to 95th percentile.
Armed with this information, I went to the design team and informed them of a huge oversight that would lead to a significant proportion of the population not achieving a secure fit with the headset. Initially they didn’t believe me, so I invited them to watch a few of the employees that had problems with fit attempt to put on the device. Immediately they understood there was a problem and that we had to make changes.
Unfortunately, the production timeline would be impacted severely if we stopped ship to change the adjustment mechanism to allow the device to support a smaller headsize, so we had to think out of the box. I first suggested making the foam a bit thicker inside of the head wrap
Participants
30 employees 50/50 gender split Ethnic diversity Mix of previous VR experience (None, Some, A lot)Method
60 minute session, participants play for 15 minutes and then answer survey questions. They repeat this process three times total for three categories of VR content (roomscale movement, Fast paced movement, Slow and detailed 360 degree movements). At the end they would answer questions on a closing survey.
Each post-task survey included the Simulator Sickness Questionnaire (SSQ) to measure the level of sickness someone is feeling. As well as Likert measures of comfort on a 7-point scale.
Results
Work In Progress