Photo showing view of google cardboard seeing self in third person.
Trying to get out of our physical bodies to experience the world differently
What it does
Undeterred in the Third enables you to see yourself in the third person using a separate camera feed.
How we built it
We set up a fisheye camera on one phone to capture ~160º FOV, serving an rtsp stream. An app then runs on the phone in the Cardboard headset to stream the video and map it to a sphere in VR.
Challenges we ran into
Mapping between the fisheye projection and coordinates on the surface of a sphere: This took a few tries, googling and glsl black magic.
Latency: By default, the android media api does not allow setting the network buffer size. This meant that while the video was smooth, there was ~7s of latency. By using LibVLC instead, we have choppier video but with only ~0.5-2s latency, depending on network conditions. We used a portable wifi hotspot to try to improve the network speed as well.
Accomplishments that we're proud of
Getting it to work
What we learned
Effectively collaborate on generating ideas + lots of GLSL
What's next for Undeterred in the Third
We wonder: if the latency could be reduced even further, could we use it to study phantom pain?