Improvement Suggestion
Posted: Thu Feb 02, 2017 9:29 pm
I've been trying to setup my trackIR for about 3 hours now, and I cannot seem to get the input just right. It's the relationship between my physical head and virtual representation of my physical head that seems to be the trouble. I yaw my physical head to the left, and the TrackIR software representation of my physical head, represented by the "raw" data, pitches forward and translates along the z axis. It's also not repeatable, as in occasionally, the compound movement will change, despite making nearly the same yaw motion with my physical head. Sometimes changing the camera rotation in the software to 180° helps...
I'm sure it's due to the physical location and orientation of my clip pro. I will probably find the best location eventually.
My suggestion is to add a calibration procedure. Having the user move their head through a prescribed motion should allow calibration of the transfer function between LED relative and absolution orientation and the virtual plane that represents the raw head. Having a fixed transfer function, which appears to be the current implementation, between LED orientation and virtual plane forces the user to find exactly the right physical relationship between the camera and IR LED's, BOTH of which have 6DOF on their own.
I'm sure it's due to the physical location and orientation of my clip pro. I will probably find the best location eventually.
My suggestion is to add a calibration procedure. Having the user move their head through a prescribed motion should allow calibration of the transfer function between LED relative and absolution orientation and the virtual plane that represents the raw head. Having a fixed transfer function, which appears to be the current implementation, between LED orientation and virtual plane forces the user to find exactly the right physical relationship between the camera and IR LED's, BOTH of which have 6DOF on their own.