I am part of a team using an HTC Vive Pro Eye to obtain medical data. Because of our use cases, we do not calibrate the eye tracker, but this has not caused significant issues with calculating gaze fixations.
However we have two problems with the eye tracker.
1) We are not obtaining some data we should be from the tracker (pupil size), and
2) the data we are getting has a hard max. Each subject begins with an eye openness of 1 and has periodic deviations from the maxima (interpreted as blinks). However the amount of deviation varies widely between subjects (almost never showing fully closed eyes), in some cases never falling below 80% open. Two example graphs showing extremes of behavior are shown below.
