In VisionGlass the view is not properly shown "in front of" the user in cases in which the user is moving their head while Wolvic is starting. The reasons are multiple:
we were incorrectly getting identity quaternions sometimes in positions where the glasses orientation was not precisely that
the recenter caused by setting the head lock state was not issued on time, so it was not doing anything in the practice
From now on we don't render any frame until a valid head pose is received and also we do perform a full widget recenter after the first drawn frame.
In VisionGlass the view is not properly shown "in front of" the user in cases in which the user is moving their head while Wolvic is starting. The reasons are multiple:
From now on we don't render any frame until a valid head pose is received and also we do perform a full widget recenter after the first drawn frame.