Discussion in 'Article Discussion' started by bit-tech, 20 May 2019.
2.3mm margin of error (heck, even the 1.4mm difference) is HUGE for the type of applications they are aiming this at!
Actual study here
Sounds about right for pointing tasks (nobody has a good technical solution for accommodation that doesn't come with a heaping pile of drawbacks yet).
However, the study did not appear to control for tracking offset beyond a single registration marker. Hololens is well known for it's... adequate at best tracking capability, so this result could well have been down more to the dots in the 'join the dots' task being offset between line sequences than an issue with vergence-accomodation mismatch. This really needs testing with either a much higher quality tracking system retrofit to Hololens (requires software written to use an external tracking server), or by a simpler method of physically fixing the HMD during the task and disabling tracking after initial setup & registration.
The lack of registration between the virtual environment (projected dots) and physical environment (drawn lines) is a critical flaw in the study, limiting its applicability to an evaluation of the Hololens 1 system and not applicable to view-through HMDs in general.
Separate names with a comma.