Purpose: A consistent stereoscopic (stereo) effect in disc photographs is important to be certain that apparent depth changes are real and are not the result of artifact caused by a variable stereo angle. The purpose of this study is to compare the accuracy of relative depth assessments using three different stereo photo viewing methods.
Methods: We evaluated the ability of 12 observers to arrange sets of simultaneous stereo photographs (Nidek 3-Dx split-frame slides, Lentec transparencies) and sequential stereo photographs (Zeiss full-frame slides) according to apparent cup depth. Observer arrangements of each set of 10 photographs were compared to a "reference standard" sequence determined from cup depth measurements made by the Imagenet optic nerve analyzer.
Results: The median number of sequence errors (n = 6) made by all examiners in arranging the Nidek 3-Dx simultaneous stereo slides was significantly lower than the median number of errors (n = 12) with the Zeiss sequential stereo slides.
Conclusions: Simultaneous stereo slides provide significantly more interobserver consistency for judgements of cup depth than sequential stereo slides.