Supervisor: Simon Watt
In real world stereopsis, accommodation and vergence movements of the eyes operate together. However, simulated depth in 3D cinema/media, requires a ‘decoupling’ of these mechanisms; accommodation is held constant (i.e. fixed at the screen) whilst the simulated distance of objects (in front of and behind the screen) induces differing vergence movements. This conflict in accommodation and vergence demand has been associated with visual discomfort and poor stereoscopic performance whilst viewing 3D media.
The current research project investigated accommodation-vergence conflict (AVC) and the limits of stereopsis. Six optometric measures were also taken to assess their predictive value for stereo-performance: Visual acuity, AC/A ratio, CA/C ratio, phoria, accommodation response, and motor fusion.
AVCs were presented using a stereoscopic multi-plane display that contained two horizontal display monitors that projected stimuli onto one of three image planes (these were either beam-splitters or, in the case of the farthest plane, a reflective mirror). The image planes were positioned 45° from horizontal, and aligned from near to far. The beam splitters were sufficiently transparent to ensure that stimulus quality (i.e. luminance, contrast, etc.) did not vary with plane presentation.
Stimuli were presented haploscopically, and consisted of a sinusoidal corrugation-in-depth embedded within a pair of random-dot stereograms. In each trial, the observer responded to the directional slant of the vertical corrugation that deviated by 10° either to the left or right. AVC conflict was created by presenting the stereogram pair at differing distances to each other. Along with AVC, the spatial frequency of the corrugations varied, so that successful performance of a given trial indicated the ability to resolve that spatial frequency, at that given conflict (i.e. the stereoscopic threshold at that conflict).
The stereo-display was used for four of the six optometric measurements. The haploscopic design allowed for the vergence loop to be opened so that phoria, and AC/A measurements could be taken. The display was such that both screens could be independently rotated on separate axes, so as to converge/diverge the monitors by a specified ° value. This allowed motor fusion to be measured. CA/C was measured using a difference of Gaussian stimulus that could not be accommodated to, thereby rendering accommodation open-loop. Accommodation for both CA/C and AC/A was measured using a Shin-Nippon SRW-5000 autorefractor.
Visual acuity was measured using a Landolt-C task that was run on Freiburg Visual Acuity Test software. This was presented on a moveable monitor that could be positioned between 0-5m from the observer. This software allows a high degree of measurement accuracy as the stimulus dimensions are specified by the screen size and resolution of the display monitor. Accommodation response was measured with a Maltese-cross stimulus positioned at 6 distances between 3.1D – 0D. Accommodation was measured with the Shin-Nippon SRW-5000 autorefractor.
Placement
Over the course of the internship, I was involved in a variety of tasks that have developed my skills and expectations for conducting vision science research. There were 3 main useful learning areas:
Practical
Theoretical
MATLAB Programming