Stereoscopic 3-D display technologies aim to provide a compelling, realistic sense of 3-D depth. Theoretically, this could be achieved by having a system that can project the same light rays coming from the real world via a display. Recent advances in rendering, display, and image acquisition have made it possible to present almost all depth cues accurately. Focus cues, including the eye’s focusing response to image blur (accommodation) and the pattern of retinal blur arising from objects at varying distances (retinal blur gradient), remain challenging to replicate accurately, however. In conventional stereoscopic displays, focus cues are
incorrect, as the depth is specified by the screen surface and not by the position of the content in the scene. This issue causes two problems: firstly, the unnatural pattern of oculomotor responses elicited when viewing the 3-D scene (inaccurate accommodation signals), and the appearance (inaccurate retinal blur gradient) may signal the unnaturalness of depth in the scene (Hibbard et al., 2017). Secondly, the inconsistency between depth specified by vergence from disparity and focus cues results in a conflict commonly known as vergence-accommodation conflict. The visual system’s effort to resolve this conflict causes visual discomfort and reduced stereoacuity (Watt et al., 2005 a; Hoffman et al., 2008; Shibata et al., 2011). There is an increasing demand for stereoscopic 3-D displays that can elicit realistic depth and have minimal to no user issues. Several technologies have tried to address the issue with incorrect focus cues, but all are computationally intensive, require specialist hardware, and compromise image quality. Hence, the value of presenting focus cues accurately needs to be fully understood. In Chapter 2, we measured the vergence distances (zone of good stereo depth perception) around the screen, where stereoscopic depth perception remains effective despite incorrect focus cues. To understand whether all users would benefit similarly from correct focus cues, we specifically examined whether age and age-related changes in the ability to accommodate predicted individuals’ tolerance to vergence -accommodation conflict. We used a fixed viewpoint volumetric display (specifically Multiple focal plane display, Mackenzie et al., 2010) to present stimulus with varying vergence-accommodation conflict to see at what conflict value individuals’ stereoacuity drop to a criterion level. We did not find any predictive effect of age or people’s ability to accommodate on their tolerance to vergence-accommodation-induced degradation in depth perception. However, we found other visual factors, such as phroia, to predict the symmetry of the centre of the zone (similar to Shibata et al., 2011). And guidelines for stereo 3-D content optimisation. In Chapter 3, we explored whether depth appears more realistic when focus cues are correct. We used a fixed-viewpoint volumetric display (specifically Multiple-focal plane display, Mackenzie et al., 2010) to present stimulus with and without natural variations in focus cues. Participants made a two-interval, forced choice response indicating which interval contained: Experiment:1, the largest depth separation or Experiment:2 which interval contained the most tangible, solid and real depth separation. The results provide strong evidence that focus cues are important for presenting scenes with realistic depth, and, that individuals can separate the quantitative and qualitative aspects of stereoscopic depth. In Chapter 4, we evaluated whether the effect of focus cues on realism is detectable in high-fidelity images and if depth-of-field (DoF) rendering can be used to substitute any effects of focus cues (using a similar method in Chapter 3). The results show that focus cues do increase depth realism in a 3-D scene for high-fidelity images, and the effects of focus cues cannot be substituted with DoF rendering techniques. Our study has shown that we can optimise certain aspects of 3-D content to suit an individual's tolerance to vergence accommodation conflict by utilising visual factors, such as their phoria. However, the technical constrains are difficult to achieve if we are to get focus cues correct. So, depending on the application's requirements, a decision must be made between stimuli with tolerable conflict and realistic depth.