The cost and complexity of capturing stereoscopic 3D video or of converting conventional 2D video to stereoscopic 3D in post-production can be considerable. With this the case, producers of stereoscopic 3D content occasionally use an inexpensive third alternative. A single 2D image captured by a single camera is used as the basis for the stereo pair.
The image presented to the right eye is the single image horizontally offset to the right. The image presented to the left eye is the single image horizontally offset to the left. The effect of introducing such a uniform disparity is to shift the planar 2D image to a position behind the plane of the screen. (The shift has to be behind the plane of the screen to avoid a “window violation.”) This approach can be called “shifted 2D”. Shifted 2D is believed in the industry to create an illusion of depth.
It is also believed that, while shifted 2D may not be as visually striking as stereoscopic 3D, it is still more compelling than conventional 2D. At “face value,” these beliefs are not unreasonable. Now, two researchers, Paul Hands and Jenny Read at the Institute of Neuroscience within Newcastle University (Tyne and Wear, United Kingdom) have conducted a study to determine the extent to which these beliefs are actually correct.
A recent article by the researchers is entitled “True Stereoscopic 3D Cannot be Simulated by Shifting 2D Content off the Screen Plane.” It was published in Displays 48 (2017) 35-40. A copy of the article is can be found here.
Some of the reasons that the use of shifted 2D seem reasonable are that “2D images contain many pictorial cues to three dimensional structure, including perspective, shading, texture cues and apparent size. These can even trigger reflex vergence eye movements, implying that the brain accepts these depth cues at a basic perceptual, rather than simply cognitive, level. 2D video content includes still more powerful depth cues, such as structure from motion.”
The reason that the researchers felt that their study was needed is that there is also evidence that suggests that the use of shifted 2D may be problematical. Apparently, “the visual system detects the shifted 2D image as a flat plane, and that perception is powerfully influenced by this.”
In their investigation, the researchers showed 30 second video clips drawn from a nature documentary to a group of participants. Some of the videos were shown in the original stereoscopic 3D, others were shown as a 2D version while still others were shown in a version that contained both stereoscopic 3D and shifted 2D content.
Although the differences are subtle and very difficult to pick out in the illustration above, the left hand pair of images compose to a stereoscopic 3D image. The right hand pair of images compose to a shifted 2D image.
Participants in the study were asked to evaluate the sense of depth produced by each type of video clip using a 7 point Likert scale. (“A Likert scale is the most widely used approach to scaling responses in survey research. It is a psychometric scale that employs questionnaires. When responding to a ‘Likert item,’ respondents specify their level of agreement or disagreement on a symmetric agree-disagree scale for a series of statements. By this means, the range captures the intensity of the participants’ feelings for a given item.”)
The researchers found an unmistakable difference between the sense of depth produced by stereoscopic 3D (mean 6.03) and the sense of depth produced by shifted 2D depth (mean 4.13). On the other hand, no difference was found between ordinary 2D presented in the plane of the screen and the shifted 2D.
Based on these results, the researchers concluded that “the shifted 2D method not only fails to mimic the depth effect of true stereoscopic 3D, it in fact has no benefit over ordinary 2D in terms of the depth illusion created.” They added the caveat: at least not without the use of an unacceptably large parallax.
In a final comment to reinforce their results, the researchers state that the “shifted 2D technique is not viable as a cheap way of making ‘fake’ stereoscopic 3D. -Arthur Berman
Newcastle University, Paul Hands, [email protected]
