The popularity of 3-D movies has brought attention to the Vision Science Program at UC Berkeley, led by Marty Banks. A majority of research in the affects of stereoscopic imagery on vision has its roots with this program, including the work of Simon Watt of University of Bangor in Wales (reported in the December 2010 mkpeReport). This month a paper was published by Banks and fellow researchers titled “The Zone of Comfort: Predicting Visual Discomfort with Stereo Displays.”
Any publication that addresses discomfort in 3-D gets the attention of the press, and this paper got its share of misquotes, several pointing to the study as proof that 3-D causes eyestrain. Few writers, however, appear to have actually read beyond the abstract to understand what the study was about. While the study’s title would indicate a focus on the display itself, the paper provides some useful guidelines for creating stereo content. Its conclusions are not all intuitive, and incrementally further the understanding of the limits of stereoscopic display.
There are many factors to consider when investigating discomfort, including eyewear, crosstalk between left and right images (ghosting), misalignment of the left/right images, and perceptual distortions that may be introduced by the display and/or viewer position. But in much of his published research in recent years, Banks has focused on discomfort caused by vergence and accommodation. Vergence refers to the rotation assumed by each eye as one views an object. One can imagine that an object too close to the eyes will cause the eyes to cross, an extreme example of vergence. Accommodation refers to the focus required to bring an object sharply into vision. Both efforts require muscular movements of the eyes, separate but connected.
Banks points out that stereoscopic displays require an abnormal response for vergence and accommodation, in that the display surface remains at a fixed distance, while the object is perceived a different distance. Accommodation remains fixed, while vergence changes, an unnatural combination. It was demonstrated in 2008 by Banks and David Hoffman, a former student of Bank’s, that vergence and accommodation are a cause of fatigue and discomfort in stereoscopic viewing. This recent study, in which Banks and Hoffman comprise two of the four authors, was an attempt to measure the degree of discomfort with different vergence and accommodation factors.
To conduct such measurements in laboratory conditions requires a special means of viewing objects in which the natural vergence and accommodation response is induced, as well as the unnatural response caused by stereoscopic displays. A clever apparatus was created in which the depth of an object could be electrically controlled for each eye, in synch with the displayed image, whether stereoscopic or not. It is the sophistication of the apparatus that gives these experiments validity.
Three experiments were conducted. The first focused on the effect of stereoscopic distance. In this experiment, the viewer would see objects placed at a certain distance. But the viewer would not know if the perception of distance was naturally conveyed, meaning solely by the optics, or if it was stereoscopically induced. The second experiment presented the viewer with objects placed an equal distance in front of the focal plane (in cinema, the focal plane is the screen), or behind it, to learn relative discomfort. The third experiment measured phoria and the zone of clear binocular vision, tests well known in the clinical assessment of vision. The goal was to learn if there is correlation between the clinical test results and vergence-accommodation-induced discomfort.
The first experiment revealed that vergence-accommodation conflict causes somewhat more discomfort when objects are farther away. The second experiment revealed that images perceived behind the focal plane, or screen, will cause more discomfort through vergence-accommodation than images perceived in front of the screen when viewed at a distance. When viewed closer up, the opposite is true. Images appearing in front of the screen cause more discomfort than images that appear behind the screen. The third experiment revealed that the clinical measurements under study had a high degree of correlation to the discomfort data produced in the first and second experiments.
The rags will tell you that this proves that stereoscopic 3-D causes eyestrain. But in truth, Banks and Hoffman demonstrated that in 2008. But the 2008 study didn’t attempt to discover a comfort zone, and that is what Banks set out to learn in this new study. The report concluded with several guidelines. The first is that content producers have the ability to minimize viewing discomfort through a better understanding of vergence-accommodation conflicts. The second is that the discomfort of a viewer is position-dependent. Near imagery is more comfortable than far imagery at long viewing distance, while far imagery is more comfortable than near imagery at short viewing distance. The third is that there appears to be a relationship between certain clinical evaluations of the eye and vergence-accommodation discomfort. If further studies take this tip, then it may be possible to clinically measure one’s ability to comfortably view stereoscopic 3-D content. More pertinent, it may be possible to develop an understanding of a population’s ability to view stereoscopic 3-D.
Further studies could spring from this work. It would be useful to learn learned about the length of time it takes to develop discomfort as a result of vergence-accommodation, for example. But it’s likely that this study only scratches the surface of our understanding of stereoscopic viewing. Vergence-accommodation discomfort assumes perfect stereoscopic images. There are many other mechanisms for discomfort related to imperfect images or displays. The value of such studies, ultimately, is to learn where the limit of comfort lies. A more cynical view would be to say that they teach us how to a better job of cutting corners. But perhaps best of all, a solid base of scientific studies on stereoscopic discomfort will educate studio executives that good 3-D doesn’t come by running one’s movie through a black box. Perhaps, just a thought, it could lead to fewer 3-D releases, and much better ones. That would make 3-D special, once again.