Stereo Disparity
Stereoscopic enjoyment is all about seeing and exploring the depth factors. Enjoyment is all about having the right relationships in place without going to a stressful extreme. This highlights one of the most difficult challenges in stereoscopic 3D because there are so many variables to account for.
In stereo photography the goal is to provide an image pair that when viewed appropriately, recreates for your senses the same visual geometry of the original scene. Lots of discussion can erupt over topics of whether or not it is OK or good to use a wider baseline between the camera lenses in what is termed hyperstereo. Hyperstereo can be very useful especially if you want to observe shape details of a distant scene which normally would appear not to have depth.
The medium of Stereoscopic imagery is actually a very flexible, stretchable medium. That is good in that it allows the artist/photographer plenty of lattitude in composing and arranging an image or scene. It can be bad in that a novice can easily be unaware of some effects which have the potential to affect people in an adverse way when they view the results.
Abstract art such as Fractal Flames has the advantage of not containing familiar and recognizable features. Thus when dimensions become stretched or distorted, who is to say it didn’t belong that way to begin with?! If the subject were a human or animal face, any distortion can be easily noticed and most often the impression would be negative.
More particularly relating to this website and stereoscopic issues, there is a huge difference between freeviewing or cross-viewing images on a computer monitor and attempting to project those same images onto a theater sized screen in stereo! The easy explanation is that cross-viewed images on the computer screen are far more compressed, which allows a much larger overall stereo disparity to be present before viewing stresses become too much. By comparison, when stereo images are projected to a large screen, the viewing geometry is stretched so much that a much smaller range of disparity is necessary. The guidelines for the NSA presentation call for a maximum of 3% disparity. The percentage is found by dividing the image width by the total measured stereo disparity between near and far points.
All of the images used for the NSA show had to be rendered especially for that viewing environment. For one thing they needed to be a particular shape and size for best performance, but also their parameters had to be adjusted so that the finished stereo pairs had no more than 3% disparity. Some of them as they originally existed had been rendered with 6% or more disparity and were quite comfortable to view on a computer screen.
Below is an illustration to help visualize the situation. The basic fractal image has been resized for website viewing and is actually 500 x 382 pixels. The ideal 3% then is 15 pixels as a total stereo disparity. When I measured the image I found that the foreground, or near point projected in front of the stereo window formed by the two sides and the plane of zero disparity. The total disparity from that near point to the most distant points turned out to be 4.4%.

Stereo Disparity Extremes
In this image various levels of disparity have been marked by the numbers. You may want to back up a ways from your computer monitor before trying to visually fuse the more extreme values represented by the range from 12% to 25%. I can personally fuse all the values, but when I view 25%, much of the rest of the image blurs out of range. For some that may cause eye strain so don’t try too hard! For others you simply may not be able to fuse the more extreme levels at all.
Projected stereo images with disparity well over 3% become for that audience what levels above 12% does for the freeviewing example.
It would be well to limit one’s creative expression to the more conservative and viewable levels of 3% to 8% when freeviewing on a computer screen. Then be aware that not all viewing circumstances will provide the same experience and no two pair of eyes are the same either!
This entry was posted on Saturday, July 11th, 2009 at 2:27 am and is filed under Stereo Disparity Percentage. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.
Related Posts
- No related posts found.


on December 4th, 2009 at 4:59 am
you just got yourself a place in my bookmarks
on February 3rd, 2011 at 8:28 am
Cool site!