Stereo Scans

Yep, that's right--stereoscopic image-pairs are easy to produce on a flatbed scanner.

I wish I had thought of this myself--but, like almost everyone else who uses flatbed scanners, I just didn't think enough about they work. Richard Schubert did.

Yesterday (8 October 1998) Benton.Holzwarth sent an email message in which he included a link to a web page in Germany,  http://www.stereoscopicscanning.de , set up by Richard Schubert--and clicking on that link really got me fired up.  Try it and see for yourself.

Here's how it works:  The scanner has a fluorescent tube that produces what we think of as a "bar" of light extending from one side of the scanner to the other as the scan head tracks the length of the glass.  I thinik that's what fools us into thinking that the scanner's sensor extends all the way across the glass, like the light.  It doesn't.  It's fed by a small lens in the center of the scan head.  What that means--and what we don't think about--is that there's parallax in scanned images.  We don't ordinarily notice it because we scan flat things.  But scan a 3-D object, and that parallax suddenly becomes apparent.

What Richard Schubert realized is that if you scan an object twice, once with it placed to the right of the lens and again with it to the left, you get two different images--because of the parallax.

When you look at one of the scanned images with the right eye and the other with the left eye, you see the object in three dimensions--stereoscopically!

Of course I had to try it right away.  Some of my experiments are below.  (If you haven't tried looking at stereograms before, you might want to read the Viewing Tips first.)

Peruvian Creche
Butterfly Paperclip
Earth Marble
P-38 Lapel Pin
Honest Abe
 NOTE:  These links open in a new window.  If you leave it open, then the other links will open in it, too.  You can use this page as a menu.

You can perceive the stereoscopic effect with no special equipment either by crossing your eyes or by letting them diverge, to go wall-eyed.  Which way you do it depends on which scanned image is on which side. In most of my experiments here, there are three images--but they're really two image-pairs, one for the walleye method and the other for the crosseye method.  The center image is in both pairs, and the two side images are identical.

It's a little tricky at first to get the hang of perceiving steroeoscopically.  For the "walleye" method, you can use two rolled up sheets of paper (or toiletpaper tubes) to guide each eye to see only one of the images.  Then sort of gaze at them without trying to focus sharply, until they merge into one image.  Then you should begin to perceive it in three dimensions.  Without the tubes, you can do the same thing.  Let your eyes relax as if you were looking at something beyond the screen, until you start to see three images.  Then try to sort of pay attention to the middle one, without really focusing on it.  After a little while it will sort of lock in, and you'll suddenly know that it has depth.

For the crosseye method, you can hold a fingertip (or a pencil or the like) about halfway between your eyes and the image pair.  Focus on the fingertip, but be aware of the images behind it.  After a while, you'll become aware that there's a third apparent image between the real ones, and eventually you should realize that it has depth.

Important:  To see the stereo effect, you must be able to see clearly with both eyes.  The images must be in focus with each eye.  You may need to wear reading glasses.

Back to Scanning Experiments page

My home page

Send me email

(Please type it into the "To:" box in an email message.)

Don Maxwell