[photo-3d] Re: Chromostereopsis

[Gabriel Jacob <3-d@xxxxxxxxxxxx>]

Earlier today Jim Norman reported some interesting observations he
made viewing Shab Levy's illustrations. I have uploaded one of the
examples I was working on yesterday to the yahoo file section.
Hopefully this will demonstrate the detrimental impact
chromatic aberrations have on stereo pairs.


The following link points to a stereo pair consisting of alternating
red and blue boxes arranged in a 3x3 array.
http://groups.yahoo.com/group/photo-3d/files/3d/chroma.gif
Looking at it with the unaided eye and not fusing the images, one
can easily see  the red boxes seem to be closer than the blue boxes.
This is a stereo depth cue and is due to chromatic aberrations in the
eye refracting the different colors to varying degrees.

Free viewing (parallel) the stereo pair, the red boxes should look
even closer now. This is because I shifted the blue boxes in the right
stereo pair, one pixel to the right. Now the interesting part. View the
image with a stereo viewer. You should see the red boxes either
being almost in the same depth plane as the blue boxes or even the
blue boxes being closer! The degree will depend on the viewer your
using and how far back you can go. If you have prism lenses (with no
convex lens component), you can demonstrate this very well. Those
lucky enough to own a PokeScope should see the effect very easily.


It's interesting to note, you get different results using different colored
boxes. This is due to the different color fringing . In the following
stereo pair,
http://groups.yahoo.com/group/photo-3d/files/3d/chroma2.gif
with the unaided eye and not fusing, the red box will be closer than
the blue box. The two white boxes will be in the same plane.

Free viewing (parallel), the red box will look even closer. The left white
box will look closer compared to the right white box. Again here is the
interesting part, viewing the stereo pairs with a viewer, the red/blue
boxes will now look like they are almost in the same plane (or the
blue box will now look closer!). The white boxes will retain the same
depth planes as before, that is the left white box will look closer.


I was working on another example that uses animated gifs and
changing the colors on the boxes only on the border (to simulate
chromatic aberrations) and I got some very interesting results. When
I get some time, I'll try to put these up too (one day!). Hopefully the
above examples will suffice to demonstrate that  chromatic
aberrations do have an impact on stereo pairs.

Gabriel