P3D Re: My 3D Litmus test

["Greg Wageman" <gjw@xxxxxxxxxx>]

From: Paul Talbot <ptww@xxxxxxxxx>


>Some subject matters are not well suited to applying this litmus
>test.  Images with "continuous depth", like a large field of
>smoothly cut grass, train tracks, or long corridors make it
>very difficult or impossible to perceive the airspace under
>normal viewing circumstances.  That has a lot to do with why
>I find stereo images with continuous depth of this sort quite
>boring.  Why not just shoot them in 2D?  The non-stereoscopic
>depth cues are so strong that stereoscopy adds little information.


This is ironic for a couple of reasons.  First of all, I've always found
Paul to be a person who is not quick to leap to snap judgements.  Paul
has always appeared to me as a very objective and acute observer who has
been a great moderating influence on this list.  The second reason is
that I've recently concluded that the most satisfactory stereo images
I've taken are the ones that exhibit a continuity of depth between the
foreground and the background, precisely the sort of images which Paul
pronounced 'boring' in the above quote.

In contrast, I've found that images that are "planar", and which would
tend to exhibit the "airspace" Paul finds prefererable, tend also to
exhibit "cardboarding".  I find that a smooth depth transition is
preferable.  So having Paul question the value of these images is
frustrating, at the least.  That statement is not meant in any way to
insult Paul, but to simply express my consternation with his preference,
if you can understand the difference, which is subtle.

>I'm curious as to whether those who have seen the ChromaDepth
>effect and the Dunkley effect can perceive the air space between
>objects in the images.


I have not personally experienced the so-called "Dunkley effect", but I
am familiar with ChromaDepth, thanks to Dan Shelley's "3D CD" and the
provided Chromatek materials.  There is no question in my mind that
ChromaDepth demonstrates a "true" binocular stereo effect.  One
anecdotal piece of evidence in its favor is the fact that it shows
'stretch'.  As you increase the distance between yourself and the
subject, the depth dimension increases.  This is completely consistent
with "conventional" stereopairs.  I would not expect this to occur with
some kind of "optical illusion" depth effect manufactured in the brain.
However I have no way to prove this theorem.

     -Greg W. (gjw@xxxxxxxxxx)



------------------------------