[photo-3d] Re: The 3-DVG effect

[KenDunkley@xxxxxxx]

Hello gentlemen: 

I would like to inform those members of the list who may not be aware of it 
that I am not the first researcher to describe the creation of a true 
stereoscopic image from a single ordinary two-dimensional picture. What I do 
claim is to have invented a much simpler method to achieve such results. 

>From Koenderink, "On so-called paradoxical monocular stereoscopy," 
Perception, 1994, volume 23, pages 583-594

"A few years after the introduction of the Verant, Carl Zeiss also patented 
devices for the binocular viewing of single pictures (von Rohr 1907; Carl 
Zeiss 1907). These devices are reputed to yield highly increased pictorial 
depth, though obviously based on the removal rather than the exploitation of 
binocular stereo information." …"The most perfect Zeiss design is the 
synopter (see figure 1), though eventually a cheaper and rather inferior 
design was actually marketed and apparently sold for the benefit for visitors 
to art galleries."…"Photographs viewed with the synopter are seen in vivid 
'stereoscopic' depth. To many observers the effect is even 'better' than that 
obtained with regular stereoscopy, because the annoying [Oops, fax smeared 
word] effect that so often spoils stereoscopic depth is absent."    (Sorry 
about the fax glitch.)

I will present a short list of other authors reporting stereoscopy from 2-D 
pictures in a subsequent mailing.

One important point that Koenderink makes, and one that I agree with, is that 
the mechanism underlying the generation of stereopsis from single pictures is 
simply not fully understood. My guess has been that persons reporting to this 
list  who profess an understanding of the effect have, more than likely, 
observed something else and assumed that what they saw was it. 

I say this because the criteria for detection are quite clear and 
unequivocal, either you see it or you don't. Once you get the hang of it you 
should be able to create perfect looking stereoscopic images from virtually 
any magazine picture from a distance of six feet, regardless of the 
distribution of colors in the picture, using only your fingers.  If you can't 
do this  you haven't yet got the hang of it. (Oh, and yes, there are some 
pictures  that display best with a nasal disparity versus a temporal)

The detection criteria are straightforward. You have detected the 3-DVG 
effect when you can answer "yes," unequivocally, to all three of the 
following questions: 

1. Is it giving you the clearest and sharpest picture you have ever seen bar 
none? Switch back and forth between the device and your normal view of the 
observed picture to ascertain this. (This is cause by the well known 
squinting effect)

2. Can you look all the way into a good scenic picture (like, out to 
infinity)? If you have trouble here, increase your viewing distance by three 
feet and try again. Also, try looking at smaller magazine photos. 

3. Does the picture appear to be truly three-dimensional (i.e., 
stereoscopic)? If you examine a scenic picture containing trees, foliage, or 
flowers you should be able to distinguish the apparent difference in visual 
depths of small individual elements of the picture such as leaves on a tree 
or flowers in a garden. 

It should appear unequivocally stereoscopic though not identical to images 
from normal stereo pairs. Direct comparison can easily be made by "free" 
viewing a color stereoscopic pair versus the 3-DVG viewing of one of the dual 
images. 

As I describe on my site, the 3-DVG effect is based on an optical feed back 
loop. You do not observe a static image like looking through a telescope or a 
microscope. When you finally detect it, it will feel like you are creating 
the 3-D image you see. I hope some of you take 15 min of you time and try it 
yourself. It really is not difficult to detect.

Best regards,

Ken Dunkley

http://www.3-dvg.com                     Make any picture 3-D using only your 
fingers.

http://www.3-dvg.com/thoughts         Arguably, the most stunning 3-D image 
ever created