P3D Re: Reply to Larry's critique of the red button

[Larry Berlin <lberlin@xxxxxxxxx>]

>Date: Tue, 6 Jan 1998 
>From: "Dr. George A. Themelis" writes:
>
>Sorry folks but I am getting a bit tired listening Larry Berlin 
>criticize my beloved Stereo Realist red button viewer...
>
>>What was worse
>>was that being able to see as far as those edges, allowed me to sense the
>>distortion created by the optics themselves. The entire view in a RED button
>>is presented as if it's painted on the inside of a bowl. 
>
>Sorry, I don't get it.  Are you talking about pincushion or barrel 
>distortion (both two dimensional) or something else?  

*****  What's hard to understand about looking at the inside of a bowl? It's
curved in 3 dimensions with all the edges closer to you than the center. All
stereo image data of a stereo image projected to such a space is placed
relative to this bowl instead of relative to a flat plane which would
correspond to the original image taking situation. Therefore this distortion
messes with and disrupts that *favorite requirement* known as ORTHO. 

>If you view a 
>perfectly flat picture, does it look distorted in two dimensions or 
>three dimensions?  Your "bowl" example indicates a distortion in the 
>3-dimensional space.  There is no question that there is a bit (just a 
>slight, hardly noticed) pincushion distortion in these lenses.  But if 
>the distortion is the *same* for both lenses then it should only by 
>seen in 2D and not 3D space.  

*****  Go try it for yourself. A flat picture of full frame dimensions
presented in a Red Button will curve towards you in 3D space on all edges
proportionately. I haven't examined the 2D distortions in each side, only
examined the effect of stereoscopically flat surfaces (two images of the
surface with zero parallax) as they are seen when using both eyes looking
into a Red button viewer in a normal fashion.

I haven't done this with many red buttons, but it was very distinct in the
red button I had access to when I had that image in hand, and no reason to
suppose it was built differently than others. (but possible I suppose)

>
>It is mainly by observing the FRAME (the perfectly straight and square 
>frame of an RBT or aluminum mount) that this distortion can be noticed.  
>This could not be possibly observed in a three-dimensional image!!!

****  When I view a stereo image I get an overall sense of relative depths,
and an absolute sense of depth placements of each element in the image
relative to other elements in the image, the normal up/down axis, and
including but not requiring, notice of the edges of the frame itself. I do
this by eye-scanning of details in the image just like I would scan the real
scene. If you view a stereo image of a person's face that occupies most of
the frame, this distortion will tend to flatten the apparent view because it
will cause the edges of the face to appear stereoscopically, as triangulated
depth wise, closer to you than they would in real life. Not completely flat,
just a tendency to flatten.  In real life the nose and central features are
closer to you than the ears, cheeks and the edges of what can be seen of the
head and hair. Map that to a bowl shaped field and it flattens the depth
features by bringing the edges forward by some amount proportionate to
distance from the actual image center.

>
>> With full frames,
>>you can definitely see it and trace the entire bowl surface with your eyes.
>
>Exactly how many times have you used a red button viewer?  How many 
>different viewers?  (YES, there are differences from viewers to viewers 
>as the quality of the lenses is not the same in all of them and possibly 
>the supplier has been changed a couple of times - also, I have seen 
>viewers with lenses reversed - talking about distortion!!!) 

****  OK, maybe it doesn't exist in all red buttons. I suspect it may exist
in most, and out of familiarity with the device it is ignored, therefore
unnoticed. I never said I don't like red buttons, only that I observed this
definite distortion, which I had *not* expected to find, based on hearing
lots of praise for the accuracy of the Red button from you and others. 

>
>Who modified it/them for full-frame?  I have worked over a hundred red 
>button and other viewers.  The magnitude of this distortion is minimal 
>and you are blowing it out of proportion.

****  I'm stating plain facts, and not exagerating any of it. I have no
reason to think that this viewer was modified in any way. If you think it's
something you don't like, I can't help that. If it doesn't bother you, fine.
If I get a chance to buy a Red button eventually I will still do so
regardless of this distortion. If I do, I likely will want one of your
modified versions. They are a great device even with this distortion. That
doesn't make the distortion magically disappear.

>...............
>They are "fun"?  This "don't prefer" RED button viewers sounds like 
>the result of some experimentation and testing.  Exactly how many 
>different STEREO SLIDE viewers have you tested and how?  Do you even 
>shoot slides?  Have you ever shot a roll of slide film???

****  Come on George, I'm not attacking you or even the Red button viewer,
just reporting on a particular instance where I had the right kind of slide
pair available and a RED button viewer, probably a Themelized one, though I
have to check that detail before you start hopping around the room! ;-)  In
my life I've shot lots and lots of slide film. In recent years I've used far
less slide film due to the difficulty in using slides. I recently bought a
couple small projectors, but have no real reason (or space) to use them just
yet. Even when I do have such reason it will be the exception rather than
the everyday situation. I spend hours and hours conveniently working on
stereo images without having to have a dark room, or projection space, or
wearing eye devices other than occasional use of LCS and anaglyphic glasses.
Once digitized, I'm free to fix problems, change alignments, and a host of
other conveniences not even remotely possible with slide chips alone.

I use prints because they are easier to scan for the computer and they are
easier to freeview as prints in their original format. It's kind of like
prints combine the slide and projector and screen all into one piece of
paper at the same time they remove the need for any kind of special glasses.
Pretty convenient when available space is at a premium!

>
>>Take particular notice of the triangulation/depth effect.
>
>What exactly is the "triangulation/depth effect" that you ask us
>to notice?

****  This has been discussed on P3D on many occasions and in various
topics. If you look at a specific image detail and notice that that detail
is closer to or farther from you relative to another image detail and by how
much, that is triangulation, as opposed to taking in the whole image without
notice of specifics. Since the rule of thumb for stereo is for a maximized
DOF allowing the viewer to look anywhere in the image, this triangulation
thing is common to most of us, whether or not we take special notice of the
activity.

I am constantly aware of relative depth factors in any stereo image that I
view. I notice when projected slides are rotated relative to each other
causing distant mountain scenes to lean dramatically forwards towards the
viewers as if about to fall on you. In such situations everything in the
image distorts along similar top-forwards directions. There always seems to
be a few of these in most of the projection events I've witnessed including
several examples at NSA in Bellevue. I didn't take notes on which
image/event, just that it did happen. In one case that was particularly bad
I happened to be standing behind the projectionist and asked about the
distortion. It was the only one in that particular series of slides with
rotation. The projectionist seemed unaware of it, but I could prove it was
there by a variety of means if pressed to do so. (the middle of a projection
event isn't the right time) It may have been 2x2 slides with one side not
seated properly or a slide pair with a rotated film chip.

I notice lots of things in stereo images that are due to interpretation of
relative parallax. (in stereo that's the predominant factor!) A lot of
discussion about stereo images deals with similar factors so I know I'm not
the only one who uses triangulation as a routine part of image
interpretation. What I'm pointing out is that you can use this common
facility to identify and hopefully fix problems in slide pairs, or evaluate
the effective accuracy in viewers.

>..............
>Now, give me a break!!!  You prefer a cardboard viewer with plastic 
>simple lenses to the red button?   A $10 cardboard viewer with plastic 
>lenses that you have to squeeze to focus (and work hard to bring both 
>lenses in focus and parallel to your image) to an all bakelite viewer 
>with metal rack and pinion focusing, interocular adjustment, internal
>illumination and achromatic lenses?  

****  It's lighter, less bulky, easier to carry around, pass around or use.
It's less expensive by an order of magnitude, and it doesn't seem to have
the bowl distortion as judged by the same slides used in examining a Red
button... When I'm rich and wealthy I'll gladly buy a RED button just to
have one, and enjoy it despite any distortions. Or maybe I'll wait till
someone creates a super-duper non distorting (color or dimension) viewer and
pay a lot more for it... Maybe it's on the way to market now? It just seems
to me that non-distortion of depth factors should be paramount in stereo
devices which are intended to show depth factors.

>
>And you don't see any distortion in the plastic lenses of this $10 
>viewer? 

****  NO

> How about chromatic aberration?  

****  NO

>Do you see any of that?  
>Don't you see the blue-colored edges?  Doesn't that bother you?  You 
>have been hypersensitized to some minimal distortion and close your 
>eyes to loads of false color?

****  As to loads of false color, NO. I've seen plastic lenses that DO have
lots of false color and the RBT viewer is not one of those kinds of devices.
As to preferring a tiny amount of false color to obvious depth distortions, YES.

>
>Better slow down in this red button critique Larry and concentrate 
>in something that you really know like computers, stereograms, etc.
>

What I know intimately is the stereoscopic phenomenon itself, not just
computers or stereograms. It has allowed me to cut through the BS in
computers and make definitive comparisons of methods, images and devices
with a high degree of accuracy based on actual stereoscopic principles. As
experience over time includes more stereo devices I have more comparisons to
make. Anyone can make the same comparisons. Most don't try. Most ignore a
host of existing problems since it's more practical to have a device that
works than to have a perfect device. Actual quality in any device should be
measured relative to solid principles, rather than the often misleading and
subjective marketing disclosures.

As to computers, do you want an analysis of the effectiveness of today's
computer? It's pretty radical compared to evaluation of stereo devices...

I will get more information about the specific RED button that I used. As I
don't own it, I can't just reach over to the shelf and examine it again. The
owner of this particular device was not previously aware of this distortion,
yet it hit me right away with the right sample slide. Maybe others can
assist the process and learn to evaluate such things for themselves and
report back. That way there will be many more samples, and many more pairs
of eyes/minds doing the evaluating. I'm willing for only 10% of RED button
viewers to be affected by this. I'm equally confident that I can
definitively help you see the distortion I observed in the exact same viewer
that I used, should such an opportunity ever arise. This is science not
subjective feelings. We just need a larger sample base.

Larry Berlin

Email: lberlin@xxxxxxxxx
http://www.sonic.net/~lberlin/
http://3dzine.simplenet.com/


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