P3D Re: autostereoscopic depth

[aifxtony@xxxxxxx (Tony Alderson)]

>Boris Starosta wrote
>Tony, could you elaborate on this please?  Are you saying that lenticulars
>and these barrier strip prints cannot under any circumstances have
>sharpness very far away from the level of the actual printed surface?  Does
>this also hold true for autostereoscopic video displays?

>Boris Starosta


I don't claim to have any deep understand of the theory and mechanics of
autostereo displays. I have some competence in preparing art for the
process, but I have no special expertise in constructing the prints
themselves--I leave that to my associates Steve Aubrey and Bob Mannle.  So
take the following with the proverbial grain of salt, it's only an artist's
impression...

I have never seen an autostereo with either great depth or extensive depth
of field. I have prepared arrays that  I know are "pinsharp" across the
image, and that have parallax beyond projectibility, depending on which two
chips of the array one selects for the stereo pair. The sample conversions
on my website, for example, exhibit half or less of the parallax from the
most extreme pair of the 21 image array.  The difference between viewing
these as a stereo slide vs. a lenticular print is quite dramatic. The
lenticular print does not show as much depth (you are viewing a pair set
closer together) and the foreground and background always goes fuzzy.  I
have not seen every autostereo ever made, and some are significantly better
than others, but they all seem to exhibit this problem.  As far as I can
tell, from experience and limited research, this is an inherent limitation.

The problem, as best I can figure, is cross-talk between the viewing zones.
Think about what is happening when you view an autostereo. The array is
interlaced, so under each lenticule (or barrier slit) there is a stripe of
substripes from each image of the array. As you look through the lenticule,
each eye sees (hopefully) only one substripe from a different image.
Theoretically, each eye sees the substripes from the same image across the
entire lens. But the light rays reflected from the print emerge in a wedge
shape, so as the distance varies, this may not be the case. (Alignment of
the lens or barrier is critical.) And as the wedge broadens, you start to
see adjacent substripes, which go fuzzy and limit the depth. Note how, if
you rock the lenticular print side-to-side, you sometimes lose the effect,
or it goes pseudo. At those points, your eyes are either between viewing
zones, or out of phase.  Note also how there is an "optimum" viewing
distance for a given autostereo--if one is too close or too far away, one
can't get each eye into a proper position to get the effect.

This is not a condemnation of autostereo. I like a lot of autostereos, and
they certainly have their place. All I am saying is that there are
trade-offs in moving the stereoscope from the viewer's eyes to the print
surface.  A well-made autostereo has a lot of impact,  people are used to
limited depth of field in photography, so the fuzzy BG is often a tolerable
compromise.  I'm glad to see the expanded availability of the process, but
I don't think it is the general answer to the stereographer's prayer.

Tony Alderson
aifxtony@xxxxxxx
www.aifx.com



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