Recent discussions regarding the stereo window have me questioning the
conventional viewer design parameters. Offhand, I can think of four likely
sources of viewing discomfort: image distortion, excessive eyeball
convergence, excessive eyeball divergence and the decoupling of focus from
eyeball convergence. Taking these one at a time:

Image distortion seems to increase as you move your eye off the center of
the lens. Say the average relaxed eye lens spacing is 62.5mm and perhaps
61mm when the eyeballs converge to look at something close up. Why then
are the lenses of my Chinese viewer spaced at 64mm and those of other
viewers as much as 65.5mm? The eyes are forced to view slides through
suboptimum points of the lenses, a situation that only gets worse as they
converge on close-up homologs. For comfort, I think it would be better to
space the viewer lenses at, say, 62mm.

Eyeball divergence can be deemed to be excessive only if the eyeballs have
to roll out to some angle beyond parallel vision. Since the eye muscles
are most relaxed when the eyes are parallel, I would think that spacing
distant homologs at exactly lens spacing would make for the most
comfortable view. The chips for my example viewer would thus have infinity
spacing of 62mm. Why then do all the common viewers have wider lens
spacing than the nominal chip infinity spacing? This forces the eyes to
start out crosseyed when viewing the most distant objects, and strains them
even more when viewing anything else on the slide.

Excessive eyeball convergence is defined by one of the 1/30 rules as
anything more than the arctangent of 1/30, or two degrees. But at a
reading distance of 10 inches each eyeball converges a whopping fourteen
degrees. Thus even several multiples of two degrees should be a piece of
cake for the eyeballs. I think I would be perfectly comfortable with
window to infinity on-film deviation of, say, 3mm, resulting in a mount
aperture spacing of 59mm for my example viewer. Actually, the limiting
factor that the rule addresses might be not the convergence per se, but
distortion in the view caused by looking through the lens at an angle. But
whatever the reasoning, I find that through most magnifying lenses my own
eyes can easily accommodate convergence double that called for by the 1/30
rule. As a result I often do double depth mounting, that is, I use a
special mount with the window twice as close as in a conventional mount.
Of course, due to limited depth of focus the close-up subject matter
usually ends up blurry. This is often acceptable because of the way it
mimics natural vision. That is, in scenes where everything that needs to be
scrutinized lies beyond any nearby objects that happen to be obscuring your
view, you naturally ignore the blur up front. However, if blur turns out
to be undesirable, you can often gain depth by using view camera movements,
or wider lenses, or wider lens separation.

Decoupling of focus seems to be a trick most people accomplish
unconsciously. In fact, for my old eyes it is an advantage in that I can
view the whole slide in focus without resorting multifocal eyeglasses (as I
would when looking at the actual scene). In any event, I suppose that the
best setting for the viewer would be the farthest distance that your eyes
can focus. Not only should that make for easier decoupling of focus (if
that is a valid issue), but your eyes would need to converge less when
looking any given object than they would at a closer focus setting.

It should be noted, however, that when it comes to viewing discomfort,
rotational mismounting can be a much worse culprit than any of the factors
discussed above. Misalignment between the two chips amounting to a small
fraction of a degree of can cause painful tugs on the eye muscles.

Please set me straight on any misconceptions I am harboring regarding
viewer design.

Oleg Vorobyoff

--- In MF3D-group@yahoogroups.com, Oleg Vorobyoff wrote:

>
> Recent discussions regarding the stereo window have me questioning the
> conventional viewer design parameters.

... Say the average relaxed eye lens spacing is 62.5mm and perhaps

> 61mm when the eyeballs converge to look at something close up. Why

then

> are the lenses of my Chinese viewer spaced at 64mm and those of other
> viewers as much as 65.5mm? The eyes are forced to view slides through
> suboptimum points of the lenses, a situation that only gets worse as

they

> converge on close-up homologs. For comfort, I think it would be

better to

> space the viewer lenses at, say, 62mm.

.... The chips for my example viewer would thus have infinity

> spacing of 62mm. Why then do all the common viewers have wider lens
> spacing than the nominal chip infinity spacing? This forces the

eyes to

> start out crosseyed when viewing the most distant objects, and

strains them

> even more when viewing anything else on the slide.

I believe you are not as much questioning the validity of 64mm viewer
spacing, but using it in combination with the placement of infinity
separation at 64mm (or thereabouts), correct?

I found that if I used short focal length lenses for MF3D, like 60mm
lenses, I preferred both a shorter lens spacing and shorter infinity
spacing. I believe it was around 62. For 75-80mm viewing lenses
however, 65mm spacing works better for me.

Interocular adjustment of the lenses in the viewer solved many viewing
problems. I did propose a slide mount with an adjustable infinity
separation a year or two ago, but never went any further with it. I
think for convenience to most it really boils down to the law of
averages. The average eye distance according to Ferwerda is 65 to 66mm
for men and 62mm to 63mm for women. This may explain why more women
seem to have problems viewing than men. But there are far more men
into stereo than women.

Sam

>The average eye distance according to Ferwerda is 65 to 66mm
>for men and 62mm to 63mm for women.

Thanks, Sam. This was the information that I spent fifteen minutes
unsuccessfully trying to find using google. My assumption, then, that the
average is 62.5mm was incorrect. Thus the 64mm lens spacing of the Chinese
viewer is on the nose (or on the eyeball). I can't trust my memory for
anything.

Oleg

Sam wrote:

>I believe you are not as much questioning the validity
>of 64mm viewer spacing, but using it in combination
>with the placement of infinity separation at 64mm
>(or thereabouts), correct?

Actually I was questioning the validity of 64mm viewer lens spacing, but
that was before you informed me that average human interocular is, in fact,
64mm. I still, however, stand by my assertion that infinity separation
should be same as lens separation.

>I found that if I used short focal length lenses for MF3D,
>like 60mm lenses, I preferred both a shorter lens spacing
>and shorter infinity spacing. I believe it was around 62

By sliding together the lenses and chips as a unit you kept eye convergence
geometry unchanged relative to what it would have been with a wider
spacing. You simply ended up looking through the lenses at axes 1 to 2mm
outside the axes of the lenses themselves, depending on your eye
interocular. I'm not sure why that helped. Using 60mm lenses to view s
tereopairs taken with 75mm lenses (and mounted in 50mm square mounts) would
have increased the amount the eyes would have to crank to converge on
subject matter at the periphery of the slide. Maybe you found "sweet
spots" in the lenses through which this additional immersion would be more
comfortable to view. It is complicated because when looking sideways one
eye will roll closer to the center of its lens, while the other eye will
roll farther away from its. Maybe the improvement in the view of one eye
trumped the deterioration in the view of the other. It might be a good
rule to have somewhat less lens spacing than eye spacing with any viewing
lens focal length.

>For 75-80mm viewing lenses, however, 65mm spacing works
>better for me.

Makes sense if your eye spacing is any greater than 65mm.

Oleg