T3D ocular specs II

[john bercovitz <bercov@xxxxxxxxxxx>]

OK, here they are again, revised and with additions

Added infinite conjugate ratio to item 1.

I added the pupil of the system which is the pupil of the eye.  
See item 3.  I made it 5 mm.  If it's any more than that, there's 
something definitely wrong with the lighting in the viewer - it's 
much too weak.

Changed item 5 to 1 moa..  I revised distortion downward because of 
Don's comments and because I read that binocular distortion has to be 
held to <1/2% to be OK for all users.  Could be even more stringent 
for stereo, I have to allow, since different objects will get ever-so 
slightly different distortions since they will lie in slightly different 
parts of the two fields.

I added "no vignetting" to item 6

Added "all lens surfaces..." to item 10.

Added item 13 but I don't like it since I don't want to jack the 
price up higher than it would naturally fall, though I doubt this 
will happen.  8-(

Changed item 12 from 2% to 1/2 % for the same reason as above.

Added item 14 - a question....

================================================================

1) Focal length 36 mm +/- 0.5 mm.  Eye focussed at infinity.

2) Field of 40 mm diagonal (across the diameter of the field).  You 
might think it should be 31 mm for 5P and 36 mm for 7P but these 
are only the diagonals of the apertures in the masks.  The 
apertures are offset from the lenses because the apertures set the 
window and the window is closer than infinity. The difference on 
the mask is an offset of (65-62)/2 = 1.5 mm  This ups the 7P 
diagonal to 39 mm and the 5P to 34 mm.

3) Eyepoint at > 20 mm (the spectacle point is at 16 mm and we need 
some clearance for a bumper on the lens holder to protect the 
glass).  Pupil diameter 5mm.

4) Distortion < 1/2% (one would suppose this would be at the edge of the 
field since the distortion should be a cubic).

5) Resolution 1 moa (the resolution on the film is maybe 40 lpmm 
including many factors such as DOF and it is located at a distance, 
optically speaking, of 36 mm.  Arctan (1/40*36) = 2 moa).  We need
to do much better than this to avoid adding much to it.  Strong 
preference for better than 1 moa.  Let's say a sine wave MTF of 50% 
constitutes resolution.  Wavelength of 555 nm.

6) Allowable offset of eye from lens axis while retaining full specs
and with no vignetting: 5 mm

7) Field flatness equivalent to a range not to exceed 0.5 diopters to 
allow for the elderly and enfeebled among us (that's me).  That's 
+/- 0.25 diopters

8) Color weighting to allow resolution resolution which is half as 
good at 700 and 400 nm as it is at 555 nm.  What should the shape 
of this weighting curve be?  Should it be mapped to the relative 
visibility (y bar) of the colors in some way?

9) How much error in magnification should we allow between the various 
colors?  I'd say what ever works out to 2 moa.  There are contrasty 
situations where this could become important.

10) Multicoat AR coating of all lens surfaces which are exposed to air.

11) Let the cost of the finished product and environmental resistance 
drive the glass choices.  (In other words, expensive glass is OK if 
it solves the problem more cheaply than many extra elements of 
cheap glass as long as the expensive glass isn't weak in its 
resistance to the environement.)

12) Magnification.  Focal length of matched pairs to match better than 1/2%.

13) Cost < ~$200 per pair.

14) What should we do about "swimming"?  This is when you look to a 
different part of the slide and the image shifts as you reorient your eye.  
How should we spec this?

John


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