Re: Sputnik at f32?

[John Bercovitz <bercov@xxxxxxxxxxx>]

>  Regarding your 50mm lens that you made waterhouse
>stops for. Repeating the above (50mm lens, an=2m, af=inf)
>
>At f11: rg=3.9 rd=.42
>At f16: rg=2.7 rd=.61
>At f22: rg=1.95 rd=.84
>At f32: rg=1.34 rd=1.23
>At f45: rg=0.95 rd=1.73
>
>  So here f45 is bad news. F32 isn't great
>unless the near object and infinity are the
>most important in the scene. F22 looks like
>the min aperture to use.
>
>  Straighten me out if I'm confused about
>what MAOFD is telling me.

Right on!  I'll bet if I plugged MAOFD for the actual
distances I used, I'd find the results I observed.  
Thanks for reminding me I could use MAOFD.

It looks like what we're getting back to is the old
thing about the angular resolution of a camera or a
telescope or whatever being a function of the entrance
pupil diameter and nothing else.  That is that the
angular resolution in seconds of arc is 114 divided
by the entrance pupil diameter in mm.  

Since we view a stereo image from the center of 
perspective, more or less, it is the angular 
resolution which is important, not the lines per
mm on the film.  If you view from across the
room, then 5 lines per mm is more than good enough
but if you view from 25 mm, you want a whole lot
more "plate" resolution.

An 80 mm lens at f/45 has an entrance pupil of 
80/45 = 1.8 mm while a 50 mm lens at f/45 has an
entrance pupil of 50/45 = 1.1 mm.  The corresponding
angular resolutions (for lines, not points) are 
114/1.8 = 63 seconds of arc and 114/1.1 = 1' 44" of
arc.  The former is tough to see while the latter is 
fairly easy (under the usual viewing conditions of 
moderate contrast).  And of course we have to add 
them in quadrature to the CoC due to fuzziness which
pushes the fuzziness up to where it's more obvious.

John B