[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Rollei] OT: pupil distorsion and MTF effects

 From Richard Knoppow, about subtel effets of pupillar distorsion in
 wide-angle lenses :

>...I've also not seen any explanation of the effect of the tilting
>entrance pupil on MTF variation with image angle. Presumably it could
>increase resolution away from the axis beause the stop is not
>vignetted in the same way as usual. However there may be other

Richard, thank you very much for your additional comments on Roosinov
principles. It is always better to proper credit the original

I have seen one thing related to pupillar effects on MTF in the edges,
namely in Rodenstock MTF charts posted by Paul Butzi in his web site.

Rodenstock large format and enlarging lens literature

Rodenstock's view camera lens MTF charts are simulated and not
measured like in Zeiss datasheet, but they are quite informative.

You will notice on those charts that the ultimate MTF limit due to
diffraction is plotted as a reference and is smaller in the edges than
at the centre of the image field for wide-angle grandagon lenses. My
understanding is that there is also no absolute compensation of light
fall-off by pupillar distorsion in the edges, since Rodenstock like
Schneider sells centre-filters to compensate for this. The elementary
theory of diffraction tells you that the angle under which you see the
exit pupil from a given image point actually yields the cut-off
spatial frequency due to diffraction

f_c = 1/(N_eff*lambda) 

where N_eff = distance/pupil diameter. In the edges of the fiels, the
distance increases, and the pupil diameter changes both by the effect
of a naturela cosine factor counter-balanced by some Roosinov effects.
In fact, this N_eff factor is the same as in photometric formulae used
for close-up (actually : (N_eff)^2 is proportionnal to the bellows
factor), so in a sense if lightfall-off occurs in the edges, the
elementary diffraction theory predicts that diffraction effects are
worse also, in the same proportion as N_eff. I'm really dubious about
the actual validity of the elementary (paraxial) Fraunhofer
diffraction theory for wide-angle optical systems, but there appears
to be at least a certain consistency with what is simulated by
Rodenstock experts.

- -- 
Emmanuel BIGLER