Re: [ATM] Diagonal Size

[Mark Holm <mdholm@telerama.com>]

I think the histogram is a poor tool for looking at resolution issues. 
Probably the correct thing would be a spatial frequency spectrum
generated by a Fourier transform.  You would want to normalize to get
the same effective exposure.  Then, the difference between the spatial
frequency spectra should show the resolution effect of the obstruction. 
Maybe from this data, one could do an MTF plot (sounds logical to me).

Examining test images can give a pretty good feel for the result.  I
don't think these targets were terribly good for showing the effects,
but you can see some differences, especially in high spatial frequencies
(fine detail).

Knowing this conceptually, and knowing how to carry it out in practice
are two different things.  I would need help to actually do it.

I believe there is a simple approximation to understand the effects of
different kinds of obstructions.  (based on Fourier theory)  Low
frequency obstructions, like a big black circle (secondary shadow), or
an aperture mask, have their greatest effect on high spatial frequencies
(fine detail).  High frequency obstructions, like spider vanes, or
scratches or pits or high cirrus clouds, have their greatest effect on
low spatial frequencies (spread light all over the image).  Cirrus
clouds are an interesting example, because the ice crystals are often of
a particular size range.  Since they have a relatively small size, but
the range of sizes is narrow, they tend to spread light over a
restricted range of low spatial frequencies, in other words, a halo. 
You could probably get the same effect by grinding a polished mirror
very lightly with grit of a very narrow particle size. or leaving pits
from only one size of grit.  Bright sources would show halos.

So, central onstruction or smaller aperture should impair fine detail,
while leaving larger scale contrast pretty much alone.  Scratches and
pits should fog the whole image, lowering overall contrast, but not
degrading fine detail more than the overall lowering of contrast.  This
of course does not include the effects of seeing.  When the atmospheric
effects are factored in, the analysis becomes quite a bit more complex,
and beyond me.

Mark Holm
mdholm@telerama.com

_______________________________________________
ATM mailing list http://www.atmlist.net/