Re: Ultra-quick evaluation of scanned photo with Mak lens

[Tom Droege <droege@wwa.com>]

Using Michael's assumptions, and a few more:

Signal to noise at the center is:

117/(sqrt(s))

Where s is the sky background component.

Signal to noise at the edge is:

71/(sqrt((71/117)*s))

Taking the ratio of signal to noise at the center and the signal to noise at
the edge and simplifying:

=(117/71)*sqrt(71/117) =1.28

log 1.28 * 2.5 = .27

So we lose .26 magnitude of sensitivity at the edge.   Seems not too bad to
me.

One problem with this biz is that you do not gain  fast no matter what you
do.  One good point is that system defects do not cause you to lose fast.
So in this case, the glass is sqrt(half) full.  = .7 full.  Wonderful
world, eh?  
Nature seldom does things in your favor.  

Tom Droege


At 06:19 PM 1/7/98 -0500, you wrote:
>  Looking at the scan of the blue-sky image Cameron sent, it appears to me
>that the difference in brightness between the center and corners is 
>something like:
>
>                  (brightness at center) - (brightness at corner)
>                 ------------------------------------------------
>                           (brightness at center)
>
>                           117 units - 71 units
>                =         ----------------------
>                                  117 units
>
>                
>                = 40 percent, approximately
>
>  
>  This assumes that the film response was linear, and that the SCANNER
>respose was linear, too.  And that the JPEG-compression didn't make
>things wierd.  And that I understood how to interpret the little numbers
>ImageMagik displayed as I moved the cursor around :-)  That's a lot of
assumptions.  
>
>  Let's see: the Mark IV chip will be 2048 pixels * (15 microns per pixel)
>wide -- that's about 3 cm side to side.  Your photograph was a little 
>more than 3 cm in the long direction ... so, if the lens really does
>produce the variations in your scanned photo, it will produce roughly
>the same variations with the Mark IV CCD.
>
>  Anyone want to speculate on the maximum amount of vignetting we may
>permit in a candidate for the Mark IV lens?  Remember that we can compensate
>for a known illumination pattern, but that the noise in the corners will
>always be higher than the noise near the center.  Arne, what would _you_
>consider acceptable?
>
>  Oh, Cameron (or other people with big lenses): please do go ahead and
>use PPF film, if you've got it handy.  For the current purposes, I would
>say it's perfectly adequate.  Good luck with unguided pictures ... a few
>of mine have come out okay.
>
>
>  
>-- 
>-----                                      
>Michael Richmond                   "This is the heart that broke my finger."
>mwrsps@rit.edu                     http://p674p06.isc.rit.edu/richmond/
>
>
>
>