Re: [APML] optimize scanning [was: noise]

[Gene Horr <genehorr@texas.net>]

Matt BenDaniel wrote:

> I wonder if anyone has tried taking dark frames and flats with a scanner to
> calibrate images.

The exposure time for each pixel is so short that dark current would be
immaterial.  Plus you would need to kill the illuminating light.

If we assume 1 minute for a 35mm scan @ 2400 dpi and dark current at 1 e/sec
@ 0 C, then you are exposing each pixel for:

	60 / (24 * 2400 / 24.5) = .025 seconds

The dark current rate @ 35 degrees C would be (assuming dark current doubles
every 5 degrees C - I can't remember whether it is 5 or 7 and so used the 
more conservative figure):

	2 ^ (35 / 5) = 128 e/sec.

Giving us an average of 3 electrons per pixel dark current.  This is probably
at most 1 ADU @ 16bit depth.  If we assume a dark current variation of even
as high as 50% it would be very hard to separate from the readout noise which
is likely over 20 e.  

A flat would be very tough to do as the reflections would be far different
with the film not present.

Gene Horr
genehorr@texas.net

--  APML Archives at <http://astro.umsystem.edu/apml/>  ---
             Unsubscribe at <majordomo@seds.org>