Re: Re: [APML] Equivalent ISO of an ST10E CCD

[Aplanatic@aol.com]

Jeez Jerry, this is a lot of work and some guessing.  Here's my attempt at an answer.

Starting from Kodak's TechPan PDF file, to get a density that's 0.1D above base+fog, it looks like the film needs an exposure of 1/100 lux-seconds if the film is developed in D-19 for 8 minutes.  This development schedule yields an exposure index of 200.

OK, now what?  Well, we have to convert lux-seconds to photons/meter^2.  After fumbling through several references I finally found that one lux-second is 1/683 Joules/meter^2.  At 550 nm, a photon has an energy of 3.56 X 10^-19 Joules, so, one lux-second = 4.11 X 10^15 photons/meter^2.

Since the ST-10E has pixels that are 6.8 X 6.8 microns, at an exposure index of 200 (roughly 1/100 lux-second) each pixel will be exposed to 1900 photons.  The QE of the ST-10e at 550 nm is about 55%, so the 1900 photons will be converted to about 1000 electrons.  This will, of course, vary with the wavelength of light being considered. 

Now comes the interesting part.  As you may recall, the exposure index is defined at a density 0.1D above base+fog.  What does this mean, exactly?  Well, 0.1D is 1.26 times the base+fog.  We must somehow turn this into a film signal-to-noise ratio to compare it with the CCD response, otherwise we just quit here and state that an exposure index of 200 = 1000 electrons for the ST-10E, roughly.

Here's my guess -- and you're welcome to challenge it.  At 0.1D, the signal-to-noise ratio is about 3 (edge of detectability for 6.8 micron pixel scans).  I don't have any justification for this value except from experience with TP.  One could find a region on TP that has a 0.1D density above base+fog and measure the S/N in photoshop to get a better answer, or better yet, look at many regions and draw a graph, etc.  Lots of work to get it close.

For the CCD the answer is easy.  One thousand recorded electrons has a S/N of about 30, dominated by shot noise.  So, its S/N is very roughly 10 times higher than TP.  A factor of 10 change in S/N is a factor of 100 change in "speed" in the shot noise limit, so I'd say that for the case analyzed (low light levels) the "equivalent ASA" of the ST-10E is something like 20,000 plus or minus a factor of two or more.

At higher light levels, I think film does a substantially better job at S/N because of the power-law relationship between density and exposure.  Very complicated and very non-linear.

If you told me I was all wet, I wouldn't argue too hard.  It's all Jerry's fault, anyway.  <g>

Dave Rowe




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