Warning: No correlation between the master dark and target frames

[robhawleyastro]

I am trying to calibrate some flat frames from my dark library.  I prepared the library using the calibration tutorial.  I am getting the following error message when calibrating the flats

Writing output file: S:/Images/ACP/images/2011 05 04-05/flat_FLAT_B_2000ms@2011-05-05T11-37-30_c.fit
Dark scaling factors:
k0 = 0.000
** Warning: No correlation between the master dark and target frames.
Writing FITS image: 32-bit floating point, 1 channel(s), 3352x2532 pixels: 100%

The documentation says that IC will scale based on temperature. I notice that the integrated files do not have either type (e.g. BIAS) or temperature info so how can that work?

[pfile]

well the theory is that temperature and exposure length correlate with noise. but what PI is doing is trying to evaluate the noise in the target frame and scale the master dark appropriately, so it does not have to know what the exposure duration or temperature. it does it all based on noise analysis.

a flat is going to be a very short duration relative to a light, and so the dark noise in the flat is going to be very small. this may be why it's having trouble correlating the two. what is the duration of your dark subexposures?

[robhawleyastro]

This particular one was 2 sec.  Depending on the filter the flats range from 130msec to 60 sec.

I noticed the this happened occasionally on light frames also, but most of the time a correlation was calculated.  Since I match my light frames with darks of the same exposure the two should have similar background noise.  Why would the correlation not be 1.0?

[pfile]

hang on - you had a 2 second dark that you were using to calibrate 130mS to 60s flats? or you mean your flat was 2s long?

PI is trying to minimize the noise in the calibrated frame, and computes a dark scaling factor accordingly. even though your dark and your light have the same duration the amount of dark signal could be different between the two, since in the end the dark noise is (infrared) photon shot noise.

what kind of camera are you using? is it cooled? is it possible that the amount of thermal noise is just vanishingly small and PI is having a hard time evaluating it?

[robhawleyastro]

Sorry poorly written answer

I am trying to match my dark flats to my flats so a 2 sec flat would be matched with a 2 second dark.

The writeup says this is not required. The exposure times on my newer flats vary a bit since they are now computed by ACP.  Thus I may be using a 30sec flat-dark for a 26.xx sec flat or a 60 sec flat-dark for a 48 sec flat.

I have a SBig 8300 that I have recently been running at -20. It is possible for the shortest flats that the noise was just too small. Unfortunately I did not record the specific case

After writing my post I completed calibration of about 12 hrs of lights.  Considering just the lights which are always matched to an identical dark-master about 10% got no correlation and many correlated at less than 1.0.  Perhaps that is because the dark master is cleaner since it is a combination of 20 images, but that is only a guess.

[Tinbeard]

Not sure, but this topic may help.

http://pixinsight.com/forum/index.php?topic=2803.0