Compute CCD Characteristics

[georg.viehoever]

Hi,

attached a script that should enable you to determine some characteristics of your CCD, such as dark current, readout noise and gain. I used the method described in described in "The Handbook of Astronomical Image Processing" (HAIP) by Richard Berry and James Burnell, Second Printing, Errata for chapter "8.2 Basic CCD Testing" (see http://www.willbell.com/aip4win/Errata%20to%202nd%20%20Printing%202nd%20Edition%20of%20HAIP.pdf). All you need are 2 bias frames, 2 flat frames and 1 dark frame.

Currently, I am not really sure if I did implement everything correctly. I am publishing this first version in the hope that someone will review the implementation. Let me know if the implementation needs to be changed. And also let me know if you find it useful.

Georg

[Nocturnal]

Looks good Georg but why allow color images? Wouldn't this be most (only?) meaningful with mono frames?

[georg.viehoever]

Well, my Canon happens to be a color camera... .

But you are right, maybe I should read the frames as monochrome without debayering, since all colors share one sensor, one amplifier, one AD-converter. The script should already be able to handle that. Do you have any idea how I can convince PI to read the RAW .CR2 files without splitting pixels into different channels? The PI Format Explorer does not give me this option (even "Create RAW Bayer picture" splits channels, see screenshot).

Georg

[Nocturnal]

Ah, didn't realize raw DSLR files get debayered by default. Maybe Juan can suggest something. I also have a color camera but I need to debayer explicitly. That's why I wrote the debayer module

[georg.viehoever]

Sander, Juan,

I guess I will have to write a small "Rebayer" script. The "-D" option from dcraw also would do the trick, but it is not available with PI.


Georg

[Juan Conejero]

Hi Georg,

Extremely good work!

I can easily modify the DSLR_RAW module to include the -D dcraw argument as an option. Unfortunately I have to go out of town right now but next Monday you'll have a modified DSLR_RAW module, which we can release with PI 1.5.3 (along with your new and nice script if you want, of course ).

[georg.viehoever]

Juan,

sounds good! The alternative to "dcraw -D" would be to do the "re-bayering" in the script, which unfortunately does not work yet, see http://pixinsight.com/forum/index.php?topic=1236.0.

You can include the script with PI when it has matured a little bit, maybe with 1.5.4 or so. I am sure it is still buggy.

Georg

[vicent_peris]

Hi!

Amazing script.  But this gives me another idea... It would be nice, to use while observing, an script that says you how long you must expose, depending on your camera readout noise and your sky background illumination.

Imagine you take one 10 min exposure. You don't know if you will go deeper if you expose 20 min, or if there's no sense in taking subs longer than 5 min. You indicate the script the file corresponding to the 10 min exposure, and one bias, and also you say the script how many times the sky background noise must be the readout one. Calculating the actual ratio, you can predict for the actual working conditions, how long you must go in your subs.


Regards,
Vicent.

[georg.viehoever]

Vincent,

this is exactly what I had in mind, although I wanted to do that with Excel first. Given the characteristics of your CCD, the maginitude of your sky background (light polution), the magnitude of your target object, exposure time and number of exposures, it should be possible to predict the signal/noise ratio of the resulting image (at least a ballpark figure). Of course, it would be possible to change any of the exposure parameters, and see how this influences the resulting image. Depending on the current seeing conditions and the available time, you would be able to predict if it is worthwhile to take pictures an object.

I am not sure if such a theoretical approach is valuable. In a quick search of the internet I did not find any exposure calculation tools that works this way, so maybe few believe it would be useful  .

What do you guys think?

Georg

[mmirot]

George,

I think Vincent is talking about optimal subexposure ( read noise in burried in the sky noise).

I think your idea is a bigger challenge.  I think everyone asks your question.

How long does it take to get a reasonable  S/N for a given object?

Max

[mmirot]

Sorry about the "e" Georg 

[georg.viehoever]

Good morning,

I found some nice material about optimal (sub) exposures. Actually, there is something like a "sweet spot" in exposure times, mainly depending on the characteristics of your CCD and the sky background. http://starizona.com/acb/ccd/advtheoryexp.aspx and the pages linked from there discuss some of the issues. They are based on the excellent notes written by John Smith in http://www.hiddenloft.com/notes/acq.htm, who also discusses that there are more issues that just the best signal/noise ratio to consider. John also provides an Excel Sheet http://www.hiddenloft.com/notes/DarkandSubExp.xls that allows to do some calculations based on those models.

I plan to add the capability to measure the sky flux and the dark signal doubling temperature to the script (when the current PJSR issues are resolved), which would allow to use the spreadsheet to get an idea about a good exposure/stacking strategy. At least for deep sky objects where we have a surface brightness, it should also be possible to estimate a resulting signal noise ratio SNR.


Georg