Dark scaling

[bpipes]

I am just getting started with PI, so this question may have a very simple answer. I created a master dark frame and saved it in FITS format. However, when I look at the FITS header of the master dark frame, there is no information about the ccd temperature or the exposure time. If I use this master dark on light frames with a different exposure time, where does PI get the information it needs to perform dark scaling?

Thanks,
Bruce

[Juan Conejero]

Hi Bruce,

Our dark frame optimization algorithm does not depend on any physical property of the sensor or acquisition conditions. It will always converge to the optimum scaling factor that minimizes noise induced by dark subtraction in the calibrated image. Exposure time and temperature are just not necessary with our calibration routines.

[bpipes]

Hi Bruce,

Our dark frame optimization algorithm does not depend on any physical property of the sensor or acquisition conditions. It will always converge to the optimum scaling factor that minimizes noise induced by dark subtraction in the calibrated image. Exposure time and temperature are just not necessary with our calibration routines.

Thanks, Juan. Does this mean I can use a master dark with a shorter exposure time than the exposure time of the light frames I am calibrating?

[Juan Conejero]

Hi Bruce,

Does this mean I can use a master dark with a shorter exposure time than the exposure time of the light frames I am calibrating?

Sure. The dark frame optimization routine will find the best scaling factor in terms of noise minimization. Ideally, only thermal noise should be present. If read noise is significant, then the dark optimization routine will have to under-correct thermal noise to find a minimum noise solution. Since the only correction applied to a master dark is bias subtraction, this means that you should take a large number of bias frames in order to get a noise-free master bias. Always take much more bias frames than dark frames —they are cheap to acquire anyway, so why not take a lot.

In case you haven't read it already, this is the reference document for image calibration in PixInsight:

http://pixinsight.com/tutorials/master-frames/en.html   

Something that must always be pointed out, is the fact that dark frame optimization comes at the very small cost of under-correction (and rarely over-correction) of hot pixels. This is due to non-linearity of CCD response at high pixel values. So don't be surprised if after dark frame subtraction with optimization some hot pixels survive in your light frames. Hot pixels can be fixed cosmetically, when necessary (we have a couple tools that work very well for this task: the CosmeticCorrection script and the DefectMap tool), and, if you use dithering between exposures —because you use dithering, don't you?—, they will be rejected automatically during integration (ImageIntegration tool).

[bpipes]

Hi Bruce,

Does this mean I can use a master dark with a shorter exposure time than the exposure time of the light frames I am calibrating?

Sure. The dark frame optimization routine will find the best scaling factor in terms of noise minimization. Ideally, only thermal noise should be present. If read noise is significant, then the dark optimization routine will have to under-correct thermal noise to find a minimum noise solution. Since the only correction applied to a master dark is bias subtraction, this means that you should take a large number of bias frames in order to get a noise-free master bias. Always take much more bias frames than dark frames —they are cheap to acquire anyway, so why not take a lot.

In case you haven't read it already, this is the reference document for image calibration in PixInsight:

http://pixinsight.com/tutorials/master-frames/en.html   

Something that must always be pointed out, is the fact that dark frame optimization comes at the very small cost of under-correction (and rarely over-correction) of hot pixels. This is due to non-linearity of CCD response at high pixel values. So don't be surprised if after dark frame subtraction with optimization some hot pixels survive in your light frames. Hot pixels can be fixed cosmetically, when necessary (we have a couple tools that work very well for this task: the CosmeticCorrection script and the DefectMap tool), and, if you use dithering between exposures —because you use dithering, don't you?—, they will be rejected automatically during integration (ImageIntegration tool).

Thanks again, Juan,and, yes, I dither.