Difference between CosmeticCorrection and Dark subtraction?

[joelshort]

I'm very new to PI...
Since the CosmeticCorrection script uses a master dark to remove hot pixels, how is that different than just subtracting darks from an image during calibration with the batchpreprocessing script? 

Remember I'm a newbie so give me the dummy version.   

[pfile]

CC pretty much just fixes hot pixels and bad columns... the dark is just used as guidance so CC can identify which pixels are hot/warm.

doing dark subtraction will remove the dark current (and bias) from your lights. CC won't do that.

CC can be useful if you have residual hot pixels left in your calibrated lights. this can sometimes happen when your master dark is scaled while calibrating lights.

rob
 

[topboxman]

You probably don't even need to use Master Dark to remove hot/bad pixels. You can "Use Auto detect" check box instead of "Use Master Dark". I've tried both methods and found no differences.

Peter

[joelshort]

So your suggestion is to use darks in the calibration process, and if needed use the cosmetic correction for any residual hot pixels?  In another forum it was suggested that I not use darks on narrowband images because it might introduce more noise and the suggestion was just to use CC and bias/flats.  I'm using a SX-694 (Sony) camera that is pretty clean.

[pfile]

yeah, the type of camera is important. supposedly the sony sensors have extremely low dark current and so many people only use bias frames.

i'd say that if you don't have bad columns, then you can probably live with the hot pixels assuming you are dithering during acquisition. with enough frames you should be able to reject the hot pixels as outliers during ImageIntegration. if you have too few subs for whatever reason, or are not dithering, then CC might be helpful to clean up those hot pixels.

rob

[joelshort]

Thanks Rob and Peter.  I'll see what it looks like with ImageIntegration and go from there.  Having been in the Photoshop/DeepSkyStacker camp for so long with a KAF-8300 camera it is a foreign concept to not need to use darks. 

[chris.bailey]

Joel

My workflow with the KAF8300 most certainly includes dark frames, especially for long exposure narrowband images. I find the darks to have a distinct gradient towards the periphery which is hard to get rid of using DBE. My workflow is therefore to calibrate with bias, optimised darks and flats. Cosmetic correction with Autodetect (Sigma of 3 does it for both) and a defect map to clean up a residual bad column. My subs have a 5 pixel dither to them which should get rid of any remaining. The problem is hot pixel clusters. My 383L has several small clusters of errant pixels. Interpolation during star alignment can sometimes cause these to grow meaning some leak through into the stacked image.

Chris