Dark Frames for DSLR

[RikM]

First thing to realise is that you cannot subtract out random noise, when you subtract two images with random noise, the resultant random noise will be higher. The only way to reduce random noise is to take many frames and average them. This is why we take many light frames, flat frames, bias frames, and dark frames. It is solely to reduce the random noise in these images to zero. Taking many bias frames leaves only the fixed component of the noise (or pixel to pixel variations), taking many flats leaves only the fixed component of noise and taking many darks leaves just the fixed component: hot and cold pixels.

So when we do image calibration we do NOT subtract the random noise, we subtract the fixed pattern components. You can ONLY remove fixed stuff via subtraction.

Paul, thanks for distilling such a complex and confusing subject down a nice succinct explanation. I do understand what calibration is actually doing, and why we take a sufficient number of calibration frames, but have been unable to explain it as elegantly as you have here. Very many thanks for this

[Alien 13]

A question?

I have problems with short term memory and often am faced with 100s of images on my sd card flats/darks etc with no real idea of what how and when i took them i then have to trawl through them trying to make sense of this and slowly organise them. 

Is there any software that can automate this process i.e. sort by date time exposure temp dark or bias f ratio used (i have a chipped t adapter) or any other tips or spreadsheets for this?

[Russe]

Usually I get around 8-15C with this sort of setup. Without the foil it got to -5 once.

[Russe]

A question?

I have problems with short term memory and often am faced with 100s of images on my sd card flats/darks etc with no real idea of what how and when i took them i then have to trawl through them trying to make sense of this and slowly organise them.

Is there any software that can automate this process i.e. sort by date time exposure temp dark or bias f ratio used (i have a chipped t adapter) or any other tips or spreadsheets for this?

APT is quite handy. Astrophotography tool - a bargain for its price!

[RikM]

Plenty of software packages can ID images at the capture stage. APT allows you to group images by exposure plan, e.g. flats, bias etc. and can add a sufix to the file name with ISO, exposure time, chip temp etc.

Maxim DL also includes frame types, exposure times, filter used etc in the filenames.

Edit: teach me to start typing, then answer the phone halfway through

[Alien 13]

Thanks Russe and Rik i will look into that much appreciated..

[ollypenrice]

Dithering, when carried out by a programme like Nebulosity, is not performed on a several-pixel scale but on a sub-pixel one. This may be counter intuitive. From what I can work out this sub pixel dithering is intended to support drizzle stacking, a process which ingeniously derives sub pixel information from a stack. However, moving between subs has to be a good thing. Unless you have remarkably serious polar alignment, though, you are going to get it anyway. If, when you stack a set of subs, the borders of the first one exactly overlap the borders of the last, then you have this kind of PA. If not, you are getting PA induced dither. I invariably get it and am happy to leave it that way rather than perfect my PA. It also means I can guide on only one Dec axis and this is usually the best way.

I don't use darks any more with my CCDs. I stack in Astro Art and this has a Defect Map facility. To make a defect map just clip 2000ADU off the bottom of the master dark and put it in AA's box. I put a master bias in the box normally used for darks. I find the defect map and bias routine has several advantages; it gives extrememly clean results. It gives very consistent results. It is remarkably fuss-free, meaning I can use a 15 minute Defect Map on a set of 30 minute subs with absolutely no perceptible disadvantages.

I used to find darks fickle. Sometimes they would 'just do it' but sometimes they seemed to make matters worse. I use the very noisy but gargantuan Kodak 11 meg chip and now get the sweetest data I could wish for.

Anyone tried a Defect Map approach with a DSLR?

Olly

[IanL]

Anyone tried a Defect Map approach with a DSLR?

Olly

Olly, as I said in my post above I have used PixInsight's CosmeticCorrection process as part of the testing I am doing.  It works at least as well as scaled dark frame subtraction (better actually since it also deals with the residual hot/cold pixels that dark frame subtraction doesn't).  The way I have done it is to use a large master dark, and also the 'Auto Detect' settings which just let you select hot and cold sigma values to reject pixels that are outliers vs. their local environment; you can also use a third setting which is a defect list that you manually create but I haven't tried that.

Visually the results are indistinguishable from the scaled master dark subtraction process.  You do have to create a master dark once for use with all your images, but you don't have to be concerned about matching it to the lights for temperature or exposure length, so in PixInsight at least there are two pretty satisfactory methods of calibrating DSLR images.  For anyone using a different package that didn't have dark scaling, I would recommend taking a look at any defect map/cosmetic correction type processes that it has, as you may well be able to avoid a lot of hassle trying to create matching dark frames.

Paul,  I don't think there is any disagreement between us save for whether it is right to call the various fixed elements that we are trying to remove during calibration 'noise'.  I know many articles refer to them as noise, but to my mind that is liable to mislead and cause people to draw the wrong conclusions, so I am going to stick with 'unwanted signal' since that distinguishes it from the noise which you cannot subtract (only minimise the effect of it by taking more samples of the unwanted signal).