11 replies to this topic

[Steve Saturn]

I'm interested in creating a library of master dark frame images so that I don't have to use a good percentage of my time in the field shooting dark frames when I could do that at home. I have a couple questions:

1) First of all, is this worthwhile? In Ron Wodaski's New CCD Astronomy he's quite adamant that dark frames should be taken as closely as possible to the time that the images are taken. On the other hand, this seems like a common practice.

2) If this is indeed worthwhile, how often sould you re-shoot your dark frame images?

3) How many images (minimally) would you want to combine for each temperature/exposure set in order to create the master dark frame for that temp/exposure combination?

4) Is median combine the best method to use to combine dark frames?

Thanks for your help!

Steve

[SleepIsWrong]

Steve:

Average combine is a bit better than median combine from a signal/noise standpoint (factor of sqrt(2) better if I recall). That is unless you have some cosmic rays in your images. Then it helps to median combine the images to make a comparison frame that you can use to find the cosmic rays in the individual images. Remove the cosmic rays and then average combine.

Its best to take the darks on the same night as your images, but a few days or so won't make much difference. Hot pixels gradually come and go so a dark taken in your refrigerator in June may not have the same hot pixels come November when you go to use it. It really doesn't take all that long to acquire darks for an evening. I usually do one set just after my first set of images and another just before the last set for the night. I typically take 10 to 20 darks in each set; exposure time for each dark frame is the same as whatever the exposure time I'm using for the evening.

Mike

[Dean]

Steve -

1) With a cooled CCD, you don't have to shoot your darks at the same time as your light frames. The main thing is that they be taken at the same temperature as the light frames and with a cooled CCD you simply set to be the same.

2) Over time the CCD can change, mine has developed new hot pixels and some columns have changed so I have had to reshoot my darks a couple of times, but generally I get about 8-10 months out of a set.

3)I use 10 dark frames. It's not ideal and I probably should use more, but 10 works fine for me.

4) Median combine is much better that summing and somewhat better than averaging for controlling flyers. I use a sigma-clip median combine because it is more agressive at removing flyers, but that's mainly because at my altitude I get a ton of flyers from cosmic ray hits.

[Chuck Anstey]

I just read that in Wodaski's book so I tried taking some dark frames after my shots a few days ago. There is one thing he forgot to mention in his book about SBIG cameras and maybe others. If you shoot your dark frames immediately after your light shots, you can see smudges created in the exact same spots as the light shot's saturated areas. It seems to take many minutes to completely drain the charge down to its normal bias after pushing the pixel to its limit.

[Steve Saturn]

Dean and Mike,

Thank you very much for your thoughts. That's a great help. I'm looking forward to starting this project then. I'll plan on a minimum of 10 frames for each temp/exposure combo and median combine.

[Steve Saturn]

Hey Chuck,

Great observation. If I remember correctly (my book is at home), it says someting about ideally you would take some dark frames before and after each exposure to use in your master. I guess your test puts the "after" shots in question. Hmm......

Steve

[Hoser]

Steve, that sounds like something you would do with a non-temperature controlled camera, like the Meade DSI series.

Chuck, very true...there is residual charge that can certainly act like you describe.

With TEC cams, like the SBIGs, you want to do your darks at other times, so as not to waste precious imaging time at the scope. For example, I do my batches of 30 minute dark subs overnight, when I am not imaging.

[Chuck Anstey]

I have a complete library of darks at home that I had done a few weeks before. I was reading the book while imaging and decided to try it. Plus I was able to go down to -25C instead of my normal -15C so I needed ones for that night anyway. Wodaski also thought that much colder is better but even though I definitely had much lower level dark frames at -25C, after applying and stacking I could not tell the difference between the shots at -15C and -25C, at least for my C11 / CGE combo.

[MichaelW]

3)I use 10 dark frames. It's not ideal and I probably should use more, but 10 works fine for me.

I am a little fuzzy on a dark library although it sure makes more sense than wasting 1/2 of you imaging time shooting "autdarks". So you have a library of 10 darks, can I ask at what temps and/or times you are using for my go by for a start up library. Is the time more important than the temp or versa visa? How critical is the timing and temps?

Thanks!

[Dean]

I'm sorry Michael, I didn't state that very clearly. I don't have a library of 10 darks, I combine 10 dark frames for each exposure time/binning that I use. I shoot everything at -20C. I could go much colder in the winter, but I use -20C because I can shoot that all year without ever having to worry about overtaxing the cooling system.

Anyhow, the exposure times I commonly use are 1, 3, 5, 8, 10, 15, 20, 30 and 45 minutes so I have a combined dark frame for each of these binned at 1x1 and some binned at 2x2 as well, all taken at -20C. I have a few others at different exposures (I'm at work so I can't say just what I have) Each of these dark frames is 10 individual dark frames (but probably should be more).

Both exposure time and temperature are important. However, thermal noise increases at a predictable rate with time, so it is possible to create a single dark frame that can be scaled for different exposures. I've never felt the need to do this though.

[Chuck Anstey]

I have 1,2,3,4,5, and 6 minutes 1x1 and 2x2 at -15C. I do not yet do H-alpha so I don't need longer than that. I shot 9 frames for each dark and median combined. I'm not sure you really need more than that. If noise is still a problem after 9 frames then the dark subtraction won't be very good for each light shot because of the large variance shot to shot.

Dean,
-20C all year? Do you have the water cooler or is it always in the 60's or lower? On the hot summer and even fall nights I can't even get to -15C. Just last week it wasn't until 10:30pm that I could get down to -15C.

[Dean]

No water cooling. My max night temperature is about 65.