let's talk dark libraries on cooled cameras

troypiggo (Troy) · #1 03-03-2011, 05:45 PM

So now I've got a cooled CCD camera. There's clouds here for the next week, so thinking I can build up a dark library. I never bothered with the 40D and took darks every outing. But since the temperature can be regulated now, why not? Some questions and thoughts I'm having:

What min and max temperatures? I understand (yet to test myself) that the QSI583 can get down to around -20 degrees max. This might be maxing out the cooling, though, and maybe -5 degrees is enough. I guess you assess it by taking some darks and having a look at the noise at that temperature for given exposure times. Is there a way of actually measuring the noise level against a benchmark to know what's acceptable?

What increments? Thinking maybe 5 degrees increments? eg -10, -5, 0, 5 degrees etc. Or is that too excessive, and 10 degree increments is enough?

How often? I understand with time the darks will change. Read one guy uses a kind of "rolling library" where he takes a few new darks every few weeks, deletes the older ones, and makes a new master. Alternatively make a whole new library every so often. But how often? 1 month? 3 months? Yearly?

Any other thoughts or things to add?

Octane (Humayun) · #2 03-03-2011, 06:07 PM

Stick to 10 degree increments.

Don't forget binning, you'll need a set for 1x1 and 2x2.

Swap 'em out every couple of months.

H

Tandum (Robin) · #3 03-03-2011, 07:03 PM

Troy, I normally shoot a cal library twice a year. One as summer starts and a new one when winter starts. I know I can easily achieve -20C in summer and -30C in winter so I shoot at the right temp for the season. I find it normally takes over 24hours to do a full set of cal files ranging from 20min to 60sec for darks binned x1 and x2, a dozen each, plus bias files for x1 and x2 about 30 each. I don't capture at x3 or x4 so don't bother with cal files for them.

I would of course reshoot for the current season if I updated drivers or similar on the system.

troypiggo (Troy) · #4 03-03-2011, 07:35 PM

Gotchya. Thx guys. Makes sense.

peter_4059 (Peter) · #5 03-03-2011, 07:41 PM

Wow - I never though I'd need to update the dark library so often? I've got -5, -10, -15 with exps of 1m, 2m, 5m, 10m. I've also shot bias and dark flats at each temp although my tests suggest the temp doesn't have much impact at these exposure lengths.

OzRob (Rob) · #6 03-03-2011, 10:03 PM

Adam Block suggests that darks should be stable for 3 to 6 months.

I have started to use scaling of my darks so that I can take two sets (for 1x1 and 2x2 binning) for each temperature. In future I will only be producing them at an exposure time of 1800 seconds. It will take some time to produce a reasonable dark but it can be done while I am sleeping.

marki · #7 04-03-2011, 06:52 PM

Like Robin I update my library every 6 months but leave them at -20 all year round as there seems to be little to gain with the KAF 8300 by going any lower. I also only take 1 set of 10 darks for 1 x 1 and 2 x 2 bins and use these for all calbration regardless of exposure time. My max exposure is about 15 mins with the KAF 8300 so I take a set of darks with 5X that (75 mins each) then set the calibration routine up for adaptive dark subtration (scaling) so they can be used for all exposure lengths. I have found this works well giving very clean images.

Mark

troypiggo (Troy) · #8 04-03-2011, 07:48 PM

Thanks again guys. This weekend (considering the weather) is to be spent creating these libraries.

RobF (Rob) · #9 04-03-2011, 08:08 PM

My 20c would be don't get too obsessed with getting it right first time Troy. Its a new learning curve with things to explore. It takes a while to get a feel for gain/offset, preferred temps, when and how to shoot darks/bias, not too mention flats (or how many filters to do those for). As Robin said new drivers can come along from time to time too. Processing becomes more complicated too, and you get variations in master cal frames between different software packages to some extent.

You can always shoot more darks or flats if you need them after a particularly successful imaging night, but just make sure you remember to do flats while the imaging train is together (or at least keep focus locked and measure the angle the camera is at very carefully).

Hope the skies part soon for your first light. Should be a memorable occasion.

Stigandi (Rob Stinson) · #10 05-03-2011, 03:53 PM

Hi guys.. Good thread. I've got a SBIG CCD. My dark library was shot a couple months ago using CCDSoft. I have recently moved completely to be Mac based. I'm now shooting the lights with Equinox Image and do the processing in Nebulosity. I usually shoot 10 or 15 minutes lights. The processed images look fairly clean. However there was at least one comment in the thread above about using different software. Will the darks be measurably different if shot with different software?

Thanks... Rob

allan gould · #11 05-03-2011, 04:26 PM

I use either 5, 10 or 15 min subs and have made master darks for these but find they are fairly stable. I also have flats (taken at the time) always of 4 secs duration. Take flats and adjust light box with a potentiometer to give what I want with a 4 sec exposure. I have master flat-darks and bias frames All done at -20C. redo them about every 3-6 months.

RobF (Rob) · #12 05-03-2011, 04:47 PM

Quote:

Originally Posted by Stigandi

Hi guys.. Good thread. I've got a SBIG CCD. My dark library was shot a couple months ago using CCDSoft. I have recently moved completely to be Mac based. I'm now shooting the lights with Equinox Image and do the processing in Nebulosity. I usually shoot 10 or 15 minutes lights. The processed images look fairly clean. However there was at least one comment in the thread above about using different software. Will the darks be measurably different if shot with different software?

Thanks... Rob

Just to clarify my comment on software Rob, I don't have much experience capturing data with different packages, but I do note differences in noise levels in cal frames and final calibrated images between different packages, most likely due to differences in integration algorithms and their implementation (e.g. median/average/sigma clipping/windsorized etc). Doesn't matter much once you settle on what works for you

You would expect drivers to be affecting data coming off the camera more than software. Certainly for the QHY9 there have been notable differences when people have changed driver version at times.

strongmanmike (Michael) · #13 05-03-2011, 06:24 PM

Hi Troy

It's hard for you because you aren't totally familiar with your camera yet but I have had great success with a library of darks and flat darks taken at just one temperature.

With the previous FLI PL11002 (KAI11000 chip) I used the same set of dark frames shot at -35C all year round for nearly three years - no problemo.

With my current FLI PL16803 (KAF16803 chip) I have been using the same dark library set at -30C (made in May last year) all year since with no problems.

The FLI has superior stock cooling to other cameras but the point is that from my experience doing regular updates of darks hasn't been necessary..?

I wouldn't bother with increments either - just complicates things and increases the chance of using the wrong dark folder during processing

. Consider your local temperature patterns and your camera cooling specs and perhaps choose two teperatures that should be just comfortably reached by your cooling in winter and in summer and do a dark set at each temperature..? I believe the QSI can go about 38C below ambient..? So given Brisbanes weather stats I would pick -20C for most of the year and -10C for some nights during the hotter summer months and you should always be able to reach these ok? Use the -20 any night you can and only run at -10C if you have to. Of course if you get the water cooling add-on you could go another 5C colder in each case.

By the way, why are you doing "bias" frames? If you use darks you don't need bias frames...? You need light darks and flat darks at each temperature, that's it . Aslo, if you take your flats at the end of the night when it is always cool you will likely always be able to use a -20C flat dark set anyway so wont need to bother having a -10C flat dark set then.

Mike

troypiggo (Troy) · #14 05-03-2011, 07:52 PM

Thanks again everyone.

Mike, I didn't mention biases. Maybe you read that in someone else's post? Totally makes sense what you're saying, though.

strongmanmike (Michael) · #15 05-03-2011, 11:24 PM

Quote:

Originally Posted by troypiggo

Thanks again everyone.

Mike, I didn't mention biases. Maybe you read that in someone else's post? Totally makes sense what you're saying, though.

..Right, who was that? huh? com'on...I know you're at the door..what?? Who's that?...where'd he go...

Tandum (Robin) · #16 06-03-2011, 12:22 AM

I shoot bias files. I believe you need them if you don't use dark flats or if you are scaling the darks.

troypiggo (Troy) · #17 06-03-2011, 10:53 AM

Yeah, I plan too shot flat-darks as well as darks. Suppose it's not that much more effort to shoot a bias library as well "just in case" I dont' get a chance to shoot flat-darks.

strongmanmike (Michael) · #18 06-03-2011, 04:15 PM

Quote:

Originally Posted by Tandum

I shoot bias files. I believe you need them if you don't use dark flats or if you are scaling the darks.

If you have darks that match your lights in temp and duration, then all you need are:

1) Lights + matching darks
2) Flats (for each filter) + matching Flat darks

A Flat is just a Light frame, so all they need is a matching dark

Mike

Tandum (Robin) · #19 06-03-2011, 04:50 PM

But I have 3 scopes and 2 reducers = many different setups and lots of different duration flats. Shooting flats with a qhy9 is a pain in the rear at the best of times never mind flat darks for all those setups, so I opted not to use them. Bias files are a lot easier to manage in my case.

So I use :

1) Lights + matching darks
2) Flats (for each filter) + Bias files

OzRob (Rob) · #20 06-03-2011, 05:19 PM

Quote:

Originally Posted by Tandum

But I have 3 scopes and 2 reducers = many different setups and lots of different duration flats. Shooting flats with a qhy9 is a pain in the rear at the best of times never mind flat darks for all those setups, so I opted not to use them. Bias files are a lot easier to manage in my case.

So I use :

1) Lights + matching darks
2) Flats (for each filter) + Bias files

If you are using MaxIm DL getting flats is very easy with the Sky Flat Assistant plugin. I point my scope at my monitor with Al's Virtual Light Box covering most of the screen. It is painless once you know the brightness levels needed for VLB.

I don't bother with matching darks for the flats. I just subtract the bias from each when I create a master.