Making a dark frame library
Posted 06 December 2012 - 10:00 AM
I've recently read Jerry Lodriguss's excellent
A Beginner's Guide to DSLR Astrophotography, and have a few questions.
I'd like to start building a dark frame library like is mentioned on the CD, and this seems like a pretty darn good idea given how you can shoot a bunch of frames on cloudy nights when there's nothing else to do.
Shooting the bias frames takes a matter of seconds, so those aren't a big deal at all, even when you're cold and cranky in the early hours...
I know that aperture, ISO, and temperature are all important (and I hope to be able to use the bias frames to scale the darks, that's why I didn't list exposure time).
I was playing with my Astrotrac the other night and could manage 5-5.5 minute exposures at 200mm on a crop Canon body with nice round stars.
Assuming that'll be my longest exposure time with any lens, can I just go ahead and start shooitng darks when I have the time at night, with the appropraite ISO, and keeping track of the temperature?
How important is focal length? And focus? DEpeiding on the seeing and temperature it will change slightly...is that minute difference at all important for the darks? If my longest exposure with any lens will be five minutes, do I need to shoot variable ISOs and focal lengths at roughly 5 degree (Celsius) intervals as Jerry recommends? Or can I just base my darks on temperature and ISO, using bias frames shot with the lights to scale the darks to the correct exposure length?
Posted 06 December 2012 - 10:26 AM
Posted 06 December 2012 - 11:16 AM
Posted 06 December 2012 - 11:36 AM
I've also put my camera into the refrigerator with varying layers of ice packs and cloth padding to manufacture a variety of temperatures. I've also let it run for days in a dark shed so that it gets the full range of day and night temperatures.
Posted 06 December 2012 - 08:44 PM
So I can just stick on my most used lens, aim it in such a way as to approximate the kind of heat loss I'd get while imaging, choose an ISO I'll use, and then let it run for 5 minute exposures (since that's what I figure my longest exposure will be)?
I assume DarkMaster or DarkLibrary then scane the dark frame library and picks those frames which most closely match the lights.
I downloaded ExifTools, which showed the tempratures of all my RAW files (I'm using a 40D), and I've downloaded DarkMaster as well. Cheers, Barry!
Will DSS automatically scale the darks to the right time once they've been added?
Posted 07 December 2012 - 01:11 AM
For my use, I have a windows 7 bat file that runs exiftool and writes the ccd temp out to a tag which I can use to sort my lights and darks. Makes it easy to see the progress over the course of a night.
Posted 07 December 2012 - 07:44 PM
The weather here has been totally overcast for the past few weeks, and looks to stay that way for a few more, but as soon as I have a chance I'll have to give all of this a try.
Posted 08 December 2012 - 04:19 AM
Posted 09 December 2012 - 03:02 AM
Posted 09 December 2012 - 04:02 AM
Here's what it looks like:
@echo off exiftool -r -P "-cameratemperature+>keywords" -ext .JPG -ext .DNG .
I put a copy of EXIFTOOL.exe in the same folder and double click the bat file. As written, it will scan recursively through sub directories, which is nice to index a whole collection of shots. Be aware that it will keep a backup of the original file so there's some need of extra storage space.
Note that your camera maker/model would have a tag for "cameratemperature" that may not be the same as mine - or may not even be used. You may wish to check what the tags are by using PhotoMe - another piece of EXIF reading software.
Hope this helps!
Posted 09 December 2012 - 08:37 PM
Posted 30 March 2014 - 10:07 PM
Posted 31 March 2014 - 12:23 AM
You need to create VERY, VERY CLEAN master bias frames. There are a few threads on the PI forums that show how stacking 30-50 bias frames just doesn't get you a truly clean result, where as stacking 160-200 gets you a nearly perfect result that shows the fixed pattern noise that needs to be removed from darks before scaling. I actually experimented with this myself by stacking 10, 30, 70, and 100 bias frames in PI...the 100 was definitely cleaner than all the others, but still not as clean as 160-200 frame master biases I've seen online. You want them CLEAN, because they are first used to calibrate the darks before scaling, and if they aren't as clean as possible, you'll introduce more random and semi-random noise.
MaxIm DL can also do dark frame scaling, in pretty much the same way, although I have not tried MaxIm yet.
For PixInsight, you can find the details here: http://pixinsight.co...ames/index.html
Now, granted, PixInsight is not free. Nor is MaxIm DL. If you don't want to spend $250-$600 on one of these two tools, then you'll have to stick with DSS. DSS can do some dark optimization, but it does not scale well. Even a few degrees difference affects your final results, especially with DSLRs and warmer temperatures.