Jump to content


Photo

Q: Lenght of sub-exposures

  • Please log in to reply
10 replies to this topic

#1 Jeff Esposito

Jeff Esposito

    Sputnik

  • -----
  • Posts: 30
  • Joined: 24 Sep 2007
  • Loc: Omaha Ne

Posted 15 December 2013 - 08:43 PM

Are there any advantages or disadvantages to using 6 10-sec sub-exposures as opposed to a single 60-sec exposure?

#2 Madratter

Madratter

    Fly Me to the Moon

  • *****
  • Posts: 7140
  • Joined: 14 Jan 2013

Posted 15 December 2013 - 08:47 PM

Yes, but the disadvantages are somewhat hard to explain without getting technical. Basically, the very act of trying to read the data from the sensor introduces a certain amount of noise. And that amount of noise is the same for a very short exposure and a long exposure. So by doing lots of little short exposures, you end up with more noise.

Advantage wise, the short exposure is much much less demanding on the mount.

#3 bseltzer

bseltzer

    Apollo

  • -----
  • Posts: 1028
  • Joined: 28 Oct 2007
  • Loc: East S.F.Bay, CA

Posted 15 December 2013 - 10:22 PM

It's a bit of a balancing act between thermal noise, background noise (AKA seeing/light pollution), and read noise. As exposures get longer the first two accumulate, which argues in favor of shorter exposures. But, as MR pointed out, every time you download an image, you add to the read noise. So the ideal exposure length depends on the inherent thermal noise of whatever chip you're using, the seeing at your location, and the read noise inherent in your camera's electronics. If all 3 can be quantified, there are formulas that can give you an idea of 'ideal' exposure times. There's a fairly detailed discussion of these factors here.

#4 Jared

Jared

    Fly Me to the Moon

  • *****
  • Posts: 5077
  • Joined: 11 Oct 2005
  • Loc: Piedmont, California, U.S.

Posted 16 December 2013 - 10:33 AM

Are there any advantages or disadvantages to using 6 10-sec sub-exposures as opposed to a single 60-sec exposure?


Under almost any sky conditions with almost any current camera, the 60-sec exposure is going to give better results. With short exposures like these unless you are shooting during the day read noise is the dominant source of noise in your image. Lots of short exposures have more read noise than fewer long exposures. Once your exposures are long enough that light pollution noise dominates over read noise, the situation reverses and you are better off keeping the exposures shorter. There are calculators available to let you know how long an exposure you need for this to happen--check the Starizona website for one of them.

My answer assumes that your tracking is good enough to provide round stars on a sixty second exposure. If that's not the case, then you are better off shortening your exposure times until you most of your subexposures produce good tracking. It's fine if you are throwing out the occasional outlier.

#5 CounterWeight

CounterWeight

    Cosmos

  • *****
  • Posts: 8237
  • Joined: 05 Oct 2008
  • Loc: Palo alto, CA.

Posted 16 December 2013 - 10:40 AM

I think it a bit more complex with DSLR as you can play with your ISO and F stop as with normal terrestrial exposure 'bracketing'. I'd ask on the DSLR imaging forum for your specific DSLR too. Focal ratio plays a huge part in imaging and in my very short experience so far my issue with the Canon 6D has been keeping the histogram in the 1/3rd from right ;)

Part of it all too is what your are imaging. Point sources (stars) expose quickly, so if you were going for open cluster like the Beehive or coat hanger or double cluster you don't really need a long exposure. Objects with faint nebulosity like exposure time with the caveats mentioned in the above posts.

In the last several months a new technique has been mentioned / suggested / shown / demonstrated that sort of affects it all as well, called 'massive dithering' which shows a lot of promise and changes a little about what I think considered basics. It requires moving the mount between exposures and some interaction between the imaging software and the mount control / auto-guiding. I know BackyardEOS has PHD (a popular free auto guiding platform) dithering integrated into the program / package. Dithering helps the post processing software remove pattern noise, hot pixels, ?

#6 bseltzer

bseltzer

    Apollo

  • -----
  • Posts: 1028
  • Joined: 28 Oct 2007
  • Loc: East S.F.Bay, CA

Posted 16 December 2013 - 12:29 PM

Jared,

I can easily understand what you're saying about the 6' X 10 vs. 60' X 1 example. Does the answer remain the same, however, if we're talking about, say, 600' X 6 vs. 1200' X 3 in a red zone. Also, how much of a role do bias frames play in controlling read noise?

I ask because I do a fair amount of imaging under truly ugly skies in terms of light pollution, and I've always been a bit foggy about bias frames.

Thanks.

#7 pdfermat

pdfermat

    Apollo

  • *****
  • Posts: 1075
  • Joined: 12 Nov 2007
  • Loc: Wisconsin

Posted 17 December 2013 - 12:10 AM

Here's 3 images of M1, all are 30 minutes of total integration. It's not a perfect comparison (different nights, slightly different conditions, different number of darks, processed as similarly as possible), but you can get some sort of idea.

Here's a stack of 10 3 minute subs:

Posted Image

Here's a stack of 6 5 minute subs:

Posted Image

Here's a stack of 3 10 minute subs:

Posted Image

#8 jbalsam

jbalsam

    Viking 1

  • *****
  • Posts: 519
  • Joined: 06 Jul 2012
  • Loc: Spencerville, MD

Posted 17 December 2013 - 09:22 AM

Bert -

There have been theses written about this question of image SNR. The best answer that I have seen in these forums and that I can offer you is: just test it for yourself. Take 20 min and 10min and 5 min frames of an object on the same night and process them with and without calibration frames. That will show you exactly what the difference are for your equipment and with your skies. No approximations needed.

As far as Bias frames go, I don't use them and I know a lot of people don't use them. That's appropriate if you are using a temperature regulated CCD and if your darks match your lights exactly (exposure time and temperature). The only time bias frames are required is if you are doing dark frame scaling (i.e. your darks don't match your lights but you want to use them anyway). Then the darks are bias-subtracted and the remaining signal is scaled linearly until the residual of its subtraction from the lights is minimized.

-Josh

#9 Ron (Lubbock)

Ron (Lubbock)

    Messenger

  • -----
  • Posts: 447
  • Joined: 17 Aug 2012
  • Loc: West TX

Posted 17 December 2013 - 01:31 PM

Sub length does indeed matter. For brighter objects like M42, short subs may work well. For a faint object like an 11th magnitude spiral galaxy, longer subs are needed. In addition, the longer subs go, the more important sensor cooling becomes.

#10 bseltzer

bseltzer

    Apollo

  • -----
  • Posts: 1028
  • Joined: 28 Oct 2007
  • Loc: East S.F.Bay, CA

Posted 17 December 2013 - 04:40 PM

Thanks to all. Your examples and explanations have been a big help. My dark frames are always set to match my lights for temperature and duration, so it sounds like I don't need bias frames, just my darks and flats. It'll be interesting to go back and re-process some old data without bias frames and see what, if any, difference it makes. And Pat, your example stacks will definitely make me re-evaluate the exposure times I've been using. Thanks.

#11 Madratter

Madratter

    Fly Me to the Moon

  • *****
  • Posts: 7140
  • Joined: 14 Jan 2013

Posted 17 December 2013 - 08:27 PM

If you read the PixInsight calibration page:

http://pixinsight.co...-frames/en.html

they believe Bias frames matter even with pretty closely matched darks and lights. I have collected the bias frames to give it a try. I have to admit, I don't see a lot of difference in their examples.






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics