Jump to content


Photo

5 minute or 30 minute exposures for broadband?

  • Please log in to reply
5 replies to this topic

#1 Mike Wiles

Mike Wiles

    Viking 1

  • -----
  • Posts: 950
  • Joined: 04 Feb 2009
  • Loc: Goodyear, AZ

Posted 30 January 2013 - 01:38 PM

Hola,

To piggyback on what has been an excellent discussion on signal to noise ratio vs exposure length in narrow..., I thought I'd start a separate topic to apply the discussion to normal LRGB or OSC imaging as it's an entirely different ball of wax than narrowband imaging. As with my post in the other thread, please allow me to start by demonstrating the depth of my ignorance.

In the other thread there has been significant discussion about shot noise limited vs read noise in narrowband images. I think we can all agree that drowning the read noise in the sky background is extremely difficult if not impossible in a narrowband image. However, it's not at all difficult to do in a normal broadband image. I think that we all agree that if it's possible to bury the read noise in the sky background, that it's always a good idea. So my overall question for this discussion is this: Is there any true benefit to shooting a longer subexposure once the read noise has been buried in the sky background?

The same general principles apply. At a less dark site, sky glow will be brighter and will cause the sky background to bury the read noise sooner. As a result, it'll take a short subexposure to get to that point. Similarly, at a dark site a longer exposure is required to overcome the read noise because the sky background is recorded at a slower rate. All other things being equal - the benefit of shooting longer at a dark site is that fainter detail will be recorded than you'll get from a brighter site. Shooting longer at a less dark site will simply cause the background to grow brighter instead of recording fainter detail. Looking at that logic from another direction it seems to imply that once you overcome the read noise in a broadband image you've also established the faintest details that you're going to record with a single subexposure. If you want to go fainter, your only option with the same equipment is to shoot more subexposures to eliminate noise via dithering and rejection. Or....move to a darker site. Shooting any longer simply begins to chop off the dynamic range of the camera as the sky background pushes the histogram to the right.

Note - this is all as I understand it and I put it out here because I'm likely wrong for the 37,899th time in my life, and I'm just trying to fully wrap my brain around it.

Put another way....let's shoot two images from the same site with the same equipment:

Image 1: 16 subs x 10 minutes each
Image 2: 32 subs x 5 minutes each

If we assume that both images are shot noise limited....which stack will go deeper and show fainter detail? Is it image 1 because the subs are longer? Or is it image 2 because the subs are "long enough", but there's less noise because of the increased number of subs for data rejection?

Mike :gotpopcorn:

#2 vpcirc

vpcirc

    Soyuz

  • -----
  • Posts: 3962
  • Joined: 09 Dec 2009
  • Loc: Merced CA

Posted 30 January 2013 - 02:51 PM

Mike what you're saying is exactly what I was told!

#3 Rick J

Rick J

    Fly Me to the Moon

  • *****
  • Posts: 5659
  • Joined: 01 Mar 2008
  • Loc: Mantrap Lake, MN

Posted 30 January 2013 - 02:55 PM

It isn't a whole new ball of wax. The concept is the same. You want to expose as long as possible without saturating bright detail or if that is unavoidable then use long exposures for the faint and short for the bright. Bright already has a good S/N so doesn't need the long times that faint does. Still it saves a lot of work if your camera has the dynamic range needed to do this in one exposure.

It isn't much of an issue if the brightest stars saturate. There are ways in processing to preserve their color and to shrink them down to size.

For most objects for my system I need at least 30 minutes with narrow band and 10 with L and 20 with RGB. Since RGB data can survive a lower S/N than L I usually use 10 for RGB as well though there have been a few times I should have used 15 to 20.

Exposure is always a balancing act between saturating bright detail and getting enough for good S/N in the faint detail you wish to capture. Once you reach sky limit there's no significant gain in going longer!

Cameras with deep wells and low read noise can make a big difference. Low read noise means you reach sky limit faster, often fast enough not to saturate much. Large wells help by holding more electrons before saturation sets in.

Well depth is usually related to pixel size. A 9 micron pixel will have a larger capacity than a 5 micron one but it also captures photons over a larger area so captures them faster. Focal length and aperture along with well size come into play here.

It's all a major balancing act dependent on a particular system as a whole as to what the answer is. No one or even 10 answers cover all situations.

You want long enough exposures for burying read noise in sky noise (dependent on camera and your sky conditions as well as focal length, aperture and even binning) yet short enough you don't burn in bright detail you wanted to retain.

You can compensate with two exposure times when necessary (best) or by taking many more (total time) short exposures and stacking them though this latter method soon reaches a limit where more is of little help.

It is harder to saturate or reach sky limit with narrow band. That's the only practical difference. This puts more demand on polar alignment and the mount as a whole.

Every aspect of this hobby is a compromise. We all make a zillion of them. What's acceptable to one is unacceptable to another. You "just" have to understand all that is involved and make the compromises you can live with. Acquiring this understanding is the hard part and a bit of math.

Rick

#4 shams42

shams42

    Apollo

  • *****
  • Posts: 1108
  • Joined: 05 Jan 2009
  • Loc: Kingsport, TN

Posted 30 January 2013 - 03:41 PM

If the exposures are sky limited, there will be no difference in SNR between 16x10 vs 32x5. The SNR will be determined by total integration time.

#5 Mike Wiles

Mike Wiles

    Viking 1

  • -----
  • Posts: 950
  • Joined: 04 Feb 2009
  • Loc: Goodyear, AZ

Posted 30 January 2013 - 04:26 PM

If the exposures are sky limited, there will be no difference in SNR between 16x10 vs 32x5. The SNR will be determined by total integration time.


So then....is it safe to make the inference that once the exposure is sky limited, you've reached the limit for faint detail in a single subexposure for that setup under those conditions? If that's the case (and it seems to my mind that it is...) then going beyond that point is inviting trouble with satellites, airplanes, wind gusts, increased dark current and any of the other things that can ruin a subexposure.

Put another way, once the exposure is sky limited - does the sky background now place a limit on how deep you can go with a single subexposure?

Mike

#6 neptun2

neptun2

    Viking 1

  • -----
  • Posts: 796
  • Joined: 04 Mar 2007
  • Loc: Bulgaria

Posted 30 January 2013 - 05:11 PM

Well i understand it this way. As the exposure time grows you capture fainter and fainter details up to a point where the sky background and other noise come to a certain level which will mask any fainter details. There is no point to go longer because this will not help.

My general rule is as follows. Go as long as possible but considering 2 things. First be careful not to overexpose the object especially if it is too bright. Then be careful not to make the background too bright. On the most places where i take pictures for faint objects this means from 5 to 10 minutes per subexposure with my equipment. Of course it also depends on what your mount will allow but let's eliminate this factor for simplicity.






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics