Jump to content


Photo

Riddle me this, please. re: multiple exposures

  • Please log in to reply
9 replies to this topic

#1 ponz

ponz

    Vostok 1

  • -----
  • Posts: 144
  • Joined: 18 Jul 2012
  • Loc: Kansas City, MO

Posted 09 January 2013 - 10:24 AM

Imaging requires x amount of subs, darks and flats.

The darks entail shooting multiple exposures into the back of the lens cap.
The lights? entail shooting multiple exposures into a white cloth.

Why can't you simply shoot one dark, for example, and "duplicate" it several time in, let's say Photoshop?

Why not?

If someone would please explai this in layman's terms that would be great.

John Ponzo
aka - Ponz

#2 NeilMac

NeilMac

    Gemini

  • -----
  • Posts: 3062
  • Joined: 25 Sep 2010
  • Loc: MedHat, AB, Canada

Posted 09 January 2013 - 11:04 AM

I think the point in regards to blacks is for the elimination of hotspots in the sensor by comparing several black samples to generate a generalization of an average black background.
Ive edited frames in PS and DSS rejected them.

#3 jrcrilly

jrcrilly

    Refractor wienie no more

  • *****
  • Administrators
  • Posts: 33724
  • Joined: 30 Apr 2003
  • Loc: NE Ohio

Posted 09 January 2013 - 11:06 AM

Why can't you simply shoot one dark, for example, and "duplicate" it several time in, let's say Photoshop?


The dark frame has two kinds of information in it. It has the desired information (differences in black levels of the various pixels) and that will appear in each dark frame captured. It also includes random noise from the process of reading the chip and that will be different for each frame captured. If a single dark frame is used, that random noise is injected into the target frame so while one form of error is reduced, another is added. Combining a number of dark frames can cancel most of that random noise so the master dark frame can do what is desired (adjust for warm pixels) while introducing a minimum of new noise into the image.

#4 JoLo

JoLo

    Ranger 4

  • -----
  • Posts: 323
  • Joined: 09 Feb 2010
  • Loc: Highland, IL

Posted 09 January 2013 - 11:10 AM

Because certain aspects of noise is random, simply duplicating a dark frame does not do the trick, you will just get the same noise repeated frame after frame. Also, dark frames introduce noise of their own, which must be subtracted out of your light frames.

Darks - cap on, same ISO and duration as your lights, and same temperature.
Flats (lights) - you can do a white cloth at dusk or dawn..the object here is to have a evenly illuminated frame in which the "hump" of the histogram is 1/3 to 2/3 of the way off the left side. I use a DIY light box, which gives even illumination, typically a 1/8 to 1/4s exposure.

Some people take many darks and flats. I have read, and adhere to the notion that after about 15 darks and flats, the return on additional frames decreases rapidly. I usually take 7 darks before imaging, and 8 when finished. I do the flats after imaging, as the focus must not change when taking flats.

#5 ponz

ponz

    Vostok 1

  • -----
  • Posts: 144
  • Joined: 18 Jul 2012
  • Loc: Kansas City, MO

Posted 09 January 2013 - 03:51 PM

So.....If I shoot the back of the lens cap multiple times with the exact same exposure/length/ISO, etc.....the noise captured will differ each time?

I'm not suggesting utilizing one dark frame, but utilizing "many copies" of an initial dark frame to eliminate the need/time photographing the back of the lens cap over and over again.

Is the imaging software smart enough to know it's tweaking multiple copies of one frame? After all, the exposure time and ISO are to be the same for each frame anyway, right?

I'm just trying to eliminate the need, time and wear and tear on all the equipment...

Ponz

#6 JoLo

JoLo

    Ranger 4

  • -----
  • Posts: 323
  • Joined: 09 Feb 2010
  • Loc: Highland, IL

Posted 09 January 2013 - 04:07 PM

You have to shoot individual frames, they have to be averaged into a master dark frame, which is then used to calibrate your stacked images of the object you are photographing. Same is true of the multiple flat field you should take.

What software are you planning to use? Yes, astronomy software is smart enough, and then some. Stacking and calibration of astronomical images requires specialized software.....I use MaximDL, but there are many options out there, some are free, some are not.

As mentioned above, if you don't take these steps you will have a noisy, messy image that you really can't process effectively. The darks will remove hot pixels and average out noise, the flats will help with uneven field illumination and dust on your optical surfaces. To get quality astrophotos, you cannot eliminate the need, time, wear or tear.

I'd strongly recommend you read up on acquiring, stacking, calibrating and processing astronomical images before attempting this. There are many good books out there, and lots of information on the internet.

#7 Peter in Reno

Peter in Reno

    Fly Me to the Moon

  • *****
  • Posts: 5784
  • Joined: 15 Jul 2008
  • Loc: Reno, NV

Posted 09 January 2013 - 04:17 PM

I had the exact same questions when I first started this hobby. I know you are trying to save time but to get good quality images, you need to spend extra time following proper procedures. Like others said, the more darks, bias and flats, the better. Also more lights are good. The more subs, the less noise and greater details in final processing.

Take a look at this good and simple to understand reading:

http://deepskystacke...lish/theory.htm

Peter

#8 Mike7Mak

Mike7Mak

    Apollo

  • *****
  • Posts: 1277
  • Joined: 07 Dec 2011
  • Loc: New York

Posted 09 January 2013 - 05:01 PM

So.....If I shoot the back of the lens cap multiple times with the exact same exposure/length/ISO, etc.....the noise captured will differ each time?

Yes exactly. Every pixel returns a slightly different value each time even if all else is kept identical. That means that the 'average' of more than one pixel will be closer to the 'true' value. The more pixels averaged the more accurate the result. That variation that you're trying to average out is considered noise.

If you just make multiple copies of the same frame there is no variation to average. For example if a pixel in 5 different frames delivers values of 1, 2, 3, 2, 1, the average value of that pixel is 1.8 (9/5). If you just made 5 copies of the middle frame (pixel value 3) then the 'average' is 15/5 or 3. There is no point in stacking multiple copies of the same frame.

#9 ponz

ponz

    Vostok 1

  • -----
  • Posts: 144
  • Joined: 18 Jul 2012
  • Loc: Kansas City, MO

Posted 10 January 2013 - 08:42 AM

OK - I get it. I couldn't, initially, accept the fact that the imaging software was that smart.

My equipment arrives today. Alas, the one thing I neglected to buy was a way to power it up! I bought a deep cycle marine battery, power outlets, etc....the other day; but have nothing to get me started tonight!

Ponz

ps - that is, of course, once I get it all assembled

#10 Alex McConahay

Alex McConahay

    Vanguard

  • -----
  • Posts: 2303
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 10 January 2013 - 10:01 AM

A few notes regarding this thread in case some newbie is reading and learning from it......

1. The term "lights" has been confused with the term "Flats." (As in--from the original poster: "The lights? entail shooting multiple exposures into a white cloth." )

"Lights" are the actual exposures of the targets. The telescope is uncovered, pointed at the object. If you take a "light" you will see a celestial object.

"Flats" (what the original poster called "lights") are taken with a cloth pointed at a bright sky. If you take a "Flat" you see a fairly uniform whitish frame, with perhaps a bit of darkening in the corners (vignetting) or faintly dark roundish spots (dust motes).

"Darks," as the OP had, are taken with the front of the camera (scope or lens) closed (and the viewfinder if using a DSLR closed also). You will get a black field, with perhaps defective pixels here and there of gray or white, and perhaps areas of the sensor slightly different due to thermal issues in the camera. (And perhaps the random gamma ray showing up white.)

2. The answer to the original poster's question regarding why one cannot use multiples of one exposure instead of taking multiple independent exposures and then averaging has been well answered. The key is that any given exposure is slightly different from another. Some of it is the same, true: A white hot pixel or a completely dead pixel will be the same in every shot (perhaps). But because of random quantum variations, gamma rays, and other anamolies, most pixels will be ever so slightly different from one exposure to the next, as has been well explained. That is why you need multiples, and need to average them out.

Alex






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics