Taking pics in short bursts

[Moonshed]

Hi, I am waiting for delivery of my Canon EOS1100D so I can get started on digital photography. I have been reading a lot of what people are saying  on this forum and being new to digital I have just got myself confused. If I understand it correctly if a deep sky object needs say 16 mins exposue you break it down to say 8 x 2 min exposures with a 4 min interval between each. I believe this a dual function, it allows the chip to cool down and the camera can take dark frames in the pauses, and because the mount cannot track accurately enough for longer periods. Okay with that. What I don't understand is that if this results in 8 different frames which are then stacked, how does the detail build up? If it is all built up in a single frame then how does that prevent star trails? I am so confused. With the old film camera the shutter was left open for 16 mins collecting all those photons and building up the detail, but this way we have 8 identical frames. What am I misunderstanding?

[SilverAstro]

30 minutes ago, Moonshed said:

If it is all built up in a single frame then how does that prevent star trails?

The software that you use to stack the images ( such as DeepSkyStacker) moves each frame (registers them) so that all the stars end up on top of each other.

[Moonshed]

25 minutes ago, SilverAstro said:

The software that you use to stack the images ( such as DeepSkyStacker) moves each frame (registers them) so that all the stars end up on top of each other.

Thanks for your reply. I understand the stacking part, I have been using Registax for about a year to stack videos into a single image when taking close up videos of the Moon using an Orion planetary camera. What I don't understand is how stacking what is basically a number of identical 2 min frames end up with a far better image than any individual frame.

[Bizibilder]

Its all to do with signal to noise.  Each frame has a signal (stars etc) and noise (background random variation).  When you stack the images they add up.  The clever bit is that when you "add" noise, it averages out to a very low value (but not quite zero  ). So you get a final stacked image with your enhanced signal and, hopefully, very little background noise.

If you only take "one long exposure" you get the signal and all the noise.

To prevent trailing your mount/scope has to be driven at the same rate that the Earth rotates (i.e. one rev per day or "siderial rate").  If this is done the camera "sees" the exact same part of the sky all the time and you do not get trailing.  Unfortunately there is no such thing as a "perfect" mount so, for longer exposures, you have to add guiding to the set-up.  This simply "locks on" to a guide star and keeps the scope pointing precisely at your target.

By the way if you are taking 2 min exposures with an 1100D you only need about 5 seconds between exposures - anything longer is wasting good imaging time!  Additionally you can take your dark frames at any time you like as long as they are taken at the same temperature as your lights.  I have built up a library of darks at roughly every 5°C that I use over and over with different light frames.  The exposure time of the darks must be the same as the lights of course.

For the same time-saving reason make sure that you have "dark frame compensation" turned off or your camera will take darks for you immediately after each light - thus wasting more imaging time!!

[SilverAstro]

47 minutes ago, Moonshed said:

Thanks for your reply. I understand the stacking part, I have been using Registax for about a year to stack videos into a single image when taking close up videos of the Moon using an Orion planetary camera. What I don't understand is how stacking what is basically a number of identical 2 min frames end up with a far better image than any individual frame.

Ah ok, well previously you asked " how does the detail build up? If it is all built up in a single frame then how does that prevent star trails? " - which is a two-parter, detail and trailing,  and I dealt with the second part first, cos that is the easy bit

Right then, how the detail builds up > The star detail is held as a digital number in each 'slot' in each frame, after registration all the slots for each star correspond 1 for 1 in each of the frames and are added together and build up kind of quickly. All the slots between the stars hold just noise and when they are added to each other they dont build up as quickly as the stars do. We should now go to the statistics of coherent signals (the star signals) and random noise (the rubbish between) to prove this !  But I'll leave you to go research the maths ,  the 'in phrase' is Signal-to-Noise ratio, abbreviated to S/N. It isnt quite as simple as that ( never is is it ! ) because there are different types of noise but -

- to a first approximation think of it as the stars adding up and the noise smearing out,  so the S/N ratio of the stars to the noise gets better  Howzat ?

Just to complicate things a bit more Registax with videos is ( a little bit ) different from DSS with longer exposures.

EDIT : I added the (little bit ), else an enormous great argument might ensue amongst the mathematicians lurking

[Moonshed]

Thank you guys, I think I understand it now it has been explained. I understand that periodic errors become a major problem with this type of photography, whereas with a film exposure the finished image was never clear enough to show the drift, well, it was there obviously, but didn't show up too bad because the finished image was not that brilliant anyway. Mine weren't anyway. Looking forward to getting starting, my wife has braced herself for a few weeks of expletives blasting out from the Moonshed, and words to the effect that the bloody manuals are all wrong and the camera doesn't work...etc.etc. All part of the joys of Astrophotography!

Thanks again, always get great info and advice from this forum, I would be lost without it!

[SilverAstro]

10 minutes ago, Moonshed said:

I think I understand it now it has been explained.

Thanks again, always get great info and advice from this forum, I would be lost without it!

Yep, it is good isnt it,, and I'm tempted to say " , come back and explain it to us just to make sure we do "

I see that I was typing away merrily at the same time as Roger, good to see that we more or less agree ! except that I was assuming a different type of trailing !

By the way there are some reasons why one single exposure can be better from the S/N point of view than subdivided exposures, depending on where the noise is from, but perhaps we'll leave that for another day. There are lots of reducto discussions on the forum about, for example, 1000 x 1sec compared to 1x 1000sec  but hey my evening meal is getting cold Good luck