Question about High ISO.

I was recommended to use the highest ISO possible when taking long exposure shots and using star stax by a fairly awesome photographer. I'm still trying to wrap my head around why this is? I understand that if I use too long of an exposure the stars or moon will start to move, I just don't understand how the pictures won't be complete noise? Will exposure stacking in post help clean this up? Thanks in advance for any help/advice.

You should PM Martin Wise (MalVeaux)

https://photography-on-the.net/forum/member.p​hp?m=372251

He's an excellent, very knowledgeable astro shooter. He'll be best to answer your question.

I have heard the same thing more than once but I'm not an astro guy so I've never fooled around with it.

Stacking removes random noise while keeping constant signal. The more you stack, the cleaner it will become as random noise is removed and the constant signal you keep stacking over and over will strength the signal to noise ratio, allowing you to pull more information from the signal you're capturing. Stacking increases your signal to noise ratio by the square root of the number of images stacked. That said, you get more signal from a longer exposure than a shorter exposure, and at the highest ISO possible you start losing out on the quality of that signal with really short exposure times. Basically, you're a lot better off at 30 seconds and ISO 1600 than you are at 2 seconds and ISO 25,600 (exposure is equivalent, but the longer exposure will yield greater signal quality). So this is my point about stressing that you should use the longest exposure time you can tolerate for your focal length and situation, then increase ISO until you get the histogram you need. It's possible to hit the maximum ISO to get there, but with wider focal-lengths it shouldn't be.

You don't have to use the highest ISO possible. You want to use the best combination of exposure time and ISO for your camera and your focal-length relative to the exposure you're trying to make (I assume this is static from a tripod, not on a tracker, so likely wide field like milky way). There's also a different histogram you're looking to achieve. You select the ISO needed to get you to that histogram with the exposure time you're limited to. There are more in depth ways to approach ISO and noise, such as read noise, etc, but if you're not on a tracker and you're limited to 30 seconds or less exposures from a static tripod, don't worry about that, and just focus on getting a good exposure and lots and lots of them to stack. It also matters what kind of light pollution you're working with in terms of how you want to expose your histogram (dark sky is different, from a well polluted street lamp infested sky as you hit a limit real fast under polluted skies).

+++++++++++++

Exposure & histogram information (this is your bread & butter to get started and directly addresses what ISO you should be using based on the histogram):

https://jonrista.com …nal-noise-and-histograms/

All the information you may ever need to know about this subject (this goes into detail about why stacking is used and its benefits):

http://www.clarkvision​.com …s/image-stacking-methods/

One of the best tutorials for processing DSO in general (milky way to DSO) from color dSLR files (this just teaches you how to process the night sky and draw detail from high signal to noise ratio final stacked images):

http://www.astronomers​doitinthedark.com/dslr​_llrgb_tutorial.php

+++++++++++++

So here's an example of my own that I learned from the above links and put into practice.

Using a Canon T4i (650D) at ISO 800 and my exposures are 240 seconds long (I'm on a tracker) with a 480mm F6 telescope (I'm shooting at F6 which is why I need that long exposure time, plus, longer exposure time means better signal quality). At my exposure time, I only needed ISO 800 to get the histogram I wanted, under a fairly dark sky where I am (minimal pollution) with a histogram that was around 1/3rd to 1/4th from the left or so. I did 21 exposures like this (21 x 240 seconds = 84 minutes of total exposure time). I only got 21 frames to work with, at longer exposure time, so my signal to noise ratio increase for doing 21 exposures is about 4.6 times more signal than I would have had in a single exposure un-stacked. I would have benefited from doing more, but an hour and a half basically of time just for the exposure was where I was comfortable at with the weather.

Here's the resulting stacked image results at 21 x 240 seconds through F6 and ISO 800 with the Canon T4i. Note the hisotgram (closer to 1/4th from the left, would have been better probably closer to 1/3rd from the left). You can see stars, but not much else. That's how its supposed to be. The real detail is buried in the signal that I'm stacking to get more of so I can stretch the histogram to reveal it. Because I stacked several images, the ISO noise is pretty much gone or at least not noticeable.

IMAGE LINK: https://flic.kr/p/RnZh​v5

IMAGE LINK: https://flic.kr/p/RgsU​H1

Thank you so much for taking the time to explain and show examples.... this is exactly what I was wanting to know. I'm still reading up on the links you gave me. Thanks again!

wantafastbusa wrote in post #18503007
Thank you so much for taking the time to explain and show examples.... this is exactly what I was wanting to know. I'm still reading up on the links you gave me. Thanks again!

If you're shooting wide angle from a tripod with an intervalometer and doing wide field astrophotography, like milky way or large constellations or groups even, you shouldn't need much. You can do it any where from F1.4 to F4 and still not likely cap out on your ISO unless your camera is over 12 years old. Even ancient 10 year old cameras with ISO 1600 caps can still do a good job here, because again, you can take several exposures and stack the results to remove random noise. You don't actually want it to be completely cleared out, because it will look unnatural without some minor noise, as in reality, there's stars virtually everywhere in the exposure, they're just too feint to show up until you expose for a very, very long time, or stack a ton to get enough signal to keep them (and differentiate them from random noise).

An easy way to start:

Setup, use Live View (or equivalent) to focus on a bright star in manual, and lock it in manual, at 10x magnification if possible.
Now, compose based on what you want in your FOV with some wiggle room on the edges.
Set focal-ratio for performance (coma handling, CA, etc). Depends on your lens. Faster helps, but it's ok to stop down a little to clear up the CA or calm down coma.
Exposure time is as long as you can tolerate before star trails.
Then start at a high ISO, 1600~3200 maybe, and see where your histogram is with an exposure.
Adjust ISO until you're at about 1/3rd histogram spiking from the left, or between 1/3rd and 1/4th.
Once you have that, start exposing over and over and get as many frames as you can.

If you really get into it, I suggest you start shopping for a small tracking mount, like the Sky-Watcher Star Adventurer ($300ish). It removes the limitation of exposure time from seconds to minutes which really opens doors. You'll see people spending hundreds to even thousands, trying to get a fast focal-ratio prime lens and a modern full frame camera, just to try to do milky way shots. Yet, a $300 tracker is all you need, then a 10 year old (or whatever it is now) Rebel XSI at ISO 1600 with a vintage adapted 50+ year old lens will produce even more signal because of the lack of limitation of exposure time. So don't get trapped into the buying of fast stuff and sensitive ISO cameras, etc. Just get a tracker (if you really want to get into this).

+++++++++++++++

Removing the limit of exposure time really helps get signal though. Say you don't have an hour to devote to grabbing lots of frames and you want to get a deep wide field exposure. Doing it in a few minutes from a single exposure can be done easily with a tracker.

Here's a cheap tracker with a little Rebel and a 35mm prime lens:

IMAGE LINK: https://flic.kr/p/nrWd​9W

IMAGE LINK: https://flic.kr/p/kRz9​wv

I'm not sure if I will jump into buying a tracker right now, living in Phx, where it is so noisy. I won't be able to venture out too far to find a cleaner area to give this a go. I have a canon 10-18mm which I know won't cut it. I was hoping to find a good deal on a rokinon 14mm. My FOV will be pretty narrow with my nifty 50, but I was going to start out with that to see what kind of luck I have.

......the pipe......nice touch.

wantafastbusa wrote in post #18503225
I'm not sure if I will jump into buying a tracker right now, living in Phx, where it is so noisy. I won't be able to venture out too far to find a cleaner area to give this a go. I have a canon 10-18mm which I know won't cut it. I was hoping to find a good deal on a rokinon 14mm. My FOV will be pretty narrow with my nifty 50, but I was going to start out with that to see what kind of luck I have.

You should still be ok with the 10-18 STM. 10mm at F4.5. You can go 30 seconds at 10mm on APS-C before trails become obvious. F4.5 is not that slow for this, especially not for wide field. Start around ISO 3200~6400 and adjust as needed to get the appropriate histogram.

As I was mentioning, you're talking about buying a 14mm F2.8, only to gain 1 stop of light. That's all you gain, and it's not that wide on APS-C. But spending $200~250 on a lens like that, when a $300 tracker removes the limitation of exposure time from seconds to minutes, the better buy for this is the tracker. Though I know of course the 14 F2.8 may be useful to you outside of astrophotography. But the point is, if you're going to try it from wherever you are and won't be venturing out far to get darker skies, all the more reason not to buy a new lens for the job and to get a tracker instead.

You can go about 6 seconds with the 50mm. This will require very high ISO and lots of exposures to stack to get enough signal to see much other than stars (if you want to start seeing DSO other than the biggest brightest ones like Orion and Andromeda).

But really, don't buy anything at all, just use what you have using the tools in the links above. You'll get results. But I can't stress enough, don't buy a lens just for astrophotography, especially not until you're very comfortable getting the data in the first place and especially not until you're proficient processing it.

Very best,

It wouldn't be a single purpose. The 10-18 isn't working well for the indoor shooting as well I wanted it to. I'd be selling the 10-18 and replacing it with the next bargain i find(tokina 11-16, rokinon, etc). Hopefully I will have time tonight to go out and see if I can grab some shots.

wantafastbusa wrote in post #18503244
It wouldn't be a single purpose. The 10-18 isn't working well for the indoor shooting as well I wanted it to. I'd be selling the 10-18 and replacing it with the next bargain i find(tokina 11-16, rokinon, etc). Hopefully I will have time tonight to go out and see if I can grab some shots.

The Tokina 11-16 is very sharp. I used one for a while. I actually preferred it at 16mm, wide open at F2.8. It's great for everything that doesn't involve direct sunlight (sun flares on the Tokina are bad, that's it's only fault).

The Rokinon 16mm F2 would be something to consider. F2 is important, because you can stop down to F2.8 and gain more sharpness, better coma control, no CA, etc, while still having F2.8 speed. And 16mm on APS-C grabs the milky way's core with plenty of room to spare. Worth looking at. Pulls double duty as a good wide lens too of course. Though it's not ultrawide the way 10mm is. 14mm isn't very wide on APS-C either though.

If I recall, there is now a Rokinon/Samyan 10mm F2.8 for APS-C that is not a fisheye. That would be very good on APS-C for this purpose. It gives you a very wide lens (ultrawide), that is fairly fast, and can be great for anything and everything, and allow 30 second exposures at night at F2.8 which would help you keep your ISO down around 6400.

Here's a Tokina 11-16 F2.8 @ 16mm and f2.8 on a Rebel T4i and ISO 1600, single exposure for 135 seconds (to do the same thing at 20 seconds would be some where over 2.5 stops more of ISO, so maybe ISO 10,000 would have been required to do the same exposure). This is where a lot of the argument for "pushing ISO as high as it goes" comes into play. I would have been at nearly the ISO 12,800 limit of the T4i even at F2.8 because of the limitations of exposure time (20~30 seconds).

IMAGE LINK: https://flic.kr/p/nS8k​eV

I must say... you have some awesome shots! I'd be stoked to get something like that!