ISO Confusion

[Ande]

Hi all. Just trying to get my head around what the ISO setting for my Canon 80D should be. Now, I understand how ISO works during regular, terrestrial photography, and how the need for varying the ISO is a factor.  However, after reading this article: http://dslr-astrophotography.com/iso-values-canon-cameras/ it is suggested that, when it comes to astrophotography, there is no advantage to be had by increasing the value beyond the optimum setting, which,  for my particular camera, is an ISO value of 200, as the SNR rises considerably alongside the ISO.
 

When I last attempted imaging, I stuck with this setting of 200 and tried, unsuccessfully, to capture the Pleiades. As it turned out, my exposure time was way too short due to my inexperience. But, in the back and forth that followed, it was suggested by a few members that my ISO needed cranking right up, which in complete contrast to what the article above is telling me.  So, is there a general consensus? Do I stick with, what appears to be, hard fact, or do I take the article with a pinch of salt?

[fifeskies]

ISO 200 is the best setting but will mean longer exposures which is fine if you have a tracking mount or are guiding.

 

My Canon has best setting at ISO 800  (600D) and at that I use 120 sec subs on the Pleiades (with my 4 inch ED scope).

at ISO 200 this equates to 480 sec (in my scope) , which I could do as I often image at 360 sec subs.

If you cant expose that long without trails then by all means crank up the ISO but I would limit it to 800 or 1600.

 

Remember you wont see a lot in the single frame and will need to stack then stretch to get a final image.

below is a single 120 sec frame (at ISO 800) compared to a stacked and stretched version (20 subs)

 

Edited January 18, 2021 by fifeskies

[vlaiv]

Depends.

Both are right, but differences might not be what you expect.

ISO is just multiplication factor - meaning that it does not change SNR - which is signal to noise ratio (multiply top and bottom of fraction with same number and fraction does not change). In that sense - after you enter region where there is no quantization effects - there is no difference between ISO 200 and ISO 800, except in full well capacity and one more thing.

What is this other thing? It is read noise. With CMOS sensors - some of the noise is injected before A/D conversion and some of it after A/D conversion.

This means that part of read noise is subject to "amplification" by ISO. This means that total read noise is actually smaller at higher ISO values. This is best seen if you look at read noise vs Gain settings for astro cameras (not sure if anyone made such graph for DSLRs):

So higher ISO offers somewhat lower read noise because of that.

Now, read noise is combated with exposure length (depending on your light pollution) - so it is not that important - unless you are forced to use short exposures.

Conclusion?

If you use short exposures - then take care of read noise and use high ISO, but if not - consider ISO to be of no importance and "fixed" for your purposes.

[Roy Foreman]

There is a trade off with ISO settings. Low settings like 200 will give higher quality and can be pushed further during post processing to bring out details, but will require longer exposure times. Higher settings like 6400 will reduce your exposure times quite dramatically but will produce a lot more digital noise. Generally you can offset this but taking several identical shots and stacking them together (I usually take 5 or 7) in photoshop.

 

As a guide for the Pleiades, try 5 x 60 sec at ISO 6400 and see what you get.

Let us know how it turns out

[Ande]

9 minutes ago, fifeskies said:

ISO 200 is the best setting but will mean longer exposures which is fine if you have a tracking mount or are guiding.

 

My Canon has best setting at ISO 800  (600D) and at that I use 120 sec subs on the Pleiades (with my 4 inch ED scope).

at ISO 200 this equates to 480 sec (in my scope) , which I could do as I often image at 360 sec subs.

If you cant expose that long without trails then by all means crank up the ISO but I would limit it to 800 or 1600.

 

Remember you wont see a lot in the single frame and will need to stack then stretch to get a final image.

below is a single 120 sec frame (at ISO 800) compared to a stacked and stretched version (20 subs)

 

That is very helpful, thank you. I am using a Skyguider Pro, so should be okay with longer exposures. I’ve just bought a guide camera too, but not had a chance to become acquainted with it as yet. Next time out I am going to make it my mission to find out what max exposure time I can reach.  That is a stunning final image by the way

[Ande]

10 minutes ago, vlaiv said:

Depends.

Both are right, but differences might not be what you expect.

ISO is just multiplication factor - meaning that it does not change SNR - which is signal to noise ratio (multiply top and bottom of fraction with same number and fraction does not change). In that sense - after you enter region where there is no quantization effects - there is no difference between ISO 200 and ISO 800, except in full well capacity and one more thing.

What is this other thing? It is read noise. With CMOS sensors - some of the noise is injected before A/D conversion and some of it after A/D conversion.

This means that part of read noise is subject to "amplification" by ISO. This means that total read noise is actually smaller at higher ISO values. This is best seen if you look at read noise vs Gain settings for astro cameras (not sure if anyone made such graph for DSLRs):

So higher ISO offers somewhat lower read noise because of that.

Now, read noise is combated with exposure length (depending on your light pollution) - so it is not that important - unless you are forced to use short exposures.

Conclusion?

If you use short exposures - then take care of read noise and use high ISO, but if not - consider ISO to be of no importance and "fixed" for your purposes.

Wow! Thank you. That is quite a lot of information there. I shall need time to digest, and understand it. I am aiming for longer exposures. My first night out I really wanted to grab some quick exposures before the clouds rolled in, mainly so I could check out the mechanics of DSS. But my exposures were way too short, to the point that DSS rejected them. That’s when I got into the ISO mini debate. I shall be aiming for up to 3 minutes at a push. We’ll see how that transpires😂

[Ande]

15 minutes ago, Roy Foreman said:

There is a trade off with ISO settings. Low settings like 200 will give higher quality and can be pushed further during post processing to bring out details, but will require longer exposure times. Higher settings like 6400 will reduce your exposure times quite dramatically but will produce a lot more digital noise. Generally you can offset this but taking several identical shots and stacking them together (I usually take 5 or 7) in photoshop.

 

As a guide for the Pleiades, try 5 x 60 sec at ISO 6400 and see what you get.

Let us know how it turns out

Next time out I shall try a quick experiment at higher ISO. However, if I can increase my exposure lengths significantly, then I would rather work with ISO 200.

[fifeskies]

Guiding will make a big difference to what you can achieve.

 

My processed image is still a bit below what I would normally aim for and needs more integration time. 

I only spent about 10 minutes processing it (in GIMP) and there are a number of steps that can be used to improve it further that have not been done yet, but shows what you can get without spending too long on it.

 

It is worth cranking up ISO to get some quick exposures to have a go with , and early on with new gear it is better to get some enjoyable results fast to encourage you to keep going.

I only use ISO 6400 for initial setting and platesolving to a target, but it is always good to see your target pop out on the first test frame before setting off on a set of exposures at the better ISO settings.

[Roy Foreman]

My camera is a Nikon D810a and is designed for astrophotography, so I think nothing of using ISO 6400 on a regular basis. Other cameras may not work so efficiently at these high ISO's, but I would suggest most would give decent results at ISO 1600. The advantage of high ISO's is shorter exposure times, which is an advantage when making the most of short imaging windows in the weather, as I often have to.

Here is an example of an ISO 6400 image :-

 

[bottletopburly]

6 minutes ago, Roy Foreman said:

My camera is a Nikon D810a and is designed for astrophotography, so I think nothing of using ISO 6400 on a regular basis.

 Suggested Sweet spot for the D810 is 200 iso  http://dslr-astrophotography.com/iso-values-nikon-cameras/  

[vlaiv]

14 minutes ago, Roy Foreman said:

The advantage of high ISO's is shorter exposure times, which is an advantage when making the most of short imaging windows in the weather, as I often have to.

Do you mean that you can use shorter subs and same total imaging time, or that there is some magic involved and that on ISO6400 you can make same quality image in say 15 minutes as with ISO400 in an hour?

[smr]

All my astrophotography images prior to the Heart Nebula were taken with my stock 80D at ISO 200 if you want to have a look

https://flic.kr/ps/2ZddwW

[Stuf1978]

I occasionally use my stock 80D for astro work on broadband targets and have found there is no real discernable difference at ISO 200, 400 and 800. The following image was taken at ISO800, 105 x 120 second exposures

 

[Roy Foreman]

To reply to the above few posts in one hit - firstly the D810a is not like a regular D810. The lowest ISO setting has been raised from 64 to 200, so the sweet spot is probably higher at something like 1600.  I seem to remember reading that somewhere. A 60 sec exposure at 1600 could probably be pushed to look like a 60 sec exposure at 6400 but I've never tried it. I have taken images of star clusters at ISO 200 and they can be pushed quite dramatically, but nebulae are not quite so forgiving.  Most modern cameras perform really well at high ISO's these days, and my personal preference is to keep exposures as short as possible to minimise tracking errors, atmospheric effects and the like.

In terms of total imaging time, to a certain extent, yes, 15 minutes of imaging at 6400 gives similar results as 60 minutes at 1600 - for  a single image. But, at 6400 you will need a minimum of 5 exposures to smooth out noise, whereas at 1600 you can get away with three.  So it's 75 minutes versus 3 hours !

I must stress that the above applies to the D810a, but I can't see why the same logic cannot be applied to other cameras.

Nice image of the heart nebula Joel, well done at ISO 200 !

[shropshire lad]

44 minutes ago, smr said:

All my astrophotography images prior to the Heart Nebula were taken with my stock 80D at ISO 200 if you want to have a look

https://flic.kr/ps/2ZddwW

Interesting .. and I'm followingh you now on flickr

 

[vlaiv]

1 hour ago, Roy Foreman said:

To reply to the above few posts in one hit - firstly the D810a is not like a regular D810. The lowest ISO setting has been raised from 64 to 200, so the sweet spot is probably higher at something like 1600.  I seem to remember reading that somewhere. A 60 sec exposure at 1600 could probably be pushed to look like a 60 sec exposure at 6400 but I've never tried it. I have taken images of star clusters at ISO 200 and they can be pushed quite dramatically, but nebulae are not quite so forgiving.  Most modern cameras perform really well at high ISO's these days, and my personal preference is to keep exposures as short as possible to minimise tracking errors, atmospheric effects and the like.

In terms of total imaging time, to a certain extent, yes, 15 minutes of imaging at 6400 gives similar results as 60 minutes at 1600 - for  a single image. But, at 6400 you will need a minimum of 5 exposures to smooth out noise, whereas at 1600 you can get away with three.  So it's 75 minutes versus 3 hours !

I must stress that the above applies to the D810a, but I can't see why the same logic cannot be applied to other cameras.

Nice image of the heart nebula Joel, well done at ISO 200 !

Well, no.

ISO is just multiplier - it can't magically increase amount of light that you gather. Upping ISO is like increasing brightness in processing. Signal to noise remains the same with both low and high ISO.

In another words - if we neglect read noise (and sometimes it makes quite bit of difference - but that is in case you have cooled camera, very dark skies or use narrow band filters - in essence when you remove other noise sources) it is total integration time that counts regardless of sub duration or ISO setting.

Modded camera makes not difference to this.

[Roy Foreman]

Yes I agree - using higher ISO is like turning up the brightness at time of capture rather than in processing. Never tried taking side by side images at ISO 200 and 6400 for the same duration and then processing each to see which reveals most detail / less noise.  When I get a chance I'll try it.  I've done it with 1600/6400 and the 6400 turned out less noisy. Can't show you the results of that because I since discarded the images as being below par.

[Alien 13]

I have the 80D and do shoot nightscapes/astro at an ISO of 200 as this camera is unique in the Canon range because its virtually ISO invariant, this doesn't mean you cant try other higher settings but there will be little to gain in theory.

I think most other Canon cameras tend to operate best around an ISO of 800 to 1600 which is close to its unity gain setting.

Alan

[rickwayne]

Note too that DLSRs usually have two kinds of amplification, and different makes and models mix and match them differently.  Vlaiv makes a cogent explanation of what's going on in the lower range, where analog circuitry is doing the driving.

But at higher ISOs (and "higher" is what's different for different cameras), the camera starts multiplying the digital output. This gains you exactly nothing that you can't achieve with post-processing, at the expense of lowering the dynamic range and possibly blowing out highlight areas.