Update:
If you are looking for a specific suggested ISO to use for your camera you can check it here:
Suggested ISO values for Nikon cameras.
Suggested ISO values for Canon cameras.
Suggested ISO values for Sony cameras.
For a better understanding what ISO exactly is and how it works and how to deduct the best value you can read on below.
The ISO on DSLR cameras is probably one of the most misunderstood settings and people don’t seem to agree on the best values to be used for astrophotography. I’ve seen articles and people recommending anything between ISO100 and ISO3200 for exactly the same camera. This is very unfortunate, especially since choosing the best setting can make quite a difference in your end result.
Much of the confusion comes from the fact that the reality about ISO for digital cameras is very counter intuitive and people have trouble ‘accepting the truth’. However, I do want to stress that I think we can all agree on the ‘best ISO’ for your model. Time to have a detailed look!
The ISO setting is NOT influencing your cameras sensitivity to light!
ISO misunderstood
In daytime photography we are taught about the ‘Exposure Triangle’; ISO, aperture and exposure time. The three are interconnected we learn, and to get the same amount of light, you need to double your exposure time if you go from F2.8 to F4. Alternatively you could double your ISO and leave your exposure time as is. If you increase one of the three, you can decrease one of the others in order to get the same exposure in terms of light captured. Or so we are told at least.
In almost every article that is explaining this, ISO is defined as the light sensitivity of the sensor. So a higher ISO value means a more sensitive sensor. hmmm. hold on. Does this make sense?
Sure, in the analog days of film you would have film that used larger grain to be more sensitive, so the film really was more sensitive for a higher ISO value. But how would that work for a digital sensor? Are the pixel sizes suddenly increasing when you increase the ISO? No, of course not. So how does the ISO setting increase the sensor’s sensitivity to light then?
It doesn’t.
It really doesn’t.
Increasing the ISO setting is NOT making your camera pick up more light. It is NOT changing your sensor and it is NOT increasing sensitivity.
Ok, so what does this setting actually control then?
The ISO setting is determining the amplification factor that the camera applies to the signal that represents the amount of photons captured in a pixel. Using a higher ISO just means you are amplifying the light you captured on the sensor. So your pixels never collect more photons, it is just that the signal is amplified. In a sensor this amplification can be done both analog and digitally. And this is the really important part of choosing the best ISO setting for your camera for astrophotography.
ISO and Read Noise
If you increase the ISO, you’ll increase the noise. This has been true for years, even though the sensors are getting better and better. It is important to really understand what this noise is that is increasing; the read noise. The read noise is quite complicated but can simply be understood as the noise that is inherent to the electronics of the camera. This means this noise is independent of exposure time. So a longer exposure will have the same read noise as a very short exposure, and this is (partly) why exposing longer will increase your Signal to Noise Ratio (SNR).
Many reviews and reports on camera performance and sensor statistics will state that the read noise is depending on the ISO. For our understanding it is necessary to look one step further.
Read noise consists of two different parts; the so called upstream and downstream noise. The upstream noise is all the noise that occurs in the electronics before the actual Analog-to-Digital conversion in the ADC. The downstream noise is anything from that point onward, which of course includes the noise of the ADC itself.
Now remember that the ISO setting controls the amplification of the output signal of the sensor. This means that all the upstream noise will be amplified with it, while the downstream noise remains consistent. In other words; we gain SNR! And this of course is a good thing right? So should we choose very high ISOs then? Well, not so fast. As always, things aren’t quite that easy;). There is a limit to the gain you will get, which is determined by the downstream noise. Once the upstream noise gets high enough OR the downstream noise gets close to zero, the upstream noise will simply swamp out the downstream noise, making it irrelevant. So once we move past this point, the read noise will be scaling linearly with the signal, gaining us nothing anymore in terms of SNR; If you double the signal, you’ll double the noise. It is this situation what people tend to call ‘ISO-less’, as there is no real gain and you can do this amplification also in post processing. This might even be better in terms of Dynamic Range (DR). I’ll explain why.
Please note this is an oversimplification, but for our understanding it will do for now. For the exact math of it all I recommend this paper.
ISO and dynamic range
There is another, very important thing to keep in mind when we are talking about the ISO setting. This is the influence of ISO on the dynamic range. Simply put, the dynamic range determines how well a camera can distinguish and record the light levels between the faintest (black) area and brightest (white) area in a picture. The dynamic range is determined for the most part by the ability of a pixel to hold a certain amount of electrons, the full well capacity. Think of our analogy of the buckets of rain again; if a bucket can hold 1000 drops, it will hold more information than if it can only contain 100 drops. This is what we call the dynamic range.
Furthermore, the dynamic range is limited by the ADCs capacity to represent the different charges in numbers, determined by the amount of bits the ADC can use.
Now, if we look at amplification we can see what will happen to the dynamic range; the signal values get amplified, thus raising the ‘floor’ of the signal. BUT, at the same time the upper limit of the dynamic range remains fixed, determined by the ADC. So, the difference between the highest measurable/recordable value and the lowest value is decreasing as we are amplifying, and thus raising, the lowest value while the upper value remains fixed. In other words, we loose dynamic range if we increase the ISO. And this is very bad for astrophotography, where we have to deal with dynamic range challenges in all our pictures because of faint versus bright stars for instance.
We loose dynamic range if we increase the ISO, which is very bad for astrophotography
ISO and Unity Gain
Unity gain is a concept that is introduced often when discussing the choice of best ISO to use for astrophotography. However, unity gain would only be relevant to consider if you are trying to capture very faint signal AND you want to do this by using only a few exposures. In all other cases unity gain really doesn’t matter, and you certainly don’t want to choose it if it means loosing Dynamic Range.
You can read more details on unity gain and why it doesn’t matter here.
1/3 and 2/3 ISO stops
Most cameras offer steps of 1/3 to change the ISO setting. Don’t use those settings for astrophotography, always use ‘full’ ISO stops!. The camera is just scaling your images to mimic these ISO settings. For 1/3 stops, like ISO 125, the camera will simply use the analog amplification of ISO 100 and scale it digitally to match ISO 125. On the other hand, for a 2/3 stop like ISO 160, the camera will use the next ISO setting and scale it back down. You’ll encounter some articles talking about using 2/3 ISO settings since it has a better SNR, but for astrophotography you want to be in the range where you already have the best SNR from your ISO setting and don’t want any digital scaling as you loose information.
Just stick to using the ‘real’ ISO stops; 100, 200, 400, 800, 1600, 3200, etc.
Summary; ISO considerations
Ok, lets take a step back and look at what we have established;
- ISO has nothing to do with the sensor’s sensitivity
- Instead, ISO is an amplification of output signal
- this amplification will be done either by an analog amplifier or digitally, depending on the brand/model and the ISO setting
- Increasing the ISO decreases dynamic range
- Increasing the ISO increases SNR until the upstream read noise swamps out the downstream read noise
Point 3 and point 5 determine the best setting for your camera.
So even though this is really counter intuitive for most of you, stop thinking about ISO as a way to increase sensitivity or a way to make your image brighter. It really really doesn’t!
The best ISO for your DSLR
The best ISO for your dslr will of course depend on your particular model, but in general it will be defined by the following;
The best ISO for astrophotography for any DSLR is the lowest ISO level from which either a.) the upstream noise swamps out the downstream noise OR b.) the amplification will be done digitally in camera, whatever value of both is the lowest.
Starting from that value the read noise will be (more or less) consistent and we can call the sensor ‘ISO-less’ and we gain nothing by increasing the ISO, while only hurting ourselves by diminishing the dynamic range.
Next, let’s learn how to find the best ISO for your specific camera model.
The best ISO for astrophotography for any DSLR is the lowest ISO level from which either a.) the upstream noise swamps out the downstream noise OR b.) the amplification will be done digitally in camera, whatever value of both is the lowest.
Defining the best ISO for your model
At this point you probably still are confused as to what ISO value is the best for your specific DSLR. It’s nice that we have a clear definition of what the best value is, but how do you determine what this means for you in real life?
Unfortunately there is no direct answer to this question, as the ISO levels from which a dslr starts amplifying digitally is generally not listed anywhere. However, the read noise levels for your camera at different ISO settings can be found at http://sensorgen.info, which uses data from DxO to calculate their curves. Digital amplification will be visible in these results as an (dramatic) increase of read noise for the highest ISO settings, since you’ll start amplifying the downstream noise as well if you do it digitally. So in general you would be fine at checking the read noise curves and see when they start to even out at the lowest values and use this as your ISO for your model. In the graph below I’ve plotted the read noise for the Nikon D7000 and the Canon 6D, not to do any qualitative comparison, but just to show you that the best ISO settings can and will greatly differ per model!
This graph makes it very clear why there can be and remain so much debate and confusion about this matter! Based on this graph I would guess that ISO1600 would be best for the Canon 6D, while ISO100 (!) or 200 would be best for the D7000. I recommend to test it out for your self to be sure, but at least these graphs give guidance on what to test; you don’t have to try the 6D at ISO100 and testing the D7000 at ISO1600 is also a waste of your time…
I’ll do a write up on the best ISO settings for the most common dslr models soon.
Update:
You can find the list for Nikon models with suggested ISO values here.
Click here for a list of Canon models with suggested ISO values
And here for the suggested ISO values for Sony cameras
Concluding the best ISO discussion
This matter is really, really complicated which is increased even more by the fact that we have to work with facts and numbers that are not generally made public by the manufacturers. I’ve simplified things (a lot) in some places to try to give you an explanation that can be understood without diving in to the math or doing lots of side research to even understand what I’m talking about here.
I hope I succeeded in doing so, and would love to hear your further questions or remarks in the comments below.
Hi,
I am still a little confused. I have a Canon 100D and I am trying to figure out the best ISO setting for it. So just to clarify, the higher the read noise for a given ISO, the better it is for astrophotography, is this correct?
Also, how does exposure time come in to this, for example, if I am imaging Andromeda galaxy I can use a short exposure time at ISO 800, but I would need to drastically increase the exposure time to image an object at magnitude 11 lets say. The way I would compensate for this is by increasing the ISO. What would be your advice for this?
Also, does temperature of the camera affect the read noise?
Thanks and I love the articles by the way, I’m learning so much!
Regards,
John Dolan
Hi John,
” So just to clarify, the higher the read noise for a given ISO, the better it is for astrophotography, is this correct?”
No this is not correct. Although there are situations where noise can actually help you, in general you can state that all noise is bad 🙂
Then the exposure time. You really need to try and get rid of the idea that increasing ISO means you catch more light. You don’t. You just amplify the signal that’s already there. So if you need to catch real faint stuff, at least you need to catch the photon’s in the first place. Choosing a higher ISO could help you with the fainter signal so that is stands out more above the noise, but ONLY if you are still increasing the SNR. And that happens ONLY if the downstream noise get’s swamped by the upstream noise.
Now I think this might need a bit more explanation.
For now it’s just important to understand that the downstream noise is the noise that occurs AFTER the amplification and the upstream noise BEFORE the amplification. So, if we are increasing the ISO, we increase the amplification and thus are increasing both the signal and the upstream noise, but NOT the downstream noise. So even though we are amplifying some of the noise, we still gain Signal to Noise Ratio and are better of doing this amplification in camera than in post processing. This of course because in post processing we would also be amplifying the downstream noise.
Now, imagine that the downstream noise is already very small or even close to zero. There is hardly any SNR gain in this case by amplifying and therefor no use to increase the ISO. And in this case it won’t help getting the faint signal above the noise floor, and will only make you clip the highlights faster.
So, increasing ISO only makes sense when we are still (significantly) gaining SNR.
For the Canon 100D I would test out both ISO800 and ISO1600 if I were you. If you look at the sensorgen.info data http://www.sensorgen.info/CanonEOS-100D.html, you can see that the read noise is levelling out more or less from ISO800. So this would be my first target to test.
Great article ! Thank You !
Thanks so much! 😀
hi
i have a canon 500d, and from the charts for my camera the read noise lowest point is between 1600-3200, so after that it starts to increase. am i correct to assume 3200 should be the maximum i use? can the additional noise not be processed out later?
thank you. this is a great article it has really helped me.
Kind Regards
Faheem
If I look at the sensorgen data I would think ISO1600 would be best.
You’re best of using dithering to get rid of the noise. Remember; it’s not additional noise, just amplified (along with your signal) 😉
How would you suggest I evaluate my Fujifilm X-T1 camera, since it, nor any modern Fuji cameras, are not reviewed in Sensorgen?
Thanks for a fascinating article!
Ted
Hi Ted,
The best method of evaluation is always to test your own camera under your own conditions yourself imho. So I would just make sure the sky conditions are more or less consistent (no lights going out from evening to night, no moon setting or rising, stable sky conditions) and then take multiple long exposure subs at different ISO settings and analyse them in your program of choice. Not sure if you are a PixInsight user, but most software is able to give you some indication of the SNR in terms of mean values and StDev.
Chris
Hello Chris, congratulations on your superb website which really demonstrates your passion for DSLR astrophotography!
I have found that lower ISO settings seem to produce results in which the post-processing is “more relaxing”, so my long exposure DSO frames tend to be at ISO 400/800. Unfortunately this is not robust evidence – to gather true comparison frames would mean sacrificing an evening of imaging, and clear skies are very few and far between here in NW England.
So I’m drawn to the fascinating figures and graphs on the sensorgen site. Would you mind giving me your opinion on the data for the Nikon D90? The read noise seems to step down quite significantly between ISO 1600 (indicated, actually 1125) and 3200 (indicated, actually 2322). I’m not sure how this could be so, but it suggests that ISO 3200 (indicated) would be a good choice, as the Dynamic Range holds up well, and saturation is almost never a problem anyway.
For brighter objects where saturation might come into play, the next best appears to be ISO 800 (indicated).
Maybe I should give an evening to testing. We do have some dark skies here.
Best wishes
Simon
Thanks for this interesting article. I wanted to argue so I went and looked up my camera (Sony a5000). From what I can tell, your prescription recommends 1600ISO setting and that seems to be where I get my best images. Always nice to see the theory match up with the results 🙂 Thank you for this article.
Thx for sharing your experience! Always good for others to read, as the best ISO will differ quite a bit per camera model.
I understand that the experience of the Nikon’s with low ISO performance can trigger some debate. Just important to understand the mechanism and check for your own model!
Hi!
Great site and brilliant article!
I am almost embarrassed to admit though, I still don’t fully understand what the best ISO is for my particular camera – even looking at the info on the link.
I’m planning going full frame with the 6D as soon as the funds permit, but I currently using a cropped-sensor Canon 70D with a Samyang 14mm f/2.8 lens. Can you put me out of my misery and give me your best guess please??
Many thanks in advance!
Gavin
@Gavin Park – for your Canon 70D camera, the best ISO setting (based on read noise) is ISO6400.
@Chris – just ignore the saturation numbers?
The saturation numbers are especially important in determining the quality of the sensor. I would say you can ignore it to determine the best ISO, since you will pick the highest saturation number automatically if you look at the curve of the read noise and pick the value when it starts leveling out.
Thanks for this, and very helpful. I currently use a Sony Alpha SLT-a55V for my astrophotography. I’ve been using it at ISO1600 to reasonably decent results, but I worry about blowing the stars out with the reduced dynamic range. Looking at the charts, it kinda levels out at ISO800, although ISO1600 is marginally lower than 800. I’m thinking that maybe I should try it at ISO800 to increase the dynamic range… especially if I can manage to get longer exposures?
I’d always recommend to test the results for your camera for a few values that lie near the point where read noise starts to level off. It is simple enough to do and you can discover what works best for your own needs in your own circumstances. This article gives you a pointer as to which ISO’s you should be testing the camera on.
Good luck!
Hi, thx for your article.
I find it hard to see the point where the read noise is levling out when i look at the chart for my 1000D. What would you suggest? Until now I always used ISO 200 and was happy with the results, but never really tried something else.
Hi, I’d just like to point out that your conclusion on which ISO to use for the D7000 vs 6D should be a little different. The D7000 read noise range is nearly linear/flat the whole way out (a hallmark of Nikon camera’s is that they have flat read noise through the ISO range) so you should find the point at which the dynamic range drops off the most. You basically lose 1 stop of DR for each EV equivalent step in ISO, you maintain a DR greater than 10 (from 13.5 at 100) out to 1600, I wouldn’t shoot above an ISO of that. With the 6D the read noise becomes linear (or flat) after around ISO 800 and you should really not shoot above that, the Canon DR drops from 11.5 at ISO 100 to 10.1 at 3200 so not much loss there.
The Nikons have always been the clear winners for read noise across an ISO range, with the Canons being able to maintain DR, but at the cost of read noise at lower ISO’s. This is a fantastic resource for people interested in seeing what their camera sensors technical specs are across different ISO’s:
http://www.sensorgen.info./
So if I look at the data for the Nikon D750 it seems like the best combinations of read noise and dynamic range seem to be for ISO 800 and 1600?
[…] about the correct ISO for AP. Be sure to read the questions and answers at the end of the article. What is the best ISO for your DSLR for astrophotography? | DSLR Astrophotography Cheers, […]
I have a Eos 450D. I had been using it at ISO400 but going to ISO1600 there seems to very little difference in the dynamic range and saturation seems to max out there per the chart. Thoughts on that?
Definitely worth a try at ISO1600 indeed, looking at the read noise.
is there any proof to show that, ‘1/3 and 2/3 ISO stops are digitally scaled’? Thanks.
Of course you can check it out yourself for your camera in specific, but there are a lot of forum posts and user tests showing this.
This link and especially the comment section might be interesting for instance;
https://marvelsfilm.wordpress.com/2009/10/10/canon-7d-noise-and-iso-test/#comment-960
Shows how partial ISO is really shot on closest full stop and then digitally processed +1/3 or -1/3
So, I’m confused by the ‘clipping of highlights’. How actually does it work? Will it increase DR than the higher full ISO stop? Will it change the linearity of RAW data? If both are no, it seems just a kind of Histogram changed RAW on the higher full ISO stop? But in my pictures, it seems ISO 640 has more DR than ISO 800 (I don’t know how to measure DR).
[…] and poor contrast, but I’d recently discovered an article explaining at a technical level what the ISO setting in DSLR cameras actually does, which gave me an idea of what I’d been doing […]
Thanks for a very interesting blog. I now know more about the best ISO settings for my camera. But how can I find the camera that perform the best in astro photography? You can do so much In post, but is Nikon superior to Canon? Or should I choose Sony for my next camera? What to look for? If not high iso performance? I’m considerating a change to Sony. My camera now is 5D mk ll but Your article made me a little bit confused?
Camera choice can be complicated if you consider all variables that you should take into account.
In general, Nikon is better for astrophotography because of the sensor performance. However, since you have Canon right now I can imagine you already own some Canon lenses. In that case I’d just stick to Canon.
Next is noise performance, but also take the resolution into account. Get a modded version or have your model modified (lots of people offering this as a service).
Not knowing anything about your gear or personal preferences and budget, I’d recommend Nikon D600/D610 which is good value for money. Canon equivalent would be the 6D(a).
If you have more budget to spent I’d definitely look into the D800A from Nikon.
Thanks for the great article. I am still confused on which ISO to use on D3300, can you help me out?
Thanks!
Looking at the read noise graph on sensorgen.info I’d say ISO400 probably is best in general.
Very good article, thank you!!
Now, unfortunately I have doubts about the exposure length 🙁
I shoot from a quite polluted environment (SQM 19-19.5), with a Newton 200/800, a Canon 700D: stacking usually 70-100 frames with 90 seconds exposure at 400ISO
This means that at 1600 ISO my Canon won’t be able to go past 20-30 seconds of exposure without having the histogram peak past 50% of the scale.
It seems to me that short exposure times may have these problems
1. Huge number of frames to achieve the same total integration time: I will need to shoot hundreds of frames! Maybe going beyond a certain number of frames is useless, I guess, because it only improves the quality of what you got in a single frame, I am right?
2. I read somewhere that exposure length matters too, because the camera has to collect enough photons from the faintest parts of a DSO and that if you increase the ISO and lower the exposure time, the risk is that you are collecting no photons from faintest regions and the result will be zero signal irrespectively of the amplification, what do you think about it?
thanks!
Excellent points! Yes the exposure length matters, mostly on faint signals.
What I left out of this article for simplification matters, is the role read noise actually plays in any particular situation. The situation you are describing is limited by other factors and not read noise. In that case I definitely go for lower ISO (more dynamic range!) and longer exposure times (more signal/ better SNR)
I’ll do a post on optimal exposure times in the future 🙂
Holy smokes, there are no apostrophes in plurals!! Basic grammar!
Hi,
Thanks for your explanation.
I have however some questions, so you say that in the D7000 the optimal would be 100/200 ISO, but assuming a 14mm Lens with a 1.5x crop factor, we could only have around 22 sec of exposure. with these specs, I would have a very dark image, and so I would lighten up the image in post processing, right?
However you say in one of your comments “his of course because in post processing we would also be amplifying the downstream noise”, so am I missing something? Shall I take photos with a D7000 in 100/200 ISO and then amplify in post processing?
Thx in advance.
Thanks for your question, which is a very valid one!
Your case of using 14mm is very different in comparison to doing (wide field) deep sky with let’s say 400+mm. In your situation the read noise doesn’t matter as much. However, what of course still matters is the dynamic range. So yes, I would be inclined to say it would be better to use ISO200 and stretch in post.
Whether you will see significant differences or not and if this is convenient compared to shooting at ISO 1600 or even 3200 I leave for you to test and find out 😉
Hi,
Thanks for your prompt response.
I have to admit that I didn’t catch that this topic was about wide field.
I usually take pictures of the milky way on 14mm (x sensor factor 1.5), and use around 3200 ISO on a D5100 (I just asked about the D7000 because I think they share the same sensor).
To be honest, when I take on a low ISO (and by low, I mean 800!!) the picture is really dark, so I usually go up to at least > 1600. What I was trying to understand was if I take the picture really dark then I would get more dynamic range and then I could put the image clearer/lighten up in post processing (I am sorry if the term clearer is not correct, I am not a native English native speaker).
So for what I understood, since that sensor just amplify the signal after ISO 200, than I could take the shots at that low ISO, even though I would see nothing in there and then take care of it in post processing. However, usually the milky way tutorials mention to take the shots in the highest ISO we can ..
I will try to make some experiences next weekend to see if it gets any good, I mean, taking some shots in raw in 200/800/1600/3200/6400 and then make some comparisons.
Thanks.
Hi pedro,
How did your tests go?
I would indeed recommend a low ISO because of dynamic range and just amplify the signal in post processing (stretching 😉 )
Hi Chris, very good article, sorry for my bad english, I just started astrophotography, judging from the graph the best ISO for my 550D camera is 800, but I am limited by my equipment, I have a portable setup using Canon 28-135/3.5-5.6 and Vixen Polarie mount, my rule of thumb so far to target the EV value to -5 (I read from an article EV value less than -5 will introduce more noise), I can see mostly your pictures has EV -6 to -7 (using exposure calculator in PhotoPills), I can’t reach that EV value also due to polar alignment precision of Polarie (can’t shot longer than 1 minute without trailing), will I ever be reaching your astro photo quality due to this?
Hi Wiama,
To be honest I never look at the EV.
If i’m not sky (light pollution) or equipment limited (in terms of tracking/guiding) I mostly determine my exposure time as follows:
first I determine how much total exposure time I’ll have. If limited I’ll make sure I end up with 25 frames for stacking because of the SNR gain by stacking dithered exposures.
If i’m not limited in terms of total exposure time, I’ll adjust the exposure based on the target I’m shooting and then it is always a trade of between keeping nice and colorful stars (not over exposing to many stars, especially relevant with targets within the milky way) and getting out the faint(est) signal.
As far as your situation; if you can’t go longer than 1 minute without trailing this is your exposure limit. But for many targets this can work out just nicely as long as you take enough frames.
Thanks Chris, one more question if you don’t mind, on my Canon 550D (suggested ISO is 800), I need to raise ISO to 1600 so that sensor read noise swamps the banding. What’s your thoughts on this?
Nothing is more important than your own tests! So if your results indicate you need to use ISO1600 because of the banding you definitely should!
thanks for sharing, it’s valuable to hear people’s experiences to check the (theoratical) graphs 🙂
Hi,
Great article which really made realise I had no idea what ISO was doing in DSLRs. I have a Canon EOS 600D. http://sensorgen.info/CanonEOS-600D.html seems to suggest my optimum ISO is 12800. Have I got my understanding wrong as this seems really high?
Hi,
Glad you found the article interesting and helpful.
I’m not sure how you are reading the sensorgen chart, but you should be looking at the bent in the curve of the read noise. The point where it starts to ‘flatten out’ so to speak. That seems to me to be at ISO800 so I’d suggest to try that next time and analyze your results 🙂
Thank you so much for your reply. I was looking at the lowest point not where it was flattening out. Dope!
Loved the two articles on DSLR and ISO selections. Learned a ton. Already had the sensorgen data imported into Excel so I graphed the date for my two Canon cameras (6D and 60Da). Very insightful. Using the logarithmic trendline, was able to calculate the exact intercept for the inflection point on the curve giving the recommended ISO’s (200 for the 60Da and 400 for the 6D). Thank you for taking the time to explain these exceeding complex ideas in a very pragmatic way.
P.S. Can share the charts on Dropbox if anyone is interested; just email me so I can pick up your email and send a sharing notice.
Hi Chris
Loved the detail and explanations you give in this article! I’m using a Sony A7S and am wondering about optimal ISO. http://www.sensorgen.info/SonyA7S.html
The read noise seems to continually drop until even 400K ISO. However, a reasonable inflection point could be ISO 25600. Does this seem right?
As my ISO goes higher, I’m also sacrificing dynamic range. Would this mean I would get faint data but bright data would quickly be blown out (bright stars, cores of nebula/galaxies)?
Hi Joe,
Glad you liked the article 🙂
As for the A7S: I’ve written about that model in particular here: http://dslr-astrophotography.com/iso-values-sony-cameras/
What I learned from other people is they tend to use ISO2000 because there is quite a big drop in read noise at that point. I would not go higher than 3200 normally as you’ll be loosing to much dynamic range.
In short the dynamic range is the ‘room’ the camera has available to store it’s color and light information in. Less dynamic range means less color information and easier clipping of signals. (on both ends!)
I hope this helps 🙂
I have a Nikon d5500, which according to the charts should be 100-200 ISO. I am new to astrophotography with a Samyang 16mm f/2.0 lens. I have had moderate success at 1600 or 3200 IS0, f/2.0, and shutter speed 20″. If I use a 100-200 ISO, would I adjust the shutter speed to a longer time (say 30″) to increase exposure, or could the “under-exposed” image be pulled up in post-processing?
Hi Joe,
Thanks for your question. I know it sounds counter intuitive, but the important thing with ISO is to disconnect it from your exposure time. You choose an ideal exposure time (based on personal preference, guiding, sky conditions etc.) and you choose ISO separately. And indeed the idea is to ‘pull it up’ in post processing. Basically because the camera is doing stretching for you with higher ISOs while giving you less dynamic range (and no control over the stretch). So better to do it in post 🙂
Hi Chris! Thank you so much for your hard work here. I shoot Milky Way landscape astrophotography with a Canon 6D and a Rokinon 14mm f/2.8 in and around a little beach town in Delaware. There are usually some artificial lights in the frame. Can you recommend an ISO setting or tell me how to test it myself please? I took a bunch of shots the last couple nights at multiple ISO’s, but I’m feeling overwhelmed in terms of post-processing – I’ve read many articles and have a workflow mapped out but don’t know where to start in terms of RAW exposure straight out of the camera. Thanks in advance!
I’m shooting wide open for 20 seconds – may be pertinent – thanks!
Just found the links at the top of the article – perfect – thanks again!! lol
Hi Chris, great article! I shoot with a Canon 77D and it’s not on the list of Canon cameras as it was released in February 2017, which is way later than your posts have been put up. Are you planning to do further testing and update your list with the newer models as well? Thanks
[…] Long story, here for details if you are interested: http://dslr-astrophotography.com/iso-dslr-astrophotography/ . One take-away is that for most Canons use ISO1600 unless ISO800 (or even ISO400) will let you […]
Great article! Given that you wrote/published this 3 years ago, I very much wish I’d found it sooner! I’ve been puzzling over the right ISO to use (Nikon D7000, D600, D750) for astrophotography. I’ve been pouring over Bill Claff’s excellent photons-to-photos http://www.photonstophotos.net/
It totally makes sense that the ‘right’ answer depends on the camera.
BTW, If you look at Bill’s graphs for the D850, there’s an interesting ‘bump’ in response. Asked about it over at dpreview: turns out it’s a ‘dual-mode’ sensor.
I’ve been trying to convince myself that ISO 100 is *fine* for my D750. Looks like you’ve confirmed my thinking. The onkly issue is that I’m not using a star tracker but instead taking a lot of 5 second exposures and stacking in Sequator – and ISO 100 doesn’t give me enough in a single exposure to judge critical focus. I’d guess the way around is a a few initial shot at high ISO for focus eval, then ISO 100 and interval mode. No?
Glad to hear you found the article useful!
You are describing pretty much the only scenario (in astrophotography) where the high ISOs come in useful; focussing.
I usually crank up the ISO for the focussing. Just be careful you don’t forget to return the ISO to the right value after focussing… (yes I’ve waited quite some exposures because of forgetting to do so 😉 )
And I’ve wasted an entire night of dark sky with a great milky way shot because, although s bright star looked OK in the liveview panel, I was a bit out of focus. I was pretty unhappy when Sequator got done and I put the results up on a big monitor. 🙁
Forgot to mention: I’m pretty sure the chart to look at over at Bill Claff’s PhotonsToPhotos is “Input Referred Read Noise vs ISO” http://www.photonstophotos.net/Charts/RN_e.htm
It’s interactive, click the camera model you’re interested in and the curve for that model appears. Lots of models, so it’d be unreadable otherwise.
Hi Chris,
I am just a filmmaker searching for a Canon 60D native ISO and I landed in this article. Great JOB!
It seems that the native Canon 60D ISO are 800, but this value is valid for astronomical photography. Any suggestions instead for normal filmmaking on planet Earth :)?
From what I read it seems to understand that ISO 800 is the native ISO of Canon 60D. Correct?
So if the native ISO are at 800, at most you can move to 1600 if you have very little light and you do not have the ability to add light.
Vice versa, you can at most go to 400 if you have too much light and you do not have ND filters. Correct?
I ask you to try and understand a little more …
Thanks for an answer!
Hi, Thanks for your excellent question. Nice to hear I’m reaching people outside of astronomers with this article as well 😉
As for ‘native ISO’ (and possibly ‘unity gain’); these are tricky concepts that are often misunderstood and given to much importance from a possibly faulty idea of reasoning.
The main reason these are tricky concepts is that a digital DSLR is always a combination of an analog and digital system. The sensor is an analog system and will ‘fill up’ with electrons. These electrons need to be translated into ‘digital units’ and registered. So there is always some form of translation going on and therefor one could say amplification. People tend to want to stick to a situation where 1e- is translated into 1ADU, but this is not necessarily the best situation. This is mainly the case because other factors come into play and it’s not as simple as translating 1e- into some (discrete!) number of ADU. Basically we are talking about read noise here then and especially for Canon this read noise tends to be high for low ISO values. Therefor I’d suggest to avoid them when working in low light situations. I’ve known people wanting to trade in there Canon because the thought it was faulty because of the noise they were seeing at low ISOs (seems counter intuitive). But it was simply a matter of a low light situation in which read noise was very clearly visible. Upping the ISO solved everything.
So basically my suggestion for (low light) film would be the same as for astrophotography; there is an ISO value for which you’ll have the best trade off between dynamic range and (read) noise and that’s the ISO explained in this article. So for your practical situation I’d say your approach sounds about right, even though ‘native ISO’ could be considered somewhere in between of 100 and 200!
And as always I strongly recommend to take this advice and articles as starting point for your own tests and see for yourself what works best!
I hope this helps
Hi Chris – great article! I found it intuitive to follow and mirrors what I see with my Canon DSLRs. The DR values on you rgraphs – are these ADUs? A true ratio or a log ratio for the Y axis? Thanks…
Excellent question 🙂
Data is from DxOMark en for Dynamic range they use Ev and Log ratio as far as I can tell
Hi Chris,
I have been searching for the best ISO info for my Canon 6D Mk-II. But, your other link (for Canon) cameras has not been updated with the latest models for a few years. Will you be updating this soon? Thank you very much for all your time, explaining this info the the masses. Take care & All the best.
[…] lens should be able to do 18 seconds without any significant trails showing. Work out what the best ISO is for your camera to make it as sensitive as possible without introducing too much noise. Something between 800 and […]
[…] lens should be able to do 18 seconds without any significant trails showing. Work out what the best ISO is for your camera to make it as sensitive as possible without introducing too much noise. Something between 800 and […]
Hi Chris,
Great series of articles. Can you verify the Canon 80D numbers? Dynamic range seems very high, and I’ve not read that the 80D is “ISO less”. I see Sensorgen is offline (the Wayback Machine shows last data from 5 Oct 2018- without the 80D). Does the info from Photons to Photos provide a verification resource? Not sure if its info is a direct correlation, but the 80D doesn’t seem to have an “ISO less” input noise/ISO level response (Input-Referred Read Noise Chart) like the Nikon 7000 series.
https://www.photonstophotos.net/index.htm
Hello.
I am trying to understand all this. I confess Is not easy.
For example. With my Nikon 5300 modified, this theory tell me I should work with 200 ISO.
So if with a CANON I can Use 1600 ISO, I would say it’s a lot better, because with 200 ISO all I see is black.
If I use a good star tracker with a big exposure time, I anderstend.
But in a fix tripod, shooting a landscape with 200 ISO . I don’t see anithing in my screen.
Even if I could get something good working in the photo edition. I only can be sure if the photo is good or not when a get home to do it in my PC.
Is hard to believe this rule is the same for all kind of astrophotography and works with all kind of optics.
Sory for my bad English.
And I am not saying this is wrong. I just need to understand to make the write choice.
Thank you.
Hi Chris
Thanks for your very informative article.
I have a Canon R5 and I am wondering which ISO settings you would recommend for optimal resolution of the faint stars.
I have already established F2.8 and 15 sec exposures work best for me.
Thanks and best wishes