ISO comparison – various Canon and Nikon cameras
I had a number of people ask about more details about the Canon 6D and whether I would recommend the Canon 6D (affiliate), or Canon 5D mark III (affiliate). It’s tough enough to give advice at best of times, since the decision to buy a top-notch camera is a nuanced one. There are so many factors that come into play – your budget, weight of the camera; ergonomics; features & specification. Everyone has a different requirement of their camera gear.
So when I was able now to get my hands on a broad enough selection of Canon cameras (Canon 5D mark II / Canon 5D mark III / Canon 6D / Canon 1Dx (affiliate) simultaneously, I decided to also add the Nikon D4 (affiliate), and Nikon D600 (affiliate) into the mix. One would expect that the Canon 1Dx would beat the Canon 5D mark II hands-down since there is a generation difference in technology as well as a massive difference in price. Similarly, one would expect the Canon 1Dx (affiliate), and Nikon D4 (affiliate) to compare favorably to each other.
In this review, we are going to compare the ISO performance of these cameras:
- Canon 5D mark II
- Canon 5D mark III
- Canon 6D
- Canon 1Dx
- Nikon D4
- Nikon D600
Now, as I said, the choice between cameras depend on a number of factors – but one of them that becomes important in certain areas of photography, is high-ISO performance. Instead of relying on my say-so, and a few 100% crops, I decided it might be interesting if everyone does a bit of homework for themselves, and scrutinize the relevant RAW files. This would help in making the decision a personal one. Download the RAW files from here.
You can download the RAW files, via Right-Click and Save-As to your computer. They have been renamed in an self-evident way. (The last 4 digits are from the original file-name.) Be prepared though that this might hit your bandwidth limits with your internet service provider, since these files are quite large!
I shot sequences of images (of the same castle), with all 6 cameras, starting at 400 ISO all the way to 6400 ISO, in full-stop increments. The cameras were on a tripod. I used the Nikon 24-70mm f2.8G AF-S (affiliate), and the Canon 24-70mm f/2.8L II (affiliate) on the respective bodies. I tried to keep the framing as exact as I could. In terms of camera settings, I changed the shutter speed in full-stop increments as I changed the ISO. I kept the aperture at a constant f/8 and do keep in mind this isn’t a lens test.
I purposely photographed the shadow side of this castle, so you can see how the high-ISO noise looks like in the darker shadow areas. There is also enough detail in the image so you can figure out how the higher ISO settings affect image detail.
You will notice that for some images, I changed the shutter speed by 1/3 stop lower. This is because despite me working as fast as possible, the light did change subtly in the 3 or 4 minutes in which I shot the initial sequences for each camera. So I repeated several sequences. Therefore, the images you see here, are images that to my eye looked to have the same brightness. In other words, I tried to compensate for the slight change in light levels as I shot the sequences. I know, I know, it’s not scientific, but this is as fair as I could make the comparison.
Also, be aware that I shot with Shade WB, and this differs quite a bit between how Canon and Nikon interprets that. So for your own comparison, change the images to some specific Kelvin setting. (The beauty of RAW files – these parameters aren’t fixed.)
A few observations
1. To my eye, the Canon 6D and Canon 5D mark III behaved similarly in this aspect – high-ISO performance.
2. There is a definite (but not that an incredible) jump from the Canon 5D mark II, to these two cameras.
3. The Canon 1Dx looks to be at least a stop better than the Canon 6D and 5D mark III.
4. The Canon 1Dx and Nikon D4 look very close in performance here.
5. The Nikon D600 performed surprisingly poorly, especially if you think of how fantastic the Nikon D3s and Nikon D4 perform, and how the Nikon D3 was the high-ISO king for quite some time. The D600 that I had, did have the latest firmware, so I can’t explain the noticeable poorer performance compared to the other cameras.
I would still recommend that anyone that is shopping and choosing between the Canon 6D or Canon 5D mk III, to do some more research and decide whether the better auto-focus specs of the 5D mark III (for example) is worth the extra investment.
That’s good advice anyway – to do some deeper investigation when you’re looking for a camera to buy. As I mentioned, the high-ISO performance is but one aspect of the whole process. Hopefully the images here will help.
- High-ISO performance – Nikon D750 / D4S / D4 / D810 / D610
- High-ISO performance – Nikon D4S / D4 / D3s / D610 / D700
- Canon EOS 5D Mark III – high-ISO noise performance
A little bit of (further) homework
How might the difference in sensor resolution between the various camera influence enlargements of a photo? For example, how will the different resolutions impact the appearance of the high-ISO noise when you enlarge to 24×36 inches at 300 ppi?
21 Comments, Add Your Own
1David Tong says
Interesting and thanks for throwing in the RAW files as well.
In practice, however, like web gallery or 30″ print, are the differences glaring enough to warrant one over another?
I got tired of comparing cameras these days as I feel were reaching diminishing returns.
2Neil vN says
I also think that we’re starting to see a plateau in this area, especially if you compare a succession like the mk II and mk III.
Yet, it would also appear like the camera makers keep their special juju for their top-of-the-line bodies, because the 1Dx and D4 are noticeably better. It also has to do with those two cameras having less resolution.
Which brings me to another point – one of the reasons I chose the D4 over the D800, is that for my work, the super-high resolution of the D800 is a negative.
So the more conservative resolution of the D4 is an advantage for me in two ways – better high-ISO noise, and more manageable file size (storage & processing speed.)
3Daimler Naranjo says
Hola Neil me causa curiosidad la canon 5d3 y la 6d veo que disparo primero la 6d y posterior 5d3 la misma configuracion y con una diferencia de casi una hora, pero la 6d se ve mas clara aunque todavia no subia el sol, y lo que me deja impresionado es que tiene mejores detalles que la 5d3. Revele en Camera Raw 7.3 y con los mismos ajustes.
Muchas gracias Neil interesante articulo.
I agree with your ISO assessment Neil. The 1DX and D4 are very close in noise. The files for the 1DX are sharper across the frame when compared to the MK III and the D4. You stated that it was not a lens test but the crisp files of the 1DX just jumped right out when I zoomed in a little. I was very surprised to see such a large difference when compared to the 5D MK III and believe that we can draw more conclusions than just noise being better on the 1DX when comparing it to the MK III since the same lens was used.
I also found the color rendition of the 1DX to be better when I pushed the shadows to 100+ and exposure to 1+ in LR4.2. The greens in the door and trees just stayed much greener than the MK III and the D4. The shadows and details in the trees for the 1DX was just on another level. Was a little surprised to see the D4 fall behind here since Canon tends to be behind Nikon in this area. The D800 is light years ahead here (DR as well) of the MK III and expected the D4 to win this when pushing shadows and exposure.
All in all, I do not suffer from GAS, but find myself wanting a 1DX based on being able to push exposure and shadows. (and that detail too!)
Did not look at the 6D files since I already own the MK III and would never trade the AF system for the 6D’s.
I forgot to mention above that I compared the 3 files at ISO 6400. Just now the ISO 400 files showed the same thing-when pushed in exposure and shadows, the 1DX pulls ahead.
Neil, do you think that the 24-70 II is just that much better than the Nikon 24-70 or is something else going on here? Again, I realize you said this is not a lens test but the difference in detail is just too great to not warrant a side discussion since both the 1DX and MK III show much greater detail/sharpness.
PS: I used a custom WB for all files. In fact, all files were synchronized for everything before exporting.
6Neil vN says
This really is not a lens test. I didn’t take particular care with the focusing, since I wanted to fire off the frames in quick succession between the different cameras. Even though f/8 should take care of DoF, I still wouldn’t make *any* quantitative observations about lens quality here.
Edited to add: I also just realized that I had a UV filter on the front of the Nikon 24-70mm lens. So there really is no way that a comparison can be made, since I didn’t have a filter on the Canon lens.
I guess there wasn’t any opportunity to acquire a Nikon D700 for this test to examine the Nikon generation gap. I suspect that since the D700 was based on a D3 sensor, the D700 also wins against the D600 in high ISO performance. (I’m still using a D700.)
I grabbed both shots from 1Dx & D4 @ 6400 ISO, then I did what Rudy stated in LR and yes, in LR the Canon looks that bit punchier, noise was roughly the same to me, but I then started to think what the files would look like edited, after all, it’s in the edit and how a file responds to that edit that is the end product you are wanting, especially in Sharpening.
So, I edited both of them *precisely* the same way in Photoshop.
First up, both files were opened in ACR then dropped exposure by -0.5 Ev, trying to bring back some sky detail, in hindsight it did not avail, but it did bring back the front part of the pavement better since I knew what I was going to do with them in Photoshop.
I left the WB “as shot” on both, used Camera Profile/Calibration Adobe Standard and 2010, took off contrast in the Tone Curve Panel [left default 50% Brightness and 25% Contrast in the Basic panel like all RAWs normally are in Profile 2010], no sharpening, no noise reduction, no vibrance/clarity/saturation; then opened them up directly out of ACR into Photoshop and edited both precisely the same way.
The D4 file looks better in my opinion, the tree punchier, and the stonework is also better.
However, the key factor here is how each reacted to my editing especially with the sharpening. Big difference.
With the D4, the noise is less, I see more noise in Canon, look at the stonework and ground floor windows near the tree on both @ 100% and tell me if I am mistaken.
So, I think the D4 resulted in a better processing and took to sharpening a lot better than the Canon. [nope, aint’ bashing, I own Canons along with D3s’]
As stated, both files edited precisely the same way to the exact percentile on Highlights/Midtones/Shadows with actions on bringing back shadows and retaining highlights, then do the main editing as normal, all done with same % settings as each other including a 40% brush of an action on the stonework on each file to punch/pop it up – Secret Sauce in Neil’s Photoshop action set, at 30% layer opacity.
In the end, you have to make up your own mind. Both are good choices.
Here are the images. Click through (or right-click / save-as) to get to the high-res processed JPGs.
Canon 1Dx – 6400 ISO (processed RAW)
Nikon D4 – 6400 ISO (processed RAW)
you made a great point so I went and processed via my preset in LR. I then looked at the files and aside from VERY minor color differences, the photos were identical unless I zoomed in. I did, however, use the same custom WB for the files since a different WB actually changes the histogram and thus exposure (to a small degree). I also matched histograms as close as possible since Neil mentioned that the light was changing. In the end, as processed via my preset, the files were close. Had the focus been dead on for the D4, I would expect do difference real difference in sharpness.
As far as the processing, your D4 version seems to be a lot darker in the shadows. You can’t see any detail in the shadows by the tree on the left when compared to the 1DX files. Also, look at the museum sign on the right-the transition from light to dark is very abrupt for the D4 file. I think this is why you are seeing less noise-the dark areas are hiding it better.
Again, when comparing mine processed, the files are, more or less the same. I will stick with my initial assessment that when pushed in exposure and shadows, the 1DX performed better here. And no bashing here either :) I am saving for a D800 for my landscape shooting. The 5D MK III is good but when you push the shadows a little, the files fall apart (other MK III shooters don’t like when I say this but it is the truth). Banding is visible and it drive me crazy because it makes exposure blending very difficult. There are also a few other aspects about the D800 I like better for landscape work but for portraits/weddings, the 5D MK III would (most likely) remain my preferred choice…unless I win a 1DX!
You are right with more shadows from D4 edit, so I re-opened both RAWs to compare again, and Doh!, although I stated everything I did in RAW was the same, in the Nikon RAW I forgot to take off the default blacks of +5 in the Nikon File which is a huge difference, as I did in the Canon file.
So, the shadows did open up more when I processed it again [re-posted above] I got a better result again, same as before but cleaner open shadows.
Yes, 5D MkIII is a beautiful camera no question.
EDIT: Rudy, I have now re-done both to same specifications as listed in my first post correcting the oversight of leaving +5 Blacks on in the Nikon File.
thanks for going back and doing that. The files are now so close that pixel peeping is pointless :) ISO 6400 just looks so good these days…
One other thing that I find interesting here is the apparent identical dynamic range of the cameras-DXO will have you believe that the DR of the D4 is 1.3 stops better than the 1DX. Yet the files look identical as far is DR is concerned. One of the reasons the D800 interests me is because of the claimed 14.1 stops of DR when compared to the 11.9 of the MK III. If it is all lab results, then I will stick with my MK III and find a work around to the less than stellar shadow recovery the MK III has in the dark areas. A rental of the D800 might prove more revealing than the DXO information.
Neil’s post here has proven quite revealing even if we went off on a tangent (no pun intended)
I do think you should hire a D800 before buying to make sure it suits you. I have one and like it very much. There is definitely a noticeable increased dynamic range at ISO 100 over any camera I have tried. But that falls off sharply as you go into higher ISOs. If you are happy shooting at lower ISOs – which is what the camera is optimised for – then you will love it. But don’t expect 14 stops of dynamic range at ISO 3200.
13David de Fortier says
The 6D climbs ahead of the 5Diii at ISO 12800 which is why i prefer it for low light shooting. Its also a hek of a lot cheaper.
This has been an interesting experiment Neil, thanks for the RAWs and some ‘homework’.
I just processed the Canon 1Dx; 5D MkIII; 6D and Nikon D4 400 ISO files.
Verdict. They are so damn close no noise at all, each image had WB adjusted to get them as close as I could possible match each other in terms of the stonework/pavement. I also had to adjust exposure slightly on each to once again bring up pavement, but also to match each other as close as my eye could detect, but I am talking 2/10ths of a stop here.
[once again all the same % on shadows/midtones/highlight].
In fact, the only thing I could say was that the Canon 6D verrrrrrry slightly nudged out the rest of them in terms of sharpness via my processing. Very slightly more detail/sharper. Miniscule in fact and had to pixel peep at 100% on all images side by side.
However, it would depend on the precise focusing point used, a window, corner of stonework, as I don’t know where the point was.
Oh, here is some more homework. In the Nikon D4 400ISO file there is a ‘hidden item’. Spot it and win . . . well satisfaction you could say. :)
15Jon Lloyd says
I reckon the hidden object is one of 3 things:
1. The man behind the tree with his back to us
2. What appears to be a shoe or a boot in the 5th small arch from the left on the tower (Left)
3. The Canon 24-70mm f/2.8 IS II L from my bag… dammit.
Hope you are well my friend!
16Jon Lloyd says
You could say I “won” my 5D3. It’s more of a beautiful thing when it is also free! I would however willingly pay the $3k for it now I have used it for 6 months. I still use my 7D – but it has been seriously demoted!
Yep, give the man a cigar, guy hidden behind tree.
Dunno what is that is though in the arch on tower.
Did not define it properly until I opened the shadows.
I am well and thanks for inquiring.
my 7D is now my wife’s camera and serves as a back-up of wildlife camera. It is still a joy to use but lacks the refinement and ease-of-use of the MK III and it really is worth every penny.
main reason for adding the D800 would be for landscape use and thus low ISO but it would need a serious real-life increase in DR for me to buy it. If it saves me from having to bracket exposures and then blend them later then I am in but I will do a side-by-side comparison first. We will be in the Adirondacks for about 10 days and I might rent the 800 for a few days to compare the DR. Thanks for the input.
19Chris K. says
Thanks Neil very nice, I have the 5D III and love it but when I get back to the States in Dec, I will be buying the 1Dx to have along side my 5D III, you know long story short I was going to switch to Nikon D800E and when my friend bought one he let me try it out and honestly while it is a great camera i really didn’t see any reason to sell off my Canon gear for Nikon, DR ok D800 is better than 5D III but since i use my GND filters does not bother me at all btw who the heck likes underexposing by 5 stops anyway? I could never understand that, what I am trying to say is don’t get caught up in that forum BS talk, ask people who own both and they will tell you just how much they appreciate the 5D III.
20Stephen Russell says
Many thanks for the comparisons. Having a small collection of Canon EF lenses alongside a small collection of (mostly) manual focus Nikkor lenses, I’ve owned Nikon D200, D700, D600 models and Canon 40D, 1DsIII and 5D II models. My own findings show the D600 is no worse than the D700 in terms of noise, especially as the D600 can be downsampled to 12MP. Both Nikons beat the 5D II which beat the 1DsIII (although my best ever portraits were taken with the 1 series). Of course the full frame cameras trounce the APS-C cameras (not least because their maximum ISO settings are so much lower anyway). But I never found the D200 to be the noisy camera many others claimed.
So I find the poor performance here of the D600 a little odd. Came here for help in finding a Canon body (having sold the 5D II) in the summer.
21Andre NYC says
Thanks for a very thoughtful article. This Canon vs Nikon comparison is so complicated to address. To add my 2 cents based on following the tech over the years: Canon has decided to focus more on the video aspects than the still photo aspects, and this does show in their cameras for photos. Despite catching up on the ISO front after a very long hiatus (and I am glad I jumped ship early, a year after the D700 intro), it still isn’t quite up there even if the D1x is close enough to the latest Nikon offerings, for example the D750. This is mostly visible in the dynamic range area, see the DxO comparison which confirms what pros have told me https://www.dxomark.com/Cameras/Compare/Side-by-side/Nikon-D750-versus-Canon-EOS-1Dx___975_753