UPDATE: Results from our reader's poll can be found here! Last week we released our head to head competition review between the Sony A7RII, the Nikon D810, and the Canon 5DsR ultra high megapixel cameras. Our test put all three cameras up against each other and compared their performance in terms of ergonomics, HD video, auto focus, ISO performance, Dynamic range, and overall image quality when used in the studio. Unfortunately a bunch of Sony users complained that our final studio test wasn't up to snuff, so we did what any respectable review site would do and brought back baby turtle. The new results might shock you.
Here is the original video that caused such an uproar among Sony A7RII users.
The main point we were trying to get across with this video was that all three cameras of these cameras are unbelievable in image quality. In our opinion, anyone who switches camera systems that they are heavily invested in, is ultimately making a very poor financial decision for a very small difference in image quality. If you like the Sony sensor then simply wait 6 months and that same sensor will be in the next Nikon camera. If you want more megapixels, hold out because both the Nikon and Sony systems will catch up with Canon very soon. Want better video quality? Yep, I have do doubt all cameras in the near future will probably take 4k video as well. Switching brands 10 years ago might have made sense but with image quality reaching near perfection in the current crop of cameras, it seems a bit foolish to think that your photography will greatly improve after switching to any one of these three cameras.
Our Readers' Results
Of course that is OUR opinion but we wanted to to hear from our readers concerning the image quality from these three ultra high megapixel studio and landscape workhorses. In the camera comparison article, we posted three near identical images from the Sony, Nikon, and Canon cameras and asked you our readers if you could accurately tell the difference between the three files (all exported at 36 megapixels for fairness and anonymity). Below our the finds from our 3 poll questions and our 1 quiz.
Which Camera Took Each Photo Poll Results
The results from our Quiz "Which Image Looks the Best" is even more shocking. Again all three images almost split the votes evenly with the Nikon D810 actually receiving the most votes. So not only did more people find that the camera with the lowest megapixel count (Nikon D810) did in fact produce the best looking file, the massively impressive (and expensive) Canon 5DsR didn't really look that much different from the Nikon and Sony cameras both sporting Sony sensor technology.
Which Image Do You Think Looks the Best Results
With a sample size of 4,000 participants, I think it is safe to say "Who Cares?!? The mass majority of people cannot tell the difference between the images these cameras produce!"
But Wait, Sony Wasn't Given A Fair Opportunity!
Many Sony users were quick to point out that because we used the Sony 24 - 70 f/4.0 lens instead of the Tamron 24 - 70 f/2.8 VC lens that we used on the other two cameras, the A7RII actually had a disadvantage in our high resolution studio test. Honestly, Lee and I thought the native Sony E Mount lens would have actually given Sony the upper hand but depending on who you ask online, their 24 - 70 lens is either the worst lens we could have picked or one of the best lenses you can buy for the A7RII (yeah, go figure).
So in an effort to be absolutely transparent and fair with our testing, we ordered the Flagship Sony A Mount to E Mount adapter (The LA-EA4) so that we could run the test again with all three cameras having the exact same lens. Many people argued that we should have used this zoom or this prime on each camera but the truth of matter is, if you want to test the actual image quality out of each of these three cameras fairly, you really need to use the same lens across the board. This isn't a lens test, this is a relative sensor quality test.
Reader Test Take 2!
Below are the three files once again resized to 36 megapixels and uploaded in their full glory. We invite all of you to download the three images, compare them closely, and give us your opinion on which image looks the best. As we mentioned in the video above, the Tamron lens was significantly better than the Sony f/4.0 lens so this should help raise the A7RII image up quite a bit from the previous test.
Camera 1
Camera 2
Camera 3
Which Image Looks the Best?
Which Camera Took Which Photo?
Not sure the focus is in the same spot on all three.
We zoomed in an used Live view to pin point the focus. Also when the lens is shot at f8 there should be huge amounts of depth of field. Not sure how to really make the test any more accurate if it isn't here.
I appreciate the reply. Not trying to complain at all. I don't have a dog in this race. If anything I'd like the Sony to be a real challenger for Canon because I want the 5DIV to kick some butt.
Camera 2 has a different focus point than the other two. It is the only one the fibers sticking out of the top of the costume are not in focus. At first I thought there was just a lack of detail in the darker costume eye (right of frame) but it is actually just out of focus.
I've missed focus at f7.1 so I bet I could do it at f8 too. :)
Something else looks strange too. There seems to be some camera shake in her lips and the front of her costume in Camera 1. I didn't notice this before. It's late, my eyes could just be shot. Maybe that it is less than ideal rendering of high contrast areas.
Thanks for the retest.
FWIW, it seems worthy to note that the Sony did so well, even when using non-native lenses, via adapters. Pretty commendable if you ask me.
One could perhaps argue, that it's still not quote an evenly done test - as it seems very safe to venture even further superior results when using *native* Sony E mount lenses - no adapters.
For example, either the 55mm 1.8, or 90mm 2.8.
The subtext of the matter being the unfortunate lack of high quality zooms in the current Sony line up. But primes... IMHO, they have some gems in that category.
Cheers...
an "even test" and an "absolute highest quality" test are very different things. I don't think you can make much of an argument that this isn't even when all three cameras used the same lens (Canon and Sony used the actual same specific lens). I'm sure all the cameras would have better image quality if we shot with their most celebrated prime lenses.
Yup, I would agree with that comparison of test types, Patrick.
That said... I still think it's still more a case of "as even as possible". Just because they use the same lens doesn't mean the one using an adapter to use the Same Lens has no chance of increased performance - should say that same make lens have come in a native Sony mount version. Know what I'm saying?
Personally, especially if its a studio test... an all out native prime test would be interesting.
Especially one which added a secondary test comparing keeper rates utilizing each systems auto-focus system. (For studio work, the Sony eye-focus system is quite an advantage)
I know how this perspective must sound, but don't get me wrong, I've loved Canon (mainly the glass), & more recently Nikon systems (<3 D810) pretty deeply, but it's starting to become hard not to notice the cool things Sony is doing.
Anyway, also want to say I applaud the effort for the test, and retest. Got love for you guys, and appreciate all the hard work reporting on these matters.
Best,
AV
"For studio work, the Sony eye-focus system is quite an advantage" the eye-AF doesn't work with adapted lenses, and in a dim studio light the A7r2 AF works very slow and unreliable in general, i.e. may not focus at all in the small spot mode.
The 90mm is a very inconsistent lens according to http://www.lensrentals.com/blog/2015/10/sony-e-mount-lens-sharpness-benc...
The cameras are pretty close to each other with this retest.
I wouldn't mind a more practical "test" then something with studio lighting, such as seeing how they compare with drastic contrast differences, or super bright highlights, low light, etc.
We did those tests in the first video. It seems these cameras are being marketed as the high megapixel studio and landscape cameras so this is a pretty good test IMO.
Forgot to mention the 100% test sample photos from those comparisons or similar like the studio tests-been a hectic day and forgot to mention that part.
Pretty tough to really tell from those with the YouTube compression. But regardless those three cameras are pretty impressive
thank you guys for your effort! I'm relieved to see that the difference isn't that big overall as I almost fell for the marketing hype of Sony...
What will you choose then? The "hype" around Sony is not about image quality benefits, there are other things to consider.
canon1,sony2,nikon3...i could be all wrong. XD
Hi guys, any chance for a comparison between strobe lightings? I have been thinking to add more lights to be lights and wonder what would be the best choice. Currently shining with the B1s on location but need more lights for semi indoor and product work. I know D1 would be the easy answer but I am interested in the Move and Siros too from Bron because they have better performance with reasonable prices. Is there anyone who have used both brands of light extensively and can make a fair comparison? Lighting is a long term investment I just want the best I could afford.
I certainly wouldn't recommend having some Profoto and some Bron gear. That would be a pain for sure. I haven't used the Bron mono lights but and I love the Profotos but I dislike how loud their fan noise is. We can't use them to shoot video.
Lee I have a couple of D1 heads and noticed the fan noise. Thought it was just a faulty batch until I realised its actually how they sound!
Thanks Lee. I don't mind to completely switch to Bron of they produce the ultimate good light. ; )
My only concern with D1 is its slow flash duration for fast fashion/action shoot and high speed product shots. But I like how easy to use they are and their form factors. Quite a bit lighter than Siros.
If you have resources to interview people who use both systems that would be awesome. There's surprisingly so little info or comparison between lights. I guess most people just stick with one brand and never consider jumping ship.
Thanks in advance.
Thank you for the real world tests and the articles Patrick! Seeing DxOMark scores and saying which camera is better is easy, looking at test shoots and guessing which one is from which camera is really hard. :D
I think cameras are like sport cars. It doesn't matter if you have a Porsche, Lamborghini or Ferrari, you won't be using them at the top speed most of the year. They are all high-end performance wise, choose which one you like better. :)
Anyway my guess is:
1. Canon
2. Sony
3. Nikon
I'm thinking canon, nikon, sony, and if that's wrong, switch canon and sony. Nikon would have the poorest color rendition out of the gate from what I've seen, though the difference is almost indistinguishable after editing, and reversed when pulling the info out of the shadows. The nikon is using an older Sony sensor, so it stands to reason that there's better color rendition in the new one, and it should theoretically be the worse of the three.
If I'm wrong, oh well. The difference between these isn't enough to warrant switching systems, at least in a controlled environment.
Nice. Would you mind sharing the RAW files?
Even if it goes against my beliefs (20ish MegaPixel are enough, over this threshold the quality decrease) could I suggest to post image further scaled down? Like 10 or 16 megapixel? Obviously the Nikon D810 is the only one where the image are not rescaled. Flaws in rescaling algorithm account a lot for perceived lower quality.
Which camera took the pic 1 is really obvious: by looking at the shadow noise.
LA-EA4? Seriously!!??? You guys know that you are putting a semi-transparent glass between the lens and the sensor? This is the very same glass that is present on A99, which effected the image quality this or that way.
The final test did not use this adapter. We used the metabones adapter with the A Mount Tamron lens so everything would be even across the board.
It isn't about finding a great camera, but finding the right camera. I think the answer is that we have better gear available than most people's ability to use it. I don't print that big or often. None of my photos deserve 40+MP. Not many others do either. It has come to be a matter of features, convenience, and workflow.
Sigma makes their 35 and 50mm lenses for Nikon Canon and Sony right? Why can't you rent those lenses so that there isn't an adapter being used? I'm a canon fan, if that matters. =)
Sigma lenses are not for the Sony E mount, they are for the A mount.
And for this test you have to use the exact same lens! One lens, three camera. :)
Ahhhhh, another little "thing" that would confuse the crap outta me if I switched to Sony! =)
Now the same with the Sony 55mm 1.8 ... what i think is that the tamron is limiting all the cameras in sharpness. i don´t want to fanboy something (i got a A7 and a Nikon D800) but if a resolution-test is the goal, choose the sharpest lens you can get for each camera-system. i know thats also unfair because the lens decides at the end which is sharper ... at the end its unpossible.
Here's a shot I did a few months back with the Sony A7II and the Tamron 24-70 2.8 (A-mount). I'd say it does a pretty excellent job and that was only 24mp. Imagine with the A7RII ;)
Who gives a crap??? They are all ok. I won't use adapters and I have about 10 Minolta G glass lenses from when I was a Minolta shooter. Sony doesn't do it for me. I'm pretty much retired now and shoot more film with my Minolta 9000 that Minolta gave me years ago. As for the images, well they all look the same. I mean Hoo Hum. Why do these "photographers" get off on comparing? Aren't photographers suppose to be out recording events of today??? Call me old and yes I am but this comparing is wierd. My megapixels are bigger than your megapixels and my camera is better than yours boo hoo hoo cry me a river. I mean get real.
Ok I'm done. Back to the darkroom.
Have fun and enjoy your comparing.
Shouldn't you resize the photos to 50MP instead of shrinking them to 36MP to give a fair test of sharpness? Seems potentially unfair to the Canon.
Shrinking them to 36MP is the correct test for noise though.
No! Down sampling is the fair test. Slight edge for the higher native MP sensor in theory this way, but the fair method is this.
I'll be shocked if the Canon isn't #1. More shadow noise, less dynamic range, but slightly more detail (evident even web sized). #2 & #3 are close (which they should be, they're both Sony sensors). #3 seems to have slightly more detail and darker shadows so I'm saying that's the Sony but #2 has the more vibrant greens which is also a Sony hallmark/flaw so it could be either based on how they were processed. I didn't agree with a lot of the review but I thought it was funny and I do agree that in the end all three cameras have reached IQ sufficiency.
Thanks! Now I know that my cat photos will be the best 640x480 images FB has ever seen!
Hi Lee and Patrick,
Are these straight exports (except for resizing of course) from RAW, or have there been colour corrections done? I've used Canon, Nikon, and Sony before and they all seem to have different default colour casts or characteristics that make it somewhat easier to determine which took what. Canon always seemed to have a more reddish tint and Nikon seem to be more on the green side. The last D800 I had really had a "too green" cast which drove me crazy. I know everything can be fixed in post but you know, it would have been nice to just have nice colours SOOC.
By the way, just became a fan of Baby Turtle. She's so cute! :-) You should have her appear on more of your videos. :-)
Oops.. never mind. Just watched the other video again, and you did say that you tweaked the colors a bit to match all photos.
If there is one thing we can all agree on it is that no matter how well thought out the test, no-one will ever agree on the results.
Detailed lab tests covering all possible parameters = not a good reflection of 'real world' performance
Sample of 'real world' examples = non-exhaustive/biased/unfair test which does not take account of x y and z
I think the majority of people here would take any of them and I would be immensely happy with whichever they ended up with. At this level it seems to be more a case of which works better for you in practice - so basically these results were useful in just removing any major doubts over general image quality - making the choice much much easier!
"So in an effort to be absolutely transparent and fair with our testing, we ordered the Flagship Sony A Mount to E Mount adapter (The LA-EA4) so that we could run the test again with all three cameras having the exact same lens."
You don't seem to get it right? Instead of being professional's you just seem like amateurs. First you don't learn to use the cameras you are testing and then you use most of the time of your orginal "review" video to bash Sony. Real professionals! Then you do retest baecause of rightful complaints about your biased "review". And what do you do? You do the retest also wrong, but what can you expect from amateurs? Instead of using LA-EA4, which adds semitransparent mirror to system, you should have used LA-EA3 adapter. Thats without semitransparent mirror. I suggest that next time you try to learn to use cameras before you do testing. Otherwise you may end up looking like idiots, like you did this time.
Did you even watch the video? The final test didn't use that adapter.
I'm going to contact Sony on your behalf and tell them that "Mika" from the Internet says that your flagship adapter should be discontinued immediately.
It's flagship adapter because of the PDAF ability it gives to cameras. That doesn't change the fact that it has semi-transparent mirror which should not be used if we are doing silly pixel peeping tests.
Completely right. Flagship adapter.... that funny. What makes an adapter a "flagship?" A bigger number? The 3 is the one to get. It has been well known since the a55 days (over 5 years now) that Sony's SLT mirror reduces resolution. How do we know this?A noted member on Dyxum removed the mirror. Look, you guys are trying hard, but you just don't know the tech. These aren't happy mirror slapping dslr's. Reach out, and ask some experts. Lots of them out there.
Once again... We didn't use that adapter for this test.
Then what am I not reading correctly? This seems pretty unequivocal.
So in an effort to be absolutely transparent and fair with our testing, we ordered the Flagship Sony A Mount to E Mount adapter (The LA-EA4) so that we could run the test again with all three cameras having the exact same lens. Many people argued that we should have used this zoom or this prime on each camera but the truth of matter is, if you want to test the actual image quality out of each of these three cameras fairly, you really need to use the same lens across the board. This isn't a lens test, this is a relative sensor quality test.
We used the LA-EA4 adapter to test all three lenses on the Sony camera alone. We found that the Sony 24-70 2.8 lens was the best of all three (it might have been even better with the EA3 as you suggest).
However, once we tested all three lenses on the Sony ALONE, we then switched the adapter to the Metabonez with no element and used the Tamron 24-70 2.8 VC (Canon version). This allowed us to compare all three cameras with the same model lens and for the Sony and Canon it was the actual same exact lens.
So yes the Sony A series glass was the best, but to keep everything fair, we compared all three cameras with the same Tamron lens. The results show that few could tell the difference in cameras but the Nikon was found to be the best looking image by number of votes cast.
Then you have a very misleading sentence in your article. In the paragraph I referenced, it reads that the la-ea4 is responsible for the use of the Tamron lens. No reference at all to Metabones.
Wow, you girls just don't get it.. It might have been unfair, if the test had actually been performed using the LA-EA4 adapter, but it wasn't. The pixel peeper test was with the Tamron lens and Metabones adapter, which to my knowledge does NOT have any mirror in it. It is just a metal adapter ring with slow AF and no glass.
I suggest you watch the video properly next time before you give one of the founders of FS shit for something they put a lot of work and thought into.
...
Oh, and I guess you owe Mr. Morris an apology your little brain fart.
so u did this re-test on my birthday :p
but honestly who r these pixel peepers? im sure half of these complaining or arguing dont even own a dslr ... like someone said, real photographers r busy doing photography and spammers, fanboys wanabeeez n keyboard warriors are fighting this war for their brands
im a canon user and everytime i see such reviews one half of my brain curses me to switch to nikon but then other half says, "ur dumb, ur pic sucks not because of camera its because u suck n u dont know how to use that camera which is in ur hands and u think that the one ur not having will make ur photography better"
so in short, who really cares? these pictures look stunningly similar and amazingly sharp and without metadata its really an impossible guess to make.
im so excited that sony is giving tough time to nikon n canon DSLRs with a mirrorless camera. 3 means more competition and better features for lesser cost. In few years Pentax might also join the race but in the end its us consumers who r getting better value for our money.
So i totally agree to ur ist review conclusion, stick to what one has or else buy all three systems and find out urself who is best. For me these pictures look similar, i just want to know one thing though, high ISO test hints that sony pics r still not 14bit raw? is it the case or it really lacks that much dynamic range
Number 2 seems to have warmer skin tone. Number one has more detail. Number 3 is cooler with more detail than 2 and seems to have most contast. Of course, the rescaling will have messed with much of these subtle differences. Bottom line they are all very good and perfectly usable. I would have thought the test should be re-done with a top flight prime, such as an OTUS 85/1.4. That would be the nearest thing to eliminating the lens from the test and let us see the detail the sensor can record rather than the lens being the limiting factor. I would love to see the photos in their full size, without the size matching, to eliminate the effects of the rescaling. I know this would have given the game away but I have cast my vote now and partly I was assuming the sharpest was the Nikon becasue I know the smallest image would not be rescaled, so it did not really work for me, even if I got it wrong I was guided by what I knew as much as what I could see.