From the mailbag:
I have difficulty choosing between the D810 and A7RII. The A7RII has many features that I like but there are things that are not crystal clear to me. Since you have selected the D810 and A7RII as your favourite gear, I assume you know them inside out. I would really appreciate if you could help me to decide. For instance, which one has better AF accuracy? If I take pains to do AF fine tuning, which one is more accurate?
The other thing is the effect of sensor stacking on adapted lenses. I have the Zeiss 135mm APO and it’s an incredible lens but I have to put a loupe on my D610 screen whenever I want to use the lens hand-held which makes the combo unwieldy. With A7RII EVF and peaking it would be much more convenient but different sensor stacking gives me the creeps. Nobody knows if Nikon and Sony have the same stacking (most probably not). I do not know if this affects the lens performance. Do you get the same sharpness when you adapt your Nikon mount Outs lenses to the A7RII? I would really appreciate it if you could give me some information on this. I will also post this in the technology forum to see if somebody has investigated the effect of the sensor stacking on the lens performance.
That’s a complicated set of questions. I’ll try to do them justice, but it may take a lot of words. Let’s get some simple things out of the way first.
The D810 is no longer one of my favorite cameras, if number of exposures per day is the criterion. It was my favorite for precise, detailed photography until the a7RII came along, but the a7RII does better the things for which I prized the D810. My most often-used other camera besides a7x cameras was the D4, and is now the D5. A Venn diagram of what the D5 does best and what the a7RII does best would not have much overlap.
Well, that’s it, right? I’m recommending that you buy the a7RII, right? Emphatically not. The D810 is a remarkable camera, and has traits that surpass the a7RII in some applications. I’ll get to those traits soon, but remember that there are many applications for which the D810 is the better tool. There’s also a consideration of what your other cameras are and how varied your photographic work is. Since I have the D5, I don’t have to try to use the a7RII for some things that the a7RII does poorly and the D810 does better, since many — if not most — of those things the D5 does even better.
The other thing that’s easy to get out of the way as a practical matter — though not, I admit, from a test-chart-measuring perspective — is whatever differences exist between the Nikon F-mount digital cameras’ sensor stack thickness and the Sony a7x stack thickness. Don’t worry about that at all. I did a series of tests a while back that attempted to answer the question of whether the a7R or the D810 were sharper when used with the Otus 55/1.4. I finally gave up. Whatever differences existed were drowned out in experimental error. I now have a electronically controlled focus stacking approach that could reduce the error and perhaps allow conclusions to be drawn, but without doing that, it’s pretty obvious that whatever differences exist are beyond mattering in normal photography.
It sounds like you have a Nikon body. I assume you have some Nikon lenses. Here’s a point of difference between the two cameras that may be dispositive for you. At present, using Nikon G lenses on the a7RII is an imprecise, unpleasant experience. There are adapters that will control the diaphragm, to be sure, but you can’t set it to any accuracy. I end up opening the lens all the way and stopping it down while looking at the exposure meter or the shutter speed in A mode to see what aperture has been set. If you have Nikon E (for electronic diaphragm control) lenses, forget about it. You can only use those lenses wide open. There is an adapter that purports to fix all that, and autofocus to boot, but it’s currently shipping in small quantities only and early user experience has been rocky. It’s possible, even likely, that we’ll get to the point where using G and E lenses on the a7RII is as seamless as using Canon lenses, but we’re not there yet.
OK. Now let’s sort through the differences between the two cameras. There are many, in spite of the fact that they are superficially similar if you’re looking at spec sheets.
Sports, wildlife, and AF
Since you brought it up, we’ll start with autofocus. In my experience, for static subjects there is no more accurate autofocus than that of the a7RII. However, in general, the D810 AF system, which has a lot in common with the D4’s, is better at dealing with rapidly-moving subjects. In addition, pro-grade Nikon lenses tend to focus faster than the Sony E-mount lenses, in some cases a lot faster.
Thinking of autofocus and action makes me think of sports, and sports makes me think of continuous shutter operation. It is true that the a7RII offers that mode, and 5 fps is fast enough for many uses, but using continuous shutter mode for more than a couple of shots at a time renders the finder nearly useless. It is odd that a camera with a flapping mirror can provide a clearer view at high frame rates, but it certainly does. In addition, there is a lag in painting the a7RII EVF or LCD screen that is normally inconsequential, but not when photographing fast action. The D810 “paints” its finder at the speed of light.
If you want to use a fast, long lens (the classic sextet: 200/2, 300/2.8, 400/2.8, 500/4, 600/4, 800/5.6) at present, you’re going to have to adapt something. With the D810, you can have just about any focal length your back and your pocketbook can stand.
Just as the D810 is the clear winner for sports photography, the a7RII is the champ when lens choice is the main concern. Yes, you should probably cross Nikon G and E lenses off your list for now, but you can add Leica R and Canon glass. Many Leica R lenses put their Nikon counterparts to shame. However, since the Sony full frame mirrorless cameras have come on the market, the price of the best of those lenses has been bid up to stratospheric levels. You can also use Pentax SLR lenses, of which there are many. Then there are rangefinder lenses, especially Leica M-mount ones. There, the story is mixed. Anything over 75mm will work just fine. When going shorter than that, it’s wise to look at compatibility on a case-by-case basis. The a7RII , with its back side illuminated (BSI) sensor, has fixed the corner color casts of previous a7x cameras, but not the sensor stack influenced corner smear. Shorter lenses will tend to have more problems, Faster ones, too, but there are plenty of surprises when it comes to testing various rangefinder lenses on the a7RII.
Both the D810 and the a7RII have many different metering modes, and, as these things go, they both do well in that department. However, I consider camera metering to be fairly crude at best, and useful only to get you in the ballpark of the right exposure. The classic digital camera exposure tool is the histogram, and the two cameras handle it in very different ways. In the a7RII, should you choose the right option, a luminance-only histogram appears in the bottom right corner of the EVF and LCD. This is a “live histogram” in that it is updated as fast as the image on the finder, allowing you to see changes immediately. While it is subject to error, both from the fact that’s it’s not a real raw histogram (which neither camera offers) and from the fact that you’re looking mostly at the green channel, with a some red and very little blue influence, it is remarkably useful.
A reader has pointed out the a live luminance histogram is available on the back LCD of the D810 if you get into exposure preview mode by pressing the OK button, then display a live histogram by pressing the “info” button (you may have to press it more than once depending on where you left it in the cycle of options). It’s not available in the optical finder, of course.
Both cameras offer luminance and three-channel histograms during playback, so you can use the traditional “make a test exposure, chimp, and look at the histogram” approach to exposure, which is far more cumbersome, but serviceable in most circumstances.
Another exposure tool that I’ve become fond of — although I resisted it early on, for reasons that escape me today — is the overexposure zebras on the a7RII. Even at 100+, which is their least invasive setting, the zebras are pretty conservative, but they point out regions of potential overexposure in a way that’s hard to ignore. After having blown a couple of shots because I didn’t notice a thin tail on the histogram, I leave the zebras on except when I’m in a static situation.
Having said all that about exposure, I’d be remiss if I didn’t point out that, at lowish ISO settings, both cameras have enough dynamic range that underexposure is rarely a problem.
I should also mention that the D810 is pseudo-ISOless, and the a7RII is so at ISO 640 and above, although I don’t think that would affect a purchase decision.
Finders and focusing
The D810 has an optical finder and live view through the LCD. The a7RII has an electronic finder (EVF) and live view thorough the LCD. While such an arrangement could allow Nikon to offer live view focus peaking, they have not done so. The a7RII’s LCD is articulated, and the D810’s is not. A world of differences in operation flow from these design decisions.
The EVF is sometimes difficult to see well in bright light, particularly if you wear eyeglasses. The EVF is much easier to use in dim light.
Sony’s auto switching between the EVF and the LCD can be frustrating when it gets it wrong, which it does a lot outdoors if you’re wearing a hat. Fortunately, you can set the camera up for manual switching.
I mentioned EVF lag above.
The articulation of the a7RII LCD makes many shots easy and comfortable that would be awkward and back-straining on a D810. In particular, using the camera at waist level or a little higher works great, allowing smaller lighter tripods to be used, and larger ones to not be extended as much, offering more stability. The way that the articulation works means that these advantages only pertain when the camera is in landscape orientation, however.
As the cameras come configured from the factory, the Sony focus magnification is easier to access than is the D810’s live view magnification, and, of course, it is available when using the viewfinder, which is not the case with the Nikon. However, as a reader pointed out in a comment to this post, it can be as easy on the Nikon if one sets menu item f2 to assign the “Zoom On/Off” function to the center button of the multi selector.
The combination of focus magnification and live view means that the Sony offers the most accurate manual focusing system I’ve encountered short of the Betterlight’s system, which requires a special focusing target. The D810’s magnified live view is a huge improvement over that or the D800, but is not quite up to the a7RII standard, even leaving out focus peaking.
The a7RII allows accurate focusing wide open with the camera hand held (though you’ll have to stop the lens down manually before tripping the shutter). The D810 does not. The a7RII allows accurate focusing at taking aperture with the camera hand held. The D810 does not.
Controls and menus
The D810 is a serious photographic tool designed by engineers who know what serious photographers want in a camera. The controls are large enough to be operated with gloves. The buttons and wheels move a fair distance before a change is activated, and the clicks and detents are robust and distinct. You don’t have to go into the menu system a lot. You can do most common functions by pressing a button, moving either the front or back wheel, and looking at the LCD display on the top of the camera, which is easily seen no matter how bright the ambient light. You can even format a card that way. I think it’s a great system; not as great as the similar ones in the D4 and D5, but really good. In my infrequent usage of top of the line Canons, I have found their controls and menus similarly excellent.
When you do go into the menu system of the D810, it is fairly logically organized, with a hierarchy of categories, and a special user-configurable menu for frequently accessed features.
The a7RII is a serious photographic tool designed by engineers who are trying to please the rawest neophyte and the most seasoned pro. They end up failing the latter. I suspect they fail the former as well.
As an example of what I’m talking about, consider the mode dial on the top of the camera. In addition to the usual A, S, M, and P modes, there are dial positions for an auto mode that’s, I suppose, even more automatic than P, and a sweep panorama mode that produces in-camera stitched, low-resolution, JPEG-only images. To be fair, there are also two positions on that dial for recalling previously-stored sets of camera settings, which is very useful. There is also a dedicated exposure compensation dial, another handy feature.
The controls on the a7RII are lightly detented, with limited motion. It’s hard to use the camera with gloves on. It’s easy to move the wheels too far, especially when you’re in a hurry.
My big problem with the a7RII menus is that you have to use them a lot. To access commonly-used functions in the Sony a7x cameras, you press the Fn key, navigate to the function you want to change, and change it in a similar way to the D810. The difference is that access is more complicated, and that you need to be able to see the back LCD screen or use the EVF to see what you’re doing. To its credit, Sony has several buttons that can be configured for most — but maddeningly, not all — menu functions, but for many of those, you still have to look at the EVF or LCD to actually make the changes; the buttons just get you there faster.
Sony’s user-configurable, quick access menu system is more comprehensive and easier to access then that in the D810, and comes pre-populated with some reasonable choices, whereas the equivalent menu in the D810 is a tabula rasa.
And the organization of items within the a7RII menus is nonsensical. At the top level, there are not category names. Instead you have a camera, a gear, a box with what looks like a cartoonist’s depiction of waves emanating from it, four boxes, a playback button, and a suitcase. Functions are assigned with abandon to those categories. Here are some examples:
- Functions that control the display of information and pictures in the EVF and the LCD panel are in various places. You turn Setting Effect on and off in Gear 3. Finder and Monitor switching is in Gear 4. Monitor brightness is in Suitcase 1. So is viewfinder brightness and finder color temperature.
- There are focus-related items in Camera 3, Camera 4, Camera 7, Gear 1, Gear 3, Gear 4, Gear 5, and Gear 6.
- There are video controls in Camera 2, Camera 8, Gear 1, Gear 2, Gear 7, Gear 8, Suitcase 2, and Suitcase 3.
For the most part, you can get used to the Sony menus, and they won’t seriously affect your photography once you do so. Also, Sony is getting better and better with each camera they introduce. The new a6300 has some improvements that I expect will appear in the a7RII in a future release. And, to give a sense of how far we’ve come, consider the abysmal menu system on the NEX-5. I gave an old one away recently, and I had to set it up for an unsophisticated photographer, and I was amazed at how bad it was. Sony has shown they can learn a lot in a short time.
The a7RII has in-body image stabilization (IBIS); the D810 does not. Even if you use stabilized lenses, IBIS can give you sharper images because it can handle roll and translation, where optical stabilization can only deal with pitch and yaw. It’s particularly nice having IBIS on a camera that is so good with adapted lenses. However, IBIS and zoom lenses that are mounted with dumb adapters need a lot of fiddling.
The D810 had dual SD/CF slots; the a7RII has a single SD slot.
The D810 has better environmental sealing.
There are obvious differences in size and weight between the two cameras. Those differences can be evened out if you have a lot of big lenses, but I think that many a7RII owners will choose their lenses with an eye to lightness.
Both cameras have electronic first curtain shutter (EFCS), but only Sony’s is practical to use with the camera handheld.
Notice that I have not mentioned resolution or dynamic range. The resolution differences between the two cameras is small, with Sony getting the nod. The low-ISO dynamic range differences are not consequential for almost all uses, with Nikon in the lead.
They are both fantastic cameras. Mated with top-notch lenses, each can deliver image quality beyond the ability of medium format film camera under almost any circumstances, equal to large format cameras in many situations, and better than almost any film camera extant in dim light. That doesn’t mean that they can do everything with ease. No camera can. But both can do an extraordinary number of things well, and both are important advances over the cameras that immediately preceded them, the 17R and the D800E.
So, dear reader, there you have it. Let me know what you decide.
Tony A. says
I’d like to add a couple of clarifications regarding ease of use of the D810. You wrote “The Sony focus magnification is easier to access than is the D810’s live view magnification”. However it can be extremely easy on the Nikon if one sets menu item f2 to assign the “Zoom On/Off” function to the center button of the multi selector.
You also wrote “I can attest that the longer you’ve had a live histogram available to you, the more irritating its absence on the D5 and D810 is.” Actually, while in live view you can display a live histogram by pressing the “info” button (you may have to press it more than once depending on where you left it in the cycle of options). This one goes in concert with being in the “exposure preview” mode in live view, which the OK button toggles on and off. Full disclosure: some of the many uses of the OK button are not immediately obvious.
I hope those tips (which might also apply to the D5) may be of some use.
You are absolutely right. I will fix the post.
That works, too. I’ll fix it. Thanks.
Tony A. says
I accidentally left out of my previous reply what’s probably a pretty obvious point: I concede that there is indeed a complete lack of histograms when using the Nikon DSLRs in optical viewfinder mode! 🙂
Thanks for your complete answer to my question. There is one more thing that I’d like to know. Is the A7RII RAW tone curve linear? Ming Thein once said that “the D800E has a very linear tonal response, with low shadow noise, and highlights that clip fairly abruptly. This means you want to expose to the right just to the precise point of clipping; your dynamic range maximization comes from bringing up the shadows – since there is minimal penalty in color accuracy or noise. The D810 is the opposite: it is nonlinear and all of the extended dynamic range appears to rest in the highlights. What appears to be clipped even on the flattest picture control setting (the histogram is read off the preview jpeg embedded in the raw file) often still has a stop or more of highlight recoverability in ACR 8x, process 2012*. So you effectively need to have quite a bit of the frame blinking to make the most of the D810’s extraordinary highlight latitude. However, it’s still subject to the laws of physics, so this headroom has come at the expense of the deep shadows – whilst you can still recover a surprising amount of information, this comes at the expense of mainly color fidelity and separation. Recovered shadows somehow just look muddy – I suspect something in the ADC conversion has forced a ‘toe’ curve upstream of writing the raw file”
First of all, did you also notice such difference between the D810 and D800E with regard to shadow recovery? Second, I’d like to know how did you find the A7RII RAW tonal curve?
From a technical perspective, as opposed to the kind of poetic descriptions found in wine tasting, I find Ming Thein’s quoted content incomprehensible. Leaving aside some lossy compression techniques which can be used in both the D810 and the a7RII and are completely reversible, a CMOS sensor is inherently linear over most or all of its operating range. In all of the cameras I have measured, including the D810, the D800E, and the a7RII, I have not encountered a camera with nonlinearities that would affect normal photography. I think what Ming is probably talking about is how the histogram and zebras are computed, not the characteristics of the sensor itself.
I suggest you read this: http://blog.kasson.com/?page_id=2387
For that matter, can you comment on the highlight recoverability of various cameras? I haven’t tried a D810 file yet, but in my experience post processing D800 raw files, you essentially cannot recover anything in the highlights, the colors go crazy. On the other hand, I can gracefully recover almost anything on my 5D, without making ugly edges and stuff.
Is this a matter of white points, or color filtration and relative channel sensitivity, or what? I doubt any of them is actually significantly nonlinear like Ming describes (as referenced in John’s comment).
Lr/ACR is very clever when faced with one blown raw channel, at looking at the others and guessing what the values would have been had the exposure been less. With two blown channels (not both green ones), things get dicier. With three, its a crap shoot.
When you talk about the differences between highlight recoverability wrt camera model, make sure you’re looking at the raw files to see if they are both overexposed the same amount.
Given the huge low-ISO dynamic range of the D810, I don’t see any reason to select exposure settings that jeopardize the highlights, and thus I don’t use the Lr recovery feature.
Jack Hogan says
Highlight recoverability has virtually nothing to do with a sensor capabilities: it’s all software and they are all pretty well capable of achieving the same recovery when equally ETTR’d in the raw data.
Different cameras cannot be necessarily equally ETTR’d with respect to highlight recovery when their white balance multipliers and matrices are different; what do you ETTR with respect to? The brightest channel? The brightest part in the whole image of the second-brightest channel at any given pixel location?
Here’s a shot I just took with full color reconstruction on my 5D (the first one) and my GR (the second one): http://imgur.com/a/BRIhg
In the raw histograms (as shown by RawTherapee), the 5D is actually clipped *more*. Both have the green channel clearly clipped, the blue channel moderately clipped in the 5D but only barely clipped in the GR, and both are not clipped on the red channel.
Raw histograms: http://i.imgur.com/ye7sUlF.png (The 5D was at f/5.6 and 28mm to produce an equivalent picture)
Yet the 5D somehow behaves much better with color propagation highlight recovery in both LibRaw and RawTherapee, while the GR with only one clipped channel turns purple.
I’d say that the 5D effectively gains back a whole stop of dynamic range versus having no or poor highlight recovery, and that’s hard to ignore.
Jack Hogan says
Sensors behave pretty well linearly with light and the raw-> sRGB color matrices at DxO are practically the same for the 5D and GR – so the processing is looking mighty guilty there in my book, starting with white balance and the different profiles that get applied upon conversion… If you would like, send me an email via the form at ‘strolls’ with the two raw files and I’ll take a look at them. I may learn something new 🙂
John Lehet says
You didn’t mention IBIS. I had a D800e for a while, along with an MFT Olympus system. Because of the Olympus IBIS, for hand holding, the Nikon needed a lot more light. On many occasions I didn’t even think of taking the Nikon out of the bag and used the Olympus — just not enough light to hand hold that beast. With an 85mm lens on the D800 hand holding with a subject close to mid field I did not get close to 100% keepers at 1/200 sec. The same frame on the Olympus was completely reliable at 1/60 and probably reliable at 1/30. Because I was getting better results in general hand-holding the Olympus and had to throw so many shots away from the Nikon, I chose an A7rii to replace my D800e rather than a D810. I have far more keepers with the Sony than I did with the Nikon DSLR.
Also you say metering is about the same, but I found with the D800e the metering was heavily influenced by the focus point, while the Sony is a more balanced matrix. From what I was able to gather in deciding on an upgrade, the D810 has the same situation. Either a real matrix or a weighted spot like the D8xx can have its advantages, but I found with the Nikon’s metering, along with needing to chimp to see the histograms and then re-expose, that I had to struggle much more with the Nikon, which was probably more suited to studio work or sports. With the focus point influencing metering, I found I had to compensate the exposure when working with the same scene far more often than I’m used to. I could not take for granted that when working with the same light I would generally be in the same ballpark; the Nikon could vary the exposure wildly based on the focus point even when the scene in the viewfinder was the same or close to the same. Expose, chimp, turn the EV dial, repeat, refocus, expose, chimp, etc. Now with the Sony it is much more the situation that I’m used to; if I have a frame in the viewfinder but change the AF point, I do not have to think about changing the exposure. In years of photographing with Nikon DSLRs and more recently Olympus MFT and the Sony A7rii, I have never had to chimp so much and also lost so many exposures when I didn’t chimp as with that D800e. If I thought the D810 was different I would have been tempted, since I had Nikon lenses.
As one with mild presbyopia, even with bifocals I see better through the EVF than the LCD, and of course that is especially true in bright light. With a mirrorless camera, knowing things are just right before making the exposure and certainly just after without having to move the camera from my eye is a great advantage as well. With the Olympus I enjoyed manual focusing (Oly’s peaking is much better than Sony’s), and now with Sony’s magnification and good EVF I find I also can enjoy manual focusing again. Which is a good thing since one of my lenses is a Zeiss Loxia, and I might get another. One would not want to indulge very extensively in manual focus with a D8xx camera.
I do miss Nikon’s menus and controls. Oh well.
Thanks for your detailed comments. I don’t know how I could have forgotten IBIS. I’ll fix that.
Theodoros Kefalopoulos says
Jim hi and thanks for the comprehensive head-to-head. My question is about the quality factor of a photo, perceived as “photorealism”. Would you say that one camera is better than the other in representing a more natural view throughout the frame? One camera’s files to be more realistic and not that “digital” compared to the other? I know that this issue also depends on the lens used but I think you get the idea behind my question.
If we leave color and the image projected by the lens out of the picture, there are only two differences between the sensors. The a7RII has lower resolution. The GFX has a smaller effective fill factor. The first makes the a7RII’s files coarser, but not by much (1+sqrt(51/42)). The second means the GFX files are sharper, but more prone to aliasing.
It’s clear the Adobe Standard Profile for both cameras yields quite different color. However, I have no reason to think that if you profiled both cameras, you couldn’t get the color pretty darned close. The Fuji camera does have a set of Fuji-designed film-look profiles, which some folks like.
To me, the big win of the GFX is the Fuji lenses. The Sony and Zeiss native E-mount glass, while quite good, doesn’t come close if you’re going to make huge prints. If you’re not, save your money.
Oops, I thought you were commenting on the recent a7RII/GFX comparison. Now that I see you’re not, the resolution part of the above applies to the D*10 vs the a7RII, but the resolution winner is the Sony. The fill factors of the two cameras, after the microlenses are considered, is about the same. Color rendering is different, but not as different as Fuji vs Sony. So the images with the same lenses are going to look similar, with the Nikon images being fractionally fuzzier. Size, weight, native lenses, ergonomics, etc are much more important in choosing between the two cameras than IQ.
I would expect both cameras to be replaced soon with higher-resolution ones, probably with the Nikon replacement coming first.
Good read. I need some thoughts. I am a Nikon shooter (not a pro in anyway) and have a D610 with few good glasses like 24-70 2.8, 85 1.4 etc. Lately I am developing problems with my eye sight as well as shaky hands. I am experiencing the problem you mentioned about focus accuracy. Most of the things I shoot are arranged or set (landscape, portrait, products etc) other than the occasional street or fun shoots I do. So very fast AF and FPS are not a requirement for me, but the accuracy is critical. Even on a tripod, with Live View, many times I find I am not getting the critical focus on eyes (or wherever I wanted on people or products). I am OK with EVF (I have used a NEX before) and like the experience of WYSIWYG. Do you think the A7 series would be better for me?
Could be. Rent one and find out!
I am looking for a suggestion to which camera should i go for (sony a7 iii or Nikon d800). As i mostly shoot fashion and editorial. So let me know with your reviews which one should i buy and which lens would be great and supportive with the bodies.
Are you planning on buying a used D800? as far as I know, you can’t buy a new one. It seems odd to be comparing a just-released camera to one that’s many years old. The only reasons I can’t think of not to choose the a7III are:
1) 24 MP isn’t enough
2) You’re going to use it in the studio with the lens stopped down
3) You have a lot of Nikon glass.