We all know the paper specs for the Nikon D800: 36MP, 4fps, full frame. Same ISO range as the D700 – 100 to 25,600. 100% viewfinder, full HD movies, and an improved 51-point AF system derived from the previous camera. And to boot, a D800E version with no anti-aliasing filter for even more resolution, as if 36MP with a weak filter wasn’t enough for you. What we don’t know is how it fares in the real world.
What follows is what I believe is the one of, if not the first, complete, real-world test by a photographer of a production D800. 8 hours of non-stop flat-out work – so, please leave a comment if you enjoyed it.
I first heard D800-shaped noises way back at a Nikon event in March 2011, both locally and from my sources in Japan. These definitely wasn’t the same information as what was going around on the popular rumor sites at the time – I was told specifically D3X replacement, slightly higher pricing. An interesting strategy; too bad the initially planned May-2011 release got derailed by the tsunami. The Sendai plant that produces the D700, D800, D3S and D4 was inundated and had all of its precision machinery replaced; an amazing feat considering the magnitude of the disaster.
All images in this review were shot as 14-bit lossless compressed NEF and converted in ACR 6.7/ PSCS5.5.
Still, the camera has finally arrived, and delivered precisely on the promised date – even in a small market like Malaysia. That’s impressive. I got mine through NPS; apparently there are around 200 members, 90 D800 orders, and…only 18 cameras to go around. Mine must be one from the very very first batch – serial number 234.
Initially, I wasn’t going to order one. Then my high-mileage D700 began to give up the ghost, and I downloaded some sample images – which in short, blew me away. They were honestly better than the output from the Hasselblad H3D-39 I used a couple of years back in mid-2010 – and quite close to what I’ve seen out of the Leica S2 (I do have access to one, I will try to do some comparisons soon) so far. I called my local NPS rep and put in an order for the D800E; however, playing with both sets of demo files further, it became clear that a) you weren’t really giving up that much getting the regular D800, and as a bonus, it would arrive sooner – an increasingly important factor given this month’s shooting commitments – and b) lenses and diffraction would be the limiting factor for me, not the camera’s sensor. Furthermore, for most purposes outside the studio environment, I intend to shoot the D800 in 14 bit compressed RAW, but downsize by half to 18MP for manageable output, lower noise, and better per-pixel detail.
At some point, I will probably source a split prism screen and have the focusing screen and mirror precisely adjusted for manual focus planarity; for now, I’m relying on AF. I do really miss the focusing snap of the custom-cut F6 type-J screen on my D700; it’s just so much easier to tell if things are in focus or not. The standard D800 screen is bright but doesn’t have much snap. This may sound odd, but I’m having trouble getting used to the 100% finder again – I’ve become so accustomed to mentally adding a little bit around the edges of the D700 frame (97% finder) that now I’m chopping things off. Just one of those little differences between the two cameras, but important nevertheless.
First impressions on lenses
But this is a camera review! Glass matters. Two big things: a) AF fine tune matters a LOT; b) the optimal set of lenses for this camera is different to the D700, again. The 24/1.4 never quite focused properly on my D700 – I was at the extreme limit of AF fine tune adjustment – but it’s bang on with the D800 with zero adjustment, and incredibly sharp all over. The 85/1.4 needs a lot of shutter speed while handheld to shine; probably double what you’d expect – in the 1/125s range or higher. I’m also seeing a lot of edge CA that wasn’t there before (1-2 pixels worth; that’s probably less than a pixel on the D700). My 60/2.8 G Micro is soft until f5.6 and focus shifts, which isn’t something I’ve ever seen before. Oddly, the 28-300VR is actually rather impressive at 300mm on the D800 – NOT something that could be said about the lens on the D700. In fact, it performs much better on the D800 than it did on my D700 – curious considering the demands of this sensor.
And a 100% crop – this is a best case scenario for CA, with the lens wide open at f1.4. There were other, much worse shots; I suspect being ever so slightly out of focus also contributes to visible CA in a big way. The older 85/1.4 D is very likely going to be unusable wide open with any subject that’s even moderately contrasty.
The sole lens that has been outstanding on every camera it’s been mounted on is the Zeiss ZF.2 2/28 Distagon – wide open, I think it has the highest resolution of any of my lenses. Makes me want to get the 2/100 Makro-Planar again, and possibly also the 4/18 Distagon. Generally, lenses I though were good wide open on the D700 are showing a slight but noticeable improvement stopped down, even if only by a stop – I’m talking about my workhorse AFS 24/1.4 G, AFS 60/2.8 G Micro and AFS 85/1.4 G here. Also, lenses that vignetted a bit before will vignette more strongly now; I suspect it’s because the individual photo sites are smaller, and there’s no trick offset micro lens array like in the Leica M8/M9 to counter it.
It was dark by the time I got the camera and had a chance to shoot with it, so take it as a worst case scenario and impressions will almost certainly improve when I have more light to work with. I will not be providing full size files, so please don’t ask. There may be crops. Clicking on any of the images will bring you to a larger version on Flickr; the EXIF data is all intact.
Seems to be about the same speed as the D700 in good light, no difference as far as I can tell in low light. Has issues focusing the 85/1.4 G accurately in low light; this may be true of the D700 but it’s a lot more noticeable here due to the higher resolution. Tracking ability seems slightly improved. More tests are required before I can reach a conclusion here. Contrast detect AF for live view subjectively seems at least twice as fast as the D700, and doesn’t require as many passes while hunting.
First thing you notice is the camera is lighter – about 100g, according to the specs. However, I personally find it not quite as comfortable as the D700; my fingers were cramping after use. This is because the lower section of the grip is thinner – not sure why, perhaps their testers had small 4th/5th fingers, or perhaps Nikon just really, really wants you to buy the vertical grip.
That’s about the only bad thing ergonomically – I don’t know if it’ll be a deal breaker for extended use. Sadly I don’t find it anywhere near as comfortable as the D4, which is pretty amazing. Oh, there IS one more thing: the mode button is a stretch to access; I feel like I’m going to dislocate my index finger by pressing it. Too often I hit the movie record button by mistake and wondered why nothing was happening. A firmware fix to make the movie button change exposure mode when shooting stills would be a nice easy fix. I do like the new angle for the shutter button, though – it’s much more comfortable.
There are a lot of nice touches though. Live view is a lot easier to access thanks to the button where the AF mode switch used to be; am I the only person who misses the AF mode switch though? That little button near the lens mount is not so easy to find, but at least you can see what the camera is set to in the finder. The new drive mode dial is also a lot easier to use – it locks and still has detents, so you can count positions and change modes in the dark – there’s a big difference between using CH and Q in a theatre, for instance. Speaking of the shutter, it’s slightly more hollow sounding than the D700; crisper, too. Sadly not as quiet as the D7000, which is nearly silent in Q mode. Interestingly, the mirror doesn’t cycle when shooting in live view – just the shutter – so the camera is actually very quiet, and doesn’t vibrate much. Although the maximum frame rate is 4fps, it doesn’t feel any slower than the 5fps D700. Mirror blackout time is the same, which is to say, effectively instantaneous.
I mention this because it seems that Nikon’s newest meter isn’t quite as accurate as the last one. My D800 definitely meters a bit hot compared to the D700, and seems a bit more erratic. Further investigation is required here.
Turning AUTO ISO on and off is an option from the button, finally! You use the front command dial to toggle on/off, and the rear one to select ISO. There’s also an option to automatically select minimum shutter speed as a 1/focal length, with some fine tuning in either direction – sadly, the fine tuning isn’t granular enough. For example, the 85mm defaults to 1/90; adjusting this to ‘faster’ gives 1/200 rather than say 1/125, which would be perfect. Back to manually selecting shutter speed again, it seems. In short: you will be needing to use higher shutter speeds than 1/focal length would suggest. Think about what you’d set on a D7000, and that’s about right – remember, the pixel density is the same.
Bearing in mind that I’ve only shot it under low light/ night/ available darkness conditions, I’m impressed. It’s doing a decent job for the pixel density – though I would not pick this over a D700 for reportage work. The few flash-based tests I have done have left me stunned. Color accuracy is slightly better than the D700, but resolution is out of this world. Dynamic range is about the same, subjectively; however, instead of being highlight-biased as with the D700/D3, it’s shadow biased – you’ve got to be careful not to blow highlights because there simply isn’t as much recoverable color information there. Still, I wish I’d had the camera earlier today for the food assignment I just shot; it would be the ultimate tool for things like that. No matter, because I’ve got several watch shoots in the coming weeks. Early impressions are that the pixels don’t have the same degree of elasticity/ integrity as the D700 (duh) and are probably somewhere between that and the D7000; probably closer to the D7000.
See the following crops; they were shot under pretty dark conditions and tungsten light, i.e. a torture test. Subjectively, I think it’s ~1 stop behind the D700 at a pixel level; if you downsize to D700 size, it’s actually a stop ahead. Now if only Nikon would give us a pixel-binned half-resolution sRAW size for low light! If you are shooting full resolution, I recommend stopping your auto-ISO at 6400; anything higher than that has to be downsized to look good. 3200 is definitely acceptable, and anything below is good. The reality of printing, however, is that because you’ve got so many more pixels, a print will look a lot better than at 100% on screen. There’s no sign of banding, but beware of strong noise in one particular channel over another in the shadows, but it depends on the temperature of your light source – for instance, heavy shadow recovery or dodging under fluorescent lights is going to give you a red cast to that area.
Note that I didn’t bother with ISO 100 and 200, they look the same as ISO 400. Click to go see larger versions on flickr – the ‘original’ size is a 100% crop.
Here’s a real world ISO 6400 example, sodium-vapor street light. Yes, it’s noisy at 100%, but I’m fairly confident it’ll print just fine.
I’m not a huge video person, though I have dabbled (to be the subject of a future article). I do know what good quality footage looks like. The D800 is excellent. Dynamic range is great; noise is low, and above all, there’s no rolling shutter effect that I can see – even while panning rapidly under fluorescent light operating off a 60Hz AC supply.
Pretty darn good, I think – I just grabbed the battery that came with the camera; 36% charged; shot about 650 frames, and it went down to 6%. Extrapolating, that’s about 2,100 shots per charge. And that was with heavy LCD use and some live view. One battery should more than get you through a day – you’ll run out of card space far, far sooner. I can’t honestly say I’m pleased about the complete battery system change (I have plenty of EN-EL3es and EN-EL4as) – but at least the new power system lasts longer, and also has a little catch that allows for a spring loaded (read: easier to replace) battery.
Buffering and file handling
The manual claims 25 images for 14-bit compressed raw – the camera shows r13, but I’m getting 17, using a UHS-I Sandisk Extreme 32GB SDHC card. Still trying to find out where the difference is; auto ISO gives back three more frames, but curiously NR makes no difference. The buffer flushes surprisingly quickly, and you never feel like you’re waiting for files to write – although there is a slight lag when playing back images, probably due to the file size.
It’s probably worth noting that file handling is a bit slower, but not 3x slower (despite 3x the resolution) – however some operations like brushes etc. and even converting in ACR definitely take longer, so budget time accordingly. I’m using a mid-2010 MacBook Pro with the 2.66GHz i7 and 8GB of RAM. I don’t even want to think about retouching files this big yet.
It doesn’t take a rocket scientist to figure out that getting the most out of the D800 is going to require a lot more care than the D700; the resolution is so high, in fact, that I think the AF system may be letting it down slightly – not from a speed point of view, but from a precision standpoint. And I’m not sure it’s the AF sensor per se, but possibly the granularity with which the lens motors can move the elements small distances. I know that in live view, there’s a point of critical sharpness that’s usually very tough to hit using the focusing rings of AF lenses; the travel is simply too fast.
I don’t think the D800 is a general purpose tool. It definitely isn’t a run-and-gun photojournalist’s camera; in fact, I find it more demanding to shoot street with this than the Leica M9-P. It’s probably at a two stop or more disadvantage to the D700 if you want critical sharpness at the pixel level – firstly, you’ve got a slightly noisier sensor, and secondly, you’re going to need higher shutter speeds to maintain pixel integrity and combat camera shake. Although downsizing the files to 12MP yields lower noise and more detail than the D700, I don’t think I’ll be using the D800 for photojournalism at the moment; I’m going to have to figure out the AF and lens foibles first.
Where the camera will shine is in the studio for work with controlled lighting, or landscapes – the resolution is outstandingly impressive, and dynamic range at base ISO seems subjectively on par with the D700 – no mean feat indeed. However, I need to do more testing in daylight (not experiments with flash) to determine for sure. Stay tuned for more images and thoughts over the next few days; at some point I want to try to get hold of a Leica S2 to do a head to head comparison. Please leave a note in the comments below if you’ve got any questions or have something you’d like me to test, and I’ll do my best. Right now, I’m going to get some sleep. MT
Visit our Teaching Store to up your photographic game – including Photoshop Workflow DVDs and customized Email School of Photography; or go mobile with the Photography Compendium for iPad. You can also get your gear from B&H and Amazon. Prices are the same as normal, however a small portion of your purchase value is referred back to me. Thanks!
Images and content copyright Ming Thein | mingthein.com 2012 onwards. All rights reserved