|Nikon D3 Review
Long time coming. Best Nikon ever?
|Add a comment or send Thom feedback on this article.|
Canon produced a DSLR with a frame size the same as 35mm way back in 2002 (the 1Ds). Surely Nikon couldn't be far behind?
Along came 2003 and a 4mp DX-based D2h. Surely Nikon would have a larger framed camera soon?
Then in 2004 we got a 12mp DX-based D2x. Surely Nikon wasn't going to stay with DX forever?
In 2005 and 2006 we got minor updates of the above. Oh dear, maybe DX is the end!
In the meantime, Canon produced a 1DsII and a 5D, both with the same large frame size. The Nikon minions were restless. Many gave up on Nikon and went to Canon, despite the financial hit of selling off one system at declining used prices and acquiring another one new.
Meanwhile, Nikon was doing strongly at the other end of the camera spectrum, with the D100 leading to the D70 leading to the D200 and eventually the late 2006/early 2007 blitz of the D40, D40x, and D80. What gives? Was Nikon giving up their traditional strength in the serious amateur and pro markets for volume in the consumer market?
The first hint of change came in August 2007, when Nikon introduced the D3 and D300 pair. And what a pair. But before I get to them, note my words carefully in the first sentence of this paragraph: "the first hint of change..." Yes, that means what you think it does. Nikon isn't done trying to re-establish its reputation at the high end. The recently introduced D700 is just one more indication of Nikon's quest to re-establish their pro and serious amateur core market. You'll know what I mean by that statement sometime just before Photokina 2008. By mid-2009 I don't think anyone is going to doubt Nikon's commitment to the high-end market.
But let's get back to that D3. Nikon brought hundreds of photo press (both print and Internet) to the D3 introduction in Japan, all at Nikon's expense. The guest list read like a Who's Who of those who write regularly about DSLRs. (No, I wasn't invited. And I would have had to ask to pay my own way if I had been in order to maintain my no conflict of interest policy.) That showed a great deal of confidence on Nikon's part. For the few months prior to the introduction, every time you talked to a Nikon executive, engineer, or manager, they would have this sly smile on their face and just say "you'll be happy soon." In other words, Nikon thought they had a winner on their hands.
Actually, they had two. As I wrote in my D300 review, the less expensive of the cameras Nikon announced in August 2007 clearly is better than the D2xs, which had been Nikon's flagship DSLR. If the lessor camera was better than the previous high-end camera, what was the new high-end camera like?
Well, like nothing Nikon produced previously. Okay, that's not exactly true. From an external standpoint, the D3 looks and feels like a D2xs with a larger viewfinder prism hump. All the mechanical and user interface things we enjoyed with the D2 series carry over pretty much intact. It's what's under the covers that brought all the surprises. Better autofocus. Better metering. Better image quality. Better high ISO performance. Better card write performance. It's as if someone took a decently regarded high-end sports car and found a miracle engine to put inside that transformed the whole package. Yes, it's that dramatic.
I should note that Nikon was worried about one thing with the D3 introduction. They knew Canon was continuing to push megapixels, and they knew that other competitors would be, too. The big risk Nikon took was leaving the camera at 12mp, same as the D2xs (and the much less costly Canon 5D, for that matter). For the D3 to succeed, it had to be so much better at everything else that even the pixel cravers would find themselves satisfied. That's a tall order.
Then came the announcement, and talk of "go ahead, shoot at ISO 6400 (or 12800, or even the previously unimaginable 25600)." Really?
I need to sidetrack for a moment. Up to the D3, I knew of no digital camera I'd really trust at higher than ISO 1600. The problems are multiple. Many of you reading this may think that it's just about noise buildup that hurts high ISO digital images. No, it's not just that. Dynamic range tends to drop about a stop for each ISO stop increase (think about that for a moment and what that implies). Color saturation and differentiation, therefore, disappears and everything goes muddy; micro contrasts disappear. At some point in every digital camera the loss of edge integrity, loss of dynamic range, block up of colors, and production of visible noise just becomes too visible and too difficult to post process out of a picture. Basketball sports photographers know what I'm talking about. There isn't a gym in the country that'll give you f/2.8 and 1/500 at anything lower than ISO 1600, even with supplemental lighting in the rafters. Some Canon shooters may say "but my 1DIII does okay in that environment." Yes, it does, with the key word being "okay." But you still have to be careful with your settings and you don't want to blow those images up too large without doing some noise reduction.
And here was Nikon claiming that you could shoot higher than that. Really? That would be something new in digital, all right.
Not to give away the punch line, but Nikon's claims have some merit: the D3 does indeed shine in ways no previous DSLR I've used has (and before the Canon crowd jumps on that, I include the 1Ds, 1DsII, 1D, 1DII, and 5D in that list; I haven't yet used the 1DIII or 1DsIII enough to make a sound judgment call yet, but am leaning towards lumping them into that list, too). I wasn't fully convinced of that initially, but probably because I was mostly doing my usual nature and landscape photography. Then I took my D3 to the dimly lit gym I play basketball in. To put it bluntly: it was a whole new shooting experience for me. But I'm getting ahead of myself and starting to talk results when we have a lot more "what is it?" detail to cover.
Beyond the high ISO claims, Nikon was making some other claims that sounded a bit outlandish, like, if I was reading it right, an autofocus system that focused beyond the autofocus sensors. It's not quite that straightforward, but that's what it sounded like Nikon was saying. And that the metering system did a better job with whites and blacks and recognizing skin tones than any previous one. And that the autofocus system acquired focus faster and held it longer than any prior Nikon system. And that the camera wrote to cards five times faster than before. The list of gee-whiz things that came out of the announcement sounded long and pretty incredible. The announcement was going to be a hard act to follow, since the camera had to deliver on so many claims that it seemed impossible.
The first D3 cameras shipped right after Thanksgiving in 2007, about when Nikon said they would at the product launch in August. But Nikon did something interesting during the interim: they had plenty of loaner D3 cameras that they got into pro user hands, at least for short periods. I was able to try one out briefly in early September. Many others got their chance between September and December when the camera shipped to those of us with NPS Priority Purchase orders in. This was another example of Nikon's confidence: they hadn't let prototype cameras into the wild for any previous pro (or consumer) camera. That they felt a prototype was good enough to withstand the high level of scrutiny their claims would generate was an act of supreme confidence. (I suppose some could say it might have been an act of supreme desperation, but I never got that impression from anyone at Nikon; they weren't desperate, they were confident.)
D3 bodies are still in relatively short supply, even though I write this review nearly five months after getting mine. I suspect they're going to stay in somewhat tight supply even after the full line of Nikon pro DSLRs is out (late fall 2008). There are couple of reasons for that. First, the camera is good. Real good. Second, it's made at the Sendai plant in Japan, which has limited capacity for DSLR assembly, and that plant is assembling other pro models (the D700, for one). Third, the camera is winning back some of the switchers, which is increasing the demand for Nikon pro products. Finally, at the moment I write this it's the only FX sensor size body Nikon makes, and there are several things that FX gives us back that we all wanted but DX never quite delivered. Fortunately, the D700 and another Nikon pro body should take some pressure off D3 availability. Those that just want an FX body will likely opt for the D700, while those that want more pixels will opt for the next model, leaving those that need the high-speed aspects of the D3 free to find bodies.
There's no question that the D3 is a hit. It's been astonishing how fast they've appeared in places where before I only saw a sea of Canon bodies. Let's find out why that is, shall we?
The only rating here that is even questionable is the rating for value. At US$4699 (street) the D3 isn't cheap. But it's good enough to argue for a full mark in value, even considering the upcoming D700 at US$2999. Long-term readers will note that this is the first time I've given full marks for any product.
The D3 is not a new body. Nikon has taken what wasn't broken about the D2xs and simply morphed it forward into the D3. That's not to say that the bodies are identical. They aren't. Specifically, you'll need new Arca-style plates because the shape of the camera is slightly different. That means you need new doors for your EN-EL4 batteries, too (though if you don't mind an edge sticking out, you can use the old doors in a pinch). Other minor changes abound, as well: the hand position on the right side of the camera has been tweaked slightly, the ENTER button is now OK, the Direction pad has a real button in the middle now (yes!), the white balance detector on the front of the prism is gone, and a few of the buttons and controls have moved a bit (more separation between AF-ON and AE-L/AF-L for example. The autofocus method switch on the bank is missing a position (Group Dynamic), the larger color LCD now dominates the back, and under the rubber cover on the left side of the camera (from the back) is a surprise or two. The prism is wider and taller, too, but the net result is still the same any way you slice it: if you've handled a D2 series body, the D3 will immediately be so familiar that you'll at first think that nothing has changed.
Indeed, even disassembly of the D2xs and D3 is similar, and reveals just as many unchanging things internally as externally (yes, unlike some reviewers, I'm not afraid to pull out my screwdrivers and torq wrenches, though I'll admit I haven't dissected far into my D3 yet, as I prefer using it to operating on it).
As I noted with my D300 review, looking inside the D3 is a bit like observing the work of a bunch of nano-bots let loose on a D2xs to improve it. Everywhere you peek you find small changes, but overall, the result is so familiar as to be uncanny. So, suffice it to say, if you liked the body design of the D2 series, you'll like the body design of the D3.
The optional wireless function is now handled by the WT-4, which is not an older wireless transmitter worked on by nano-bots. No, the big bots took a whack at this option, coming up with a much different design (hip mount, not camera mount), and throwing in lots of bells and whistles (wired Internet should you want it). The WT-4 requires an EN-EL3e to operate, and draws less power from the camera than the older WT's did. The WT-4 is also shared with the D300, meaning you only need one wireless accessory if you use both bodies. I'll come back to that thought later.
But you're saying "get to the sensor, Thom." Okay, I will. The sensor in the D3 is a CMOS part designed by Nikon. The part was not fabbed by Sony, though just who manufactures it for Nikon is still unverified. Just after the D3 was announced, I used both the official and unofficial channels to ask Nikon a number of questions about this new sensor. Curiously, Nikon isn't answering most of those questions, neither via the official channels nor through the back channels I sometimes get information from. That's not to say they're not responding. The Japanese always respond if you ask them a question. But they have a way of giving non-answers. Even when you hit the nail right on the head--which in their culture forces them to give you a real answer--they will often give you something that requires a great deal of deciphering and reading between the cultural lines to figure out what it means.
This is a curious departure from when they introduced their last wholly-Nikon designed sensor, the LBCAST JFET sensor in the D2h. Nikon went out of their way to talk about the underlying technologies in that sensor. And they got burnt by doing so. Not so much because the things they talked about were wrong or misleading, but because some ancillary decisions made the D2h Nikon's one prominent failure in the digital age. (What decisions? 4mp, incorrect IR filtering, bad metering part, etc.). What's difficult to assess is whether Nikon is being quiet this time because they felt burnt last time, whether they have secrets they don't want to reveal, or whether they see an advantage to staying quiet and letting the sensor speak for itself. I suspect that it's a little of everything, actually. I'd be willing to wager that the D3 sensor uses lateral buried techniques or other aspects inherited from the D2h sensor, for example. One part they seem to want to keep secret is exactly how the microlenses are designed and deployed. At present all we know is that Nikon claims the microlenses are continuous. However, that doesn't seem to explain all the difference in light capture ability the D3 sensor gains. I suspect that the real trick is something more than just continuous microlenses.
12 effective megapixels mean 4256 x 2832 pixel images, so you actually lose a few pixels compared to the D2xs (and D300 for that matter). If you're keeping score, at 300 dpi that means 9.5x14" prints native. The base ISO of the sensor is 200, with third-stop increments up through ISO 6400. You can also boost ISO two more stops, up to an effective ISO 25600. If that weren't enough, you can un-boost it by a stop to ISO 100, however I'll warn you that you must be careful doing this with high contrast scenes, as the method Nikon is using can truncate highlight detail in such situations. The sensor can provide 12-bit or 14-bit NEFs (and it doesn't slow the camera in any way to shoot the higher bit rate).
Of course, that brings us to "is 14-bit really better than 12-bit?" A definite yes on that (unlike the "maybe" I said with the D300). If you shoot NEF, you get to choose whether to shoot 12-bit or 14-bit, and whether to shoot Compressed, Lossless Compressed, or Uncompressed (the nano-bots aren't very good at naming things ;~). To some degree, your converter is going to impact whether you can pull anything additional out of a 14-bit image that you can't out of a 12-bit. First, the place the difference is most likely to show up is noise in the near black realm, and in my experience, there's a lot of variability in converter capability in this range. Second, the difference really isn't enough at the low ISO values to be visible without you making big linearity changes to the shadows (e.g. D-Lighting). Finally, I'm not sure most people know what it is they're looking for, so they just won't see it even if it's there. Even with my trained eyes it took me awhile before I could consistently see the modest difference 14-bit makes. This is no different than the old "does Compressed NEF create a visible difference" question we grappled with for years. Same answer: for most of you, no, it doesn't matter. However, if you're really trying to get "optimal data" then you'll not shoot Compressed NEF nor 12-bit. I can't see any real need for Uncompressed NEF now that we have Lossless Compressed. Increasing card capacities and lower card costs have made maximum compression not so important an issue.
For JPEG and TIFF shooters, the D3 uses Nikon's Picture Control system to change image parameters (sharpening, contrast, gamma [which everyone has taken to calling Brightness lately], hue, and saturation). The camera comes with three presets (none of which are very well chosen) and Nikon has provided downloadable options to mimic the NikonD2X Color Mode defaults (a better starting choice). You can also roll your own by varying the ones you're given, though to really take this to heart you need ViewNX or Camera Control Pro so that you can create custom tone curves, as well. I'll give you a couple of clues: the color LCD brightness is set too high to correctly evaluate images, so a lot of you are setting Picture Controls and getting upset when they don't look right on the LCD, especially if you've also set AdobeRGB (the camera's display is sRGB at best). This new facility is a little trickier than the old Image Optimization choices. Nikon's choices, as I noted, aren't well chosen--they produce overly bright, contrasty images with over-saturated color. Couple this with Active D-Lighting, and you get a very faux look, not nearly as good as the camera can produce with proper setting of the controls. Fortunately, I think we'll all figure it out fast enough, and there's enough flexibility in the system to satisfy most people. However, note that Nikon kept at least one trick to themselves: we can't create new starting points, only modify existing ones. The D2XMODE downloads are new starting points, not modifications of existing ones already in the camera, thus you should download them and add them to your camera to give you more flexibility in what you can produce out of camera. Nikon needs to allow us to create our own Picture Control starting points, though. Indeed, that would bring the Fujifilm DSLR users back to the fold and probably capture a few of the Canon crowd, as well (right now it's difficult to get the exact same JPEG look out the Nikon bodies as with Fujifilm and Canon bodies).
Metering has changed on the D3, as has autofocus, and the two are now having relations together. Yes, we still have the same 1005-pixel CCD in the viewfinder doing the metering, though its position has changed (still vulnerable to light coming through the viewfinder, by the way) and it's now linked in real time to the AF system. The critical change comes for matrix meter users: the D3 pays more attention to what's under the autofocus sensor being used than the previous cameras did, though not nearly as much as some of the low-end cameras are now doing (D80 comes to mind). Enough so that you need to pay closer attention to your histograms. Nikon has changed the mid-tone gamma at the default settings, which gives the appearance of brighter images, but in a stable, moderate contrast scene with something neutral under the AF sensor, all my Nikons give the same exposure. Curiously, I find that my D3 gets whites and blacks "more right" than my D300 does. Indeed, I'm extremely impressed at how well the D3 matrix metering handles scenes that previously almost always needed exposure compensation to get right. I've shot quite a bit of glacier work now with the D3, and most of the time the meter just gets it right, something no previous Nikon ever managed to quite do.
About that autofocus system: it rocks, but it's going to take some getting used to for anyone coming from previous Nikon DSLRs. Group Dynamic AF and Closest Subject Priority are gone. That alone is going to bother some people who counted on those features. But here's the thing: once you spend the time to really learn how the new system works, you'll find out that it works well. The real trick is to understand when you should be in 51 point 3D and let the camera do all the lifting, and when you should drop down into one of the Dynamic AF options (9 point, 21 point, or 51 point). Unlike the D300, the D3's sensors do not cover a large portion of the frame, which can be problematic for some shooters. However, also unlike the D300, the D3's Auto Area works faster and better when subjects move in and out of the autofocus sensor area. In the Auto Area AF mode (and 3D tracking mode) the AF sensor and matrix meter get together in interesting and useful ways. The system works well on anything that has a flesh tone in it, even if the subject moves off the autofocus sensors. There's some serious computational stuff going on in the focus system now, and it has much more "magic" than before. On the other hand, magic isn't foolproof, so when the system flops, it flops. I was blown away to find that, while shooting basketball if I could get the camera to start autofocus on a sensor, it really didn't matter much where that player went in the frame. The camera simply figured out that my subject moved out of the autofocus sensor area and calculated correctly using the viewfinder CCD how it should probably adjust focus. This certainly doesn't work with anything approaching 100% confidence, but that it works at all is amazing.
I'll repeat what I said before: spend time studying your autofocus options. With practice you'll start to understand the situations where the magic won't happen and where you need to step in with a different AF choice. Once you get to that level of understanding, you'll have no problems at all with the system. But it is enough different than anything that came before it that you must spend time learning it.
Another nice addition to the AF system is the ability to tune it for individual lenses. If you've ever had a Nikon (or Canon) body that back-focused or front-focused with a lens consistently, AF fine tune is the answer you've been waiting for. In return for a few hour's work, you'll get more precise focus than you had before. A few lenses that I had put off as being good but not great performers suddenly went up a notch after I tuned them on my D300 and D3. Unfortunately, a focus tweak isn't enough for some of my dog lenses, though. Of course, you're probably asking why lenses might perform differently. That's a huge can of worms, actually, as there are so many variables involved in "correct" AF that you really don't want to go there. The short answer is this: a phase detection AF system like the Nikon DSLR bodies use expects all lenses to perform identically. The camera detects the current focus situation, calculates where focus should be (yes, it knows not only how much out of focus, but which direction, the error is when you initiate a focus operation), then tells the lens to move a set amount. To prevent hunting, if the new error after the focus is performed is within a reasonable tolerance, focus isn't attempted again. Tolerances are the bane of accuracy, and with fast lenses on close subjects, even a small error is too much. Fortunately, most things that would cause a lens to not make it to the desired position are repeatable (i.e. the lens always goes to the same position given the same input). Thus, the new tweaking feature allows you to just adjust the lens to where it should be, and then everything magically is right with your focus.
Semi-related to that is Live View, the ability to see what the sensor sees prior to taking a picture. Indeed, I see one of the better uses of Live View as being to verify (and or tweak) focus. But the best use of Live View is to tweak white balance. The D3 has two Live View options: handheld and tripod. These differ primarily in how focus is handled, with tripod using a compact camera-like contrast focus (slow) and handheld using a quick drop-the-mirror blast to use the regular focus system. Both are nice features to have, but don't get too excited about them--it's just another way of shooting. The D3 can display a histogram and a virtual horizon in Live View mode.
On the back of the D3 is a big 3" color LCD. But it's changed in more than size. First of all, it's gotten about three times more "dots" (you don't say pixels with LCDs, as they use individual colored dots close together to produce the perception of more colors). This brings the view on the D3's display up to VGA levels (yes, that means that the older cameras were very sub-VGA). A second change--going from delta to striped layouts--also adds to the perception of a less pixeled display. The net result is that you can zoom in on images on the D3's display and get a very good sense of focus and noise in the final image. For those of you that are counting, that's not all the way in. When you zoom as far as you can go on an image on the D3 display, you're at about 400%. Hit the Zoom out button twice and you're looking at actual pixel view. (You can also specify a magnification for the center button on the Direction pad via the Custom Settings.) As I noted earlier, the display is too bright (and slightly too blue). Set LCD brightness to -1 if you're trying to get the best view of your image. Curiously, the D3 is the only Nikon camera that doesn't come with a screen protector (the D300 and D700 do, and they use the same LCD). Buy one from Hoodman if you want one.
Okay, it's time to wrap up the feature list. The D3 has voice annotation, a dual card slot that allows you to backup, copy, or overflow to the second card, a 100% viewfinder (though the white balance indicator has been dropped from the viewfinder), and a host of other small, useful features. The latest firmware allows you to set Artist and Copyright information in the EXIF, though it's not documented well. None of the D2 series features like GPS support, wireless support, AI and AI-S lens support, multiple exposures, etc., went away. In fact, several of them were updated. These changes didn't come without their own downsides, though. Entering manual lens data is no longer fast or optimal, the interval timer menu is more obscure than ever, and even histograms have taken a slight turn for the worse: by making the luminance histograms white on white instead of yellow on red, Nikon has made it tougher to see blowouts, as the frame is too close in color to the data.
The D3 has the same intelligent battery life tracking as the D2 series, and battery life has been improved a bit yet again, if my use is any indication.
Overall, there's an awful lot to digest with the D3. However, with the recent introduction of the D700, we have one last thing to digest:
To many, the D700 feature list and price are shockers. But here's the way I see it: the D3 is squarely targeted at the photo journalist and sports shooter who needs absolute performance, the D700 is more targeted to the serious amateur and pro backup market. If I were shooting all day for a newspaper, that second slot and the voice annotation alone would be the clincher for the D3. If I were shooting sports, the faster frame rates make the D3 the choice (yes, you can get close with the D700 and MB-D10 grip, but you end up with a larger combination that isn't quite as robust). As much as the differences are minor, it should be pretty obvious immediately to you whether those differences put you in the D3 or D700 camp.
In the (admittedly long) time it took for this review to appear, the D3's street price has dropped several hundred dollars from its original US$4999 price. I give it full marks for value at that price, that's how good the camera is.
I think handling is now a good news, bad news situation. The good news is that Nikon continues to tweak and pay attention to basic ergonomics. The D3 body design, for instance, has had another full pass from the D2x and fits the hand a bit better.
The D2x is on the left, the D3 on the right in both images. Note a couple of things. First, the Command dials have a bit of a tilt to them--they're no longer perfectly horizontal. Bingo. This makes them easier to find and produces a slightly better hand position. Moving the AF On button on the vertical grip was a wise choice, too. And even though those curves on the grips look the same, they're subtlely different, all in ways that are a real improvement. The addition of a real center button on the Direction pad is also welcome. About the only questionable choice is the relocation of the AE-L/AF-L button, which I think is too far left for smaller hands than it should be. Still, all the small ergonomic changes make the D3 a pleasure to use, despite its size and weight. The D2 series was no slouch in this respect, but it's nice that we continue to get refinements.
The bad news is this: the complexity of the system now overwhelms it. Indeed, this shows even in Nikon's own second guessing: with the most recent software update they've added back the RECENT SETTINGS menu option for those that don't like MY MENU.
I dealt with this problem at length in my D300 review, so I won't belabor the point (much) here. The issue is simply this: we've got over 30 buttons, dials, and switches on the camera; we have 40+ custom settings and can have four variants of that; we have 22 things on a very disorganized SHOOTING menu and that too can have four variants; plus we have four different displays for information (viewfinder, top LCD, rear LCD, and the new Info display on the main LCD). And I'm not counting button overload or settings that have triggers (like High ISO NR). Having just written and re-written an 800-page guide to the camera, I can say without hesitation that the combinations and permutations start to get mind-boggling. It doesn't help that Nikon punts on telling you some of the detail in their manual (quick, which things can you assign to which buttons via Custom Setting #F4, #F5, and #F6? and which are button only, button+dial, or dial only combinations?). And then there's the "can't get here from there" problem that putting Live View, Self-Timer, and MUP on the Shooting Method dial added (e.g., you can't use some things on the SHOOTING menu, like Interval shooting with some things on the Shooting Method dial).
Long-time Nikon pro camera users sneak by confronting the full level of this complexity because this has been a long term feature creep. Thus, the complexity increased a bit when we upgraded from D1 series bodies to D2 series, and again now with the D3. But even I still get caught every now and then in forgetting something is set or not optimizing my settings the way I want to. And something important like white balance is now missing from the viewfinder, which is a BIG step backwards and makes this setting complexity more likely to sting you when you don't want it to. The problem is a lot of little things that have crept into the mix and now make for more potential for error, or slower-than-should-be setting combinations. As I've suggested earlier, a fix is absolutely necessary now, and at a minimum requires:
If Nikon is going to hire a professional Italian designer to work through subtle hand position ergonomics, I think it's more than high time that they also do the same thing for the user interface. As I've suggested before, my old colleague Alan Cooper makes his living doing just that sort of thing. Nikon needs to bring on someone at least as good as Alan for a full review and fix.
It's a good thing I don't give marks for handling, otherwise the D3 wouldn't get five stars for everything in my ratings. Instead, it would have one item that gets only four stars. Don't get me wrong, we can live with the current complexity in handling, but it's not optimal and becoming less so with each passing revision. Moreover, it just doesn't sustain the addition of more features. A perfect example of this is the addition of Copyright information in the latest firmware revision. Both Image comment and Copyright information do the same thing: they let you fill in a field in the EXIF Maker's Tags that will appear in the Metadata area when you use any of Nikon's software. (This, by the way, is also wrong. They should fill in IPTC standard data fields, not proprietary fields, but that's a different problem.) Now we have two items on the SETTINGS menu that essentially do the same task. This is a place where a second hierarchy level should have been used (Image comments which would have sub-menu items of Note, Artist, and Copyright).
The definitive D3, D3s, and D3x Guide is here
Almost a 1000 pages of reader-friendly text, dozens of extra useful support files, plus an unrivaled volume of detailed information for D3, D3s, and D3x users! No other work comes close to providing you the detail and advice on this camera that my Complete Guide does. Find out why other Nikon book authors buy my Complete Guide when it comes out.
Writing to Card
One thing to note is that file sizes can be large. NEFs range in size from 11MB (Compressed 12-bit) to 24.7MB (Uncompressed 14-bit) in Nikon's numbers, while a JPEG Fine Large is 5.7MB on average, a bit smaller than a D2xs JPEG. The way I shoot (Lossless Compressed 14-bit NEF most of the time) I get around 100 images on a 2GB card, which is a bit low for me (too many card swaps when I'm doing HDR panoramas, where I've been known to take 100+ exposures for a single image), so I've switched to 4GB and 8GB cards. (There's a dynamic here that you need to consider: the bigger a card, the more your images are at risk to a card failure or loss; the smaller the cards, the more you have to change them and the more likely you'll misplace one). I should note that all the numbers I just presented are variable with the amount of detail in the scene. These numbers represent "typical use." I've seen NEFs that are 50% larger than the expected number, and 30% lower. Obviously, the former had tons of small detail in it, the latter had very little detail in it at all. The same will be true of JPEGs. This, of course, means that the Frames Remaining counter on the camera can be about as reliable as a gas gauge on a car, perhaps less so, as variations in what you shoot will impact the number. Fortunately, the D3 has that second card slot, so if you're shooting the spillover method, you can just swap the first card out when you see the second card start to being used.
I've now got a pretty large mishmash of cards I can test in cameras. I've yet to see any that misbehave in my D3. I've gotten the best performance so far from my Hoodman Professional 280x (UDMA-enabled) and SanDisk Extreme IV (UDMA-enabled). The worst performance was split between my oldest 1GB Microdrive and a generic PNY 1GB flash. In my experience so far, I'd say you want at least a SanDisk Extreme III (IV is preferred) or a Lexar 133x WA (300x is preferred), though. The Delkin and Hoodman UDMA cards, along with the SanDisk Extreme IV and Lexar 300x, are the best performers I've found so far.
On the D300, the new CAM3500 sensor covers a very large portion of the frame, which means that the system is very good at following subject motion and managing highly off-center autofocus. On the D3, the same sensor covers a smaller portion of the frame, meaning that autofocus capabilities are restricted to a smaller portion of the frame. For some subjects, this is a problem, for others it usually isn't (hint: subjects with skin tones fare better outside the autofocus sensing area than others).
Coupled with the scene recognition that is being done by the metering CCD in the viewfinder, the new focus system is sometimes so uncanny in 51 point 3D or Auto Area focus mode that it boggles the mind. This happens most often when there are faces or other skin tone in the area covered by the autofocus sensors. As it turns out, while various skin tones can look fairly different to us humans, to an RGB metering system they all are in the same narrow range and thus easily detectable. However, note that when you're in non-white light--some fluorescent and sodium vapor, for example--this system seems to not work anywhere near as well, probably because the color tint from the lighting is polluting the skin tones the camera is trying to detect. On the D3, recognition of the colors of an object is clearly taken into account in Auto Area focus mode and somehow dealt with even when that object goes out of the autofocus sensor area! The D300 is less capable at this than the D3, and Nikon does talk about the fact that the focus system in the D3 has the benefit more CPU capability in the camera. Shooting basketball at the end line, for instance, I was surprised to find that the D3 in Auto Area AF would more often than not get focus right even when my subject left the autofocus sensor area and kept moving towards me. The D300 fared less well when set and used the same way (same angle of view lens, same settings, etc.). The problem you're going to have is trying to figure out when to trust the D3 and when not to. Simple answer: give it a modest amount of trust in Auto Area, trust it far less when in 51PT3D Dynamic Area. Note that the camera has to acquire focus on the subject with an autofocus sensor first for it to track outside the sensor area in these modes (none of the autofocus modes do this). While the D3 will surprise you with sometimes getting the right focus when an object goes off the autofocus sensing area, it will downright amaze you in situations where the object returns to the sensing area. This is the real trick in the D3's autofocus sleeve: it watches the color and contrast of the item you focused on, and follows that object with the viewfinder metering CCD even when that object is off the autofocus sensing area. When the object returns to the autofocus sensing area, the camera simply picks right up where it left off. Yes, that means you could be focusing on something at the right edge of the autofocus sensing area, have it move up above the sensors and come back in on the left side of the sensing area and the camera will figure that out. This system isn't foolproof, but it's pretty darned impressive. Again, the D300 does something similar, but not nearly as well as the D3 does.
Low light focusing is also surprisingly good. While only the center 15 autofocus sensors are cross hatched and more sensitive to low light, if there's enough contrast under the outlying 36 sensors the D3 still seems to be able to focus. A few people have noticed that this is less true with vertical pictures, probably due to the line orientation of those outer sensors. But in general I find only the inner autofocus sensors confident and reliable in low light with Single Area AF.
The D3 does a very good job with erratically moving objects, such as flying birds and some sports. Indeed, I'd place it a bit better than the D2x in this respect. Only when the light gets dim and the subject gets decidedly out of the central area did I find any problems with tracking moving subjects.
So far, all good news, right? Well, there's some bad news, too. Be prepared to completely forget how you used to set autofocus on previous Nikon DSLRs and prepare to take considerable time figuring out how to optimize your use of the new system. I'll have much more to say in my eBook, but here's the learning approach that I think works best:
By the time you've completed that progression of testing, you should have a good idea about what things work for which subjects. Which brings me to my point: more so than any previous Nikon camera, I find myself changing autofocus settings more often with the D3 (and D300) than ever before. There's no "does it all" setting you're going to dial in and forget with this system.
White balance, curiously, seems better on the D3 than on the D300, especially in mixed lighting. The direct Kelvin settings once again didn't match my Minolta Color Meter and my Imatest measurements, but they're closer than the D300 was. The range over which Auto WB works well is more limited than Nikon suggests, though from about 4000K to 6500K it can probably be trusted. Over and over on my Patagonia trip I found that I could guess the setting better than the camera could. True, I've been setting white balance on cameras since the early 1970's, and have learned how to recognize different lighting. But you'd think by now a camera that does a good job of detecting white in the exposure meter ought to be able to figure out color temperature better than it does.
The white balance system is also different in how you bias it, which is going to confuse a lot of people. No longer can you set Cloudy -2. That would be closer to Cloudy A2 now, though the changes Nikon made means you can't set exactly the same thing as Cloudy -2 anymore. While I appreciate that the new adjustments are a constant MIRED apart (they weren't before), it took me a while to get used to the new adjustments. Overall, there seems to be something slightly different about white balance in the D3 (and D300): it's as if the rotation of colors is being done slightly differently, so the changes to reds and blues with basically the same white balance setting looks ever so slightly different to me on the D3 than it does with the D2.
Overall, color is good once you get a handle on the new white balance settings. Be careful of saturation, though. Especially if you crank up the saturation controls you can end up with images that are cartoonish and don't easily adjust back. Also, contrast tends to get set too high by the camera. Consider manually controlling and shooting for a lower value than the camera will give. It's far easier to put contrast back into an image than it is to take it out.
Curiously, a similar problem crops up when using DX and 5:4 image sizes (pardon the pun). Using the exact same settings, the full FX frame and a DX crop have different and more visible hot pixeling patterns.
I start with the problems because, other than those, the D3 simply blows away any DSLR Nikon has previously produced. I'm almost tempted to say that you can shoot at any ISO up to 1600 at any camera setting (sharpening, JPEG compression, etc.) with impunity. I said almost. As with all digital cameras, increasing ISO does have other impacts, most notably on dynamic range and noise. If I really work at it, I can force JPEG compression artifacts into the faint noise and make them visible at 100% view at as low as ISO 800. But I have to select the worst possible JPEG compression (Size priority and JPEG basic) and push the contrast and sharpening to do that. So, here's how I'll put it: if you don't overdo the camera settings you can shoot at impunity through ISO 1600. Yes, there will be a bit of noise if you pixel peep, but not enough to worry about. And, yes, you'll lose a modest amount of dynamic range, but again not enough to worry about except in very high contrast scenes.
The big surprise for me was my dimly lit basketball gym. To date, I've not found any DSLR that I'm 100% comfortable shooting at ISO 3200 in that gym (and you have to in order to get even a modestly usable shutter speed at f/2.8). Well, not any more. The D3 does just fine in that gym. We're going to do this a little different than I've presented it before, though. Here's a shot at ISO 3200 from the far end of the court using the 24-70mm:
Even at this small size you should be able to see that we've got good whites and what appears to be decent definition in the dark tones. But what about noise? Well, try this 100% view of the same image:
Straight from camera, camera defaults. But let's clean that up a bit in post by pushing it through Neat Image and then use a very aggressive sharpening and see what we get at 100%:
Can you tell that one player had a band on his right ring finger? And did you notice the small sign on the far wall? Remember, I'm at the other end of the court, at 70mm, shooting at ISO 3200. I really don't expect to read 12 and 14 point type on a wall at the other end of the court with a 12mp camera. I'm a little surprised, however, that I'm seeing clear black-on-orange contrast and not a blur of noise, though. I'll bet you with the 70-200mm at 200mm from this position we would be reading what's on that small sign.
Simply put, you can go further with ISO on the D3 than you can with any other Nikon body. ISO 3200 is quite usable. ISO 6400 is usable in some situations with the right camera settings. I'm less excited about ISO 12,800 and 25,600. Dynamic range loss, contrast and noise build-up are just too much at the highest ISO values to be anything but emergency and experimental settings.
Should You Get a D3?
Bottom line: the D3 a great camera, but make sure you need what it offers.