D3/D300 Q&A

My best answers to the questions that are being asked

Original: 8/25/07
Updated: 8/27/07 added WT-4 question, fixed compression QA
Updated: 8/28/07 added focus calibration question, changed compression answer, edited landscape answer, flash syn
Updated: 8/29/07 answers from Nikon Japan on flash and NEFs
Updated: 8/30/07 answer to ISO 100 question, added comments to right column (see end of column)
Updated: 9/4/07 part numbers, Bayer question, HDMI question

As usual with a Nikon preannouncement, there are a lot of unanswered and partially answered questions. I'll give it my best shot at filling in some blanks.

What happened to LBCAST?
Unknown. Though Nikon did say that the D3 CMOS sensor derived from work on the LBCAST sensor used in the D2h. And I'll point out that LBCAST is a form of CMOS: it differs mainly in the type of transistor used at the photosite. LBCAST is produced on CMOS fabs. There's a possibility that Nikon has simply downplayed the underlying technology so that people won't associate the D2h liabilities (color cast, IR pollution) with the D3. But I suspect that the answer is much more simple. Nikon's sensor group seems to both be receptive to experimental ideas and quick to take up real advances rather than sticking to a single design path. Put another way: "Best ideas win." Note: I asked Nikon about what the underlying transistor technology was in the new sensor and received the answer "we cannot disclose this information."

Where's the D3x?
Higher resolution FX models have been prototyped, so the answer is: still in the design labs. Whether they appear, when they appear, what they use for a sensor, what price they might have, and what they'll be called is still not decided, AFAIK. Next year is Photokina, and I wouldn't be surprised if a higher end product gets announced just prior to that show.

What got replaced?
Nikon Japan still shows the D40, D40x, D80, D200, D2hs, and D2xs as current models. But that's a little deceptive. Those indeed are the current models, as the D300 and D3, while announced, aren't available. As with any fast-changing product line, the end life of most products is not just determined by the introduction of new products, but is also dependent upon final sell off of existing inventory. D2hs inventory seems to be running low, and I don't believe it is still being produced at Sendai, so we're at the tail end of its sales cycle. The D2xs could possibly survive a bit longer if priced between the D300 and D3, so it'll be interesting to see what happens there. The D200 is a different story. A D200 at US$1299 and a D300 at US$1799 would bracket the Canon 40D very nicely, I think. Given the commonality of parts between the D200 and D300, it wouldn't surprise me to see a continuation for the D200 production for awhile.

Why no sensor cleaning in the D3?
The D300 gets sensor cleaning because Nikon is basically licensing an APS (DX) sensor directly from Sony (though with changes; see below), and Sony has that design already built for their sensors. The D3 sensor is a Nikon design and I suspect that the critical path to introduction just didn't allow time for integration of a sensor cleaning function. Remember, the parts would be a different size due to the larger area of the FX sensor. This introduces quite a few subtle issues that may not have been solved in time for the D3 intro.

Why no sensor stabilization?
Nikon seems committed to lens VR. Indeed, they produced slides showing how much better lens VR is than sensor VR at the D3 launch and introduced the remaining exotic telephoto lenses in their VR versions (and it's at the telephoto end that the differences become more pronounced). Curiously, the 24-70mm didn't have VR, a critical mistake as far as I'm concerned, as Nikon continues to send mixed signals to users. Will there be a VR mid-range zoom (that covers FX) at some point? There should be if Nikon continues to stand by lens stabilization. But the fact that we haven't got it yet is a silly marketing mistake and one that leaves them out of parity with Canon. That said, I wouldn't be surprised to see sensor stabilization eventually enter the consumer line for Nikon DSLRs. Why? Because consumers just don't use tripods, don't hold their cameras steady, and are using lighter products that don't stay as steady in the first place. Thus, even a small benefit in stabilization would be useful for that market (and yes, this would mean that the cameras would have to recognize when VR lenses are mounted and default to the lens stabilization; there's one small fly in that ointment: if VR is turned off on the lens can the camera override that if the user turned stabilization on in the camera [i.e. default to the "best" VR solution]?).

Where's the landscape camera?
Actually, the D3 should make for a better landscape camera than the D2xs most of us have been using. Same resolution, a better diffraction limit, and lower noise mean somewhat better results. Remember, as pixel counts get higher and higher, it takes more pixels to make for any dramatic difference. The Canon 1DsIII nets you 5616x3744 instead of the 4992x3328 of the 1DsII, or 624 more pixels horizontally and 416 vertically. All that from what seems like what should be a more hefty boost from the 5mp increase between the cameras. On a 30" print that works out to 187 dpi versus 166 dpi. Are you really going to see that difference? Remember, you've got diffraction working more strongly against you. Meanwhile, a D3 nets you 142 dpi on the same print. That's perhaps enough to make a visible difference, but not as much as the difference between the numbers 12 and 21 seem to suggest. At more normal sized prints--say up to the limit of the desktop inkjets--the D3 should do just fine.

What happened to the ambient light sensor?
It isn't needed anymore (and didn't work that great in the first place). Basically, the only time it was used was when you pressed the FUNC button while doing a PRE white balance measurement. That not only complicated the white balance function, but in practice it didn't always do as good a job as using the sensor.

Why no ISO 100?
The D3 (and D300) have a new low setting, which is essentially ISO 100. Nikon has been very consistent in following proposed ISO guidelines for ISO: the base ISO should be the optimal light gathering point for the sensor, and in both sensor's cases, that would be ISO 200. Normally with a lower-than-optimal ISO setting the camera tries to keep the midtone curve the same and the highlights tend to clip faster, essentially reducing usable dynamic range. And for what it's worth, there are intermediate settings at 1/3 stops between 100 and 200, as well (L0.3, L0.7, and L1; you're going to have to think backwards to understand this, though!). But expect ISO 100 to be slightly inferior to ISO 200 in image quality due to highlight clipping.

Is the D3 sensor Bayer?
Despite speculation to the contrary, it appears that the D3 sensor is indeed Bayer. If it was three-layer, as one radio interview with Vincent Versace has been interpreted to say, the raw files would be enormous, and they don't appear to be. Moreover, the number of bits moving around at 11fps would be beyond what current designs can handle.

What's with the 4-camera HDMI thing Versace mentions?
I think Vincent must have mixed up a couple of things. Most of the Nikon pro bodies back to the F4 have been able to be "linked and sequenced" via the 10-pin connector so that you can shoot with multiple cameras at precise moments. The Matrix (the movie) type of visual effects are often (but not always) done this way, using an array of still cameras that fire in a specific and very carefully managed sequence. To do a stop-action-and-revolve sequence, for instance, you array still cameras in the dolly pattern you want to emulate, spaced however far apart you want the "virtual" movie camera to move for each frame. You then fire these still cameras simultaneously, which the pro Nikon bodies guarantee to syncronize if linked properly. In the editing room, you sub in the individual still frames for what would be the movie dolly sequence; the lead-in and the lead-out are from movie cameras positioned at the start and end of the still camera array and which are recording the action continuously. Back when both the movie and still camera used film, you could simply match film stocks used, and no one was the wiser. For awhile you had to post process the digital stills to match the film movie stock. Now, with more movies being shot digitally, we're once again getting to the ability to quickly match still and motion sequences easily. As to Versace's apparent claim that the HDMI port allows something new, I don't believe this to be the case. First, the HDMI port simply mimics the camera's LCD, and is a video feed. Thus, there's a delay in getting the still out the port (the EXPEED engine has to process the data first). Second, the resolution of that video is lower than the camera can achieve, and lower than some of the digital video cameras being used in features. Third, you'd have to--in real time--excerpt a frame from each HDMI source to the digital recorder taping the sequence. You actually don't need still cameras to do what Versace suggests: just array a bunch of HD video cameras that are timecode linked and extract the same time-coded frame from each. HDMI might be nice in studio situations, where you have a big HD monitor for the clients to view. You'll be able to push Live View to the HD monitor if you want, while the camera is linked to a PC (wired or wirelessly) for capture and edit.

Can I create uncompressed NEFs?
Yes. Even though the cameras support uncompressed, the new lossless compression means that all the raw converters out there have some work to do to understand the new format. Nikon's track record in facilitating such understanding is underwhelming, which means we could see converters lag a bit from the camera intros. Note to Nikon Executives: this is the wrong posture and hurts Capture NX sales because it potentially slows D3 and D300 sales. What helps Capture NX sales is that it's the best converter for NEFs, period. You might want to help us folks out in the trenches get that message out more loudly rather than spend time trying to slow down software competitors. In the digital world, every successful hardware platform manufacturer I know of benefits greatly from a healthy "eco-system" of third party software that increases the value of the manufacturer's product by providing functions or ease of use that the manufacturer didn't supply. In this specific case, pro shooters--especially the kind that will buy the D3--have established workflows that they're not going to want to disrupt. Not being open and supportive of third-party software would be another marketing dissonance that holds you back and keeps some pros from switching (see right column).

Does the WT-4 mount to the camera?
No. It's a standalone unit now, so I guess belt carrying systems will make a comeback (it's really too big to fit in most pockets, as it's 5 inches tall; Nikon supplies a belt case and neck strap). On the plus side, you can connect the WT-4 via Ethernet cable for very fast LAN transfers (supports both 10 and 100Mbps). For wireless, you have 802.11a/b/g support, and because of the larger antennae of the WT-4 you have an extended range (specified as 590ft [180m] in a/b and 853ft [260m] in g mode). To control the camera or make settings from the computer, you need Camera Control Pro 2. The WT-4 uses its own battery, but fortunately that's the existing EN-EL3e (it can also run from the EH-6 AC adapter).

Is there anything Nikon didn't mention?
Yes, actually. The D2 series supported D-TTL flash units. The D3 specifications curiously only mention i-TTL, so I initially suspected that D-TTL support is gone. Nikon Japan, however, says that the support is still there, which is good news.

What's with flash sync?
The answer from Nikon Japan is as follows: Auto FP is supported by both the D3 and D300 with external flashes that support it, such as the SB-800. The full ability of Auto FP is not supported by the internal D300 flash, though the internal flash supports a form of Auto FP at reduced power in the shutter speeds above 1/250 and up to and including 1/320 (I think I'm reading the chart they sent me correctly).

How is focus calibration done?
I don't know the answer to that question. I've neither seen anyone else describe nor seen anything in Nikon's published documentation that describes it. Personally, I'm a little leery of such user-managed precision options. It's often just as easy to get things wrong as it is right. A lot of its usefullness will depend upon how well and thoroughly Nikon documents the process.

Does the D300 really drop to 2.5 fps when using 14-bit NEFs?
Yes: frame rate drops on the D300 if you shoot 14-bit NEFs. We'll have to wait and see via testing whether that is a meaningful limitation. I suspect that the primary beneficiary of the extra bits will be shadow detail, especially if any post processing moves shadow values much. The primary folk complaining about this specification are the wildlife and sports shooters (the rest of us realistically don't need more than 2.5 fps). But they're making complaints without knowledge right now. I suspect that the D300 at 12-bits and full speed is a better wildlife camera than the D2xs is (due to lower noise). If that's true, then the D300 is a bargain to start with. Moreover, 14-bits increases file size, something that, at 8 fps means filling cards more rapidly. Every photographic decision is a trade-off. I'm willing to give Nikon the benefit of the doubt on this one until I've shot with the camera and can see whether the two extra bits are so compelling that I always want them (in which case the drop in frame rate is unwelcome) or whether they're just somewhat visible (in which case I can make the decision between frame rate or bit rate, depending upon which is more useful in any given situation). I asked Nikon why 14-bit was slower than 12-bit, speculating that external circuitry might be involved. There answer was that the "14-bit sequence needs more time to convert A/D and for data read out. We don't have external circuitry for 14-bit sequence."

The D200 battery sucks, so why didn't they switch?
Actually, the D200 battery does just fine for JPEG images on a D200, it's only when shooting NEF that performance drops. But Nikon says that battery life has been substantially improved on the D300, perhaps as much as double (after all, a CMOS sensor should be lower power than a CCD). But if that doesn't cut it for you, try this: get the MB-D10 grip and use the battery from the D2/D3 series, which has much more capacity. The more curious thing that isn't being discussed is why the switch from the EH-6 AC Adapter to the EH-5 or EH-5a? That seems like a decision that flies in the face of all the other good decisions Nikon made about compatibility.

Is the D300 sensor the Sony IMX021?
Two possible answers: (1) yes, or (2) a slightly modified version. Given Nikon's historical relationship with Sony's sensor division and the slightly modified specifications, I'd have to say #2 is the correct answer. Also, note that like previous Sony APS sensors, the sequence has remained the same: (a) Sony announces sensor samples; (b) Nikon announces a camera using a variant of that sensor; and (c) Sony and/or other companies announce cameras using the sensor Sony announced in step a (that step hasn't happened yet, but will shortly).

What's with the MB-D10 grip?
There are unconfirmed reports that the MB-D10 grip will fit on other new Nikon cameras (some sites reported that it would work on the D3, which has to be wrong, as the widths of the cameras are different). But why would the grip be called the MB-D10 instead of the MB-D300 as has been the norm in past Nikon naming? I suspect that this is because there will be another D300-sized body that uses the grip in the future. Note also that you must use something other than an EN-EL3e battery in the grip to get the boost in frames per second on the D300 (the grip also supports the D3 battery, by the way).

Are there part numbers for the new products?
Yes. The NikonUSA part numbers for the new products are as follows:

D3 - #25434
D300 - #25432
MB-D10 - #25359

14-24mm - #2163
24-70mm - #2164
400mm - #2171
500mm - #2172
600mm - #2173


Nikon should have...

...produced a road map of where they're going with the new DX and FX split. They did say that the two lines will continue to develop in parallel, but that's not saying anything, and it doesn't even have the same weight as saying, for example, "we are committed to continuing to develop more and better DX cameras." Ironically, we got more of a roadmap from Nikon at the D2h announcement than we did with the D3/D300 intro.

The usual reason given for not offering a road map forward is the fear that it will reduce current sales. For example, had Nikon said they will produce additional FX models in 2008, one with higher resolution, one smaller in size, they might think that they would sell fewer D3 bodies in the short term as people simply wait for the other products to appear.

In my experience, that's simply not true. If anything, the inverse is true: when a buyer has a clear idea of how the current offerings fit into a continuum of offerings (both past and future), they are more likely to make current buying purchases. Indeed, this is at the root of the Canon/Nikon switch trend. Many pros think that Canon has revealed their lineup (going from I to II to III with a model makes it a bit obvious, right?), while Nikon made vague commitments and slipped the ball with at least one model (the D2h) and now hasn't said anything about their future. The way to get someone solidly in your camp is to make sure they know why your camp is preferred both today and in the future. Nikon simply hasn't done that.

So once again we have vague statements and seemingly contradictory products from Nikon. The vague statements are the "develop in parallel" ones. The contradictory products are new lenses without VR. Nikon's marketing is once again fighting against itself. Are we going to get a full VR lineup or not? What if I buy that new 24-70mm and Nikon comes out with a mid-range zoom with VR next year? These dissonances need to be fixed, and fast.

As I've long said, Nikon's product engineering is top notch. There's nothing wrong with the D3 and D300 offerings--they're state of the art. But Nikon's positioning and marketing is still crude and sometimes self-defeating.

So, just in case anyone at Nikon is surfing my site, here's the minimum statements (or some variant thereof) you need to make:

  • We will have at least one additional FX body in the next year that clearly compliments but does not obsolete the current offering.
  • We will continue to produce DX bodies with additional features and performance to keep our consumer DSLR lineup state-of-the-art at competitive prices.
  • We are committed to VR in the lens. We may supplement this with additional stability controls, but all telephoto lenses will have VR as it is more effective.
  • We are committed to producing both DX and FX lenses that fill out our product offerings. (Help us out: let us know what's missing!)
  • Our product update cycles for cameras are 18 months at the consumer end and 36 months at the pro end. On rare occasions we may shorten those cycles to remain competitive. But note that updates tend to reset prices upward, so even if you buy a camera from us towards the end of its cycle, you will still be getting value for your money.
  • If any of the above changes, you, our customer, will be the first to know (another reason why you want to register your camera purchase!).

Update: It seems a lot of folk took exception to the above. Some said it is crazy to announce products in advance (I didn't ask Nikon to announce any products; read it again carefully). Some said that plans have to be flexible and subject to change (if you change "will have" to "currently plan" in the above there's no lack of flexibility). Some disagree with the basic contention that more knowledge of possible future system directions is useful. Let me elaborate on two points to better illustrate what I'm talking about:

1. Most of the emails and calls I've gotten from confused pros (yes, there are plenty of them after the announcement) stem from one simple problem: Nikon didn't put an "h" after the "D3." Nikon clearly said the D3 is the replacement for the D2hs. But the problem of going from D1h/D1x and D2h/D2x to D3-only is that many people are making a quick interpretation about that naming: there won't be a high-resolution equivalent to the D3. This is part of the problem people are having with the vague Nikon statements about the future. Obviously, the naming changed, so expectations changed with it.

2. Pros have timeframes and budgets for upgrades and replacements. Most of us in the US do that at the end of a fiscal year for tax reasons (some every year, some every couple or few years). The first problem is that Nikon's products are coming awful late in the year, and because supplies are likely to be limited, you'd have to commit now to an unknown (and it's a double unknown because we don't know about a high resolution pro body from Nikon but do know about that from Canon). The second is that we simply don't have the full picture about Nikon's pro plans still, and probably won't when the buying decision time comes in November and December.

bythom.com | Nikon | Gadgets | Writing | imho | Travel | Privacy statement | contact Thom at thom_hogan@msn.com

All material on www.bythom.com is Copyright 2007 Thom Hogan. All rights reserved.
Unauthorized use of writing or photos published on this site is illegal, not to mention a bit of an ethical lapse. Please respect my rights.