|Q&A with Nikon D4 engineer Toshiaki Akagi|
|Saturday, January 14, 2012 | by Rob Galbraith|
We sat down this week with Toshiaki Akagi, an engineer and senior manager at Nikon in Japan, as well as a key figure in the development of the company’s newest flagship digital SLR, the D4. In an hour-long chat at the 2012 International CES trade show in Las Vegas, Akagi, whose title is General Manager, 1st Designing Department, Development Headquarters, Imaging Company, answered our questions about D4 autofocus, image quality and more.
The answers are not verbatim transcriptions. Our questions were posed in English, and Akagi replied in Japanese, with Nikon USA’s Kenji Suzuki and Yasuhiro Nozaki acting as interpreters. Interviews conducted in this way require constant clarifying of meaning, of both the questions and the responses, which leads to fairly extensive back and forth discussion about each answer. From this we distilled what we asked, and Akagi’s replies, into the text below.
Q. Has the speed of the autofocus system improved, specifically the time it takes to acquire focus when first engaged or when rapidly switching to a new target at a different distance? For example, when shifting from the quarterback throwing the football to the receiver catching it, the D3S’ AF system is a bit too slow to always bring the receiver into focus in time to get him making the catch.
Yes, it is faster. First, the D4 is able to autofocus at 10fps, vs the D3S at 9fps, so that’s one improvement. Second, in situations where the D3S might have taken until the third frame in a burst to bring the subject into focus, the D4 will get it on the first frame. Third, some professional photographers in Japan, ones who told us the D3S was a little bit slow to acquire focus before the first frame, have tried the D4 and said it is improved, said it is faster.
(Akagi then showed on his computer an example sequence of a significantly defocused subject coming into focus by the fourth frame with the D3S, and by the second frame with the D4. The photos appeared to have been taken with a long lens, and showed exactly the sort of AF system speedup that should, in our football example, lead to more properly-focused pictures of the receiver receiving the ball.)
The autofocus calculation is done in the camera’s main CPU, which is faster in the D4 than the D3S and which contributes to the AF system’s better speed.
Q. Has tracking accuracy been improved in sunlight vs lower light? This is a quirk we’ve experienced with supertelephotos such as the 400mm f/2.8 when photographing the same team at a night game and then at a day game: the in-focus percentage is consistently somewhat higher at the night game (as long as the field lighting isn’t completely terrible).
This has not been reported to us as a problem.
(Nikon USA’s Lindsay Silverman spoke up to say that, at Nikon USA, they have received some reports of this phenomenon.)
The new 91,000-pixel RGB metering sensor, which also plays a role in autofocus tracking, we believe will improve autofocus accuracy generally. This is true not just when the AF system is set to Auto-Area AF or 3D tracking, but also, for example, with 9-point Dynamic AF too.
Q. How would you compare the high ISO image quality of the D4 to the D3S?
Overall, D4 high ISO noise levels are very similar to the D3S, though photos shot with the D4 will have reduced colour noise (thanks mainly to noise reduction improvements in the D4’s EXPEED 3 processing hardware). The main difference you’ll see in pictures from the two cameras at high sensitivities is in the fine detail: D4 photos are noticeably better in this way.
(Akagi then showed several 13 x 19-inch prints of the same scene, captured as NEFs with the D3S and D4 at ISO 12,800, in which the overall colour, tonality, visibility of noise and shadow rendering couldn’t have been more similar. But, the D4 photos looked sharper and more detailed. The D4’s crispness advantage in Nikon’s comparisons was significant, suggesting that we were seeing the effect of more than just the new camera’s higher pixel count.)
Q. Among the image quality improvements cited for the D4 is better skin tones. What difference would we notice in a portrait shot with the D3S and then the D4?
You will notice better tonality. We have reshaped the tone curve to give more natural-looking shadows and shadow transitions in faces.
(This change presumably will be apparent in photos processed to JPEG and TIFF in the camera, as well as NEFs converted to finished files in Capture NX2.)
Q. Have there been any changes to the optical low pass filter?
The D3S has a resolution of 12MP, and the D4 has a resolution of 16MP, so we had to make the filter thinner in the D4, to match its finer pixel pitch. But, the effect of the filter in the D3S and D4 is the same.
Q. Have there been any changes to the self-cleaning mechanism?
No, it’s comparable to the D3S.
Q. Are there any changes to the long exposure noise reduction, include a shortening of the delay between frames?
No, the dark frame subtraction works the same as the D3S, and takes the same amount of processing time.
Q. Why did you choose the XQD format for one of the D4’s card slots?
The main reason is the XQD format is faster than CompactFlash. We worked with Sony (the maker of the first XQD cards) to get better write speed in the D4. Sony says their first XQD cards will be capable of 125MB/s, and the D4 will be able to take advantage of that level of performance. In fact, the camera’s XQD slot can go much faster than that, well over 200MB/s, as higher-speed XQD cards come out in the future.
This is faster than CompactFlash cards today, though the D4 will be able to take advantage of fast UDMA 7 CompactFlash cards as well.
Q. Both SanDisk and Lexar have told us that XQD is not on their product roadmap currently. Have any companies other than Sony told you they will be making XQD cards?
At this moment we don’t know which companies will be making XQD, other than Sony.
Q. Why does the D4’s EN-EL18 battery have a lower capacity than the D3S’ EN-EL4a?
Nikon’s adoption of a lower capacity battery for the D4 was a bit of a head-scratcher, especially in a camera that’s overflowing with specification improvements. We sat down this week with Toshiaki Akagi, an engineer and senior manager at Nikon in Japan, as well as a key figure in the development of the company’s newest pro digital SLR, and got the skinny on why, in the case of the EN-EL18, less is more.
The following information came from an hour-long chat with Toshiaki Akagi that took place at the 2012 International CES trade show in Las Vegas. Akagi, whose title is General Manager, 1st Designing Department, Development Headquarters, Imaging Company, answered our many questions about image quality, autofocus, the XQD slot and more. We’ll get the full interview published in the next short while, and it will have some tantalizing bits of information for prospective buyers of the D4. For now, we’ll tackle just the matter of the 10.8V/2000mAh EN-EL18’s seemingly inferior specs when put against the D3S’ 11.1V/2500mAh EN-EL4a.
Akagi cites two reasons for the EN-EL18’s emergence. The first is straightforward: a new battery, and new charger for it, was required to be compliant with new battery regulations in Japan. The second is that Nikon sought to build a battery that would not only meet their home country’s revised regulatory requirements, but would also surpass the performance of the EN-EL4a.
Less is more
The particular Lithium battery chemistry they ultimately selected for the EN-EL18 delivers greater actual shooting time than the EN-EL4a, says Akagi, despite its lower milliamp-hour rating, when the camera is autofocusing and firing constantly. For example, when a photographer is covering a soccer match with a 400mm f/2.8, or any other situation where the camera is spending more time fully active than idle. When the battery is being continuously pushed to provide current to drive the lens motor, shutter, mirror, image sensor, processing circuitry, memory card and more, Akagi indicates the EN-EL18 will power the D4 through more frames than the EN-EL4a could have, and more frames than the D3S and EN-EL4a too.
Akagi accompanied his explanation with hand-drawn approximations of the discharge curves of the EN-EL18 vs the EN-EL4a at low, medium and high current draws, which showed the EN-EL18 outstripping the EN-EL4a when current demands were highest. The high-current discharge characteristics of the two batteries, he says, are different enough that it enables the 10.8V/2000mAh EN-EL18 to outlast the 11.1V/2500mAh EN-EL4a when the D4 is doing lots of autofocusing and capturing lots of pictures. In this scenario, says Akagi, the EN-EL18 will provide roughly 10% more runtime than the EN-EL4a, at normal temperatures. In cold environments, the EN-EL18’s runtime advantage is described as being even greater.
Put simply, if you’re using the D4 to take a steady stream of photos, Nikon’s contention is the EN-EL18 will give you more frames per charge than would have been possible with the EN-EL4a. Conversely, if the camera is kept awake but is spending far more time idle than it is taking pictures, the EN-EL4a would last longer on a single charge than the EN-EL18.
That said, what sounds like a potential EN-EL4a strength is probably nullified by the fact that the Auto Power Off feature of Nikon cameras will usually be set to kick in after a short period of inactivity. In this ultra-low power state, either the EN-EL18 or EN-EL4a is going to last an extremely long time. So, the fact that the EN-EL4a might last the longest gives it little practical advantage over the newer battery.
The previous paragraphs bring some context to the 2600 frames-per-charge specification for the D4 and EN-EL18, which is a precipitous drop from the 4200 frames a D3S and EN-EL4a are specified to give. These figures are generated using a standardized test developed by Japan’s CIPA industry group, and is comprised of shooting one frame at a time with a lengthy pause in between each frame. This synthetic test probably reflects typical consumer-level camera usage patterns well enough, but at the expense of properly simulating how a fast frame rate pro digital SLR tends to be used. It has the effect of emphasizing the low-current charge life of the battery, and results in a much superior number for the EN-EL4a.
But, if Akagi’s description of the characteristics of the EN-EL18 are correct – and based on our understanding of how batteries work, his is an entirely reasonable and logical explanation – the real-world charge life of the EN-EL18 in the D4 should end up about matching or slightly exceeding the D3S and EN-EL4a (or the D4 and EN-EL4a, if that pairing were possible).
This does not, however, explain why Nikon chose to prevent the D4’s EN-EL18 and Battery Charger MH-26 from being interoperable with the D3S’ EN-EL4a and Quick Charger MH-22. By placing the new battery’s and new charger’s connector in a different spot, any measure of backwards or forwards compatibility was eliminated.
The reason for this, says Akagi, is the MH-22 (and presumably the MH-21 also) would not be able to properly charge the cells in the EN-EL18 and could cause the battery harm (we didn’t get into the specific damage that could occur, or why, though a too-fast charging rate is the most likely thing that Nikon was trying to avoid). To prevent this from happening, the company opted to make a clean break beginning with the D4 and its battery and charger.
Q. The Nikon 1 J1 and V1 demonstrate that a camera without a mirror can be made to autofocus quickly and accurately, and the ultra-quiet operation of these cameras is really useful in certain situations. Do you see a mirrorless design eventually coming to Nikon’s F-mount digital cameras, including flagship models?
We cannot comment. What do you suggest we do to improve our cameras in the future?