Need a camera for those Classic car memory’s?

Wednesday, March 07, 2012


Why Nikon HAD to make the D800 ultra high resolution

Mike Curtis

Megapixels: Go Big or Go Home was their only choice


Theory/Supposition/Personal Opinion: Nikon realized Canon had video zipped up. They couldn’t get a decisive win there. If they made their low light performance too close to the D4, they’d cannibalize sales as they had with the D3/D700. So how could they respond effectively and get some mindshare back? Megapixels: Go Big or Go Home was their only choice left.


OK, so Nikon clearly realized they needed to have a response to the incredibly popular Canon 5D Mark II. Canon really knocked it all of the park with the 5D MkII, especially once they (finally) got 24p video working on it. The D700 was an excellent stills camera, and with the grip even shot a higher burst frame rate. But with no video capabilities, it just got overrun in the marketplace by the 5D Mark II.
Based on their experience with the D700 eating sales of the too similar but more expensive D3, they realized they needed a different way to compete with Canon without cannibalizing D4 sales. How to do this?
Nikon must have sat back and wondered “How will we compete with that?” Nikon actually had the first HDSLR out – the D90, at a lower price point and some months before Canon’s salvo (pun?). Canon ran away with the success and got not only marketshare, but mindshare. Why? I think three reasons:

1.) Canon had 1080p vs Nikon’s 720p

2.) Canon had the better H.264 codec, not M-JPEG

3.) Canon had historically been perceived as the generally “better” camera (until the D3/D700 came along for stills)

(EDIT – and another reason, it was full frame not APS-C/Super35, so it was a bigger sensor than any electronic video camera out there)

Example of the last point: I went down to Chinatown to shoot the dragons and fireworks for Chinese New Year recently, and went full-on photo nerd with a BlackRapid double shoulder sling rig with two pro camera/lens setups – a Nikon D3S with a big 24-70 lens on it, and a D300 with grip and the even bigger 70-200 lens. While walking around, a young guy (high school/college age) excitedly asked “What are you shooting? 5D Mark II? 1D?” He clearly knew a bit about cameras. I said over my shoulder “Nope. Nikon – D3S and D300S.” And he kind of laughed and said “Lame….gay.” Letting go the sexual implications of Nikon vs Canon and my technonerd desire to correct the young man as to the relative merits of D3S vs 5D Mk II….I saw where he was coming from. Canon gets ALL the attention for video DSLRs. Nikon announced their D300S right before Canon announced the popular 7D, and Nikon’s next response was the D7000 – their first DSLR that was 1080p24 H.264 capable. The public response? Whatevs. While a nice little camera (I bought one last year), NO ONE I know ever mentions shooting Nikon for video – it is 7D or 5D all the way. Canon has just WON this round in the public’s (or at least camera nerds’) mind. Knockout punch, no contest.
So (and this is all conjecture) I think Nikon realized this. Especially after the D7000 didn’t set the world on fire in terms of getting video pros to take them seriously. Canon, in the majority of the pre-HDSLR years, had generally been considered the pro’s choice for quite some time. While the D3/D700/D3S helped change minds in the stills world, minds take time to change, and Canon had a fantastic rep to ride on. So when cinematographers started hearing about how great Canon DSLRs were for low cost cinematic video, the wave just kept on rollin’.  And Nikon, as much as they might not have liked it, probably knew this. No matter how Awesome Sauce the D4 and D800 were going to be, if they didn’t absolutely blow the Canon 5D Mark III and 1DX out of the water, they’d be perceived as “Me To” products, or worse yet “Yeah…whatever. Still shooting Canon.” (or renting).

And speaking of renting – another detail that separates Nikon and Canon, especially for the professional shooter, has all to do with luck. Way back, when Canon and Nikon picked what direction to turn the lenses to focus, the direction Canon chose happened to be the same direction as used for PL mount lenses – the standard now used for cinema shooting. Nikon’s don’t. So even if Nikon starts making cameras that shoot better video than Canon, there is an uphill battle to overcome in terms of operation – the lenses focus the “wrong” way compared to how serious shooters are used to – you turn the expected way, and focus pulls closer rather than farther (or vice versa). This is readily addressed with a reversal gear for a follow focus, but is just One More Thing Nikon has to contend with. Back to point –

So they needed something BIG. Nikon needed to have a category or capability that they could just OWN, and stomp Canon into the ground with, the way the 5D Mk II had to the D700 (which lacks video). So what were the options? Especially since they (probably, hopefully) didn’t know exactly what Canon was up to, and didn’t know exactly what the new products would be like, and everybody is on a long term development timeline anyway. What were their possible choices? Here’s some possibilities to go for:

1.) Video capabilities – Canon was doing quite well with this, and likely to improve, and other divisions in Canon had been making video cameras for decades, so going up against Canon for video capabilities was a risky proposition – not only for this camera, but for future generations as well. What happens if you poke the Video Monster with a stick and tick it off? Canon’s R&D resources (and pockets) were deeper, this wasn’t a good option. And, gallingly, while Nikon had been first to market with this, not only did it not make a sizeable difference after the 5D Mark II shipped, nobody really even remembers that Nikon was first. Double lose.

2.) Dynamic Range – while I think this would be a good place to start competing, it isn’t a priority in the public’s eye – everyone talks about megapixels and video and maximum shooting frame rates, but rarely is this discussed. So even if they ended up technically superior….it probably wouldn’t be rated as important by the measurebators of the internet. What is the total or usable dynamic range on these cameras? I don’t know off the top of my head, even after reading tons about them.  Exactly. This is a below the fold feature.

3.) Burst shooting frame rate – this requires more processing power, bigger buffers, and other things that cost real money to develop and produce. Plus, the faster you make this mid-priced model, the more you’ll start competing with your higher end cameras – and often the shooting frame rate and buffer size is one of the primary factors separating models, such as the D3 and D700.  The baseline D700 shot 5fps vs the D3’s 9fps. But with the grip and big battery, they were suddenly only one frame apart. Buffer depth suddenly was one of the few features dividing them. Which takes us to….

4.) Low Light/High ISO Performance – which takes us to what happened the last time Nikon tried to succeed via this metric. In late 2007 Nikon announced the D3 camera to good success. It was their first full frame DSLR, and it had (for the time) amazing low light performance. The D700, based on the exact same sensor, was announced a few months later, and the primary differences were processing speed, buffer, and shooting speed. If you bought the grip and big battery, the frame rate gap closed to 1fps difference. The D700 reportedly cannibalized the hell out of D3 sales, since if all you wanted was ONE pretty picture, and not 10 in a row 1/9th of a second apart, the D700 worked just fine instead of the pricier D3. Add the grip and you got 8fps – almost as good, and an incremental upgrade to boot (so was an AND purchase, not an OR purchase.) Thus the D3S was developed, markedly improving the low light performance of the D3 and adding HD video recording capabilities (at the lesser 720p24 M-JPEG spec). Note, however, there was no D700S – the D3S created the performance gap between itself and the D700 to justify the price once again. History lesson over. So if they upped the frame rate or low light performance too much, they’d be crowding out their high end model again – that’s no good.

5.) Other – so what other top-of-mindshare feature could you try to compete on? Dynamic range, as mentioned above, doesn’t grab enough attention. Buffer depth? Same problem, as well as sales cannibalization and implementation costs. HDR? Canon saw the demand rising as Nikon did, what if they did a better job at it than you did? (Hint: they did anyway.) As I look down the rest of the improved specs between D800 and D700, and D800 and 5D Mark II, you’re only left with one, above-the-fold headline feature to compete with:



Megapixels, baby – Go Big or Go Home. I think this was the one feature that would satisfy Nikon’s needs to both compete with Canon in a way that Canon couldn’t (or wouldn’t), and to not undermine sales of their other new model, the D4.
But in doing so, this meant they’d be setting themselves up for some challenges. The bigger the pixel, the more light it can gather, thus gaining more dynamic range, implying better low light sensitivity and less noise. By going with more resolution, that mandates smaller pixels. So your per-linear-micron-of-sensor-pixel performance has to be better than the other guy’s if you want to up the resolution, since you’ll be sacrificing in other areas. So Nikon did this, and just came out and stated that this camera is not for extreme low light challenges – spend more money on a D4 if you want that. But they had a good clue that this might be possible – if you look at the D7000 introduced a year and a half ago, it has almost the same pixel pitch/pixel density as the D800, just cropped down. Using that as a starting point, that must have given them confidence (it it wasn’t already the plan) to design/request/build a sensor of similar density yet greater size.
The upside of this is that for those using it for its intended purpose, and wanting very large prints – this camera will probably be competitive with medium format cameras from a few years ago…IF your technique is spot on. (See end of this article for discussion of how excellent shooting technique is required to get the best out of these systems)

The downside of this decision? Boy, the D3x got totally thrown under the bus. This new camera has 50% more megapixels, looks likely to at least meet if not exceed the low light performance, and be close enough in fps, buffer depth, etc. to shut down at least 95% of the reasons I can think of to buy an $8000 D3x rather than a $3000 D800. Oh, it has built in grip for vertical shooting? Fine fine fine, drop $1000 on grip/big battery/cover/charger and you have longer battery life, comfortable vertical shooting, and 6fps in DX mode on top of all the other benefits – at HALF the cost of the D3x. But how well were those selling anyway? How long until they are discontinued at this point? Finish a production run, shut down that production line. Boom and DONE.

As for the argument that this camera is getting into medium format territory..well, sorta. A little Googling shows that 37-40 MP for recent medium format digital cameras is normal these days. Pentax’s 2010 release of the 645D at 40MP seems a typical example, although resolution goes up to 60.5MP on a Phase One P65+ digital back with Dalsa’s 53.9 × 40.4 mm sensor. So are the megapixels in that range? Just barely for the contemporary stuff it would appear. Can the camera resolve enough detail to compete with those cameras?

NOW it gets more interesting. According to Nikon’s own technical guide, you start to lose sharpness around f11 on the D800 or D800E.  What stop is that the case for on medium format cameras, that have generally twice the physical size and only slightly more pixels, thus pixels about twice as big? How much of a difference does twice the imaging area make on how relatively easy it is to achieve a given level of sharpness?

Napkin math would imply twice as easy, since if, for instance, 60 lp/mm (line pair per millimenter) might generate a sharp image on the D800, with twice the target area, would 30 lp/mm be sufficient to resolve the same detail for medium format framed similarly? I’m not enough of a optics guy to say for sure, but that seems a resonable surmise. Therefore, would a 60 lp/mm lens on a medium format camera resolve twice the detail as a 60 lp/mm lens on the D800? 60 lp/mm is on the sharper end of the spectrum for Nikon glass, according to The sharpest lens I saw on DXOmark for Nikon was the 85mm f1.4D (the older one, not the new f1.4G) at 66 lp/mm. I’ve read that Leica nearly doubles that on some of their lenses….ah, but this is getting into turf for another article. Hang on, I’ll come back and have more to say – and actually do some math and fact checking and whatnot.

I’d just be really, REALLY curious to know when Nikon decided to make the D800 an Amazingly Huge Gob of Pixels, at what point they committed themselves to that. We know product development takes years, and if you rush it, compromises have to be made. Look at the Canon C300 – they clearly state that they lifted the X305 hardware codec engine for the C300, because to develop a newer codec would have pushed the entire project back, what? Six months? A year? Dunno for sure, but it mattered.

Ridiculously Required FTC Disclaimer: I own Nikon DSLRs and Canon point ‘n shoots that I have, painfully, paid for entirely with my own money. I have no working relationship with either company. By the way – did you know bloggers are legally required to declare their affiliations, but “celebrities” are not? True story.

Leave a Reply