Tuesday, January 22, 2013

Nikon D4x? Baseless Speculation... Based on the D5200

Comments about my photo editing skills will be judiciously moderated....

More objective testing coming in for the Nikon D5200, this time from DxOMark, who have ranked this model as the best ever for APS-C cameras. Ordinarily I'm not a big fan of DXOMark, (and even less so of their lens rankings) because I find the reduction of a number of different variables down to one specific numerical rating to be overly simplistic and terribly misleading. The art of comparing things is more than just ranking what is better than another; writing and reporting should always be more nuanced than that. However, if you dive deeper into their reviews you will get that more nuanced view, and what they have said is in line with other samples being put up on the web. I've already stated that the sensor in the D5200 is the best Nikon APS-C, and have felt very comfortable doing so, almost from the get-go.

However, this is almost a no-brainer. Nikon has consistently met or exceed expectations with each new sensor generation. After the D3, nobody expected (let alone was asking for) the extra dose of high ISO power of the D3s; the D4 then consolidated that. After the D90, the sudden jump in high ISO, resolution and dynamic range of the D7000 was a shock and a delight. And then we have the mother of all generation jumps: the D800. Nobody was expecting 36mp, and nobody would have thought that so many megapixels could be so useable. So with this history, one would have expected that Nikon wouldn't just release a maintenance product that merely evolved the DX camera line; too much time has gone by. The game isn't about giving people what they expect because most people expect too little of the future. We all wanted the D3s in a D700 body at D300 prices. That would have been an awesome camera, but think of how outdated that would like in comparison to the D4 and D800 of today.

It's not that we can't dream big, it's that we aren't willing to. The ingredients are right here before our eyes. Think back to the D90; the photosite diameter was about 5.5µm. You know what other camera had pixels around that size? The D3x, at 5.94µm. If you take the 12mp of an APS-C sensor and multiply it by 2.3x times to get the area of an FX sensor, you will get about 27.6mp. The difference is off a bit; the photosites aren't exactly the same between concurrent generations of DX and FX sensors. With the massive increase in pixels on a FX sensor, they can afford to use a larger photosite to keep dynamic range and up and noise low.

However, my point is this: For the past three generations of FX sensors, the photosite size hasn't been something utterly crazy and new, but instead has been close to, if slightly larger, than a photosite size that is already familiar in the DX format. The D3 had largish photosites like the D70s, the D3x would have corresponded to something like the D200/D300, and it's easy to think of the D800 as being double a D7000. I'm not an engineer, but I can't see this as being coincidental. Sensors seem to cluster around similar photosite sizes ostensibly because the development work and R&D are already in place. It wouldn't have made sense for Nikon to have gone straight to a 24mp FX camera if they had only developed DX to 6mp... the difference in photosite size would have meant more research and development than first introducing a 12mp sensor with photodiodes closer in size to the smaller sensor.

You see where I'm going with this, don't you? The D5200 has shown us what state of the art DX can do, and also what the future will be like as more sensors move from aluminum to copper fabs. So to recap:
  • Principle 1: New product generations are more than just shuffling existing parts around into new combinations
  • Principle 2: FX and DX sensors tend to congregate around similar photodiode sizes 
Which brings us to a camera that nobody is talking about: the D3X successor. There may in fact, never be one. Canon unintentionally eroded their high end full frame 1Ds line with the introduction of the 5DmII. It was no where near as sophisticated as the fully professional 1Ds line, but it created a product positioning situation that eventually pulled the whole full frame market into a cheaper but more broad and accessible space. Same goes with the D3x... it was as exotic as anything when it came out, but the plebian (by comparison) D800 out performs it and the pedestrian (by comparison) D600 matches it. Still....

... Applying Principle 2: Double the D5200 and you get 55.2 megapixels. For the sake of argument, let's call that overkill and round it down to 50mp. Now, going back to Principle 1: putting the D800 sensor in a D4 body might be useful for some people, but it doesn't move the game on. Put 50mp in the D4 body, with technology that is already demonstrated in the D5200, and suddenly you have people talking.

Say that with me: 50mp D4x, with similar per-pixel noise and dynamic range characteristics as the D5200... it sounds crazy, but even if it's a speculative dream, it's not a fanciful one. The ingredients are already here in front of our eyes.

No comments:

Post a Comment