Following up on my last two articles on the subject of improving sensors and closing windows, let’s contemplate the following chart:
This is rough approximation of how I’d expect DxO overall ratings to improve in the near future given the D90 to D7000 progression (actually, the last four generations of Sony sensors are very close to following this line). Yellow is FX, green DX, and blue CX.
Notice I drew a bar across at about the current 24mp FX sensor mark. Really, do you need better image quality than a D750 produces? Would you really keep trading your D750 (and eventually D7300 body) for better image quality than that?
The likely answer is no, not really. To some degree we’ve already hit that bar, and maybe even with current DX cameras. 24mp DX produces perfectly fine 20” prints even at ISO 3200 these days. Do you need better than that? (Hold that thought, I’ll get to it some more below ;~)
One of the things I learned to do very early on in the tech business is just map assumed improvement curves and guess at threshold tolerance. You can actually get quite complex analyses going with multiple trend lines and intertwined thresholds. But they’re also pretty good predictors for the end of iteration and the need for disruption to continue growth.
And this is where I have the problem with where we are today. I think we’ve hit the basic threshold with DX/FX (APS/full frame) cameras and maybe even at the m4/3 level. Iterative features don’t really drive the update cycles these days, they’re a side benefit that gives more marketing points to add to the basic “better performance” one. And “better performance” is often just a number (megapixel count, highest ISO setting allowed).
I pointed out last year that focus is the one performance factor that the camera makers really need to address, even on the pro cameras. While better than what we’ve had before, I think virtually every pro photographer can point out things they’d like their current focus system to do better. So there is room for iteration to move things forward some more. Case in point, the Canon 7DII. Basically, that moves Canon’s crop sensor camera performance in focus up to the highest pro level we’ve achieved to date. Akin to D4s and 1Dx performance, in other words.
So while I was bouncing around the African landscape this August enjoying my current set of Nikon gear, the back of my mind was asking this question: if I were starting from scratch today, what would I buy and why? Here’s what I came up with:
- Really need pixels and high image quality (e.g. Landscape) — If you’re in the business of staying at the top of the game, we’re back to where we were with film: buy medium format. With the Sony CMOS sensor tech now deploying in the big cameras, we’re back to the “bigger is better” scenario. Even though photo editors don’t review images on light boxes anymore, which gave big advantage to larger formats, close examination will pick the better pixels every time. Honorable mention: Nikon D810 (and maybe Sony A7r, though they need to change their raw format compression and shutter to really compete).
- Really need low light performance and focus speed (e.g. Sports) — The area where DSLRs still reign supreme. Basically you choose a Canon 1Dx or a Nikon D4s and fire away. Both have slight advantages over the other in minor areas, so it really comes down to lenses, I think. Honorable mention: the Canon 7DII (and maybe the Nikon d7100 if you’re not a buffer buster and can tolerate consumer build).
- Want a carry everywhere camera better than your smartphone (e.g. Shirt-Pocket Camera) — An area where we’re getting some intense competition these days as we finally are getting big sensors in small cameras. Having used most of them and had a little hands on experience with the latest round announced at Photokina, At the moment it’s a bit of a toss up between Fujifilm (X100T), Panasonic (LX100 and GM5), and Sony (RX100III), but all have some things they need to do better. Honorable mention: the EVF-less Canon G7 X.
- Everything else (e.g. the general purpose, travel, jack-of-all-trades type of camera) — I now believe this has shifted away from DSLR to mirrorless. Sony alone has proven that with the a6000 and A7 series cameras, I think. I’ll bet that 80% of the enthusiast world could get by with one of those cameras and a full set of f/4 zooms and a prime or two. But Fujifilm, Olympus, and Panasonic have cameras and lens sets now that also compete well. The Sony RX10 and Panasonic FZ1000 are the non-interchangeable lens versions to consider. Honorable mention: the poorly iterated Nikon 1 V series, which have real pluses in some areas, but a lot of minuses in others.
Some people claim I’m a Nikon fan boy. Go back and see how many times I mentioned Nikon in those four cases versus, say, Canon or Sony, and in what context. In actual practice, I compromise on my landscape choice a bit (D810), but that keeps me using the same lens mount in those first two categories. I’m not sure I’d make the same choice starting from scratch though.
It’s really a shame that the Sony A7s didn’t make a real breakthrough in focus performance, nor did Sony realize their liability in their shutter shock and raw format compression on the A7r. Run those changes into the cameras and you could really see an A7 future for a shooter getting started today. Assuming, of course, that the lenses keep coming ;~).
That just shows how vulnerable Nikon really is at the moment. There’s nothing at all wrong with the D750 (see related article), but it’s getting more and more difficult to see how Nikon is clearly serving a broad set of customers, or a customer with a broad set of needs. Nikon’s mostly missing the point in the latter two categories and relying upon the legacy lens set to drive their product line still in three of the four. Which means big FX lenses on moderately big FX bodies for most people, while the trend is towards smaller/lighter.
Bringing things back around to where I started this essay: where’s your bar? I’d argue that most people shooting FX cameras at this point aren’t going to upgrade on a one-generation sensor change any more, and maybe not even on a two-generation sensor change. I suspect that’s pretty true of DX shooters, too (and certainly would be if a D300s replacement that pushes one more sensor improvement into a 7DII competitor shows up).
So where to next? For Nikon, there’s plenty of room to work in the short term period: (1) D300s replacement that pushes DX over the bar, coupled with a D7300 that fixes the few D7100 lapses; (2) an answer to the a6000 and A7 series, the XT-1, E-M1, and GH4; and (3) fix what’s wrong with the Nikon 1 V, plus the Coolpix B ;~), and push those closer to the bar.
