(commentary)
Here we go again.
Do you need a 50mp sensor? No.
Do you want a 50mp sensor? For some of you, yes.
Some of this is just “the game.” Camera companies, like auto makers, TV set manufacturers, et.al., need to have something that they can point to that is buzz worthy, that makes you want to replace your perfectly fine product with another, and that’s simple to understand. There’s a constant need for consumer products that make customers believe that not only are they getting better every generation, but that they’re disposable (thus freeing you up to buy the next one).
50mp versus 36mp is an 18% increase in resolution, best case. Versus 24mp, it’s a 37% increase. That latter number would certainly be enough of an increase for most people to perceive a difference, according to studies (typically 15-20% is the median threshold of visibility). It also means that if your lens were producing 2 pixels of chromatic aberration on high contrast edges, it’s now producing 3 ;~).
I updated my How Big Can You Print article recently, and it looks like I’ll have to do so again (50mp would still only be “Good” at 24x36”, with about 230 dpi). Yes, I’m a bit hypercritical about quality for prints. A lot of people think that even as low as about 188 dpi is “excellent,” and it might be if you’re viewing at anything that most people would consider normal distances.
But look at the chart I presented in that article again: virtually any camera 21mp or higher is going to produce what picky old me considers an excellent print at the largest size you can make with a desktop inkjet printer. Do you really need more? Unlikely. Do you want more? Of course you do; bigger is better, except in beer guts.
Think back to my D810 review. There was a camera that, perhaps, could have been the first to feature a 50mp sensor. But no, it was pretty much the same two-year old 36mp sensor that was in the original D800. Yet I and every other pro I know that’s used a D810 immediately says the D810 is a better camera. So it must not always be about the sensor, right?
So let’s not get too caught up in numbers. (We Nikon folk will have that chance soon enough, anyway.) Let’s make sure that our photographs are worth printing large before we worry too much about the pixel count.
As usual, one of the questions that I get every time we have these megapixel bumps is whether or not our lenses will hold up to the increased sensor resolution. Yes, many of them will. A few won’t. You get what you pay for, basically. I have no doubts, for example, that my Nikkor 400mm f/2.8 will continue to shine even when we get to 100mp sensors. It’s about the best lens I’ve tested (the Zeiss Otus and the Nikkor 200mm f/2 are arguably better).
More pixels just give me more sampling of what a lens does. So a great lens continues to look great, a not-so-great lens looks less great the more we sub-sample it. Another issue that raises its head with more pixels is diffraction. While diffraction produced by a lens at any given aperture is the same, smaller photosites record it better. A general rule of thumb has been that when the photosite is less than 2x the Airy disc, the diffraction is recorded well. So a 4 micron photosite will show clear diffraction impact slightly earlier in the aperture series than a 5 micron photosite. And it will certainly record the diffraction impact better than a 12mp camera with 8 micron photosites.
Personally, I’d already weeded out my lens closet in preparation for the move to 36mp, so I feel I’m probably in great shape for the next generation of pixel bumps or two. Those of you thinking about jumping from the 12mp world to the 36mp+ world should do some real soul-searching about how well your current lenses fare. Many will make the cut, but some won’t. And if you don’t care how your lenses fare, then you don’t need 36mp+ ;~).
