With this week's Phase One 80mp medium format back announcement and Sony's rumored 24mp APS sensor cameras nearing, it ought to be obvious to everyone that there is no end in sight to the More Megapixels race. As I've noted before, we're in the land of diminishing returns, though. The doublings from 3mp to 6mp and 6mp to 12mp produced highly visible differences to most people, but 12mp to 24mp and 24mp to 48mp and whatever else is in store will (do) produce less visible image quality increases.
So is there an end to this madness? After all, 24mp is enough for a 20" print without compromising dpi, and how many people really go beyond that?
Again, the answer is no. High tech is a little bit mindless at times. Once you get large numbers of engineers working on refinement of a basic technology, they just keep doing that over and over. Every now and again we hit "flat spots" where the refinements don't really produce any real user benefit without something else happening. For example, in computer CPUs, just pushing the clock speed wasn't enough at one point. Word processors and email programs rarely wait for the CPU, for example--they're waiting for the keyboard, memory, or storage. Parallel processing, faster memory buses, and faster peripherals all opened up new abilities for our computers. My word processor doesn't go any faster than it did (indeed, it might be slower), but I can now have image editing, cataloging, email, databases, Web creating and browsing and far more going on simultaneously without impacting my writing speed.
We're going to see similar things happen with image sensors. Just pushing pixel quantities up means we hit limits with lenses, for example. But if we put multiple sensors behind that same lens or clones of the same lens, new things happen. Fujifilm gave us a taste of that with the dual-pixel sensors. We're going to need more of that in the future to break through image quality barriers. Ctein proposed recently using an array concept rather than pushing up a single sensor pixel count. That's exactly the right path to increase image quality, IMHO. We've been trying to brute force image quality by making better photosites, but there are alternative strategies that actually take us further, faster.
But does it matter? Do we need better word processors or better image quality? Probably not. In both cases most people are bound by other constraints: time and output. The time and output equations are why cell phone cameras are becoming so important: the collect-and-distribute equation is faster for a cell phone than it is for any other camera. By far. Given that our output is slowly changing to 1920x1040 HD TV displays, just how many megapixels do we need in the future?
Yet the sensor and camera makers keep toiling at reducing photosite size while increasing performance. So this year we'll see 24mp APS/DX cameras and heaven knows how big an FX sensor (32mp? 38mp? 44mp?). Those new devices will be on the same steady linear line of progress we've seen for the past 20 years in sensors: more pixels, more efficient. But to what end? If your output is your Facebook wall, even the 5mp iPhone 4 does just fine. If your output is the news wire (photojournalism), the D3 was just fine and the D3s is sweetness. If your output is a desktop inkjet, pretty much any of the current DSLRs are fine.
The camera makers know this. That's why we're seeing the experiments with 3D (Panasonic and Fujifilm), more video (everyone), and the tweener mirrorless cameras. They're looking for traction to continue their growth in unit sales. As I've written before, I think it'll take a complete break from what we currently consider a camera to something new (communicating, programmable, and modular) to regenerate rapid growth (mostly due to replacing outdated generations of equipment, much like DSLRs replaced SLRs).
What I think will happen with those More Megapixel efforts is this: narrower and narrower niche users will get charged more and more money. The Leica S2 is a good example of the future at the top end: more megapixels, higher quality lenses, higher price. The majority of the market is going to opt for smaller and lighter cameras with "enough pixels and performance" that integrate better into their everyday life (emails, social networking, Web, etc.).
But that's in the future. The present is more of the same. After all, it worked for the camera companies in the past, it has to work in the future, too? But just like adding megapixels gives customers less and less visible improvement in image quality, adding pixels gives camera companies less and less improvement in sales and profit. Eventually they'll figure that out. Until then, the engineers will toil at increment improvements regardless of whether they're useful or not. After all, that's their job.