(commentary)
When you’re offline for a couple of weeks as I’ve been, you get the chance to process information without the constant ping of new data points arriving at your senses. Too many people—including some who were at the photo workshop I was teaching—try to stay too connected and thus never get relief from the constant data stream the Internet, TV, radio, and other media is targeting at us. So while there was cellphone and Internet access available to me on probably four or five days in the last two weeks, I simply left my phone turned off. I prefer to let my brain do subconscious processing on data already received every now and then.
One of the things that immediately struck me upon re-emerging into the fire hose that is the worldwide Web was this: image sensors aren’t really the issue anymore, yet they’re all anyone seems to write about. I’m not saying that we won’t continue to see the constant stream of sensor progress we’ve seen in the past, only that, at the moment, there’s almost a complete standardization well above the “acceptable level” bar. Consider:
- 50mp Medium Format, mostly Sony sensors now
- 24-36mp Full Frame (FX), all Sony or Canon sensors
- 16-24mp Crop Sensor (DX), mostly Sony, Nikon, or Canon sensors
- 16-18mp Smaller Crop Sensor (m4/3, CX), Sony, Panasonic, Nikon sensors
Even within the different sensors there’s some standardization: Nikon, and to a smaller degree Fujifilm, seem to have cross-pollinated a number of advances in sensors so that previous gains are now available across a range of makers instead of one (e.g. Nikon/Aptina’s phase detect pixels in Sony sensors; the original Nikon/Sony photosite ADC now promulgated to Toshiba and others). The only exception is Canon with its APS and full frame sensors, where they seem to be going it alone. Thus, to expect much sensor-level differentiation in cameras at this point would be looking at the wrong target. Sure, you’ll see short-term differentiation as one company adapts the latest-and-greatest sensor advance first. But if you look over a longer time period—say a year or two—you’ll find that everyone is basically jumping on the same bus going to the same location.
Moreover, pixel counts are plenty for 99% of what people want to do, high ISO capability is probably plenty for 95% of what people want to do, and dynamic range is beyond what we had in film and currently headed into HDR territory.
So while I’ll continue to analyze sensor changes and what they mean for my photography, I have to say that at this point I’m not going to get excited about any incremental or evolutionary change, as I’m pretty sure those will all accrue to the gear I use within the next update cycle.
On the other hand, the imaging ASIC that deals with the data coming off the sensor is still an every-company-for-itself game. The fact that we can take an identical sensor and get different results from different companies almost always is due to the dedicated ASIC that is processing the data (e.g. Sony A7r versus Nikon D800). Yet even that’s getting more mature and lookalike, though it still is a special sauce that differentiates in small ways for the time being. I expect that differentiation to mostly disappear soon, though. Or at least fade into the realm of unimportant.
Which brings us to the Million Dollar Question: what is it that does differentiate cameras in the coming cycle? Price. Size. Weight. How the camera interacts with the user. Integration into the users’ other digital world. Oh, and one more thing: focus performance.
That last one really struck home in my recent Galapagos trip. I had three complete systems with me: Nikon D7100, Fujifilm X-T1, and Olympus EM-1. What was my overall response to these cameras? It all boiled down to one and only one thing: focusing.
Let me start out by saying that none of the three were perfect. The Nikon DSLR requires a lot of attention to detail to extract focus performance. AF Fine Tune. Setting the proper focus area mode for the task at hand. Getting the Track-On crossover point right for the subject. And more. However, I’m used to all those things and can usually dial in focus to where I need it. Still, on the fastest and most sporadic moving birds near at hand in the Galapagos, I sometimes struggled to get high hit rates on focus. On the more predictable birds (e.g. Albatross), no real problems. However the truly bad news is that we really haven’t had a significant increase in focus performance in DSLRs for at least seven years, maybe more. The focus issues I struggle with on a DSLR today are what I struggled with in previous generations of DSLRs.
The Fujifilm and Olympus were a different story. I’d say that the Fujifilm struggled partly because of lens, partly because its phase detect system isn’t yet up to the performance level of DSLRs. The lens part is because all we really have at the moment for telephoto work of the kind I was doing is the 55-200mm f/3.5-4.8. It’s not a snappy lens when it comes to focus. So when you’re trying to dial in a bird that’s coming at you at warp speed, even if the Fujifilm manages to attain focus initially, it struggles to maintain it. Getting continuous sequences of in-focus birds was near impossible.
The Olympus had a different problem. I don’t think lenses are its problem in this situation. Most have been designed for fast movement commands from the camera (e.g. MSC). More often than not I was able to get fast, sure initial focus. The problem was more this: if the motion wasn’t too fast or too random, the camera would get slightly more winners than losers in the continuous focus realm. But past a certain point in each type of motion, the camera would fall out of focus and not catch up.
I mention all this because of one thing that I realized upon resurfacing back to the Internet: the real issue with digital photography these days has moved mostly from cameras to lenses and focusing. The sensors and ASICs are ALL up to the job of producing great looking shots, at least if you put a great lens up front and get it to focus correctly. Almost always, the reason I don’t get the shot I want devolves to lens (including focus issues) or not dialing in the right settings quickly or consistently. That’s it.
To some degree, that’s why the Nikon 1 is so intriguing. It’s the only small camera that has DSLR-like focus abilities. And before I get attacked by the fan boys repeating camera company marketing claims of “fastest focusing camera” hyperbole, let me point out that focusing on a static subject should be fast. If a camera can’t do that, don’t buy it. It’s subjects in motion that are the killer. Even for DSLRs. But DSLRs are still at a level that hasn’t been achieved by anything else, though the Nikon 1 sure comes close. (Part of the Nikon 1’s “comes close” is that it doesn’t have any real ways to tune/adjust the focus system as you shoot without dipping into menus, and even then the choices are minimal.)
All this was reinforced during image review during the workshop. The number one thing that we’d talk about was composition, which is as it should be in a workshop. The number two thing was almost always not quite nailing focus or getting a exposure or focus-related setting slightly wrong. I rarely needed to talk about what the sensor/ASIC were doing, even on the 1” sensor cameras that were being used on the trip (mostly AW1’s). And then is was usually about dealing with the extreme dynamic ranges you encounter in a place like the Galapagos, and strategies for dealing with that in post processing. Even a D800 with its large dynamic range capabilities can’t necessarily do full justice to white bird on black rock in bright sun. The only difference between the low dynamic range cameras and high ones was that you might be able to preserve more lava detail in the latter, but only with post processing.
Funny how when you go out and shoot you have different problems than when you’re sitting at your desk writing about photography ;~). If the camera companies want to break through to a new set of users, I’m almost sure that focus is what has to improve. Speed and accuracy of focus, especially for moving subjects. Even on DSLRs. Second place would be keeping the user in constant and direct control of what the camera is doing using the most simple and straightforward manner (in other words, don’t add more buttons!).
Sure, keep up the sensor improvements, but that’s not where you’re going to win any new customers these days.
Perhaps if the camera companies were actually out shooting with their cameras, they’d realize the problems they really need to address.
