We’ve had three highly visible all-in-one-Android cameras launched to date (Samsung Galaxy NX, Coolpix S800C, and Zeiss ZX1) and a few with Android tagalongs (early Sony Alpha PlayMemories). With the Zeiss retirement, all have now been terminated.
I remember a heated argument I had with a Nikon executive over a decade ago: he asserted that Android was the future for dedicated cameras, I told him he was dreaming. So far, I’m winning the argument.
The problem is a simple one. It has to do with the word “dedicated.”
Android is Google’s primary thrust into the generic operating system world: it’s not specific to any one task, it’s a huge layer (or more accurately, layers) of APIs sitting on top of a kernel. Android works for a mobile phone or even tablet, because those are all-in-one devices that are expected to perform a huge range of functions.
My argument would be that if you wanted to make an “Android camera,” just improve the camera in an Android-based phone, a tactic that has also resulted in a lot of experimentation—e.g. Panasonic’s 1” image sensor phone—but hasn’t yet produced the level of results that we expect out of our dedicated cameras.
I note in almost every review I’ve seen of the Android cameras some form of “do we really want to browse the Web or play games on our cameras?” type comments. Meanwhile, in some commentary on mobile phones you see the “do we really want our phone to be bigger in order to accommodate more/better sensors and lenses?”
You see, there’s a design crevasse between phone and camera, and no one has really managed to cross it (yet?).
Some will argue that computational execution makes those small image sensors on phones equal to the large sensor dedicated cameras. Those folk haven’t actually been pixel peeping. Heck, you don’t have to pixel peep, just bring the phone image up in its full resolution instead of looking at it downsized on the tiny mobile display.
Smearing, artifacts, edge effects, contrast misappropriation, and a host of other problems show up in every phone camera implementation I’ve seen to date, and they get worse when the advanced computational abilities are applied. It isn’t for no reason that the notion of applying “filters” is a common one in the smartphone social media usage. That lets you say “see, it was intentional.”
The dedicated camera makers—Canon, Fujifilm, Nikon, OMDS, Panasonic, Sony—should be pushing to make that crevice wider and deeper. I have no idea, for example, why Nikon thinks it can avoid putting pixel-shift into their cameras. That one feature alone makes the crevice far more difficult for mobile to cross. It isn’t about just using larger image sensors; you also have to do the right thing with the larger image sensors.
One of the reasons why the Nikon executive many years ago was arguing with me about Android has to do with one of my pet peeves: cameras don’t communicate in 21st century ways. His position was that “just let the Android app developers deal with the problem.” Uh, no. Foisting a problem off on others means that you have no control over the solution. Losing control was the whole point of a presentation I made in Tokyo that triggered this discussion.
Camera makers need to give camera users control. That’s a bit antithetic to the cultural norm in Japan, so it doesn’t come easy. But ceding control to another organization or company (app makers) seems even more antithetic to the long-established Japanese Consumer Electronics strategies. The Japanese CES companies tend to get decimated when they cede control.
Most of you know that I have a fairly narrow view of camera gear. My camera is a data collection device. I want my camera to collect optimal data. Then I want the camera to push that data to my optimal data processing device. In the field that might be my iPhone, iPad Pro, or MacBook Pro. In the studio and office, that would be my desktop Mac.
So the question I have for those claiming to want Android cameras—admittedly fewer of you today than a few years ago—is this: how is adding Android to a device making my data collection more optimal?
Update: fixed typo
