My interest in photography started in the 1980s with a small disc camera. At some point, my brother, a professional photographer, looked at my pictures and commented: “You shoot in color, but you think in black-and-white. Here’s some TMAX film.”
I started with “digital” photography by scanning my paying for high-quality scans of black and white negatives of that TMAX that I shot with my Canon Elan II. The results were mostly good, but the process took a couple of weeks and cost more than I care to remember. At some point, I started using a Canon 20D digital camera, with the same lenses. Now all I had to do was copy the files from the flash card to the PC, and tinker with them in Photoshop. While this produced much better digital pictures, I had to give up my black and white photography because no matter what I tried, I couldn’t get these images to rival the scanned negatives. The tradeoff between convenience and quality swayed toward convenience.
Next came the iPhone, and the pendulum swung even further: the convenience of a camera that was always in my pocket combined with its ability to distribute the images quickly to waiting grandparents around the country made it much more likely that I would use it rather than the camera. Of course the latency sucks, the resolution is poor, the light sensitivity isn’t what I’d like, and there is no optical zoom, but it’s always at hand.
Of course what I’d really like (short of a Sherpa for my camera gear) is a lightweight SLR with a network connection, a large display (the 2oD isn’t large enough), and an auxiliary display in my glasses so that I don’t have to look through the lens when I don’t want to (such as when trying to be unobtrusive while doing candid photography).
It’s not clear what restrictions physics places on such a device, but then again a few years ago, the iPhone-style camera was the realm of science fiction. But there are some interesting trends, both in terms of usability and capability. The current generation of point and shoot cameras is producing good images (although with latency that makes them unsuitable for action photography), augmented with clever computation such as face — and now even smile — detection. These features while not perfect, will continue to improve, and to make it easier to do the routine, just as anti-lock brakes and traction control have made driving more routine in a range of situations.
In parallel to these products, there is an interesting development in camera-related research. A group at Stanford has combined Canon optics and sensors with a Linux box to produce what they call the Frankencamera. The idea is that computation can be applied at the device before, during, and after imaging, to help improve the quality of the image and to create images that might not be possible without such tools. This is an interesting development not only because it makes experimentation with computational photography more widely available, but also because it pushes the notion of what digital art photography can become.
Just as anti-lock brakes don’t make people better drivers, neither will face-, smile-, or strike-out detection make people better photographers. While technology can improve the mechanics of human expression, it cannot (and should not) improve artistic expression. But that is not mean that art and automation cannot be combined. In fact, it will be interesting to observe these “Camera 2.0″ tools in the hands not only of researchers in image and signal processing, but also in the hands of artists who will find ways to use this dynamic medium in new and exciting ways.
Now if only the Frankencamera would fit in my pocket.