A New Vision


Still from concept video for the Visual Media Reasoning (VMR) software package,
currently under development by the Defense Advanced Research Projects Agency

Photography is changing, but some of the most significant changes remain like an undeveloped picture, latent and unseen. Paul Virillio once suggested that in a world of ubiquitous cameras and photographs, we will no longer dream of images, but instead will dream of being blind. It’s an aphorism I’ve repeated many times here because it remains for me a very profound one, and it’s one that seems closer than ever to being made true.

Photography is no longer the physical medium of old. Gone are the dark slides, polaroids, roll films. A minority of people still use these things, but they are akin to the living historians who spend their weekends dressing up as German soldiers. It might be eye catching, but they have still lost the fight, and every time they don their uniforms they replay an old defeat. The photograph of today is something immaterial and innately non-visual. The photograph is now a document of text, numbers and symbols which must be conjured into a visual form each time we want to view it. This transformation has implications far beyond an imagined death of photographic artistry or skill that it is often seen to represent. It has consequences for the very definition of photography, consequences which most who discuss the medium appear to still be waking up to.

At the same time there is a vast and growing deluge of imagery in the world, a source of angst even for those who simultaneously herald the realisation of photography’s supposed democratic potential. What we forget in our anxiety about this flood of photographs is that not all photographs are created equal. Much of this deluge is unseen, and what amongst it we notice and wring our hands over is a small proportion, and a very particular type of image. Many more people despair, and many more column inches deplore, the plethora of inane selfie taking than do, say, the huge mass of satellite mapping reconnaissance imagery generated each day. If we agree that we need to talk about the production of imagery in our world, it would seem that to date we have mostly been talking about the wrong producers.

Part of what causes us to fear, or even turn away from this mass of imagery is maybe that it’s quantity short-circuits the thing that photography is supposed to be about. It becomes painful to look when we can never do more than scratch the surface of these petabytes of pixels. But the gradual rise of sophisticated algorithms able to really understand photographs suggests that the onus may soon shift from humans looking at these images, to machines doing it for us. From there, the further application of new or ‘big’ data principles to imagery might mean it becomes possible to analyse vast numbers of photographs in real time. Something of a primer for this idea can be found in a recent piece An All Seeing Eye. In the future the individual iconic photograph of the burning monk or the terrorist attack will perhaps become far less important than the aggregated data of countless citizen produced images of the same event.

Already mass spectacles like 9/11 or the Boston Marathon bombings are recorded by countless photographic eyewitnesses, and those images are later used to forensically reassemble the events. But this is manual, and in the future it may be automatic. As computer vision becomes more sophisticated the photographs we rely on to diagnose, judge, and perform a host of other functions may come to never even be viewed by human eyes. All of this is heralding a world where machines are not used as portals through which people might view, but as viewers in their own right, no longer providing us with imagery, but with ready prepared insights into those images and what they mean. Exactly what this new vision might mean for us is something I will be exploring more in forthcoming posts.


Leave a Reply

Your email address will not be published. Required fields are marked *