Rangefinder

Views on the photographic universe by Erwin Puts

The algorithm threat



When the digital tsunami blew over the photographic landscape and killed Kodak, Agfa and many others (including Leica that hardly survived the storm) I concluded that the photography as we had known it for a century was dead. I was severely ridiculed for this statement and in hindsight I have to say that I was only half-right. Photography is not dead, it has simply vanished, like dinosaurs and Renaissance painters. Some relics and artifacts may be found, but in essence they are gone. So is photography. There are may persons who claim and indeed believe that current digital photography is simply classical (that is chemical) photography with new technical means. This approach is also visible in Leica's advertisements: old values can be renewed and rejuvenated in a digital coat.
Why has photography vanished? There are two facts that undermine the belief in the phrase: new technique - old values.The first fact is the composition of a digital file: it is nothing more than a bunch of data, like any statistics data base. It is also accessible by the modern technology of Big Data. Face recognition is only one example. Many cameras have this feature and adjust focus and white balance and exposure to this data selection. Lightroom and the likes search the database of image data and select the files with the same face. Behind this process that seems to be a convenient feature lurks the algorithm. An algorithm is a mathematical set of rules that work autonomous and without human interference and are self-correcting. Many computer programs include software that can debug the program and correct it. In digital cameras many algorithms are active in focusing the lens, compacting the picture, enhancing the captured detail and the base sensor sensitivity. Even exposure is decided by algorithms. This is all very helpful, but the bad point is that we do not know what happens. We see only the result, not the original input and certainly not the manipulations during processing the file. This is the fundamental even paradigmatic change from old (documentary) to new (fluid experience) photography. It may still be called photography, because it is a convenient and recognizable term, but who now would call a modern automobile a horseless carriage? Algorithms correct optical aberrations (like distortion and vignetting), but do we have any idea what is happening. Programmers and their code are becoming even more important. The frequent updates in camera software are an indication. And almost every problem detected in cameras will be solved soon in the next update, is the standard response by camera manufacturers.
The algorithm is a powerful tool, but also a dangerous tool. This technology enables machine communication and automation 'under the bonnet'. We do not have a clue what programmers have constructed and we do not know how the many algorithms work inside a camera. In the past it was Kodak's slogan: you press the button, we do the rest. Now it is: you press the button and the program takes over. What the program does, we do not know and it seems that many photographers do not care. But a photograph as a fixed physical print, the heritage and genesis of which can be controlled and judged is gone. Now we have a data set that is fluid (continuously changeable) and can be manipulated by unknown algorithms. This difference may be downplayed and many experts in the field do dismiss it as useless philosophical musings far removed from the photographic practice.The photographer does take the picture and does upload it (current) or print it (less current) and therefore is in complete control.Technical details (the algorithmic processes) are irrelevant and not interesting. We are however burying our heads in the sand.