Traces Of The Real

In this essay, written in 1995, Lev Manovich explores the ramifications of digital technology and photography. He asks if such a thing as digital photography really exists, and to what extent this really differs from traditional photographic practice.

Manovich starts by referring to a range of digital innovations that have transformed the practice of image production and manipulation, innovations that would lead most people to the conclusion that the fundamental nature of the photograph has radically changed.  The aim of his essay is to question whether this is really the case, and to expose a number of paradoxes at the heart of digital photography that become apparent when we attempt such questioning. His position is that it does not – that, in fact, digital photography does not exist.

He starts by claiming that digital technology is being used, not to establish new ways of working, but rather to re-inforce existing ones. In support of this he cites examples of how digital movie making software incorporates terminology from cinema into it’s operation, how computer games increasingly employ cinematic codes and practices, how the film “look” (e.g. graininess) is simulated in both digital video and photography, and how digital modes of distribution (e.g. CD-ROM) are simply used to distribute digitised versions of traditional films. The point is that in spite of the innovations of the digital, more often than not, digital technology is used to simulate and continue codes of practice that are inherent to the traditional pre-digital world. He sees this as a paradox of digital imaging.

He then considers the question of whether the physical difference between a digital and a traditional photograph has any significant bearing on it’s fundamental nature. William Mitchell’s The Reconfigured Eye tries to answer this by presenting a number of justifications for why it is correct to state that it does. For example, Mitchell claims that there is a core difference in that digital images can be reproduced ad infinitum without loss of quality whereas traditional ones cannot. Manovich refutes this by pointing to the fact that digital images are encoded with lossy compression, so every time the image is reproduced there is a loss in quality. Mitchell also claims that there is a huge difference in the amount of information encoded with a digital image, in comparison to a traditional one. The digital image has a fixed number of pixels whereas the continuous tone photograph has an indefinite amount of information contained within. Again, Manovich refutes this by saying that a digital image contains enough information that it makes no difference to the eye, which is reasonable, but then makes some spurious statements regarding the possibility of using software to go beyond the limits of a pixel based representation (see comments on this below). Mitchell’s third distinction is that the digital photograph is manipulable in ways in which the traditional photograph is not – it can be edited, altered, reworked and so on. Manovich’s counter-claim is that manipulation of photographs has always been part of photographic practice, for example the photographic montage of Soviet realism.

Having decided that the differences between the digital and the analogue photograph are nowhere near as significant as normally assumed, he moves on to the second major theme of the essay. This is concerned with the methods of production of digital photographs and in particular with the use of 3D graphics technology to generate synthetic imagery. He suggests that this is in fact a radically different means of image production that is enabled by digital technology, and that if we are to speak of a digital photography, then it really should be this. The goal of technologists involved in this work has always been to try and produce “realistic” imagery but once again there is a paradox here. What in fact they are doing, is not trying to simulate reality but rather to simulate reality as experienced through photographs or film. As evidence of this we need only to point to the use of the metaphor of a camera within any 3D graphics applications and the simulation of photographic effects such as depth of field. Once again, we have a sophisticated digital technology being used to mimic a traditional form. He concludes by stating that the aesthetic of films such as Jurassic Park is closely related to Soviet realism – it is an attempt to present a vision of the future, rather than traditional photography which points to the past.

Manovich makes a number of really interesting points in this essay, but there are some things about which he is clearly and simply wrong. For example, he refutes Mitchell’s claim that a fundamental difference between digital and traditional photography is that digital photographs can be infinitely reproduced without loss of quality, whereas film-based photographs suffer a loss in quality upon reproduction. Manovich’s objection here is that this is true in theory, but in practice lossy compression is employed (e.g. for a JPEG image), and that therefore each time “a compressed file is saved, more information is lost”, meaning that “there is actually much more degradation and loss of information between copies of digital images than between copies of traditional photographs” (my emphasis in both cases). The problem here is that Manovich is confusing “saving” with “copying”. I can of course make as many copies of a JPEG image as I like, and I, or anyone else can view those copies as many times as they like, without any loss of information. They are infinitely reproducible. The only difficulty that arises is if I wish to manipulate the photograph in some way during this process. In that case I need my software to decode the image, carry out the manipulations, and then re-encode it – when the image is then saved (which will have to be the case for my manipulations to take permanent effect) loss of information will ensue.

This is not the only spurious technical information in the essay. Manovich also suggests that the limitations of pixel-based representations have been overcome and that image-editing programs can convert a pixel representation into a “set of equations” which effectively remove the limits of the pixel. What he is referring to here is software that converts a bitmapped image into some sort of vector graphics representation. This is in fact extraordinarily difficult to do, and 15 years on from the time of writing of this essay, is still only possible in severely constrained circumstances. Unfortunately the pixel still does exist, and every working photographer is keenly aware of its limitations.

This is not to say that his basic point that the fundamental nature of the photograph has not changed much is not correct. Whether the internal representation of an image is a grid of pixels or whether it is light impressions on an emulsion makes little difference to the viewing experience.  What I think he glosses over however is how the production process of digital photographs has changed, and how the dissemination of these photographs has changed – both of these factors have radically altered how people consume and produce photographs. He ignores the digital innovations that have led to entirely new picture-taking technology within cameras  (though to be fair this was maybe not as apparent in 1995 as it is now) and he ignores how the Internet has radically changed how photographs are disseminated and consumed around the world.

Photograph by Andreas Gursky

I will use a photograph by Andreas Gursky to illustrate this post. Gursky’s work is heavily digitally manipulated but his initial shots are created on large-format film, before being scanned into the computer. I think Gursky’s brilliant images are sufficient evidence that we don’t have to look at 3D synthetic imagery to find examples of new types of photographic work that are dependent on, and enabled by, digital technology. This digital photograph certainly exists.