Digital post-production, the photographic document, and truth
The Power of Images, Sir Peter Herbst seminars, 4 September, ANU.
Earlier this year Paul Hansen’s image of two children killed by Israeli missiles, Gaza Burial, won the World Press Photo contest. The image attracted attention because it had a cinematic feel, as though an expert director of photography from films such as Kathryn Bigelow’s Hurt Locker or Zero Dark Thirty had lit it with movie lights. A suspicious forensic image analyst called Neal Krawetz, asked to examine the submitted JPEG file, and tracked the history of the file’s Photoshop ‘saves’. He found that, on the day it was taken, 20 November 2012, a JPEG image had been converted from the original RAW file. (RAW is the format of the first readable image file written as the camera uses firmware algorithms to convert the various voltages generated by its CCD sensor into digital data. In the RAW file image information like contrast and colour is stored as separate metadata, rather than saved and compressed within the image itself as in a TIFF or JPEG file.) Then, about six weeks later, and two weeks before the competition’s submission date, a further two images were converted from RAW formats and added to the first JPEG file. This analysis led Krawetz to accuse Hansen of breaking the implied rules of the World Press Photo composition by collaging three different images together.
The photographer replied to this accusation by admitting that he had given the image a post-production treatment similar to a High Dynamic Range photograph, where three different camera exposures of the same scene are superimposed. The JPEG files, he explained, were each ‘save as’ from the one RAW file, giving maximum tonal range and different chromatic saturation in turn to the shadows, mid-tones and highlights, before being superimposed in Photoshop. This had the effect of modulating the otherwise harsh lighting across the whole image, de-saturating the distracting light on the walls, giving the skin of the subjects a smoother, more ‘inner’ glow, balancing the viewer’s attention equally between all of the men in the alley, and heightening the dramatic illumination on the faces and shrouds of the two dead children. The superimposition was then merged. Hansen submitted the original RAW file, which he had neglected to do when he originally entered the competition, to another image forensics analyst, Eduard de Kam, who declared: ‘all of the pixels are in exactly the same place’. So, Hansen claimed, rather than shifting pixels, he had merely modulated each pixel’s colour and intensity in situ, acceptable to World Press Photo rules. An analogy that springs to mind could be to a beauty pageant, which would allow entrants to use make-up to enhance their natural beauty, but not to undergo plastic surgery to artificially create beauty.
Krawetz was not deterred, however, and subjected the JPEG image to a further ‘Error Level Analysis’ that indicated which pixels had been altered to which degree. The outlines around all the figures showed the systematic operation of Photoshop’s sharpening algorithm but also, according to Krawetz, betrayed some localized pixel modification. Meanwhile, an original reproduction of the image, before its submission to high dynamic range post-processing, had surfaced and been brought into the argument, and Krawetz noticed that some pixels had in fact been shifted. For instance the bruise on the right hand corpse’s forehead had been shortened to emphasize the glow of light on her round forehead.
Fifty years ago Roland Barthes identified six different connotational procedures at work in the press photograph, which were working away to inflect the ‘natural’ denotation of the image with cultural meaning. One of these he called ‘photogenia’, defined as: ‘the image itself ‘embellished’ (which is to say in general sublimated) by techniques of lighting, exposure and printing.’ Hansen’s post-production was pure photogenia, sublimating the brutal facts of the Gaza funeral within the elevated cinematic aesthetic that could be called: ‘the universal tragedy of contemporary warfare’ or ‘the nobility of the oppressed Palestinian people’. At the conclusion of his discussion of photogenia Barthes suggests that perhaps in the press photograph ‘there is never art, but always meaning.’ (IMT 23-24) Perhaps if Hansen had used as his photogenic reference point not Hollywood movies but, say, the gritty black and white style of old school photojournalism with its historical connotations of ‘meaningful concern’, his image would have passed without notice — but perhaps, also, it wouldn’t have won.
The furor over this image, and the subsequent digital forensics of Krawetz and de Kam, indicates that my neat analogy of plastic surgery versus make-up just doesn’t hold up any more; and nor can Barthes’ fifty-year-old mutual exclusion between ‘art’ and ‘meaning’ in press photography be sustained either. There is no ‘original’ data, which is subsequently modified in a computer. Even before the image is extracted from a camera’s CCD sensor and turned into a RAW file, firmware algorithms have been at work, sharpening edges and interpolating colour. Programs such as Photoshop provide further semi-automatic modifications along the same line. In this environment there is no single point where pre become post production, where denotation become connotation. Clearly there is a point, somewhere, where the image we see in our newspapers or on-line, in which we may still be happy to invest belief as being ‘true’, becomes just another Photoshop job; but where is that point? Clearly there is also a point where the aestheticization of the image shifts a photograph from the genre of news or reportage, to the genre of personal universalized meditation on the state of contemporary war, from specific referential meaning to generalized aesthetic art; but where is that point?
This make it much harder for people such as myself to stay up on our high horses, looking down on the plebs below unable to appreciate the different valencies and experiential nuances of various photographs. We will have to perform prettier and prettier dances in the future to stay ahead.