Skeptics about the merits of digital photgraphy for architectural documentation have noted the ease with which images can be manipulated in Photoshop. The classic complaint is that if it’s so easy to remove things like power lines, how can we trust that any digital image is an accurate representation of a building? Well, I suppose we can’t, if somebody is especially determined to pull the wool over our eyes.
This is, according to the Chronicle of Higher Education, a much larger problem in the sciences, where journal editors, working with the Federal Office of Research Integrity have developed software tools to detect digital fakery in submitted evidence.
I’m reminded of the widespread concern in the 19th century over spirit photography, which cast much doubt over the veracity of photography in general, because its images of ghostly figures seemed so real. The New York trial of William Mumler was an attempt to rescue photography from the stain of fakery and to establish its status as a truthful form of representation and a legitimate, scientific tool (see Michael Leja’s Looking Askance for a full and nuanced account of the Mumler trial and its context) .
For the purposes of documentary architectural photography, it is certainly legitimate to ask what the appropriate limits are to digital manipulation. Is it alright to remove dust spots? What about adjusting color balance? What about cropping and stretching to remove parallax? All of these are easily done (though sometimes time consuming) within Photoshop, which includes tools specifically for this purpose. The “Magic Healing Brush” allows a user to select a spot–a piece of dust in an otherwise spotless blue sky, or a bird, say–and remove it with one click. The crop tool includes a perspective adjustment, which makes it simple to turn buildings that are afflicted with parallax distortion (ie, they look like trapezoids) into perfect upright rectangles. This is to say nothing about adjusting exposure and contrast, or adding sharpening, or creating panoramas, or assembling high-dynamic-range composites.
So what are the acceptable limits to image manipulation for documentary photography (here i refer to images shot digitally, not scans of slides or prints, which require a different, much tighter standard)? Clearly, insisting on preserving the image just as it came from the camera makes little sense. No one has any problem with a film photographer’s choosing a particular film for its highly saturated colors, or a particular chemical bath, or for using a high or low-contrast paper for processing, etc. So clearly, we should be comfortable with a certain amount of manipulation that allows the photographer to present the image as he or she intended to capture it.
I, for one, routinely tweak exposure settings slightly, as well as contrast, and frequently color balance, at least on interior shots, where multiple light sources (flash, incandescent, daylight) complicate color considerably. Since I always shoot in raw format, which produces a soft image, I add some judicious sharpening as well. Sometimes I find that, despite my best efforts to get vertical lines to be vertical, that I have some parallax distortion, as well, and I need to do some minor perspective adjustments. For a discussion of the merits of and means to perspective correction in Photoshop, see this discussion. Finally, when I had a relatively cheap lens, I found that it was almost always necessary, when photographing buildings, to remove pincushion or barrel distortion, a task that is automated by the marvelous ptlens tool (if you need it, get it. There’s nothing better).
That’s really about it. I don’t fiddle around trying to remove power lines, or cars, or people. Every now and then, I might burn in and blur someone’s face but on the whole, I try to get the shot right in the camera, framing my views carefully and adjusting exposure in the field. Here’s one where I was in too much of a hurry and didn’t get it framed right. Cropping the lines out would have removed too much sky, so needless to say, I was sorely tempted to try to zap them. Sometimes, of course, there’s nothing you can do. And a lot of people, it turns out, are actually interested in photographs of power lines.
I should emphasize that all of the adjustments I make tend to be subtle. It quickly becomes obvious when you fuss with an image a lot, and even when manipulations are well done, they tend to call attention to the technique, as opposed to the building. I don’t do perspective adjustments that require distortions of more than 10%, and I try not to do them at all (as David Ames advises, it’s always preferable, when trying to capture a tall building, to frame your shot vertically, but sometimes that’s not enough). Exposure adjustments above about one stop start to reveal and emphasize noise. And trying get rid of people, cars, etc. convincingly is a fool’s errand, unless you’ve got tons of time on your hands (digital image manipulation in the context of architectural documentation is seriously subject to the law of diminishing returns).
So yes, it’s technically possible to fake a photograph of a building, and photorealistic images are now possible of buildings that don’t even exist, thanks to high-end CAD and rendering software, but wouldn’t you rather spend your time doing research, or in the field? I would.
Edit: For what it’s worth, in the most recent New Yorker, Sasha Frere-Jones discusses Auto-Tune, one of many goodies in the modern music-producer’s tool box, in terms that are analogous to the documentary photographer’s use of Photoshop. The photograph is always a transformation of a thing, in the same way that a recording, even of a live performance, is a transformation. The degree and nature of that transformation varies, but photographs, like recordings, are always mediated.