Objectivity

Over the last week or so I've been observing and remarking on some shoddy forensic analysis of photos. Never mind where or who. Several academics with alleged expertise in documentary photography and ethics weighed in. It was not a good situation. This is the photo in question:

It's shot by Ralph Pace, and was a contest winner at World Press Photo (I gather that it's done well more broadly). Obviously, it looks weird to the untrained eye. The mask really pops. This is why it's a contest winner: contest winners often look weird. The point is to stand out, after all. Anyways, one of the academics decided to Investigate Using His Skills.

First he applied Error Level Analysis (ELA) which, as it always does, detected the sharp edges. The astute observe will notice that this mainly means the mask in this picture. ELA is thoroughly discredited trash, even by design it only detects a narrow band of composites that no serious photographer would ever produce, and it is notoriously hard to interpret. It is, for all practical purposes, a false-positive farm. That said, our Serious Academic read the results completely wrong. The tutorial on ELA's inventor's web site specifically calls as not relevant out the features our hero notes as evidence.

Moving on, our boy zoomed in on the mask:

He notes that the KN95 marking is backwards, and speculates that this could be a flipped image. In passing, he notes that mayyyybe we're looking at the inside of the mask, but there's no way to know for sure.

There is, in fact, a remarkably obvious and clear way to tell that you're looking at the inside of the mask. Do you see it?

Regardless, it doesn't matter and this particular shitshow is not my point.

The point is that the first academic and then a second one, looked at this closeup of the mask, wondered if we were looking at the inside, and completely failed to notice the obvious tell. What the hell is going on here?

Maybe they were both having a bad day. Maybe they suffer from cognitive or visual deficits that prevent them from seeing things. It could be a lot of things. My guess though is that they both approached the picture assuming that it was a composite, and that the closeup mask was therefore a flipped photo. Despite handwaving toward the other possibility, they were not able to take it seriously. They never made even a silent, internal, feint at "ok, what else is there in the picture that suggests that it has not been flipped."

If true, this exhibits a remarkable degree of bias. The closeup simply hasn't got many features, the tell is one of a very small number of things that can actually be seen. It's not hiding. If you were to enumerate the Things Which Can Be Seen in this thing you'd get, I don't know, you might be able to stretch it out to a dozen elements and that only by counting both elements of some pairs. You have to really not try even slightly to prove that you're looking at the interior of the mask.

So what we have is a pair of academics (one is doing post-doctoral research in aspects of photojournalism, the other is in a PhD program on other aspects of photojournalism) who apparently cannot work out the ground truth of a remarkably simple photograph. I don't expect everyone on earth to be able to solve this puzzle, so if you can't, don't feel bad. But this is literally their wheelhouse, this is their area of expertise. I do not see how you can study photojournalism seriously and be unable to figure out what the hell you're looking at.

Tragically, I am convinced that this is in fact normal in the field. I am surprised only at how far personal bias can take you into the land of blindness.

The academic community has correctly observed that bringing a true and unambiguous objectivity to a photo is a pipe dream, it is not realistic to think one can do this. Your default posture is intensely personal, you bring your biases, beliefs, history, etc to the table. This influences which things you notice in a photo, and which you do not. This, in turn, influences what kinds of extrapolations you build upon the photo, to recreate the world that you're looking at. This is inevitable.

The proper response, one I have advocated at some length here and there, is to do your best. There isn't any choice, you have to try, because otherwise you can't do anything. You have to develop the skill of looking at a picture with a moderate degree of neutrality, and seeing what is actually there. There are techniques you can use, and practice always helps, naturally. Can you make it perfect? No. Can you do better than these guys? Unquestionably yes.

The response from the academy appears to have been to simply abandon all hope of objectivity. They make no serious attempt to bring anything but their own selves to the picture. All you can learn from their discussion of any photos is what a kind of prog-left weirdo/mediocrity might make of the picture. This could be fine, I have no problem with people explaining their personal reactions to things. It can be lovely and fun and enlightening.

We run into trouble when they present, as they all too often do, their own biased and personal reading out as some kind of truth. Either they claim their reaction as more or less universal, or if not universal at least the right one (who cares what fascists thing, amirite?) and sometimes as ground truth.

In the case we began with, we have a remarkable situation: a combination of ignorance, bias, and laziness (nobody involved actually went and located a KN95 mask to see what they look like, or photographed one to see how light falls on them. well, except for me. the farthest anyone else went was, literally, a google search.) leading to an accusation of fakery which is in the end founded on literally nothing. There is no evidence presented that stands up even slightly. The only thing we have is "well, the photo looks weird" which is does only if you don't know what white things photographed underwater with a flash look like.

The original accusation has been tempered with a lot of "well, I'm not sure so everyone should keep an open mind" which is insane, because there is, again, literally no reason to believe one of the two sides. While this thing could be a composite, literally no evidence of any kind has been presented that it is, and the World Press Photo people have attested that their forensics team has passed the photo as not-composite. At this point it borders on ludicrous to hew to the "faked" theory, at least if you're paying attention.

What we have is a biased, personal, reading being presented as not even as a universal, or some kind of cultural truth, but as literal ground truth.

This is where photo criticism has gone to die. In the hands of incompetents leaning on incoherent, unworkable, theories to produce nonsense results. Not quite the results we're looking for.



The tell, by the way, is the straps of the mask. These are always adhered to the outer surface of the mask. Placing them on the inside would break the seal, such as it is.