Why do people think the way they do?

[apparentparadox]

A while back, someone recommended that I read a Scienterrific American article on photo-tampering (see this, and this, and this. I finally did, and it kind of blew me away - not the article itself (which was good), but the realization that people -- including me -- treat photographs as trustworthy

Comments

[markens]

Here is another interesting link:

Iran missiles

[apparentparadox]

Yes, it's getting more clear to me that print media can't be trusted either with their pictures (or stories).

[apparentparadox]

But, in the world of digital cameras, there is no negative. Only bits on a disk.

[anonymous]

Legally, at least, my understanding is that it is the photographer who is the witness rather than the photograph that is the evidence.

I'm wondering how long it'll take before we have cameras that can digitally sign their photographs. The technology certainly exists.

On a side note, Mark's moral compass for how far he could modify his underwater photos was "What would Ansel do?" Ansel, it turns out, did quite a bit of manipulation.

[apparentparadox]

It's actually amazing how much just cropping (or choosing your viewpoint when shooting the picture) can do.

[fyellin]

I've had many conversations with Paul A about this: how easy, given the bits, is it to recognize tampering? In most cases, it is surprisingly easy. Every camera has a certain predictable pattern of noise that gets lost when a picture is faked. A good fake is hard to create.

Have you ever seen the book "The Commisar Vanishes"? It's about people disappearing from photographs during the Stalin era as they fell out of favor.

[apparentparadox]

I've heard of the book, but haven't seen it.

The Scienterrific American article talks a bit about this, and also compares it to the spam/anti-spam "arms race" -- as people develop algorithms to determine fakes, the fakers get better.

There's some evidence that the gov't is using noise signatures to track where pictures come from -- that they can link pictures from the same camera, and so gather more info in order to find the person who took the pictures.

[fyellin]

I'm wondering how long it'll be before cameras digitally sign their photographs. The technology certainly exists to make this feasible.

[apparentparadox]

And how long before someone cracks that algorithm . . .

[pklexton]

I've also often wondered why people treat photographs as relatively reliable. I suppose it depends on what the alternative is - a photograph as compared to what? If the alternative is an eyewitness narrative description, I can perhaps understand. While both are far from infallible, I would think it's easier to detect fakery with a photograph. Though perhaps not by much.