Featured Post

Pinned Post, A Policy Note:

I have made a decision to keep this blog virus free from this point forward, at least until the smoke clears. This is not a judgement about ...

Friday, March 29, 2013

Similarity

I occasionally opine that all the photographs, in certain genres at any rate, have been taken. What does this mean, exactly?

What I mean is that if you were to take a photograph in that genre, you would be able to find another photograph made by someone else which is very similar to yours. So, what do I mean by similar? After unpacking this a bit, I have decided that it means the following.

Suppose we have two photographs and you wanted to measure how similar they are. You could label than A and B, and then ley them down side by side. If you cannot distinguish them when confronted simultaneously by two prints, obviously they are very similar. Suppose instead you had a chance to inspect both prints for a period of time, and then the prints were taken away. Some interval of time, T, passes. Then you are shown one of the two prints, and asked to identify which one it is.

Obviously as time T grows, this task may become more difficult. If one print is of the Eiffel Tower, and the other of your mother in law, it probably does not matter how long T is. If, however, there is some similarity between the two prints, T begins to matter. Over time, most likely, fewer details remain in your memory and what's left is an overall sense of the image. The subject, usually, remains. The ideas, if any, probably remain as well as a general notion of how it looked. The overall placement of forms, the way the light fell, the color palette. At some point your probably cannot distinguish between two mirror images, unless there's something special about the left-to-right orientation, for instance.

This measures similarity in a way which I think is particularly relevant, since what it's really measuring is how similar two prints are in terms of how much of the content sticks in the viewer's head. Since I am mostly interested in photography which sticks in the viewer's head, I find this measure apropos. Your mileage may vary.

When I say that all the photographs have been taken, what I mean is that if you shot a photograph in the genre, there is very likely to be another photograph in existence which would get confused with yours on a timescale of, say, T equals 24 hours or so.

There are subjects which have been shot over and over. Landmarks, geographical features, cityscapes for which particularly fortuitous vantage points exist. Everyone sees the same shot from that vantage, and shoots it. Nearly pixel identical duplicates of every photograph of the Eiffel Tower surely exist.

Pictures of peeling doors, pictures of factory windows (one pane broken, natch), pictures of driftwood; these things have many duplicates "similar" on the 24 hour time scale. You might have to hunt a bit on flickr to find one "24-hour similar" to yours, but there's a good chance you can.

Photographs of mountain ranges you do not recognize probably tend to all blend together after a few weeks, in the absence of some very distinctive element in the frame, as do park benches.

Why does it matter?

If you want to create an image with impact, which connects with the viewer and communicates something unique to the image, you probably don't want yours muddled up with a whole bunch of other similar ones in the viewer's mind, six months hence. This uniqueness property pretty precisely measures how impactful your image is in the minds of the viewers.

Depending on what you want to achieve, then, you may need to consider this. Maybe, though, your goals do not rely on uniqueness, or perhaps your goal is in fact the subvert the idea, and you're deliberately making images which get muddled up with other ones after a while. I don't presume to judge, at least not out loud.

No comments:

Post a Comment