Bond Street & 64th Street, Southwest Corner

Sunday, May 2, 2010

The 1:3 apect ratio

All but three of the photographs in New York in Plain Sight were taken with digital cameras with a sensor aspect ratio of 2:3 — the "classic" ratio of 35mm film, dating back to the Ur-Leica of nearly 100 years ago. The three that were not were taken with a borrowed camera with a sensor aspect ratio of 3:4 — curiously, perhaps, this was a Leica-branded (actually a Panasonic) digital camera.

So if uncropped and otherwise geometrically uncorrected, the photographs would have all (with three exceptions) had an aspect ratio of 2:3, that is, 50% wider than tall, since they were all taken with the camera held horizontally.

I have cropped all of them to an aspect ratio of 1:3.

Lexington Avenue & 49th Street, Northeast Corner

Because in general I composed the photographs with the horizon or eye-level line in the middle of the frame, in most all of them, this cropping amounted to eliminating, more or less, the upper and lower fourths of the image. Often, though, I found the photograph as I took it to have too wide a field of view — too "wide-angle" — and so I cropped it in from one or both sides as well, with a corresponding narrowing top to bottom as a consequence.

It's not easy to hold a camera perfectly level, and so in the processing, I also corrected for that, i.e., rotated the images so the horizon line would be parallel to the top and bottom edges of the frame. And then corrected the converging verticals that are an unavoidable optical result of not having the camera level in the sense of parallel to the plane of the buildings. (In some cases the camera was deliberately held pointing up or down, but even so I intended to, and did, correct the resulting converging verticals).

Otherwise, except for exposure correction and color (white) balancing (when obviously needed), I left the images "as is." And often enough I left the color balance alone too; less often the exposure. Had I been working with a tripod and a shift lens, the geometric corrections would have been done at the time the photograph was made, by leveling the camera and using shift to avoid the need for pointing the lens up or down.

Which would have been better, I think, but vastly more time consuming — I'd still be photographing!

So in a certain sense the biggest "editorial" decision I made — other than choosing which frame to use when there were several, as there usually were — was where to crop. And this decision was constrained by my having settled, early on, on a 1:3 apect ratio. (I should say, at this point, that the original "raw" files are unaffected, so it's always possible to go back to the image as shot.)

The difficulty, of course, is that if I come in a little more from the sides in order to eliminate X or to concentrate the image more on Y, then I'm bringing it in from the top and bottom as well, and may then have to choose between getting X out of the image on the right or left or not having A or B near the top or bottom.

Well, such is life. The real question was, why 1:3? Actually, there's a logically antecedent question, or even two questions: Why crop at all? And why crop uniformly? Why not have different aspect ratios for different corners? After all, there's nothing sacred about the sensor aspect ratio.

The answers are pretty straightforward, I think, though the thoughts aren't inherently compelling except in regard to my intentions for the project.

The aspect ratio 1:3 is simply what you get if you put two pictures together, each of which has the classic 2:3 aspect ratio. So in my ideal corner picture (which occurs relatively rarely — here as so often the "ideal" in theory is anything but ideal in practice), there would be in effect a 2:3 image to the left of the center of the corner and a 2:3 image to the right.

And there didn't seem to be much point in including a lot of asphalt in the photographs, so cropping that out made a lot of sense to me — indeed, as the project progressed, I found myself cropping out more and more of it, and moving the horizon / eye-level line down from where I initially placed it, at the midpoint (vertically).

Correspondingly, these moves eliminated a lot of the upper part of the frame, so that, more often than not, the upper floors of the buildings are elininated as well. (There's a lot variation in this across the set). Which leaves a picture that is a little street at the bottom, then mostly sidewalk and whatever is happening on it, and a little architecture above the sidewalk (ground floor) level.

I wanted the photographs to be "about" everyday life at street level, and this cropping helps, I think, to narrow the frame to include that and not much more. But I also wanted a lot of variety, since I felt that there was a danger of getting too mechanical, which would call attention to the photographs at the cost of seeing the subject.

Which is also why I wanted a uniform cropping: it would be distracting, clicking through hundreds or thousands of images to have the aspect ratio changing all the time. I want the viewer looking at the subject, not distracted from it by anything else about the pictures.

If I reshoot New York in Plain Sight as I hope to, next year, I think I'd aim for rather less variation, but I'd still try to avoid complete uniformity. And I think I'd stick with the 1:3 aspect ratio, or something very close to it, say 1:2.5.

Just thought maybe you might want to know ….

No comments:

Post a Comment