A couple of weeks ago, Jane and I went down to the Getty Center in Los Angeles. If you’re not familiar with the Getty, it’s a sort of an art museum, sort of a public gardens, and sort of an interesting pile of architecture. Among (many) other things, the Getty has a massive collection of photographs; and though I didn’t really hope to see any of them (the art galleries proper are rather small) I figured the museum bookstore would have a neat photography section. Also, I hoped to take some nifty pictures of the grounds with my new camera. We didn’t have as much time as I’d have liked, but I did get some good shots…or so I thought until I got home.
One of the differences between the Nikon D80 and my old camera is that there’s no live LCD–you have to look through the viewfinder. (“Have to”! “Get to” would be more accurate.) The LCD is used for menus, which I don’t use all that much while I’m out shooting, and for reviewing the pictures you’ve taken. Reviewing them is neat, but extremely limited. You can tell if a shot is egregiously awful…but a shot that looks good on the LCD, that is exposed properly and well-composed, might still look awful when you see it full-size on your computer screen. Consequently, I don’t tend to look at the D80’s LCD all that often, or I’d have noticed the problem quicker.
Not to put too fine a point to it, I had a horrible problem with over-exposure–not by a little, but by an awful lot. Most of the shots were simply unusable; some were exposed properly; and some few were over-exposed, but in an interesting way. Here’s an example:

What you’re seeing are a couple of giant “bouquet” shapes made of steel rebar with bougainvillea planted inside so that it spills out the top. I was trying to get the colorful tops along with part of one the Getty buildings and some sky; instead I got what you see. But why was I having so many exposure problems? I’d been experimenting with manual exposure and different ways of metering, so I assumed it was my fault, made a note that I’d have to be more careful in the future, and moved on.
The next Friday I went to my son’s soccer game, and took a bunch of pictures using semi-auto exposures (shutter priority or aperture priority), and again, many of them came out slightly or horribly over-exposed. I was nonplussed: in semi-auto mode, the camera sets the exposure; it surely ought to be able to do a better job. I got on-line at photo.net, and with my friend Ted the Test Lead, to ask about what I was seeing. Both Ted and the gang in the on-line forum eventually nailed it: it wasn’t just me, there was a problem with the camera or lens–and the evidence pointed to a sticky lens diaphragm.
When you take a picture, you need a certain amount of light to expose it properly. This gives you a choice: you can let a lot of light through a large hole for a short time, or a little light through a small hole for a longer time. The size of the hole is called the aperture; the length of time is the shutter speed. The mechanism that controls the aperture is the lens diaphragm; and the consensus was that sometimes the diaphragm wasn’t closing to a smaller size when it was supposed to, that it was getting stuck part way.
That seemed to explain what I’d been seeing; so I went out that afternoon, and started taking some test pictures. What I found surprised me. The pictures I took with a wide-open aperture were fine; the pictures I took with a smaller aperture were over-exposed. The smaller the aperture, the worse the over-exposure. It began to look like the lens wasn’t stopping down the aperture at all! Quickly conjectured, quickly tested. The D80 has a depth-of-field preview button, which stops down the aperture to the current setting (the aperture is usually wide-open except when you’re actually taking a picture). I stopped it down to F/22 (which is very small), and pressed the DOF preview. No change. Nothing. Suddenly, it all became clear–the only pictures I’d taken that worked over the previous week were those taken with the aperture at or near wide-open.
The folks at Samys Camera were very good about it, and swapped the lens for one that works properly; the salesguy said he’d never seen such a thing (with that lens, anyway), and he looked honestly shocked.
So, problem solved!
Well, maybe. I now have a lens that’s working the way it’s supposed to, and that’s a very good thing. It turns out, though, that the D80’s “matrix metering” system does tend to over-expose by about 2/3 of a stop in some circumstances. The conjecture is that it’s a marketing thing, with the intent of making high-contrast scenes look more “punchy”; the less expensive consumer-oriented D50 does the same, but the more expensive D200 (their low-end pro DSLR) does not. I’ll need to get used to that…or perhaps I’ll just experiment with spot and center-weighted metering.
Man, have I got a lot to learn!