In a recent post about silly extreme macro photography, I commented that some of the small-aperture f/22 shots weren't good for much except illustrating that I need to clean my image sensor. Just as a small light source like a flashlight casts a sharper shadow than, say, a large light source like a picture window, a small aperture lens setting highlights any dust on the camera's sensor by allowing it to cast a sharper shadow on the sensor's photosites, yielding noticeable spots in the resulting picture. Some of my silly macro shots had blotchy dust spots on them.

I clean my D200 sensor about twice a year. There are many methods one might use to clean a sensor, and for some reason debates about them can become quite lively. I happen to use this popular method, but whatever method you choose, you need a concrete way to determine just how dirty or clean your sensor is. This is useful both to know when it's time to clean it, and to ascertain the results of a cleaning you just performed.

The best way to test your sensor is to take an out-of-focus picture of a smooth field of brightness, with a very small aperture (high “f” number) lens. Common methods are to take a picture of a clear sky, or of a brightly-lit white wall. I generally have neither handy, so make my computer screen white, and use that.

As a joke for something I no longer remember, I once made a dust reference page that was totally white, except it had a few instructions on how to take a test shot. Well, it turns out that I return to that joke page twice a year when I clean my sensor, as sort of a checklist of things to do to ensure a reasonable test:

Set camera to lowest ISO and aperture-priority to highest-numbered f-stop (e.g. f/16, f/22, ...). Set lens to its highest zoom and manual focus to infinity. From close enough so that only white fills the viewfinder, aim at white of screen and take a metered exposure.

The first time I cleaned my sensor, it was a hair-raising event full of stress and uncertainty that entailed perhaps a dozen cleanings over the course of half an hour before I felt it was much better than when I started. A little experience goes a long way, and when I cleaned my sensor today, it took about 10 seconds. I took a reference shot (using my test reference page, of course ) and was satisfied to find more or less a perfectly clean sensor.

As silly as a white page with a few common-sense instructions is, it's actually useful, so I thought I'd post about it.