I read Ken Rockwell's site even though I violently disagree with some of his conclusions and opinions. I mean, he doesn't shoot Canon, so I'm questioning if he's got the right to have the same first name as I. :) But, sometimes he's really right. Like when he talks about the importance of timing in getting a good image.
Recently, he wrote about his technique for getting properly exposed long exposure images at night and I've been meaning to write about how I do it for some time now. This other Ken talks about digital cameras and certain models of Nikon. It's worked for him, but I have another way of working that suits my thinking and my equipment much better, so I figured that I should still write about my technique.
Growing up as a photographer (My dad got me a Canon TX when I was in grade school) I always saw night photography in the books on how to take pictures, but never felt like it was something I could do. I was unnecessarily careful about wasting shots, which I really didn't properly get over until 2005. I never had a tripod and I didn't really have the confidence to brave the area outside of the TX's coupling range. Eventually, I got over it and started figuring it out for myself.
As far as I'm concerned, there are two types of night scenes to meter, where the dividing line is around -3EV @ 100 ASA. Most scenes with any sort of lighting in them are brighter than -3EV. When it's just the stars and moon, it will probably be darker than -3EV.
For "bright" night scenes, I use a digital camera as a meter. My Mamiya RB67 has no meter at all and the TX meter isn't especially handy at night, so I take a small point and shoot digital camera -- a Canon A95 -- and use that. It's got some issues, however. The longest it will ever hold the shutter open for is 15 seconds. I'll set it to "M" mode, set the aperture to 2.8, set the ISO to 400, and take some pictures at various shutter speeds to figure out how much light I want. From there, I work out the reciprocity in my head. So, at f/2.8 / 15s / 400ISO gets adjusted for aperture to f/5.6 / 60s / 400ISO and then for film speed to f/5.6 / 4m / 100ISO (I usually either use Provia 100F or Velvia 100 while shooting). I've reached the point where I've got the progression of aperture settings memorized, but if you can't remember them, just figure out how many clicks of the aperture ring or knob it is between whole-stops of exposure (if it takes 2 clicks to go from 2.8 to 4, each click is a 1/2 stop, if it takes 3 clicks, each click is a 1/3 stop), transfer the settings, and then use the aperture knob or ring to keep track of the aperture changes as you double or halve the shutter speed in your head.
Also, the A95 seems to lie in these situations. Because of accumulated dark-noise, I find that I have to add a little fudge factor for the A95's dark noise, plus the usual adjustments for reciprocity.
Obviously, this takes a little more effort than using one of several specific models of Nikon camera or a digital. However, when your two main cameras won't fit the bill and you don't want to spend loads of money on a camera, you end up doing things my way.
There is one advantage to doing this. The test exposures are too noisy and poorly composed to be displayed. Sometimes I don't even bother putting the A95 on the tripod. However, they serve as a great record for the approximate time and exposure of a picture for later analysis and recordkeeping, so I get some of the advantages of shooting digital on a film camera.
Now, given that the A95 will do, at best, a 15 second exposure at f/2.8 at ISO 400, there are limits to how din I can go, hence why I divide things at -3EV. -3EV also happens to be around the light of a full moon, which most scenes with any sort of artificial light are above.
This is where things get hard. Many exposure meters aren't specced this dim, and I've been told that it's not really worth bothering with anyway.
Beyond this, I use Marcus Kazmierczak's exposure calculator, even though it's not necessarily accurate, as a good starting point. From there, I make an educated guess based on the cloud cover (if the clouds look like the sort of clouds that would cost you 2 stops during the day, they'll probably cost you 2 stops under moonlight, too) and everything else.
Generally, you want to do the same sort of lighting conditions over and over again to develop a feel. If you shoot 2 days from the full moon, every time it's 2 days from the full moon, you will eventually discover what exposure settings to consistently use.