Kodak grey cards used to have instructions about opening up half a stop to allow for the difference between the 18% grey card reading and the 12-ish%. 18% grey corresponds to an L* of 50 in CIELAB (L*a*b*, max L* is 100). A perfect 100% reflecting diffuse surface is, of course, only 2½ stops brighter than a grey card. That's a bit of a waste of the dynamic range of many films.
There is no reason to restrict your choice of key metering tone to the highlights or to the shadows. You may wish to place the most important tones on the right part of the curve. In many cases the most important tones are skin tones. Exactly where you place them on the curve is up to you. This is perhaps getting a little less important as the graininess of recent colour neg films is not increasing with decreasing density as much as it used to (ie the graininess of the film is more consistent with regards to density, with the exception of the increase in graininess at the toe).
Standardised example of this, for good reason:
The Aim Density method keys your exposure to an 18% grey card. This is a standard method in cinematography. EK give you the Status M aim densities in each layer for an 18% grey card exposure. You do your tests with your lenses and your meter and find which meter setting gives you the closest density in the three layers. It is likely that the three layers will not give the same result, but I've never known any layer to be more than a third of a stop from the chosen speed. If in doubt, err on the side of the blue-sensitive layer, because that is likely to be the most grainy*. From the manufacturer's curves or, better, from your own tests you know the latitude above and below that midtone key. There is an exact video equivalent of this film technique using a waveform monitor.
* one reason why a single RMS value for a colour film is not the whole truth.
Originally Posted by Roger Hicks
Thanks for your very through explanation, and one that is easy to follow.
I shoot almost exclusively transparencies, and I find that I generally make 2 or more meter readings of my subject area. The first is to get a reading of my highlight; from there I determine if I want the hightlight to be lighter or darker than the meter indicates. A lot of times I am shooting towards the sun, and I may need to just let it blow out. From that I meter my shadow areas, to determine the amount of latitude that I have to work with. If I determine that the shadow areas are too important to allow to fade away into blackness, I will use a 1 or 2 stop split neutral density filter (I don't like 3 stop filters) over the top area and compensate. This can be a juggling act, since I normally use a center filter with my widest lens.
I admit that if the light is changing too fast, I will sometimes cheat and use my Nikon F5 as a meter.
Thanks again for your excellent explanation,
Then there's the fascinating observation from Garry Coward-Williams, who remarked that no two photographers get exactly the same reading, even with the same meter, and when they use different meters, agreements of +/- 2/3 stop are not unusual. And yet, we all get 'correct' exposures, even with slides...