Kodak.com and light meter use: an error?

[cyrus]

According to the Kodak manual on "Accurate exposure with your lightmeter":

For example, with either a reflected-light meter or an incident-light meter, if the main subject is very dark or very light, the indicated exposure will make the subject appear as a medium tone in the picture. The result will be incorrect exposure unless you apply your own judgement to the information the meter gives you.

I think this is a reference to the old "white cat in snow" or "black dog in coal" issue. If you're using a reflected lightmeter, you have to made adjustments for the fact that the lightmeter would give a reading that would make either subject come out medium gray.

However, while it is true that both incident and reflected meters are calibrated on the medium gray standard, I always thought that the this was not a concern when using an incident lightmeter (since you're measuring the light falling on a subject, not reflecting off of it.) The rest of the article seems to confirm that too.

Was I wrong?

[Kevin Crisp]

I think the key qualifier in the Kodak statement is "very." People are under the impression that if you take an incident reading and then photograph something black or white that it will look black or white. It will actually look darker than the reference grey or lighter than the reference grey, but not white or black as though you'd placed it with the zone system.

This specific point was gone into at length in one of the old Zone VI newsletters, where Picker used the 'it will come out black or white' saw as an example of something said by people who think they get the theory but don't know from practical experience what they are talking about since it doesn't work that way. Yes, the "very" dark or light subject will be closer to the realistic tone than if you'd just taken a straight reflected reading or it, but it won't look "very" light or "very" dark unless you further adjust to get it there.

So this is what I think Kodak is driving at, they just didn't explain the justification for what was being said in detail.

[cyrus]

I think the key qualifier in the Kodak statement is "very." People are under the impression that if you take an incident reading and then photograph something black or white that it will look black or white. It will actually look darker than the reference grey or lighter than the reference grey, but not white or black as though you'd placed it with the zone system.

This specific point was gone into at length in one of the old Zone VI newsletters, where Picker used the 'it will come out black or white' saw as an example of something said by people who think they get the theory but don't know from practical experience what they are talking about since it doesn't work that way. Yes, the "very" dark or light subject will be closer to the realistic tone than if you'd just taken a straight reflected reading or it, but it won't look "very" light or "very" dark unless you further adjust to get it there.

So this is what I think Kodak is driving at, they just didn't explain the justification for what was being said in detail.

SO in other words even an incident light meter reading has to be adjusted for "very" dark or bright objects. Right? So they're talking about a different problem than the "white cat in snow" coming out gray using a reflected meter reading. Got it.

[Kevin Crisp]

Yes, when the subject is very bright or very dark, I think that is a generally accurate way of looking at it. I have many times heard people say that if you use an incident reading black will be black and white will be white and that isn't really the way it works. It does to some considerable extent avoid the problem of a reflected meter moving a dark subject up to a middle grey or a light subject down to a middle grey.

The limitations of using this method make the benefits of film speed testing, development time testing and placing things where you want them with a spot meter using the zone system apparent. Which is not to say that somebody who can use a hand reflected or incident meter thoughtfully can't get consistently excellent negatives.