When Black and White Are Gray
So how do you know if you are looking at a gray object in average light, or a white one in dim light, or even a black one in very bright light? Generally this dilemma isn't a very challenging one since we can see that object in context and our brains automatically help us figure such things out. But what about your camera?
Last week when discussing meters and histograms I mentioned that a camera meter has no idea of anything outside its field of view. It doesn't see "context" and has no real idea of what it is looking at. Indeed, it's really not all that smart about the world around it. It's great for measuring how much light is being reflected from something, but that skill can only take it so far. Your meter can't really tell how bright something is, only how bright it appears to be under the prevailing light. In any meaningfully real sense though, it has no idea what's in front of it.
Camera makers long ago standardized meters on the assumption that an average scene should be recorded at an average brightness. They had to pick something, so this seemed as good as any standard. In other words, the only thing your camera meter really knows is that on average, things are, well... "average." That's how it's calibrated. Specifically, "average," or "medium toned," is the same color as a standard gray card. It reflects precisely eighteen percent of the light that falls on it and is therefore known as 18% gray.
You might think that medium would be 50% gray, but keep in mind that our eyes are sensitive to doubling and halving of brightness, not absolute changes in brightness. That's what a "stop" is, after all. Whether you double the exposure time of the make the aperture opening twice as big, you have increased the exposure by one stop. Eighteen percent really is medium to our eyes. It just works out that way.
So back to our original question: How does your camera meter know if it is pointed toward a gray object in average light, or a white one in dim light, or even a black one in very bright light? As you probably have guessed by now, it doesn't. To prove the point, here are images of a standard gray card together with shots of white and black poster boards. All three shots were made on aperture priority with the shutter speed picked by the camera meter. All three images came out basically indistinguishable from one another. That's because all three looked basically the same to the camera meter. It automatically compensated for what it perceived to be lighting differences to make all three images come out eighteen percent gray. That's all it can do.
So if you want the white card to look white and the black one black, what do you do? It's up to you to adjust the exposure compensation to make this happen, or shoot in manual exposure mode and set the meter accordingly. As you can see from the above, the difference between the gray and black images was the difference between 1/8 second and1/2 second — that's two stops (1/8 to 1/4 to 1/2). And the difference between gray and white was also two stops: 1/8 second down 1/5 and then to 1/30 second. So if you are metering on white and want it to appear white rather than gray you need to add those two stops back in. Your camera isn't going to do it since it thinks the entire world is medium gray. You have to do it. And if you meter on white and want it to appear white not gray, you'll need to increase the exposure by two stops to get where you want to be.
Of course the same applies to changing how any of the other subjects gets recorded. Want a white card to look black? No problem. The camera will get you half the way there automatically by rendering it medium gray. All you have to do is subtract another two stops and you'll end up at black. To the camera, all three cards are the same so you'll need to subtract the same two stops from each to make them look black. Or add two stops to each to make them all look white. The choice is yours.
If you've ever ended up with gray dingy snow pictures rather than white you know the two stops I'm talking about here. Don't overdo it though. If there was any snow brighter than what you actually metered on, adding two stops to it risks rendering it as burned out white without any detail. When in doubt it may be safer to add only a stop and a half if the lighting is uneven.
Keep in mind too that you are rarely metering on something so uniform in brightness. Most scenes are more varied so be careful about what you are actually metering on and how your camera will see it versus how you want it to come out.
It does take practice to get good at this but digital makes things much easier than it once was since you can check the histogram after you take the shot. You remember the histogram, right?