What is "correct" exposure?

Question

Photography has been my hobby for far too many decades — first on film, and now digital (for about seven years). I recently had occasion to use exposure bracketing seriously for the first time, when I was shooting difficult subjects (high contrast, large light value range, intense lights in shot, etc). Much to my surprise, the one stop underexposed shot was virtually always the best shot. I have since been experimenting with this and find that it applies to virtually all the shots I take. It also applies if the is a lot of sky in shot, and if there is very little. The colour range is more realistic, the level of detail is greater, and the balance much better.

So what is going on? Am I just getting old and decrepit with my vision crumbling? Do digital sensors react better to low exposures? Is the manufacturers calibration suspect? Something else?

What is "correct" exposure for a modern digital system?

FWIW I use Nikon equipment; with all but one lens (a Tokina extreme wide-angle zoom) being Nikon. All the shots I refer to above were taken by natural light in the daytime, and a substantial portion of them were landscapes.

Answer 1

I'm not sure if this is a "how is exposure defined" question or an "is my camera busted" question, so I'll try to address both. :)

Definition of proper exposure

ISO standard 1271 contains a definition for photographic exposure.

Bypassing the math, "correct" exposure averages a scene's luminance and renders that luminance at a particular (but arbitrary) level, measured in lux-seconds, at the image plane.

That level has been chosen based on an assumption that the the average scene's peak luminosity is <=7.8x its average luminosity (again, an arbitrary figure).

The standard provides manufacturers with a small amount of wiggle room (it specifies a constant K, the value of which the manufacturer can select, within a narrow, defined range) to compensate for transmission light losses in the optical pathway, as well as for a rendering a particular manufacturer feels is most pleasing.

In simpler language, "correct" exposure maps a particular shade of "average" grey in a scene to a specific RGB value in the image.

Anything in the scene brighter or darker than this "average" simply falls where where it falls in your image. Or put another way, depending on the average luminosity of the scene, dynamic range of the scene, dynamic range of your imager, etc., etc., it is entirely possible to experience clipping (in shadows and/or in highlights) in a "properly exposed" image.

Is my camera exposing correctly?

In practice, manufacturers have developed sophisticated metering systems to properly weight or discount areas of an image to achieve a higher rate of pleasing images. In effect, this adjusts the the shade of grey the camera considers to be "average" in the scene.

The fact that you prefer the image your camera delivers when its meter says the scene is one stop underexposed may indicate that your camera meter is out of calibration, or may simply be revealing your personal preferences.

Your histogram may offer some clues (be sure to evaluate a RGB histogram, not just a single-channel histogram), but it would be best to shoot a calibrated grey reflective target to see where your meter places the grey in your image file--should be at least 2.96 stops below saturation for 12.8% grey target or 3.46 stops below saturation for 18% grey target.

Hope that helps,

Answer 2

I'd dare to say the correct exposure is whatever is needed for the artist to get the effect he or she desires. That might be technical perfection, but it might also be deliberate over- or underexposure used to get specific artistic effects.
I've used this myself to get seriously blown out highlights, causing a winter beach scene to look like a desert under a blazing sun. Or pushing the darker background of a photo to be completely black so that the light face and clothes of the baby resting on a dark blue blanket pop out without any distracting background elements. Technically both images are bad, but they achieve the desired effect so the exposure is correct.

Answer 3

The "correct" exposure may vary some from one image processing system to another, but the general goal is to make the darkest and brightest parts of the image both fall within the dynamic range of the camera with a good white balance and natural contrast.

The bright and dark part is easy if the scene doesn't exceed the dynamic range of the camera since it can simply make sure the brightest white and darkest black are not clipped. If the scene uses less than the full dynamic range, it becomes a bit more subjective and the metering mode comes in to consideration. A variety of different metering modes are supported by many cameras. Some choices are things like "Evaluative", "Center Weighted Average", "Average" and "Spot".

Spot metering is probably the easiest to explain, you simply choose a spot in the scene to evaluate the exposure at. It looks at that spot and sets that point to make sure the exposure is good (if it's black, make it dark but detailed, if it's white, make it bright but detailed, if it's color, make it detailed around that color.)

Average is similarly fairly easy, take the average intensity of the scene and focus the ability to resolve detail there. Center Weighted Average is similar but it gives more value to the center of the shot than the outsides.

Finally Evaluative is the most difficult to describe. The camera makes an evaluation based on some internal criteria and what the scene looks like. Typically it will try to identify the subject of the shot and focus on getting a good exposure of them.

As far as why your camera would consistently under expose (in your opinion), this may be to try to avoid blowing highlights. In general blown highlights are more noticeable than lost detail to dark. Our eyes have a much wider dynamic range than a camera and our pupils adjust to avoid things getting too bright. We aren't used to washed out images in real life, but we are used to things being too dark to see.

There is also a school of thought that under-exposing (particularly when shooting raw) to give more headroom can be a better way to shoot and then bring it up after the fact. If you are shooting RAW, it could also be that the adjustments are being made for the RAW file but then the render of the RAW is based on what the camera thinks is good. Generally, what is displayed on the camera has very little to do with if a RAW was well exposed. You don't know what information is really there until you get it in a RAW processor and play with fine tuning the exposure.

Answer 4

Ansel Adams developed the Zone System to allow him to select the exposure levels of specific objects in his photos in relation to other specific objects with different luminosities rather than basing the exposure on a single meter reading of an object with the approximate average luminosity of the overall scene. We often forget that cameras have only had built in exposure meters for much less than one half the history of photography. The availability of a built in meter with more than one metering pattern is an even more recent development.

As others have covered the technical standards very well, we'll just say that your camera's meter should operate similarly to other cameras when metering the same scene. But realize the technical standards are applied to a scene with a standardized level of variation between the brightest parts of the scene and the average brightness of the entire scene.

Also realize that exposure compensation will vary from one metering mode to the next. I've got a photojournalist friend who shoots Nikon gear owned by his employer and Canon gear when shooting personally or for a side job. He hated Nikon's matrix metering mode in all the Nikon bodies he has used prior to the D4. He feels that even the D4's matrix mode, while more usable than prior Nikon bodies, considerably trails the evaluative metering mode in his Canon bodies that are several years older. In both cases these modes base the exposure calculation upon contextual clues in the scene. For instance, if the upper two thirds of a scene is much brighter than the lower one third, the calculation will be based on the assumption that the bright area is sky.

Answer 5

The camera wants it to be Zone 5 if you smush all the tones together.

The histogram wants it to be solid black at it's darkest pixel and pure white at it brightest (though thats more of a contrast thing)

You can have it anyway you choose.

Kudo's to guy who mentioned the Ansel Adams system. The king of the correct exposure.