The "problem" is that your camera's meter is reflective type, meaning that it reads the light being reflected off the subject(s). The camera's system interprets everything as "average gray", so in a bright white scene it's going to underexposure. How much depends upon how bright the scene is... it's a lot different on an overcast day or in shade, than it is on a cloudless, sunny day... and it's different early in the morning or late in the afternoon, than it is at midday. (Note: You'd have the opposite problem if shooting an unusually dark subject, such as a black bear in a coal mine. The camera will want to over-expose.)
As you can see from the previous responses, there are several ways to deal with this: exposure compensation, bracketing, or by using a different type of light meter.
To really "nail it" with exposure compensation, you can learn to judge the scene and set the amount of compensation by eye. Depending upon you, this can be pretty accurate... or a bit spotty in results.
Bracketing is also a bit hit and miss. For a snow scene, I migth do both some compensation and some bracketing... say three shots +1, +1.5 and +2. One of those should be pretty close.
I agree with a previous post that, used right, an incident meter is probably the most accurate way to arrive at an accurate exposure. This type meter measures the light falling onto the subject, instead of what's being reflected off it. So it's not influenced by the tonality of the subject. This type meter can't be built into a camera to provide auto exposure, unfortunately. When you use one, it's pretty obvious why (normally you point it toward the light source, instead of toward the subject).
If you only have a reflective meter (such as the camera's), you also can use an 18% gray target (some say 15% gray, if you want to split hairs). There are "gray cards" for this purpose, as well as other types of targets (I use a Lastolite EZ Balance folding target that also serves to set custom white balances). As some suggested, it's also possible to use the palm of your hand... it's probably one stop lighter than 18% gray (that's almost always true, with most people).
Be it a gray card or your palm or some other type of target, it's important it be oriented to the light source the same way as your subject when you take the meter reading.
All these can be used in Manual exposure mode. Or, if using some form of auto exposure, most cameras have AE Lock that can be used to hold an exposure setting for a short period of time (the Canon cameras I use hold it for 16 seconds... check what your cameras do).
Actually with purely manual exposure you can probably use the "Sunny 16" and "Shady 8" rules pretty accurately, too. These simply say that for any given ISO, on a sunny day you should set f16 aperture and a shutter speed that's the reciprocal of the ISO. For example, if using ISO 500, set 1/500 shutter speed. If using ISO 100, use 1/100. Etc.
Shady 8 applies to deep shade or an overcast day and f8 simply opens up the aperture to give two stop more exposure. Lighter shade might only need f11.
You can vary these as needed. For example, ISO 500, f8 and 1/1000 give the same exposure as ISO 500, f16 and 1/500, say if you wanted a faster shutter speed or a larger aperture for some reason.
aellman wrote:
Not so sure about this. The incident meter only knows how much light is hitting the subject. It doesn't know what the subject is. So you end up compensating anyway...
No, you don't.
Well, you might tweak exposure slightly over or under if you want to render more of a high key or silhouette effect, or some variation of these. But it's usually far less tweaking and compensation than is needed with reflective metering.
But an incident meter that's properly calibrated and used correctly will give you a very accurate reading. You do have to take into account what you're photographing.... For example, if you are shooting cows standing out in a sunlit field, you don't want to be taking a meter reading while standing in the shade. Or, if you are shooting the shady side of a sunlit subject, the meter should be used so that it's reading similar shade. Basically, if possible walk up to the subject and point the incident meter back toward the camera. This also works when using manual flash or studio strobe lighting, if you have a means of tripping the lights from the meter (most flash/incident meters have a PC socket for this purpose, and some can be fitted with radio modules).