Understanding Metering, Part Two: What To Use, When By Ming Thein
This is part two of Ming Thein's series on Understanding Metering.
In part one we examined why metering is important, and how the basics of how meters work. In today’s article, I’lltake a closer look at the different types of metering, how they differ, and under what situations they should be deployed.
A sample viewfinder – in this case, a rough representation of the Nikon D2H/ D2X finder.
With that background out of the way, let’s look at how the various metering options work, and what typical situations they might best be deployed under. Cameras typically have three options, or some variation upon that. Within these options, it’s also usually possible to fine tune various aspects of the meter’s operation. I’m going to leave out handheld meter operation since this is something that’s almost never encountered today. An important point to note is that all meters can be fooled by situations of uniform luminance, so don’t trust the readout blindly. Remember, meters function by averaging the entire evaluated area out to middle gray; this means if your evaluated area is meant to be black or white, you’re going to need to add or subtract some exposure compensation. For predominantly light/ white scenes, you need to add; for dark scenes, subtract. This holds true for every one of the different metering methods detailed below.
The simplest form of metering evaluates the frame as a whole, and tries to expose it to middle gray – under the assumption that there will be shadows and highlights, but these will average out. Seldom used today because you will almost always require exposure compensation (making it unsuitable for the point and shoot crowd which constitutes most of the global camera market), but has the one enormous advantage of behaving predictably under every situation.
The simplest form of meter is the spot meter. This evaluates luminosity at the desired point only, ignoring everything else in the frame. There are two important things to be aware of with a spot meter: the location and size of the spot. The metering spot’s location is either in the center of the frame, or tied to the selected or active autofocus point; the logic there is that you would typically want to ensure your subject is both in focus and properly exposed. Variations on the spot meter include types that are biased for highlights or shadows – i.e. you meter a shadow or highlight and it doesn’t turn out over or underexposed. Don’t forget to add appropriate exposure compensation.
The size of the spot is also very important – don’t be fooled into thinking that it’s a tiny, precise eyedropper the same size as your autofocus area box – it isn’t! Most consumer cameras have a spot size that’s about 2.5% of the frame area, which is actually quite large – imagine your frame divided into six vertically and horizontally, i.e. a grid of 36 boxes; a 2.5% spot meter is the size of one of these boxes. Professional cameras might have a 1% spot meter; imagine a 10×10 grid of 100 boxes, and this is pretty much what you’ve got. In our sample viewfinder above, the cyan box is a 1% spot meter, tied to the active (red) AF point. Keep this in mind as you’re moving it around. If your spot meter is tied to the center of the frame, then you’ll need to assign another button – perhaps the shutter half press – to lock exposure once you’ve metered for your subject (unless it is of course dead center, which is highly unlikely).
The obvious question would be why spot meters aren’t smaller – firstly, you don’t actually want them to be that acute, otherwise moving the camera by a fraction of a degree might yield a vastly different (and incorrect) exposure – they’d be too sensitive to use. Secondly, some averaging is still a good thing – you can move the camera around a bit until the spot falls onto the right mix of light/ dark to give the desired exposure. With practice, this can be much quicker than using exposure compensation.
Use the spot meter in situations where your subject is in very different light to the rest of the frame – either much brighter or much darker – in order to ensure that the focus of your shot is properly exposed. It’s great for high key or low key images – put your subject in the shadows or highlights respectively, and spot meter there – or even general situations under which the luminance of your composition doesn’t vary that much. I don’t generally use it for street photography or fast moving situations, because it requires precision and/ or a little meter-and-recompose dance that can cost you valuable time.
One tip: the way I use the spot meter is always either covering my subject, if the subject is darker than the rest of the frame; or, on the highlights plus a bit of dark area if your subject is lighter than the frame. This effectively tricks the meter into adding a bit of exposure compensation to average out the bright/ dark areas – you need to do this to prevent your highlights from falling into middle gray and consequently completely losing your shadow information. It also adds a bit of speed in operation since you don’t have to muck around with exposure compensation.
Spot meters only came about when the metering cells in cameras could be made small enough to evaluate only a portion of the frame; they’re common now because our metering sensors are made up of hundreds, if not thousands, of discrete individual elements.
In our sample viewfinder, the circle around the center AF point represents the centerweight meter area border. That sounds like a bit of complex mouthful, but in reality it’s not. A centerweighted meter divides the frame into two areas – the circle in the middle, and the border. The circle in the middle is presumably roughly where most subjects are going to be framed, which in turn you would like to expose properly etc – it is metered separately from the border area. The two metering values are combined in a predetermined ratio – usually 70-30 in favor of the central portion, sometimes 60-40 – to determine the final exposure value.
Centerweighted meters are the predecessor to matrix metering – they try to average things out over the entire scene, and make a sensible assumption or two about what you would like to expose for. Modern cameras allow you to change the size of the center area – the D800E, for instance, allows a spot anything between 8mm and 20mm in diameter. The default center area is usually etched onto the focusing screen for reference. Note that centerweighted metering was the successor to evaluative metering, and shares its advantage of predictability: if you put your subject in the circle, chances are the exposure will be right; the advantage it has over evaluative metering is the ability to bias the exposure towards your subject.
In situations where spot metering would not be suitable – action, for instance – I actually prefer using centerweighted metering to matrix in unfamilar cameras; at least I have some idea of how the meter will respond. There’s nothing more frustrating than missing a shot to over or underexposure because matrix metering has gotten things very, very wrong.
Matrix metering is either a miracle or a curse, depending on where you stand. For those who don’t want to take control of their cameras, matrix metering provides a higher ‘hit rate’ than evaluative or centerweight; the problem is, you have absolutely no idea when it’s going to get it wrong, and how much by. This can be rectified with experience with a certain system; as you encounter more situations, you get a better idea of when the camera is going to miss. It’s for this reason that the only time I use matrix metering in a situation where delivery is critical is when I’m shooting cameras I’m familiar with – the Nikons, and the OM-D. Everything else is either spot or centerweight.
That doesn’t of course explain how it works. The frame is divided up into a number of areas – up to 100,000 of them in the Canon 1Dx – and a reading taken of each area, for both luminance and color. The camera then either compares this to a database of similar situations (i.e. photographs converted into 100,000 or however-many pixel maps, along with exposure values) and then determines the exposure. If the camera can’t find a matching situation, then it makes an intelligent guess about what the exposure should be based on a combination of overall scene luminance, color, and the current AF point. With this many variables, it’s actually surprising that the meters get it right such a high percentage of the time – perhaps there are only so many possible luminance maps?
In any case, matrix metering tends to be more reliable in situations that don’t have extreme contrasts, or bright point sources in the frame, or very small subjects. Under quickly-changing circumstances, it’s the method of choice – it might get things wrong, but most of the time it will save you from having to move around the spot or use exposure compensation. For most users, matrix metering is sufficient, and you can always add or subtract exposure compensation and take another shot. It’s also worth noting that matrix meters that use the imaging sensor are much more accurate and reliable than those that have separate metering sensors simply because the tonal response characteristics of both match, making overexposure almost impossible. Presumably, these should also run some form of ‘expose to the right’ algorithm for digital cameras, but then again perhaps not as it would only be useful for RAW shooters.
I think considering some examples would be useful at this point. Let’s take a few of the images from my recent Introduction to Wildlife workshop:
This image could be taken care of by either spot or centerweight; I have no idea if matrix would have been accurate or not. For centerweight, you would need to ensure the central spot is over the subject area, like so:
This implies a lock-exposure-and-recompose is necessary – or, perhaps not seeing as I intended to crop the final image to a more square aspect ratio anyway. You might wonder whether the 70-30 distribution – specifically the metered portion falling on the black water – would throw things out; in this case, actually it helped. The center portion would have metered the white bird to middle gray, i.e. too dark; the outer portion metered the black water to middle gray, i.e. too light. They averaged out.
We could also have used the spot meter, in a few different ways. For location A, no compensation would be required so long as we took a bit of the dark portion and a bit of the highlight portion – i.e. enough to average out to middle gray. Location B would have required some positive exposure compensation as it is a highlight, in zone VII-VIII or so. Location C falls in zone V anyway, which is middle gray – so no exposure compensation would haven been required. In this case, I would have picked location C if using an AF lens (I wasn’t) as it’s of both the right luminance value and subject distance – alternatively, the head would have been a good choice, too.
Here’s our second example. This is a much trickier situation because of the thin rim of backlight around the bird; you don’t want to overexpose that else you’ll lose all tonal detail in the feathers.
You can see here that centerweighted metering wouldn’t work; the highlight areas – in this case, the subject from the meter’s point of view – is just too small. It would expose for the dark area and result in blown highlights. Spot metering, on the other hand, is ideal:
Location A is obviously nonsensical because although it might be the same luminance value as most of the bird, that isn’t the part we’re exposing for; using location A would result in huge overexposure. Location B is fine, and the highlight area is small enough that it wouldn’t require any exposure compensation since some of the dark background is also included – this is actually what I used – C and D are also workable options, though C might require a little negative compensation.
How about a few more examples?
Clearly, spot metering on the eye is the only way to go – all other options would have resulted in overexposure and both detail loss and an imbalance in the composition caused by the eye of the viewer not going to the intended area.
Actually, any metering option would work fine here – the scene is divided into relatively large portions of different luminance. If you used spot on the feathers, you’d have to add a bit of exposure compensation to keep things white; if you used center, you’d have to lock exposure and then recompose.
Our frame is fairly consistent in luminance, so once again, any metering method would work. However, all would require a bit of positive exposure compensation as the overall tone of the subject is light, and should be kept high-key.
Small white subject against a dark background, intense contrasts – spot meter.
Here’s one situation where matrix metering would actually work better than the other options: you have relatively even luminance across the frame, a strong colored background (making centerweight possibly inaccurate) and a fast moving subject (making spot metering impractical).
Of course, knowing which metering method to use in a given situation is quite useless unless you have things set up so that it’s easy to switch between them; otherwise, pick one and get used to the way it operates. If you can lock exposure separately from focus, then you don’t really need to use exposure compensation most of the time – the spot meter is all you need. If you can’t be bothered to do the finger dance, well, that’s why matrix was invented. Needless to say: as ever, practice is the key to mastery. MT
About The Author
Ming Thein is a photographer based in Kuala Lumpur, Malaysia. He specializes in watches, food and architecture / interiors. He is also the author of the Photography Compendium app. You can follow his excellent blog here, his photography on Flickr and his musing on Facebook. This article was originally posted here.