evaluative metering vs spot-metering

The three metering modes offered on modern cameras are,
- center-weighted,
- evaluative / matrix metering,
- spot-metering.

In a broad sense, the choice of metering is often linked to which exposure metering mode you’re using. Essentially, the sophistication of matrix metering is ideal for when you’re relying on an automatic metering mode, whereas spot-metering is a metering mode which is used with more thought and specific intent - which most often implies manual exposure metering. (That said, I use manual metering mode nearly exlcusively, and switch between any of the three metering choices dependent on the situation.)

With matrix metering, the picture area is divided into different sections, which are then metered with individual cells, and the exposure values evaluated and averaged out according to algorithms which are based on information gleaned from thousands of typical photo scenarios.

A simplistic example: if the top sections of the picture area (when viewed horisontally) are all much lighter than the bottom sections, then the matrix algorithm will ‘assume’ that the scene being photographed might be a landscape with lots of over-bright sky - and then would reduce the importance of the over-bright areas in the final exposure calculated. In other words, the matrix metering algorithm would then bias the exposure toward the bottom sections.

Yet, as sophisticated as matrix metering may be, it is still only a ‘guess’, and there are certain situations where the photographer will be able to make a better exposure than the camera’s automatic metering would.

As an example, let’s look at the photograph of the two men standing in the strip of sunlight again. With matrix metering, even though various sections of the scene would have been evaluated, the entire scene would have been taken into account in assessing the exposure - and the dark areas would still have affected the eventual exposure, and hence thrown the exposure settings off from what would have been ideal. A much more specific and consistent way to measure for a scene like that, would’ve been to take a selective meter reading - either with spot-metering off an appropriate area - or as was the case here, metering off another area which was similarly lit as the section I wanted to meter for.

Some magazine articles suggest that the photograher should use the spot-meter to measure off various areas and average it out to find the proper exposure. This may sound like a good idea, but is most likely unworkable in a real situation, requiring mental calculation by the photographer. Once again, a much better method in my opinion, would be to simply meter off an appropriate area as explained in the main Exposure Techniques page.

The choice between whether to use center-weighted or matrix metering is a more personal one. Some photographers claim to get more consistency with center-weighted metering because it is more predictable than matrix metering where the camera uses algorithms that we as photographers have no real insight to. I have found instances where I do prefer center-wieghted metering over matrix metering, for example: when I am photographing a bride and groom, the vertical or horisontal framing positions the white dress or black tux differently in the frame, and my TTL flash is very inconsistent with matrix metering because of that. In cases like that, I switch over the center-weighted metering, and use exposure compensation to reduce or increase the flash a bit - but my TTL flash exposures are more predictable then.

Other than if you have a specific reason to use center-weighted metering, you’re probably better off generally using matrix metering in favour of center-weighted metering. So that really leaves a choice between two modes - when to use matrix metering and when to use spot-metering.

Matrix metering is a much better approximation than center-weighted metering, and will give great results, but thinking about what you want, will enable you to overcome a lot of situations where even matrix metering will be fooled.

No comments: