The theory behind matrix metering (better named as "evaluative multipoint exposure automation") is this:

- The metering sensor system splits the scene up into some number of zones and takes an independent reading at each.

- In the camera's ROM is a library of "typical" scene dynamics factored into zones matching the array of metering sensors.

- The camera does a pattern match between the actual readings and the typical scenes in its ROM, picking an exposure based upon the total light in the scene and adjusting that to compensate for backlit or dark-backed scene types.

In average scene dynamics, the result of such exposures is virtually identical to a full screen averaging reading. In scene types with a bright subject on a dark field, the camera should reduce exposure so as not to overexpose the primary subject. In backlit situations, the camera should add exposure to prevent underexposure of the primary subject. Same goes for a standard daytime scene with a bright sky and dark ground, etc.

This is similar to the basic system that Nikon first released on the FA in the late 1970s/early 1980s. All others are derivative, with new twists added over the years like more pattern cells, distance information, color information, etc.

On the DS, "Linking AE and AF" with multipoint AF setting as well helps prioritize to the exposure evaluation system where the interesting, primary subject is by adding distance information and focus sensor info into the parameter mix.

If the reference table is good and the lookup function accurate, what you're essentially getting is 'smart' auto-correction of the metered exposure based on scene dynamics. However, since any 'smart' metering system like this is going to be wrong some percentage of time time due to both ambiguous targets and lookup function failure, EV compensation overrides and simple CW averaging and Spot meter patterns are provided to allow the photographer to take control.

Godfrey


On Dec 27, 2005, at 6:46 PM, Fred wrote:

OK, so what's the big deal with matrix metering?  How is measuring 16
exposures in 16 zones all over the frame any different from (or better
than) simple average metering of the entire screen?

I ask this as an ol' manual focus film dinosaur, used to using center-
weighted metering all the time "back in the old days", but who now has a DS and is finally getting around to reading about such things in the manual (although changing the DS from matrix to center-weighted metering was one
of the first things I did when getting it.)

Fred


Reply via email to