Archived posting to the Leica Users Group, 2001/06/22[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]
Ray Moth writes: > ... having never read a coherent explanation > of what matrix metering does, I prefer not > to use it anyway. Matrix metering (as implemented in the F5, which has the best matrix metering of any camera using this technique) uses an array of sensors to examine the scene to be photographed and compare it against a database of known lighting situations in order to determine both the best overall exposure and the appropriate exposure compensation to apply. In the F5, the image is divided into just over 1000 segments, each of which is examined in color (red, green, and blue). In addition, the lens provides the F5 with the distance of the primary subject of the photo, and the AF sensors tell the camera where the subject is in the image. All of this is used to carry out a search against a database of 30,000 or so specific exposure situations, and the nearest match is selected and used to compute exposure. Matrix metering of this kind yields appropriate automatic exposure in at least 99% of all situations, in my experience. It has only two drawbacks: (1) the matrix metering always exposes for the _most likely_ intent of the photographer, which usually means maximizing detail in all areas of the image, and (2) there is no easy way to know exactly what the selected result sought by the camera happens to be, although watching the exposure change can give some idea of what the camera has selected. For example, the camera knows enough to recognize a subject that is in shadow against strong backlight, and compensate appropriately to ensure correct exposure of the subject. However, if by chance you want the background exposure correctly instead, the camera has no way of knowing this, and so you must change exposure manually. Also, in more complex situations, it may not be obvious which of several "appropriate" exposures the camera favors. Still, these are rare situations, and overall, matrix metering is vastly more accurate than any other kind of automated exposure control. It is so accurate, in fact, that bracketing is generally a waste of time with this type of metering. Manual intervention is only very rarely required, when the camera's choice is not the same as the photographer's. For example, I recently took a photo of a building on an overcast day. I tried one exposure with the matrix metering, and one exposure with incident metering for the front of the building. The F5 chose a slightly lesser exposure than I got with my incident metering. The camera decided that some degree of detail needed to be retained in the overcast sky, and so reduced exposure (by less than a stop) to keep the sky from washing out; this darkened the building in the foreground, but only slightly, and the overall photo was extremely well exposed, with maximum detail. However, in this case, I wanted maximum detail on the building only, and I didn't care at all about the sky, so I used the exposure indicated by the incident meter, which was a bit greater than that chosen by the F5. As a result, I got slightly better detail in the building, but none at all in the sky. I used the shot made with the incident meter. And I used the incident meter because I've figured out how the F5 will respond in situations like this. After a while, you get used to this, and you know in which situations you'll have to adjust exposure (and they are exceedingly rare, rest assured). The F5 and its matrix metering will _always_ produce a usable exposure (I've never seen an exception to this), but on very rare occasions its idea of the _ideal_ exposure may not match yours. The F5 tries to provide a good general exposure that will suit any photographer's purpose to a large extent, but in doing so, obviously it also provides an exposure that is unlikely to be ideal for more specialized requirements. In my case above, the camera had no way of knowing that I wanted maximum detail in the building even if it blew out the sky--I might just as well have wanted maximum detail in the sky at the risk of letting the building get too dark. Since the camera can't read minds, it compromised, with good (but not ideal) results. The difference between what the camera decides and the "ideal" exposure for a specific intent is typically inferior to one stop, and often only 1/3 stop. And, as I've said, in 99% of cases, the exposure is right on the mark. In particular, for situations in which I can't figure out the best way to compromise exposure for a good overall result, I let the F5 do the work, as it is better at working with complex scenes than I am. > * I would never give up the R8's flash metering (F) > mode for any of the other fancy bells & whistles > the competition may have to offer. For what it's worth, the flash metering on an F5 is superlative. It has invariably produced an essentially perfect result for me. I hate flash and try never to use it, but if I know I must use flash, I take the F5 with me, as it can handle flash better than anything else. Note that, at least with a Leica M, I consider the use of flash to be "cheating," as I've always seen the M series as intended for "real" photography by ambient light only, without flash. Why else would Leica offer a Noctilux for the M?