Originally posted by Ryan Trevisol Agreed. The detail was preserved where it counted; the shot was more successful by making the metal darker and more detailed, but was surprising compared to how it actually looked to the eye.
And this is going to be pretty common in high-contrast situations like this. The eye is amazingly good at opening and closing its iris as it darts from the light to the shadow area and the brain then conspires to produce the illusion that there is not such a great difference. So when we see an *accurate* rendition in a static photograph where our eyes and brain can't play those tricks, we're often surprised. Artists use tricks of their own to outsmart their eyes to avoiding being fooled by this - the most natural beginner mistake in the world is to underestimate the difference between light and shadow when working from life (as opposed to working from photographs).
BTW, I meant to mention before - looks like the maker notes have been stripped from the "shiny metal" photo, but I am curious about the setting of the "link AE and AF points" or whatever its called. If this is option is set, then the camera will bias the exposure toward the part of the image it has focused on in this case, the shiny metal. I'm guessing that option was off, though, because otherwise the shiny metal probably would have been rendered darker (closer to the middle of the histogram, rather than half a stop from the right).
I'll also add that I switched from matrix/pattern/multi-segment to center-weighted metering years ago. The former mode- the default - does try "a little" to guess what I might want, but does a pretty poor job of it. And it's guessing is such that I don't find it easy to predict what it's going to guess. With center-weighted metering, it makes no attempt to guess a thing - which makes it much easier for me to predict how it will behave. The end result is that I have to apply compensation at least as often, but it's way easier for me to predict even before I take the shot what kind of compensation I'll need. And the "center" is actually tight enough that I can use this a kind of a spot meter, allowing me to meter the light and shadow separately and remove pretty much any doubt. And better than a true spot meter, I'll get an "average" of a bunch of values in the light (and the same in the shadow), meaning I don't even have to find something "average" to point at.
Quote: What you can't count on with Pentax Cameras in general is them bending ISO rules to make pleasing pictures without the photographer giving it much thought.
Now, to be fair, ISO doesn't actually care how a multisegment metering system sifts through the data to come up with an exposure in the case of scenes with high dynamic range - it really only cares about how a flat surface (eg, a gray card) would be metered. And some multi-segment meters will indeed deliberately overexpose the sky in order to render a backlit subject well, and ISO doesn't forbid this. But the issue is that people coming from an SLR background (as opposed to a P&S background) won't be expecting this, so they'll tend to naturally expect to need exposure compensation in these settings. So basically, you can either please the folks accustomed to P&S cameras, or the folks accustomed to SLR's, but it's tough to please both at once, because they really do have different sets of expectations. Obviously, Pentax could provide multiple modes - one to please people with SLR backgrounds, one to please people with P&S backgrounds. but you're never going to nail it completely. At some point it still becomes a matter of learning why the camera does what it does, and then anticipating this so you know how to meter effectively given how the camera actually works.
BTW, something else a lot of P&S cameras do is play tricks with the exposures curves when producing their JPEG files from the raw sensor data. A P&S camera probably wouldn't have metered the "shiny metal" picture any different (at most half a stop brighter), but it might well have decided to artificially boost the shadows to make them not so dark, creating the impression of a brighter picture. The K-x and some other newer DSLR's do have modes for this - "shadow compensation" or something like that. But for the most part, this is the kind of stuff one can do better oneself in PP, particularly if one shoots RAW, so you and not the camera is in control of how much brightening the shadows get, how dark the shadow has to be before it gets brightened, etc.