goldtech wrote:
> This is probably basic exposure how-to...but I was trying to take a
> series of pics w/Digital SLR and stitch them together w/software. I
> noted a problem w/exposure which I'll try to explain. I have it set to
> aperture priority/sunlight. In every pic there's some beige sidewalk
> at the bottom but each pic 's center is different - some have sky vs.
> some trees. Looking at pics - some have dark sidewalk vs. others have
> OK exposed sidewalk. Maybe camera is exposing per different center
> areas so sidewalk is not taken into account re the exposure(?)
In auto-exposure, it evaluates the scene and decides what exposure to
use each time you push the release. Exactly what it takes into
account depends on various setttings -- matrix metering for example
looks at many points across the frame and tries to figure out what's
going on, whereas center-weighted is rather simpler, and spot reads a
much smaller area. These details are partly camera-specific, and I
probably have different cameras than you do anyway.
Different cameras, and to some extent even different manufacturers,
may have different goals in the auto-exposure algorithms. Some try to
make as many pictures look really good as they can, for example, and
perhaps others try to take as few pictures as possible that can't be
rescued in post-processing (the first might be useful for snapshots,
the second for professional event and news pictures).
I, and various sources I've read, pretty consistently recommend manual
exposures for panoramas, precisely so that the exposures are uniform.
> This brings up related exposure problem I'm having as I get more into
> digital photography, with auto exposure settings. Took pic of white
> cat in center and pic was underexposed. Took flash pic of black cat w/
> white wall in immediate background - cat looks OK - maybe a bit
> underexposed - but wall was is too white/washed out.
>
> What are strategies? How do you deal with stuff like this?
The camera is not as smart as you are, is what it comes down to. It
has no idea what the image actually "looks like", just what the
brightness is at a few spots.
Here's the simple-minded model of exposure meters. It's roughly true
for older equipment; it's drastically non-true for matrix metering in
modern cameras. Okay; simple-minded meter: the most basic exposure
meter averages the whole range of what it looks at and makes it come
out middle gray on the average. (The exact density is widely given as
the classic 18% gray, but recently I've been seeing people say that
was never exactly right. So, we'll set aside the exact number; but
any given simple meter just makes the average brightness of the scene
come out to a constant value). This works fine for "average scenes".
It's badly wrong for the naked redhead (very pale skin) on the
white-sand beach, and it's badly wrong for the black cat in the coal
cellar. Those scenes should not in fact come out to any "middle"
gray.
> Please explain grey card - how to use. Would it help?
A grey card is useful for a couple of things.
One is to take incident light readings. By reading what middle grey
reflects in the lighting on your main subject, you get an exposure
that will render your subject "right". Kind of. What this doesn't
take into account is scenes with a wide brightness range, one that
exceeds the capability of your camera. Incident light readings are
easy, but they're not a really good substitute for deciding yourself
where to place the key values in your scene.
(Spelling of gray/grey differs in different regions; I think both are
now "acceptable" in the USA. I mention this since I suspect you of
being a non-native speaker of English, and I was using the other
spelling above.)
> My camera can change meter areas w/in a frame? Bracket with over and
> under shots?
Bracketing is a mechanical way to increase the chances of getting a
"good" exposure. By trying a variety of exposures centered on the
meter reading, you increase your chances of lucking out. Especially
with film, bracketing was very frequently used by serious
photographers. These days, you can take a shot at the metered
exposure, and then look at it (and, on many cameras, examine the
histogram, which gives you a LOT more reliable info than just looking
at the image), and decide if it's good enough or not, and if not,
which direction it's off and about how much. And adjust your exposure
(either in manual, or by using exposure compensation), and take
another exposure, and check it to see if it's good. Repeat as needed

.
Very occasionally you have a situation that you can't test in
advance and don't have time to mess around in the middle of, but
that's rare (one example is a rocket launch; the brightness of the
exhaust flame is key, and it's not there to measure or test until the
actual launch).
> How else? Thanks.
>
> One thing I do find is that even if histogram is skewed to the
> underexposed left I can adjust with software and get good results.
Yes, you can. Basically, if you haven't clipped significant
highlights or shadows, all the information is there, and you can
adjust to make a good-looking image in post-processing.
The less-exposed parts of the photo have more noise in them; if you
have to boost those a lot, that noise becomes more visible. Thus, you
get a moderately better-looking photo when you don't have to brighten
it in software. But it's not a drastic thing; brightening in software
is very useful.