M-M <nospa...@ny.more> wrote:
> In article <
rbel6a-...@ID-52418.user.berlin.de>,
> Wolfgang Weisselberg <
ozcv...@sneakemail.com> wrote:
>> > An uncropped photo is exposed for the frame. When you crop out a
>> > portion, the exposure is likely to be off.
>> You're not talking sense in either sentence.
> I'll try to rephrase it.
That won't help.
> When a camera auto-exposes a photo, it takes the entire frame into
> account. If you crop out a portion, it would not be the same exposure as
> if you zoomed in on that portion and exposed for it.
What you're saying sounds reasonable, but is wrong.
> At least _I_ understand what I'm saying. :)
I didn't say I didn't understand you ...
Let's start with "An uncropped photo is exposed for the frame."
This is obviously easily wrong if one uses the M(aster) mode
(including manual ISO) --- maybe with test shots, maybe with
guessing, maybe with a handheld meter.
This can also be argued wrong if you use a significant amount
of light from flash --- there you underexpose (the ambient
light) and set the flash (in conjunction with the aperture
and flash-object distance) to the right value. ETTL (version
I) was tied too tightly to the focus point(s) used --- it
basically ignored much of the rest of the frame. With direct
flash and a lens reporting the distance ETTL I and II would
set the flash power according to the distance and aperture and
mostly ignore the returning light of the measuring flash ---
basically, ignoring most of the frame again.
Let's go on to "When a camera auto-exposes a photo, it takes
the entire frame into account". Please look up partial and
spot metering. Only the centre is being used (smaller if
spot, larger if partial, at least with Canon), the rest of
the frame doesn't even count.
Further, even with centre weighted average mode the border
counts for very little, so framing the image such that
it's border is reduced often causes very little change ---
not enough to be relevant and often not enough to choose
a different amount of light to fall on the sensor (i.e. an
equivalent ISO/exposure/aperture triple).
Worse, with matrix the camera decides where to use and where
to ignore the frame's brightness, and in fact can detect
"it's sunshine outside" from a comparatively small part of
the frame and then may use a "sunny 16" or "sunny 8" rule, no
matter what the rest of the frame is. (Of course it can also
detect backlit subjects and then make them be exposed right,
not the larger backlit area. So a change in framing may or
may not have the same amount of light on the sensor.
So: no, the camera does *not* take the entire frame into
account in most cases.
Let's look at "If you crop out a portion, it would not be the
same exposure as if you zoomed in on that portion and exposed
for it." Now, the obvious counter example is the evenly lit
brick wall filling the frame. Zoom in and the exposure stays
the same. OK, you said "the exposure is likely to be off",
not "always off".
Now, see above: with M(aster) mode, assuming you know what
you do, and with spot and partial mode (again assuming you know
what you do) the exposure on the main subjects will be right
--- even if you crop severely and let only those subjects in.
With center weighted average mode, cutting off the borders
doesn't make an appreciable different --- you'd have to
cut much and that much needs to be either very dark or very
bright compared to the rest of the image to make a difference.
With matrix it only matters if the camera thinks the zoomed
in frame means a different subject or switches from/to backlit
subject mode or does/does not switch to "full sunlight" mode.
In short: if you don't crop lots, even with full auto exposure
settings the esposure is likely to be right for the cropped
image. If you intend to crop much and don't trust the matrix
mode, M or spot or partial mode will see you through.
In fact, some cameras have/had other metering moded: average
(not centre weighted), in which case the whole frame would be
important, or multi-spot, or they replace the 'centre
weighted average' with a 'active focus point weighted average',
on the general idea you probably want that what you focus on
to be important in the image.
-Wolfgang