«

Hi,

Take a look at these shots
http://gallery129375.fotopic.net/c395008.html
(WARNING: The originals are 2+MB each)

When I was setting the camera for these shots, the indicator on the EV
scale had gone below -2EV and so was blinking. I would say the
exposures were -3EV to -4EV from the camera recommended "0" EV.

Everyone in my office who looked at the photos liked them so I am
assuming the exposure is right. The question is that if the exposure is
right then how come the camera meter was WAY off?
Canon 300D
Canon EF-S 18-55mm f/3.5-5.6

- Siddhartha

The camera metering assumes that the average illumination is 18% (or
13% depending...), so when you take an image of a large swath of dark
stuff the camera metering will try to make the dark stuff (well)
darkish (18%). This is about 3 stops from perfectly exposed.

Conversely, when you take an image of the moon at night, the metering
thinks the average illumination is the sky and will overexpose the moon
by about 3 F/stops.

White cats in a snow storm and black cats on a black bed spread have
similar problems.

The metering does not and can not understand the image, and that is
what EC is for. You, the photographer, need to understand what the
metering does (e.g. how it thinks), and what kind of circumstances you
should wrest control over the camera metering.

MitchAl...@aol.com wrote:
> The camera metering assumes that the average illumination is 18% (or
> 13% depending...), so when you take an image of a large swath of dark
> stuff the camera metering will try to make the dark stuff (well)
> darkish (18%). This is about 3 stops from perfectly exposed.
>

Got it. I assumed that as soon as I switch to manual mode the metering
goes to partial. But it doesn't work that way. One has to press the AE
button to switch from centre weighted to partial in the "M" mode.
My bad, should've read the manual more closely

- Siddhartha

Todd H. wrote:

> Also with long exposures, there's this bit about reciprocity failure
> that comes into play as I recall.

"Reciprocity failure" is the tendency of color films to shift color over
long exposures; the real problem with it is that it's inconsistant and
largely unpredictable and varies from one brand and formulation of film
to the next. It DOES NOT affect digital photos (see
http://www.pbase.com/bret/image/34152731 and note the exposure settings).

Todd H. wrote:
>
> I'd be curious if your photos would look punchier if they were
exposed
> even longer, or at higher ISO's.

I kept the aperture at f/8 coz I figured (from some previous subjective
tests) that this is where the lens would be at its sharpest. Kep the
ISO at 100 coz didn't want any noise. Hmmm ... But I could've
definitely set to BULB and taken longer exposures. Or bumped up the ISO
to 400 or 800.
Anyways, thanks. I'll keep all that in mind in future

- Siddhartha

"Siddhartha Jain" <> writes:

> Hi,
>
> Take a look at these shots
> http://gallery129375.fotopic.net/c395008.html
> (WARNING: The originals are 2+MB each)
>
> When I was setting the camera for these shots, the indicator on the EV
> scale had gone below -2EV and so was blinking. I would say the
> exposures were -3EV to -4EV from the camera recommended "0" EV.
>
> Everyone in my office who looked at the photos liked them so I am
> assuming the exposure is right. The question is that if the exposure is
> right then how come the camera meter was WAY off?
> Canon 300D
> Canon EF-S 18-55mm f/3.5-5.6

Camera meters aren't all that smart. They evaluate a scene, and more
or less want the result to be middle gray, and in digital perhaps
they're doing math to minimize the number of pure white (all 1's)
pixels. If you shoot a black cat in a coal mine, a proper exposure
will be a few stops below middle gray. If you shot a white cat in a
snow storm, your exposure will have to be a few stops over middle gray
to look accurate.

Your shots are a nighttime scene, and are very dark--and I'd agree
with your camer that they are several stops below middle gray.
Techincally, they are somewhat lacking in contrast and could stand to
punched up a bit, and I think they might benefit from having been
exposed longer. I don't see any areas that are at risk of becoming
overexposed anytime soon.

Also with long exposures, there's this bit about reciprocity failure
that comes into play as I recall. I don't know enough about the
subject to expound upon it at great length though, but what I remember
of it tells me to not expect any metering to be terribly accurate when
you get into 15 and 30 second exposures. "Bracket like hell--those
buildings aren't moving away anytime soon" is the advice I tend to
follow to get a good one.

I'd be curious if your photos would look punchier if they were exposed
even longer, or at higher ISO's.

Best Regards,
--
Todd H.
http://www.toddh.net/

In article <. com>,
Siddhartha Jain <> wrote:
>
>Everyone in my office who looked at the photos liked them so I am
>assuming the exposure is right. The question is that if the exposure is
>right then how come the camera meter was WAY off?

They're night shots, dominated by dark skies. Assuming you were using
evaluative metering, it never really stood a chance. Try partial on the
buildings, with maybe -1 EV for this sort of thing, and experiment around
there. You should find you get better results.

On Mon, 10 Jan 2005 17:05:44 +0000, Siddhartha Jain wrote:

> Hi,
>
> Take a look at these shots
> http://gallery129375.fotopic.net/c395008.html (WARNING: The originals
> are 2+MB each)
>
> When I was setting the camera for these shots, the indicator on the EV
> scale had gone below -2EV and so was blinking. I would say the exposures
> were -3EV to -4EV from the camera recommended "0" EV.
>
> Everyone in my office who looked at the photos liked them so I am
> assuming the exposure is right. The question is that if the exposure is
> right then how come the camera meter was WAY off? Canon 300D Canon EF-S
> 18-55mm f/3.5-5.6
>
They are pretty good shots - but of course beauty is in the eye of the
beholder. Shots 1 & 2 are technically underexposed but the shot number 3
is fine.

When taking shots of illuminated buildings, the best way I found is to use
a spot meter (partial metering of Digital Rebel will do). Select the area
which you want to be properly exposed. Take a reading from that area alone
and then set the exposure manually. Quite often a bit of overexposure, by
half to a stop, improves the atmosphere, specially if you are taking the
maximally illuminated area as your target (that is often a natural
tendency). Alternatively, take several spot (partial) readings from two or
three interesting areas and then use the middle one.

--

Gautam Majumdar

Please send e-mails to