Shawn Wright wrote:
> I was talking over lunch with our Photo instructor yesterday, who is a
> longtime N*k*n user. He told me one of our students came back with a new F5
> body! He also said he wished that he had made manual cameras mandatory, as he
> spends too much time trying to help students with their wonderbricks. I
> mentioned I'd heard even the F5 didn't have a quick simple, yet sophisticated
> multi spot system like the OM-4's... his answer was maybe it's not needed
> since
> the matrix metering is so good.
One of the claims made for the F5 is that, with experience, the
operator can learn what exposure the matrix metering is choosing. But
it uses a neural network to map the focus distance and exposure reading
matrix (erroneously called "3D" metering by Nikon) to an overall
exposure, and one of the well-known limitations of neural networks is
that they *can't* provide a justification for the answers they give.
The F5 has a training database of 32,000 , so is a fairly complex
machine learning application, and well beyond the point at which
humans can guess the answer.
In other words, when the F5's metering fails, it does so without
warning or pattern in a way which cannot be predicted or even
rationalized afterwards. But it may just get you that shot of
a princess bleeding to death beside her boyfriend's body. There
has been a TV series "Models" running here, made by David Bailey.
In it, one of the fashion editors comments that modern cameras
are so automated that there are now quite a few photographers
working professionally who lack basic technique. IMO, also morals.
As for multi-spot metering, I think students would be better off
learning to (spot) meter a mid-tone and choose a tone (Zone) or
meter a known tone and allot it to a tone. Both the OM-4 and the F5 can
be used for these more classical methods, though probably both are
too complex for students.
< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
|