John Adams wrote:
Doug Jewell wrote:

Now go to the drop down list and choose "Barbie, detail without
flash". The Nikon is at ISO 3200, the Olympus is only at ISO 800.
Despite the ISO on the Nikon being 4x the Olympus, they are about the
same as far as noise. The Nikon colouring looks better too.
Actually, I wouldn't put a lot of faith in that site. Just looked at a
few other cameras, and it appears they don't make much effort to
ensure the cameras are shooting in consistent settings. I picked a few
cameras at random, and they all had the main scene rendered at
different sizes, and the main scene seems to change quite a bit too.
If they can't maintain consistency in the scene, how are we to know if
they are using comparable camera settings? Other than the barbie
without flash, they don't say what ISO, what white balance setting,
what lens, what aperture, what shutter speed etc, they have taken the
photo with.

Well, regardless of their testing methods I still agree with their
bottom line so I don't think the smaller sensor means all that much in
the end, especially considering it is quite a bit cheaper than what
Nikon and Canon offers in this range.
A testing method without consistency, is no test at all. I
could very easily make a $99 :&S look better than a $10000
SLR, simply by having a poor combination of shutter,
aperture, ISO & white balance.
How do we know that in that first shot, the focus point was
set the same? How do we know that the softness in the Nikon
wasn't because it was shot wide open, while the Olympus was
stopped down? How do we know the softness isn't because for
whatever reason it didn't focus accurately? As much as
dpreview is biased, at least with their tests they try to
ensure consistency in camera settings.

"The E-510 is the most complete general public reflex: stabilization,
anti-dust, LCD aiming, and 10 MP. With excellent ergonomics and very
good image quality, this camera rivals and sometimes surpasses the
current leaders, the Canon 400D and Nikon D40x."
I have no doubt that most of the above is true, but what is
also true is that in some circumstances (and ISO above 200
is one), the Nikon and Canon wipe the floor with the
Olympus. The testing site that you linked to doesn't provide
enough information to truly make the comparison. There is
only one shot where they post information about the camera
settings, and that is barbie detail without flash. Here the
Nikon is tested at ISO 3200, the Olympus at ISO 800. I
noticed the K10D was at ISO 100 - that is no way to make
comparisons. In a lot of the shots, the image size is
different - how can you compare sharpness between 2 images
where the subject is at different sizes? What you say about
the E510 may be true, but that site doesn't provide proof of
the matter.