> Shots posted by forum members extolling the virtues of SR are
> essentially meaningless. Even I can sometimes fluke shots handheld at
> exposures / focal lengths that shouldn't work.
I agree you can't read too much into one or two examples. But it's
tough to ignore the sheet number of such posts. Keep in mind that
*many* people tested the SR system on the K100D fairly extensively
practically first thing out of the box, because it as the thing they
were most curious about (just as we see a plethora of posts from new
K10D owners testing the ISO 1600 performance, since *that* is what
everyone is most curious about there). So I think it fair to assume
that many of the posts you see there are not from people posting their
one fluke shot, but from people who really sat down and did some
testing. In many cases, the post itself makes this clear. Also,
considering how quick people are to complain about, say, perceived
underexposure issues, I would say the overwhelming amount of data in
support of the SR system seems signifcant - eg, if there were something
to complain about, people *would* be complaining.
Oh, and FWIW, it is my perception that most of the underexposure
complaints come from people new to DSLR's who expect their camera to
behave like their P&S - to try to "guess" what exposure you might
aesthetically prefer, rather than carefully calculating an exposure
designed to preserve the maximum of amount of detail. In particular,
many P&S cameras will automatically expose a backlit scene for the
foreground and hence overexpose the background, or will happily ignore
and hence overexpose a bright reflection in a scene in order to render
the rest of it in a pleasing way. Some DSLR's may do the the same, but
the Pentax ones decidedly don't. Don't get me wrong - they do have an
intelligent multi-segment metering system - but it seems it chooses to
interpret this data somewhat conservatively in terms of deciding how
much highlight detail it can afford to lose.
---------------
Marc Sabatella
marc@outsideshore.com
Music, art, & educational materials
Featuring "A Jazz Improvisation Primer"
http://www.outsideshore.com/
frederick - 06 Dec 2006 21:53 GMT
>> Shots posted by forum members extolling the virtues of SR are
>> essentially meaningless. Even I can sometimes fluke shots handheld at
[quoted text clipped - 14 lines]
> support of the SR system seems signifcant - eg, if there were something
> to complain about, people *would* be complaining.
Okay - if that is correct (and it could be - I don't know), then what
does Phil Askey have to say about it?
If there was a specific problem with the camera that he tested (which if
your argument is correct, could be possible), then I expect that he'd
retest or point out that a faulty camera may have been involved in the
result. Perhaps he'll retest when he tests the K10d?
As far as iso 1600 noise goes, as usual mostly people compare jpegs, and
argue about how much detail is lost by in camera NR etc. Shooting raw
is the only way to go IMO when pushing things - whether it's colour
balance, tough exposure conditions, or situations where high ISO is
needed. It's so much better to apply NR in PP - when you know what
you're doing with the image (printing, to what size etc) than letting
the camera do it, or even worse, farting around with buttons to set
options for different in-camera NR settings when you should be taking
photos.
> Oh, and FWIW, it is my perception that most of the underexposure
> complaints come from people new to DSLR's who expect their camera to
[quoted text clipped - 9 lines]
> interpret this data somewhat conservatively in terms of deciding how
> much highlight detail it can afford to lose.
Sure. I hadn't seen the underexposure complaints about the K10d. I use
a D70 - and the similar complaints about underexposure also seem endless
and misinformed. Nikon apparently reacted to this with the D80, or
perhaps just figured that there would be less whining from users who
knew to set EV when needed, and just make it work like a P&S camera for
the rest who probably shoot jpegs all the time. Probably another reason
to go the way that they did was also from all the whining and forum
comparisons about noise - from people who put far too much emphasis on
this because pixel peeping it is far more visible than it ever is in
print, so seems like a much bigger deal than it is. Lightening a
slightly underexposed shot is the best way to accentuate noise. A
camera that slightly overexposes (compared with what I prefer) is going
to be thought less noisy. Just my opinion, but a little noise detracts
far less than blown highlights in a print.