On 4/21/2014 9:00 AM, Ken Norton wrote:
I would like to steer the conversation towards "red management". There
are so many individual ways of interpreting what a proper "red" should
be, but there are several underlying issues which we ALL struggle to
get our arms around.
Exposure. Like, seriously. We have a massive disconnect with exposure
management when it comes to reds. Why is it? Why can't we use the
in-camera histogram? Why is it that we need to give reds as much as
two stops headroom from ETTR?
The LCD image is a JPEG, downsized to LCD size, so many pixels have already been sacrificed to averaging or whatever.
Then, the histograms are just too small to show a relatively small number of clipped pixels.
So, know that, and compensate for it. I know what my cameras need, and I'll bet you know what yours need. Those with
Belief in Authority who can't bring themselves to mess with the meter reading to make enough EV adjustment will continue
to have to live with blown reds or only shoot in dull light.
The human eye does not actually "see" red. It sees orange and red is
actually a derived color based on the absence of green. Camera sensors
have detectors which are more narrow-banded than the human and doesn't
do the mix-minus that the human vision does. The algorithms used to
combine the sensel information from the camera sensor is either
additive or averaging based. There are a tiny handful of converters
that are now starting to mimic the human vision in this regard, but
the bulk of this development has been done by Phase-One. Adobe is
light years behind in this subject. Even Silkypix uses mix-minus,
although, rather crudely. As a result, I've been able to use the
DMC-L1 with Silkypix to work reds very well without having to resort
to heroics of exposure latitude during capture.
But, does what you know about the L1 have usefulness for more current sensor systems? EV -1 and run it through the ACR
defaults for the GX7 works awfully well for me. Much the same for the E-M5, but maybe more likely to need -1.25 or a
touch of the Highlight slider.
Just as in television, we used to have unsafe colors, the same is
still true to a certain extent with digital still cameras. To
understand why, we need to resort to charts and bench tests.
Real-world shooting is certainly the "defacto standard", but we need
to understand what is going on behind the scenes to understand why a
camera will respond a certain way in the wild.
What happens if you have a pure red subject that falls outside of the
spectral response of the green and blue sensels? In RGBG terms, let's
say your subject is 255,0,0,0. What will happen to that? Will the
resulting converted file be red? Likely not. Let's back it off a bit
to 125,0,0,0 and see what the raw converters do with it? I can
guarantee you that you'll not end up with what you expect.
In the real world of shooting flowers, we have other factors in play.
UV/IR will certainly wreak havoc. That red flower is reflecting a ton
of UV light which does fall under the sensitivity of at least the blue
channel. Oh, and did I mention that all three color sensels see IR?
So, that flower, which you so carefully ETTR'd to give a stop of
headroom (128,0,0,0) is actually more like 128,50,10,50. Or is it?
Is the flower pure red, in the real world? This is all nice speculation. But when I can shoot a red poppy, etc. in the
yard, process it, go out and look at the original, and feel the processed image is pretty much what I see in the
original -- do all the flaws in the process mean anything?
Again, a little thing about how histograms work and why they don't
reflect what is going on at the sensor. Unless you are using a
computer program that analyses the RAW file directly, everything you
look at is post-conversion.
As above. You can't rely on the LCD histogram, for perfectly straightforward
reasons - so don't try.
One way to illustrate this is to take color contrast filters for B&W
film and place them on your camera. Carefully meter the scene and take
a sample image without the filter using manual exposure mode. Without
adjusting the exposure AT ALL, shoot it again with a Red filter. The
results will boggle the boggled.
Why? I would expect weird results, mixing filters with filters, none with sharp cutoffs, so they all overlap, with Bayer
processing. I just don't see how that plays into the practical art of capturing red flowers on RGB, Bayer array digicams.
So you don't like the process, but if the sausage is good, does it matter?
AG (just shoot film and all these headaches go away)
Hahahahahahahahahahha :-)
Gossamer Moose
--
What if the Hokey Pokey *IS* what it's all about?
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|