On 4/20/2014 6:20 PM, DZDub wrote:
On Fri, Apr 18, 2014 at 3:58 AM, Moose <olymoose@xxxxxxxxx> wrote:
...
A second effect, both of clipping and of JPEG engines/conversions that compress
the top of the histogram, is a loss of
visual detail. The variations in the red channel that help define details
of the subject all go to the same or only very
slightly different colors, and the detail disappears.
You can see the effect pretty clearly in the detail of the vein structure.
Also in the rough, grainy structure of the
underside of the orange ones. In both cases, there are still areas where
tonal detail was not recoverable, and thus they
remain without detail.
Does that answer your question?
Yes, in the sense that it is what I would have expected you to say about
it. The point in my questions was to ascertain that you do truly believe
that you are able to render an image "true to the original subject" without
having actually having seen the original subject. This seems to be what
your words mean. Am I missing something?
Would "less false" read better to you? I think I read once about a capture process, using mercury?, that captured the
actual wave forms. Short of that, all captures are less than accurate. Worse, both subject and light cannot be exactly
reproduced for comparison. Flowers and light change rapidly, mountains slowly, and the exact subject never exists
outside that moment.
Nevertheless, a clipped channel is demonstrable proof that a color is less true than it could be/have been, as I
explained in some detail in one of my later posts on reds, color spaces and browsers.
But I suspect this is mainly how you respond to other people's
images, rather than how you work with your own. We have had this
conversation before, and my recollection is that you admitted to being as
locked onto the memory of a subject as your guide as I am.
I am actually surprised at how relatively accurate my color memory seems to be. When I use a white balance reference in
difficult light, the corrected version generally agrees with my memory. This is sort of weird, as I was viewing in
unnatural light, so some sort of correction was going on in the mental process of creating the virtual image I 'see'. I
think mostly I 'remember' color based on clues in the subject and its situation. And, of course, I am doing the same
thing again when I view it later on the screen.
In my case, I
would not argue that my memory is perfect or that my rendering is perfect,
but the memory prevents me from doing some things that other people might
do if they had no experience of the original subject in its original
setting.
I should not have mentioned that the reds looked more natural as revised. Although true for my eyes and on my screen,
it's really irrelevant to my technically defensible point about clipping changing colors. Mea culpa.
These are your studies of the red and orange orchids:
<http://www.moosemystic.net/Gallery/Others/Wilcox/3159746w.htm>
< http://www.moosemystic.net/Gallery/Others/Wilcox/3159740w.htm>
Red orchid: I think your examples sufficiently demonstrate how far one can
go with what you call "recovery." However, your version of the red
orchid looks less like what I saw than mine. The red is just OK (as
a representation of the scene) in yours, but the weakening of the yellow is
not acceptable to me. (I realize you were mainly focused on "taming the
red.")
I hope I was clear that not everything could be recovered, only the parts compressed enough in the JPEG to be invisible
until stretched down. I wasn't trying to say the 'corrected' version is 'correct', only something to give an idea how it
might look without red channel clipping.
I was quite aware of the changes in the yellow. I saw it mostly as revealing more luminance/subject detail, at the cost
of lost brilliance. As you may have seen in my tulip example, it is possible to retain both, with careful capture and
post processing technique. And as you say, I wasn't there for the yellow any more than the red.
Orange orchid: My original of the orange flower already had some reduction
to match color to detail, I hoped without sacrificing the glow. Your
version is a bit lifeless, IMO. If you don't feel that, it's not
surprising. You don't have the crucial information (experience of the
original subject) to guide you, though I think you succeeded in what you
meant to demonstrate. I obviously failed to make my vision compelling to
you. It is not perfectly realized for me either. In my version, there is
already more detail visible than I actually had seen at the point of
capture, and the right flower was lit up like a Chinese lantern. While
successful to your point, your demonstration takes the actual photo in the
wrong direction if representation of the original is the goal.
Ah well, there's photography as representation and photography as recreation of impression. Even if the channels in my
eyes are blown out, a 'proper' exposure won't be. :-)
I am slow replying because I was working to a deadline for a book project. Finished it yesterday before the special,
low, prepaid voucher expired. In preparing images, I was reminded again and again of the difference. What the camera
captures and soft and hardware present to me is often different than the scene as I remember it.
The scene I've used for the cover almost knocked me out, as I came to a screeching halt and leapt out of the car with
camera in hand. I of course took a lot of shots, but light and clouds change, and the first one is the best. The result
out of ACR is perfectly nice, but rather flat. The printed one will be much nearer my vision.
So, I'm not taking issue with interpretation, only with technical aspects that may, or may not, affect the realization
of the photographer's vision.
I reworked the red one, this time using Viewer 3, which I think produces a
more accurate red (more accurate to my memory of the orchid):
http://jfwilcox.jalbum.net/Wallace%20Garden%20Center%20Orchid%20Show%202014/#_3159746vw.jpg
I think you will approve of the histogram, but I am not sure you will like
the overall effect any better. Because I made the background nearer to
black in this one using the curves tool, I edited the previous version made
with ACR to match it:
http://jfwilcox.jalbum.net/Wallace%20Garden%20Center%20Orchid%20Show%202014/#_3159746w.jpg
This is, to my eye, much better. How much of the improvement is Viewer 3 and greater care and how much is posting an
aRGB JPEG, I don't know, but I like it. Did you check the before and after color space conversion and downsizing and
after on the original?
Not too much to choose between them, but the glow and color of the new
version made with Viewer is my final preference.
Oh, new is much nicer to my eye! The detail in the yellow veins and spots is much better, yet still bright yellow. Reds
are better, too.
I studied these flowers pretty hard. (I don't mean the photos of them -- I
mean the actual flowers.) The color of these flowers is so intense that it
obscures most detail. In fact, it is surprising how much detail the sensor
captures despite that amazing intensity of color. I did do some "recovery"
adjustment -- consulting the histogram, of course, but making my
determination more by the visual impact on the monitor. I stopped when I
got the balance of detail-to-color that triggered, to the extent
possible, the feeling I had when I made the photograph.
Works for me.
Actually, I felt as though this would be the most difficult image for most
folks:
http://jfwilcox.jalbum.net/Wallace%20Garden%20Center%20Orchid%20Show%202014/#_3159756w.jpg
That little paph is a corker!
I had already viewed the gallery and enjoyed it, pretty successfully ignoring some blown red channels. when Chris'
casual comment led me back to take a closer look, I just chose a couple of thumbnails that looked likely as subjects.
My in-camera setting on the E-5 is "vivid," but that is primarily to
enhance playback on the LCD. In raw development, ACR ignores that,
'Cause Oly doesn't give them the "magic sauce" recipe. :-)
and if I use Viewer 3, I switch to "normal" in raw development. On most of the
flower shots
Yup, yup.
I tend to use a moderate amount of "vibrance" and just a dab
of saturation if vibrance doesn't get the result where I want it, but I
don't use either when the main color is red, usually. If the exposure is
right on, the reds often are the best they are going to be, to my eyes.
Occasionally, even a bit of saturation reduction improves the balance
between detail and intensity. Sometimes the highlight slider helps a bit,
but not too much. I think exposure is the most important factor, and even
lighting. In short, I doubt there is too much difference between us in
methodology,
Weeelll, yes, a fair amount. I do use the exposure sliders in ACR, (same as LR, less pretty interface) if needed, but
all later adjustment is with the, often quite different, tools in PS. For example, the Highlight/Shadow and
Brightness/Contrast tools act differently than and have more control options than the sliders in LR.
but in these cases I had something in mind that you didn't
have and really couldn't have. Your versions do look like they would print
nicely on an inkjet, which is a problem I would have to face with mine if I
wanted to print mine.
Your reworked Viewer version is the best yet.
It's a pain that I can't quite get away from Viewer 3, but it still seems
to be the best for reds. With my previous E-system bodies, ACR tended to
take the red slightly towards yellow. With the E-5 it seems to err a
little more towards blue.
Does sound a pain. As you say, correct exposure avoids many problems, but when it's over, a bit or a lot, Viewer is most
unhelpful.
Staunch Moose
Joel (let 'em bleed) W.
--
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/
|
