On 3/14/2022 7:16 PM, Wayne Shumaker wrote:
At 3/14/2022 03:29 PM, Moose wrote:
On 3/14/2022 8:49 AM, Wayne Shumaker wrote:
I am wondering if anyone on the list has used Cobalt profiles? There is
discussion on FM. One of the examples at the bottom of this page:
https://www.cobalt-image.com/basic-pack/
of the red flower looks pretty dramatic.
The big change there, and in the larger one at the top of the page, is in
pulling down highlights in the red channel, to avoid clipping and loss of
detail. I don't need a profile to do that. It is useful to bracketed EV comp in
situations known to lead to that problem.
Some further discussion in this link:
https://www.cobalt-image.com/in-the-heart-of-our-basic-profiles/
I found interesting to read.
There are times when the adobe curve can create problems where you can no
longer pull it down. To quote the article:
"We are therefore in the unfortunate situation of being able to photograph a
high-contrast scene up to 14 stops, but not being able to represent it as an output
in a credible way for our senses.
This isn't a new problem. Even with vast exposure bracketing taking care of tech limitations, finding a way to map huge
DR onto real life display media is a problem. It seems to me as much or more an aesthetic problem as a tech one.
A good part of the "problem" is that our eyes work differently than a camera. When viewing a high DR subject, our eyes
adjust aperture constantly while building up the "image" in our minds. As we focus attention on any part, the eye aims
at it for best resolution and color, and adjusts aperture.
Reproducing what I "see" becomes a matter of art. I think this example dramatically illustrates the issue.
<http://www.moosemystic.net/Gallery/tech/Dynamics/DR&EV.htm>
Compressed, or even partially compressed DR simply doesn't recreate anything
close to what I saw.
I suppose deep learning AI could do a pretty good job of making human visual sense of 14 stops of DR with a high DR
subject. Topaz' next product?
The practical solution to this problem was the introduction of a tonal compression
curve."
...
Well . . yes, and no. Exposure and the limits of the hardware rule. If the exposure clips a channel, that tonal detail
is gone, squashed to one value, as in the above example.
Years ago, DPR did some comparisons of Raw converters (5D review?). It showed that ACR-PS managed to intuit/interpolate
some tonal values based on surrounding values better than anyone else. ACR/LR is still pretty amazing with some images.
Most converters simply don't try to recover those exposure errors.
So, yes, if a curve pushes Raw vales into clipping, a different curve may correct that. But no, overexposure of
highlights is not recoverable via different curve.
The big change in recent years is what DPR terms "ISO invariance". That is Raw files where lifting shadows in post gives
the same results in shadow areas as shooting at higher ISOs. That allows one to underexpose mid and low tones, to avoid
clipping highlights, and pull up the middle and bottom in post, without IQ penalty.
"There is no absolute scientific reason for choosing this specific curve, it is
an empirical choice.
I would have said aesthetic or artistic choice.
Other software manufacturers have chosen slightly different or even significantly different
curves. In fact, in Colorimetry there is no concept of “curve” in the space of
the XYZ tristimulus; all transformations should be linear. The presence of this curve makes
the representation of the photos on our monitors credible to our senses, both in terms of
exposure and contrast ratio.
There are, however, some secondary effects, typical of RGB curves, which are
undesirable:..."
Worth noting some of the problems. Pulling down the highlights may fix the
problem that was created in the first place by the profile tonal curve. Which
means one is fighting against the profile.
I have not encountered this as a problem I have to contend with.
We have often discussed different cameras and their colors. But every color
produced by the camera has a profile used to render those colors. As Jim Kasson
has said:
https://blog.kasson.com/the-last-word/roles-of-camera-and-raw-developer-in-determining-color/
in particular the color profile intent.
I assume that you are aware that Adobe, in LR and ACR, has several camera
specific profiles available, as well as many non-specific ones.
Their default profiles seem to be intended to minimize differences between cameras, pushing all
toward their idea of correct color. A few years ago, they changed defaults from "Adobe
Standard" to "Adobe Color". AGAK commented at the time, opining that the new profile
pleased him more.
If a Raw file has been opened before, in LR/ACR, with the "Standard" profile,
when it is opened again, it will still use that profile, which is also available by
choice for any file.
In any case, there is a way that using the Cobalt profiles is paying to use
profiles designed by a small outfit in preference to those done by a big one,
which has been at this profiling business much longer, and pay extra for the
privilege.
I like the Adobe default profiles. The look a lot better to me, for example,
than those in DxO.
I think adobe is trying to cover the widest audience.
An alternative is that they are trying to accommodate the widest range of cameras by adjusting them to all be similar in
colors. And yes, that standard is meant to look pleasing to as wide a rang of photographers as possible.
I also don't necessarily think the Cobalt profiles are the ultimate solution.
They even advertise the basic profiles as just a starting point. I just find it
interesting to explore and try to understand the reasoning behind it all.
Most photographers don't want to deal with it anyway. And most people don't
like realistic color rendering.
I'm not so sure about this statement. Are we back to "what the camera sees" vs.
what I see?
What is the standard for "realistic"? What chemistry or sensor creates? I'm betting that you, and virtually anyone who
views these examples, will prefer the color of the profiled scans to straight scans of one of the most "realistic" films
ofthe era. <http://www.moosemystic.net/Gallery/tech/Scan/VuesProf/>
Which is most accurate? I dunno, and I was there, pushing the shutter release. I know
what seems most realistic to me. 😁
I'm sticking to "realistic" as somewhere in the range of "what it looked like to me" through "what that sort of thing
looks like to me". Perhaps "looks like" should be "feels like"?
Perhaps if one of the Cobalt basic camera specific calibrated profile is used
for camera comparison, will there is still a significant difference when
comparing recent camera models?
I don't know. If you use the Adobe profiles, they tend to minimize differences
between cameras.
I'm not the best color judge by any means. I only notice real differences with
side by side comparisons.
As Meher Baba said, I use the Adobe profiles and "Don't worry, be happy."
As Moose says, no one else here was there, and even if they were, and had perfect visual
memory, how does one know whether yours or theirs is "correct".
Everything is going to be viewed on a range of display devices and/or prints,
with various biases of their own. Even if one color controls the whole chain,
the light under which the results are viewed changes them.
Cameras "see" differently than I do. Quite often, when I push the button, I already know what the
result is going to look like, and how I will likely modify it to mirror what I saw. St. Ansel became famous
doing this. ðŸ˜
My books look different under different natural lights, and under various sorts of
interior light. And yet, to the extent that they are of familiar sorts of things and
people, viewer's visual systems correct for all that. I've never had anyone say "all
these pictures are reddish" when viewing under incandescent.
If I were shooting for catalog repro, I'd be using Sigma Foveon bodies or Oly
bodies in HR mode, and profile like crazy, to avoid demosaicing color errors -
but I'm not . . .
Other Things Moose
I have mentioned in the past using a linear tonal curve in the profile to avoid
having to fight against the existing tonal curve.
You have. And if I haven't already mentioned it, I disagree. Not in theory, but in practice I just don't feel this is a
problem I've encountered.
And I do use camera specific profiles, which offers a choice for a starting
point.
After reading more about these profiles, it does seem one is paying for
something that may not be that much value add. You have to pay 39 euro per
camera just for the basic profile.
As you say, our eyes have a way of compensating for a lot. But I still wonder
about the comments of one camera is better than another, when it may just be
the profile.
The examples I've done are now so old . . . Yup, I've compared ACR-PS, Canon DPP, variously named Olys, RawTherapee,
SilkyPix, DxO, CaptureOne, Exposure and others. They all do make variously different colored images with different mid
points and contrast. I've always decided, so far, that I like Adobe profiles not as best of a flawed lot, but as good in
their own right.
I'm always interested in what goes on behind the curtain.
And if anyone has sleep problems you can read this - Kodak 2000:
http://www.photo-lovers.org/pdf/color/romm.pdf
which I found trying to understand what RIMM ROMM means -
"Reference Input/Output Medium Metric RGB Color Encodings"
Good luck with that. I'm saving my remaining brain cells for other things.
Synaptic 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/
|