Reading mode: Light Dark

Monitor Colour

Go to last post
klinikl said:
That's a very good explanation. But Photoshop (and everything else
colormanaged) still has to make adjustments to reflect those
curves....
Click to expand...
That is correct, and is the reason for creating the profile. However, these adjustments are done using LUTs and matrices, and cannot be characterized by a single number "Gamma", which explains (I believe) why Eye One Display does not list the "native gamma".
klinikl said:
Dominic, as an unrelated question, which utility have you used to
produce those gamut comparisons on your pbase site? They look like
excellent illustrations.
Click to expand...
Those are screen-shots of the online tool at
http://www.iccview.de/index_eng.htm

The Microsoft Windows Color Applet has a similar gamut comparator, but I don't quite like the presentation.

Dominic
--
http://www.pbase.com/goggled/rythmic_gymnastics
 
Dominic.Chan said:
The Microsoft Windows Color Applet has a similar gamut comparator,
but I don't quite like the presentation.
Click to expand...
Thanks man, that's why I asked.

The adjustments in Photoshop are not done using videocard LUTs AFAIK. I admit I don't know anything about matrices. But obviously in Chris's case there was a problem in this pipeline somewhere.
 
klinikl said:
Dominic.Chan said:
The Microsoft Windows Color Applet has a similar gamut comparator,
but I don't quite like the presentation.
Click to expand...
Thanks man, that's why I asked.

The adjustments in Photoshop are not done using videocard LUTs
AFAIK. I admit I don't know anything about matrices. But obviously
in Chris's case there was a problem in this pipeline somewhere.
Click to expand...
While the values for the video card LUT's are stored in the monitor's profile, once the video card is loaded with them they are ignored by both color managed apps and unmanaged apps since they are automatically applied by the video card. Further, they are not really a part of the functional ICC profile but reside their for convenience in an extension tag.

The ICC profile (see http://www.color.org ) does have bidirectional LUTs than can take many forms including simplified ones giving only the gamma value. These allow matrix based profiles (most small monitor profiles) to easilly convert back and forth between colorspaces such as A98 or sRGB and the monitor.

For instance, if you convert from sRGB to the monitor's profile with a 2.2 gamma there will be a small difference in the shadow RGB triplet values since it takes a larger RGB set to represent dark gray in gamma 2.2 than it does in sRGB's modified gamma 2.2. This is about the only effect easily noticed when using Photoshop vs an unmanaged app since the video card LUT's fix the bulk of tone curve irregularities.

Things can get more complex still. Some monitors allow you to create 3D-LUTS and these can have some use when there is an interaction between the intensity of differing color channels. This is always used with printers but is generally not useful with monitors as their RGB channels are pretty independent.

marty
 
I only understood the first paragraph of technoid's post... And I can't see how it addresses the problem at hand (Native Gamma target producing artifacts in colormanaged applications). Sorry.
 
Dominic.Chan said:
klinikl said:
That's a very good explanation. But Photoshop (and everything else
colormanaged) still has to make adjustments to reflect those
curves....
Click to expand...
That is correct, and is the reason for creating the profile.
However, these adjustments are done using LUTs and matrices, and
cannot be characterized by a single number "Gamma", which explains
(I believe) why Eye One Display does not list the "native gamma".
Click to expand...
Not so. profiles that have a true gamma do not require LUTs at all. The formula is actually specified in the profile. When you use the "Profile Inspector" and click on the xTRC tags if "gamma=n.nn" comes up in the top right you have a gamma defined by an equation. OTOH, if you have "points=nnnn" then you have an LUT based tone curve. There is no problem at all storing a native gamma in a profile.

marty
Dominic.Chan said:
Dominic.Chan said:
Dominic, as an unrelated question, which utility have you used to
produce those gamut comparisons on your pbase site? They look like
excellent illustrations.
Click to expand...
Those are screen-shots of the online tool at
http://www.iccview.de/index_eng.htm

The Microsoft Windows Color Applet has a similar gamut comparator,
but I don't quite like the presentation.

Dominic
--
http://www.pbase.com/goggled/rythmic_gymnastics
Click to expand...
 
Marty - that's the kind of information we are looking for. Links and more information are appreciated. Thank you.
 
Well the first paragraph was the important one. Sorry for the techie talk.

In any case there is no problem storing a "native gamma" in a profile. What is usually meant by "native gamma" is that porint where the general graph of the video card LUTs doesn't bend upwards or downwards. The usually provides the best results in terms of banding. Other than that it doesn't really matter what gamma you use (within reason) and it should not affect the picture in any way other than perhaps somewhat worse banding. It certainly should not cause color casts and such.

One reason to use 2.2 would be to get accurate images when viewing in non color managed apps.

marty
 
technoid said:
Not so. profiles that have a true gamma do not require LUTs at all.
The formula is actually specified in the profile. When you use the
"Profile Inspector" and click on the xTRC tags if "gamma=n.nn"
comes up in the top right you have a gamma defined by an equation.
OTOH, if you have "points=nnnn" then you have an LUT based tone
curve. There is no problem at all storing a native gamma in a
profile.
Click to expand...
Marty,

It was a rather long thread, but the question being discussed is this:

The latest version of Eye-One Match (v3.6) has an option of producing monitor profiles using Native White Point and Native Gamma. On completion of the profile generation, the software reports only the native white point, but not the native gamma.

My theory is that such profiles do not have a gamma associated with them.

Perhaps you can confirm that with Profile Inspector.

Dominic
 
Dominic.Chan said:
technoid said:
Not so. profiles that have a true gamma do not require LUTs at all.
The formula is actually specified in the profile. When you use the
"Profile Inspector" and click on the xTRC tags if "gamma=n.nn"
comes up in the top right you have a gamma defined by an equation.
OTOH, if you have "points=nnnn" then you have an LUT based tone
curve. There is no problem at all storing a native gamma in a
profile.
Click to expand...
Marty,

It was a rather long thread, but the question being discussed is this:

The latest version of Eye-One Match (v3.6) has an option of
producing monitor profiles using Native White Point and Native
Gamma. On completion of the profile generation, the software
reports only the native white point, but not the native gamma.

My theory is that such profiles do not have a gamma associated
with them.
Click to expand...
I have some monitor GMB PM5 profiles targeted to native gamma. They do indeed have gamma's defined parametrically rather than by LUTs. But surprise! What I didn't know until now is that GMB software fits each of the 3 channels separately resulting in three different gammas! On reflection it makes sense but that is wild!

marty
Dominic.Chan said:
Perhaps you can confirm that with Profile Inspector.

Dominic
Click to expand...
 
Sorry Marty - are there somewhat less technical links?... I was thinking along the lines of something that can be observed using the over the counter colormanagement software.
 
klinikl said:
Isn't it supposed to be separate if you want to measure the actual
output?
Click to expand...
What's astonishing is that when you ask the program to use native gamma, it actually measures that gamma that is the best match for each of the 3 color primaries. Each gamma is then stored in the profile. Probably it doesn't display the "native gamma" since there are three of them and people would find listing the three even more confusing.

marty
 
That's what David suggested and I think he's right. I think it's a very smart explanation.

Now... Why would this cause problems with Photoshop?

Plus - how do you see the actual curves and numbers? ICC Inspector?
 
Dominic.Chan said:
technoid said:
Not so. profiles that have a true gamma do not require LUTs at all.
The formula is actually specified in the profile. When you use the
"Profile Inspector" and click on the xTRC tags if "gamma=n.nn"
comes up in the top right you have a gamma defined by an equation.
OTOH, if you have "points=nnnn" then you have an LUT based tone
curve. There is no problem at all storing a native gamma in a
profile.
Click to expand...
Marty,

It was a rather long thread, but the question being discussed is this:

The latest version of Eye-One Match (v3.6) has an option of
producing monitor profiles using Native White Point and Native
Gamma. On completion of the profile generation, the software
reports only the native white point, but not the native gamma.

My theory is that such profiles do not have a gamma associated
with them.
Click to expand...
GMB should say "use native gammas" I guess that would be begging for more questions though.

Indeed they don't have "a" native gamma. I had taken your comment that they lacked a specific gamma and hence had to use LUTs and such.
Dominic.Chan said:
Perhaps you can confirm that with Profile Inspector.

Dominic
Click to expand...
 
Dominic.Chan said:
technoid said:
Probably it doesn't display the "native gamma" since there
are three of them and people would find listing the three even more
confusing.
Click to expand...
Not only that, with parametric encoding, the "gamma" for each
colour is not just one number, but 3 or more (gamma, a, b, etc.)
Click to expand...
Indeed, this makes it possible to spec sRGB using "equations" rather than LUTs though I notice the HP sRGB profile uses LUTs. The GMB profiles I have all use simple gammas.

marty
 
klinikl said:
That's what David suggested...
Click to expand...
I meant to say Dominic...

Whatever - come to think of it you need separate gamma descriptions to describe the actual output. Why does it cause problems in colormanaged applications?

Guys can you come back to earth again?
 
klinikl said:
That's what David suggested and I think he's right. I think it's a
very smart explanation.
Click to expand...
I had wrongly assumed that the program picked the gamma closest to all three tone curves, used that and adjusted the video LUT's accordingly.
klinikl said:
Now... Why would this cause problems with Photoshop?
Click to expand...
Well, it's possible that programmers implimenting the TRC lookup in Photoshop simply used the first one they found under the assumption that each channel would be the same. An extremely bad assumption but one that would work good enough almost all the time.

The recent update to Photoshop fixed a big that was pretty blatant. When one used the marquee tool to select an area and try to do a "fill" with a color specified in LAB, the wrong color was used. It wasn't even close. Bugs happen. Maybe a programmer made the same assumption about one common gamma (or LUT curve).

marty
klinikl said:
Plus - how do you see the actual curves and numbers? ICC Inspector?
Click to expand...
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top