Colour Management / Calibration / Profiles ?

In message , John A.
writes
On Wed, 16 Dec 2009 18:05:25 +0000, Chris H
wrote:
[...]
SO you have a sort of approximate calibration..... how is that better
than using the generic printer/paper profiles and adjusting your screen
to suite?

Or for a few pounds you have an accurate one. Actually all you need to
do is borrow a screen calibrator. Then use the generic printer profiles
which will do a job better than the method you are suggesting.

The OP is suggesting making his own colorimeter and spectrometer. Again
pointless.

Unless, of course, it *is* the point of the whole exercise for the OP.

I'm sure most of us could just go out and buy prints of the sort of
subjects we photograph for far less than the cost of our kit. But
that's not the point, is it?


What the OP is suggesting is spending a lot of time and money to do a
calibration system that is not accurate when more accurate methods are
available for less.

What is the point of spending a lot of time and money on camera
equipment and then spending more time and money to screw up the colour
calibration?
--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills Staffs England /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

On Thu, 17 Dec 2009 09:28:23 +0000, Chris H
wrote:

In message , John A.
writes
On Wed, 16 Dec 2009 18:05:25 +0000, Chris H
wrote:
[...]
SO you have a sort of approximate calibration..... how is that better
than using the generic printer/paper profiles and adjusting your screen
to suite?

Or for a few pounds you have an accurate one. Actually all you need to
do is borrow a screen calibrator. Then use the generic printer profiles
which will do a job better than the method you are suggesting.

The OP is suggesting making his own colorimeter and spectrometer. Again
pointless.

Unless, of course, it *is* the point of the whole exercise for the OP.

I'm sure most of us could just go out and buy prints of the sort of
subjects we photograph for far less than the cost of our kit. But
that's not the point, is it?


What the OP is suggesting is spending a lot of time and money to do a
calibration system that is not accurate when more accurate methods are
available for less.

What is the point of spending a lot of time and money on camera
equipment and then spending more time and money to screw up the colour
calibration?

If everyone thought like that there wouldn't be any nifty new cameras
or color profilers to buy.

It reminds me of a guy I ran into once in an employment agency. When
told I wanted to be a programmer he said "what would anyone need
programmers for when they can just buy software off the shelf?"

Anyway, it's good to see the spirit of learning & exploration is alive
in someone. It's such folk who propel us forward.

In message , John A.
writes
On Thu, 17 Dec 2009 09:28:23 +0000, Chris H
wrote:

In message , John A.
writes
On Wed, 16 Dec 2009 18:05:25 +0000, Chris H
wrote:
[...]
SO you have a sort of approximate calibration..... how is that better
than using the generic printer/paper profiles and adjusting your screen
to suite?

Or for a few pounds you have an accurate one. Actually all you need to
do is borrow a screen calibrator. Then use the generic printer profiles
which will do a job better than the method you are suggesting.

The OP is suggesting making his own colorimeter and spectrometer. Again
pointless.

Unless, of course, it *is* the point of the whole exercise for the OP.

I'm sure most of us could just go out and buy prints of the sort of
subjects we photograph for far less than the cost of our kit. But
that's not the point, is it?


What the OP is suggesting is spending a lot of time and money to do a
calibration system that is not accurate when more accurate methods are
available for less.

What is the point of spending a lot of time and money on camera
equipment and then spending more time and money to screw up the colour
calibration?

If everyone thought like that there wouldn't be any nifty new cameras
or color profilers to buy.

Of course there would. As it happens my business is the development of
new technology like that.

However the poster was looking at something that was not "too accurate"
and therefore pointless.

If he was looking at something better, faster, easier than we have now I
can see a point but to cobble together something slower, more expensive
and less accurate than the systems currently available (and the free
generic profiles etc) there is no point.

A lot of effort for something worse.

It reminds me of a guy I ran into once in an employment agency. When
told I wanted to be a programmer he said "what would anyone need
programmers for when they can just buy software off the shelf?"

That is some one who does not understand the problem

Anyway, it's good to see the spirit of learning & exploration is alive
in someone. It's such folk who propel us forward.

Thanks I work hard at it. So far managed a few useful innovations.

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills Staffs England /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

On Thu, 17 Dec 2009 15:22:32 +0000, Chris H
wrote:

In message , John A.
writes
On Thu, 17 Dec 2009 09:28:23 +0000, Chris H
wrote:

In message , John A.
writes
On Wed, 16 Dec 2009 18:05:25 +0000, Chris H
wrote:
[...]
SO you have a sort of approximate calibration..... how is that better
than using the generic printer/paper profiles and adjusting your screen
to suite?

Or for a few pounds you have an accurate one. Actually all you need to
do is borrow a screen calibrator. Then use the generic printer profiles
which will do a job better than the method you are suggesting.

The OP is suggesting making his own colorimeter and spectrometer. Again
pointless.

Unless, of course, it *is* the point of the whole exercise for the OP.

I'm sure most of us could just go out and buy prints of the sort of
subjects we photograph for far less than the cost of our kit. But
that's not the point, is it?


What the OP is suggesting is spending a lot of time and money to do a
calibration system that is not accurate when more accurate methods are
available for less.

What is the point of spending a lot of time and money on camera
equipment and then spending more time and money to screw up the colour
calibration?

If everyone thought like that there wouldn't be any nifty new cameras
or color profilers to buy.

Of course there would. As it happens my business is the development of
new technology like that.

But how? Where would you get your employees?

However the poster was looking at something that was not "too accurate"
and therefore pointless.

If he was looking at something better, faster, easier than we have now I
can see a point but to cobble together something slower, more expensive
and less accurate than the systems currently available (and the free
generic profiles etc) there is no point.

A lot of effort for something worse.

For learning.

It doesn't only happen in school.

It reminds me of a guy I ran into once in an employment agency. When
told I wanted to be a programmer he said "what would anyone need
programmers for when they can just buy software off the shelf?"

That is some one who does not understand the problem

Exactly.

Anyway, it's good to see the spirit of learning & exploration is alive
in someone. It's such folk who propel us forward.

Thanks I work hard at it. So far managed a few useful innovations.

Good for you! And how did you become able to do such things?

eatmorepies wrote:

Much better is to get someone in your photoclub to lend you their monitor
profiling hardware (e.g. a Spyder 3). This will generate a much more
accurate colour profile.

Seconded.
Personally, I'd choose profiling hardware to fit the software
you have (or can obtain), so you can actually make use of the
hardware. Unfortunately, the software bundled with the hardware
usually has license restrictions, so your friend may not lend
it to you (legally, at least). Fortunately, there is argyll
(http://argyllcms.com/) which understands quite a number of
hardware (though not the spyder *3*) and is not restricted to just
Windows and/or MacOS --- and for those who don't like to dig deep
into the stuff and understand what's going on, several GUIs exist.

If you are using Epson ink and paper then you should get reasonable prints.

Assuming his monitor is displaying properly, yes.

It is also best to set your camera and Photoshop to the Adobe 1998 colour
space - that way you have matched colour spaces and a bigger range of
colours.

Disagreed. If your monitor supports the Adobe 1998 (aRGB)
colourspace (in the sense that it can display those colours in
AdobeRGB which are not inside sRGB) you *know* it because you
*paid* for it through the nose.
His Syncmaster doesn't.

Anyway, the camera's native colourspace of a Canon 20D (and
thus most likely also of a 40D) is quite a lot bigger than sRGB
or aRGB, so you're better off using RAW and doing all colour
shifting changes within the RAW converter (or using a really
huge colourspace and programs and devices that support it).

His printer doesn't do aRGB, either (and it's not a CMYK device
in that the driver is fed RGB, not CMYK). Most photo labs don't
even understand aRGB (and mangle colours, assuming it's sRGB),
and many that do at least grasp that it's aRGB, not sRGB just
convert it to sRGB.
So there's little, if any, plus in shooting aRGB JPEGs versus
sRGB JPEGs and none versus shooting RAW and developing in whatever
colourspace you find you need and can effectively use.

Third: If you just stay with sRGB as your output format,
(i.e. use RAW, develop, but save the result as sRGB), most
any image viewer and browser will manage to display the image
correct --- this is not true for aRGB, which can result in
muddy colours when an image viewer doesn't get it's not sRGB
it's displaying.

So unless you *know* you have advantages by aRGB (or the much
much larger and denser 16bit ProPhoto colourspace, for example)
and *understand* the drawbacks, don't.

-Wolfgang

John A wrote:

I'm not familiar with printer
profilers, though, so you'll have to research that or depend on the
good graces of another poster or fellow club member. I figure they're
not the same as monitor profilers, though, given that paper doesn't
produce its own light.

There are some units that can do monitor, paper and a few even
transmission, but many do only one thing.

Basically, for paper you have spot readers (you have to point the
instrument to each spot seperately, by hand), strip readers which
either pull in paper strips by themselves or are slid over them on
a guide by hand and finally x/y tables where you tell the program
where 3 patches are and the rest of the page is run automatically.
Oh, and of course, you could use a profiled camera or scanner as
a poor man's solution --- that would be option 4.

Take, for example, the i1 pro / eye one pro (Gretag MacBeth,
now bought by X-Rite). No, not the display or monitor one,
the more expensive unit. (The cheapest paper reader is Photo
Pro LT or something like that, the differences then are mostly
the software and the revision (Rev. D, samples twice as fast and
reads ambient light)). That unit can do both monitor and paper
(strip mode).

Or take the (discontinued) Spectrolino (also Gretag MacBeth).
It does monitor, paper and transmissive, but comes only with a
(well thought out) spot reader 'clip-on'. But it also attaches
to the SpectroScan x/y table (also to the SpectroScanT, which
adds a lamp for transmittive measurements), turning it into a
quite automatic solution for the home vendor.

(Both units also work with the open source Argyll CMS
(http://argyllcms.com), so e.g. the limited software of cheaper i1
pro versions may not be as large a drawback as one might assume.)

Now look at the prices for these beauties (and most other
reflective (paper) spectrometers and weep. Little wonder there
are companies that will do the profiling for your printer, you
just have to send them a printed out sheet or two and from ~20
EUR upwards...

-Wolfgang

Bernhard Agthe wrote:

For quite a while I've been thinking about some kind of home-brew color
profiling for hobby use. I don't expect to be ultimately accurate,
rather I want to be just better than nothing...

Someone already tried, it seems:
http://www.homecinema-fr.com/colorimetre/index_en.php

My question is, whether some tool exists that allows me to manually
profile my monitor(s) and printer by using a sample card I have printed
by the photo studio I use most.

Argyll CMS would allow that from what I know, if you add your
camera (or scanner), have some real colour reference to profile
your camera and live with the inherent restrictions.

Basically an image file that I have printed at the studio which I could
use to adjust my monitor against. This would include patches of white,
black and gray and patches of different colors, I suppose. I would
require a GUI to display something on my screen and some sliders to
adjust what is being displayed until it resembles the printed version.

That wouldn't be impossible, just very inconvenient and hard
to repeat properly

Finally I would get a profile I could use within my raw converter,
whatever...

First, is this possible? (I know it won't be too exact, but closer than
nothing ;-)

http://www.cipho.de/img/dqt/DQ-Tool_Kontrast.jpg
The image contains 6 'Q's.
Set ~6.000K.
Contrast to maximum.
Lower brightness so you just just see the dark 'Q's.

Print the following image at 13x18cm (5x7 inch) @ 300 ppi (at
your preferred printing lab!):
http://www.cipho.de/img/dqt/DQ-Tool_Print_13x18cm.jpg
View on screen, next to your print:
http://www.cipho.de/img/dqt/DQ-Tool_Monitor.jpg
See that you see 3 slightly brighter squares in the black
rectangle on the bottom left, and 3 slighty darker ones in the
white rectangle on the bottom right.

Adjust your monitor, only adjust via the graphics card if the
monitor doesn't have the adjustability needed.

Choose one (or more) of the images to tune your monitor to natural
looking images, portrait (skin and gray balance), neon colours,
and/or grays.

Try to keep the grayscale huefree and the hue-bar (marked are pure
magenta, red, yellow, green, cyan. blue) looking, well, proper.

Now your monitor image looks (almost) like your printing lab.

Second, where do I find the software (including the reference image)?

Sorry, no such software, but ...

Something that runs on Linux would be great ;-)

Argyll CMS runs on Linux (and Windows and MacOS and probably
others), but that's the command line solution using colourimeters.
If you build your own, you could probably use that for a basis.

-Wolfgang

John A wrote:
On Sat, 19 Dec 2009 04:23:51 +0100, Wolfgang Weisselberg
John A wrote:

One drawback many don't understand is that if you use a larger gamut
when a smaller one will contain all the colors (that you care about)
in an image, you've effectively reduced the color depth (bits) of the
image.

Only matters if you use 8 bits, with 16 you have more than
enough reserves.

Depends on how much manipulation you plan to do.

For any limited number of bits there is always a manipulation
that will cause them to collapse to one value. However, that
manipulation isn't necessarily sensible or useful.

-Wolfgang