Colour Management / Calibration / Profiles ?
Mark Spencer
photography.
My images are distinctly 'muddy' when reproduced on something other than my
monitor at home. This is immediately obvious when I print, but I'd assumed I
needed a new printer (it's an Epson Stylus Photo 895), but as I got the same
'muddy' quality when I submitted three images for projection at my local
camera club last night, I'm wondering whether it could be something to do
with Colour Calibration/Management.
When I first installed Photoshop ~6 months back, it said (something like)
the monitor's profile appears to be corrput, did I want to use it anyway? I
said no. This seemed to be born out by the fact that images I viewed in
Windows Photo Gallery and Powerpoint all had a terrible sepia tone until I
subsequently deleted the monitor profile. Vista Color Management is now
configured to use sRGB IEC61966-2.1 as default for monitor and printer.
Have I set this up wrong somehow, or should I be using a different colour
profile/s?
My graphics card is INNO3d NVIDIA GEFORCE 7300GS and monitor is SAMSUNG
SYNCMASTER 205BW 20.1" TFT-LCD. I have the latest NVIDIA graphics driver
installed. My camera is a Canon EOS 40D
Thanks for any help.
Mark
John A.
You should be able to download a profile from Samsung for that
monitor. It won't be tuned exactly to yours but it will probably be
closer to it than sRGB. Note, though, that it will likely be tuned for
the monitor's default factory settings.
Alternately you can see if the monitor settings include an sRGB mode.
If so you could switch to that and stick with the sRGB profile you're
using. It probably won't quite give you access to the full range of
colors your monitor can produce, but it's more of a known quantity.
The best option is to get a monitor profiler and produce a custom
profile to match your preferred monitor settings. Unlike the options
above, this is not free (unless you can borrow one from a friend or
fellow club member.
John A.
wrote:
Oh, and you should also get a profile for your printer. Download or
custom - same deal as the monitor. 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.
eatmorepies
"Mark Spencer" <mspenc...@hotmail.com> wrote in message
news:U7WdnQdVovID7rrW...@brightview.co.uk...
Start by using Adobe Gamma - this can create a profile for your monitor.
Your copy of Photoshop probably put Adobe Gamma onto your machine. If so,
it's in the control panel. Run and follow the instructions. If you want more
information on Adobe Gamma there are plenty of references to it on Google.
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.
If you are using Epson ink and paper then you should get reasonable prints.
They may well be darker than what you see on the screen because screens
transmit light whereas paper reflects it. I run my monitor at 50% brightness
to edit photographs and get a good brightness match with the prints.
If you are not using Epson ink/paper then you will have to get a print
profile made (or fiddle about with printer settings a lot). There are lots
of sites that carry information about the profiling process. There is
probably someone in your photoclub who can give you a potted tutorial.
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.
The topic is rather large and occupies two tutorial evenings in my photoclub
every year or so.
Regards
John
nsbm
problem.
First set the monitor to its default settings for brightness, contrast and
RGB (read the manual). I have one of these and it is a very good monitor for
its price point.
If you are shooting jpegs they are in sRGB and will have color that is as
"correct" as the camera exposure, white balance and processing algorithms
can manage. If you are shooting raw you should choose the Adobe RGB color
space, which leads to other issues.
If you have your monitor set to default settings, you shoot jpegs and you
use sRGB as the default setting in PS (or instruct PS to preserve the
embedded profile) you should be seeing reasonably accurate color unless your
monitor is defective. Non color managed images, which is everything outside
of Photoshop, should also look reasonably accurate. The generic PnP monitor
profile is not any different than any Samsung would supply, based on my
experience.
If you do not make any adjustments to your jpeg images they should project
and even print correctly. If you make adjustments in Photoshop, based on the
overly bright LCD images, problems start.
If you want accurate color and brightness levels you need to calibrate your
monitor with a reasonably sophisticated device, like the latest Spyder, but
you have to aim for a brightness level of around 100 (using both brightness
and contrast adjustments)--that is what CRTs provide and reasonably mimics
the reflectivity of glossy print paper. This avoids the needs for other
workarounds.
Most people do not seem to understand that they are working on a consumer
grade monitor that is a 6 bit device that fakes an 8 bit gamut and printing
with an 8 bit device that cannot fully reproduce 8 bit color and usually
exaggerates parts of the gamut exaggerated by the mismatch between inks and
lcds.
me
<mspenc...@hotmail.com> wrote:
>I'd appreciate some advice if anyone can help? I'm relatively new to digital
>photography.
>
>My images are distinctly 'muddy' when reproduced on something other than my
>monitor at home. This is immediately obvious when I print, but I'd assumed I
>needed a new printer (it's an Epson Stylus Photo 895), but as I got the same
>'muddy' quality when I submitted three images for projection at my local
>camera club last night, I'm wondering whether it could be something to do
>with Colour Calibration/Management.
This might be caused by images in AdobeRBG being shown/printed without
using a color manged application (properly). What color space is the
camera set to if shooting jpeg or what is the working colorspace of
PS? When printing do you let PS or the Epson driver control color
management and with what settings, including the profile you use for
the paper/ink combo?
Bernhard Agthe
sorry for pirating the thread, but I think the question fits well...
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...
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.
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.
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 ;-)
Second, where do I find the software (including the reference image)?
Something that runs on Linux would be great ;-)
Thanks!
Chris H
<dark...@gmx.net> writes
>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...
Then there is no point at all... there is plenty of hardware about that
is inexpensive AND calibrated. Monitor profiles can be downloaded from
the monitor makers, most printers come with generic profiles for the
printer makers paper, most decent independent papers come with free
generic ICC profiles for various printers.
>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.
You need colorimeter HW to calibrate the monitor. This needs to be
calibrated HW. The low cost entry level calibrators cost less than it
will cost you to make the HW.
BTW you do not calibrate "the printer" You calibrate the printer and
ink on a particular paper. Most printer manufacturers provide generic
ICC profiles for their inks and papers. NOTE these profiles are in ASCII
To do your own profile you need a test image with all the colour blocks
on them, there are plenty about you can download (try anywhere that does
the calibrator HW). Then you need spectrograph HW to read the colours
on the paper you printed. This need to be calibrated HW. Again this
will cost as much to buy as make.
SO there is little point in DIY HW particularly as generic ICC profiles
are free and will be as good if not better.
>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.
Yes usually at least 256 colours. These need to be "exact colour"
blocks.
> 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. 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 ;-)
Yes... Lots of people have done it. You can buy the HW and SW from lots
of places. The trouble is it is a LOT cheaper to buy the HW than do your
own "approximate" versions.
>Second, where do I find the software (including the reference image)?
>Something that runs on Linux would be great ;-)
The SW and images can be downloaded from most of the sites that do the
HW. There are many test images about.
What I have done is use an Eye-one profiler for the screens. Then I use
an Epson printer. This comes with generic profiles for their paper. I
use Permajet paper and their generic profiles for their paper and Epson
Inks.
If you use Permajet Inkflow system then there are not only the Generic
ICC profiles for their ink and paper on the Epson printers but they do a
FREE bespoke profile service. The difference is they use professional
"printer" profiling HW that costs thousands for a very good custom
profile.
So building my own is pointless (and as I am an embedded Engineer with
30 years experience and have lots of tools and HW at my disposal it is
something I could do. )
Look at it another way. What are your camera and lenes worth? How much
do you spend on your printer and computer? So why skimp on the one thing
you need to make sure the colours are accurate?
Something like the
http://www.pantone.com/pages/products/product.aspx?pid=79
or the
http://www.robertwhite.co.uk/product.asp?P_ID=1610&PT_ID=413
is peanuts compared to the rest of your kit.
--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills Staffs England /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/
Chris Malcolm
> Something like the
> http://www.pantone.com/pages/products/product.aspx?pid=79
> or the
> http://www.robertwhite.co.uk/product.asp?P_ID=1610&PT_ID=413
But for hobby use you can use something like the following simple
cheap suck it and see method. Find a good on-line downloadable colour
chart. Get a print made of it by someone whose printing you like and
trust for good colour accuracy. Then fiddle with your printer and
monitor profiling until you've got a reasonably good approximation to
that good print on each of them. You'll need to arrange a good white
high colour temp light on the print which is not too bright and shaded
from your monitor in order to compare print colour side by side with
monitor colour.
You will need a good colour matching eye and some experience in colour
matching to do this well. If you're old enough they may even have
taught you how to do that in school art classes in the old days before
children became too badly behaved to teach them much.
This is far from a perfect or lab quality accuracy calibration, but if
it produces results which you find it hard to find fault with, what do
you care? :-)
--
Chris Malcolm
Chris H
<c...@holyrood.ed.ac.uk> writes
>
>But for hobby use
It is pointless. To get something accurate it is still far cheaper to
buy a system. Unless you want to make an inaccurate calibrator....
>cheap suck it and see method. Find a good on-line downloadable colour
>chart.
There are dozens of those about to download.
>Get a print made of it by someone whose printing you like and
>trust for good colour accuracy. Then fiddle with your printer and
>monitor profiling until you've got a reasonably good approximation to
>that good print on each of them.
> You'll need to arrange a good white
>high colour temp light on the print which is not too bright and shaded
>from your monitor in order to compare print colour side by side with
>monitor colour.
>You will need a good colour matching eye and some experience in colour
>matching to do this well. If you're old enough they may even have
>taught you how to do that in school art classes in the old days before
>children became too badly behaved to teach them much.
>
>This is far from a perfect or lab quality accuracy calibration, but if
>it produces results which you find it hard to find fault with, what do
>you care? :-)
>
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.
Chris H
<jo...@nowhere.invalid> writes
>On Wed, 16 Dec 2009 18:05:25 +0000, Chris H <ch...@phaedsys.org>
>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.
>
>
>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?
John A.
wrote:
>In message <inaji5l4c0ksobvme...@4ax.com>, John A.
><jo...@nowhere.invalid> writes
>>On Wed, 16 Dec 2009 18:05:25 +0000, Chris H <ch...@phaedsys.org>
>>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.
Chris H
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.
John A.
wrote:
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?
Wolfgang Weisselberg
> 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
Wolfgang Weisselberg
> 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
Wolfgang Weisselberg
> 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
Wolfgang Weisselberg
> 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.
-Wolfgang
Wolfgang Weisselberg
> On Sat, 19 Dec 2009 04:23:51 +0100, Wolfgang Weisselberg
>>John A <jo...@nowhere.invalid> wrote:
> 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