Camera Color Calibration

I am a new lcms user. I want to calibrate the color of a
camera for use 
in image processing. If I understand correctly, I should
develop an ICC 
profile for the camera and then transform images into a
standard space 
for color processing. To do so, I would like to have an
"identity" 
output profile so that I can run the algorithms in the
standard 
(absolute calorimetric?) space instead of in the space of
some output 
device. I understand how to create the basic profile, but
how do I 
create an RGB->RGB identity profile for output?

Thanks,

-- 
Michael Shneier
Intelligent Systems Division
National Institute of Standards and Technology
Michael.Shneiernist.gov
Tel: (301) 975-3421
Fax: (301) 990-9688



------------------------------------------------------------
-------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and
a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Lcms-user mailing list
Lcms-userlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user

Michael,

often a software does the profile creation with the
transformed data you 
obtained from taking shots with the camera of measured
patches. For 
instance with a IT8 test chart. But possibly you want more
than this.

The resulting profile can be used together with a standard
RGB profile to 
form the chain for transforming your cameras image colours
to the standard 
RGB space. Lcms is right for this step including the final
conversion.

Chapter 2. Step-by-step Example: in the
lcms-1.xx/doc/TUTORIAL.TXT file 
will be a good starting point for learning about lcms' C
programming 
interface.

The absolute colorimetric rendering intent is just one way
to influence 
the conversion. It forms in no way a color space.

kind regards
Kai-Uwe Behrmann
--
developing for colour management 
www.behrmann.name + www.oyranos.org + www.cinepaint.org

------------------------------------------------------------
-------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and
a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Lcms-user mailing list
Lcms-userlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user

Michael,

I believe you are mixing some things. I'm also a beginner to
CMSs and  
I am currently reading Real World Color Management. Is is
really a  
great book and it cleared many questions I had.

For what you want to do you need 2 profiles:
one is the input profile - you get that picturing a target
and  
running that through a profiler (take a look at Lprof for
that, its  
open source and is said to deliver very good profiles), the
other is  
any standart RGB profile as you said (that would be
ProPhoto, sRGB,  
AdobeRGB,...). Profiles for standard RGB color spaces are
shipped  
with operating systems, apps, or downloadable on the web -
so you do  
not have to care about these, just have to find them

Once you have the profiles you would create ONE transform
out of TWO  
profiles: LCMS does exactly that. At last you would then
apply the  
transform.

Things can get a lot more complicated as you look at which
image  
manipulation works best in which colorspace... and which is
then the  
most appropriate colorspace for your application.

HTH
Mark

On 18.10.2007, at 20:00, Michael Shneier wrote:

> I am a new lcms user. I want to calibrate the color of
a camera for  
> use
> in image processing. If I understand correctly, I
should develop an  
> ICC
> profile for the camera and then transform images into a
standard space
> for color processing. To do so, I would like to have an
"identity"
> output profile so that I can run the algorithms in the
standard
> (absolute calorimetric?) space instead of in the space
of some output
> device. I understand how to create the basic profile,
but how do I
> create an RGB->RGB identity profile for output?
>
> Thanks,
>
> -- 
> Michael Shneier
> Intelligent Systems Division
> National Institute of Standards and Technology
> Michael.Shneiernist.gov
> Tel: (301) 975-3421
> Fax: (301) 990-9688
>
>
>
>
------------------------------------------------------------
---------- 
> ---
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems? 
Stop.
> Now Search log events and configuration files using
AJAX and a  
> browser.
> Download your FREE copy of Splunk now >> http://get.splunk.com/
> _______________________________________________
> Lcms-user mailing list
> Lcms-userlists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/lcms-user



------------------------------------------------------------
-------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and
a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Lcms-user mailing list
Lcms-userlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user

On Thursday 18 October 2007 12:59:37 Mark wrote:
> Michael,
>
> I believe you are mixing some things. 

One of the things being mixed up is the difference between
calbration and 
characterization.  It is a common misconception that by
creating a profile 
for a device that you are "calibrating" the
device.  A profile is at it's 
root a detailed description, or characterization, of the
devices color and 
tonal charaterisics.  Calibration is the process of making
adjustments to the 
device to make it behave in a predefined way.   When you
profile a device you 
are characterizing the device not calibrating it.

> I'm also a beginner to CMSs and 
> I am currently reading Real World Color Management. Is
is really a
> great book and it cleared many questions I had.
>
> For what you want to do you need 2 profiles:
> one is the input profile - you get that picturing a
target and
> running that through a profiler (take a look at Lprof
for that, its
> open source and is said to deliver very good profiles),
the other is
> any standart RGB profile as you said (that would be
ProPhoto, sRGB,
> AdobeRGB,...). 

Most photo editing software will have the functionality to
assign profiles to 
images and to convert images between profiles built to the
software.   For 
exmple, Photoshop and recent versions of GIMP have this
functionality.  All 
you have to do specify what profiles you want to use.


> Profiles for standard RGB color spaces are shipped 
> with operating systems, apps, or downloadable on the
web - so you do
> not have to care about these, just have to find
them
>
> Once you have the profiles you would create ONE
transform out of TWO
> profiles: LCMS does exactly that. At last you would
then apply the
> transform.

Again your photo editing software should do this for you and
you only need to 
worry about the details of how this happens if you are
writing software that 
needs to use this functionality. 

>
> Things can get a lot more complicated as you look at
which image
> manipulation works best in which colorspace... and
which is then the
> most appropriate colorspace for your application.
>
> HTH
> Mark
>
> On 18.10.2007, at 20:00, Michael Shneier wrote:
> > I am a new lcms user. I want to calibrate the
color of a camera for
> > use
> > in image processing. If I understand correctly, I
should develop an
> > ICC
> > profile for the camera and then transform images
into a standard space
> > for color processing. To do so, I would like to
have an "identity"
> > output profile so that I can run the algorithms in
the standard
> > (absolute calorimetric?) space instead of in the
space of some output
> > device. I understand how to create the basic
profile, but how do I
> > create an RGB->RGB identity profile for
output?
> >
> > Thanks,
> >
> > --
> > Michael Shneier
> > Intelligent Systems Division
> > National Institute of Standards and Technology
> > Michael.Shneiernist.gov
> > Tel: (301) 975-3421
> > Fax: (301) 990-9688
> >
> >
> >
> >
------------------------------------------------------------
----------
> > ---
> > This SF.net email is sponsored by: Splunk Inc.
> > Still grepping through log files to find problems?
 Stop.
> > Now Search log events and configuration files
using AJAX and a
> > browser.
> > Download your FREE copy of Splunk now >> http://get.splunk.com/
> > _______________________________________________
> > Lcms-user mailing list
> > Lcms-userlists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/lcms-user

>
>
------------------------------------------------------------
-------------
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems? 
Stop.
> Now Search log events and configuration files using
AJAX and a browser.
> Download your FREE copy of Splunk now >> http://get.splunk.com/
> _______________________________________________
> Lcms-user mailing list
> Lcms-userlists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/lcms-user




------------------------------------------------------------
-------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and
a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Lcms-user mailing list
Lcms-userlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user

On Thu, 18 Oct 2007, Kai-Uwe Behrmann wrote:

> often a software does the profile creation with the
transformed data you
> obtained from taking shots with the camera of measured
patches. For
> instance with a IT8 test chart. But possibly you want
more than this.

A fundamental assumption here is that camera operation is
repeatable 
and not very sensitive to the environment.  Typical consumer
cameras 
are not very repeatable and are very sensitive to the
environment 
because they always try to adjust for the best looking
high-contrast 
image.  So the camera makes many automatic adjustments based
on what 
it sees.  Given the same scene, just removing the test chart
or 
changing the camera view slightly is likely to cause a
change in 
behavior.

Scientific cameras, or professional cameras with an entirely
manual 
mode, are likely much more repeatable, at the expense of far
more work 
on the part of the user to get a good shot.

Bob
======================================
Bob Friesenhahn
bfriesensimple.dallas.tx.us, http://w
ww.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMag
ick.org/


------------------------------------------------------------
-------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and
a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Lcms-user mailing list
Lcms-userlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user

Hello Bob,

> A fundamental assumption here is that camera operation
is repeatable
> and not very sensitive to the environment.  Typical
consumer cameras
> are not very repeatable and are very sensitive to the
environment
> because they always try to adjust for the best looking
high-contrast
> image.
...
> Scientific cameras, or professional cameras with an
entirely manual
> mode, are likely much more repeatable, at the expense
of far more work
> on the part of the user to get a good shot.

since you mentioned it.... I have a industrial camera I'm
profiling.  
The camera has an auto shutter mode, that in most cases does
really  
enhance the image.

Into what category would an auto shutter (no additional
manipulation)  
fall:
a) breaks all profiling efforts
b) can be used together with profiling

I've asked myself this question some time ago, but I really
can't say  
yet...

Mark

------------------------------------------------------------
-------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and
a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Lcms-user mailing list
Lcms-userlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user

Mark <devcine-scan.com> writes:

> Into what category would an auto shutter (no additional
manipulation)  
> fall:
> a) breaks all profiling efforts
> b) can be used together with profiling

It's not yes/no - it's what degree of accuracy you get.  For
best
accuracy, you should create a profile for the same situation
that you
will encounter.  For cameras, the characteristics of the
illuminant are
critical.  If you mean that only the shutter speed is
automatic, then
that shoudln't matter as long as the spectral power density
of the
illuminant is the same.  But, if you are changing the
character of the
light, that's something else.

You may still get more repeatable results with a profile and
lighting
variance than with no profile and the same lighting
variance.

------------------------------------------------------------
-------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and
a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Lcms-user mailing list
Lcms-userlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user

On Thursday 18 October 2007 13:41:00 Mark wrote:
> Hello Bob,
>
> > A fundamental assumption here is that camera
operation is repeatable
> > and not very sensitive to the environment. 
Typical consumer cameras
> > are not very repeatable and are very sensitive to
the environment
> > because they always try to adjust for the best
looking high-contrast
> > image.
>
> ...
>
> > Scientific cameras, or professional cameras with
an entirely manual
> > mode, are likely much more repeatable, at the
expense of far more work
> > on the part of the user to get a good shot.
>
> since you mentioned it.... I have a industrial camera
I'm profiling.
> The camera has an auto shutter mode, that in most cases
does really
> enhance the image.
>
> Into what category would an auto shutter (no additional
manipulation)
> fall:
> a) breaks all profiling efforts
> b) can be used together with profiling

Auto exposure functionality as long as it only adjusts
exposure is not an 
issue for using camera profiles.  The main issue is cameras
that do auto 
white balance.  As Bob pointed out these can change the
white balance of 
photos taken in exactly the same lighting conditions simply
because some of 
the content of the photos is different.  In addition some
comsumer grade 
cameras also adjust image "contrast" for each
image which also breaks any 
profiles created for the device. 

The difficult part for users of consumer grade cameras is
that the automaic 
featuers can not be disabled which makes using profiles
problematic at best.  
In other words to use profiles with your camera you need a
camera the YOU 
control and you need to take control of the camera.  With
current cameras you 
will get your best results using RAW images since this by
passes any in 
camera processing.

>
> I've asked myself this question some time ago, but I
really can't say
> yet...
>
> Mark
>
>
------------------------------------------------------------
-------------
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems? 
Stop.
> Now Search log events and configuration files using
AJAX and a browser.
> Download your FREE copy of Splunk now >> http://get.splunk.com/
> _______________________________________________
> Lcms-user mailing list
> Lcms-userlists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/lcms-user




------------------------------------------------------------
-------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and
a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Lcms-user mailing list
Lcms-userlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user

On Thu, 18 Oct 2007, Mark wrote:
>
> Into what category would an auto shutter (no additional
manipulation)
> fall:
> a) breaks all profiling efforts
> b) can be used together with profiling

My guess would be the former.

Using a consumer camera in manual mode, I have played around
with 
exposure times and notice that there are profound
differences with 
color balance.

You pretty much need a raw CCD with a fixed aperture and
fixed shutter 
time.  The scene light levels and illumination
characteristics need to 
be similar as well.

Bob
======================================
Bob Friesenhahn
bfriesensimple.dallas.tx.us, http://w
ww.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMag
ick.org/


------------------------------------------------------------
-------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and
a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Lcms-user mailing list
Lcms-userlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user

On Thu, 18 Oct 2007, Hal V. Engel wrote:

> With current cameras you will get your best results
using RAW images 
> since this by passes any in camera processing.

It seems that there is often some in-camera processing
before the shot 
is even taken.  Typically you depress the button a bit while
the 
camera evaluates the situation, and press the rest of the
way so the 
shot is taken.  Even with raw image formats, the camera may
still be 
free to make other hardware-oriented adjustments such as CCD
bias 
voltages, aperture size, and shutter time.

Bob
======================================
Bob Friesenhahn
bfriesensimple.dallas.tx.us, http://w
ww.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMag
ick.org/


------------------------------------------------------------
-------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and
a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Lcms-user mailing list
Lcms-userlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user