Hi,

 first of all thank you both for the quick and informative answers. 
It is clear, that I get the best results with individual profiles for every 
illuminant, but I fear, I have to deal with the common behavior of the people 
working with microscopes and CCD-cameras doing this "white balance" on the 
illuminant. I got to find the best realizable solution and accept the lower 
precission. It will still be a great enhancement of the color quality!
I think, we will try Marti's "quick and dirty"-solution, but how do I estimate 
the value for the TempSrc in �K? What I have are RGB-values, which are 
normally not equal. They should be my input white point, which I want to be 
mapped to R=G=B at the end. Additionally I can't be sure, that they lie 
exactly on the line of temperature dependent white points (I do not know, how 
this line is called). And what are my TempDest-values, how can I calculate 
them?


Regards
Dirk Str�ker


> Hi,
>
> I fully agree the good solution is to have a different profile for each
> illuminant,
> as that would be the *only* way to get decent results.
>
> However, there is a quick hack to do if you really don't care about
> precission.
> Again, Hal has suggested what is the "good" way, and what I am here stating
> is
> just quick and dirty, but anyway here we go:
>
> There is a specialized, virtual device link profile for doing some specific
> tweaks.
> One of those tweaks is white point translation.
>
> cmsCreateBCHSWabstractProfile(int nLUTPoints,
>                                                      double Bright,
>                                                      double Contrast,
>                                                      double Hue,
>                                                      double Saturation,
>                                                      int TempSrc,
>                                                      int TempDest);
>
> Keep Bright, Hue and Saturation to zero. Keep Contrast to 1, and set
> TempSrc and TempDest
> as the temperatures in �K of input and output white points. That function
> would give you
> a abstract devicelink handle. Do a multiprofile transform joining input
> (microscope) profile, that one
> and finally sRGB (or whatever output).
>
> Regards,
> Marti Maria.
>
> ----- Original Message -----
> From: "Hal V Engel" <[EMAIL PROTECTED]>
> To: <[email protected]>
> Sent: Tuesday, March 01, 2005 2:23 AM
> Subject: Re: [Lcms-user] White balance
>
> Dirk,
>
>  As you have found out when doing color profiles for a device everything in
> the color processing chain matters and has an impact on the resulting
> profile.  A different light source or a light source adjusted differently
> will change the profile.  As a photographer I do not expect my day light
> camera profiles to work with photos that were shot using tungsten light. 
> So I have profiles for my camera for each type of light that I might shot
> with as well as for each film type.  On some occasions I may shot a custom
> color target for a specific shoot because I am not sure of the lighting. 
> So there is no easy answer as you can not reasonably expect a profile that
> is created using one device to work perfectly for another similar device. 
> Let alone for
> a similar device with the light adjusted differently.
>
> This all boils down to having a consistant workflow so that you can
> reproduce
> the same results consistently.   Since my situation is such that I only
> have to deal with my personal workflow this is not a big issue.  I
> understand how important this is and I always follow a consistant routine
> so that my results
> are reproducible.  It sounds like you are dealing with others that may need
> training to understand how this works and how their workflow influences the
> results they get.  In addition it may require that you have a number of
> profiles that are correct for different devices and conditions.
>
> Hal
>
> On Monday 28 February 2005 01:22 pm, Dirk Str�ker wrote:
> > Am Montag, 28. Februar 2005 22:18 schrieb Dirk Str�ker:
> > > Hi everybody,
> > >
> > >  I have a problem, which might is simple to solve, but in the moment
> > > I'm a bit confused and can't find a simple answer.
> > > Before I explain my problem I will give a short description of my
> > > setup. I'm working with a CCD-camera on a microscope. The color quality
> > > of the system was very bad, so I did some research, found ICC-profiles,
> > > color management...and so on (I learned a lot in the past weeks  ;-) 
> > > ). We bought software from Gretag MacBeth to create profiles for the
> > > cameras
> > > and we were pretty impressed over the results. After that we found
> > > "littleCMS" and tried to use it to achieve the improved color quality
> > > directly, which worked fine.
> > > The problem is, that it is a common technique, to make a so called
> > > "white
> > > balance" on the "illumination", the yellowish ligth of the tungsten
> > > halogen lamp should be white or gray (R=G=B). I understood the ICC
> > > color-management-technique in that way, that this is normally done by
> > > the
> > > chromatic adaption tag (often the "Bradford matrix"). This is matrix is
> > > calculated while creating the profile.
> > > But not every individual microscope lamp is the same. Some people dim
> > > the
> > > lamp, or the lamp is older... but everyone wants to make the pure lamp
> > > light gray.
> > > How can I do this
> >
> > Oops, I wasn't ready.
> >
> > Should I use "cmsCreateRGBProfile" and make a "transform" with the
> > "normal"
> > camera profile + the new RGBProfile?
> >
> > Maybe there is a simple solution and I'm just to blind to see...
> >
> > I hope my explanations weren't to confusing ;-)
> >
> >
> >
> > Dirk Str�ker
> >
> >
> > -------------------------------------------------------
> > SF email is sponsored by - The IT Product Guide
> > Read honest & candid reviews on hundreds of IT Products from real users.
> > Discover which products truly live up to the hype. Start reading now.
> > http://ads.osdn.com/?ad_ide95&alloc_id396&op�k
> > _______________________________________________
> > Lcms-user mailing list
> > [email protected]
> > https://lists.sourceforge.net/lists/listinfo/lcms-user


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_ide95&alloc_id396&op=click
_______________________________________________
Lcms-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/lcms-user

Reply via email to