[Openicc] XICC specification draft

[Chris Murphy]

On Jun 25, 2005, at 3:27 PM, Craig Bradney wrote:

> Hmm, Scribus would need similar code into Aqua too. As of Scribus  
> 1.3.0cvs, we
> are now running natively there. Our 1.2.x series does run via Fink/ 
> X11 tho.

The case of a page layout application running remotely is an  
interesting one. Each object in page layout can have a different  
source ICC profile.

In color managed page layout applications on Windows and Mac OS such  
as QuarkXPress, PageMaker, and InDesign, this is done by the  
application. The OS has next to nothing to do with it other than  
providing information like where profiles are located, and what the  
current display profile is. The application is sending device  
dependent information to the OS because prematching (display  
compensation) is done by the application.

This could still be done with Scribus running on one machine with the  
result displayed remotely, but it means the x11 client has a means of  
sending the client ICC profile to Scribus. Given that you can't be  
sure of what services/functions are available on any given OS or  
machine, application-level color management seems to be more  
realistic at this point. The color management can occur on "vector"  
or prerasterized data, which I'm guessing is what the server provides  
to the client which then rasterizes it.

In this case, all that would be needed is a way for the x11 client to  
grab the current display profile, and then send it to the server- 
side. This single display profile is less than 4k in most instances.  
Normalizing on the server side avoids the possibility of having to  
send multiple 3MB output device profiles to the x11 client over  
unknown bandwidth.


Chris Murphy
Color Remedies (TM)
www.colorremedies.com/realworldcolor
---------------------------------------------------------
Co-author "Real World Color Management, 2nd Ed"
Published by PeachPit Press (ISBN 0-321-26722-2)