[Openicc] Hello *and* (was): LINUX, Gutenprint / CUPS / Color policies

[Chris Murphy]

Hal V Engel hvengel at astound.net
Wed May 11 12:26:59 PDT 2005

> Don't know about how this works an a Mac but on Windows the printer  
> driver,

With few exceptions, none I can think of off hand, but the  
manufacturer printer drivers on both platforms default to the same  
settings: proprietary color management. The exception to this is a  
setting for image gamma which defaults to 2.2 on Windows and 1.8 on  
the Macintosh.

However, this is somewhat misleading because the operating system can  
still become involved in conversions. On Windows, if ICM is not  
explicitly selected, it's not invited. On Mac OS however, if it sees  
a profile supplied by the application it will use it even if  
ColorSync is NOT selected in the print driver. If ColorSync is not  
selected, the assumed destination space is Generic RGB (for raster  
printers), so ColorSync converts the data from application tagged  
source space to assumed Generic RGB and then hands it over to the  
printer driver for proprietary color management to the printer. In my  
view this isn't a bad scenario, even though it's exceptionally  
misleading that ColorSync is being used even though it wasn't  
requested (or possibly explicitly not requested).

> My experience is that the CMM mode in the printer driver is totally  
> borked and
> as far as I know is not used by anyone doing serious color work and  
> because
> it is not the default more is not used by "normal" users.

It can work, the problem is missing black point compensation at the  
OS level.

> Of course, we are replowing old ground and I think for the most  
> part everyone
> here understands what the problems are on Windows and the Mac.  I also
> believe that everyone here would like to avoid these problems when  
> CM is
> implemented in our open source projects.

Agreed!


>   To address these problems an
> overall CM framework is needed so that everyone implementing  
> systems knows
> what they need to do to fit into the CM framework.  This requires a  
> very
> broad spectrum of participants while beyond just those involved in
> implementing the printer subsystems and graphics applications.  So  
> far I
> would estimate that 80% to 90% of the activity on this list has  
> been centered
> on the printing subsystem and those involved with systems like  
> windowing,
> desktop environments, scanners and cameras have not been engaged in  
> this
> discussion.  But to design a CM framework we need to have these  
> other groups
> involved.   Anyone have any ideas on what we can do to get these  
> other groups
> involved?

The mistakes on Windows and Mac OS are not vast in quantity. But they  
are high quality mistakes :) The general model of apps informing the  
OS of a specific source profile for image data, OR the OS assuming  
something else is a rational one. Missing are discussions on how to  
stamp out untagged RGB (all apps writing out RGB data must be tagging  
the data with something, or logically it will be assumed to be sRGB).  
The assumed source space for untagged data today is generally agreed  
upon to be sRGB (except by Apple), but that won't always be the case.  
When do we expect that to not be the case and how do we differentiate  
between untagged data for which sRGB applies and something newer  
applies? (A big part of the dilemma of untagged data.)

We also have HDR files to contend with, something the ICC hasn't even  
dealt with yet, but they are here and in a widely available  
application that just shipped - Photoshop CS2. These are 32-bits/ 
channel floating point files. Granted it's so new it's hard to  
predict exactly what will happen, but there are HDR cameras (multiple  
exposures in a single shot), and HDR displays that exist today. They  
aren't widely available yet but they will be. So some discussion  
needs to be thinking about future as well, without getting totally  
mired in the marketing fluff.

Does Open ICC have representation on the ICC?


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