Re: The MESS at the PRESS campaign (ICC as a standard)
Re: The MESS at the PRESS campaign (ICC as a standard)
- Subject: Re: The MESS at the PRESS campaign (ICC as a standard)
- From: Graeme Gill <email@hidden>
- Date: Tue, 13 Apr 2004 13:39:06 +1000
- Organization: Color Technology Solutions
Henrik Holmegaard wrote:
Encouraging the notion
that proprietary is better encourages underlying notions which are
factually incorrect. It also cuts off the ColorSync community from the
printing process.
Factually incorrect ? Care to back up this type of claim ?
The facts are that, as much as it would be a good thing to have
a standard that allows the interchange of color information in
a way that allows open interopretation of systems to a degree of
accuracy sufficient to satisfy proofing demands, the ICC profile
doesn't currently accomplish this.
The reasons for this are varied. Part of the reason is inherent :-
it's impossible to standardize a moving target, and the advancement
of knowledge, foundation standards (ie. CAM's), and technology
is still developing at a fast pace. It's possible to quickly experiment
with, and add such new developments to proprietary systems, but it's not
possible to do so to standards. Standards generally only accept widely
proven approaches, not new ideas that may not stand the test of time.
Other reasons are that the ICC standard is deficient in many
areas needed for high precision print proofing work. A brief summary
as I see it:
* The standard is confused between the aims of encoding measured
device/colorspace characteristics as a firm baseline (leaving
use of the device information up to the CMM), and providing
device/colorspace characteristics in terms of a "cooked",
CAM like PCS. [CAM = Color Appearance Model, PCS = Profile Connection Space].
* The workflow implicitly "recommended" by the ICC standard is not
sufficient for high accuracy. Gamut mapping has to be implemented
without knowledge of any particular source gamut. B2A tables are
a fast and convenient way of inverting the native device characteristics,
but much accuracy and flexibility (ie. black generation) is lost
compared to inverting the A2B table on the fly.
* Ambiguous definition of profile interpretation, particularly around
the transformation from absolute information (ie. instrument readings)
to the profile storage format which is media/D50 relative.
* Lack of support for essentials such as: viewing conditions (The information
that is standardized is only a sub set of what is required for modern CAMs),
ink limits (TAC), spectral characteristics (needed for fluorescent whitener
compensation), alternate illumination spectrums, alternate observer curves, etc.
* Device Calibration, essential for maintaining accuracy, is outside
the scope of the ICC standard, and is even hindered by the patent claims
surrounding the support that the ICC standard does have in this area.
Proprietary systems can address all these issues, and make sure that
the whole system works cohesively and accurately. Addressing the
current limitation of the ICC standard is not easy, as after all,
like most standards it's being supported by a committee of people,
each with their own interests and points of view :-)
Graeme Gill
_______________________________________________
colorsync-users mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/colorsync-users
Do not post admin requests to the list. They will be ignored.