[Openicc] Color Management for Ubuntu
ku.b at gmx.de
Wed May 4 12:11:43 PDT 2011
Am 04.05.11, 12:38 -0600 schrieb Chris Murphy:
> On May 4, 2011, at 9:31 AM, Graeme Gill wrote:
>> Scott Geffert wrote:
>>> bodies can champion best practices of the entire industry. I look at this list and I
>>> worry because I see no overall vision, just a handful of people bantering about
>>> minutia. Where is the big picture? What party can we turn to to begin to take these
>>> great ideas into a broader industry wide event like the adoption of f stops in
>> Sketching big pictures is easy. Making things work in a way that matches a big picture
>> is where things get hard, particularly when many elements from different parties
>> are meant to work together.
>> Many of the parties that establish standards, formal and defacto, simply don't
>> understand many of the issues with color, and largely don't care about them.
>> A lot of the big players put effort into color some time ago, and now regard
>> it as "done", even though it's only half done. [How many developers do you
>> think Microsoft and Apple have currently working full time on color management ?
>> My guess would be 2 - one for each of them!]
>> I'm not at all comfortable with all the calls for the ICC to "fix it either" -
>> they aren't necessarily in the position to have such influence, and the
>> track record of the ICC is mixed. In my opinion there are many notable
>> examples of the ICC producing confused and slanted elements to the standard.
>> I currently feel much more comfortable with the ICC striving to producing color
>> mechanisms (ie. the profile standard) that are clean and usage neutral, and leaving
>> policy alone.
> I'm in agreement on all of these points. The ICC is ultimately a collection of companies in competition with each other. Collaboration has occurred on an individual and personal basis, but I see no overall incentive that encourages collaboration in the service of users. It was originally a way for the member companies to create a sub-industry with products for sale a lot of which worked poorly.
> And now we've arrived at an inevitable amount of market conglomeration and departure, to the point we have only a few companies producing commercial products that build custom profiles. And each does everything they can to tie this function to their hardware, which limits user choice in hardware and software purchases, as well as workflow. Now we're running up against a real IP dilemma with ICC profiles containing baked in transforms rather than describing device behavior.
> I prefer the ICC profile file format be clean and usage neutral, but even this is difficult for the ICC to accomplish. I think there may be too many conflicts of interest.
Has'nt the last popular attemt to create colour transforms in realtime
simply failed? A profiler which builds ICC transforms in 0.01 seconds
would find surely its market. But I did yet not read about any success
story. Thats maybe one reason why we still see "baked in transforms".
developing for colour management
www.behrmann.name + www.oyranos.org
More information about the openicc