Re: Changing black generation in standard profiles
Re: Changing black generation in standard profiles
- Subject: Re: Changing black generation in standard profiles
- From: Marco Ugolini <email@hidden>
- Date: Sat, 27 May 2006 17:11:01 -0700
In a message dated 5/27/06 7:11 AM, Roberto Michelena wrote:
> Inside some of the profiles you can find embedded the measurements
> file, and some tools (for example Gretag's MeasureTool) allow you to
> "open" a profile and extract the data provided it's there.
Hi Roberto.
Unfortunately, none of the standard Adobe profiles (US Web Coated (SWOP) v2,
US Sheetfed Coated v2, etc) contains a measurement tag (something like the
"CIED" tag that ProfileMaker builds into its own profiles).
I wonder if it's possible to get those measurement data from Adobe. Or
perhaps they are proprietary.
> And finally, for any profile, you can recreate the data (with some
> loss but not significant) by using Gretag's freeware ColorLab. You
> open a target reference file (in CMYK) as for example ECI2002. Then
> you go to the "Filter" menu and there you find "ICC Profile
> Conversion"; you select the ICC profile in question, select "absolute"
> mode and destination Lab. This will yield the Lab version of the
> chart, which you can then "save as" and use in ProfileMaker, or after
> some easy file reformatting, in PrintOpen. It can also be used in
> Monaco but it takes a little more fiddling with the text formatting.
In a message dated 5/27/06 9:28 AM, Terry Wyse wrote:
> One method to do this that I've used successfully is to take the
> existing profile, convert it to Lab using absolute colorimetric
> rendering and then use this dataset to build new profiles with
> different amounts of K generation, total ink, etc.
>
> The method I've used is to open the CMYK profiling chart of your
> choice in LOGO ColorLab (open the reference/device values data file)
> and simply convert to Lab using the Filter:ICC Profile Conversion
> menu. Use your profile as the source. If you don't choose a
> destination profile, it will default to Lab. Save the resulting Lab
> data as text and then use this in ProfileMaker.
>
> (snip)
>
> Using this method, I've done EXACTLY what you've described. I've even
> done this to verify that SWOPv2 is indeed using TR-001. The only
> "issue" is that your somewhat at the mercy of the accuracy of the ICC-
> to-Lab conversion but anytime I've checked my profile-to-Lab
> conversion back to the actual measurement data, the accuracy has been
> much less than 1 delta E which seems reasonable to me.
Many thanks to both of you, Terry and Roberto, for sharing this technique.
Using that ColorLab function appears easy enough.
This procedure is the same (or should be) as opening the CMYK testchart in
Photoshop, assigning the CMYK profile to it, then converting the testchart
to Lab using the absolute colorimetric intent, correct?
Fact is, though, that doing this in ColorLab and in Photoshop yields results
that do not coincide quite exactly. Using ColorLab's "Comparing" function, I
have measured a difference of up to a maximum of 0.93 Delta E 2000 between
the results obtained using these two separate methods, with about 1/10 of
the patches differing between 0.5 and 0.93 Delta E 2000 (171 out of 1780
patches). I guess this difference is mainly due to the different CM engines
used in each case (LogoSync for ColorLab, Adobe (ACE) for Photoshop).
At any rate, a difference of no more than 0.93 Delta E 2000 doesn't seem
like something to be overly concerned about, specially given the myriad
vagaries and variables on press, compared to which a Delta E of less than 1
appears truly negligible.
Thank you again, and all the best.
--------------
Marco Ugolini
Mill Valley, CA
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden