how to assess the quality of a profile?
how to assess the quality of a profile?
- Subject: how to assess the quality of a profile?
- From: Randy Norian <email@hidden>
- Date: Thu, 24 Jun 2004 13:15:30 -0500
So- I'm sending to this inkjet proofer through two workflows. The
inkjet is set up to emulate a halftone dot proofer (Polaproofer)
workflow #1)- files postscripted and sent through Wasatch SoftRip
Workflow #2)- This shop uses Nexus for trapping and so some jobs are
submitted through Nexus, and on to the HP5000 directly. It depends on
whether spot colors are being used, I believe.
At any rate-
both workflows were profiled, and set to emulate Polaproof.icm
color is similar between both workflows, yet not identical (what can
you do) the neutrals are not great from either workflow, and so I
wanted to figure out whether my source profile , destination profiles,
or other needed to be improved. Note that there are 2 destination
profiles, one generated through wasatch, one generated through Nexus,
but both use the same source, Pola.icm.
How to assess a single profile?
I have ColorPursuit, which looks at and ranks profiles with a quality
factor, but that's too abstract for me and anyhow I don't know what's
happening inside the software. So I went for a hands-on method.
So I created a file in Lab color space. I picked 4 swatches of color,
while using out-of-gamut preview for USSF, to carefully pick 30 colors
that fell well within the Sheetfed coated gamut. I also added 10
shades of neutral gray (L, 0, 0)
I converted this file to the Hp5000 color space, using "convert to
profile" in photoshop- relative colorimetric. I repeated the
conversion from Lab to cmyk using absolute colorimetric, with both
Apple and Adobe CMM. Shouldn't 'relative' work, since colors were
in-gamut?
So, I have 3 files that have been separated specifically for the
Hp5000. I sent them to the HP5000 with color management turned off,
and only the base calibration curve applied- the same state the
printer was in, when I profiled it. The profile was apx one week old
at this time.
I also separated one file to the Polaproofer's colorspace, Pola.icm.
I then read the Lab values of the swatches that came out- I have only
read the relative colorimetric conversions, thus far- to my surprise,
they were off a bit- these numbers were read with a Spetrostar
Spectroscan, as it does this kind of thing easily. I've checked it
against my Spectroscan and it's pretty close, in the neutrals usually
within delta-E of 1 from what the spectroscan reads. Anyhow-
Polaproofer:
the average error overall was delta-e76 of 8.1 in just the
neutrals, the avg error was 4.2 in the neutrals, setting delta L = 0
and looking at just the a & b (color) error, the avg error was 2.2
HP5000:
the average error overall was delta-e76 of 7.9 in just the
neutrals, the avg error was 5 in the neutrals, setting delta L = 0
and looking at just the a & b (color) error, the avg error was 2.6
but somehow I combine these 2 in Wasatch RIP, and the HP5000 matches
Polaproof output with an average delta-E of 4. Closer than when I only
use 1 profile, and do a conversion specifically for that output device.
How is that possible?
I still need to read the prints that came from absolute colorimetric
conversion. How accurate are my profiles? How accurate is the CMM in
Photoshop?
Part of this exercise is so they can work in photoshop with a soft
proof of the full HP5000 gamut, and separate files to the HP5000 for
one-off posters, etc, and access the full color range of the device.
I just figured doing this conversion ( Lab--> HP5000 ) in photoshop
would have given an extremely close result. Where did I go wrong?
Any thoughts or comments gladly accepted!
Randy N
_______________________________________________
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.