Re: Photoshop 6/7 problems with OptiCAL Monitor profiles
Re: Photoshop 6/7 problems with OptiCAL Monitor profiles
- Subject: Re: Photoshop 6/7 problems with OptiCAL Monitor profiles
- From: "Bob Chase" <email@hidden>
- Date: Wed, 12 Feb 2003 16:34:47 -0500
The following are my observations regarding Photoshop's use of OptiCAL
profiles. (OptiCAL v3.7, Windows 2000, Matrox Parhelia video card, CRT
monitor, DTP92 colorimeter):
1) At least for my calibration target, D65/gamma 2.2, OptiCAL-produced
monitor profiles are NOT fully compatible with any version of Photoshop, and
they haven't been for at least as far back as OptiCAL v2.5 for Windows,
where I originally checked. When you look at the values of OptiCAL produced
profiles using the Photoshop Color Settings "Custom RGB..." Working Space
Panel to display the OptiCAL profile information as Photoshop sees it, the
phosphor primaries numbers bear only distant similarities to the OptiCAL
Information palette numbers. (ColorVision support indicated that this
primaries disparity arises for non-5000K Whitepoints, growing larger with
increased distance from 5000K. Although I haven't verified that personally,
since I'm interested in D65 that is not a satisfactory "resolution".)
To check for yourself, launch Photoshop and go into Color Settings and click
on the Working Spaces RGB drop-down list and select the most recently
produced OptiCAL profile - probably also your Monitor RGB profile. This
loads the profile values into the Custom RGB settings so that when you next
select "Custom RGB..." you're presented with a panel displaying all the
profile's values. In theory, the profile phosphor primaries values should
match the numbers presented by OptiCAL's Tools/Information Window, aside
from rounding. In practice the primaries numbers don't match at all, so I
don't know how they're derived. Baffling...
In OptiCAL v2.5, there was an optional profile output choice of a Photoshop
.AMS file which could then be used in PS to produce a fully compatible
profile, using PS's "Load RGB..." and choosing file-type of Monitor Setup
(*.ams), then creating a profile with "Save RGB...". (An alternative
workaround is to manually type in the OptiCAL information palette primaries
numbers into the Photoshop Custom RGB working space definition and then save
it out as a monitor profile. The drawback is that the numbers are
rounded/truncated in the OptiCAL info palette, although I'm not sure to what
degree this precision-loss impacts actual observable results.)
Unfortunately, the .AMS output feature was removed in later OptiCAL
releases - a shame. The .AMS files produced by OptiCAL v2.5 can still be
used by Photoshop, even in v7, to generate completely compatible PS monitor
profiles. The .AMS files also contained a more accurate gamma assessment
than the seemingly current OptiCAL assumption of 100% calibration match to
target. For me, with changes to Matrox's video drivers, OptiCAL v2.5 is no
longer able to update my video card's LUT, so I had to upgrade to the latest
OptiCAL releases to calibrate the LUT. No more .AMS output. So currently I
use OptiCAL v3.7 to calibrate the LUT, and then run v2.5 to obtain a
profiling .AMS file for creating the profile in PS; twice as much work.
2) OptiCAL profiles produced through the File->Profile menu choice are not
even physically the same structure as those made with the calibration
routine. File sizes of File->Profile produced profiles are ~2000k versus
calibration-produced profiles of ~500k. Neither of which are fully
compatible with PS, as mentioned in 1). One would expect they'd have the
same structure and values as those created as part of the
calibration/profiling process.
3) It seems that OptiCAL presumes that if calibration is performed, then the
target has been matched 100% - somewhat presumptuous & probably not entirely
accurate. This may be implied in the Curves tool - where the Calibrated
checkbox is disabled after calibration because it's assuming that
calibration result = target? This is also implied by looking at the OptiCAL
"calibration" output profile values in Photoshop - the gamma is ALWAYS
dead-on the same value as the target - despite the minor differences that
invariably exist between calibration target and result. However,
File->Profile produced profiles do not necessarily show gammas that exactly
match the previously used calibration target. Older versions of OptiCAL
didn't force-match the profiled gamma to target after a calibration. Again,
I'm not sure how much real-world impact these discrepancies cause.
I've had great results with OptiCAL/Precal by using the OptiCAL v2.5
.AMS-file workaround. Maybe someday the process won't need this kludge, but
in the meantime I wish ColorVision would bring the .AMS file output option
back.
_______________________________________________
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.