[argyllcms] V0.70 Beta 3 Mac OSX bus error
- From: Andrej Javoršek <Andrej.Javorsek@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Thu, 19 Apr 2007 17:44:10 +0200
Hello!
I'm testing new V0.70, and I'm impressed with iOne Pro support
(reflection, emission, transmission... increased spectral sampling
rate...).
Many thanks for good work!
But I'm haveing problem with all display utilities (dispcal, dispwin,
dispread). I'm geting bus error when I run any of them (with or
without instrument).
Read a Display, Version 0.70 (BETA3)
Author: Graeme W. Gill, licensed under the GPL
Diagnostic: Too few arguments
usage: dispread [options] outfile
-v Verbose mode
-d n Choose the display from the following list
(default 1)
Bus error
I can observe same problem with V0.60 but not with (I believe) V0.5x
I'm running Intel based MacBook Pro with Mac OSX 10.4.9
Best Regards!
Andrej
On 2007.4.13, at 17:18, Graeme Gill wrote:
Klaus Karcher wrote:
I did not find the time to test the eye-one driver, but made some
tests with profile an icclink (argyllV0.70Beta3_osx10.3_bin).
I found some strange discrepancies between the -kp parameters in
icclink and profile:
I made two proflies:
profile -D"F39-test" -kp 0.0 0.2 0.9 1.0 0.6 -l320 -L98
-S adobe_sRGB.icc -cmt -dpp FOGRA39
mv FOGRA39.icc F39-test-dev.icc
icclink -G -ip -cmt -dpp -kp 0.0 0.2 0.9 1.0 0.6 -l320 -L98
adobe_sRGB.icc F39-test-dev.icc F39-test-lnk.icc
then I tested the results by converting a RGB gray ramp using cctiff:
cctiff -ip adobe_sRGB.icc -ip F39-test-dev.icc RGB-ramp.tif test-
dev.tif
cctiff F39-test-lnk.icc RGB-ramp.tif test-link.tif
The separations look completely different, see attached plot "dev
(thin)_vs_link(bold).pdf"
The basic curve is the same, but the K generation L* control level
is a bit different (mainly offset) between the two approaches.
This turns out to be due to a collection of minor bugs,
the most significant being that the white and black points
of the newly created profile aren't being recognized for
the inversion and gamut mapping steps. There were also
issues with the exact colorspace the K generation curve
L* control level was in. This should be fixed in the
next release.
BTW: is there no -D switch for icclink?
Not at the moment. I'll see about adding them.
Graeme Gill.
Other related posts:
