608
583
w:x:y
610
point as x, y</span><br style="font-family: monospace;">
611
<span style="font-family: monospace;">
612
613
a:adaptation Adaptatation
616
(default 50.0)</span><br style="font-family: monospace;">
617
<span style="font-family: monospace;">
618
619
b:background
620
Background % of image luminance (default 20)<br>
621
622
l:scenewhite Scene white in
623
cd.m^2 if surround = auto (default 250)<br
624
style="font-family: monospace;">
625
</span><span style="font-family: monospace;">
627
628
629
f:flare Flare
630
light % of image luminance (default 1)</span><br
631
style="font-family: monospace;">
632
<span style="font-family: monospace;">
633
634
635
f:X:Y:Z Flare color
636
as XYZ (default media white)</span><br style="font-family: monospace;">
637
<span style="font-family: monospace;">
638
639
640
f:x:y Flare color as x,
642
<a href="#P">-P</a>
643
650
</span></small><small><span style="font-family: monospace;"> <a
651
href="#O">-O outputfile</a>
652
Override the default output filename & extension.</span></small><br
653
style="font-family: monospace;">
654
<small><span style="font-family: monospace;"></span><span
655
style="font-family: monospace;"> </span><a
656
style="font-family: monospace;" href="#p1"><i>inoutfile</i></a><span
657
style="font-family: monospace;">
658
Base name for input.ti3/output.icc file</span></small><br>
661
<b><a name="v"></a>-v</b> Turn on verbose mode. Gives progress
662
information as the profile is created. Since colprof can take a long
663
time to generate, this is often useful to monitor progress. If used in
664
combination with the <b>-y</b> flag, the error of each test point to
665
the resulting profile will be printed out.<br>
667
<a name="A"></a>The <b>-A</b> parameter allows setting of the device
668
manufacturer description tag. The parameter should be a string that
669
identifies the manufacturer of the device being profiled. With most
670
command line shells, it will be necessary to enclose the parameter
671
with double quotes, so that spaces and other special characters are
672
included in the parameter, and not mistaken for the start of another
673
flag, or as a final command line parameters. By default no manufacturer
674
description string tag will be generated for the profile.<br>
676
<a name="M"></a>The <b>-M</b> parameter allows setting of the device
677
mode description tag. The parameter should be a string that identifies
678
the particular model of device being profiled. With most command line
679
shells, it will be necessary to enclose the parameter with double
680
quotes, so that spaces and other special characters are included in the
681
parameter, and not mistaken for the start of another flag, or as a
682
final command line parameters. By default no model description string
683
tag will be generated for the profile.<br>
685
<a name="D"></a>The <b>-D</b> parameter allows setting of the profile
686
description tag. The parameter should be a string that describes the
687
device and profile. On many systems, it will be this string that will
688
be used to identify the profile from a list of possible profiles. With
689
most command line shells, it will be
690
necessary to enclose the parameter with double quotes, so that spaces
691
and other special characters are included in the parameter, and not
692
mistaken for the start of another flag, or as a final command line
693
parameter. Many programs that deal with ICC profiles use the
694
description tag to identify a profile, rather than the profile
695
filename, so using a descriptive string is important in being able to
696
find a profile. By default, the base name of the resulting profile will
697
be used as the description.<br>
699
<a name="C"></a>The <b>-C</b> parameter allows setting of the profile
700
copyright tag. The parameter should be a string that describes the
701
copyright (if any) claimed on the profile being generated.. With most
702
command line shells, it will be necessary to enclose the parameter
703
with double quotes, so that spaces and other special characters are
704
included in the parameter, and not mistaken for the start of another
705
flag, or as a final command line parameters. By default a generic
706
copyright string will be generated for the profile.<br>
708
<a name="Z"></a>The <b>-Z</b> parameter allows setting of the profile
709
attribute flags. There are four flags: <span style="font-weight: bold;">t</span>
710
to set Transparency, the default being Reflective; <span
711
style="font-weight: bold;">m</span> to set Matte, the default is
712
Glossy; <span style="font-weight: bold;">n</span> to set Negative, the
713
default is Positive; <span style="font-weight: bold;">b</span> to set
714
BlackAndWhite, the default is Color.<br>
716
<a name="q"></a> The <b>-q</b> parameter sets the level of
717
effort and/or detail in the resulting profile. For table based profiles
718
("cLUT" profiles), it sets the main lookup table size, and hence
719
quality in the resulting profile. For matrix
720
profiles it sets the per channel curve detail level and fitting
721
"effort". It is <span style="text-decoration: underline;">highly
722
recommended</span> that <span style="font-weight: bold;">-qm</span> be
723
used as a starting point, and
724
other settings only tried after this has been evaluated. <span
725
style="font-weight: bold;">NOTE</span> that <span
726
style="font-weight: bold;">-qu</span> is a <span
727
style="text-decoration: underline;">test mode</span>, and shouldn't be
729
except to prove that it is not worth using.<br>
731
<a name="b"></a> The <b>-b</b> flag overrides the <b>-q</b>
732
parameter, and sets the lut resolution for the BtoA (inverse) to a low
733
value. The creation of the B2A table is fairly time consuming, and if
734
the profile is only going to be used by <a href="targen.html">targen</a>,
742
as an input space profile, or if it will
744
be linked as an output profile using Argyll's <a href="collink.html">collink</a>
745
tool using the <b>-G</b> option
746
(inverse AtoB option), then a high quality BtoA table is not required,
747
and some time and profile space can be saved. If the profile is to be
749
as an output space profile with another CMS, or is going to be linked
751
the simple (-s) or mapping mode (-g) options, then a good quality B2A
752
table is needed, and the -b flag should
753
<span style="font-weight: bold;">NOT</span>
754
be set. Optionaly, a specific B2A table quality can be set.<br>
756
For input devices, the presence of a B2A table is not mandatory,
757
and it can be omitted entirely from the profile by using <span
758
style="font-weight: bold;">-bn</span>. Note that input profiles and
759
matrix profiles will only contain a colorimetric intent table or matrix.<br>
761
<a name="y"></a> The <b>-y</b> flag does a verification check on the
762
AtoB profile. This is done by comparing what CIE colors the
763
profile predicts for the test chart test patches, and comparing them
764
to the actual values. A summary of the average and maximum Lab delta
765
E's will be printed out if this flag is set. If the <b>-v</b> flag is
766
also set, then information for each patch will also be printed.<br>
768
<a name="ni"></a><a name="np"></a><a name="no"></a>Normally cLUT base
770
generated with three major elements:-
772
device channel (shaper) input curves, the multi-dimensional lut table,
773
and per PCS channel (shaper) output curves. The Using the <b>-ni</b>
774
flag disables the creation of the per device channel curves, while
775
using the <b>-no</b> flag disables the creation
776
of the per PCS channel curves.<br>
777
For cLUT based profiles, the input curves that are written to the
778
profile are composed of two components, a shape to best match the
779
detailed shape of the device behavior, and a shape to distribute the
780
input values evenly across the LUT input indexes. The <span
781
style="font-weight: bold;">-no</span> flag disables the former, while
782
the <span style="font-weight: bold;">-np</span> flag disables the
785
<a name="nc"></a><span style="font-weight: bold;">-nc </span>Normally
786
the device and CIE/spectral sample data and calibration curves used to
788
stored in the <span style="font-weight: bold;">'targ'</span> text tag
789
in the resulting ICC profile. To suppress this and
790
make the resulting profile smaller, use the <span
791
style="font-weight: bold;">-nc </span>flag. <span
792
style="font-weight: bold;">Note</span> that this will then preclude
793
final calibrated device value ink limits from being computed for the
794
resulting profile in subsequent use (ie. <a href="collink.html">collink</a>,
795
<a href="xicclu.html">xicclu</a> etc.).<br>
797
<a name="k"></a> -<b>k</b> parameter sets the target level of black (K)
798
when creating a B2A CMYK output tables. This is often called a black
799
level, a black inking rule, black generation, or under color
800
removal. These set the target black level.<br>
802
Possible arguments to the <b>-k</b> flag are:<br>
804
<b> -kz</b> selects minimum black (0.0)<br>
805
<b> -kh</b> selects a black value of 0.5<br>
806
<b> -kx</b> selects the maximum possible black (1.0)<br>
807
<b> -kr</b> selects a linear black ramp, starting at minimum black for
808
highlight, and maximum black for shadow (equivalent to -kp 0 0 1 1 1).
809
This is the default.<br>
811
<b><a name="kp"></a>-k p stle stpo enpo enle shape</b> allows an
812
arbitrary black value ramp to be defined, consisting of a starting
813
value (stle) for highlights, a breakpoint L value (stpo) where it
814
starts to transition to the shadow level, an end breakpoint L (enpo)
815
where it flattens out again, and the finishing black level (enle) for
816
the shadows. There is also a curve parameter, that modifies the
817
transition from stle to enle to either be concave (ie. the
818
transition starts gradually and and finished more abruptly) using
819
values 0.0-1.0, with 0.0 being most concave, or convex (the transition
820
starts more abruptly but finishes gradually), using values 1.0-2.0,
821
with 2.0 being the most convex.<br>
823
Typical black value generation curve with parameters something
824
like: -kp 0 .1 .9 1 .5<br>
826
<tt> 1.0 K |
827
enpo<br>
828
829
|
834
835
|
839
840
|
844
845
|
848
849
|
852
stle
854
855
+-------------------<br>
856
0.0 K
857
0.0 stpo 1.0<br>
858
590
x, y</span><br style="font-family: monospace;">
591
<span style="font-family: monospace;">
592
593
a:adaptation Adaptatation
594
luminance in cd.m^2 (default 50.0)</span><br style="font-family:
596
<span style="font-family: monospace;">
597
598
b:background Background %
599
of image luminance (default 20)<br>
600
601
l:scenewhite
602
Scene white in cd.m^2 if surround = auto (default 250)<br
603
style="font-family: monospace;">
604
</span><span style="font-family: monospace;">
605
606
607
f:flare Flare light % of
608
image luminance (default 1)</span><br style="font-family:
610
<span style="font-family: monospace;">
611
612
613
f:X:Y:Z Flare color as XYZ
614
(default media white)</span><br style="font-family: monospace;">
615
<span style="font-family: monospace;">
616
617
618
f:x:y Flare
620
<a href="#P">-P</a>
621
624
Create gamut gammap_p.wrl and gammap_s.wrl diagostics<br>
625
</span></small><small><span style="font-family: monospace;"> <a
626
href="#O">-O outputfile</a>
630
the default output filename & extension.</span></small><br
631
style="font-family: monospace;">
632
<small><span style="font-family: monospace;"></span><span
633
style="font-family: monospace;"> </span><a
634
style="font-family: monospace;" href="#p1"><i>inoutfile</i></a><span
635
style="font-family: monospace;">
636
Base name for input.ti3/output.icc file</span></small><br>
639
<b><a name="v"></a>-v</b> Turn on verbose mode. Gives progress
640
information as the profile is created. Since colprof can take a long
641
time to generate, this is often useful to monitor progress. If used
642
in combination with the <b>-y</b> flag, the error of each test
643
point to the resulting profile will be printed out.<br>
645
<a name="A"></a>The <b>-A</b> parameter allows setting of the
646
device manufacturer description tag. The parameter should be a
647
string that identifies the manufacturer of the device being
648
profiled. With most command line shells, it will be necessary to
649
enclose the parameter with double quotes, so that spaces and other
650
special characters are included in the parameter, and not mistaken
651
for the start of another flag, or as a final command line
652
parameters. By default no manufacturer description string tag will
653
be generated for the profile.<br>
655
<a name="M"></a>The <b>-M</b> parameter allows setting of the
656
device mode description tag. The parameter should be a string that
657
identifies the particular model of device being profiled. With most
658
command line shells, it will be necessary to enclose the parameter
659
with double quotes, so that spaces and other special characters are
660
included in the parameter, and not mistaken for the start of another
661
flag, or as a final command line parameters. By default no model
662
description string tag will be generated for the profile.<br>
664
<a name="D"></a>The <b>-D</b> parameter allows setting of the
665
profile description tag. The parameter should be a string that
666
describes the device and profile. On many systems, it will be this
667
string that will be used to identify the profile from a list of
668
possible profiles. With most command line shells, it will be
669
necessary to enclose the parameter with double quotes, so that
670
spaces and other special characters are included in the parameter,
671
and not mistaken for the start of another flag, or as a final
672
command line parameter. Many programs that deal with ICC profiles
673
use the description tag to identify a profile, rather than the
674
profile filename, so using a descriptive string is important in
675
being able to find a profile. By default, the base name of the
676
resulting profile will be used as the description.<br>
678
<a name="C"></a>The <b>-C</b> parameter allows setting of the
679
profile copyright tag. The parameter should be a string that
680
describes the copyright (if any) claimed on the profile being
681
generated.. With most command line shells, it will be necessary to
682
enclose the parameter with double quotes, so that spaces and other
683
special characters are included in the parameter, and not mistaken
684
for the start of another flag, or as a final command line
685
parameters. By default a generic copyright string will be generated
688
<a name="Za"></a>The <b>-Z</b> parameter allows setting of the
689
profile attribute flags. There are four flags: <span
690
style="font-weight: bold;">t</span> to set Transparency, the
691
default being Reflective; <span style="font-weight: bold;">m</span>
692
to set Matte, the default is Glossy; <span style="font-weight:
693
bold;">n</span> to set Negative, the default is Positive; <span
694
style="font-weight: bold;">b</span> to set BlackAndWhite, the
695
default is Color.<br>
697
<a name="Zi"></a>The <b>-Z</b> parameter allows setting of the
698
profile default intent. The default intent can be one of the four
699
standard intents: <span style="font-weight: bold;">p</span> to set
700
Perceptual, <span style="font-weight: bold;">r</span> to set
701
Relative Colorimetric, <span style="font-weight: bold;">s</span> to
702
set Saturation, and <span style="font-weight: bold;">a</span> to
703
set Absolute colorimetric.<br>
705
<a name="q"></a> The <b>-q</b> parameter sets the level of effort
706
and/or detail in the resulting profile. For table based profiles
707
("cLUT" profiles), it sets the main lookup table size, and hence
708
quality in the resulting profile. For matrix profiles it sets the
709
per channel curve detail level and fitting "effort". It is <span
710
style="text-decoration: underline;">highly recommended</span> that
711
<span style="font-weight: bold;">-qm</span> be used as a starting
712
point, and other settings only tried after this has been evaluated.
713
<span style="font-weight: bold;">NOTE</span> that <span
714
style="font-weight: bold;">-qu</span> is a <span
715
style="text-decoration: underline;">test mode</span>, and
716
shouldn't be used, except to prove that it is not worth using.<br>
718
<a name="b"></a> The <b>-b</b> flag overrides the <b>-q</b>
719
parameter, and sets the lut resolution for the BtoA (inverse) to a
720
low value. The creation of the B2A table is fairly time consuming,
721
and if the profile is only going to be used by <a
722
href="targen.html">targen</a>, or if it will only be used as an
723
input space profile, or if it will only be linked as an output
724
profile using Argyll's <a href="collink.html">collink</a> tool
725
using the <b>-G</b> option (inverse AtoB option), then a high
726
quality BtoA table is not required, and some time and profile space
727
can be saved. If the profile is to be used as an output space
728
profile with another CMS, or is going to be linked using the simple
729
(-s) or mapping mode (-g) options, then a good quality B2A table is
730
needed, and the -b flag should <span style="font-weight: bold;">NOT</span>
731
be set. Optionaly, a specific B2A table quality can be set.<br>
733
For input devices, the presence of a B2A table is not
734
mandatory, and it can be omitted entirely from the profile by using
735
<span style="font-weight: bold;">-bn</span>. Note that input
736
profiles and matrix profiles will only contain a colorimetric intent
739
<a name="y"></a> The <b>-y</b> flag does a verification check on
740
the AtoB profile. This is done by comparing what CIE colors the
741
profile predicts for the test chart test patches, and comparing them
742
to the actual values. A summary of the average and maximum Lab delta
743
E's will be printed out if this flag is set. If the <b>-v</b> flag
744
is also set, then information for each patch will also be printed.<br>
746
<a name="ni"></a><a name="np"></a><a name="no"></a>Normally cLUT
747
base profiles are generated with three major elements:- per device
748
channel (shaper) input curves, the multi-dimensional lut table, and
749
per PCS channel (shaper) output curves. The Using the <b>-ni</b>
750
flag disables the creation of the per device channel curves, while
751
using the <b>-no</b> flag disables the creation of the per PCS
753
For cLUT based profiles, the input curves that are written to the
754
profile are composed of two components, a shape to best match the
755
detailed shape of the device behavior, and a shape to distribute the
756
input values evenly across the LUT input indexes. The <span
757
style="font-weight: bold;">-no</span> flag disables the former,
758
while the <span style="font-weight: bold;">-np</span> flag disables
761
<a name="nc"></a><span style="font-weight: bold;">-nc </span>Normally
765
device and CIE/spectral sample data and calibration curves used to
766
create a profile is stored in the <span style="font-weight: bold;">'targ'</span>
767
text tag in the resulting ICC profile. To suppress this and make the
768
resulting profile smaller, use the <span style="font-weight: bold;">-nc
770
</span>flag. <span style="font-weight: bold;">Note</span> that this
771
will then preclude final calibrated device value ink limits from
772
being computed for the resulting profile in subsequent use (ie. <a
773
href="collink.html">collink</a>, <a href="xicclu.html">xicclu</a>
776
<a name="k"></a> -<b>k</b> parameter sets the target level of black
777
(K) when creating a B2A CMYK output tables. This is often called a
778
black level, a black inking rule, black generation, or under color
779
removal. These set the target black level.<br>
781
Possible arguments to the <b>-k</b> flag are:<br>
783
<b> -kz</b> selects minimum black (0.0)<br>
784
<b> -kh</b> selects a black value of 0.5<br>
785
<b> -kx</b> selects the maximum possible black (1.0)<br>
786
<b> -kr</b> selects a linear black ramp, starting at minimum black
787
for highlight, and maximum black for shadow (equivalent to -kp 0 0 1
788
1 1). This is the default.<br>
790
<b><a name="kp"></a>-k p stle stpo enpo enle shape</b> allows
791
an arbitrary black value ramp to be defined, consisting of a
792
starting value (stle) for highlights, a breakpoint L value (stpo)
793
where it starts to transition to the shadow level, an end breakpoint
794
L (enpo) where it flattens out again, and the finishing black level
795
(enle) for the shadows. There is also a curve parameter, that
796
modifies the transition from stle to enle to either be concave
797
(ie. the transition starts gradually and and finished more
798
abruptly) using values 0.0-1.0, with 0.0 being most concave, or
799
convex (the transition starts more abruptly but finishes gradually),
800
using values 1.0-2.0, with 2.0 being the most convex.<br>
802
Typical black value generation curve with parameters something like:
805
<tt> 1.0 K |
806
enpo<br>
807
810
|
811
_______ enle<br>
812
815
| /<br>
816
819
| /<br>
820
823
| /<br>
824
827
| /<br>
828
829
stle | ------/<br>
830
831
+-------------------<br>
832
0.0 K
833
0.0
834
stpo 1.0<br>
835
859
836
White
863
For minimum sensitivity of printed output to the lighting spectrum, it
864
currently seems best to use the maximum possible black, but other black
865
generation levels (ie. 0.3 to 0.5) may well be preferred if one wants
867
minimize the noisy appearance of black on an inkjet device, or
868
if the banding behaviour or other rendering flaws of the printer is to
872
The <a href="xicclu.html">xicclu</a> tool can be used to plot out
873
the resulting black level for a given set of parameters, by using the <a
874
href="xicclu.html#g">-g</a> flag of a profile already created from the
877
<a name="K"></a> <span style="font-weight: bold;">-K parameters.</span>
878
Any of the <span style="font-weight: bold;">-k</span> options above
879
can use the <span style="font-weight: bold;">-K</span> version, in
880
which rather than a black value target being defined by the inking
881
rule, a black <span style="text-decoration: underline;">locus</span>
882
target is defined. For each lookup, the minimum possible black level
883
and the maximum possible black level is determined, the former
884
corresponding to a locus target of 0, and the latter corresponding to a
885
locus target of 1. For instance, at
887
white point, no black will be used in the output, even if the black
888
locus specifies a maximum (since the maximum amount of black that
889
can be used to print white is actually zero). Similarly, at the black
890
point, black may well be used, even if the black locus specifies
891
zero black (since a certain amount of black is needed to achieve the
892
desired density of color). <br>
894
<a name="l"></a> The <b>-l</b> <i>tlimit</i> parameter sets the total
895
ink limit (TAC, Total Area Coverage) for the CMYK separation, as a
896
total percentage from 0% to 400%, and overrides any ink limit specified
897
in the .ti3 file. The limit value should generally be set a little
898
below the value used in the test chart generation, to avoid the very
899
edges of the gamut. If the test chart ink limit has been chosen to be a
900
little beyond an acceptable level, then this number should be the
901
acceptable level. Although
902
limits can be set below 200%, this will generally restrict the color
904
noticeably, as fully saturated secondary colors will not be reproduced.
905
Values are between 220% and 300% for typical printing devices. Ink
906
limits will be in the final calibrated device values if the <span
907
style="font-weight: bold;">.ti3</span> includes the calibration table.<br>
909
<a name="L"></a> The <b>-L</b> <i>klimit</i> parameter sets the black
910
channel ink limit for the CMYK separation, as a total percentage from
912
to 100%. For printing press like devices, this can be used to prevent
914
black channel screening pattern "filling in". Typical values might be
916
95% to 99%. Note that with the current implementation this can slow
917
down the creation of the profile quite noticeably, so do not use <span
918
style="font-weight: bold;">-L</span> unless you really need to. Ink
919
limits will be in the final calibrated device values if the <span
920
style="font-weight: bold;">.ti3</span> includes the calibration table.<br>
922
<a name="a"></a> The <b>-a</b> parameter allows choosing an alternate
925
By default (equivalent
926
to <b>-al</b>) profile creates a <span style="font-weight: bold;">cLUT</span>
927
based table profile with a PCS
928
(Profile Connection Space) of L*a*b*, which generally gives the most
929
accurate results, and allows for the four different rendering intents
933
A cLUT base table profile using a PCS of XYZ can be created
934
if <b>-ax</b> is used, and this may have the advantage of better
935
accuracy for additive type devices (displays, scanners, cameras etc.),
936
may avoid clipping for displays with a colorant chromaticity that can't
937
be encoded in L*a*b* PCS space, and may give a more accurate white
938
point for input devices by avoiding
939
clipping of values above the white point that can occur in L*a*b* based
940
cLUT input profiles. By default cLUT XYZ PCS Display profiles will also
941
have a set of dummy matrix tags included in them, for better
942
compatibility with other systems. The dummy matrix deliberately swaps
943
Red and Green, so that it is obvious if the cLUT tables are not being
944
used. If it is important for both the cLUT and matrix be accurate, use <span
945
style="font-weight: bold;">-aX</span>, which will create shaper/matrix
948
For RGB input or display profiles, a simpler
950
of profile using either a gamma curves or a general shaper curves,
952
with a matrix can be created, although such a profile cannot support
954
or saturation intents. Gamma curve and matrix profiles can be created
956
specifying <b>-ag</b> or <b>-aG</b>, the former creating three
958
gamma curves, one for each device channel, and the latter creating one
959
common curve for all the device channels. The latter may be needed with
960
certain applications that will not accept different gamma curves for
962
channel. General shaper curve and matrix profiles (which are superior
964
gamma curve profiles) can be created by specifying <b>-as</b> or <b>-aS</b>,
973
device channel, and the latter creating one common curve for all the
974
device channels. The latter may be needed with certain applications
976
different shaper curves for each channel.<br>
978
The <span style="font-weight: bold;">-am</span> option will create a
979
matrix profile with linear (i.e. gamma = 1.0) curves. This may be
980
useful in creating a profile for a device that is known to have a
981
perfectly linear response, such as a camera in RAW mode.<br>
983
<a name="u"></a> <span style="font-weight: bold;">-u:</span> cLUT
985
input profiles will normally be created
986
such that the white point of the test chart, will be mapped to perfect
987
white when used with any of the non-absolute colorimetric intents. This
988
is the expected behaviour for input profiles. If such a profile
989
is then used with a sample that has a lighter color than the original
990
test chart, the profile will clip the value, since it cannot be
991
represented in the lut table. Using the <b>-u</b> flag causes the lut
992
based input profile to be constructed so that the lut table contains
993
absolute color values, and the white of the test chart will map to its
994
absolute value, and any values whiter than that, will not be clipped by
995
the profile, with values outside the range of the test chart being
996
extrapolated. The profile effectively operates in an absolute intent
997
mode, irrespective of what intent is selected when it is used.
998
This flag can be useful when an input profile is needed for using a
999
scanner as a "poor mans" colorimeter, or if the white point of the test
1000
chart doesn't represent the white points of media
1001
that will be used in practice, and that white point adjustment will be
1002
done individually in some downstream application.<br>
1004
<a name="un"></a><span style="font-weight: bold;">-un</span>: By
1005
default a cLUT input profile with the <span style="font-weight: bold;">-u</span>
1006
flag set will extrapolate values beyond the test chart white and black
1007
points, and to improve the plausibility of the extrapolation, a special
1008
matrix model will be created that is used to add a perfect device white
1009
and perfect device black test point to the set of test patches.
1010
Selecting <span style="font-weight: bold;">-un</span> disables the
1011
addition of these extra extrapolated white and black patches.<br>
1013
<a name="U"></a><span style="font-weight: bold;"> -U <span
1014
style="font-style: italic;">scale</span>:</span> Input profiles will
1016
such that the white point of the test chart, will be mapped to perfect
1017
white when used with any of the non-absolute colorimetric intents. This
1018
is the expected behavior for input profiles. Sometimes the test chart
1019
white is not quite the same as the media being converted through the
1020
input profile, and it may be desirable in these cases to adjust the
1021
input profile white point to compensate for this. The <span
1022
style="font-weight: bold;">-U</span> parameter allows this. If the
1023
media converted is a little darker than the test chart white, then use
1024
a scale factor slightly less than 1.0 to make sure that the media white
1025
comes out as white on conversion (ie. try 0.9 for instance). If the
1026
media is a little lighter than the test chart white and is "blowing
1027
out" the highlights, try a value slightly greater than 1.0 (ie. try 1.1
1030
<a name="R"></a><span style="font-weight: bold;"> -</span><span
1031
style="font-weight: bold;">R</span><span style="font-weight: bold;">:</span>
1032
Normally the white point, black point and primary locations (for matrix
1033
profiles) are computed so as to create profiles that best match the
1034
sample data provided. Some programs are not happy with the resulting
1035
locations if they have negative XYZ values, or if the white point has a
1036
Y value > 1. The <span style="font-weight: bold;">-R</span> option
1037
restricts the white, black and primary values, so as to work with these
1038
programs, but this will reduce the accuracy of the profile.<br>
1040
<a name="i"></a> The <b>-i</b> flag allows specifying a standard or
1041
custom illumination spectrum, applied to spectral .ti3 data to compute
1042
PCS (Profile Connection Space) tristimulus values. <b>A</b>, <b>D50</b>,
1043
<b>D65</b>, <b>F5</b>, <b>F8</b>, <b>F10</b> are a selection of
1044
standard illuminant spectrums, with <b>D50</b> being the default. If a
1045
filename is specified instead, it will be assumed to be an Argyll
1046
specific <a href="File_Formats.html#.sp">.sp</a> custom spectrum file.
1047
This only works if spectral data is available. Illuminant details are:<br>
1049
A CIE tungsten
1050
filament lamp 2848K<br>
1051
D50 CIE daylight 5000K<br>
1052
D65 CIE daylight 6500K<br>
1053
F5 CIE Fluorescent
1055
F8 CIE Fluorescent
1057
F10 CIE Fluorescent 5000K,
1060
Custom illuminants are most often used when a fluorescent tube base
1061
viewing booth is going to be used to view results. Other
1062
illuminant reference files could be created using a suitable measuring
1063
instrument such as a spectrocam, or an eyeone, although such
1064
instruments do not provide the necessary response down to Ultra Violet
1065
that is needed for accurate operation of Fluorescent Whitening Agent
1067
Note that if an illuminant other than D50 is chosen, the resulting ICC
1068
profile will not be standard, and cannot be freely interchanged with
1069
other profiles that that us the standard D50 illuminant, particularly
1070
if the absolute rendering intent is used. Profiles
1071
should generally be linked with other profiles that have the same
1075
<a name="o"></a> The <b>-o</b> flag allows specifying a tristimulus
1076
observer, and is used to compute
1077
tristimulus values. The following choices are available:<br>
1078
<b> 1931_2</b> selects the standard CIE 1931 2 degree observer.
1080
<b>1964_10</b> selects the standard CIE 1964 10 degree observer.<br>
1081
<b>1955_2</b> selects the Stiles and Birch 1955 2 degree
1083
<b>1978_2 </b>selects the Judd and Voss 1978 2 degree observer<br>
1084
<b>shaw</b> selects the Shaw and Fairchild 1997 2 degree
1087
Note that if an observer other than 1931 2 degree is chosen, the
1088
resulting ICC profile will not be standard, and cannot be freely
1089
interchanged with other profiles that that use the standard 1931 2
1091
observer. Profiles should only be linked with other profiles that have
1092
the same illuminant and observer. The <b>1978_2</b> observer or <span
1093
style="font-weight: bold;">shaw</span> observer may give slightly
1094
better results than the <b>1931_2</b> observer.<br>
1096
<a name="f"></a> The <b>-f</b> flag enables Fluorescent Whitening
1097
Agent (FWA) compensation. This only works if spectral data is
1099
allows the effects of different levels of Ultra Violet in the viewing
1100
illuminant from that used by the instrument, be compensated for. This
1101
will only work accurately if you specify the <span
1102
style="text-decoration: underline;">actual
1104
you are using to view the print</span>, using the <span
1105
style="font-weight: bold;"><span style="font-weight: bold;">-i</span></span>
1106
flag. If you are doing
1107
proofing, you need to apply this to <span
1108
style="text-decoration: underline;">both your
1109
source profile, and your
1110
destination profile</span>. Note that it is not sufficient to specify
1111
an illuminant with the same white point as the one you are using, you
1112
should specify the spectrum
1113
of the illuminant you are <span style="text-decoration: underline;">actually
1114
using</span> for the proofing,
1115
including its <span style="text-decoration: underline;">Ultra Violet</span>
1116
spectral content, otherwise FWA compensation won't work properly (but
1117
see the note above about non-standard illuminants and observers). This
1118
means you ideally need to measure your illuminant spectrum using an
1119
instrument that can measure down to 300nm. Such instruments are not
1120
easy to come by. The best alternative is to use the <a
1121
href="illumread.html">illumread</a> utility, which uses an indirect
1122
means of measuring an illuminant and estimating its UV content. Another
1123
alternative is to simply try different illuminant
1124
spectra in the <span style="font-weight: bold;">ref </span>directory,
1125
and see if one gives you the result you are after, although this will
1126
be fairly a tedious approach. The ref/D50_X.X.sp
1127
set of illuminant spectra are the D50 spectrum with different levels of
1128
U.V. added or subtracted, ref/D50_1.0.sp being the standard D50
1129
illuminant, and may be somewhere to start. Note that using the
1130
ref/D50_0.0.sp spectrum with <span style="font-weight: bold;">-f</span>
1131
gives a result that is comparable to that of a U.V. cut filter. See
1132
also the discussion <a href="FWA.html">About Fluorescent Whitening
1133
Agent compensation</a>.<br>
1134
[Note: Generally using <span style="font-weight: bold;">-f</span>
1135
with the standard D50 illuminant spectrum will predict that the device
1136
will produce bluer output than the default of not FWA compensation.
1137
This is because most instruments use an incandescent illuminant, which
1138
has lower relative levels of UV than D50, so the FWA compensation
1139
simulates the effect of the greater U.V. in the D50. Also note that in
1140
an absolute colorimetric color transformation, the more a profile
1141
predicts the output device will have blue output, the yellower the
1142
result will be, as the overall color correction compensates for the
1143
blueness. The opposite will happen for an input profile.]<br>
1145
<a name="r"></a> The <b>-r</b> parameter specifies the average
1146
deviation of device+instrument readings from the perfect, noiseless
1147
values as a percentage. Knowing the uncertainty in the reproduction and
1148
test patch reading can allow the profiling process to be optimized in
1149
determining the behaviour of the underlying system. The lower the
1150
uncertainty, the more each individual test reading can be relied on to
1151
infer the underlying systems color behaviour at that point in the
1152
device space. Conversely, the higher the uncertainty, the less the
1153
individual readings can be relied upon, and the more the collective
1154
response will have to be used. In effect, the higher the uncertainty,
1155
the more the input test patch values will be smoothed in determining
1156
the devices response. If the perfect, noiseless test patch values had a
1157
uniformly distributed error of +/- 1.0% added to them, then this would
1159
an average deviation of 0.5%. If the perfect, noiseless test patch
1160
values had a normally distributed error with a standard deviation
1161
of 1% added to them, then this would correspond to an average deviation
1162
of 0.564%. For a lower quality instrument (less than say a Gretag
1163
Spectrolino or Xrite DTP41), or a more variable device (such as a
1164
xerographic print engine, rather than a good quality inkjet), then you
1165
might be advised to increase the <span style="font-weight: bold;">-r</span>
1166
parameter above its default value (double or perhaps 4x would be good
1167
starting values.) <br>
1169
<a name="S"></a><a name="s"></a><span style="font-weight: bold;">-s
1170
-S </span>In order to generate perceptual and
1171
saturation intent B2A tables for output profiles, it is
1172
necessary to specify at least one profile to define what source gamut
1173
should be used in the source to destination gamut mapping. [For more
1174
information on <span style="text-decoration: underline;">why</span> a
1175
source gamut is needed, see <a href="iccgamutmapping.html">About ICC
1176
profiles and Gamut Mapping</a>] The
1177
<b>-S</b> parameter is used to do this, and doing so causes perceptual
1178
and saturation tables to be generated. If only a perceptual
1179
intent is needed, then the <b>-s</b> flag can be
1181
and the saturation intent will use the same table as the perceptual
1183
Note that a input, output, display or device colororspace profile
1185
specified, not a non-device colorspace,
1187
link, abstract or named color profile.<br>
1188
If no source gamut is specified for a cLUT Display profile, then an ICC
1189
Version 2.2.0 profile will be created with only an A2B0 and B2A0 tag.
1190
If a source gamut is specified, then an ICC Version 2.4.0 profile will
1191
be created with a full complement of B2A tags to support all
1192
intents. The source gamut is created from the corresponding intent
1193
table of the provided profile to the output table being created. A TIFF
1194
file containing an embedded ICC profile may be supplied as the argument.<br>
1195
<span style="font-weight: bold;">Note</span> that input profiles and
1196
matrix profiles will only contain a colorimetric intent table or
1197
matrix, and hence the <span style="font-weight: bold;">-s</span> and <span
1198
style="font-weight: bold;">-S</span> option is not relevant.<br>
1200
<a name="nP"></a><span style="font-weight: bold;">-nP</span>: Normally
1201
when a source profile is provided to define the source gamut for the
1202
output profile perceptual table gamut mapping, the perceptual source
1204
determine this gamut. This is because some profile have gamut
1205
transformations in their perceptual A2B tables that is not in the
1206
colorimetric A2B table, and this needs to be taken into account in
1207
creating the perceptual B2A table, so that when the two profiles are
1208
linked together with the perceptual intent, the gamut mapping works as
1209
intended. The <span style="font-weight: bold;">-nP</span> option
1210
causes the source gamut to be taken from the source profile
1211
colorimetric table instead, causing the perceptual gamut mapping
1212
created for the perceptual table to be from the natural source
1213
colorspace gamut to the output space gamut.<br>
1215
<a name="nS"></a><span style="font-weight: bold;">-nS</span>: Normally
1216
when a source profile is provided to define the source gamut for the
1217
output profile saturation table gamut mapping, the saturation source
1219
determine this gamut. This is because some profile have gamut
1220
transformations in their saturation A2B tables that is not in the
1221
colorimetric A2B table, and this needs to be taken into account in
1222
creating the saturation B2A table, so that when the two profiles are
1223
linked together with the saturation intent, the gamut mapping works as
1224
intended. The <span style="font-weight: bold;">-nS</span> option
1225
causes the source gamut to be taken from the source profile
1226
colorimetric table instead, causing the saturation gamut mapping
1227
created for the saturation table to be from the natural source
1228
colorspace gamut to the output space gamut.<small><span
1229
style="font-family: monospace;"></span></small><br>
1231
<a name="g"></a>The <span style="font-weight: bold;">-g</span> flag
1232
and its argument allow the use of a specific source gamut instead of
1235
profile. This is to allow optimizing the gamut mapping to a source
1236
gamut of a particular image, which can give slightly better
1237
results that gamut mapping from the gamut of the source colorspace.
1238
Such a source image gamut can be created using the <a
1239
href="tiffgamut.html"> tiffgamut</a> tool. The gamut provided to
1240
the <span style="font-weight: bold;">-g</span> <span
1241
style="font-weight: bold;"></span> flag should be in the
1242
same colorspace that <span style="font-weight: bold;">colprof</span>
1243
is using internally to connect the two profiles. For all intents except
1244
the last one (no. <span style="font-weight: bold;">7</span>),
1245
the space should be Jab appearance space, with the viewing
1246
conditions generally being those of the input profile viewing
1247
conditions. The input profile will normally be the one used to create a
1248
source image gamut using <span style="font-weight: bold;">tiffgamut</span>.<br>
1250
<b><a name="p"></a></b>The <b>-p</b> option allows specifying one or
1253
profiles that will be applied to the output tables, after any gamut
1254
mapping. An abstract
1255
profile is a way of specifying a color adjustment in a device
1256
independent way. The abstract profile might have been created using one
1257
of the <span style="font-weight: bold;">tweak</span> tools, such as <a
1258
href="refine.html">refine</a>.<br>
1259
If a single abstract profile is specified, then it will be applied to
1260
all the output tables (colorimetric, perceptual and saturation). To
1261
specify different abstract profiles for each output table, use a
1262
contiguous comma separated list of filenames. Omit a filename between
1263
the commas if no abstract profile is to be applied to a table. For
1264
instance: -<span style="font-weight: bold;">p
1265
colabst.icm,percabst.icm,satabst.icm</span> for three different
1266
abstract transforms, or: <span style="font-weight: bold;">-p
1267
,percabst.icm,</span> for just a perceptual table abstract transform.<br>
1269
One strategy for getting the best perceptual results with output
1270
profile when using ICC
1271
profiles with systems that don't accept device link profiles, is as
1272
follows: Specify a gamut mapping profile of opposite type to the type
1273
of device being profiled, and when linking, use the relative
1274
colorimetric intent if the two profiles are of the same type, and
1275
perceptual intent if the two profiles are of the opposite type. For
1276
instance, if you are
1277
creating a CMYK output profile, specify an RGB profile for the <b>-s</b>
1278
or <b>-S</b> parameter. If linking that profile with a CMYK source
1280
use relative colorimetric intent, or if linking with an RGB profile,
1282
the perceptual intent. Conversely, if creating an RGB output profile,
1284
a CMYK profile for the <b>-s</b> or <b>-S</b> parameter, and if
1286
that profile with an RGB source profile, use relative colorimetric
1288
or if linking with a CMYK profile, use the perceptual intent.<br>
1290
(Note that the perceptual and saturation table gamut mapping doesn't
1291
make any allowance for the application of the abstract profile. This is
1294
<a name="t"></a><a name="T"></a><span style="font-weight: bold;"></span><span
1295
style="font-weight: bold;"></span>Normally, the gamut mapping used in
1296
creating the perceptual and
1297
saturation intent tables for output profiles is set to perceptual and
1299
mapping (as would be expected), but it is possible to override this
1300
default selection for each intent using the <b>-t</b> and <b>-T</b>
1301
flags. The <b>-t</b> flag can be used to set the
1302
gamut mapping for the perceptual table, and the <b>-T</b>
1303
flag can be used to set the gamut mapping for the saturation table. A
1304
more detailed description of the different intents is given in <a
1305
href="collink.html#i">collink</a>. Note that selecting any of the
1306
absolute intents will
1307
probably not function as expected, since the perceptual and saturation
1308
tables are inherently relative
1309
colorimetric in nature.<br>
1311
<a name="c"></a><b><a name="d"></a></b>Since appearance space is used
1312
in the gamut mapping (just as it
1313
is in <a href="collink.html"> collink</a>), the viewing conditions
1314
for the source and destination colorspaces should really be specified.
1316
source colorspace is the profile specified with the <b>-s</b> or <b>-S</b>
1317
flag, and the destination is the profile being created. The <b>-c</b>
1318
and <b>-d</b> options allow specification
1319
of their respective, associated viewing conditions. The viewing
1321
information is used to map the profile PCS (Profile Connection Space,
1323
us either XYZ or L*a*b*) color into appearance space (CIECAM02), which
1325
a better colorspace to do gamut mapping in. The viewing conditions
1327
the conversion into appearance space to take account of how color will
1329
seen under particular viewing conditions.<br>
1331
Viewing conditions can be specified in two basic ways. One is to select
1332
from the list of "pre canned", enumerated viewing conditions, choosing
1333
one that is closest to the conditions that are appropriate for the
1334
media type and situation. Alternatively, the viewing conditions
1335
parameters can be specified individually. If both methods are
1336
used, them the chosen enumerated condition will be used as a base, and
1337
its parameters will then be individually overridden.<br>
1339
Appearance space is also used to provide a space to map any remaining
1340
out of gamut colors (after a possible gamut mapping has been applied)
1341
into the device gamut. <br>
1343
<b><a name="P"></a></b>The <b>-P</b> option causes diagnostic 3D <a
1344
href="File_Formats.html#VRML">VRML</a> plots to be created that
1345
illustrate the gamut mappings generated for the perceptual and
1346
saturation intent tables.<br>
1348
<a name="O"></a>The <span style="font-weight: bold;">-O</span>
1349
parameter allows the
1350
output file name & extension to be specified independently of the
1352
filename. Note that the full filename must be specified, including the
1353
extension.<span style="font-weight: bold;"></span><br>
1355
<a name="p1"></a> The final parameter is the file base name for the <a
1356
href="File_Formats.html#.ti3">.ti3</a> input test point data, and the
1357
resulting <a href="File_Formats.html#ICC">ICC</a> output profile (.icm
1358
extension on the MSWindows platform, .icc on Apple or Unix platforms).
1359
The <span style="font-weight: bold;">-O</span>
1360
parameter will override this default.
1362
Note that monochrome profiling isn't currently supported. It may be
1363
supported sometime in the future.<br>
1365
If the <b>-v</b> flag is used (verbose), then at the end of creating a
1367
the maximum and average fit error of the input points to the resulting
1369
will be reported. This is a good guide as to whether things have gone
1371
in creating a profile. Depending on the type of device, and the
1373
of the readings, average errors of 5 or less, and maximum errors of 15
1375
less would normally be expected. If errors are grossly higher than
1377
then this is an indication that something is seriously wrong with the
1379
testing, or profile creation.<br>
1381
Given a .ti3 file from a display device that contains display RAMDAC
1382
calibration information (generated by <a href="dispcal.html">dispcal</a>,
1385
<a href="dispread.html">dispread</a>), <span
1386
style="font-weight: bold;">colprof</span> will convert this into a <span
1387
style="font-weight: bold;">vcgt</span> tag in the resulting profile,
1388
so that the operating system tools can configure the display
1389
hardware appropriately, whenever the profile is used.<br>
1391
Given a .ti3 file from a print device that contains the per-channel
1392
calibration information (generated by <a href="printcal.html">printcal</a>,
1395
<a href="printtarg.html">printtarg</a> and <a href="chartread.html">chartread</a>),
1396
<span style="font-weight: bold;">colprof</span>
1397
will save this along with the .ti3 file in the <span
1398
style="font-weight: bold;">'targ'</span> text tag in the profile, <span
1399
style="font-weight: bold;"></span>
1400
so that subsequent evaluation of ink limits can compute the final
1401
calibrated device values.<br>
841
For minimum sensitivity of printed output to the lighting spectrum,
842
it currently seems best to use the maximum possible black, but other
843
black generation levels (ie. 0.3 to 0.5) may well be preferred if
844
one wants to minimize the noisy appearance of black on an inkjet
845
device, or if the banding behaviour or other rendering flaws of the
846
printer is to be minimized. <br>
848
The <a href="xicclu.html">xicclu</a> tool can be used to plot out
849
the resulting black level for a given set of parameters, by using
850
the <a href="xicclu.html#g">-g</a> flag of a profile already
851
created from the same .ti3 file.<br>
853
<a name="K"></a> <span style="font-weight: bold;">-K parameters.</span>
854
Any of the <span style="font-weight: bold;">-k</span> options above
855
can use the <span style="font-weight: bold;">-K</span> version, in
856
which rather than a black value target being defined by the inking
857
rule, a black <span style="text-decoration: underline;">locus</span>
858
target is defined. For each lookup, the minimum possible black level
859
and the maximum possible black level is determined, the former
860
corresponding to a locus target of 0, and the latter corresponding
861
to a locus target of 1. For instance, at the white point, no black
862
will be used in the output, even if the black locus specifies a
863
maximum (since the maximum amount of black that can be used to print
864
white is actually zero). Similarly, at the black point, black may
865
well be used, even if the black locus specifies zero black (since a
866
certain amount of black is needed to achieve the desired density of
869
<a name="l"></a> The <b>-l</b> <i>tlimit</i> parameter sets the
870
total ink limit (TAC, Total Area Coverage) for the CMYK separation,
871
as a total percentage from 0% to 400%, and overrides any ink limit
872
specified in the .ti3 file. The limit value should generally be set
873
a little below the value used in the test chart generation, to avoid
874
the very edges of the gamut. If the test chart ink limit has been
875
chosen to be a little beyond an acceptable level, then this number
876
should be the acceptable level. Although limits can be set below
877
200%, this will generally restrict the color gamut noticeably, as
878
fully saturated secondary colors will not be reproduced. Values are
879
between 220% and 300% for typical printing devices. Ink limits will
880
be in the final calibrated device values if the <span
881
style="font-weight: bold;">.ti3</span> includes the calibration
884
<a name="L"></a> The <b>-L</b> <i>klimit</i> parameter sets the
885
black channel ink limit for the CMYK separation, as a total
886
percentage from 0% to 100%. For printing press like devices, this
887
can be used to prevent the black channel screening pattern "filling
888
in". Typical values might be from 95% to 99%. Note that with the
889
current implementation this can slow down the creation of the
890
profile quite noticeably, so do not use <span style="font-weight:
891
bold;">-L</span> unless you really need to. Ink limits will be in
892
the final calibrated device values if the <span style="font-weight:
893
bold;">.ti3</span> includes the calibration table.<br>
895
<a name="a"></a> The <b>-a</b> parameter allows choosing an
896
alternate profile type. <br>
898
By default (equivalent to <b>-al</b>) profile creates a <span
899
style="font-weight: bold;">cLUT</span> based table profile with a
900
PCS (Profile Connection Space) of L*a*b*, which generally gives the
901
most accurate results, and allows for the four different rendering
902
intents that ICC profiles can support.<br>
904
A cLUT base table profile using a PCS of XYZ can be created if <b>-ax</b>
905
is used, and this may have the advantage of better accuracy for
906
additive type devices (displays, scanners, cameras etc.), may avoid
907
clipping for displays with a colorant chromaticity that can't be
908
encoded in L*a*b* PCS space, and may give a more accurate white
909
point for input devices by avoiding clipping of values above the
910
white point that can occur in L*a*b* based cLUT input profiles. By
911
default cLUT XYZ PCS Display profiles will also have a set of dummy
912
matrix tags included in them, for better compatibility with other
913
systems. The dummy matrix deliberately interchanges Red, Green and
914
Blue channels, so that it is obvious if the cLUT tables are not
915
being used. If it is important for both the cLUT and matrix be
916
accurate, use <span style="font-weight: bold;">-aX</span>, which
917
will create shaper/matrix tags.<br>
919
For RGB input or display profiles, a simpler type of profile using
920
either a gamma curves or a general shaper curves, combined with a
921
matrix can be created, although such a profile cannot support
922
perceptual or saturation intents. Gamma curve and matrix profiles
923
can be created by specifying <b>-ag</b> or <b>-aG</b>, the former
924
creating three independent gamma curves, one for each device
925
channel, and the latter creating one common curve for all the device
926
channels. The latter may be needed with certain applications that
927
will not accept different gamma curves for each channel. General
928
shaper curve and matrix profiles (which are superior to gamma curve
929
profiles) can be created by specifying <b>-as</b> or <b>-aS</b>,
930
the former creating three independent shaper curves, one for each
931
device channel, and the latter creating one common curve for all the
932
device channels. The latter may be needed with certain applications
933
that will not accept different shaper curves for each channel.<br>
935
The <span style="font-weight: bold;">-am</span> option will create
936
a matrix profile with linear (i.e. gamma = 1.0) curves. This may be
937
useful in creating a profile for a device that is known to have a
938
perfectly linear response, such as a camera in RAW mode.<br>
940
<a name="u"></a> <span style="font-weight: bold;">-u:</span> Iinput
941
profiles will normally be created such that the white point of the
942
test chart, will be mapped to perfect white when used with any of
943
the non-absolute colorimetric intents. This is the expected
944
behaviour for input profiles. If such a profile is then used with a
945
sample that has a lighter color than the original test chart, then a
946
cLUT profile will clip the value, since it cannot be represented in
947
the lut table. Using the <b>-u</b> flag causes the input profile to
948
be constructed so that the profile model contains absolute color
949
values, and the white of the test chart will map to its absolute
950
reflectance value, and any values of higher reflectance than that,
951
will not be clipped by a cLut profile, with values outside the
952
reflectance range of the test chart being extrapolated. The profile
953
effectively operates in an absolute intent mode, irrespective
954
of what intent is selected when it is used. This flag can be useful
955
when an input profile is needed for using a scanner as a "poor mans"
956
colorimeter, or if the white point of the test chart doesn't
957
represent the white points of media that will be used in practice,
958
and that white point adjustment will be done individually in some
959
downstream application.<br>
961
<a name="un"></a><span style="font-weight: bold;">-un</span>: By
962
default a cLUT input profile with the <span style="font-weight:
963
bold;">-u</span> flag set will extrapolate values beyond the test
964
chart white and black points, and to improve the plausibility of the
965
extrapolation, a special matrix model will be created that is used
966
to add a perfect device white and perfect device black test point to
967
the set of test patches. Selecting <span style="font-weight:
968
bold;">-un</span> disables the addition of these extra
969
extrapolated white and black patches.<br>
971
<a name="U"></a><span style="font-weight: bold;"> -U <span
972
style="font-style: italic;">scale</span>:</span> Input profiles
973
will normally be created such that the white point of the test
974
chart, will be mapped to perfect white when used with any of the
975
non-absolute colorimetric intents. This is the expected behavior for
976
input profiles. Sometimes the test chart white is not quite the same
977
as the media being converted through the input profile, and it may
978
be desirable in these cases to adjust the input profile white point
979
to compensate for this. This can happen in the case of a camera
980
profile, where the test chart is not perfectly exposed. The <span
981
style="font-weight: bold;">-U</span> parameter allows this. If the
982
media converted is a little darker than the test chart white, then
983
use a scale factor slightly less than 1.0 to make sure that the
984
media white comes out as white on conversion (ie. try 0.9 for
985
instance). If the media is a little lighter than the test chart
986
white and is "blowing out" the highlights, try a value slightly
987
greater than 1.0 (ie. try 1.1 for instance).<br>
989
<a name="R"></a><span style="font-weight: bold;"> -</span><span
990
style="font-weight: bold;">R</span><span style="font-weight:
991
bold;">:</span> Normally the white point, black point and primary
992
locations (for matrix profiles) are computed so as to create
993
profiles that best match the sample data provided. Some programs are
994
not happy with the resulting locations if they have negative XYZ
995
values, or if the white point has a Y value > 1. The <span
996
style="font-weight: bold;">-R</span> option restricts the white,
997
black and primary values, so as to work with these programs, but
998
this will reduce the accuracy of the profile.<br>
1000
<a name="i"></a> The <b>-i</b> flag allows specifying a standard or
1001
custom illumination spectrum, applied to spectral .ti3 data to
1002
compute PCS (Profile Connection Space) tristimulus values. <b>A</b>,
1003
<b>D50</b>, <b>D65</b>, <b>F5</b>, <b>F8</b>, <b>F10</b> are a
1004
selection of standard illuminant spectrums, with <b>D50</b> being
1005
the default. If a filename is specified instead, it will be assumed
1006
to be an Argyll specific <a href="File_Formats.html#.sp">.sp</a>
1007
custom spectrum file. This only works if spectral data is available.
1008
Illuminant details are:<br>
1010
A CIE
1011
tungsten filament lamp 2848K<br>
1012
D50 CIE daylight 5000K<br>
1013
D65 CIE daylight 6500K<br>
1014
F5 CIE Fluorescent
1016
F8 CIE Fluorescent
1018
F10 CIE Fluorescent
1021
Custom illuminants are most often used when a fluorescent tube base
1022
viewing booth is going to be used to view results. Other
1023
illuminant reference files could be created using a suitable
1024
measuring instrument such as a spectrocam, or an eyeone, although
1025
such instruments do not provide the necessary response down to Ultra
1026
Violet that is needed for accurate operation of Fluorescent
1027
Whitening Agent compensation.<br>
1028
Note that if an illuminant other than D50 is chosen, the resulting
1029
ICC profile will not be standard, and cannot be freely interchanged
1030
with other profiles that that us the standard D50 illuminant,
1031
particularly if the absolute rendering intent is used. Profiles
1032
should generally be linked with other profiles that have the same
1033
illuminant and observer.<br>
1035
<a name="o"></a> The <b>-o</b> flag allows specifying a tristimulus
1036
observer, and is used to compute tristimulus values. The following
1037
choices are available:<br>
1038
<b> 1931_2</b> selects the standard CIE 1931 2 degree
1039
observer. The default.<br>
1040
<b>1964_10</b> selects the standard CIE 1964 10 degree
1042
<b>1955_2</b> selects the Stiles and Birch 1955 2 degree
1044
<b>1978_2 </b>selects the Judd and Voss 1978 2 degree
1046
<b>shaw</b> selects the Shaw and Fairchild 1997 2 degree
1049
Note that if an observer other than 1931 2 degree is chosen, the
1050
resulting ICC profile will not be standard, and cannot be freely
1051
interchanged with other profiles that that use the standard 1931 2
1052
degree observer. Profiles should only be linked with other profiles
1053
that have the same illuminant and observer. The <b>1978_2</b>
1054
observer or <span style="font-weight: bold;">shaw</span> observer
1055
may give slightly better results than the <b>1931_2</b> observer.<br>
1057
<a name="f"></a> The <b>-f</b> flag enables Fluorescent Whitening
1058
Agent (FWA) compensation. This only works if spectral data is
1059
available, and allows the effects of different levels of Ultra
1060
Violet in the viewing illuminant from that used by the instrument,
1061
be compensated for. This will only work accurately if you specify
1062
the <span style="text-decoration: underline;">actual illuminant
1063
spectrum you are using to view the print</span>, using the <span
1064
style="font-weight: bold;"><span style="font-weight: bold;">-i</span></span>
1065
flag. If you are doing proofing, you need to apply this to <span
1066
style="text-decoration: underline;">both your source profile, and
1067
your destination profile</span>. Note that it is not sufficient to
1068
specify an illuminant with the same white point as the one you are
1069
using, you should specify the spectrum of the illuminant you are <span
1070
style="text-decoration: underline;">actually using</span> for the
1071
proofing, including its <span style="text-decoration: underline;">Ultra
1074
Violet</span> spectral content, otherwise FWA compensation won't
1075
work properly (but see the note above about non-standard illuminants
1076
and observers). This means you ideally need to measure your
1077
illuminant spectrum using an instrument that can measure down to
1078
300nm. Such instruments are not easy to come by. The best
1079
alternative is to use the <a href="illumread.html">illumread</a>
1080
utility, which uses an indirect means of measuring an illuminant and
1081
estimating its UV content. Another alternative is to simply try
1082
different illuminant spectra in the <span style="font-weight:
1083
bold;">ref </span>directory, and see if one gives you the result
1084
you are after, although this will be fairly a tedious approach. The
1085
ref/D50_X.X.sp set of illuminant spectra are the D50 spectrum with
1086
different levels of U.V. added or subtracted, ref/D50_1.0.sp being
1087
the standard D50 illuminant, and may be somewhere to start. Note
1088
that using the ref/D50_0.0.sp spectrum with <span
1089
style="font-weight: bold;">-f</span> gives a result that is
1090
comparable to that of a U.V. cut filter. See also the discussion <a
1091
href="FWA.html">About Fluorescent Whitening Agent compensation</a>.<br>
1092
[Note: Generally using <span style="font-weight: bold;">-f</span>
1093
with the standard D50 illuminant spectrum will predict that the
1094
device will produce bluer output than the default of not FWA
1095
compensation. This is because most instruments use an incandescent
1096
illuminant, which has lower relative levels of UV than D50, so the
1097
FWA compensation simulates the effect of the greater U.V. in the
1098
D50. Also note that in an absolute colorimetric color
1099
transformation, the more a profile predicts the output device will
1100
have blue output, the yellower the result will be, as the overall
1101
color correction compensates for the blueness. The opposite will
1102
happen for an input profile.]<br>
1104
<a name="r"></a> The <b>-r</b> parameter specifies the average
1105
deviation of device+instrument readings from the perfect, noiseless
1106
values as a percentage. Knowing the uncertainty in the reproduction
1107
and test patch reading can allow the profiling process to be
1108
optimized in determining the behaviour of the underlying system. The
1109
lower the uncertainty, the more each individual test reading can be
1110
relied on to infer the underlying systems color behaviour at that
1111
point in the device space. Conversely, the higher the uncertainty,
1112
the less the individual readings can be relied upon, and the more
1113
the collective response will have to be used. In effect, the higher
1114
the uncertainty, the more the input test patch values will be
1115
smoothed in determining the devices response. If the perfect,
1116
noiseless test patch values had a uniformly distributed error of +/-
1117
1.0% added to them, then this would be an average deviation of 0.5%.
1118
If the perfect, noiseless test patch values had a normally
1119
distributed error with a standard deviation of 1% added to
1120
them, then this would correspond to an average deviation of 0.564%.
1121
For a lower quality instrument (less than say a Gretag Spectrolino
1122
or Xrite DTP41), or a more variable device (such as a xerographic
1123
print engine, rather than a good quality inkjet), then you might be
1124
advised to increase the <span style="font-weight: bold;">-r</span>
1125
parameter above its default value (double or perhaps 4x would be
1126
good starting values.) <br>
1128
<a name="S"></a><a name="s"></a><span style="font-weight: bold;">-s
1129
-S </span>In order to generate perceptual and saturation
1130
intent B2A tables for output profiles, it is necessary to specify at
1131
least one profile to define what source gamut should be used in the
1132
source to destination gamut mapping. [For more information on <span
1133
style="text-decoration: underline;">why</span> a source gamut is
1134
needed, see <a href="iccgamutmapping.html">About ICC profiles and
1135
Gamut Mapping</a>] The <b>-S</b> parameter is used to do this,
1136
and doing so causes perceptual and saturation tables to be
1137
generated. If only a perceptual intent is needed, then the <b>-s</b>
1138
flag can be used, and the saturation intent will use the same table
1139
as the perceptual intent. Note that a input, output, display or
1140
device colororspace profile should be specified, not a non-device
1141
colorspace, device link, abstract or named color profile.<br>
1142
If no source gamut is specified for a cLUT Display profile, then an
1143
ICC Version 2.2.0 profile will be created with only an A2B0 and B2A0
1144
tag. If a source gamut is specified, then an ICC Version 2.4.0
1145
profile will be created with a full complement of B2A tags to
1146
support all intents. The source gamut is created from the
1147
corresponding intent table of the provided profile to the output
1148
table being created. A TIFF or JPEG file containing an embedded ICC
1149
profile may be supplied as the argument.<br>
1150
<span style="font-weight: bold;">Note</span> that input profiles and
1151
matrix profiles will only contain a colorimetric intent table or
1152
matrix, and hence the <span style="font-weight: bold;">-s</span>
1153
and <span style="font-weight: bold;">-S</span> option is not
1156
<a name="nP"></a><span style="font-weight: bold;">-nP</span>:
1157
Normally when a source profile is provided to define the source
1158
gamut for the output profile perceptual table gamut mapping, the
1159
perceptual source table is used to determine this gamut. This is
1160
because some profile have gamut transformations in their perceptual
1161
A2B tables that is not in the colorimetric A2B table, and this needs
1162
to be taken into account in creating the perceptual B2A table, so
1163
that when the two profiles are linked together with the perceptual
1164
intent, the gamut mapping works as intended. The <span
1165
style="font-weight: bold;">-nP</span> option causes the source
1166
gamut to be taken from the source profile colorimetric table
1167
instead, causing the perceptual gamut mapping created for the
1168
perceptual table to be from the natural source colorspace gamut to
1169
the output space gamut.<br>
1171
<a name="nS"></a><span style="font-weight: bold;">-nS</span>:
1172
Normally when a source profile is provided to define the source
1173
gamut for the output profile saturation table gamut mapping, the
1174
saturation source table is used to determine this gamut. This is
1175
because some profile have gamut transformations in their saturation
1176
A2B tables that is not in the colorimetric A2B table, and this needs
1177
to be taken into account in creating the saturation B2A table, so
1178
that when the two profiles are linked together with the saturation
1179
intent, the gamut mapping works as intended. The <span
1180
style="font-weight: bold;">-nS</span> option causes the source
1181
gamut to be taken from the source profile colorimetric table
1182
instead, causing the saturation gamut mapping created for the
1183
saturation table to be from the natural source colorspace gamut to
1184
the output space gamut.<small><span style="font-family: monospace;"></span></small><br>
1186
<a name="g"></a>The <span style="font-weight: bold;">-g</span> flag
1187
and its argument allow the use of a specific source gamut instead of
1188
that of the source profile. This is to allow optimizing the gamut
1189
mapping to a source gamut of a particular image, which can
1190
give slightly better results that gamut mapping from the gamut of
1191
the source colorspace. Such a source image gamut can be created
1192
using the <a href="tiffgamut.html"> tiffgamut</a> tool. The gamut
1193
provided to the <span style="font-weight: bold;">-g</span> <span
1194
style="font-weight: bold;"></span> flag should be in the same
1195
colorspace that <span style="font-weight: bold;">colprof</span> is
1196
using internally to connect the two profiles. For all intents except
1197
the last one (no. <span style="font-weight: bold;">7</span>), the
1198
space should be Jab appearance space, with the viewing conditions
1199
generally being those of the input profile viewing conditions. The
1200
input profile will normally be the one used to create a source image
1201
gamut using <span style="font-weight: bold;">tiffgamut</span>.<br>
1203
<b><a name="p"></a></b>The <b>-p</b> option allows specifying one
1204
or more abstract profiles that will be applied to the output tables,
1205
after any gamut mapping. An abstract profile is a way of specifying
1206
a color adjustment in a device independent way. The abstract profile
1207
might have been created using one of the <span style="font-weight:
1208
bold;">tweak</span> tools, such as <a href="refine.html">refine</a>.<br>
1209
If a single abstract profile is specified, then it will be applied
1210
to all the output tables (colorimetric, perceptual and saturation).
1211
To specify different abstract profiles for each output table, use a
1212
contiguous comma separated list of filenames. Omit a filename
1213
between the commas if no abstract profile is to be applied to a
1214
table. For instance: -<span style="font-weight: bold;">p
1215
colabst.icm,percabst.icm,satabst.icm</span> for three different
1216
abstract transforms, or: <span style="font-weight: bold;">-p
1217
,percabst.icm,</span> for just a perceptual table abstract
1220
One strategy for getting the best perceptual results with output
1221
profile when using ICC profiles with systems that don't accept
1222
device link profiles, is as follows: Specify a gamut mapping profile
1223
of opposite type to the type of device being profiled, and when
1224
linking, use the relative colorimetric intent if the two profiles
1225
are of the same type, and perceptual intent if the two profiles are
1226
of the opposite type. For instance, if you are creating a CMYK
1227
output profile, specify an RGB profile for the <b>-s</b> or <b>-S</b>
1228
parameter. If linking that profile with a CMYK source profile, use
1229
relative colorimetric intent, or if linking with an RGB profile, use
1230
the perceptual intent. Conversely, if creating an RGB output
1231
profile, specify a CMYK profile for the <b>-s</b> or <b>-S</b>
1232
parameter, and if linking that profile with an RGB source profile,
1233
use relative colorimetric intent, or if linking with a CMYK profile,
1234
use the perceptual intent.<br>
1236
(Note that the perceptual and saturation table gamut mapping doesn't
1237
make any allowance for the application of the abstract profile. This
1240
<a name="t"></a><a name="T"></a><span style="font-weight: bold;"></span><span
1241
style="font-weight: bold;"></span>Normally, the gamut mapping used
1242
in creating the perceptual and saturation intent tables for output
1243
profiles is set to perceptual and saturation gamut mapping (as would
1244
be expected), but it is possible to override this default selection
1245
for each intent using the <b>-t</b> and <b>-T</b> flags. The <b>-t</b>
1246
flag can be used to set the gamut mapping for the perceptual table,
1247
and the <b>-T</b> flag can be used to set the gamut mapping for the
1248
saturation table. A more detailed description of the different
1249
intents is given in <a href="collink.html#i">collink</a>. Note that
1250
selecting any of the absolute intents will probably not function as
1251
expected, since the perceptual and saturation tables are inherently
1252
relative colorimetric in nature.<br>
1254
<a name="c"></a><b><a name="d"></a></b>Since appearance space is
1255
used in the gamut mapping (just as it is in <a href="collink.html">
1256
collink</a>), the viewing conditions for the source and
1257
destination colorspaces should really be specified. The source
1258
colorspace is the profile specified with the <b>-s</b> or <b>-S</b>
1259
flag, and the destination is the profile being created. The <b>-c</b>
1260
and <b>-d</b> options allow specification of their respective,
1261
associated viewing conditions. The viewing condition information is
1262
used to map the profile PCS (Profile Connection Space, which us
1263
either XYZ or L*a*b*) color into appearance space (CIECAM02), which
1264
is a better colorspace to do gamut mapping in. The viewing
1265
conditions allow the conversion into appearance space to take
1266
account of how color will be seen under particular viewing
1269
Viewing conditions can be specified in two basic ways. One is to
1270
select from the list of "pre canned", enumerated viewing conditions,
1271
choosing one that is closest to the conditions that are appropriate
1272
for the media type and situation. Alternatively, the viewing
1273
conditions parameters can be specified individually. If both methods
1274
are used, them the chosen enumerated condition will be used as a
1275
base, and its parameters will then be individually overridden.<br>
1277
Appearance space is also used to provide a space to map any
1278
remaining out of gamut colors (after a possible gamut mapping has
1279
been applied) into the device gamut. <br>
1281
<b><a name="P"></a></b>The <b>-P</b> option causes diagnostic 3D <a
1282
href="File_Formats.html#VRML">VRML</a> plots to be created that
1283
illustrate the gamut mappings generated for the perceptual and
1284
saturation intent tables.<br>
1286
<a name="O"></a>The <span style="font-weight: bold;">-O</span>
1287
parameter allows the output file name & extension to be
1288
specified independently of the final parameter basename. Note that
1289
the full filename must be specified, including the extension.<span
1290
style="font-weight: bold;"></span><br>
1292
<a name="p1"></a> The final parameter is the file base name for the
1293
<a href="File_Formats.html#.ti3">.ti3</a> input test point data, and
1294
the resulting <a href="File_Formats.html#ICC">ICC</a> output
1295
profile (.icm extension on the MSWindows platform, .icc on Apple or
1296
Unix platforms). The <span style="font-weight: bold;">-O</span>
1297
parameter will override this default.
1299
Note that monochrome profiling isn't currently supported. It may be
1300
supported sometime in the future.<br>
1302
If the <b>-v</b> flag is used (verbose), then at the end of
1303
creating a profile, the maximum and average fit error of the input
1304
points to the resulting profile will be reported. This is a good
1305
guide as to whether things have gone smoothly in creating a profile.
1306
Depending on the type of device, and the consistency of the
1307
readings, average errors of 5 or less, and maximum errors of 15 or
1308
less would normally be expected. If errors are grossly higher than
1309
this, then this is an indication that something is seriously wrong
1310
with the device testing, or profile creation.<br>
1312
Given a .ti3 file from a display device that contains calibration
1313
curves (generated by <a href="dispcal.html">dispcal</a>, passed
1314
through <a href="dispread.html">dispread</a>) and the calibration
1315
indicates that the VideoLUTs are accessible for the device, then <span
1316
style="font-weight: bold;">colprof</span> will convert the
1317
calibration into a <span style="font-weight: bold;">vcgt</span> tag
1318
in the resulting profile so that the operating system tools can
1319
configure the display hardware appropriately, whenever the profile
1320
is used. If the VideoLUTs are not marked as being accessible, <span
1321
style="font-weight: bold;">colprof</span> will do nothing with the
1322
calibration curves. In this case, to apply calibration, the curves
1323
have to be incorporated in the subsequent workflow, either by
1324
incorporating them into the profile using <a
1325
href="applycal.html#p1">applycal</a>, or including them after the
1326
profile in a <a href="cctiff.html#p2">cctiff</a> profile chain.<br>
1328
Given a .ti3 file from a print device that contains the per-channel
1329
calibration information (generated by <a href="printcal.html">printcal</a>,
1330
passed through <a href="printtarg.html">printtarg</a> and <a
1331
href="chartread.html">chartread</a>), <span style="font-weight:
1332
bold;">colprof</span> will save this along with the .ti3 file in
1333
the <span style="font-weight: bold;">'targ'</span> text tag in the
1334
profile, <span style="font-weight: bold;"></span> so that
1335
subsequent evaluation of ink limits can compute the final calibrated