[cups] cups Digest, Vol 37, Issue 2
Alan Corey
alan01346 at gmail.com
Sun Dec 4 06:49:17 PST 2016
icm files are like TIFF files with a header, internal directory
structure, an endianness independent of the operating system. Most of
it is mac-endian. In a printer icm I see these tags:
# esig isig offset size
0 desc mluc 300 178
1 wtpt XYZ 480 20
2 bkpt XYZ 500 20
3 chad sf32 520 44
4 cprt mluc 564 72
5 A2B0 mAB 636 558608
6 A2B1 mAB 559244 558608
7 A2B2 mAB 1117852 558608
8 B2A0 mBA 1676460 344976
9 B2A1 mBA 2021436 344976
10 B2A2 mBA 2366412 344976
11 gamt mBA 2711388 104744
12 DEVS MSBN 2816132 692668
13 targ text 3508800 79704
In a display icm I see these tags:
# esig isig offset size
0 desc desc 288 153
1 chad sf32 444 44
2 cprt text 488 31
3 rXYZ XYZ 520 20
4 gXYZ XYZ 540 20
5 bXYZ XYZ 560 20
6 wtpt XYZ 580 20
7 lumi XYZ 600 20
8 rTRC curv 620 16
9 gTRC curv 636 16
10 bTRC curv 652 16
11 vcgt vcgt 668 1583
12 targ text 2252 27043
I'm working on writing handlers for each type of tag because inside
each there's another structure, like the first 4 bytes are a
signature, next 4 blank, then data starts.
A BA curve is the inverse of an AB curve, that is applying it undoes
what an AB does. I'm not sure why there's more than 1 set. I
extracted some as tab-delimiited data and fed them through Gnuplot,
they look like this:
http://ab1jx.1apps.com/misc/test_all_a_lines.gif
The colors are just the colors assigned by Gnuplot, not related to RGB
or CMYK colors. Notice I plot 6 data sets and get 3 lines, so 3 of
them exactly overlay 3 others. And there are some AB and some BA
here. The numbers are 16 bit unsigneds. But I'm getting curves like
I'd expect.
--
Credit is the root of all evil. - AB1JX
More information about the cups
mailing list