[argyllcms] Re: Something is wrong
- From: Yves Gauvreau <gauvreau-yves@xxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Mon, 9 Dec 2019 05:05:20 -0500
On 12/9/2019 3:44 AM, Graeme Gill wrote:
Yves Gauvreau wrote:
I probably missed it but when chartread ask you to measure column "A" is the
first patch
we are expected read near the column name or a the other hand? I use a I1
Studio, if it
helps and with I1 Studio software we need to start at the bottom of the column
with the
name at the top. I'm a lefty and I often do things the other way around maybe
it's what
I've done here.
You should normally start at the label end. By default chartread will recognize
a strip read in the reverse order, but occassionally this automatic
recognition may fail. You can check by taking notice of the messages
after each strip.
Yes, I read the strip backwards. It happened a few time that I needed to
read a strip more then ounce. I didn't notice this reverse direction thing.
*****
Can the strip be not "OK" whatever the direction and not force me to
read it again?
*****
Ready to read strip pass Q
Press 'f' to move forward, 'b' to move back, 'n' for next unread,
'd' when done, Esc or 'q' to quit without saving.
Trigger instrument switch or any other key to start:
Strip read OK (Strip read in reverse direction)
Here I've used profcheck-v2 -k -s
[19.184549] 680 @ G6: 0.05677882 0.76758980 0.64295410 -> 52.888806 -45.135408
-17.880791
should be 67.714625 -43.671131 5.752487
[17.332769] 744 @ A8: 0.06602579 0.63141830 0.65458150 -> 45.229806 -31.856842
-30.463268
should be 59.342730 -33.916699 -11.389247
[17.117341] 885 @ O16: 0.09134051 0.61879910 0.14782940 -> 42.173443 -48.528894
26.790647
should be 57.310958 -50.925333 48.683499
[17.001343] 939 @ J16: 0.07797360 0.68389410 0.52254520 -> 50.198264 -44.325153
-8.034188
should be 63.203707 -44.651316 12.355741
[16.471298] 320 @ R7: 0.14998090 0.66372170 0.41561000 -> 48.838919 -44.600747
3.032622
should be 61.551652 -45.411698 24.949179
[16.441445] 750 @ AV13: 0.00000000 0.75222400 0.25815210 -> 45.622928
-60.646602 20.524097
should be 30.150628 -65.155564 0.777215
[15.722360] 714 @ N3: 0.08830396 0.85450520 0.58712140 -> 58.597183 -52.483819
-2.945300
should be 70.442940 -47.895117 19.475484
There doesn't seem to be a pattern to the patches with the largest errors, so
that tends to point to some other problem. Other problems would be
a .ti2 that doesn't match the actual chart (i.e. say if you accidentally
ran printtarg again and overwrote the .ti2), or some sort of serious
randomness to the printers behavior.
Cheers,
Graeme Gill.
That sounds like a possibility I'll need to check. I prepared in advance
a series of chart of around 1000 patches, to be printed on 8.5" x 11",
11" x 14", 13" x 19" paper, both a 8 bit and a 16 bit version and it's
very likely I screwed up.
For some reason it seem to have worked the first time I tried with
another paper, I'll check that.
Profile check complete, errors(CIEDE2000): max. = 14.422612, avg. =
0.269652, RMS = 0.558867
Here the results of my first try.
No of test patches = 1083
[14.422612] 613 @ AM1: 0.09977874 0.17987340 0.36504160 -> 42.244709
9.981661 2.920885 should be 50.966356 16.507345 20.297320
[4.066399] 740 @ A15: 0.12938120 0.27325860 0.37624170 -> 36.474260
0.542155 -6.190852 should be 34.008105 -1.078270 -9.929441
[2.608639] 791 @ AO3: 0.06689555 0.06626993 0.33185320 -> 27.341045
11.289271 -7.760844 should be 25.769894 10.369048 -10.603439
[2.442639] 1033 @ O1: 0.03489738 0.24901200 0.34206150 -> 28.555831
-3.235494 -10.202095 should be 26.967353 -4.621916 -12.826872
[2.082944] 213 @ P18: 0.18631270 0.19951170 0.31204700 -> 34.920035
6.929366 -1.618548 should be 33.199169 6.116486 -3.122124
[1.942514] 978 @ V12: 0.10043490 0.19508660 0.47919430 -> 32.670243
7.804357 -16.070804 should be 31.635832 7.139928 -18.023709
[1.547879] 862 @ K16: 0.08216983 0.09567407 0.46074620 -> 28.530661
13.491464 -15.525868 should be 27.688269 13.133342 -17.450767
[1.364980] 461 @ X1: 0.11918820 0.10437170 0.25444420 -> 28.625617
8.951205 -2.746744 should be 28.116005 8.388118 -4.234038
[1.228797] 712 @ AU12: 0.17523460 0.09637598 0.34157320 -> 31.634338
14.872989 -4.795932 should be 30.941091 14.444659 -6.295695
[1.173805] 1078 @ S3: 0.13083090 0.25862520 0.25784700 -> 33.224340
-1.904582 -0.952816 should be 32.626657 -2.547817 -1.722755
[1.087076] 136 @ M19: 0.13203630 0.13075460 0.00000000 -> 25.026343
1.246897 9.157339 should be 24.461197 1.000995 7.810861
[1.067391] 861 @ BE7: 0.08364996 0.20590520 0.02624552 -> 26.852863
-5.289168 8.187793 should be 27.276806 -5.091671 9.491578
[1.041022] 413 @ AD2: 0.11833370 0.20901810 0.13134970 -> 29.486803
-1.604195 4.433384 should be 29.589866 -0.979040 5.042973
[0.997226] 307 @ AO12: 0.18243690 0.14427410 0.45468830 -> 34.416005
14.137403 -11.075593 should be 34.011831 13.868320 -12.376160
[0.962169] 397 @ AS11: 0.12230110 0.28526740 0.47396050 -> 35.806098
1.322707 -15.191924 should be 35.559409 1.070368 -16.541766
[0.890763] 597 @ AS13: 0.06944381 0.17416650 0.21448080 -> 27.501955
0.092903 -2.400039 should be 27.272072 -0.283421 -3.172635
[0.859840] 1049 @ AT9: 0.01683070 0.07231250 0.09417868 -> 21.033865
0.297197 -0.945154 should be 20.848769 -0.275522 -0.962411
[0.825104] 457 @ AG3: 0.86613260 0.90133520 0.89927520 -> 87.815082
-2.067034 1.634558 should be 88.085572 -1.501480 1.887168
[0.806664] 450 @ AR13: 0.70663000 0.55269700 0.27733270 -> 65.476104
7.078447 37.096898 should be 66.335363 6.956785 36.035013
[0.734050] 524 @ AM9: 0.19577330 0.89442280 0.58084990 -> 58.380897
-42.678122 -9.141781 should be 58.564873 -42.568730 -7.953482
[0.723245] 1075 @ AH18: 0.19642940 0.79740600 0.38960860 -> 54.023597
-39.892375 4.826883 should be 53.773823 -40.592357 5.946479
[0.721149] 676 @ Q15: 0.93311970 0.09536889 0.54024570 -> 55.289921
63.560458 21.187680 should be 55.930500 63.446188 20.338593
I'm curious to find out if the highest values we see here are treated as
outliers by colprof algorithms? If not then do you have an idea of their
impact on the results?
In general, to get the best results possible, should we just pruned out
these outliers? Would taking multiple measures of the patches and use
the average or the "median" be a good practice or just a waste of time?
This particular paper is from Red River and it seems the profile was
made using an M3 filter, I compared the gamut of both and the Chromix
gamut is much larger but the test image I printed seems to give results
more inline with the Argyllcms made profile. In general, should we
consider M3 profile too unrealistic to be really useful in normal practice?
Thanks,
Yves
Other related posts:
