[argyllcms] Question regarding Profile Validation
- From: <robert@xxxxxxxxxxxxxxxxxx>
- To: <argyllcms@xxxxxxxxxxxxx>
- Date: Wed, 5 Nov 2014 11:40:12 -0000
This is a follow-up from previous posts.
I am trying to validate a profile using the commands below:
>targen -v -d2 -G -f100 iPFTest
>copy iPFTest.ti1 iPFRef.ti1
>printtarg -v -r -ii1 -a1.0 -T300 -M6 -pA4 iPFTest
>cctiff -v -ia -e iPF6400-Canson-Baryta-310-Argyll-2584.icc iPFTest.tif
iPFTestO.tif
>move /Y iPFTestO.tif iPFTest.tif
>Pause Print iPFTest.tif using no color management.
>chartread iPFTest
>fakeread -v -Ia -l iPF6400-Canson-Baryta-310-Argyll-2584.icc iPFRef
>Pause The test results will be in iPFValidate.txt
>colverify -v2 -N -k -s -w -x -L iPF6400-Canson-Baryta-310-Argyll-2584.icc
iPFRef.ti3 iPFTest.ti3 > >iPFValidate.txt
Here is a summary of the results:
Total errors (CIEDE2000): peak = 2.775012, avg = 0.717430
Worst 10% errors (CIEDE2000): peak = 2.775012, avg = 2.054905
Best 90% errors (CIEDE2000): peak = 1.440833, avg = 0.584831
avg err X 0.005690, Y 0.006274, Z 0.004678
avg err L* 0.623982, a* 0.641691, b* 0.639550
The differences seem way too high, even allowing for a max dE00 of, say, 0.6
due to scan/print error.
I've tried the same test with cctiff and fakeread set to Relative and the
results are no better (but no worse).
If I run colverify without the -N flag, the results are off the scale (dE00
of 40+).
I've checked two of the spot colors using spotread with the following
results:
Spot 7 (had a dE00 of 0.5 from colverify):
Chartread: 25.19, -53.19, 18.86
Spotread: 25.40, -52.85, 18.65
(dELab of 0.45)
Spot 4 (had a dE00 of 2.31 from colverify)
Chartread: 45.18, -63.73, 36.68
Spotread: 45.25, -63.30, 36.73
(dELab of 0.44)
And, finally, I've scanned the same 100-patch target twice (with a
calibration in between) to see how repeatable my scanning/instrument is, and
here are the results:
No of test patches in worst 10% are = 10
No of test patches in best 90% are = 90
Verify results:
Total errors (CIEDE2000): peak = 0.232322, avg = 0.096521
Worst 10% errors (CIEDE2000): peak = 0.232322, avg = 0.214428
Best 90% errors (CIEDE2000): peak = 0.151862, avg = 0.085107
avg err X 0.000351, Y 0.000359, Z 0.000714
avg err L* 0.041514, a* 0.047190, b* 0.149554
Very repeatable, as you can see.
As a final bit of information, here is how I made the profile:
targen -v -d2 -G -e8 -B8 -f2688 %PROFILENAME%
printtarg -v -ii1 -a0.83 -T300 -M6 -pA4 %PROFILENAME%
chartread %PROFILENAME%
colprof -v -A"Canon" -M"iPF6400" -D"%PROFILENAME%" -qh -cmd -dpe
-SBetaRGB.icm -O"%PROFILENAME%.icc" %PROFILENAME%
So am I doing something wrong in the validation commands or the
profile-making commands above?
Thanks
Robert
Other related posts:
