[argyllcms] Re: ArgyllCMS: shadow details and blue color issues. What I'm doing incorrectly?
- From: "Xavi" <aruiz@xxxxxxxxxxx>
- To: <argyllcms@xxxxxxxxxxxxx>
- Date: Thu, 25 Oct 2012 01:58:39 +0200
Hi Nikolay.
"I'm not sure that Your editing software didn't convert pixel values while
opening and saving."
I also have analized this case. I have compared the original tiff file
generated by printtarg versus the resized, with Ps, tiff file nad the values
of every patch is exactly the same.
To empiricaly verify I have printed the original file and the modified
resized tiff file and I have calculated the deltae 2000 for some patches:
0,51
0,14
0,15
0,11
0,21
0,14
0,42
0,42
0,25
I imagine that these differences are due no a not perfect repeatibility of
the Colormunki reads and due to the different times on print and reading
each sheet.
But seems to demonstrate that resizing the original TIFF file, created by
ArgyllCMS, on Ps, doesn't alters the patch values. I checked it when I read
that someone resized the test page and I see that Ps don't alterate de Lab
values, after resizing. Now, I have corrobored, also comparing the prints.
On another email you commented " I'm using ImageMagic "convert" utility to
crop and resize Argyll targets into desirable format."
What I see that your patches, after this converssion, haven't a regular
color. For exampl, if I analize the yellow color patch with Ps, the
information of each pixel varies. A pixel has a Lab 243,222,88 and the pixel
that limites has a value 242.221,87 [Lab]. This not occurs when resizing
with Ps. All the pixels, of a same color patch, have the same Lab value.
I don't know if it can affect to the overall results obtained using the
verify command, but it is an alteration of the color.
Yesterday I printed another test chart to create a newprofile and compare
with the habitual used profile on tests. I'm letting to dry. But I used a
resized TIFF file, created without the -D ( Dither 8 bit TIFF values down
from 16 bit)
Do you think that's important to use the -D option when create the TIFF
file?
Thank you for your appreciate help.
Best Regards,
Xavi.
_____
De: argyllcms-bounce@xxxxxxxxxxxxx [mailto:argyllcms-bounce@xxxxxxxxxxxxx]
En nombre de Nikolay Pokhilchenko
Enviado el: miércoles, 24 de octubre de 2012 21:10
Para: argyllcms@xxxxxxxxxxxxx
Asunto: [argyllcms] Re: ArgyllCMS: shadow details and blue color issues.
What I'm doing incorrectly?
By the way: try to generate new image (by printtarg) for Your instrument
type and print the target with ACPU without opening the image by something
else. Please, use -D key for printtarg.
I'm not sure that Your editing software didn't convert pixel values while
opening and saving.
Wed, 24 Oct 2012 20:32:29 +0200 от "Xavi" <aruiz@xxxxxxxxxxx>:
Hi Nikolay.
Thank you for doing a ti3 file for me to try to check that everything is
correct on my case.
I have resized a little your tiff file because I have a colormunki and the
size of the patches were too small.
The patch 29,14,50 [Lab] is out of gamut but the original patch that I print
is diferent. The original was the blue color patch of the simulation color
checker chart: 29,14,-50 [Lab]
I haven't modified the colors of your file. Only the size to be read with
CM.
I have printed yout test4.tif using the ACPU utility from Adobe and exactly
the same media settings (I have recorded them as a preset). The test target
was also printed with ACPU. I use ACPU because after some tests, seems that
ACPU prints different than the PS CS4 disabling color management.
I have let dry for about 8 hours and I have read it using the same spectro
that was used for reading the test patches, and on strip mode.
The results of the verify.exe command are:
.\..\verify.exe -v -D test4_grans_ACPU_2.ti3 Test4_fakeread.
ti3
No of test patches = 4
1: 55.685061 1.050502 1.474861 <=> 50.000000 -0.000200 -0.000300 de
5.966573
2: 50.098214 15.311162 47.347346 <=> 46.397000 13.620000 42.779000 de
6.117908
3: 83.059447 2.768188 76.146682 <=> 82.000000 4.000400 80.000000 de
4.181966
4: 43.876601 14.647638 20.301919 <=> 38.000000 14.000000 14.000000 de
8.641068
No of test patches in worst 10% are = 0
No of test patches in best 90% are = 4
Verify results:
Total errors: peak = 8.641068, avg = 6.226878
Worst 10% errors: peak = 0.000000, avg = 0.000000
Best 90% errors: peak = 8.641068, avg = 6.226878
Are the expected results? If not, when do you think that can be the issue?
I attach the used ti3 file.
Thank you for your appreciate help.
Best Regards,
Xavi.
_____
De: argyllcms-bounce@xxxxxxxxxxxxx [mailto:argyllcms-bounce@xxxxxxxxxxxxx]
En nombre de Nikolay Pokhilchenko
Enviado el: miércoles, 24 de octubre de 2012 8:39
Para: argyllcms@xxxxxxxxxxxxx
Asunto: [argyllcms] Re: ArgyllCMS: shadow details and blue color issues.
What I'm doing incorrectly?
Wed, 24 Oct 2012 02:06:02 +0200 "Xavi" wrote:
Then, on Photoshop I create 4 rectangles, that I fill, one by one, with the
RGB values obtained. I save them on Tiff format and I print the file with
the same program and the same printing options that I have used to print the
test chart.
Hello, Xavi!
I think You can introduce the error at this stage because there may be a
difference between AgyllCMS chart image generation and the Photoshop.
Try to print the target in attached Test4.zip the same way as You have
printed the targets for MP_B01_396p64g3bArgyllv140_121011.ti3.
Then read the chart the same way as You've read
MP_B01_396p64g3bArgyllv140_121011.ti3 (I'd recommend the reading in the
strip mode, not patch-by-patch).
After reading, compare the targets:
verify.exe -v Test4.ti3 Test4_fakeread.ti3
And You'll see actual workflow errors.
By the way, the patch 29.000000 14.000000 50.000000 [Lab] is out-of-gamut
for Your printer. You can't check the tolerance by out-of-gamut colors. So,
it's Lab value was replaced by the nearest possible (by xicclu -fif then
-ff): 46.397 13.620 42.779
Other related posts:
