The driver appears to be working ok and I can now print from Adobe Reader, but the colours are very dark and in some cases appear over-saturated. I can definitely adjust this by tweaking settings such as the brightness, hue, saturation and gamma values, or perhaps by trying the halftone algorithm, but I can see this taking a lot of test prints...

I started out with using well tempered screens, the hp-cp1025.icm colour profile and perceptual intent. Brightness, hue, saturation and gamma are all at the default value of 0. This gave me very dark blues and over-saturated reds. Greens if anything were also a little over-saturated as well.

I tried a few changes to the halftone algorithm, colour profile and intent. So far the best result has been from disabling the ICM colour profile and using no ICM colour profile at all. This still gives dark blues, although not as dark as before. Reds look more correct and greens are pretty close to correct, if anything slightly under-saturated.

Has anyone encountered similar colour adjustment issues in linux with the HP CP1025nw? If so, what settings have you used to fix the colour adjustment? Is there perhaps a better ICM colour profile out there? The weird thing is that the windows driver is producing great output and from what I can tell from the settings in the device dialogs, it's using "no profile", however, compared to the linux output with "no profile", the linux output is still significantly darker, as noted above.

I'm having a similar problem using Kubuntu 10.10. I have the latest version of the source from today - last date in ChangeLog is 2011-02-10, so I assume I have the rickrich changes from 2010-12-13 mentioned above.

I have uninstalled the driver that ubuntu installed and installed according to the INSTALL file. The printer is working correctly over USB from my Windows XP virtual machine. The results in linux vary...

- when trying to print the ubuntu printer test page, I get black and white "snow" all over the page. The output is slightly different depending on the halftone algorihtm I use.
- when trying to print the CUPS printer test page, it looks mostly correct, however the entire page seems shifted down about 3.5mm - just enough that the bottom of the test page border gets cut off.
- when printing more complex output from Adobe Reader, I get an error - 'lpr: Bad job-sheets value ""!' and nothing gets printed.

I've uploaded scans of the output at http://www.bakuna.net/hp_cp1025nw_issues.pdf. The first three pages are the output for the ubuntu test page using Well Tempered Screens, Accurate Screens and Standard halftone algorithm respectively. The 4th page is a black and white printout of what the test page should look like from a working b+w printer. The next two pages show the CUPS printer test page output from the CP1025nw followed by the correct output from my b+w printer.

I'm guessing that maybe this could be ghostscript related or something? My ghostscript is the 8.71 build provided with Ubuntu 10.10. I'm still digging around to figure out where the problem is, but if someone has any suggestions that might save me a few test prints, it would be much appreciated. Alternatively, if there's any way I can get more useful debugging output, please let me know.

Works fine on Fedora 14 and Ubuntu 10.10.

$ lpr -Pcp1025nw -o media=a4 testpage-ubuntu-10.10-a4.ps
{printed fine in color, wts, ICM file hp-cp1025.icm}

$ lpr -Pcp1025nw -o media=Letter testpage-ubuntu-10.10-letter.ps
{printed fine in color, wts, ICM file hp-cp1025.icm }

Don't have any idea. Maybe:

$ grep foo2zjs /var/log/messages

and post here. i.e.

Feb 16 09:48:03 dual foo2zjs-wrapper: foo2zjs-wrapper -z3 -P -L0 -qwts -c -C10 -r1200x600 -p9 -m1 -s7 -n1 -Ghp-cp1025-rgb-392-bpp1.icm -I0
Feb 16 09:48:03 dual foo2zjs-wrapper: foo2zjs-icc2ps -o /usr/share/foo2zjs/icm/hp-cp1025-rgb-392-bpp1.icm -t0 > /tmp/icc12878.crd.ps
Feb 16 09:48:14 dual foo2zjs-wrapper: gs -sPAPERSIZE=a4 -g9920x7016 -r1200x600 -sDEVICE=bitcmyk -dCOLORSCREEN -dMaxBitmap=500000000 /tmp/icc12878.usecie.ps /tmp/icc12878.crd.ps /tmp/icc12878.selcrd.ps
Feb 16 09:48:14 dual foo2zjs-wrapper: foo2zjs -r1200x600 -g9920x7016 -p9 -m1 -n1 -d1 -s7 -z3 -c -u 192x2 -l 192x96 -L 0 -T3 -B -A -P

"Brightness, hue, saturation and gamma are all at the default value of 0."

Huh??? Don't have a way to change it in my driver!!!

$ ./getweb 1025
hp-cp1025-rgb-392-bpp1.icm

(c) Copyright Rick Richardson 2010. All Rights Reserved

Works fine for me.

I downloaded and installed the foo2zjs code to get my HP CP1025nw working. I had a similar issue with colors on Fedora 12. Using ghostscript version 8.71. The code I installed had foo2zjs version string "$Id: foo2zjs.c,v 1.107 2010/12/15 23:06:41 rick Exp $".

I fixed my color issues by downloading the Argyll v1.3.2 binaries to generate my own custom profile. I got a color target from http://www.targets.coloraid.de/ to calibrate my HP Scanjet 4070 for use as the colorimeter.

The result was pretty good.



Edited 1 time(s). Last edit at 03/14/2011 10:09PM by wolf.

Well how about sharing some more info. Did you create a different icm? I'm not able to get my cp1025 working here and as it is no cp1025nw that might not be a surprise :-D

Yes, I used the Argyll stuff with my scanner and a calibration target to create a custom ICM profile. I believe it is a printer RGB profile.

You can overwrite the downloaded one if you create one. I believe you can also save it in the /usr/share/foo2zjs/icm directory as "testing.icm", "user1.icm", etc in the same directory, if you don't want to overwrite the default one. If you do that, you just have to choose that particular file as your profile in the advanced tab of the printer.

I'd have to look around for a spot to upload the icm for easy access, or I'd have put a link in my original post. Also, it seems ghostscript's color handling seems to be in a state of flux. I suggest the profile be closely matched to the ghostscript version you're running.

The profile I created was working on a Fedora 12 system running ghostscript 8.71. I tried the same process on a Fedora 15 system with ghostscript 9.04. I summarized my unpleasant experiences with that project in a separate post. Bottom line is that you don't want to use ghostscript 9.whatever right now.