Mike Chaney's Tech Corner
April 09, 2020, 01:05:32 PM *
Welcome, Guest. Please login or register.
Did you miss your activation email?

Login with username, password and session length
News: Feb 2013: Qimage Ultimate Challenges... have fun and explore features!
 
   Home   Help Search Login Register  

Download and develop photos from your flash cards with one click!
Get a trial of
FlashPipe today and stop fumbling with explorer windows to transfer photos and videos
Pages: [1]
  Print  
Author Topic: Colors print issue with Qimage on HpZ9  (Read 668 times)
hedes
Newbie
*
Posts: 6


View Profile
« on: January 19, 2020, 10:28:56 AM »

Hello
i have an issue with Qimage when I print to my Hp Z9 :
1) first
- in photoshop, I export in ProPhoto RGB to Qimage (automate -> qimage)
- in Qimage, I select the ICC profile of the paper of my HP and select "colors managed by software" in the driver
- I print
=> printed colors are not accurate, the red seems to be desaturated

2) I do the same but in photoshop
=> colors are ok

3) in photoshop, I convert in the ICC profile of the paper of my HP
- I export to Qimage
- in Qimage,  I select 'OFF' in the printer profile and I select "colors managed by software" in the driver
=> colors are ok

Is there anything I missed in Qimage ?

Thanks and regards
Hery
Logged
admin
Administrator
Forum Superhero
*****
Posts: 2994



View Profile Email
« Reply #1 on: January 19, 2020, 01:29:55 PM »

After you use "File", "Automate", "Qimage Ultimate" from Photoshop and Qimage opens, hold your mouse cursor over the thumbnail of the image that was exported from PS and see if the EXIF status bar on the bottom of the main window indicates the photo is in ProPhoto RGB color space (see example: circled in the attached).

Regards,
Mike
Logged
hedes
Newbie
*
Posts: 6


View Profile
« Reply #2 on: January 20, 2020, 06:39:18 AM »

yes it's ok: i checked that point and the right ProPhoto  color space is displayed. I tried also with several options in the color conversion and still the issue.
Logged
admin
Administrator
Forum Superhero
*****
Posts: 2994



View Profile Email
« Reply #3 on: January 20, 2020, 02:02:28 PM »

yes it's ok: i checked that point and the right ProPhoto  color space is displayed. I tried also with several options in the color conversion and still the issue.

Then your 1 and 3 will produce the same results (send the same data to the printer).  If not, then one of the following must be true:

(1) When you converted to the printer profile in PS in your step 3, your profile conversion settings were different than those set in Qimage (for example, relative colorimetric, BPC on).

(2) The printer profile you are using is corrupted which can cause PS and QU to behave differently when trying to handle the bad profile.

Where did you get the printer profile you are using?  Is it one of HP's profiles (came with the driver)?  Is it from the manufacturer of some third party paper brand?  Did you make the profile yourself?

Finally, a third possibility comes to mind:

(3) It's possible that the image is tagged ProPhoto RGB in Photoshop but it hasn't been properly converted to ProPhoto RGB, making PS think it's in ProPhoto RGB but it really isn't.

When you hold the mouse cursor over the thumbnail of the exported image in Qimage and press the spacebar to open the image full screen, does color look OK on screen?

Mike
« Last Edit: January 20, 2020, 02:19:41 PM by admin » Logged
hedes
Newbie
*
Posts: 6


View Profile
« Reply #4 on: January 21, 2020, 06:16:01 AM »

Hello

Quote
When you hold the mouse cursor over the thumbnail of the exported image in Qimage and press the spacebar to open the image full screen, does color look OK on screen?
yes colors in Qimage are ok on screen

Quote
Where did you get the printer profile you are using?
I built my own one with the HPZ9 spectro. I repeated the operation 3 times..

Quote
When you converted to the printer profile in PS in your step 3, your profile conversion settings were different than those set in Qimage (for example, relative colorimetric, BPC on).
using 16 bpc in both cases.

Thanks for your reply.
Logged
admin
Administrator
Forum Superhero
*****
Posts: 2994



View Profile Email
« Reply #5 on: January 21, 2020, 02:19:28 PM »

OK, your original steps 1 and 3 are doing exactly the same thing (converting from ProPhoto RGB to your printer profile) with the only difference being that in 1, Qimage does the conversion and in 3, PS does the conversion.  Since you are using two different programs to perform an ICC profile conversion, we have to be sure the conversion parameters are exactly the same and that the profile itself is intact.  With that in mind, we have to check a few things:

(1) First, print again in Qimage with your printer profile selected in "Printer Profile" and color management turned off in the HP driver but before you print, click the tool button just to the right of the drop-down box for Printer Profile and select "Perceptual" intent with BPC on.  Then repeat and select "Relative Colorimetric" with BPC on so that you can compare one print in Perceptual intent with the same print in Relative Colorimetric intent: do they differ greatly?  Does one look better than the other?  You can print smaller sizes like 5x7 just to compare color.

(2) What ICC profiling tool did you use to create your custom profile?  I've seen some tools produce profiles with broken LUTs (look up tables) where only the Perceptual intent table or only the Relative Colorimetric table is valid and the other is invalid.  In cases like this, or even when there are just internal problems in the profile), Photoshop will "prune" the profile, ignore it, select a different intent when it finds a bad one, etc. and will give no message to alert you to the fact that is isn't showing you what the profile truly specifies.  Of course, this is a bad thing because you would never know your profile is corrupted in some way and would be led to believe it is fine... when it is not.  QU never does this type of pruning and will show you exactly what the profile specifies, so "bad profiles" can cause a difference in color when using PS vs QU.

(3) What does the soft proof (eye button under the live view page) look like in QU when you have it set up to use the printer profile?  Does it show muted reds too?  This could give us an indication of whether the problem lies in the profile or driver/printer.

Regards,
Mike
Logged
hedes
Newbie
*
Posts: 6


View Profile
« Reply #6 on: January 21, 2020, 06:47:08 PM »

Hello
1)  I'm going to do that in the week.
2) I use the tool provided in the HPZ9.
3) it does not show the muted reds. When I activate the 'GMT' button and I compare to PS, there are some huge differences: for PS there are much less unprintable colors. But the differences are not on the places when the mutated colors are the most obvious.
Logged
Ernst Dinkla
Sr. Member
****
Posts: 384


View Profile Email
« Reply #7 on: February 03, 2020, 09:32:45 AM »

If the HP Designjet Z9+ has the same incompatibility with QU's integration of the HP printer driver as the HP Z3200 has then I fear you have more issues with the Media Preset setting in the driver not sticking instead of issues with the color management choices. With the HP Designjet Z3200-PS my last step in setting up a print job is opening the real driver settings again and checking the Media Preset again, the portrait/landscape mode and the print quality settings (300/600PPI + gloss enhancer modes: off/economy/whole page) and then an OK.  If QU's printer menu window then shows another media preset I am not worried as I know the real driver settings are OK.  This way the other QU's printer driver menu windows (size, portrait/landscape, 300/600PPI) will show the settings as done in the real printer driver menu. The other way around, setting everything from QU's virtual printer driver menu can give problems on Media Preset and portrait/landscape setting in the real driver and the print quality setting dropping to the default. This applies as well on jobs taken from the log file.

If the HP Z9+ + driver behaves the same .......... I have no experience with that printer and driver.  I am very interested to know whether that is the case or not though.

Met vriendelijke groet, Ernst

http://www.pigment-print.com/SpectralPlots/SpectrumViz_1.htm
March 2017 update, 750+ inkjet media white spectral plots
Logged
admin
Administrator
Forum Superhero
*****
Posts: 2994



View Profile Email
« Reply #8 on: February 03, 2020, 03:17:34 PM »

My money would be on the HPZ9 printers having all the same driver bugs as the Z3 series.  A year ago I outlined a number of serious bugs in the Z series drivers and included proof via specific examples that showed how the HP drivers were not following the Windows driver specs... along with the simple code fix that was needed to make them comply with the specs.  I sent this info to someone I had worked with in the past who I knew worked on the Z series drivers for HP; turns out he no longer works for HP but was able to give me some information.

From that interaction, my understanding is that every HP Z driver includes the code for every Designjet printer back to the first one released.  They don't write new drivers: they simply take the old one from a prior model and add in new features.  Because of this, the Z series drivers are now over a million lines of code and no one wants to touch the baseline code that has been there for years: I was told no one dares touch it and it is considered "cryptic and obscure".  Given this, I would suspect that the bugs that plague the underlying Windows framework (API) support will just propogate from one HP DJ printer to the next and will never get fixed.

I've expressed my concern that I'm able to support features on every other printer that cannot be supported (or supported fully) on HP Z series printers because their drivers don't comply with Windows standards.  But I'm convinced they've built such a rats nest of code that they now do not have the resources to fix it: if you can get a print to the printer... leave it alone!

Mike
« Last Edit: February 03, 2020, 03:32:13 PM by admin » Logged
Ernst Dinkla
Sr. Member
****
Posts: 384


View Profile Email
« Reply #9 on: February 03, 2020, 04:11:04 PM »

My money would be on the HPZ9 printers having all the same driver bugs as the Z3 series.  A year ago I outlined a number of serious bugs in the Z series drivers and included proof via specific examples that showed how the HP drivers were not following the Windows driver specs... along with the simple code fix that was needed to make them comply with the specs.  I sent this info to someone I had worked with in the past who I knew worked on the Z series drivers for HP; turns out he no longer works for HP but was able to give me some information.

From that interaction, my understanding is that every HP Z driver includes the code for every Designjet printer back to the first one released.  They don't write new drivers: they simply take the old one from a prior model and add in new features.  Because of this, the Z series drivers are now over a million lines of code and no one wants to touch the baseline code that has been there for years: I was told no one dares touch it and it is considered "cryptic and obscure".  Given this, I would suspect that the bugs that plague the underlying Windows framework (API) support will just propogate from one HP DJ printer to the next and will never get fixed.

I've expressed my concern that I'm able to support features on every other printer that cannot be supported (or suppoerted fully) on HP Z series printers because their drivers don't comply with Windows standards.  But I'm convinced they've built such a rats nest of code that they now do not have the resources to fix it: if you can get a print to the printer... leave it alone!

Mike

Thank you Mike,

I saw the Designjet Z5200 name flashing by when I installed the Z3200 PCL3 driver on Windows 10 last week. Which intrigued me. One thing you could do is grey out QU's media preset window when QU finds an HP Designjet printer driver, it is really the main risk for CM issues. Sizes and portrait/landscape settings can also result in problems but the HP Print Preview stop in the workflow shows it when a print page meant landscape lands as portrait on a roll, most common issue then. Can still happen though when the roll is not 1118mm or 914mm wide but one mm less, so I have my custom media size settings at 1mm less.

I can manage but last week I tried a 2017 version of QU to get rid of the maze routes I am used to now. Too many new features and a better general GUI in the QU 2020 versions though. What helps is making the default settings of the HP driver to one that is most used in practice and having it set at 600PPI requested input quality. The last takes more printing time if it was not intended but the quality does not suffer then. The other way around it can lower quality on the best papers. Next to that I have extra Printing Shortcuts in the HP driver for more usual print jobs.

Is Qimage One coping better with HP Designjet drivers?

Met vriendelijke groet, Ernst

http://www.pigment-print.com/SpectralPlots/SpectrumViz_1.htm
March 2017 update, 750+ inkjet media white spectral plots
Logged
admin
Administrator
Forum Superhero
*****
Posts: 2994



View Profile Email
« Reply #10 on: February 03, 2020, 04:23:03 PM »

After writing that, it occurred to me that the tech-minded might appreciate some inside info on what is going on under the hood: what causes the HP Z series driver bugs.  I want to be clear: these driver bugs do not come from a decision by HP to simply not support some features in the driver.  The problem is worse than that.  The features are supported, but they are miscoded/mis-used.  The two biggest issues are outlined below, the first of which is the cause of what Ernst was talking about.  These bugs have nothing to do with Qimage or "QU's integration with the HP driver" since the latter is nothing more than the driver's integration with Windows itself.  They are simply driver bugs where the driver is improperly using Windows functions:

(1) Improper enumeration of media types.  The driver needs to communicate the media types to Windows so that Windows knows the paper types supported by the driver.  In the Designjet drivers (Z drivers), the table has cross-linked (duplicate) and even missing entries.  Every media type is supposed to have an index (number) and a name (text) and each entry (at least the index) should be unique.  So an HP Glossy paper type that you'd find in the driver might have an index of 110.  Matte might have an index of 127.  And so on.  The problem is that some of the entries collide, particularly if you create user media presets: over time the driver's media type table can become a mess of conflicting entries.  You may end up with a media type that isn't even found in the table: you select "My Paper 1" in the driver, the driver gives that an index value of 133, but when you enumerate the media types in the driver, the driver gives you the media type table and there is no index 133 listed: so QU can't get the name for that media type.  Worse, depending on what you do and the media presets you've created, you could end up with two listings for index 110... like 110 - HP Glossy and 110 - Red River Glossy.  Now, when you select "HP Glossy" in Qimage's drop-down list on the main UI and QU sends the 110 index to the driver to tell the driver to switch to that media type, which media type will the driver pick?  That is, which 110 is it going to use?  There should not be duplicate or missing indices: this is the root of the problem.

(2) Mis-use of media sizes.  Wonder why the Z series printers are the only ones that can't support the QU auto-roll-length feature and why you can't just pick "Custom" or "User Defined" in the sizes on QU's main window and specify the paper size outside the driver?  It's because once again, the driver's enumeration of media sizes (to Windows) is broken.  Here again, sizes have an index and size info.  Indices under 256 correspond to standard sizes that Windows knows about: Letter, A4, and other common sizes.  Sizes greater than 256 are predefined sizes that the driver can use - usually more obscure sizes or larger sizes.  But index 256 is special: that index is dmpaper_user.  If software (like QU) specifies a media size index of 256, that is a user defined media size and the driver is supposed to allow you to specify whatever size (width/length of paper) you like.  The problem is, HP's drivers have a collision here as well: they have a hard coded size called "ANSI B" that has an index of 256.  That should never happen!  Media size 256 is reserved by Windows to specify a user defined paper size yet HP has an entry 256 called ANSI B in their media size table that conflicts and basically tries to "overwrite" the custom size you specify.  So that bug is the only reason you can't select a media size of "Custom" or "User Defined" outside the driver (in QU's main UI) and specify any size you want.  It's also the reason the auto-roll-length feature won't work: because unlike every other printer on the market, you cannot set the paper length outside the driver due to this index 256 collision.

So... for inquiry minds, there you go.  I also have other motives for posting the details about these bugs: in the slim chance that at some point, someone at HP will find this post and care enough to actually fix these driver bugs that have been around for a decade or more.  Someone with a little knowledge and motivation should be able to fix these two bugs in a day's work:

(1) The driver should always enumerate ALL media types including ones the user creates, and those media types should have unique names and unique (non duplicative) indices.

(2) The driver should always enumerate ALL media sizes with unique indices and HP must STOP using media size index 256 to specify a pre-defined size (ANSI B).  There should be no entry/listing for media size index 256 and you should use that like the Windows spec specifies: it is to be used to specify a user-defined media size.

Mike
Logged
admin
Administrator
Forum Superhero
*****
Posts: 2994



View Profile Email
« Reply #11 on: February 03, 2020, 04:40:32 PM »

One thing you could do is grey out QU's media preset window when QU finds an HP Designjet printer driver, it is really the main risk for CM issues.

I did look into that about a year ago as I thought that might be a better option too.  I decided not to remove that feature for DJ drivers because I found that at least the media type settings worked OK when the driver is first installed.  The custom media size problem is an issue from the start but the problems with media type collisions in the driver don't seem to surface until some number of user media presets are created (and saved) in the driver.  So I didn't want to disable the feature for all and at least at the time, I thought I had a reliable way to determine if the media type bug existed at any given time.  I do detect collisions or missing indices in the media type table, so I thought that should eliminate any CM issues where the wrong paper type is chosen; with your latest report, however, I may have to take another look.  The steps that I took to "work around" the driver bugs may not be as complete as I thought.  When I find a duplicate media type index, QU should report "Select in driver" for example.  And when there is NO index found in the table for the paper you chose in the driver, it should list "UNDEFINED" along with the index number.  That seemed to prevent any problems where QU could indicate one media type but the driver have selected a different one.  Worst case, QU should show "UNDEFINED" or "Select in Driver" and you'd know you would have to open the driver.  But I'll try to take another look: it wouldn't surprise me if there are other bugs that I wasn't aware of (and don't detect yet).

Mike
Logged
Ernst Dinkla
Sr. Member
****
Posts: 384


View Profile Email
« Reply #12 on: February 04, 2020, 02:06:42 PM »

One thing you could do is grey out QU's media preset window when QU finds an HP Designjet printer driver, it is really the main risk for CM issues.

I did look into that about a year ago as I thought that might be a better option too.  I decided not to remove that feature for DJ drivers because I found that at least the media type settings worked OK when the driver is first installed.  The custom media size problem is an issue from the start but the problems with media type collisions in the driver don't seem to surface until some number of user media presets are created (and saved) in the driver.  So I didn't want to disable the feature for all and at least at the time, I thought I had a reliable way to determine if the media type bug existed at any given time.  I do detect collisions or missing indices in the media type table, so I thought that should eliminate any CM issues where the wrong paper type is chosen; with your latest report, however, I may have to take another look.  The steps that I took to "work around" the driver bugs may not be as complete as I thought.  When I find a duplicate media type index, QU should report "Select in driver" for example.  And when there is NO index found in the table for the paper you chose in the driver, it should list "UNDEFINED" along with the index number.  That seemed to prevent any problems where QU could indicate one media type but the driver have selected a different one.  Worst case, QU should show "UNDEFINED" or "Select in Driver" and you'd know you would have to open the driver.  But I'll try to take another look: it wouldn't surprise me if there are other bugs that I wasn't aware of (and don't detect yet).

Mike

Mark, I will manage as it is.  Though I am still curious whether the Z9+ driver might cope better in practice. It is the machine to upgrade to one day. Hardware wise I do not know a better one for me than the Z3200. Software wise HP can improve a lot, not just drivers, its websites as well.

Met vriendelijke groet, Ernst

http://www.pigment-print.com/SpectralPlots/SpectrumViz_1.htm
March 2017 update, 750+ inkjet media white spectral plots
Logged
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.21 | SMF © 2015, Simple Machines Valid XHTML 1.0! Valid CSS!