Using the HP Z3200 spectrophotometer to generate profiles for other printers

[Ernst Dinkla]

Good posts never die.  As doesn't my Z3200. 

2. I've never understood why only RGB (and not CMYK profiles) are used to profile printers that have both sets of inks.  If someone could lift this burden, I'd be grateful.

Thanks

With the RGB inks in the 12 ink set of a Z3xxx you can not make a full color print. Papers can not reflect enough light for additive RGB mixing and that is the only way RGB colors can create full color images. That in contrast with light emitting/transmitting displays where enough light is emitted through the RGB pixels. It would be interesting though whether the RGB inks of the Zxxx ink set could be addressed independently (with a very flexible RIP) and print on a transparent foil in additive dithering and displayed with light thrown through it. The inks are actually red, mint green and violet so not ideal either for that task. "interesting" as in having too much time and experimenting as an artist. Like I have done ages ago with RGB inks silkscreen printed in a line raster on foil.

The RGB profiling is aimed at OEM drivers that have their color management at the RGB side of the conversion from RGB images to CMYKetc ink sets, so called RGB-device printer color management. The CMYK profiling is aimed at RIPs that have their color management at the other side of the RGB>CMYK (and CMYK space>Printer CMYK) conversion or maybe better described; have their color management on the conversion itself.

The RGB inks in the Zxxx printers have a different task; they replace mixes of the LcMY inks at certain color hue angles when the required saturation gets too high for resp MY, LcY, MLc ink mixes. So less of the LcMY inks are used there and more of the RGB inks. It is comparable to the use of the grey inks where the low saturation mixes are gradually replaced by the grey inks till neutral only contains grey/black inks. At that side to create more consistent neutrality.


Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
March 2017 update, 750+ inkjet media white spectral plots

[Geraldo Garcia]

Good posts never die.  As doesn't my Z3200. 

Thanks! 

1. I am unsure I understand Geraldo's quote further below (I am certain it's my ignorance).  Does this mean the Argyll profile is useful only in printing/converting using relative colorimetric?  Should I create both ICC sets and use one for strictly relative colorimetric work and another for conversions to perceptual?  As background, I soft proof ProPhoto RGB TIFFs using relative colorimetric in Photoshop, but render them usually out of PS or LR in perceptual, and sometimes using HP's 16 bit web-based RIP.

That statement/behavior from Argyll almost fried my brain too. The thing is, their profiles work perfectly for RelCol conversions. For perceptual intent you should specify the input colorspace during the profile creation. That means you will have a profile that can only (properly) render images from that colorspace, lets say, AdobeRGB. Of course you can create several profiles from the same data file, one for AdobeRGB, one for sRGB, one ProPhoto... and choose the right one when you need.

Now, I can't understand why you softproof using RelCol and prints using Perceptual. You should use the same!

Regards.

[Brad P]

Thanks Geraldo.  I think I got it.  I always render my files now out of Phocus to ProPhoto RGB, then take them into LR, PS and other ProPhoto compatible plugins in my workflow.  And I pretty much always print in relative, so unless you or someone corrects me, I'll follow method 2 in your first series of posts. 

Proofing in RelCol and printing in perceptual is a small part of a relatively new larger proofing process for me.  Since you asked, I use ColorThink to identify out of gamut colors of an image I'm about to print. Then I manipulate those specific OOG colors using PS hue/saturation layers with luminosity curve masks to push the offending colors into print gamut, checking on the results of that against the soft proofing gamut warnings in PS.  Because PS's gamut warnings do not identify all OOG colors and round tripping to CT is a pain, when I get most or all of the PS gamut warnings to disappear, I proof print in Relcol first to see if I care about the differences, then aim to final print in perceptual because at least in theory that preserves at-the-margin hidden color detail.   Having typed that out and reflecting on my experience so far, the RelCol print is the least important part, and I might just as well rely on the perceptual rendering proofs since I'm getting pretty darn close using PS gamut warnings anyway.  Will save paper too.      

Thanks to Ernst too!

[Brad P]

I now have everything uploaded and have a couple of comments.

Importantly for those following this thread, the original intent of this thread is to use the Z3200's internal spectrophotometer to create profiles for other printers.  In contrast, I like some other followers am attempting to create a better profile for my Z3200ps (version B) using an expanded color profile described in the first few posts.  This is a larger gamut profile that natively prints and scans by the Z3200, but does not appear at least to me to natively process and upload. Manual tweaks like this seem capable to improve the Z3200's printing gamut a bit.   

1. Looking with ColorThink my "before" printing gamut (created with the Z3200's canned hardware and programming) and the "after" gamut (created with the much larger sample, run through Argyll), I'm just barely over the fence that rendering of the larger profile is worth the time, money and effort.  It seems for now to be.  The new gamut appears well matched to the old and maybe up to 3-5% bigger overall.  Worthwhile to explore. 

2. I ended up following Geraldo's original instructions rather than a few well intended offshoots here and elsewhere.  Very well done Geraldo.  For Mac users, you may need to include root directories in the terminal program names and file names as I did.

3. I am having soft proofing issues with the new gamut. All appears normal in Photoshop, but Lightroom (which I usually print from) now shows the printer gamut blown out. Probably user error, possibly something else. 

4. Simultaneously I am working with CHROMiX on a custom built profile for the same paper I am profiling here. The first round of that separate process is encouraging in that the printer gamut appears to be much larger (Maybe 15% in ColorThink, smaller in places too), but appears from tonight more difficult to install. 

We'll see how it goes.

[NikonD850Boy]

Hello Guys;

I have been trying to follow the above steps but am getting a lot of errors with both ICCGen and the Argyll software when trying to convert the .txt file to .icc...I am using Windows 10 64bith.

I have printed my color patch and also got the part of scanning and creating the .txt color profile. I just cannot convert this into an ICC profile.

Both are attached...

Is there an easier way to convert the .txt to .icc Or if someone is kind enough to do it for me so I can get started at the least for now while I figure it out.

[MHMG]

Is there an easier way to convert the .txt to .icc Or if someone is kind enough to do it for me so I can get started at the least for now while I figure it out.

I have attached an ICC profile made from your .txt file using BasicColor dropRGB software. DropRGB works seamlessly with the Z3200 generated .csv files, and it produces outstanding results. It doesn't get any easier than dropRGB to make high quality profiles. However, it's pretty expensive, so there's that.

Other software, for example, i1Profiler, doesn't like some of the formatting in the Z3200 text file. You can edit the formatting in Excel or other text editing app to get it to work. Primarily, the Z3200 adds data beyond the 730nm band that needs to be removed for i1Profiler to accept it. I don't know what your issues are with ICCGEN/Argyll, but chances are it's a similar file format irregularity that may need to be fixed in Excel, etc.

Note that I didn't change the name of your target when building the ICC profile, so the attached profile doesn't reveal any info about what printer/ink/media this ICC profile is for. If you simply edit the filename, the internal tag won't reflect your changes, and Photoshop displays only the internal tag not the filename, so your ICC profile is going to show up in the list with its original filename.

[NikonD850Boy]

Can't thank you and Mark enough!!! If it wasn't for you guys I would have discarded this machine!!!

Attached is a photo of my first print from the z3200 after color calibration etc

What do you guys think?

I must say though I spent half a roll of Canson Infinity Platine Fibre Rag to get to this stage lol Very expensive effort I must say :-)

[MHMG]

Looks like your Z3200 has taken flight again! I admire your persistence

Make just a few dozen 24x36 or larger prints to suit your image quality expectations, and your Z3200 investment will soon have earned its keep compared to jobbing the work out to a high quality lab.

cheers,
Mark
http://www.aardenburg-imaging.com

[kers]

I have attached an ICC profile made from your .txt file using BasicColor dropRGB software. DropRGB works seamlessly with the Z3200 generated .csv files, and it produces outstanding results. It doesn't get any easier than dropRGB to make high quality profiles. However, it's pretty expensive, so there's that.

Other software, for example, i1Profiler, doesn't like some of the formatting in the Z3200 text file. You can edit the formatting in Excel or other text editing app to get it to work. Primarily, the Z3200 adds data beyond the 730nm band that needs to be removed for i1Profiler to accept it. I don't know what your issues are with ICCGEN/Argyll, but chances are it's a similar file format irregularity that may need to be fixed in Excel, etc.

Note that I didn't change the name of your target when building the ICC profile, so the attached profile doesn't reveal any info about what printer/ink/media this ICC profile is for. If you simply edit the filename, the internal tag won't reflect your changes, and Photoshop displays only the internal tag not the filename, so your ICC profile is going to show up in the list with its original filename.

If i understand correctly- when having measured the colorspots you get a text-file with the measurements.

Now the next thing is to build a profile and if i understand that can be done in many ways.
Is 'Drop RGB' just one of those methods and produces good results?
I have a non postcript Z3100 and bought APS for profiling.
Could it be that 'Drop RGB' might give me a significant different icc than with APS?
Since i have never been very happy with the APS profiles i am interested.

[MHMG]

If i understand correctly- when having measured the colorspots you get a text-file with the measurements.

Now the next thing is to build a profile and if i understand that can be done in many ways.
Is 'Drop RGB' just one of those methods and produces good results?
I have a non postcript Z3100 and bought APS for profiling.
Could it be that 'Drop RGB' might give me a significant different icc than with APS?
Since i have never been very happy with the APS profiles i am interested.

You've asked one of those "it depends" kind of questions. Optimized ICC profiles start first and foremost with choosing best media settings on the printer, then printing and measuring a color chart (usually 900+ patches at a minimum) to get clean and verifiable measurement data (this often requires comparing two or more sets of measurements then eliminating fliers and averaging the good data sets together), finally running though a profiling software known to produce good profiles from good data sets. DropRGB is one app that does it with utter simplicity and has data optimization algorithms as well (presumably to detect and ignore fliers in the measurement data). i1Profiler, the older Profilemaker 5, Argyll, etc., also have a reputation for building great profiles as well, and no doubt others can be added to this incomplete list. HP's APS solution is now over a decade old, so may well go back to Profilemaker 4 days (before Xrite bought out Gretag Macbeth and Monaco). As such, it may or may not be as good as the state of the art today.

Given wise media settings and clean measurement data, the biggest difference in how an ICC profile is mapping the source image colors for any given printer/ink/media combination, IMO, tends to occur with apps that have proprietary mapping algorithms for the perceptual rendering tag, but even recol/wBPC Is not standardized to the point where the enduser won't find differences between the output from various profiling apps. As such, I don't endorse one app over another. I use more than one myself to build custom profiles, and I tend to think of the results as a suite of ICC profiles where dropRGB is my "go to" profile, but others may help to settle a problematic image file in a way that makes the final edits easier, say for example, banding in a blue sky gradient that shows up when using one profile or rendering intent but not another.

If you are comfortable with a command line interface, I'd download Argyll, and compare its output to what you are getting with APS. If you see useful differences, then welcome to the club of OCD printmakers who like to add various profiling flavors to their image editing toolkit

cheers,
Mark
http://www.aardenburg-imaging.com

[Pouln]

………...

Now it is time to create the profile. On the command prompt type:
colprof.exe -v -qh -S ProPhoto.icm -cmt -dpp FILENAME
………...

I know this is an old post and apologies for reactivating it, but I wanted to give Argyll a try.
I think it is an excellent guide and kudo's and thanks go to Geraldo.

Following it, all seems well.

However executing the above mentioned command in this form:
  colprof.exe -v -gh -S ProPhoto.icm -cmt -dpp profile_name
leads to the following error:
  Creating Gamut Mapping
     Finding Source Colorspace Perceptual Gamut
     Finding Source Colorspace Saturation Gamut
     Loading Image Source Gamut 'h'
  Input file 'h' error : Unable to open file 'h' for readingcolprof.exe: Error - Reading source gamut 'h' failed

Now, looking at the syntax there is no -gh parameter. There is a -g parameter, but it expects a subsequent src.gam filename that it can read (from ARgyll CMS help command):
-g src.gam      Use source image gamut as well for output profile gamut mapping

So my question is what to do with the -gh in Geraldo's post.

[Pouln]

Ahh, Tony noticed that the command contains a parameter -qh and I typed -gh.
That solved the issue.

Thanks a lot, Tony!

[Alan Goldhammer]

Following it, all seems well.

However executing the above mentioned command in this form:
  colprof.exe -v -gh -S ProPhoto.icm -cmt -dpp profile_name
leads to the following error:
  Creating Gamut Mapping
 

I am an Argyll user and would note that you should not use '-S ProPhoto.icm' in the colprof command as it will skew your profile and make your prints look a bit dull.  See Graeme's notes:  http://www.argyllcms.com/doc/colprof.html#S    I use AdobeRGB in this command.

[Pouln]

OK, good to know. Thanks Alan

[Kyle D Jackson]

Hi, necro an old topic... (but a key one!).  I don't see a "Color Measurement" menu option.  I am using HP Utility 1.17.0.3 (the latest, I believe) on Win10 x64, and I have a Z3100ps GP.  (I also have the APS but I never installed/used it, because my monitors have their own colour management options.) 

When I open Color Center > Paper Preset Management, the last option in the right-hand column is "Restore Factory ICC Profile". I don't have the "Color Management" option with the "i1 color" icon that is shown in the screenshot in Geraldo's original post.

Any ideas why I don't see this option? Is it somehow limited to the Z3200 series only and not avialable for the Z3100? Do I need to install the (outdated) APS package?

Thanks