Z3200 - Better ICC Color Profiles

[Ethan Hansen]

Ethan, it's good to know that item 7 was the winner because I don't see any way using i1Profller to do what you did in item 6, i.e., save out the target image as 16-bit. It seems to want to generate 8-bit only. Am I missing a preference setting somewhere in i1P, or did you use another piece of software to generate a 16-bit target image from the 16-bit reference file?

You're right - i1Profiler won't make 16 bit images. I created the 16-bit images manually using ImageMagick scripts.

[Doug Gray]

Regarding i1Profiler data truncation: The last time I thoroughly vetted the consequences of using fractional data in i1Profiler was with version 1.63 (late 2015 vintage). I dug through my Excel archives and here were the conclusions.

I ran several cases:

• The default i1P 8-bit target image with RGB values rounded to the nearest integer paired with the 16-bit reference. This is the default i1Profiler configuration.
• Default 8-bit i1P target image with reference file truncated to 8-bits (i.e. values truncated down to the nearest integer).
• Default 8-bit i1P target image with reference file rounded to the nearest integer.
• 16 bit target image, 8-bit reference file rounded to the nearest integer value.
• 16 bit target image, reference file truncated to 8-bits.
• A true 16 bit target image having the fractional values in the measurement reference file (target image 16-bit, reference file 16-bit).
• A version with both image and reference truncated to 8 bits.

Options #1, #2, and #3 all made similar but not quite identical profiles, although the differences barely exceeded the measurement tolerances of an iSis. The target images were the same for all three tests. If i1Profiler truncates or rounds 16-bit reference values prior to the profile calculations the profiles should differ. The differences were statistically significant at only a 90% confidence level; these results don't definitively prove i1Profiler makes use of 16-bit reference data. A simple statement by X-Rite would be greatly appreciated.

Options #4 and #5 produced profiles that were statistically different at a 95% confidence level both from each other and from the previous three tests. Having different target image values ended up producing a larger effect than fractional differences in reference files.

Finally, options 6 and 7 - where the target image and reference files should be matched exactly - gave profiles with outputs with no statistically significant differences. This again hints that i1P makes use of 16-bit reference values.

I can confirm that I1P uses fractional RGB values when generating a profile. A side effect is being able to create an 8 bit RGB file with truncated fractionals.

Here's the process:

• Generate a patch set with I1P using "Smart Patch Generator"
• Print the chart and make a profile using the normal process. The printed chart will have rounded RGB values but the profile will be made with the 16 bit RGB values.
• Now, drag the newly generated profile into the I1P "Measurement" tab. It will populate the RGB and spectral data that was saved in the profile. This data doesn't round but truncates the RGB values to 8 bits.
• Make a new profile. The new profile will be made from the truncated RGB data. I then made a histogram of the dE00 of 10M random RGB values in device space converted to LAB for both profiles through the AtoB1 tables. Ave. dE00 was .15, with 99% falling below .3.  Attached is the histogram.

One can then click on "Patch Set" and save the 8 bit values. Then any profiles made with that saved patch set will produce matching tif and/or print files and I1P will use the proper RGB values.

We went on to quantify the profile accuracy by printing a target through each profile and measuring the resulting colors vs. predicted values. The winner was the final option above, with matched 8-bit target image and reference file. Average dE2000 was 0.68 on a Fuji Frontier contone printer and 0.91 on an Epson inkjet. The other options gave Fuji errors in the 0.95 - 1.32 range and Epson errors of 1.18 - 1.67 dE2000.

The larger question is whether these mathematical gyrations made and perceptible difference in the final print. For the most part the answer was no on both printers. The exception was in B&W images where the clear winner again, particularly on the inkjet, was the profile made with matched 8-bit target and reference. Highlight transitions had visibly fewer crossovers and smoother transitions. The matched 16-bit image/reference profiles made prints that some of us thought looked better than from options 1-5 while the rest couldn't see any difference.

I concur. Your observations are in complete accord with what I've seen though I've been able to achieve ave dE00s of better than .6 on my 9800 with patch sets of 2553 and higher. Much of the improvement comes from redundant patches for colors that show significant change when printed in different locations on the paper.

ADDED:  Turns out it's even easier to get rid of fractional values from the Smart Patch Generator. Just save the patch set. Then reload the same patch set. Voila! the fractional values are gone!

[Doug Gray]

You're right - i1Profiler won't make 16 bit images. I created the 16-bit images manually using ImageMagick scripts.

I wrote a Matlab function to create 16 bit Isis target files. It also refines the black "diamond" registration marks which are rather crude being composed of .25mm squares and can be selected for either 600DPI or 720DPI to conform with the printer's native resolution. It also prints a time stamped title with relevant info which makes it easier for me to keep track of stuff.

[MHMG]

ADDED:  Turns out it's even easier to get rid of fractional values from the Smart Patch Generator. Just save the patch set. Then reload the same patch set. Voila! the fractional values are gone!

Cool trick! Who woulda thunk it? But it seems to work

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

[Ernst Dinkla]

Is Babel Color's Patchtool offering solutions for making the Color Center <> Profile Creators linking easier?
I bought it for other purposes but did not use it enough so far.


Met vriendelijke groet, Ernst

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

[Brad P]

Is Babel Color's Patchtool offering solutions for making the Color Center <> Profile Creators linking easier?

Interesting question.  If it does generate reliably uploadable/printable CGATS text files (and maybe smoothed RGB triplet data), I’d spent the USD 125 to buy it in a heartbeat.  I just downloaded a demo version, but the demo version doesn’t export CGATS text files, so hopefully others can test it.  But since everyone has had reliability issues uploading the differently formatted CGATS text files so far, I’m a hopeful skeptic.

I did the unthinkable and read the Z’s manual.  It cryptically states that it is compatible with CGATS and MonacoPROFILER inputs. The latter was acquired by X-Rite some time ago and discontinued, but people with that may have some success.

[Ernst Dinkla]

Interesting question.  If it does generate reliably uploadable/printable CGATS text files (and maybe smoothed RGB triplet data), I’d spent the USD 125 to buy it in a heartbeat.  I just downloaded a demo version, but the demo version doesn’t export CGATS text files, so hopefully others can test it.  But since everyone has had reliability issues uploading the differently formatted CGATS text files so far, I’m a hopeful skeptic.

I did the unthinkable and read the Z’s manual.  It cryptically states that it is compatible with CGATS and MonacoPROFILER inputs. The latter was acquired by X-Rite some time ago and discontinued, but people with that may have some success.

AFAIK MonacoProfiler needs older versions of Windows to run. Doable with W7 and XP emulated there but it is getting complicated that way. Beyond W7 that solution is lost. So similar to Kodak Colorflow for profile editing.

Edit: Danny Pascale of Babelcolor has been approachable for questions in the past and used to write on the Colorsync list. We could ask him whether there is a compatible export choice in Patchcolor.
Meanwhile;
http://www.babelcolor.com/patchtool_specifications.htm


Met vriendelijke groet, Ernst

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

[Mark Lindquist]

Actually, I think MHMG and I are getting close.  Understanding that the issue with the file being rejected by the ESP is due to formatting issues in delimited text files which Excel modifies automatically is partly in large part, the problem.  The i1 profile ESP in the Z3200ps printer reqires a specific format for the data and just balks at anything not conforming. Having come to an understanding regarding rounding/truncating files to be certain of 8 bit accuracy, potentially, with certain formulas applied and importing the files corectly in Excell, specific mixes of special files can be successfully exported that the Z will, in theory accept.  MHMG is working on a few today, but so far there is no “single magic bullet”.  Our approach at this point is to begin creating special CGATS custom files, test for certain they work flawlessly, then ultimately publish them on Z3200.com for all to access.  I’m hesitant to introduce Monaco/Gretag archaic tech into the present mix, as it is known that i1 profile is the specific tech that the Z3200ps currently plays nice with, as verified by MHMG.  After a think-tank session last night, three of us MHMG/ML/Mrs. ML)) came to certain conclusions that should lead to success.  Not saying exactly what, yet, to prevent further muddying of the waters.
So now, altough pressing on full steam ahead, we’re slowing down to concentrate on this specific issue of proper formatting and MHMG will be working on adding more neutral and low frequency chroma near-neutral ramps for smoothing.  It is our conclusion that offering the optimized files is in the end a far more practical solution than offering a convoluted process that most will not attempt to use, although that info will also be made available.  Just an update on what our (MHMG/ML) direction currently is. 
Best -
Mark L

[John Nollendorfs]

Mark et. al:
Got to hand  it to you guys for persistance in potentially cracking this nut. But is a 4000 patch file profile significantliy better? Just asking. It might be worthwhile buying the Z3200 just for the automated feature of generating huge patch  profiles. ;-)

John Nollendorfs

[Mark Lindquist]

Mark et. al:
Got to hand  it to you guys for persistance in potentially cracking this nut. But is a 4000 patch file profile significantliy better? Just asking. It might be worthwhile buying the Z3200 just for the automated feature of generating huge patch  profiles. ;-)

John Nollendorfs

Remains to be seen John.  That's something that MHMG will be able to address and it's on the program. 
As far as worth the price of the Embedded Spectrophotometer being worth it, I figured something out last night that is really interesting.

Looking down the road - I hope WAAAAAY down the road, let's say HP discontinues the Z3200ps  (hopefully to be replaced by something better).  And there's no ink.
Printer becomes door stop, yacht anchor, etc., right?

NOPE!

Actually, as long as the belt and Spectrophotometer are still good, even at the point that the printer can't print, it can still be used as a stand alone Automated Spectrophotometer.
It can generate a TIFF file to be used to print a color target print on another printer, then it can read scan and measure the chart automagically then create a .TXT file for use to be able to convert to ICC profile.  It should also be able to install the profiles right in your library automagically.

So all will not be lost even at the end.  Potentially lots of use out of this wonder.

So it is worthwhile, I believe.

Best -

Mark

[Mark Lindquist]

Is Babel Color's Patchtool offering solutions for making the Color Center <> Profile Creators linking easier?
I bought it for other purposes but did not use it enough so far.


Met vriendelijke groet, Ernst

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

On it Ernst.  Thank you.

Mark

[Doug Gray]

I've used patchtool to average CGATs spectral files and they load into I1Profiler. Can't say about uses for the Z3200 tools. One thing you have to do AFTER averaging is to replace the RGB data with the original as patchtool will replace the RGB with sRGB. There is a separate menu option to for that. One of the nice things about the average CGATS option is that you can select outlier threshold to ignore values exceeding a deviation you specify. That requires at least 3 scan files of course.

[MHMG]

The printing bug on Mac version of HP's color utility for the Z3200 persists, but the effective work-around is to export the HP color target as a tiff file and "print as color target" with Apple ColorSync utility (ACSU) at 100% magnification. I created a 4357 patch target with i1Profiler, then using Doug Gray's observation on how to strip fractional decimal points by reopening and re-saving the reference file in i1Profiler, I then imported that chart file into the HP color utility, and lastly, I then exported a tiff file from HP's color utility based on that reference file.

I then printed this tiff file on my Z3200 using a 44 inch wide roll of Moab Entrada Natural. The baseline media setting was "Fine Art paper (more ink)" Note that this setting invokes a rather unique HP Z series "quad tone" gray ramp which exploits both matte black and photo black inks in the printed output.  The final print was 34 x44 inches, and I successfully measured the 4357 patches twice in a row, the first measurement taking place about an hour after printing, the second measurement at approximately 24 hours after printing. I then dropped back to PM5 measure tool, to compare and average those two files. Average dE error was .62 with largest single patch error of dE=2.12 (some of this error may be attributed to dry down effects). I consider the repeatability of the measurements over 4357 patches to be quite good.

The averaged measurement file was then imported into i1Profiler. I set I1 profiler settings on large LUT size, 16 bit, with default contrast and saturation sliders, and "neutralize gray" slide set to 100. Also, I boosted the "smoothing" slider to 75. After the profile was made, I next printed the Aardenburg 60 patch QC target using perceptual rendering, and measured the result (see attached graphs). The tone curve shows that i1Profiler's default contrast setting adds a subtle S-shaved curve to increase midtone contrast back to gamma=1, which indeed differs from relcol/wBPC in a desirable way. The grayscale neutrality plot is extraordinary!  It is the smoothest grayscale neutrality graph I have ever seen printed on an inkjet printer. It matches the very gentlest and smoothest grayscale neutrality curves I have previously measured on traditional silver gelatin print processes.

Mark L. and I will soon write up a tutorial that covers the steps involved and make it available at Z3200.com, but suffice to say, the HPZ3200's automatic scanning spectrophotometer enabled us to reach new heights in profile quality, that we probably would not have attempted otherwise

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