Pages: [1]   Go Down

Author Topic: To a (for me) new B&W workflow, small gamut CM shell added  (Read 3899 times)

Ernst Dinkla

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4005
To a (for me) new B&W workflow, small gamut CM shell added
« on: February 22, 2015, 09:10:37 am »

In theory I would go for a more abstract concept with a small gamut color space to assign to "RGB" color toned B&W images, based on eciRGB with the "L" tone curve and for greyscale images I would use the existing QTR Lab greyscale space. Practice with my printers etc says an sRGB color space will be easier, gamut small enough, tone curve quite perceptual, compatibility alright besides the D65/D50 issue. I made an sGRAY space for greyscale images to keep the relation correct.

For B&W printing through a driver's color mode (10-11 inks) and with driver's color management on I like to add an extra profile in the application color management for a more precise perceptual tone range + more precision in the color toning of the B&W prints. Kind of dual profiling on purpose.

The route I have gone so far is using sRGB as a smaller gamut color space to get more precision on the colors in the conversion. The printer is the HP Z3200PS with HP PLC3 Windows driver. The driver's color management, when ON, expects (in default) sRGB. For target printing I use sRGB (perceptual rendering) for the printer profile in Qimage Ultimate. So in this route no profile is active but the driver color management LUT for the paper specified in the driver media preset. The printer is first color calibrated with the normal HP Z3200 calibration tool/target.

There are some reasons why I want to do this. One of them; I discovered that the B&W Dmax is as good (high) using driver's color mode + driver's color management as it is when I use driver's B&W mode + driver's color management (HP's recommendation for B&W printing). With application color management ON (driver CM off) + ICC profiling there always seems to  be a loss on the Dmax.

Going this route I printed a QTR linearisation target (2x34 random steps in my case), measured it and created a QTR B&W (RGB kind) profile. That profile replaced the sRGB printer profile in Qimage Ultimate CM for actual printing, perceptual rendering kept. The next 34 step target printed through that path showed an improved perceptual tone range and no loss on the Dmax compared to the first target print and to the best HP calibration black patch.

QTR's profiles do not address more precise color control in the print. In the profile creation I can see the Lab a and b numbers of the target patches and no real a b changes happen between the two target's measurements. No surprise as the driver's color management is only in control of that and acts the same in both print cases.

I would like to create a small gamut target for a small gamut printer profile that is sufficient to color manage color toned and neutral B&W images to prints. Based on my paper white spectral measurements and measurements on a variety of color toned prints, -sepia-cold-brown-split-tone, target patches with Lab values not going beyond a b -11 + 11 from the neuiral spine would be enough. I like to keep the neutral patches on 34 steps, not more, not less. 16 bit depth preferred.

In ArgyllCMS's Targen, what are the settings for creating an RGB-device printer profile target which gives me that added precision on the tone range and color toning? What are the settings for the creation of that profile, measuring settings, white point, etc? This is a kind of last minute iterative profiling in the workflow and should not be different from the use of the iterative profile enhancement step on the near neutrals as already possible in color profile creation. The rendering will stay on perceptual. Measuring device an Eye1 Pro UV enabled.

Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2014 update, 700+ inkjet media white spectral plots

« Last Edit: February 22, 2015, 09:53:44 am by Ernst Dinkla »
Logged

Ernst Dinkla

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4005
Re: To a (for me) new B&W workflow, small gamut CM shell added
« Reply #2 on: February 22, 2015, 11:08:53 am »

Thank you, not to steep a slope for me to get there.

Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2014 update, 700+ inkjet media white spectral plots


 
Logged

Ernst Dinkla

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4005
Re: To a (for me) new B&W workflow, small gamut CM shell added
« Reply #3 on: April 25, 2015, 10:13:49 am »

Follow up on this alternative B&W printing route.

Using the HP Z3200 or Z3100 driver color mode with printer color management on and Qimage Ultimate color management also on. As explained in the first message. After the paper color calibration with the integrated spectrometer the perceptual linearity of the prints made in either Color or B&W mode is already quite good using driver color management.

The Zs use only their 3 or 4 grey inks (so sans color inks) when Greyscale or R=G=B B&W images or image areas are present. So it is possible to use either the B&W or Color printing mode for printing B&W and still use monochrome inks (Epson ABW mode doesn't do that normally). With printer CM on for B&W printing (as HP recommends) there is no loss of Dmax. So I use the color mode as it allows me next to an improved perceptual linearisation with a QTR profile in Qimage's CM, to adjust the Lab ab deviations just enough with Photoshop ACV curves added to the QTR (RGB type) profiles. I thought this would not work but it does and the 'manual' tweaking is precise. For glossy prints hardly needed though as the three Z Vivera monochrome inks used then are neutral by themselves and the color toning runs roughly from paper white to the neutral black quite straight (the last shifting a bit on the underlying paper white too). With matte papers the Z Vivera MK black ink is warmer than the 3 to 2 grey inks below it, so depending on the paper white adjusting the Lab ab values can make the color toning more consistent from paper white to black. Taking out the PK ink by selecting another Z matte paper media preset allows the MK black ink to cover more of the tone range downwards, suits warm papers better.

The double profiling used could have been done with a kind of iterative greyscale biased ICC color profile too in Qimage Ultimate's CM (my original intention) but I found it worked better with a QTR (RGB) created profile (+ Lab ab tweaked ACV curves embedded in that profile) in QU's CM . Some of this B&W profiling can actually be done with the Z's integrated spectrometer (made some greyscale targets) but I prefer my UV enabled Eye1 Pro as the target is small. Over the last weeks I switched from Measure Tool to ColorPort to Patchtool. The last integrates well with QTR B&W ICC profile creation and has nicer features for building targets. ArgyllCMS's greyscale (R=G=B) patches randomising a first step in that process.

Instead of the additional color toning possible in the printer driver for B&W I have made some color curves in Qimage Ultimate that I can assign to the B&W image in QU or just use as an on the fly printer filter. Not only to RGB images, even to a Greyscale image (with a trick as Mike has taken out color adjustments etc. on Greyscale images). Selecting the right paper white papers for warm or cool prints etc with the aim of no color adjustment on the black keeps the Dmax uncompromised. Applying a suitable Qimage color tone curve then to that paper choice is less harsh on the linearity than forcing a wrong paper white through heavy color curves. The original Greyscale image file is not corrupted, is small and if 16 bit will compress well. Qimage Ultimate's interactive logfile will recall the job settings for both QU and the driver so consistency in time is also guaranteed. The HP Z's integrated calibration tools another base for that.


Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2014 update, 700+ inkjet media white spectral plots
« Last Edit: April 25, 2015, 10:17:28 am by Ernst Dinkla »
Logged
Pages: [1]   Go Up