i1 Profiler - grey ramp too dark

[gfsymon]

I'm using i1 Profiler for the first time, with an upgrade on a Rev D i1 Photo to profile a Z3100.

I have built a profile using the 2033 patch set and was impressed to see how well the colours match using the Color Checker Proof chart (the one with the holes in it) when laid on top of the printed test chart.  Even the red square which was marked as out-of-gamut in i1 Profiler, is actually pretty close (no doubt perfect with the Z3200 ... ho hum).  So colours are good.

However, I find that the gray ramp is increasingly too dark towards to black end.  I'm printing on Hahnemuhle Fine Art Baryta 325gm Glossy, using Gloss Enhancer on the Z3100 (image area only).  The black especially, is noticeably darker the the patch on the Checker chart.

Does anyone have any ideas of why this may be happening?

[gfsymon]

No one got any thoughts on this?

[digitaldog]

You may need to adjust the print settings (media settings), it may be the driver is laying down too much ink and the profile just can't adjust for that.

[gfsymon]

Hi Andrew and thanks for the reply.

Do you think that adjusting the Z3100 presets itself, for how much ink is being used is a good way to go?  I've never ever touched this setting with the Z series, since they are already so amazingly conservative with the inks.  (Guessing you're familiar with the Z series, since they haven't changed for rather a long time).

[digitaldog]

Do you think that adjusting the Z3100 presets itself, for how much ink is being used is a good way to go?  I've never ever touched this setting with the Z series, since they are already so amazingly conservative with the inks.  (Guessing you're familiar with the Z series, since they haven't changed for rather a long time).

Never used that printer. What you can do is download my Ink Density Test file from my site or build your own, basically a 21 step wedge and adjust until you get good separation in the tones. Then build a profile over that output behavior. If the behavior of the printer is so non linear that it lays down too much ink, there isn't much the profile can do to help here.

[gfsymon]

I found that somewhat counter-intuitively using a relative intent gives excellent results.  I say counter-intuitive because the profile was made as perceptual, but it would seem that when printing with these profiles, relative is more accurate ... certainly in terms of mid to darker grey/black.  Anyway, I'm now very happy with the profiles from the i1Profiler app.

[Scott Martin]

However, I find that the gray ramp is increasingly too dark towards to black end.  I'm printing on Hahnemuhle Fine Art Baryta 325gm Glossy, using Gloss Enhancer on the Z3100 (image area only).  The black especially, is noticeably darker the the patch on the Checker chart.

That's totally normal. The black and white points are completely relative to each paper's capabilities. If you print on Hahnemuhle Fine Art Pearl you'll see a lighter gray ramp that's scaling ot that paper's lesser DMax. If you print again on a matte surface cotton rag paper like Photo Rag you'll seen an even lighter black and gray ramp.

In each of these cases you're just seeing the inherent differences in the media and not a problem with the profiling process.

Interestingly, Apple's glass surfaced displays (like the MacBookPro's Retina displays) also have richer blacks that do a better job at simulating Baryta's blacks than a matte surface display. Viewing on a matte surface display and printing on a Baryta paper can be inherently problematic....

[digitaldog]

I found that somewhat counter-intuitively using a relative intent gives excellent results.  I say counter-intuitive because the profile was made as perceptual, but it would seem that when printing with these profiles, relative is more accurate ... certainly in terms of mid to darker grey/black.  Anyway, I'm now very happy with the profiles from the i1Profiler app.

I'm not sure what you mean by the profile was made as perceptual. There are options to alter that table. And there are sometimes options to pick that table as preferred, in workflows where one can't select an intent so the preferred intent is used. Otherwise, you really want to view images with both intents via a soft proof and select the one that produces the preferred rendering. This is somewhat subjective. ICC profiles don't know anything about images and how you would prefer one rendering over another. They just understand lots of solid pixels of fixed numeric value. The correct rendering intent to use is the one you prefer visually on an image by image basis.

[gfsymon]

I'm not sure what you mean by the profile was made as perceptual. There are options to alter that table. And there are sometimes options to pick that table as preferred, in workflows where one can't select an intent so the preferred intent is used.

When I'm profiling the Z3100, I'm selecting it as an RGB printer.  Are you suggesting I profile as CMYK?  When RGB is selected, there is no longer a choice of intent.

[gfsymon]

That's totally normal. The black and white points are completely relative to each paper's capabilities. If you print on Hahnemuhle Fine Art Pearl you'll see a lighter gray ramp that's scaling ot that paper's lesser DMax. If you print again on a matte surface cotton rag paper like Photo Rag you'll seen an even lighter black and gray ramp.

But isn't this the whole point of profiling?  The device and software recreate what you have on your monitor.  I admit my monitor raises the bar pretty high with colours beyond AdobeRGB (HP LP2480) but the Z series are no slouches when it comes to extended colour gamuts (e.g. greens beyond AdobeRGB).

The only issue I'm having is the deep greys/black, with not enough separation.  Printing with a relative intent totally fixes this.  I just wish I understood more, why perceptual wouldn't get this right?

[Scott Martin]

When I'm profiling the Z3100, I'm selecting it as an RGB printer.  Are you suggesting I profile as CMYK?  When RGB is selected, there is no longer a choice of intent.

When you're using a driver you must profile as RGB. When you're printing the profiling target there's no conversion and thus no rendering intent. Only when you're printing *with a profile* can you choose a rendering intent. So you've profiled your printer and are comparing the Perceptual and Relative Colorimetric intents with that profile.

[Scott Martin]

But isn't this the whole point of profiling? 

No it's not. It's more complicated than that.

The device and software recreate what you have on your monitor.

No - that's an over simplification. The whole idea is to have predictable color that optimizes the quality on each device. There's always inherent differences and limitations to work within - which you're seeing. In this case getting an excellent soft proof requires calibrating your display for matching black and white points and enabling the soft proofing mode in PS or LR. Doesn't sound like you're there yet.

The only issue I'm having is the deep greys/black, with not enough separation.  Printing with a relative intent totally fixes this.  I just wish I understood more, why perceptual wouldn't get this right?

That doesn't sound right - Perceptual should give you excellent blacks/shadows that's not that different from RelCol with BPC enabled.

It's hard to troubleshoot this without seeing your prints and knowing more about how your profile was made. Can you elaborate more on all these details? Are you increasing the Perceptual rending contrast when making the profile? Are you printing RelCol with or without BPC? Etc, etc.

[gfsymon]

No it's not. It's more complicated than that.

No - that's an over simplification. The whole idea is to have predictable color that optimizes the quality on each device. There's always inherent differences and limitations to work within - which you're seeing. In this case getting an excellent soft proof requires calibrating your display for matching black and white points and enabling the soft proofing mode in PS or LR. Doesn't sound like you're there yet.

Well ... perhaps I could have used the word 'goal' instead of 'point'.

The thing is, the software reads the printed charts and should be calculating from there, exactly how to produce what is being seen on the monitor, which it has already profiled.  That's what it's trying to do.  Mid to dark grey and black are not outside the monitor's or the paper/printer's gamuts.

If a profiling software can't manage it first time around, I wonder why they don't let you re-print the charts, this time using the profiles it has just created and then from this new data, create an even more close interpretation (profile).  That would seem like such an obvious way to refine a profile.

That doesn't sound right - Perceptual should give you excellent blacks/shadows that's not that different from RelCol with BPC enabled.

Precisely what I think.  Nevertheless there is a significant difference.

It's hard to troubleshoot this without seeing your prints and knowing more about how your profile was made. Can you elaborate more on all these details? Are you increasing the Perceptual rending contrast when making the profile? Are you printing RelCol with or without BPC? Etc, etc.

I made the profiles with i1Profiler 1.4.2 and an i1Pro Photo. 16bit profiles, larger size.  icc v2 (v4 doesn't work on the Z3100).  I'm not touching anything to do with tone/colour.

[Rhossydd]

exactly how to produce what is being seen on the monitor

You keep using terms like this which suggest you lack an understanding of the whole colour management process.

Very simply put;
The printer profiling software doesn't know what monitor you're using or how it's set up. All it does it is try to reproduce the IMAGE data as best it can on the paper.

A monitor profile will just try to reproduce the IMAGE data on screen as best it can.

The two are different and you'll need to appreciate that to understand the CM process properly.

[Ernst Dinkla]

If one aims at reproducing the Color Checker Proof Chart that way, the choice of one of the Colormetric renderings should be the first decision in my opinion. The profiles describe the gamuts of the individual devices, the rendering choices how each gamut/color space is translated in the other one. With that the dynamic range. The last must be  wider in the paper print.

That the shadows-blacks are not separated enough with the Perceptual rendering despite a deeper black in the print could have two causes: first, your viewing light is not giving enough light for your eyes to make that separation and second, Perceptual rendering is still a manufacturer's solution. That extra light will make the divide between the print and the chart even wider though. The OBA content of the Fine Art Baryta is probably higher than what is in the chart base so any viewing light with UV content will throw in another spoke.

If you want to compare results with other sources:
Babelcolor has some information on the chart's colors, averages and deviations (whites too), and a variety of CC image files in different color spaces:
http://www.babelcolor.com/main_level/ColorChecker.htm#ColorChecker_charts_images
Hahnemühle must have the media preset + profiles for that paper and the Z3100.

There's no mention of the OS you use which could also be important when you do not use the HP Z integrated profile creation system but make profiles through another path.

--
Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
July 2013, 500+ inkjet media white spectral plots.

[gfsymon]

The printer profiling software doesn't know what monitor you're using or how it's set up.

It should. i1Profiler has produced the monitor profile and the printer profile.  Doesn't the monitor profile contain the information that is required when translating colour to a printer's profile?

[Rhossydd]

It should. i1Profiler has produced the monitor profile and the printer profile.  Doesn't the monitor profile contain the information that is required when translating colour to a printer's profile?

No, not at all. There's no inter-relation between the printer profile and monitor profile, they independent.

[gfsymon]

Thanks Ernst.  This makes good sense.

That the shadows-blacks are not separated enough with the Perceptual rendering despite a deeper black in the print could have two causes: first, your viewing light is not giving enough light for your eyes to make that separation and second,

Viewing is good and ambient light around the monitor corresponds to paper white (one of my pet things about too deep blacks).  I'm not concerned about the depth of the black on Bartya325 ... I like it ... I'm concerned about the tonal separation in greys approaching the Baryta325's achievable black.

Perceptual rendering is still a manufacturer's solution.

This is a very good point. 

That extra light will make the divide between the print and the chart even wider though. The OBA content of the Fine Art Baryta is probably higher than what is in the chart base so any viewing light with UV content will throw in another spoke.

I'm using an i1Pro-non-UVcut.

If you want to compare results with other sources:
Babelcolor has some information on the chart's colors, averages and deviations (whites too), and a variety of CC image files in different color spaces:
http://www.babelcolor.com/main_level/ColorChecker.htm#ColorChecker_charts_images
Hahnemühle must have the media preset + profiles for that paper and the Z3100.

Thanks for this.  Amazingly the Hahnemühle site's profiles are a dead link.  I'm actually running the developer release of 10.9, but it's not different to 10.8x.

[gfsymon]

No, not at all. There's no inter-relation between the printer profile and monitor profile, they independent.

But on a Mac they're being linked together by ColorSync at the OS level using the ICC system developed/put in place by Apple in the early 90s.  That's what ColorSync does.  (Don't know what MS calls its colour management).

Maybe we're talking at cross purposes?

[Rhossydd]

Maybe we're talking at cross purposes?

You need to grasp they're independent of each other totally.

Give it some thought; How could manufacturers provide printer profiles if they were linked to user's displays ? How could Epson know the characteristics of anyone's monitor ?