Softproofing not accurately reflecting out of gamut reds

[Ryan Mack]

I'm trying to print some landscape images with some dark saturated reds. It's not a surprise that they're not printing correctly. But what I don't understand is why soft proofing still shows them staying a lot more red than they actually print. Luminosity is bang on, saturation is all off. I *believe* I've got everything set up correctly although it's a new machine with new software so I may have messed something up. Display is a NEC PA271W-BK-SV calibrated for prepress with BasICColor Display using the monitor bundled NEC SpectraSensor Pro. I'm using ImagePrint to print to a 3880 on Harman by Hahnemuhle Gloss Baryta 320, profiles via ImagePrint.

[Mark D Segal]

It could be a problem with the profile. I suggest you ask ColorByte to examine the profile, or recommend another one that you could test with for verification.

[Mark D Segal]

Also, use another paper that requires all the same settings except for the profile and using its profile see whether that disconnect persists. If it does, suggests a settings issue; if it doesn't, suggests a profile issue.

[Ryan Mack]

Printing on another paper with only a different profile looks pretty much the same. That leads me to believe I've chosen the incorrect settings when calibrating my displays, but I'm a bit confused by that. The only real difference between various options when running basICColor display is the color temp for the display, and the options are basically 5000, 5500, or 6500. The documentation suggests 5000 for prepress. Regardless of that choice, shouldn't the soft proofing color management compensate for the display's temperature? Or does the display color temperature basically declare "this is the color of light I'm going to be viewing the print under" in which case the soft proofing would allow the display's color temperature to have an effect on the image?

[Mark D Segal]

For profiling your display the calibration settings before you profile should allow you to set White Point, Luminance and Gamma (at least). Once you set that calibration, you profile the display and always use that profile with those calibration settings. As long as you don't touch the display controls, the calibration settings remain the same regardless of the profile (parhaps a bit of drift over a long period of time). The softproof, through the lens of the display as influenced by its calibration, shows you the combined effect of your edits using the active paper/printer profile.

[Doug Gray]

Printing on another paper with only a different profile looks pretty much the same. That leads me to believe I've chosen the incorrect settings when calibrating my displays, but I'm a bit confused by that. The only real difference between various options when running basICColor display is the color temp for the display, and the options are basically 5000, 5500, or 6500. The documentation suggests 5000 for prepress. Regardless of that choice, shouldn't the soft proofing color management compensate for the display's temperature? Or does the display color temperature basically declare "this is the color of light I'm going to be viewing the print under" in which case the soft proofing would allow the display's color temperature to have an effect on the image?

Separate the monitor color management from the printer. Trying to match a printer and monitor is non-trivial even if both the printer and monitor color management is correct.

Grab an image of a colorchecker from a good source like Babelcolor. Then print it using Absolute Colorimetric. It should very closely match an actual colorchecker when both are viewed side by side in daylight. This is by far the easiest way to identify the printer and profile color management path as problematic.

[Ryan Mack]

Thanks all for the advice. I printed some color check images and found that even there the reds were too orange. A bit more inspection led to me discover my PK and vivid magenta ink cartridges were 4 years past their expiration date. Replacing them, shaking all the cartridges, and running a cleaning cycle significantly improved the situation!

[Mark D Segal]

Thanks all for the advice. I printed some color check images and found that even there the reds were too orange. A bit more inspection led to me discover my PK and vivid magenta ink cartridges were 4 years past their expiration date. Replacing them, shaking all the cartridges, and running a cleaning cycle significantly improved the situation!

It could take quite a while for the expired ink to completely run through the channels and draw only on the new ink. So if this is the cause, the quality should show more improvement over time.

[Ryan Mack]

It's me again. I printed 3 14x21 prints today and a few smaller proofs so I imagine the ink should be flushed now. I tried printing the colorpicker (absolute colorimetric) to see where things stood. The top (pastel) and bottom (grays) look great. And about half of the saturated colors look great too. There is still an issue with red, pink, purple, orange not being as vivid as the pre-printed colorchecker. The solid red (third row, third col) in particular is coming out a bit dark and unsaturated. I suspect this is the natural gamut limitation of the paper and inks, but I thought I'd see if one should expect a good match for those colors.

Thanks again, Ryan

[Doug Gray]

It's me again. I printed 3 14x21 prints today and a few smaller proofs so I imagine the ink should be flushed now. I tried printing the colorpicker (absolute colorimetric) to see where things stood. The top (pastel) and bottom (grays) look great. And about half of the saturated colors look great too. There is still an issue with red, pink, purple, orange not being as vivid as the pre-printed colorchecker. The solid red (third row, third col) in particular is coming out a bit dark and unsaturated. I suspect this is the natural gamut limitation of the paper and inks, but I thought I'd see if one should expect a good match for those colors.

Thanks again, Ryan

I'm not sure what you mean by "pre-printed colorchecker."  Actual X-Rite Colorcheckers are not printed. They are made with specialized ink pigments for each patch on a matte finish which makes them more consistent in production.

Presuming you are referring to a physical Colorchecker, and that you printed a high quality image from Babelcolor using Absolute Colorimetric Intent then yes, almost* all printers are capable of printing all the patches with visually close matches. However, you have to compare apples to apples so the print and the Colorchecker should be viewed in daylight. It is also important to use low OBA paper because Colorcheckers have virtually no fluorescence.

Assuming these factors have been addressed, then it looks like the printer profile does not match what it is capable of. Perhaps due to wear, production variation, ink age. No way to really know. This uncertainty is why I long ago got a spectrophotometer so I could actually measure things to make sure my printing was consistent and make custom profiles when the OEM ones were not sufficiently close.

* If, in fact your printer can't produce the same saturated red as on a Colorchecker then this should be evident in Softproof as the profile will report back what color is actually printed for those that are out of gamut.

Also, perceptually, reds are very sensitive to luminance and visually reds will appear much darker and less saturated when viewed in light where the reflected luminance is less than emissive luminance on a monitor. Which is why one should compare the print and a colorchecker together in the same light, preferably daylight.

[Mark D Segal]

All good advice and as you probably know from my various review articles on printers and papers why I routinely use the i1Pro2 and a spreadsheet for results evaluation of the print pipeline. But of course not everyone is equipped with the kind of gear to do this, so at least visual side by side inspection helps to see the most egregious variances.

The two most important limitations of this approach with the ColorChecker is that its maximum Black is only about L*20, so evaluating the B&W luminance range from say 1~20 needs a separate set of patches, easy enough to construct in Photoshop, but here the visual evaluation of the deep quarter-tones can be really tough, so an area where measurement is all the more useful. The other limitation in respect of the colours in the GMCC is that their gamut doesn't really challenge our latest generation of gloss/luster papers, inks and printers. One needs another kind of target to challenge the limits of these materials and equipment. At the same time, one wants to be careful not to print much for testing that is out of gamut, because that introduces the whole business of how OOG colours (and in Perceptual Intent, others) get shifted, which is not normally the purpose of these tests.

Finally, a "mental factor" - one shouldn't get hung up on small variances that for most intents and purposes one wouldn't perceive on a stand-alone basis in a normal photographic print. There are tolerances, unless one requires pretty close accuracy for special purposes.

[Ryan Mack]

Finally figured out the second issue (after the stale ink problem). Turns out I was misinterpreting the color emulation setting as soft proofing settings in ImagePrint. So I had incorrectly set it to try to emulate the look of a different paper when printing. This was resulting in some color swatches being darker and less saturated than the actual X-Rite ColorChecker. Turns out RTFM is still good advice :-) So I'm all sorted now. Thank you all for your help!