Pages: [1]   Go Down

Author Topic: Softproofing: Differences in Photoshop, it's Print preview and Print-Tool  (Read 486 times)

MichaelKoerner

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 65
    • www.m-werkstatt.at

Please help me finding the reason for inconsistent softproofing in Photoshop's Softproof, Photoshop's Print Preview and Roy Harrington's Print-Tool of a rather blueish image.

Attached "Screenshot 1" shows huge differences between Photoshop's Softproof and it's Print preview, especially in the dark areas of the wood.
Attached "Screenshot 2" shows even bigger differences between Photoshop's Softproof and Print-Tool's, concerning hue/saturation also.

The final print (out of Photoshop) falls between Print-Tool's and Photoshop's Print preview, but is far away of Photoshop's Softproof - which I would like to use primarily, of course...

As I understand, images with a strong cyan/blueish cast can be problematic for color managed workflows due to LABs insufficiencies in this area (and manufacturers diverse workarounds to compensate). BUT: Is there a way to produce better Softproofs in Photoshop? Perhaps via profile editing?

Thanks in advance, Michael

PS: I use Photoshop 21.0.2 on Catalina 10.15.2, an Epson SC-P800 with current drivers and Print-Tool (2.1.1).

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 16261
    • http://www.digitaldog.net/

Simple. The soft proof in Photoshops PRINT dialog is buggy, has been for awhile. Ignore it and soft proof using the Customize Proof Setup.
Logged
Andrew Rodney
Author ďColor Management for Photographers"

MichaelKoerner

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 65
    • www.m-werkstatt.at

Andrew,

thank you, that was the result of my research I did before posting. But the resulting print matches PRINT dialog better than Customize Proof Setup (and Print-Tool's Softproof best).

So I wonder whether this behavior could lie within the printing profile (wrongly done, editable...) or some general color management issues with blueish casts.

BTW: Usually, printing works fine within my workflow.

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2000

As I understand, images with a strong cyan/blueish cast can be problematic for color managed workflows due to LABs insufficiencies in this area (and manufacturers diverse workarounds to compensate).

What "strong cyan/bluish cast" issue are you referring to? I'm aware of a few issues. As we age our eyes light sensitivity changes and blues are increasingly attenuated. Also the "blue turns purple" effect along a constant hue angle in Lab. But neither of these things should have any effect on color management or how well a soft proof matches a print.

The key factor in getting such a match are a proper viewing setup with luminance and white point matching between the screen and the illuminated print. Andrew (DigitalDog) has excellent videos on a proper setup.

Another factor is the cognitive effect of seeing something on a monitor v a print. Just knowing that one is a monitor can change how one perceives things. This is often stated as the difference between emissive and reflective images but it's not. Photons are photons. But our brains know the difference.  There have been a few experiments where a monitor was physically masked in such a way that the image coming from it looked the same as an adjacent, illuminated print. Then people that didn't know one side was a monitor would not see any differences.

As an aside, the screen captures show a minimum L* of around 16 on the Photoshop softproof but about 10 on the print driver preview. Assuming you are printing to a matte paper, the former would probably be more colorimetrically accurate.

As for the hue shift, I don't see anything problematic in this image. There are issues with very light bluish areas being accurately shown when soft proofed (show paper color) with an M1 profile and a high OBA content paper. This happens because the monitor can't show these light blues accurately. it's an artifact of high OBAs causing the shorter wavelengths to have well over 100 reflectance in the blues due to uV pumping the OBAs. This is actually one of the reasons proofing paper for use in M1 (full D50) viewing environ has some OBAs.
But this is never a problem with M2 profiles or low/no OBA paper.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 16261
    • http://www.digitaldog.net/

Andrew,

thank you, that was the result of my research I did before posting. But the resulting print matches PRINT dialog better than Customize Proof Setup (and Print-Tool's Softproof best).

So I wonder whether this behavior could lie within the printing profile (wrongly done, editable...) or some general color management issues with blueish casts.
If it matches but itís wrong, the issue is the display calibration NOT showing you the correct preview results. The Print dialog isnít where you want to ever be soft proofing.
Logged
Andrew Rodney
Author ďColor Management for Photographers"

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2000

If it matches but itís wrong, the issue is the display calibration NOT showing you the correct preview results. The Print dialog isnít where you want to ever be soft proofing.
The attached profiles from his screen grab indicates a PA271Q profile dated 12/31/19 with an L* tone curve. and a Apple copywrite. Tone curves are all the same for RGB and the VCGT tag info is pure linear (unadjusted power up default). Interesting. It would appear he profiled the monitor for an L* profile using the monitor's internal LUTs. That would be consistent but I can't verify since I use Windows.

Question to the OP: Did you calibrate/profile your monitor? If so what colorimeter/spectrophotometer did you use?
Logged

MichaelKoerner

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 65
    • www.m-werkstatt.at

Andrew, Doug,

thank you for your detailed and very informative posts.

I actually battle my way through "Real World Colormanagement", where I read about the LAB issues Doug mentioned above (the "blue-turns-purple-effect" for example). For lack of own experience, I wondered whether this could lead to the effects I see. Following your explanations, it does not. Btw, I print on Canson Baryta Prestige, which is glossy.

My viewing environment is D50 with Just Fluorescents. I follow Andrew's instructions on monitor calibration and am aware of the differences of reflective/emissive light and the psychological effects when comparing print and monitor.

Doug, I calibrate/profile my NEC (yes, hardware LUT, thus linear vcgt-tag in the profile) with Spectraview II and a Discus to a viewing light specific white point (x0.35, y0.356) and, as you already found out, L* gamma, Mac/Catalina. Recently I experimented with BasICColor Display 6 (which is pre-release), but usually do a factory reset in my NEC when switching profiling software.

I am not sure whether I grabbed the screenshot when the above mentioned print settings were active or with D65/2.2 which I normally use for photo editing/general work.

Nevertheless: Leaving the assumed buggy Photoshop Print preview aside, I still see two - very! - different softproofs in Photoshop and Roy Harringtons Print-Tool. Screenshot 2 shows them side by side. How are such huge differences possible? Same printer driver, same settings.

Print-Tools softproof matches the printing result - although I printed via Photoshop.

Of course I could decide to change my monitor calibration to match Photoshops softproof and forget about Print-Tool - but how can I get assure that the problem lies not within the specific colour range, or Photoshop, or some quirky mistake on my side?

Can someone point me to a description on how to address this questions in a structured way?

Thank you!
« Last Edit: January 14, 2020, 02:52:38 pm by fineartelier »
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2000

Andrew, Doug,

thank you for your detailed and very informative posts.

I actually battle my way through "Real World Colormanagement", where I read about the LAB issues Doug mentioned above (the "blue-turns-purple-effect" for example). For lack of own experience, I wondered whether this could lead to the effects I see. Following your explanations, it does not. Btw, I print on Canson Baryta Prestige, which is glossy.

My viewing environment is D50 with Just Fluorescents. I follow Andrew's instructions on monitor calibration and am aware of the differences of reflective/emissive light and the psychological effects when comparing print and monitor.

Doug, I calibrate/profile my NEC (yes, hardware LUT, thus linear vcgt-tag in the profile) with Spectraview II and a Discus to a viewing light specific white point (x0.35, y0.356) and, as you already found out, L* gamma, Mac/Catalina. Recently I experimented with BasICColor Display 6 (which is pre-release), but usually do a factory reset in my NEC when switching profiling software.

I am not sure whether I grabbed the screenshot when the above mentioned print settings were active or with D65/2.2 which I normally use for photo editing/general work.

Nevertheless: Leaving the assumed buggy Photoshop Print preview aside, I still see two - very! - different softproofs in Photoshop and Roy Harringtons Print-Tool. Screenshot 2 shows them side by side. How are such huge differences possible? Same printer driver, same settings.

Print-Tools softproof matches the printing result - although I printed via Photoshop.

Of course I could decide to change my monitor calibration to match Photoshops softproof and forget about Print-Tool - but how can I get assure that the problem lies not within the specific colour range, or Photoshop, or some quirky mistake on my side?

Can someone point me to a description on how to address this questions in a structured way?

Thank you!

Here's a technique that creates a soft proof.

Taking the original image then for:

Making a soft proof image with "Show paper color"
Select Convert profile then convert to the printer profile using the intent and BPC settings you would print. Then convert the result to ProPhoto RGB selecting Abs. Col.

Making a soft proof image with "Show black ink"
Select Convert profile then convert to the printer profile using the intent and BPC settings you would print. Then convert the result to ProPhoto RGB selecting Rel. Col.

Making a soft proof image with neither show black ink or paper color.
Select Convert profile then convert to the printer profile using the intent and BPC settings you would print. Then convert the result to ProPhoto RGB selecting Rel. Col. with BPC selected.


You can compare the images side by side to a soft proof of the original image. Note there is a bug in Windows that can mess up soft proofing (but not these created soft proofs) that can be worked around by unchecking the desaturate checkbox in color settings and putting in 0% for the amount to desaturate. The problem shows up in deep shadows where the blacks are lower than the paper's DMax. No idea if this is a problem in iOS.

Logged

MichaelKoerner

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 65
    • www.m-werkstatt.at

Doug,

you are my hero of the day, if not the whole month :-)

In fact, this bug affects MacOS as well, your (non)desaturation routine fixed it. On my system it produced a light reddish cast in the quarter tones as two screenshots illustrate:

Screenshot A: Photoshop-Softproof vs. Print-Tool-Softproof, Bug active >> PS cast
Screenshot B: PS-Softproof vs. PT-Softproof, Bug not active >> Both identical

I can't recall how many hours I spent last year trying to get rid of inconsistencies between Softproof and Print in Photoshop, especially in blueish, dark, snowy scenes (living in the Alps did not help ;-).

Is there somewhere a list of such - in my view critical - bugs to take care of concerning CM?

Regards and many, many thanks,
Michael

PS: Interestingly, the MUCH lighter appearance of Photoshop's Softproof viewable in the screenshot of my first post at the top of this thread disappeared miraculously after restarting my Mac. Must have been some sort of digital hiccup.
« Last Edit: January 15, 2020, 06:16:05 pm by fineartelier »
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2000

Doug,

you are my hero of the day, if not the whole month :-)

In fact, this bug affects MacOS as well, your (non)desaturation routine fixed it. On my system it produced a light reddish cast in the quarter tones as two screenshots illustrate:

Screenshot A: Photoshop-Softproof vs. Print-Tool-Softproof, Bug active >> PS cast
Screenshot B: PS-Softproof vs. PT-Softproof, Bug not active >> Both identical

I can't recall how many hours I spent last year trying to get rid of inconsistencies between Softproof and Print in Photoshop, especially in blueish, dark, snowy scenes (living in the Alps did not help ;-).

Is there somewhere a list of such - in my view critical - bugs to take care of concerning CM?

Regards and many, many thanks,
Michael

PS: Interestingly, the MUCH lighter appearance of Photoshop's Softproof viewable in the screenshot of my first post at the top of this thread disappeared miraculously after restarting my Mac. Must have been some sort of digital hiccup.

Michael,

Glad it fixed things. Nothing more frustrating than tools that don't do what they are supposed to do.

I'm not aware of a list of Photoshop bugs like this. It would be nice if there were or better yet, if Adobe would fix the damned things. It was something I encountered when crosschecking (I use that a lot whenever something doesn't make sense) the effect of desaturating on expanding the effective "gamut" of the monitor to better see gamut clipping effects when printable colors exceed my wide gamut monitor. And that's when I found the "fix." It's clearly a different color rendering path. BTW, when desaturate is checked, 10 bit paths are also disabled and a 10 bit monitor will use only 8 effective bits. Fortunately, this effect is rarely if ever visible with photos unless you are working in gamma=1 colorspace. I sometimes use that when resampling because Photoshop's resampling doesn't account for gamma. Sharp changes between close pixels get messed up. Only a problem on super sharp images with things like white picket fences against a dark background and mostly affects downsampling like making 5x7's from a 45M image and downsampling to 720 PPI.


I'm an engineer/coder professionally and I can empathize with how hard it is for the Adobe crowd. Photoshop is a huge beast.

Here's a link that provides more detail on the softproof bug.
https://forum.luminous-landscape.com/index.php?topic=133309.msg1144272#msg1144272

« Last Edit: January 15, 2020, 07:01:18 pm by Doug Gray »
Logged
Pages: [1]   Go Up