Soft Proofing confusion - need to relearn all my color management

I had my system dialed in for years using a NEC PA240 monitor calibrated with the Spectraview software. I always used PS (might have been PS 6) to soft proof and had good accuracy to my Epson 4000.
Now, I have the same monitor (just recalibrated it), but I'm using LR Classic and PS via the online creative suite and a brand new Epson P900 (My Epson 4000 finally kicked it).. What perplexes me is when I soft proof in LR, there is a subtle change (as I'd expect) going from Adobe RGB color space to the Epson glossy paper profile. However, when I soft proof to the same paper in PS, there is a dramatic change in contrast and color saturation (both significantly reduced). The change is so much that I don't trust it all. (However, see below - turns out it was fairly accurate).

Any thoughts on why there would be such a difference between these two programs using the same monitor settings and same paper profile? That seems very odd to me.

So, I just printed a print test target. One from the LR print module and one via PS. Neither are spot on (but they are not way off either), but the PS is one much more washed out (color wise) VS the LR one. However, on a good note I guess that is what the soft proofing sort of suggested would happen. I just am not exactly clear on why it varies so much from one program to the other keeping all the variables the same (I think). 

So:

A - why there would be such a difference between these two programs using the same monitor settings and same paper profile?

B - I read a lot about QImage for its soft proofing. If it's worth it I'd drop the $100, though surely I shouldn't "need" another program outside of LR and/or PS. I can sharpen and DeNoise with my Topaz Lab stuff so if I got it, it would be solely for printing.

C - I need to re-educate myself about color management. Open to suggestions and links to get me up to speed. It's been years but I recall DigitalDog had some good stuff so I will check out his site again.

You should always test output using good color reference images designed for that task. The color reference images RGB values are such that they are set for output and are editing and display agnostic. Test the output this way and examine for the same color issues so we know it's not your image-specific issues causing the problems:

http://www.digitaldog.net/files/2014PrinterTestFileFlat.tif.zip

This and other such documents can also be downloaded at http://www.digitaldog.net/

Next: check that the difference isn't related to a GPU issue by examining with it disabled in both products.

Then examine if any of the profiles are V4 spec instead of V2. Not all V4 profiles play well with all apps and are unnecessary anyway.
Still a big mismatch between the apps?

17 hours ago, digitaldog said:

You should always test output using good color reference images designed for that task. The color reference images RGB values are such that they are set for output and are editing and display agnostic. Test the output this way and examine for the same color issues so we know it's not your image-specific issues causing the problems:

Thank you for the suggestions and info. I did use a test target, but downloaded the one you linked too. Un fortunately I ran out of paper 😞 - but have some order. I did a screenshot recording though, hopefully I can post it. The monitor is set to Adobe RGB so when I proof to that, as expected no change. The difference in the crushing of the colors in PS VS LR is quite evident going to the paper profile. MUCH worse than in LR.  Oddly if I do ProPhoto in PS it expands the colors, but in LR here is not much change.

Anyway, It will be interesting to see what happens printing from both programs once I get the paper. I am pretty sure I have the color management set to printer=off and PS manages color in PS. I'll have to double check the settings in LR. I will also have to figure out the GPU on or off that you mentioned.  I very much appreciate the input! 

Won't let me upload the video. I'll post some screen shots. I notice the biggest changes in the Latin woman's face (left) and color swatch.

As I look at the images and the test prints I did do last night that pinks, reds and blues seem to take the biggest hit. Both in the printing (PS worse vs LR) and the soft-proofing as noted above.

The Photoshop screen capture* doesn’t look right to me but see comments below. The middle one doesn’t make sense to me, we want apples to apple comparisons in all cases. 

The two applications when setup correctly and bug free (GPU/profile handling etc) should preview identically. 

* The screen captures are ending up in sRGB to upload so I can't comment on the quality of the colors seen, only how the differ, side by side from each other and indeed, for some reason, the printer profile softproof looks really bad. 

Do you see oddities like this to printer profiles if you pick a totally different profile, even for a different printer?

5 hours ago, digitaldog said:

The Photoshop screen capture* doesn’t look right to me but see comments below. The middle one doesn’t make sense to me, we want apples to apple comparisons in all cases. 

The two applications when setup correctly and bug free (GPU/profile handling etc) should preview identically. 

* The screen captures are ending up in sRGB to upload so I can't comment on the quality of the colors seen, only how the differ, side by side from each other and indeed, for some reason, the printer profile softproof looks really bad. 

Do you see oddities like this to printer profiles if you pick a totally different profile, even for a different printer?

The middle one was just to show that the PhotoRGB was in fact capable of showing larger gamut (but even here the magenta gets cut and banded). Yes, not apples to apples but I didn't have PhotoRGB available in LR.

I thought they should preview identically, thus the confusion. I need to figure out the GPU aspect.

Opps, I exported as sRGB for the web. I can export again in Adobe but i think you can see what I am seeing in that with the color paper profile the difference is dramatic in PS and less so in LR.

I will re-try it tonight using different paper profiles. (I think) the PS softproofing is not printer dependent - it just soft proofs a specific profile (such as a paper type or color space). Though, I need to double check that - clearly color management is not my forte 🙂 . I actually felt like I had a good handle on it some years ago when i got the new monitor (thanks to your help I might add - many years ago on this very forum) but since it was dialed in I have not revisited this topic in years.

PS is paper dependant with paper profiles. As is LR and both should match each other all settings being equal.

Ok, so I got the paper. I made 4 prints. 2 from PS and 2 from LR, one with the GPU disabled and one with it Enabled from each program. All of the prints look nearly identical. The PS ones seem a "tad" brighter, but only in the shadows of the people's faces. On the graded B&W strip across the top it's actually hard to discern a difference. They all seem slightly darker than my display which is set to 110cd/m.

I attached copies of my print dialog settings, but i seems like that all is working ok and setup OK, but a quick glance my prove me wrong..

The soft proof for the Epson Gloss Paper in PS, even with the GPU off has the same harsh banding and crushing of the magenta and blue.

 

Sounds like all you need is minor addjustments to display brightness.

Big fan of Qimage here, but you might or might not really need it. It's strong points are complex image layouts and, if you use Canon or Epson printers, very good control over setup. Qimage makes it almost impossible to have your printer driver set up wrong. It also has some editing and RAW capability, but I don't use those most of the time. The editing is non-destructive if you use it. The user interface has a learning curve, but there are videos to help with all the stuff it can do, which is a lot. Support is second to none because you're dealing directly with the author on his site.

Oh, one thing to check on this printer (P900) is the media setting on the printer control panel is in 'sync' with what you've selected in the print driver. Looks a lot different on Mac, but there is an option called " Media Type: Use Printer Settings" and they need to both be correct. 

Well, maybe, maybe not. The main issue is still there. I can clearly print a file and get a consistent response from either program. However, Soft proofing is still markedly different from one program to another (GPU on or off). In PS the soft proof crushes the reds and some blues. In LR, the red loses some saturation, but only a fraction of what happens in PS. I suppose I can compensate by boosting the red saturation in PS, but why would there be SUCH a dramatic difference? Should I have to boost saturation in one program but not another for the same image using the same profile? In PS with the first test target from Digtaldog I still get the banding in the red GPU on or Off.

I will have to print this test target and see if it desaturates the PS  image as much as the soft proof suggests it will. 

Well, maybe, maybe not. The main issue is still there. I can clearly print a file and get a consistent response from either program. However, Soft proofing is still markedly different from one program to another (GPU on or off). In PS the soft proof crushes the reds and some blues. In LR, the red loses some saturation, but only a fraction of what happens in PS. I suppose I can compensate by boosting the red saturation in PS, but why would there be SUCH a dramatic difference? I will have to print this test target and see if it desaturates in PS as much as the soft proof suggests it will. 

Here it is side by side. It is especially apparent in the bottom landscape image where in PS the pink sky is almost void of any pink hue. In LR it is diminished, but not gone. Also the chiilies in the upper right and even the green in the skyline is significantly less when proofed in PS VS LR. 

Photoshop side by side:

Lightroom Side by side:

So just the soft proof with a printer profile between the two apps don't match right? And with any printer profile?

First, the preview architecture in Develop is unique to all other modules and the most accurate preview of the data. You should always view previews there and elsewhere at 1:1 or greater and when comparing, use the same zoom for each. So compare at 100% zoom in PS and 1:1 in Develop.
It is normal to see slight differences at 1:1 but if really excessive, you might have a corrupted display profile. Recalibrate the display or whatever you usually do to create a display profile. When given the option to build this profile, pick Version 2 (not V4) and Matrix not LUT. As to viewing Develop module and Library and outside of LR, it helps to soft proof in Develop using the same profile used in Library (Adobe RGB (1998).
 

For an update, it was something with the PS software. I printed from PS and the print look fine, totally different (better) than the preview. There was a update out for PS so I just uninstalled it and re-installed it and "voila" - spot on soft proofing. LR has been good from the get-go. 

Thanks for the thoughts and input.

Given how easy uninstalling and reinstalling adobe products is now that it's a subscription model, sometimes the simplest thing to do with a problem that won't go away is to uninstall and reinstall