Soft-Proofing in Photoshop

[RobertArdill]

Hi,

I've spent an inordinate amount of time (and even more money!) on colour managing my workflow and ensuring that my equipment is properly calibrated throughout. To give you an idea of how obsessed I've been, I returned 3 Dell U2713H monitors because I just could not calibrate them with an X-Rite i1 Display Pro, and I now have an Eizo CG277 monitor that really is a dream to calibrate. I also have an HPZ3100 printer that has a built-in spectrometer, and I use an X-Rite i1 Pro 2 with ArgyllCMS to profile the printer and to cross-check the display (the Z3100 does a pretty good job of creating a profile, but it's a question of hit the button and get the profile, whereas Argyll gives a lot of flexibility ... in particular for soft proofing, as I've found out).

I would be really interested in hearing the experience and views of any of you who've played around with soft-proofing. At this point I think I know what I'm doing, but I'm really not sure.

I'll just describe my workflow for comparative purposes. BTW, I found the Cambridge in Colour article on soft-proofing https://www.cambridgeincolour.com/tu...t-proofing.htm really brilliant.

I calibrate my printer using its built-in spectrometer. I then profile it using ArgyllCMS (if anyone's interested, I have some batch files that you can download here: http://www.irelandupclose.com/custom...argyll-bat.zip.

Part of the profiling - and this is where Argyll does a brilliant job - is to measure the illuminant and reflected paper white. I guess this information is stored in the printer to PCS mapping (it has no effect on the print, only on the soft-proofing). If anyone knows how this works I would be delighted to have it explained to me!

I calibrate my monitor to 6500K, 80cd/m2, native, gamma 2.2 (standard for this monitor).

I then post-process my image as usual with Lightroom and Photoshop.

Before printing, I soft-proof, normally in Relative Colorometric but sometimes in Perceptual, BPC on, Simulate Paper Color off, Simulate Black ink off. I adjust the image if necessary (sometimes I will have a side-by-side original to see if and where the rendering has messed things up).

Then I darken my room and turn on my Solux lighting contraption, made up of a Solux lamp and tripod – which you can see here: http://www.irelandupclose.com/customer/softproofing/ . I shine the light onto a sheet of white paper (the paper I’m going to print on) and I display the image full-screen on the monitor. The white paper, which is side by side with the monitor and is in my field of view, allows my eyes to adjust to the paper white, so that the image on the monitor doesn’t look yellowish.
If everything looks good, that’s the end. If I need to adjust something, for example the white in clouds, what I do is to bring up the Properties panel for an adjustment layer I’ve already created (say a Color Balance adjustment) while still in full-screen mode; I adjust the highlight to add more blue, for example, make the mask black and paint in white where I want the change … and so on). The Properties panel can be hidden by a shortcut, so there’s very little monitor white shown on the display during this process.

Then it’s time to print (with the same settings as the soft-proofing).

I’ve tried this method with a few images and printed them at approximately monitor size so that I have been able to display the printed image side-by-side with the soft-proofed image. The match is really good, so I think I’m doing it right … but I wouldn’t be surprised if it’s just working by luck and isn’t a method that can be relied on.

I apologise for this very long-winded explanation. But what I would really like to know is if any of you have tried this technique, or some variation on it, or if you have a better way of doing it. My objective is to produce beautiful prints, and I would really, really appreciate advice and help.

Thanks!

Robert

[Manfred M]

Hi Robert - You certainly go through a very elaborate workflow to get your images to come out "right". My question would be why?

Normally prints should be evaluated:

a) Under the lighting conditons that they would be viewed under; or

b) Printed to look right when viewed in a 5000K viewing booth. The ones I used way back in the wet colour darkroom days were made by Macbeth (now owned by x-Rite).



I personally find soft-proofing of very limited value; it will show you out of gamut issues, and to me that is really the only value one gets from the process. Trying to compare a transmitted light, additive RGB image from your screen and getting it to be identical to a reflected light, subtractive CMYK on your printed image seems like wishfull thinking.

I might be a bit more careful if I were a product photographer (and there I would match the important considerations; for instance the colours in a company's logo against the appropriate Pantone swatch in a viewing booth. For portraiture, I tend to print a bit "warm" to improve the skin tones, and for landscapes or cityscapes, I print for the look I'm after. Ultimately getting an image that is pleasing is far more important than trying to get one that is technically spot on. I will take reference shots with either an x-Rite Color Checker Pro card or a white balance card and trust my RAW converter (either ACR or if I'm really fussy DxO Optics Pro 9) to ensure my starting point white balance / custom profile is bang on.

If I'm doing a landscape, at "golden hour"; I'm going for a pleasing look because there is no "right" white balance.

[chauncey]

I generally agree with Manfred...you want prints that appear pleasing to the eye, which may/may not be technically correct. That said, read this...http://www.luminous-landscape.com/tu...too_dark.shtml

[Colin Southern]

I was going to write-up a big reply, but I see that Manfred has kindly done it for me - so in summary "what he said".

(thanks Manfred!).

At the end of the day, printers and monitors are as different as caulk and cheese - one is additive with about a 6 stop DR and the other subtractive with around a 4 stop DR. Having just said that, if any monitor is going to address those limitations then it'll be an Eizo (which typically have a contrast ratio of 800:1 instead of the (cough, stupid, misleading, and inaccuarte) 10,000 -> 1,000,000:1 frequently advertised - plus it'll no doubt have a much Closer gamut than the average monitor.

At the end of the day though, soft-proofing is still just a simulation - and just like RC helicopter simulators that I fly - they're useful - they have their place - but they're never going to be 100% accurate.

For what it's worth, I went through the colour management "best practice" thing too, but in the end I found having a monitor at 80cd/mq needed room lighting akin to living in a cave - and 5000 kelvin was just too "meh/blah" for my liking. These days I post-process at 200 cd/m2 and at (I think) 6000 kelvin.

With regards to printer profiles, the build in spectrophotometers are dead in the water unless you feed the media back in once the colour has stabilised and any over-coatings applied (more so for canvas) (the over-coating lowers the black point considerably, but even then I have to over-ride the default profile settings or I end up with a hopelessly flat image.

My suggestion is stick with what you're doing if it's working for you, but remember in the real world some of the "rules" aren't as cast in stone as one might first think (the sun will still rise the next day if you upped the brightness and room lighting!). With regards to soft-proofing, again, I'm with Manfred on this - I use it to show me OOG areas, but that's it; simulating things like canvas on a monitor is just rubbish IMO. Nothing beats just printing the darn thing and seeing what you've got.

[RobertArdill]

Hello all, and thank you for your replies!

I do agree with what you say ... by and large. However, not entirely. I've been printing since 2002 (Epson 4000, 4880, HPZ3100) and I've never been able to get anything close to a realistic soft-proof on my monitor, so what I've always done is to print, correct, reprint, correct, reprint ... which is very time-consuming and very expensive, especially if you print on papers like the Canson Platine or Hahnemuhle Photo Rag. Of course I did use soft-proofing with the gamut warning to get an idea of where the photo might be affected, and getting to know the paper and printer very well does help a lot (so I pretty much only use 3 high-quality papers and one printer now).

One of the problems I had was that the gamut of my monitor (a Lacie Photon20VisionII) was effectively sRGB, whereas the printer gamuts are way bigger. Now, with wide-gamut monitors the monitor and printer gamuts are pretty close.

The second problem I had was that I didn’t have a 5000K viewing booth and the Solux lamp I use is significantly warmer (about 4300K, even though it’s rated at 4700K). So the monitor proof was always way cooler than the print.

With ArgyllCMS, the viewing conditions are taken into account, so the soft-proof and print are almost a perfect match. The only problem is that the eye doesn’t adapt to the monitor white as it does to the print white (because the monitor has no illuminant for the eye to ‘calibrate’ on). However, if you have the illuminated white paper side by side with the monitor while soft-proofing, and go to full-screen mode … then the eye adapts to the paper white and the soft-proof no longer looks yellowish.

So (and I qualify this by saying that I’ve only been using this technique for a short time, so I may come to eat my words ) I can now pretty well rely on my monitor soft-proof and I don’t need to do repeated prints.

To be honest – I didn’t start off with attempting a perfect soft-proof at all. What I was working on is trying (and failing) to calibrate the Dell wide-gamut display I purchased … and this lead me to borrowing an i1Pro … which turned out to be out-of-calibration … which lead me to using ArgyllCMS to see if I could improve the calibration using my i1D3 etc. Then, having finally given up on Dell and bought an Eizo at over twice the price , imagine my amazement when I turned on soft-proofing with paper white on … and, my goodness, the monitor print looked almost identical to the print I was holding in my hand!

So this is where I’m at – I don’t think soft-proofing is the holy grail … but it certainly seems to me to be getting quite close, and I would be really interested if someone else would try this and see what result they get.

Interestingly, Graeme Gill, who is the author of ArgyllCMS, does not think that the soft-proofing can work as I describe it because he says that only Absolute Colorimetric can give a simulated paper white … but he does admit to not knowing Photoshop. As it turns out, Photoshop does give a simulated paper white in all the rendering modes … and I would certainly like to know by what mechanism.

Thanks!

Robert