soft proofing confusion

[stuck]

I have sRGB profile 16 bit TIFF images. When I open them in my (ancient CS2) version of Photoshop and turn on soft proofing (correct paper set, 'preserve RGB numbers' ticked, 'simulate paper color' ticked) the image darkens considerably. It doesn't darken if I UNtick 'preserve RGB number's but experience shows the darkened 'preserve RGB numbers' view is a fair reflection of what comes out of my printer so that's the option I use, so I can judge how to get a better print.

However, if in Photoshop I change the mode from RGB Color to LAB Color and then turn on soft proofing (correct paper set, simulate paper color ticked) there is hardly any change in the way the image appears. Before I start 'wasting' paper learning how an image in LAB Color mode actually turns out as a print, is this difference in what I'm seeing on screen between this mode and 'RGB Color + preserve RGB numbers' to be expected?

[DanK]

I rarely print from Photoshop, and never from L*a*b, so I can't directly answer your questions. Moreover, I don't know whether CS2, which was released 17 years ago, behaves like the current versions. However, if I'm correct, the purpose of checking "preserve RGB numbers" checkbox is to tell Photoshop NOT to map colors to the output device, so it's normally not used.

Is your monitor brightness set correctly?

[stuck]

Is your monitor brightness set correctly?

Almost certainly not and I'm aware it's far from ideal in all respects when it comes to photo editing but it's all I've got. My workflow is a kludge derived from trial and error (so it's bound to be riddled with bad practise) but on the whole it allows me to press print and get something I'm reasonably happy with first time.

It was more 'trial and error' poking around that led to to my observation that while in L*a*b mode turning on soft-proofing seemed to have no effect on what I saw on screen.
I suspect I'm missing the point of L*a*b and my question is not even sensible. I just wondered if bigger brains than mine could offer any thoughts.

[Manfred M]

First of all, I have a problem with your workflow. I do use the L*a*b* colour space to edit in from time to time and do a lot of printing from Photoshop. I have been using Photoshop since CS came out, so did use CS2 at some point a long time ago. As Dan points out, it is a very old piece of software and I really cannot comment on any idiosyncrasies it may have had any more.

First of all, you should NEVER print from the L*a*b* colour space. I am not aware of any print drivers that support it. If you are working with most photo printers, you need to convert to an RGB colour space, as this is the input the drivers expect (unless you are working with digital presses or offset printing processes, where CMYK is used by the drivers).

When you finish post processing and are ready to print in Photoshop, you need to convert your image to an RGB one. I suggest either Adobe RGB or ProPhoto RGB. If you are using a gloss or luster paper (including Baryta) and have highly saturated colours, ProPhoto is often best because the gamut of the printer / paper can exceed the AdobeRGB paper. If you are using a fine art cotton rag paper, stick with Adobe RGB as the gamut of these papers is far lower. As for rendering intent, look at whether the Relative Colorimetric or Perceptual gives you a better looking print to make a call there.

If I recall correctly, you should NEVER use the preserve RGB numbers as this functionality was there to support the pre-press proofing process, not the printing processes that we use on photo inkjet printers. I have heard the "Simulate paper color" setting as the "make my image look like crap" setting; something else I never use. I primarily use soft proofing to check for out of gamut issues and make tweaks there as required to get an acceptable print.

Every competent printer I have ever me swears by test prints to do the final tuning before making the final print. Some never use soft proofing.

Finally, something Dan wrote about is the setup of your computer and work environment. Dan and I both use wide gamut screens (AdobeRGB compliant) that are calibrated and profiled. Unless your screen is at least sRGB compliant (it will say if it is in the specs; if it doesn't say so you can safely assume it is not.) I have my screen output set to 80 candela / square metre; the general recommendation is to stay in the 80 - 120 candela / square meter range) and the colours I see on my screen come out very close in my final print. My workspace is lit to around 29 lux, which I have checked with my ambient light meter, so I am also in the recommended range there. My workspace is in my basement, so the light intensity is constant throughout the day.

I generally produce at least 1 or 2 finished prints every week.

I hope this all makes sense to you.

[stuck]

...you should NEVER print from the L*a*b* colour space...

I had a feeling I was way off. Thanks, duly noted.

...If I recall correctly, you should NEVER use the preserve RGB numbers as this functionality was there to support the pre-press proofing process, not the printing processes that we use on photo inkjet printers...

I see, as I said, I've been making this up as I go along. Not a great approach I know.

...Dan and I both use wide gamut screens (AdobeRGB compliant)... ...I hope this all makes sense to you.

I'd love such a monitor but alas I don't have the money at the moment.

Meanwhile, yes, it makes sense. Many thanks for your comments.

[DanK]

I'd love such a monitor but alas I don't have the money at the moment.

Wide-gamut monitors (Adobe RGB or therabouts) tend to be expensive. However, many inexpensive monitors reproduce most of the sRGB gamut. Yours may well. But even if it does, you need to calibrate it and set the ambient light at a reasonable level to get what's on the screen to come close to prints.

This is a balancing act (brigher ambient levels require a brighter screen), and I think it varies from person to person. I set my monitor toward the high end of the range Manfred suggested, usually either 100 or 110 cd/m2. I bought an old incident light meter on eBay solely to help me set the ambient light in that room, which is all on a dimmer. It takes some trial and error to find the settings that work best for you, but I'd start with the ranges Manfred suggested.

A cruder approach is just to make a bunch of test prints, darkening the monitor between prints until the print and image on the screen seem reasonably close. They will never be identical because one is reflective and the other emissive.

[billtils]

... Every competent printer I have ever me swears by test prints to do the final tuning before making the final print. Some never use soft proofing....

More than 10 years ago when I got my first digital camera I used Apple's Aperture software for processing. I could never get it to produce a print that matched what I saw on the screen and, being new to digital processing, the stop-gap solution was to make test prints. Aperture is long gone but the test-print process is hard-wired in my workflow (at least for prints that matter) and I now keep packs of small size paper for test printing.

Sometimes simple (a.k.a. Dan's "cruder approach") is best and the test print process using small sized paper is an economical and effective was to compare the effect of different papers on the final product. Yes soft-proofing does that but what it doesn't do is show how ambient light influences how the print looks.