Soft proofing with darktable

I’m new to image processing (been using darktable for less than a year), and I’m trying to figure out how to soft proof my images for printing. The printing service I would like to use offers their ICC profiles and recommends that, when using Photoshop or lightroom, rendering intent should be set to relative colormetric. However, I don’t see an equivalent option in darktable soft proof mode. Am I missing something? Can’t we select the rendering intent?
If I simply select the ICC profile when soft proofing, all images loose a lot of contrast and saturation, which I can’t recover (simply increasing saturation or contrast produces really strange results). I couldn’t figure out yet what I’m doing wrong.

Sorry if these are dumb questions, but I would really appreciate some help.

Thanks!

Have you used the search facilities in this forum?

darktable soft proofing

As far as I know darktables internal rendering method doesn’t support intent, but you could use littlecms for that, although I’m quite happy with the darktable internal one. I could imagine that your monitor calibration is off or not switched on for darktable…

https://www.darktable.org/usermanual/en/color_management.html

1 Like

Yes, but nothing particularly helpful about it so far.

It is probably off, since I do not have a colorimeter for calibrating it. The white point is set to 5000K and Gamma to 2.2. Do you know if there is some way to achieve a reasonable (and just reasonable) calibration of the monitor without a colorimeter (using a Macbook)? I’m just confused about the great difference I’m seeing between the edited images and the soft proof.

I also find color management confusing. The soft proof profiles I got from my lab also result in a washed out image. But I thought that was intentional, since printing to a matte paper reduces the contrast of the image. In other words, I did not think much of it since the soft proof actually made the image image look more like the final print.

1 Like

Yes! That is exactly what softproofing is supposed to do.

What I did when I didn’t have screen calibration was to make a cheep small test print with the shop to check how the print compares to the softproof on screen.

And then adjust the photo accordingly.

It takes a couple of days and for important prints maybe more than one iteration, but this will guarantee good results for the final print.

Edit: For my screen calibration changed really a lot, major colour shifts from yellow to green - so, depending on your screen of course, expect some strange results for the first try… No Idea why the manufactures sell the screen with shitty profiles… Anyway, the bonus of calibrating is that you can order stuff online and know the colour in advance :wink:

1 Like

@Clerian Thanks for the tip, that’s probably the most simple and practical way to go right now!

What I have always wondered is if you have a softproofing profile, wouldn’t you want to use that as your working profile so that the image you develop is what you will get when it is printed?

No. Your working profile should be fairly wide, so as to avoid unnecessarily clipping values while editing. Profiles like ProPhoto, Linear Rec 2020, etc etc are good working profiles. Printer profiles tend to be of a smaller gamut and probably correct for some slight color/tonal shifts that the printer has.

Work in a wide profile and then softproof. What do I do if I click gamut check and clipped colors show up? Wouldn’t that be more likely to happen working in a wide profile?

Depending one what options you choose at export, the out of gamut colors will be mapped to in gamut colors.

I suppose so, in the sense that if you set your working profile to Adobe RGB and your output was also Adobe RGB, then you’d never have the gamut conversion at output, but, you’d be loosing a lot of color while editing. Once you push outside the gamut space, if you make another manipulation to try and bring it back in to gamut, then how does the pipe know what to do?

For web display, although the only place any of my photos get displayed is if I post one in a Play Raw, but in general, don’t you have to use sRGB? Plus, I do get photos printed, and all the printers I have used, Printique, for example right now, only print in sRGB. So I work in sRGB and try to keep everything as much in gamut as I can.

I notice many times when I use xmps downloaded from a Play Raw that there is a lot out of gamut, at least when checked against my sRGB working profile. (My monitor is calibrated.)

Edit: I see that oog is always “fixed” when exporting. Have you an opinion regarding perceptual vs. relative colorimetric? My desire is to know what a picture is going to look like when printed. I’m thinking that exporting and then comparing the exported image against the image seen in darkroom should show me if there are noticeable differences. Then I would have to figure out how to fix differences.

I always use perceptual, as I want to preserve the relationship of in and out of gamut pixels in my photograph.

If you’re interested in the conversion, go ahead and compare the two. But I really only check to see that my output is acceptable looking.

Thanks, Mica, for your answers. I have a better grasp of what is happening and why.

I know this topic is a little bit stale by now, Bit in the meantime I have discovered what issues the original poster might have had. So I just writhe my thoughts here for reference:

Some print shops offer their soft proofing ICC profiles only as CMYK profiles and then the darktable has an issue with displaying the soft proof (I stumbled upon this issue recently). As a quick workaround it helps to use gimp for soft proofing in this case.

e.g. the ICC profiles here: cyberlab.at | Das Hi-Tech Fotolabor exhibit this behavior.