Teaching and some special features: Darktable and / or Gimp or Photoshop and Co?


my name is Andreas and I am quite new to darktable. I observed darktable since a while but sticked to rawtherapee in rare cases where I needed to read raws and did everything else by gimp. Usually I did use jpgs out of the camera. So far I never had ambitions to get to a bigger public or to use any of my pictures for prints, so that best technical quality was no issue.

I have however now the chance to take over a photography lecture for students of “Packaging Design” starting in winter 25/26 and I like to accept this challenge.

In this context I struggle now which tools to teach. Most of my students never will earn money by photography, so this speaks for free tools. Also a point is that the algorithms are open and it is not just magic for those who want to understand, so why I’d prefer to use free software, even if there are education programs and offers of Photoshop & Co. With darktable I like the more modern approach of working non destructive.

Unfortunately Gimp and Darktable are quite different in usage. So my feeling is it may make sense to focus on one of them, either Gimp or Darktable.

I also like to teach “graphical fotography” like:

(which is an overlay of the same picture, one slightly moved and subtracted, totally b/w)


the same as negativ of the first; in analog times the slightly shifted overlay of a positiv and negativ and then directly on paper.


Here not shifted but blurred and overlay again. In similar technique those “Neon arts” of the 80ies can be done step by step without having a super smart effect filter (by adding this now with a certain color on top of the original picture).

Or here two pictures in one:

So my question is what you would recommend for this kind of stuff: Sticking with Gimp, or waiting that there e.g. comes a filter which allows to overlay a non editable pixel pipe?

I also have problems so far to work with the masks in Darktable. Maybe because I am still not used to it. Especially how to use such a a picture as mask and how to change / extend an existing mask I struggeled so far.

Or in general better teaching Photoshop and Co what would be so far the mainstream?

Best regards

1 Like

if you want to focus on raw processing, i.e. get the best from the raw output of your camera, then darktable ist the tool for you.
If your focus is on creative manipulation of an image, gimp or krita is the better choice.
darktable won’t ever get that flexibility to manipulate images as gimp, since it‘s pixel pipe isn’t designed for such usecases.

1 Like

Thanks Martin! I was not aware of Krita. That makes a more modern impression than gimp, and seemingly can do what I’d need with respect to the stuff above.
Best regards, good start into the new year!

Given the date of your course and the rapid growth of AI tools in photography…current software might not be that relevant. by then or at least might be very different. The days of modules and sliders is perhaps not going away but if you can type “make the girls jacket yellow” you might not be messing with hue sliders… It will be interesting to see where things go over the next couple of years and how fast it changes

1 Like

Hi Andreas,
I actually have a teaching background in photography and imaging. I get really annoyed at the economic barrier put in place for students by forcing them to use Photoshop. I personally have taught and promoted the use of GIMP instead of Photoshop. GIMP was used in the making of films like Harry Potter by the way so it is no slouch. GIMP is possibly the more suitable tool for you and your students to concentrate on. The use of layers and layer masks is what makes it so suitable for your students needs. In my opinion DT is really about getting the most out of RAW files and your students are likely to be shooting JPG images. Of course you could process a RAW file in DT and then export it and open in GIMP, but keep it simple for your students at this time and focus on GIMP.

In my opinion with the advent of AI there may come a time when PS gets a real advantage of GIMP if you want AI generated images.

If you would like some lesson notes from me to help with planning your GIMP class then private message me.

1 Like

Thanks Todd and Terry!

Terry, I’ve set you a private message-hopefully that worked.

I looked into Krita, and yes it can do what I did with Gimp. So I almost suspect that Krita to a certain extent is Gimp. At least it is not so different:

Seemingly it can do different workflows: Either to apply directly on the layer and change it, as in gimp or to add filters to the layer as I’ve done above, so that the layer is not changed. The file is quite huge. So I believe if I copy a layer then the data is really copied and held twice in krita. So maybe internally more gimp as photoshop.

With gimp itself I see some two to three problems:

It looks lets say not very nice. Krita is at least a little more “photoshopish”.
I struggle with its 8 bit colors. Seemingly there are possibilities to work with different resolution but I am not very sure and confident about this.
No vector graphics style of work.

AI: Yes, quite impressive what it can do. Good question whether there also will come a real open AI. Would be desireable. And maybe something like darktable would be a good field trial. So that the moduls are selected and parameterized automatically from what you type in that you want to have and later click on scales of fulfillment for what you have typed in.

Best regards

Krita supports 32bit float … see Bit Depth — Krita Manual 5.2.0 documentation

and like darktable: Scene Linear Painting — Krita Manual 5.2.0 documentation

This is the sort of thing I was talking about…it still has a way to go but imagine in 2 years what this sort of thing will be like… We will essentially be asking the software to do the editing and then likely switching out to a manual mode from time to time… well we were also supposed to all be in flying cars by the year 2000 or something but in any case I think the way we edit photos might be substantially different in 2 years from now… time will tell I guess.

There already is, it’s called stable diffusion. Can be installed from git and run locally if you have decent hardware, particularly gpu vram. I think Krita has also started incorporating it.

And it can support LAB/CMYK, and in some cases mixed color models with file layers.

No, Krita will never incorporate AI tools with the possibility of replacing real artists.

Correction: It can be incorporated into Krita as a plugin.

Happy new year to everyone!

AI and replacing real artists:

I believe that we need to learn what to give over to AI and what to keep in hand as some 80-150 ys back what to give over to automatic control and what to keep in hand. There were e.g. discussions whether it makes sense to replace police men with traffic lights. But actually it turned out that the at that time stupid traffic lights just working time based often behaved better than the policemen which obviously worked somehow “traffic actuated” but not necessarily in the sense of an optimum.

So what may be the question is whether AI can replace the craftmanship involved in digital image processing as in earlier time quite some of the famous fotografers did their chemical lab work for themselfs but by far not all of them.

E.g. the picture examples of the last days, e.g. blackrock cottage: I wonder how to deal with the noise of the clouds (quite prominently visible in some of the examples) in fineprints. Does that noise disturb? My feeling is yes, because it is clearly visible at 100% and so it may be seen in the print as grain. So somebody should get rid of that noise. Why not to give this task over to an AI?

I’d even say that cleaning out disturbing details, as seen in this linked video, could be left over to AI. Creating new picture content in the sense of new picture attractions (shown by some photoshop advertisement) is maybe what should not be handed over to AI. Here I’d even go so far that there should be laws that pictures in which this has being done must be taged visible in the picture having done so.

Best regards

I was just watching CNN I think and they had a story about an artist…this time it was music but she released the song as the AI rendition of her voice…you know just because you can I guess…why would you sing with your own voice (some might argue that few artists have truly sung with their “own” voice for some time :slight_smile: ) in any case there is so much experimentation and nuance ahead. The problem is this thing can change and evolve faster than we can adapt to manage it and those adept at using it will have a unique skill set for both good and bad… If I have a tool and I can say or type make the house yellow or the flowers pink and its does so with fidelity then soon I will not be masking and using an hsl control or channel mixer however if a deep fake video is made that incites riots or massive disinformation and/or propaganda then that is bad for everyone… I guess we all just have to stay tuned and see where this goes and pick and choose how we adopt it or push back against it where we see fit as individuals, government’s and societies…

however if a deep fake video is made that incites riots or massive disinformation and/or propaganda then that is bad for everyone… :

Yes. thats why I would say that a law is needed that such changes need to be visibly marked.

But it cannot be avoided, thats for sure.

Last weeks I observed two men, maybe farmers, talking about electromobility.

The one talking to the other: “I am still not convinced about those electrical cars, because so the batteries burn”. The other to the first: “150 years ago already two men sat on this bench, and the one said to the other: I am still not convinced about steam engines as so often that they explode”. Then the first started to laugh and answered: “That may be the reason that we still have the flails at home.”


People often mix AI generated creations with AI powered tools, like selections and masks that can automatically select your subjects out of almost everything with extreme precision. And they are huge time savers producing fantastic results. But they don’t alter the image itself.

In that regard Adobe has no competition, be it LR or PS.

As a photographer for gettyimages, I can share my workflow. When I want to experiment, have some fun, I’m all Linux, GIMP and Darktable.
But when something needs to be done for real work, with time pressure, max. features and stability, it’s always W11, Adobe Lightroom Classic and Photoshop. And there’s no exception to that rule.

Majority of people are unfamiliar with Darktable and GIMP so if you want to kind of ‘spread the love’ and make people discover those programs, please do so. But from the usability perspective, you’d be much more approved with Adobe apps, as those are industry standards. They do cost, yes and the other ones are free.

1 Like

Thanks. Yes, difficult decision what to teach. I still have time to decide; if I needed to decide today it would be Adobe.
Best regards,

Background info: I have been working as a photographer for 25 years, held a few workshops myself and I am currently a student of Software Design, writing my bachelor thesis.

“Photography” is a very broad topic to teach.
This can start from the original CAMERA – a black box with a hole – up to current digital photography with remote cameras and fully automated workflows.

What photography does not include in my opinion:

  • Painting except for masks
  • AI tools – that would be Promptography

For package designers I would go hardcore mode and teach them about straight up oldschool photography - starting with JPGs from the camera so they really see and feel how the physical part behaves. (Graphic) designers are used to making up stuff from their head, they are not bound by reality. And if they get a design involving images really wrong it is most often because they do not understand how perspective, vantage point and most importantly light work.

If the students are to do some photography on their own and hand it in: using a non-destructive editor like darktable together with RAW-files should be a good choice. And most phones can do decent raw files these days, making it a zero-invest – something students will appreciate.

The PlayRAW section of this forum is a really nice example of a modus operandi: here’s the raw, here’s my edit settings, this is the result. Give them a chance to stay honest. :wink:

my 2cts.

1 Like

That’s a good term! :slight_smile:

I agree. I think even the phones that don’t have a RAW option can sometimes use the OpenCamera app to do it.
I use that app, although solely for video, as the stock app on my phone applies some noise reduction to the video audio that gives very strange results in some situations.

I played with open camera a bit but it caps the raw at 10 bit and the DNG from the pixels were very weird… looked almost like log footage rather than a raw file…certainly also very different from the ones that came from the google app…

1 Like