I'm starting this off by saying that I'm looking for any type of reasonably advanced photo manipulation tool, that runs natively under Linux. It doesn't have to be FOSS.
I switched to Linux, from Windows, about three years ago. I don't regret the decision whatsoever. However, one thing that has not gotten me away from Windows entirely, is the severe lack of photo editing tools.
So what's available? Well, you have GIMP. And then there's Krita, but that's more of a drawing software. And then...
Well that's it. As far as I know.
1. GIMP
Now, as someone migrating from Photoshop, GIMP was incredibly frustrating, and I didn't understand anything even after a few weeks of trying to get into it. Development seemed really slow, too. It's far from intuitive, and things that really should take a few steps, seemingly takes twenty (like wrapping text on a path? Should that really be that difficult?).
I would assume if you're starting off with GIMP, having never touched Photoshop, then it'd be no issue. But as a user migrating, I really can't find myself spending months upon months to learn this program. It's not viable for me.
No hate against GIMP, I'm sure it works wonders for those who have managed to learn it. But I can't see myself using it, and I don't find myself comfortable within it, as someone migrating from Photoshop.
2. Krita
Krita, on the other hand, I like much more. But, it's more of a drawing program. Its development is more focused on drawing, and It's missing some features that I want - namely selection tools. Filters are good, but I find G'MIC really slow. It also really chugs when working with large files.
Both of these programs are FOSS. I like that. I like FOSS software. But, apart from that, are there really no good alternatives to Photoshop? Again, doesn't need to be FOSS. I understand more complex programs take more development power, and I have no problem using something even paid and proprietary, as long as it runs on Linux natively.
I've tried running Photoshop under WINE, and it works - barely. For quick edits, it might work fine. But not for the work I do.
So I raise the question again. Are there no good alternatives to Photoshop? And then I raise a follow-up question, that you may or may not want to answer: If not, why?
Thanks in advance!
GIMP is beyond stale and it's frustrating to see people recommend it as an "alternative" to Photoshop when it's about as actively developed as X11. The fact it's making rounds on FOSS news channels/sites because they ported the UI to GTK 3 (Which was replaced by GTK4 3 years ago now) is really a sign of how bad the project has gotten.
Photopea is a near feature-for-feature clone of Photoshop, designed around the superior UX and UI of photoshop, and all within a webapp that leverages hardware acceleration. All done by a single person. The downside is that it's a proprietary webapp that costs money to use without ads clogging half the screen.
And you know what? I STILL prefer Photopea to GIMP, after using the latter for years. GIMP is old, slow, and pretty much dead in the water and I'm certain that they'd have produced 3.0 faster if someone had rewritten it over a weekend instead of trying to port the godawful mess of tech debt that must be going on inside the GIMP project atm.-
Photoshop getting better support via WINE/Proton is more likely than GIMP ever returning to its hay-day of being a true competitor to PS.
Downside of Photopea is it's not open-source (mainly because the creator needs ad revenue to run it, but I digress)
I have to agree, Photopea really is the best alternative to PS for Linux users. It's honestly good! I wish Affinity would consider launching a Linux Verizon, as I actually like that a lot more than PS, but that seems equally as unlikely...
So for now, Photopea seems the best option overall. One plus, being written in WASM (probably using Rust?) it's really speedy and fast. It feels faster than Gimp anyways, which is definitely not a good statement on the state that Gimp is in...
Prompt move to GTK3 and now 4 adds very little value to gimp. Using it as benchmark is completely useless. If I understand correctly there are major changes happening under the hood and the effort may not have much effect until the work is finished.