Personal Thoughts on RawTherapee 6.0

I was just over at Github reading through the roadmap for 6.0. I notice that 6.0 will break backward profile compatibility by removing several features I greatly rely on. I completely understand and respect the reasons for this — just my luck that I’m likely one of the few people who use them extensively.

It looks like I might be stuck at 5.9 or 5.10, which is a bit of a shame but not a huge problem. One of the things I’ve always loved about RawTherapee, though, is that I could always look forward to new features without worry of things breaking. As I said, though, this really isn’t a complaint — I completely respect the logic, and I’m just kind of expressing my thoughts more than anything else. As long as I’m able to keep running the version I need, that’s at least an option — and that’s one of the reasons why Linux and FOSS in general is so fantastic: there’s always an option; no one is forced into doing anything.

And who knows? Maybe one day I’ll learn enough to fork my very own custom version that retains the tools that I need (a bit ambitious on my part, granted; but it would certainly be an inteteresting challenge).

1 Like

Hello, using AppImages you can have RT 5.8, 5.9 and 6.0 side by side.

Very true. And certainly a great work around. To be entitely selfish, though, I 'd prefer the existing tools to remain.

Pardon the question, but I’m on my phone right now which is hardly an optimum research context… :slight_smile:

Does the roadmap include “full” support for CR3 metadata? That’s the primary issue that keeps me from trying it more these days.


I found this?

Yeah that appears to be it. It was never slated for 5.9 since it required a pretty significant underlying change.


1 Like

We are testing that functionality in the metadata-exiv2 branch. And it’s coming to 5.10.


Are you aware of the ART hard fork? It is somewhat similar to RawTherapee and may or may not drop tools you like.

Which tools do you consider keep your hands off for 6.0, and which tools could you let go?

I believe the tools under consideration were under-performers compared with other methods which were later developed in RT… meaning the effects can be better implemented with the other newer tools.

Yes, I’m aware of ART, but it lacks a lot of what I use. To be honest, I think RT is perfect the way it is — I wouldn’t remove any of the tools at all. But I understand if it’s not an option — I’ll just continue to use the pre-6.0 version.

Maybe we should have a 5.10LTS branch.


That sounds fantastic! Yes please! :slight_smile:

That’s the weird thing. But artistically speaking, they work best for me persinally — I’ve experimented with RT extensively and, while everyone is absolutely right in this regard, for some magical reason I can only get that “Wow! I like that look in your images; how did you do that?” from using those ‘older’ options. Don’t ask me why — I supose it’s more by happy accident than science. Sometimes, things that are flawed, when used in the right way, can produce something special.

But, whatever happens, I’ll always be forever greatful to you guys for RawTherapee in whatever form — it it wasn’t for all your hard work, I wouldn’t be able to do what I love. I’ll always be a huge supporter no matter what the future holds. :heart:

1 Like

It’s interesting… as usual my ‘ideal’ blend of RT and ART would be, well… a blend! :slight_smile: I’ve been using ART exclusively for almost a year now and feel reasonably competent with it (within my context). Alberto is responsive and helpful – as are the RT devs – so it’s worked well for me.

Ideally? I’d like to have RT’s or darktable’s breadth of local adjustments with the ART UI. I understand the ‘spot’ paradigm but prefer the parametric / color / shape / brush mask paradigm instead. I also like darktable’s multiple module instance ability, although it’s possible to approximate some of that with ART’s local adjustment layers. What would be really nice in ART would be a variable soft-edged brush mask. “Adjustment brushes” in a general sense would be very useful and convenient: Just manually and specifically brush-in exposure, sharpness, etc. only where needed.

But on-topic it’s also good to hear about CR3 in 5.10.


Not to hijack this thread, but what is the link for a projected RT roadmap? I’d be interested to see what’s going away. TIA.

1 Like

I believe that this is it. .


Thanks Mike. I’ll take a look.

In an ideal world, tools could be added or not subject to the individual user’s requirements — something akin to the way plugins work in GIMP, for example, or addons work in Firefox. That way, any ‘removed’ tools could be added optionally and possibly maintained seperately, and new tools could be developed independantly. Of course, I’m well aware that, in reality, this would be a monumental task.

I’m considering how to tackle this, and it looks as though sticking with a 5.9/5.10 release via an Appimage is likely the best way forward for me over the longer term — over the shorter longer term, a 5.10 LTS, though, would be brilliant. It’s a bit of a shame that I won’t be able to benefit from any advantages rolled out in future releases, but the current arrangement works so darn well for me that I can’t let such a good thing go — things are pretty much perfect the way they are.

Besides, one of the best lenses I have in my collection rolled out of the factory a good ten years before we walked on the moon. A friend of mine once commented on a shot I’d taken with that lens by saying “I’d love to get shots like that, but I haven’t got around to updating the autofocus firmware in my camera yet.” I just stood there and blinked at him.


That reminds me of a story of when I used to play in a church band. A guy had about $5k USD wrapped up in keyboards. He kept talking about sounding better when he picked up some new gear. He was a decent but not great musician. HT: It wasn’t the gear that was holding him back.

I love to use old glass as well. I even try to incorporate it into my paid work occasionally. I’m doing a senior photo shoot tonight and will use the opportunity to break out a lens I just bought a few months ago.

1 Like

What you’re missing is that those “plugin” features are often in an interpreted language, e.g. lua for darktable, python or scheme/lisp for gimp, and thus are relatively slow compared to the “in program” modules, which are C/C++.

There is likely, more or less, already a “plugin/module” in RT, but since they’re all in C/C++ they need to be compiled and input into the pixelpipe, so maintaining them all together makes sense. It saves everyone from having to build the whole application.

Yeah, I know - just wishful thinking is all.

Unfortunately, I’ve also just noticed that backward compatibility will be broken as well - in short, this would mean I’d have to start completely from scratch without the tools I need to do it (I have custom PP3s that I use for a very specific workflow).

Well, that’s it for me. No more RT beyond 5.10. I’m absolutely gutted, but it was great for all the years I used it. But at least I can still carry on with 5.9/5.10 for as long as I can, and maybe It’ll even outlast me (who knows). While I won’t get any new updates, bug fixes or features, things are pretty much perfect for me as is.

To be honest, I don’t have much money anyway, and my hardware (both computer and camera gear) are very old. I suppose it would likely have got to a stage eventually where it would no longer run on such an old machine anyway. But I’ll keep on going for as long as I can.

From now on, I shall hereby be known as the 5.9 kid. :slight_smile:

1 Like

… maybe I’m wrong or I haven’t read in depth, but the tools that are removed don’t seem so fundamental.