Future rawtherapee features

I realize RT is open source and as such lacks some of the advantages of software that is developed in a more organized environment.

That being said I have a question or 2 about new features and enhancements to RT.

  1. Is there a forum to take suggestions from the user community?
  2. How are features chosen for implementation, is there a vote?
  3. Is there an long term (a year or 2) overall plan for the direction of RT?

The RT site directs me here for Forum access. I have to say, I don’t see much activity on this site about the three questions I raise.

Open source,m like everything else has its strengths and weaknesses. Things like local adjustments seem to be wanted but unfortunately don’t make it (at least I don’t see them) into the main build.

Many thanks

Answers are based on my experience with FOSS and I’m not a developer. That being said:

True and some of the weaknesses are also its strength: Developers do this in their free time, mostly during the weekends or evenings. The downside being that things progress somewhat slowly and the upside is that they must really like working on a specific problem to invest time and effort on it and the result is often quit nice/impressive.

It is not yet available in the latest stable release, but a fleshed out version is in the development branch. No need to build your own to access it if that isn’t your thing, you can download one of the fairly recent development (appimage) builds.

I do just about all my edits on the latest self build development version and except for a very occasional crash when using the wavelets levels tab I don’t have any issues with it. Major bugs are, most of the time, resolved rather quickly.

The general discussion about a specific feature request should be done here on pixls in this rawtherapee section. This to make sure that normal users can have a say and to get a feel about the general response. The actual request, assuming there is enough interest, should be done on GitHub.

There’s no democratic voting of any kind. If a developer has the knowledge, time and feels like picking up a certain issue/request s/he will. I’m assuming that the main developers do communicate about what is feasible/wanted to make sure that RawTherapee stays on its path.

I’m not aware of their being a long(er) term plan.

EDIT: Fixed a word and a “stretched truth”.

5 Likes

Hi @dellaaa, as one of the people actively contributing to RT (not much through code, but through support and GitHub ‘triage’), I think I am qualified to comment.

@Jade_NL hits most nails on the head. I see RT developers as a rag-tag bunch of enthusiasts. We operate mostly independently without a clear organizational structure and without clear aims for the future. There is really no plan (that I know of), there are only unwritten rules of conduct on GitHub, our primary platform for development and discussion. There is some sort of democracy when it comes to what is and isn’t added to RT, but, unless I am stepping on toes now, there is no one really ‘in charge’ except for maybe @Morgan_Hardwood who owns the repo, who has so far called the shots for new releases, and who is probably the maintainer of the official website (I’m not even sure!) and surely does some other stuff.

There is only a handful of other active developers and contributors. It’s Ingo @heckflosse and Jacques @jdc who are responsible for the vast majority of commits in the recent year. Ingo focusses mostly on improvements (codebase, performance), but throws in a new feature once in a while. The Local adjustments toolbox and other algorithmically heavy tools are the brain child of Jacques. He has had a lot of help from @Pandagrapher and @Wayne_Sutton along the way.
Then there’s @Floessie who applies their excellent knowledge of modern C++ for critical reviews. And there’s me, concerned with little things such as camera support, triaging issues on GitHub and trying to manage expectations, keeping the userbase happy and informed. Finally we have incidental (or potentially upcoming regular!) contributions of people like @rom9 (film negative module), @Lawrence37 (perspective correction) and an excellent bunch of other contributors who may not be interested to fully commit to the cause (whatever that may mean…), but still like to add something of their own.

Where does that lead us? I have no clue. Do I care? Well, certainly. If I could, I would not hesitate and spend more time on development, fleshing out the good into something even better. I would not be afraid to try and pick up a more ‘project management’ role to see where that gets us. But real life gets in the way, as is the case for many of us; active, continuous involvement in such a project is hard.
You might argue that RawTherapee is starting to get out of touch with its competitors, that it might be getting dated, that development has stalled. I would probably agree with you. Then again, as a project without clear goals and without any real concern for market share, what is there to lose? The project is what people bring to it. Could it be something else? Certainly. Can I change it? No, not at the moment. Can anybody else change it? Sure, come on board and we’ll have talk. Collaboration remains key.

7 Likes

(and to add, but Discuss did not let me mention more than 10 people :open_mouth: )

We should definitely not forget the work of @XavAL and @jdc and others for documenting RawTherapee on RawPedia. And a special mention for @Andy_Astbury1 for making excellent tutorial videos on YouTube.

3 Likes

@Thanatomanic: Thanks for the insights!

I have a better idea who’s who now. For example: I never knew where to place @Morgan_Hardwood in the RT hierarchy…

I think one might need to look at that on a function-by-function basis; with respect to demosaicing, for example, I don’t think anyone, FOSS or commercial, compares to the performance of RT…

If I hadn’t done my own thing, I’d likely be using RT to develop linear TIFFs to take advantage of all the raw goodness there, and opening these TIFFs in darktable to do the color and tone transforms…

1 Like

Don’t forget @Hombre who’s been coding on RawTherapee for many, many years.

Hi @paulmatthijsse, absolutely, there have been many extremely valuable contributions to RT! @Hombre, @agriggio and many others. These graph does it better justice: https://github.com/Beep6581/RawTherapee/graphs/contributors
I did not mention them, since their last direct contributions to the project were mid 2019 or earlier, so I don’t consider them ‘active’ contributors at the moment. Of course that may change in the blink of an eye. For example, lately Adam Reichhold has returned to contribute again, which he did in 2016 for the last time. Never say never :slight_smile:

1 Like

Thanks for all the responses. I use RT for a lot of raw development (Fuji and Nikon), sad to say it doesn’t support my Sigmas.

In many ways much better than anything available.

Very strong points: Demosacing for Fuji, sharpening (contrast by details). Now that I think of it IDK what you could add. Layers? IDK, Brushes for some of the tools?

If I could make a humble suggestion, it would be that it is at times overly complicated. I think several controls have little or no effect and other like wavelets are difficult to master.

Many thanks to all. I am an Oracle DBA and C++ coder btw.

Foveon or non-Foveon?

Foveon has some really unique challenges that make it extremely difficult to support. It’s significantly harder than X-Trans (which gets supported because a few RT developers do shoot Fuji on occasion, I think most X-Trans support has come from contributors who own the camera.)

I have a few tiny contributions to RT myself, I keep on intending to contribute more but I have this bad habit of hopping between various projects, it’s only recently that I started doing photography/videography stuff again after quite a long break. There’s literally one github issue where @Thanatomanic provided me some data to look at 6-9 months ago and I STILL haven’t gotten around to grabbing the files to see if I can notice whatever phenomenon he was describing.

That is actually @patdavid (and I think I have maintainer rights also).

1 Like

When you describe the circumstances / process, it really is impressive that such a great piece of software emerges from it all!
Seems like darktable currently has the most rapid development, but I find RawTherapee / ART still have some of the best quality algorithms, especially for sharpening and detail.

4 Likes

Yeessss!! :+1:

2 Likes

I have 2 Foveon cameras, SD1 and DP1 and I have the SD Quattro. RT will process the files from the DP1 but the corners have a weird cast.

I cannot find any RAW examples online for the camera types you mention so maybe this might be a Foveon issue, but my guts tells me this is probably a lens thing.

Are the lens(es) you use supported/present? Transform tab -> Lens /Geometry -> Profiled Lens correction

And have you tried using the Vignetting Correction , also in the Transform tab -> Lens /Geometry section?

Sigma RAW format is X3F and last time I checked RT (or DT) could not handle them. In fact, not many editors can.

I’ve read that in the corners, Foveon sensors get crosstalk from angled light.

Just checked the libraw supported cameras list, and all three of the OP’s cameras are listed (although libraw lists “SD 1 Merrill” and “DP1 Merrill”, don’t know if “Merrill” indicates a limitation). Now, I downloaded a DP1 Merrill raw from dpreview, .X3F file extension, and tried to open it in rawproc with the latest libraw, and it didn’t work…

@ggbutcher:

I did find some X3F examples [sigmaphotopro] and they don’t load (are ignored) by RT, DT and Filmulator.

The xtrans demosaic was coded by Frank Markesteijn for dcraw.
Then I started to port it to RT and at the same time Dan Torop from darktable ported it to darktable.
That was a very nice collaboration resulting in a much faster implementation than the one in dcraw.

Doing this, RawTherapee got a lot of help from @sguyader by providing test shots and also testing to get a good (though not perfect) xtrans support in RT.

At least I never used an xtrans camera :wink:

2 Likes