One nice perk of SilkyPix for Fujifilm users is that the software includes the film simulations. I think learning more about SilkyPix and darktable will be on my agenda this year. I did try running SilkyPix in a VMWare Workstation virtual machine a couple of months ago. It seemed to work fine, but moving the program window across the desktop was very slow. I don’t know if the issue was with VMWare or with SilkyPix code.
Using the appimage from digikam’s homepage is not an option?
Cool. Thanks, @wallie
That has issues with workload I guess, but can benefit the general user experience. In darktable some modules have an confusing overlap, example exposure and basic adjustments.
I’ve had a similar journey over the past 15 months or so. Up to that point I’d used Lightroom/Photoshop since Lightroom was in beta, upgrading every second version. When I moved to Fuji late in 2019 I needed a solution as I wasn’t going to go down the subscription route.
As a software engineer I’m a Linux user on a day to day basis, so I’d been running Lightroom either in dual boot or in a VM, so I wanted to move completely to Linux and reduce the Windows dependency.
As far as management is concerned, I knew pretty much from the outset that Digikam was going to be the solution. I know Darktable has tagging support built in, but coming from the Lightroom catalog, I wanted something that could measure up to that. Digikam does, albeit with a bit of a learning curve to get figure out getting around the UI and configuring it properly. I also use Rapid Photo Downloader to pull in the photos from the camera or card as it allows me organise the photos as I want
Linux and indeed Unix has the general philosophy that tools should do one thing and do it well, These basic tools can be composed into pipelines to do more complicated tasks. This kind of applies here, as though there is some overlap between the various individual programs, they can interact with each other (via the sidecars for example). The advantage is that if you do want to try something else (e.g. Darktable vs ART) you’re not starting from scratch for the whole pipeline.
Having initially done a fairly quick comparison between Darktable and RawTherapee over a few days, I decided to go with Darktable. There’re a lot of resources on the net - plenty of excellent tutorials on YouTube and Aurelien produces a wealth of material in the form of videos, and concise articles/posts here and elsewhere. I don’t know where he gets the time to write any code! I put in a fair amount of time getting to know the way around it and the masks in particular are really powerful. I did find issues however. I found I was having to put a lot of work into each image, and just wasn’t getting the results I wanted. The initial render bore no resemblance whatsoever to JPEG from the camera, and try as I might, I could never get it anywhere near that (I tried the Fuji simulation profiles to no avail). In particular colour management was an issue that I never got to the bottom of. There are also a lot of legacy modules which apparently should be avoided if using the newer RGB workflow, but it’s not obvious which ones. There’s also frequently multiple ways of doing something, which can sometimes be a good thing, but if memory serves there’re I think 5 different ways of converting to black and white. Some of this may be a Fuji XTrans files thing as older Canon and Panasonic files were much less problematic, but Fuji is what I’m working with.
Thus after about 6-8 months I decided to give ART a try (I think it was on around version 1.2 at that stage). I was immediately impressed. Initial rendering was usually close to what was needed and usually only needed a bit of tweaking for shadow and highlight detail. Even difficult contrasty images at the edge of the dynamic range limits of the sensor could frequently be rescued (Fuji files are very good at pulling back shadow detail). The concerns above are valid I guess. There is little documentation though I find that most things can be figured out from the RT documentation and RawPedia, and I find the UI fairly intuitive anyway after spending a couple of days with it. Yes, Alberto is the sole developer, and assuming that continues to be the case, the future is obviously tied to his enthusiasm or ability to continue development or support. But this is a risk for even commercial software. Google in particular are notorious for terminating services or products with little notice. Even Adobe with their subscription model (yes I know I can pay them every month from now to whenever, but…). What I will say it that he’s very responsive and helpful when asked questions, and I’m very happy with the results I’m getting (and the ease with which I’m getting them) for now.
The beauty of this is that I can try something else. vkdt
mentioned elsewhere in the thread looks interesting. I hadn’t been aware of it and I imagine I’ll have a look at it at some stage. That’s the beauty of open source - if something is unmaintained or bloated, anybody can come and fork it and potentially create something better
Until two days ago I didn’t know I could contribute to this thread, as I have been for a few years now a happy darktable user.
But as I’m looking to replace my shitty Fuji camera that I’ve started to hate with a Canon mirrorless, I was aware of the inability to read .cr3 files in dt. So I downloaded Art just to check a few raw files made with Canon EOS RP, R5 and R6 that I got off dpreview.
And well, apart from my evaluation of the Canon files (*), I was truly impressed with Art! A nice UI, fast and responsive, no need to dig into manuals to understand how it all worked, I was able to modify and make a truly nice black and white conversion with grain and mood in no time for example…
I mean, the point of all this is to say: good work @agriggio!!! (and of course all the RT developers out there…)
If I hadn’t already invested time and resources into organizing and tagging my library in dt, Art would be my first choice! And even if I struggle sometimes to recommend dt to friends and people that are not nerds as me, I’d have no hesitations in recommending Art instead.
(*) I can’t believe that people truly believe that Fuji files are better than these full-frame Canons! And I’m referring to the EOS RP, the ugly duck…
Reading the first post I don’t understand why RawTherapee isn’t in the equation.
The last two years I’ve been dabbling with open-source options to LR/PS. I’ve tried darktable extensively. Heck, I even think I finally get Filmic RGB.
The fast development of darktable inspiring. And while there is a lot of power to be found in the scene referred modules I have also found them to be quite complicated for what I can achieve using them. Having worked as a full-time retoucher for close to 10 years I’ve come embrace tools that helps me reach the results I’m after as fast a possible. Sometimes I feel like darktable’s complexity almost becomes an end in itself.
I recently gave RT a serious go again and I must say I have such an easy time getting the results I want. In many ways I felt that this thread kind of summed up my experience with RT and darktable. Users are fighting to get good results tweaking darktable and then an ART user comes by and pulls off excellent results with not much effort. It should perhaps be added that I’ve found highlight recovery in darktable to be an Achilles heel, so it might not be the best example.
A raw converter without local adjustments is unthinkable to me, but local adjustments will be landning in RT 5.9. Some people complain about the near infinite scroll that all the different tools in RT introduces. Having collapsed the ones I don’t use, quite literally, I don’t see the problem.
darktable is awesome and vkdt will be expand on this awesomeness, I’m sure. I’m just saying, don’t count out RT. It’s quite excellent (and will be even better when LAs land). We should be happy to have three (or more!) very complete open-source raw converters.
My experience is similar to others.
Darktable is a very well engineered programme and I have got the hang of setting up a very pleasing starting point using filmic with not many adjustments. However I often find that I seem to have to spend a long time fiddling around to adjust images to how I want them to look. Most likely, it is my lack of familiarity.
With ART, my experience seems to be that I can nearly always fix any image in less than a few minutes. I don’t know what algorithms are used in the tone equaliser (I believe it was taken from an earlier incarnation of darktable) but used in combination with log tone mapping and contrast, I seem to be able to fix even the most difficult images producing very few artefacts.
I just hope that the one man band act continues to be supported because it is an amazing effort.
The Local Adjustments in the dev version have recently been enhanced by @jdc with a full-image mode and the addition of Ciecam to the Log Encoding module. It gives very nice results imho and you can easily fine-tune the result by simply moving the RT Spot around slightly.
If you dont mind me asking, what are raster drawn masks and local editing? I’m new to photo editing in general, so I don’t know what I’m missing from Lightroom and Capture One.
The way you are telling it, it sounds like a big jump if those things come to darktable/Art.
Raster masks are the typical masks that you’d find in programs like Gimp. Proprietary raw editors usually have those or offer the same experience with their vector brush masking.
This allows you to layer opacity and layer strokes of varying opacity to achieve the look you want. It’s fast, it’s precise, it can be done with the mouse or graphics tablet etc. Basically you paint the mask instead of having to deal with countless overlapping paths.
Local editing is just refers to the ability of the software or the user to focus one one particular area of the image and work on that.
As an example I’ll use tone equalizer. When you use it without a mask it works fine, it takes into account the whole image and create the tone mask and histogram based on the entire image.
But if you use it with a mask, let’s say you just use it on a face of some person wanting to brighten the eyebags underneath their eyes, you’d still get a global histogram and the tone mask based on the entire image and you wouldn’t be able to focus on the face and brighten the eyebags.
Oh, you better believe it would be a big jump. If would be an “anything is now possible” feature.
Currently I’ve found (and many may disagree) Darktable suitable for global edits and subtile changes. But as soon as you start really working on the image you’ll find out how hard it is to get it to mask something the way you want. But if you had a proper brush mask it would be the easiest thing in the world to just create any mask you want to do whatever you want.
Many things that are trivial to achieve in Lightroom like brightening blacks and shadows or bringing down highlights and whites are incredibly hard to do in Darktable without loosing all the detail or color or both.
A proper brush masking would make anything possible. Some dodging and burning would still be easier in Lightroom but at least there would be a way to achieve it in Dakrtable. You can do anything if you can just paint the mask. You wouldn’t need to fight the software all the time to do what you want because there is no algorithm to fight, you just paint and that’s 100% predictable and reliable way of doing masking and local editing.
And while parametric and vector masking is very nice and in some cases saves you a huge amount of time that you’d have to spend with a raster brush, it still only allows you to do what it can compute based on it’s design, and more often than not you’ll be fighting it.
Thanks for the explanation @KristijanZic, need to give it a try on GIMP now.
Hope these features come with that vulkan version of darktable.