darktable has seen some major changes in the past few years, moving away from a "display-referred" to a "scene-referred" workflow. Aurélien contributed a lot of code to make that work, most significantly, the Filmic module. darktable is not as user friendly and as polished as other commercial tools (Lightroom, Affinity, Capture One) but it is capable if you take the time to learn it.
What is that?
Edit: they have a page about it: https://ansel.photos/en/workflows/scene-referred/
https://www.youtube.com/@AurelienPIERREphoto/videos
I had no idea he was working on this.
Having come across his rants on Darktable last year, I think he does have a point on the UX front. A lot of the filtering in the light table (where you organize your photos) is a bit of a mess of confusing buttons, options, and weird convoluted abstractions. I never liked that part of the software. I can work with it but years in, it remains counter intuitive how you get your files in there and manage them. It's just convoluted and weird all over. A lot of bad ideas layered on top of each other. Aurelian kind of freaked out when last year some pretty major changes were just pushed through without much debate. And I agree with him, it didn't really improve much things.
Anyway the magic is in the darkroom part, which is the part where you edit photos. There is a wide variety of modules aimed at different expertise levels. Scene referred mode is basically a big upgrade over what a lot of other packages do, which is to blindly apply pre-defined curves for cameras without much regard for the actual pixels in the raw image. Filmic and other modules do this a bit more intelligently by actually looking at the pixels, using some heuristics and working from the lowest levels of the pipeline all the way up to do the right things. It adds up to a lot less work when editing photos for me compared to earlier versions. Mostly photos come out pretty OK without much tweaking. I might tweak perceptive saturation a bit, add some contrast in filmic, etc.
Basically, the workflow is roughly: 1) tweak the exposure as needed for the gray point. Filmic adapts with sane settings for black and white points and you typically don't have to tweak that. 2) add some contrast in filmic 3) maybe add some local contrast 4) in the color balance module fiddle with perceptive saturation. Done. There are a few more things I do for sharpening, profiled denoise (as needed), etc. But that's pretty much it. One nice thing is being able to apply defaults based on rules to photos.
I might play with Ansel a bit if I can find some time over Christmas. Kind of curious to see what he's done to lighttable and the rest.
I'm perpetually in the process of beating this dead horse, but FOSS would have so so many banging gold-standard user-facing apps if they enlisted the help of experienced UI or maybe UX designers and really worked to make them part of the community. To do their job right, they need to talk to the community to figure out what their needs are, and if maintainers shrug their shoulders while the few people who speak up are skeptically bikeshedding everything they say into oblivion, then we're also going to shrug our shoulders and walk away. That's what happened to me the several times I tried to contribute to various FOSS projects as a(n experienced, professional) designer rather than a(n experienced, professional, former) developer. Often, the response you get for merely intimating that something could work better if it was set up differently is like calling someone's kid ugly.
It would be like a team of civil engineers working on a restaurant design scoffing at an architect that specializes in restaurants offering to help make an effective kitchen layout. From the civil engineers' perspective, the architect's input is superfluous and would probably slow down progress. Meanwhile, everyone else that has to interact with that kitchen suffers.
Darktable offers a few really powerful, generic tools that you can use in different ways to get different effects – things like equaliser, parametric masks, LAB curves, etc. It makes little sense to use it without reading up on some of those more advanced tools first.
Lightroom, in contrast, focuses more on offering a small selection of pre-defined tools for specific purposes. But once you want to do something outside of that (parametric masking is one of those things I really missed) you're shit out of luck.
But as a Darktable user, I found the implementations of strong opinions frustrating because I value actual workflow above potential technical superiority.
Display referred worked fine for me because the important work happens before the shutter is clicked. The skill I want to develop is fixing things in the lens not fixing them in post.
Breaking changes suck.
But again open source developers don’t owe me anything.
What this gives you isn't necessarily the ability to "fix" things in post, but the ability to decide how you want your image presented in a certain medium or format when "developing". Even master photogs like Ansel would take liberties when developing to realize their vision from the negatives they were able to capture.
It is, at least of I go by my dad who is a Photoshop veteran, quite different to what he used to.
The only things I could see as, if I am really really critical, are:
- sharpening, usually decent enough but SharpenAI from Topaz is another league. As I said, good enough for all except the edge cases
- de-noising, same as sharpening, astro-denoise works like a charm so
- masks, which I haven't really figure out yet, so my problem and not darktable's, Lightroom and Photoshop seem a tad more intuitive so
One thing I'd love, but again maybe I just didn't find it yet, is the possibility to export the database of picture edits. For now, I have darktable create dedicated files once a photonis edited and I make back-ups of those. I did loose edits for around 100ish photos due to some unrelated issues which required reinstallation of darktable so, which again was on me for missing darktable stopped creating said files after an update. Being to just dump the darktable database of edit data would be really nice so.
Summary, I can only recommend it. Especially for people starting post processing, the learing curve for each software is basically the same after all, and without prior knowledge they wont recognize any differences.
Shots fired!
It has so much promise but the 'lead by committee' approach just resulted in some kind of collective 'demand avoidance' from the devs who seem to revel in delivering an unusable product. And I mean unusable for those not willing to learn an entirely new paradigm of interacting with a piece of photographic software and dive into the docs for everything, including stuff as silly as a keyboard shortcut or move between modules.
I've yet to read all the Ansel blurb but I'm pretty sure this is from the guy who's made the most improvements to Darktable in recent releases. So it's incredibly exciting to see.
I doubt it'll get any DAM capability though, even for a fork that is asking too much :-)
The whole "if you don't like how a program works, you can fork it and change it" thing.
Amazing to see this actually working in practice. You might not like Ansel, but he does, and more power to his elbow!
But then I care about the results, a nice picture to print or look at on a screen, way way more than about the tools used to get that result. Goes for cameras and lenses and whatnot as well.
Anyone know what happened to the Darktable project? I've only used it a few times, but it seems nice for someone who knows how to use it (which isn't me!) but curious what drama happened.
Personally, I really like a number of the features that this fork removed (like the timeline). The software does have a bit more of a learning curve than the commercial variants out there, but darktable is impressively capable and I personally jive more with its theory-based approach.
I _do_ agree that deprecating the display-referred mutations is a good thing, but the darktable documentation is already pretty adamant about scene-referred being the future; though there's no way people are going to go through their entire library and update their past tweaks from display to scene-referred, so I don't know that those modules can ever be removed.
I guess if I was going to be as petty as the author of this fork, I'd say that the Ansel logo looks awful.
That's a far cry from what I'd find acceptable in any project.
As I said in my other comment, Aurélien has strong opinions. He felt he could not effectively work with the other darktable devs and decided to fork the project.
Are you referring to the fact of there being no releases in the last year or so?
Edit: Ignore the above - it's incorrect.
4.4 was released in June 2023: https://www.darktable.org/2023/06/darktable-4.4.0-released/
4.6 is slated for December 2023: https://github.com/darktable-org/darktable/blob/master/RELEA...
Their release schedule has been remarkably consistent.
The documentation actually has a pretty good example workflow that flows through mutations in an order that makes sense for most cases. Though it's not really that simple, and so you'll find yourself trying to relearn each time if it's something you do only on occasion.
src = fetchgit {
url = "https://github.com/aurelienpierreeng/ansel.git";
rev = "f7669af89a71882ebad15982d698b8df7e6c6ce8";
sha256 = "sha256-FI6dKUrmtTG7DIV0MmY6XdqlUpqdt7boKuXKU6CywjA=";
fetchSubmodules = true;
};
An example of friction in darktable:
- I have an external strobe which means I have to put the exposure down to its lowest when shooting, otherwise everything is washed out. Darktable in newer versions has "Compensate camera exposure" on by default, which washes out all the images until I click it off. I'm sure there's a way to make this checkbox disabled by default, but why can't it accept what comes out of the camera?
- No favourites: there used to be a way to have all your panels in a favourites tab. This was great as I usually only use a handful of modules that I use. It's gone in later versions
- The "color balance" panel, not to be confused with "color balance rgb", it's not in any of the default tabs but useful for saturation adjustments. Why are some of these useful modules hidden? Shouldn't all modules be available by default. The only way you can get to it is by searching.
- White balance: there are now two modules and it warns you if you adjust one or the other: "white balance" the standard one on the "base" tab and "color calibration" tucked away on the "color" tab. Both modules are turned on by default, but if you adjust one or the other without turning one off it has a big red warning.
- One upgrade decided to reset export settings, and so my EXIF data was stripped out when exporting. It took me way too long to figure it out.
You can organize modules into profiles and simply hide all the ones you don't use. The default profile hides some of the deprecated or display referred modules. You can change this.
White balance indeed has a deprecated variant for display referred and a scene referred one that works completely different that you typically use together with color calibration (which is where you should do most of your color correction, including color temperature changes). The reasons are mathematical and beyond me to explain properly (Aurelian does a great job on his Youtube channel). It boils down to not throwing away the baby with the bath water in terms of rounding errors accumulating and switching color model (to the one used by your display) too early in the pipeline. It might look pleasing but then it bites you when you want to tweak tone or do other things. This is the whole point of working with the scene referred modules.
Having all the legacy modules around is indeed somewhat confusing and Aurelian solves this in Ansel by hiding all the deprecated modules now. They are there for legacy files still.
I wish I could use it just for the development and use something else for the management, but last time I tried that wasn't really possible. I have looked into contributing several times, but it's written in java and I really don't the have time to invest in getting up to speed first.
https://petapixel.com/2022/07/30/ansel-adamss-interview-with...
“There’s no end in sight. Electronic photography will soon be superior to anything we have now. The first advance will be the exploration of existing negatives. I believe the electronic process will enhance them. I could get superior prints from my negatives using electronics.
“Then the time will come when you will be able to make the entire photograph electronically. With the extremely high resolution and enormous control you can get from electronics, the results will be fantastic. I wish I were young again!”
Ah, the oft-repeated refrain of those who never stop learning.
My current Linux photo processing tool is Another RawTherapee [1], which is a wonderful blend of the power of RawTherapee with a UI that has a lot of similarities to Lightroom.
I wondered if that is something hard to avoid: Over time features get added, usability worsens. This might be general, but it could apply to FOSS in particular, which often takes a modular approach (cue UNIX philosophy) and rarely includes structures where potentially useful features are rejected in order to keep the whole experience simple.
``` $ sh build.sh --build-type Release --install --sudo --clean-all;
In file included from /Library/Developer/CommandLineTools/SDKs/MacOSX14.0.sdk/usr/include/cups/http.h:39: /Library/Developer/CommandLineTools/SDKs/MacOSX14.0.sdk/usr/include/netinet/ip.h:189:2: error: unknown type name 'u_char' u_char ipt_code; /* IPOPT_TS */
```
...etc
I don’t mind a few gtk bugs.
Jeez, wish I had 1/10th of the confidence of this guy
The rawspeed lib is at github here: https://github.com/darktable-org/rawspeed
I'm also glad he is taking donations, since the darktable project won't take any. It gives me hope that this would give him the freedom in actually implementing his vision.
I think Aurélien's snarkiness is quite distasteful and might prevent getting some of them on board. But he has a proven track record and I believe he can execute his vision, so I will be donating as well.
About the Ansel fork, I've used Darktable for 6 years and tried Ansel the last two weeks: if you have never used Darktable, Ansel UX is better for a first time user imo, more streamlined. Keep in mind that I'm biased as I agree with most of the author's points here: https://ansel.photos/en/news/darktable-dans-le-mur-au-ralent...
I sincerely hope that this is an effort that will bring the state of the art forward - the author's stance on scene referred workflows is a good premise
- UI is snappier - Integrating syncing lets me leave my laptop at home when travelling and taking photos.
Wow, I'll have what he's having for confidence.
On the other hand, open-source photography and bitmap editing is still waiting for its own Blender or Godot and I applaud anyone willing to have a go at it. What's available (GIMP, RawTherapee, darktable, ...) can mostly sort of get the job done, but if you're the kind of person who seeks relaxation, pleasure and aesthetics in photography, the open-source software feels just too geeky, too unfocused, too unrefined.
I'm currently on DxO + Affinity and even though they cost quite a lot of money, I'd probably shoot and enjoy the whole thing significantly less if I had darktable and GIMP waiting at home.
It's pretty common with open source projects. The people with time, energy, and an "itch to scratch" and the people who really know an area aren't disjoint sets, but people who fit in both are rare. Lot's of projects like e.g. GIMP start off making mistakes that would have been obvious to someone who understood the domain well, but the team makes them because they are learning as they go.
The amount of control and physics-based editing you're doing is very unique.
This wasn't what darktable was like a few years ago, but this guy has been making incredible strides to implement his vision.
He means well and is deeply passionate about this stuff. He's worked his ass off in Darktable for many years. One of the few (only?) developers that was trying to do this full time on donations, etc. He's the closest thing Darktable had to a benevolent dictator style leader (like Linus Torvalds). Big loss for the project to not have him around. I hope they can reconcile their differences somehow. Would be best for all.
Basically, what happened is somebody pushed through something that in Aurelian's mind were some severely misguided, low quality changes without much debate or process. It just showed up and he flipped out and did not appreciate being bypassed/ignored like happened.
Nikon "High-efficiency RAW" support is missing. This is IMHO the fault of Nikon and their vendor TicoRAW. If you're going to come up with a new RAW format, then it is your responsibility to commit the decoder to open-source libraries! Sure, patent and license the encoder, but if you keep the decoder closed-source and proprietary, then you're a <insert expletive>.
The installer is an EXE instead of an MSI. Publish an MSI! Use the "wix" tool in your build pipeline, it's not that complex.
The installer and the deployed app are not digitally signed, which throws up a litany of scary warnings and errors. Other open-source developers have gone to the effort of signing their builds.
On first launch the app flashed a command-prompt window up and then disappeared.
The "Exposure" tool has an automatic setting to compensate for the camera offset. Okay... then why did I bother to offset the exposure just to have ansel undo my intent by default? Then... it layers on a default +.7 EV for like no reason...
The overall GUI is completely non-standard and bizarre. I've never seen anything even vaguely like most of the UI controls in this application anywhere else... ever. It's like someone who's never seen a GUI in their entire lives invented everything from scratch.
The non-standard UI elements like the combo boxes aren't just weird, they're buggy too: they don't work consistently with the mouse. I can see the option being highlighted, I can click it... and then it'll pick something else! They only work reliably with the keyboard. In general the "selected item highlight" appears to be off-by-one, but not consistently. It's bizarre.
Colour management is a mess. On a HDR OLED monitor, Lightroom can display wide-gamut and HDR images correctly if the desktop is set to HDR. E.g.: if toggle an image from SDR mode to HDR mode then the only difference is that highlights get brighter and some extreme colours become more saturated. Everything is correct by default and SDR tones don't "shift". In ansel, the colours look wrong and any setting I choose in the menu makes them even more incorrect.
There's a lot of talk of HDR support on the site, and the menus even "suggest" that PQ/HLG support is there... but not really. The export formats are all from the 1990s and modern HDR formats like HEIC, AV1F, JPEG XL, etc... are missing in action.
A lot of the options/alternatives seem like developers being too reluctant to get rid of old, bad code. For example, if a new superior demosaicing algorithm is added, then it's usually best to just drop the worse ones! I tried every option available, and all but one was broken, to the point of returning super-green results or just all-black. Lightroom for example just uses a really good demosaicing algorithm and doesn't burden users with a choice of a bunch of bad alternatives.
In general, the same criticisms apply to ansel that applied to Darktable: over-complicated, too many options, most of which are either wrong or useless. Internal details exposed to the end-user in the UI that should be debug traces for developers, not permanent GUI elements. Easy to inadvertently "break" the whole thing by reordering pipeline elements by dragging and dropping something accidentally.
I suspect that on top of the ~30K lines of code Aurélien Pierre deleted from Darktable 4.0, another 60K should be deleted...
https://chantal.aurelienpierre.com
might be the real hidden gem here. It's neat!
Tried the reverse: One related keyword for "bokeh" is "out-of-focus". I typed "out of focus" into both Google and Chantal. Google: first result mentions bokeh. Chantal: Also mentioned somewhere, but not as a related keyword.
Interestingly, using it as an exact keyword seems to improve some of the links but just empties the related keywords list.
This feels like an approach that can work really well for topic-specific searches (despite the hole you spotted) but that probably won't ever generalize. Still... I'd have liked to have it for a couple of topics that I used to patrol forums and post basic answers for.
My camera can send raw files to my phone wirelessly, with an ios app I can edit them on site and have a draft ready instantly. This is especially useful for events or anything time sensitive - but convenient in other use cases as well.
Still, I have good memories of Darktable and the author’s filmic module, I hope to try Ansel anyway at some point.
It is good to see some other alternatives out there.
I couldn't get over the steep learning curve of darktable, along with what I perceived to be bloated ui and function
So much pent up anger - and I understand! Darktable has a strange approach to evolving their software. They somehow keep everything old and bad as they add improved versions and you end up with tons of ways to do very similar things, and to add to this mess, Darktable often uses its own nomenclature, or a very very technical one. If you're coming from Lightroom or Capture One, you feel like taking your first steps on a new planet. For no reason!
So, I'm sold! I'll definitely take a look at this. He gets it!
Then perhaps that's not the best choice for cross-platform GUI framework?
not super related but I made a simple tool to publish apple photos to the web.
[edit] Make that quite a lot of the changes, actually, including getting rid of some really dangerously bad bits of UI design. Still dislike the guy's attitude but I've got to admit he has a point.
To me, Darktable has a 'feel' of a highly technical editor for people who really care about colourspaces etc, and although I hardly know anything about colour I rather appreciate that. It's interesting to note that almost all the modules that contribute to that 'feel' are written by Aurélien.
[1] https://ansel.photos/en/news/darktable-dans-le-mur-au-ralent...
EDIT: Am a professional photographer who frequently interacts with other pros on assignments and their own art projects. We're discussing tools constantly and have no qualms between paying for commercial tools or paying in the form of dedication to ascend the learning curve of a less friendly tool. We'll figure out a tool if it's worth it.
Source: https://github.com/aurelienpierreeng/ansel/blob/master/READM...
Maybe Pierre can pull it off himself. He’s clearly knowledgeable and has strong convictions and Darktable is incredibly powerful, so I’m hoping the good bits from either camp can be reconciled into something lasting, minding the bus factor.
Darktable’s issues and lack of coherent stewardship is quite noticeable.