So, I’ve been slowly working on improving the Czech translation of Darktable, as a lot of the messages, names, labels etc. do not make much sense and/or have been machine translated.
It’s a tough job. Not because of the amount of dirty work but because the professional terms Darktable uses may not have a 100% correct counterpart in the target language. So far I’ve been able to find some ways around it with the help of the mighty ChatGPT, which (or perhaps who) is happy to come up with much more ideas than I am capable to think of.
But then I’ve stumbled upon the display and scene -referred terms, and that’s where even this machine-assisted approach has failed. They just don’t translate very well into a sensible form.
The thing is that although I know what those terms mean in practice, neither the originals nor the long list of more or less vague translations do a good job at being self-explanatory. And I know there’s always the RTFM! approach, but that I’d like to be the last resort option for a newcomer.
So this has made me thinking. Are the display-referred workflow and scene-referred workflow the exact terms you’d want to throw at your best friend who only wants to start editing their photos? Could any of you think of any better options?
Now, I know that it’s unlikely for this to change in the English version of DT, and I’m not asking for that (or am I?). But I’d like to get this discussion going anyway because we all are aware of DT’s learning curve being very steep, and every helping hand is appreciated.
And also, I’m hoping to collect a couple more ideas for the translation!
Thanks for reading and posting your thoughts on this!
Scene-referred and display-referred are terms commonly used in the video world, so if you can find a Czech language video editing forum, they should have that part of the answer You could also try to find articles and such that cover the ACES workflow. Doing so would probably also help with many other terms.
While there are some other synonymous terms, those two are by far the most common. Honestly, it’s one of those things where, no matter what you call it, it will require the same amount of explanation, so may as well just follow the crowd. And the common alternatives aren’t really any clearer, while potentially being more context dependent.
I know the basics, and I’ve even found some Aurelien’s explanation on Reddit, and passed it to ChatGPT for it to better understand my query, but the result still wasn’t anything to write home about.
I’m trying to place myself in the position of someone who comes from a less advanced software. So far, my best was something like “limited dynamic range workflow/pipeline” and “unlimited dynamic range workflow/pipeline” accompanied by an explanation tooltip, which would then include either of the original terms, and maybe even a reference to YRGB vs ACES. That one I would be able to translate directly. But it’s too long for my taste.
Do you know of any other /perhaps commercial/ editors having the both of these options so that I could see how it’s dealt with in there?
As explanatory terms, that works pretty well I think, even if they don’t have the exact same meaning.
Darktable is to my knowledge the only raw editor with a scene-referred workflow (though ART has some parts). Certainly it’s the only one that exposes it to the user. You could have a look at DaVinci Resolve, although I don’t know if it’s translated into Czech or even uses those terms in the UI.
While those terms cover an important part of the difference, there’s another related difference:
in a scene-referred workflow, a pixel value is linear to light energy;
in a display-referred workflow, a pixel value is linear to the log of light energy (which is closer to how our eyes treat light energy difference).
A direct result is e.g. in adding brightness (say you want to add 1EV): in a display-referred workflow, that means adding a fixed value to all pixels, in scene-referred, you multiply pixels with a fixed value.
(A similar difference happens with partially transparent masks. That caused artifacts with the display-referred workflow, and is one of the reasons the scene-referred workflow was developed).
And of course, the “unlimited” part is only valid within the pipeline, the output is still limited to the same range in both workflows (at least with current implementations of filmic and sigmoid)
As you’ll probably have to add explanation tooltips anyway, it might be better to stick to terms as close as possible to the original English ones, unless those would have unfortunate associations in your language. Translated terms which stay as close as possible to the original help when users later go to videos in e.g. English.
may help gpt to find a proper translation by giving some more information into the prompt - and also let gpt explain stuff to checkt, if the translation is aware of the content ;):
e.g. you are a technical documentation translator working for photographic and video applications. in some applications theres a differenciation between scene referred and display referred workflow. https://ansel.photos/en/workflows/scene-referred/ contains a quite detailled explanation of these terms. please give an explanation of the differences and give an appropriate czech translation of "scene referred workflow" and "display referred workflow"
In French the terms are “relatif à la scène” (relative to the scene) and “relatif à l’affichage” (relative to the display). Maybe you can use a similar construction in Czech.
Guys, one more thing. Display has several meanings in English. In the display-referred workflow, am I correctly assuming that it means “in relation to how things are displayed” and it has nothing to do with screen?
In Czech, the direct translation (as in DT 4.8) is “pracovní postup s odkazem na displej”, and not only it is unacceptably long, but also, when translated back to English, it is more like “workflow with a link to the display” which makes little sense.
I think “referred” is the main problem with the terminology for me. I’d go with “display-based process”, “scene-based process”. Or use workflow instead of process assuming that’s ok for translating. Or if “display” is a problem, use “monitor”? Actually I think “monitor-based” is indeed better, for one thing, we display prints and that’s a red herring. Can you translate red herring?!
I don’t think we should invent new terminology. Anyone wanting to search for background information is more likely to find it if the naming is consistent with other software / fields (e.g. video), even if the current names are not ideal. Also, the folks who came up with that name probably also considered other terms, and had their reasons to settle on their choice.
This does not fix my problem, though, because there’s no correct Czech equivalent (that I know of), and I’m only fixing the existing machine translation.
I understand your pain. 30 years ago, when we translated OS/2 to Hungarian, much of the terminology simply did not exist in the DOS/Windows world, and we had to invent a lot ourselves. Some of it survived, most of it did not.
The only thing I found online in Czech was a thesis, and that used the English terms:
Display referred workflow
Obrazová data jsou transformována do barevného prostoru displeje (např. Rec709), ve kterém pak dochází ke všem úpravám, což znamená omezení dynamického a barevného rozsahu dat a tedy i omezenější možnosti úprav. Toto je nejčastější pracovní postup při color gradingu, kdy kolorista pracuje s obrazem tak, aby výsledek vypadal co nejlépe na jeho referenčním kalibrovaném monitoru.
Scene referred workflow
Barevná reprodukce je založena na skutečných světlotonálních informacích zachycených kamerou nebo generovaných počítačovou grafikou. Účelem je přiblížit se co nejvíce přesné barevné reprodukci zaznamenané “reálné scény,” aniž by došlo k omezení dynamického nebo barevného rozsahu. Teoreticky by jakákoliv kamera namířena na tuto scénu byla schopna v tomto workflow vytvořit identickou barevnou reprodukci (limitována jen svými technickými parametry). Scene referred workflow jsou obvykle v lineárním barevném prostoru, kde jsou hodnoty úměrné skutečným intenzitám světla.
(https://theses.cz/id/kkwq60/1_1_Pavlicek_Bakalarska_prace_Archive.pdf)
Maybe there is no point in finding native translations. I feel that by translating everything on the UI, we rob users of the possibility of learning, and make it harder to search. The main text of the users’ manual is a different question, of course.
An appropriate term that’s not typically used is ‘magnitude’. ‘Scene-referred’ and ‘display-referred’ is about the image data, and the relationship of the data magnitudes to the phenomenon/device. ‘workflow’ is about the ordering of operations to respect the magnitude relationship.
Well another little nuance is that DT is now a sort of hybrid. At least that is how I see it… A true display referred pipeline would be more what you saw in older versions of DT when the Basecurve was early in the pipeline and subsequent module edits worked with the transformed data…Now it is actually in the same place as Filmic and Sigmoid so really you just have to “manage” exposure as it won’t remap that as sigmoid and filmic do but before that it is the same if you also use the CC module. So legacy display is likely more like a true display referred and the version today is slightly modified…at least that is how I see it…
Thanks everyone for your opinions. I’ve replaced the translations with the originals, and hopefully, nobody will miss them. There’s no localized manual to support their usage anyway, and as far as my experience goes, Google wouldn’t help, too.
