Hi, in CMD you need to be in the same location as your images.
Alternatively, copy the photos+xmp to: DT-Debug.
The command you are using looks fine.
Hey @priort, sorry for the very basic question, but I downloaded your dropbox link and the installer creates a new version of the app that contains the two new modules discussed here. How do I move those modules into my existing install (portable on an ssd)? Sorry for the dumb question, Iām a photographer not a great computer user. Thanks in advance.
You should wait for those modules to be added to dt in the future. The links above are a proof of concept and it is not ready for production. I personally think we should not provide them until it is closer to a beta product stage.
2 Likes
As @g-man notes these are experimental and changing sometimes several times per day. If installed locally or on an external it should really be a separate install using its own config folder and with xmp writing turned off. Then you can safely experiment with it.
I run WIndows so what I do is when I go to install it I say no to have it uninstall the current version because this is referring to your current install and you dont want to nuke that. Then when it comes to the path I just add LC to the end of \darktable so \darktablelc and install this version there.
I then use a duplicate shortcut that refers to this install and modify the command line with --configdir āpathā where path is some empty folder that I can use for the test config folderā¦
This is what I do for testing and its the safe way as edits made with this version may have a short half lifeā¦
Oh, now I get it.
There is no significant difference in speed between DT 5.4 and the trial version.
However, DT 5.4, which is now reinstalled, runs extremely smoothly, while the entire interface of the trial versions is slow and laggy.
grakaLaptop - Dt 5.4.txt (18.9 KB)
grakaLaptop - Dt 5e1bf6cf5c.txt (20.2 KB)
1 Like
Ok, this sounds like the prudent move. Iāll play around with the modules some in the new install that youāve shared, but await a formal release in an update to 5.4 before applying to anything in my catalog.
Thanks for the info and innovation.
No thanks to me, My contribution is essentially net zero. Iām just trying to make these available for testing to Windows users and its a pretty automated thingā¦ The guys behind the modules are doing all the heavy lifting and intellectual workā¦ 
7 Likes
Itās gotten a bit quiet around this topic so I wanted to chime in. I feel like the explorative phase has consolidated and maybe itās time to make some decisions on how to proceed?
The three approaches pyramidal, scaled and colour all have their fans. Would it be feasable to include all three with a dropdown similarily to how we currently have pyramidal and scaled to choose from?
After exploring the visual adaptation (CSF) I personally have not found too much benefit for my photos but would be curious what others experiences were.
I have no horse in this race, and I am happy to leave the choice to @Wilecoyote, who put in the effort to make the PR.
What I am really hoping is that some of the edits in the discussion end up in the docs in some form, eg with a heavy crop, to demonstrate how versatile the module is.
4 Likes
Your efforts are much appreciated Todd. Your contributions help people like me test out these features. I think itās important that users on platforms other than Linux get a say.
@paperdigits, just a side note, but is there a guideline for use of translation tools on this forum? Would it make sense to request that āTranslated by XYZā be added to all posts where the text has been fully translated by such tools? (Some users already do add this note.)
Iām wondering if this would be helpful as a way to notify users when the tone of the text may not be exactly what the author intended, and likewise it helps the author by essentially saying āthis is not necessarily how I would phrase itā.
For info, Iām a professional translator and have extensive experience in this field and with these tools. They can be very good and quite accurate a lot of the time, but they do not use the authorās voice. They often default to āpoliteā phrasing, which is helpful, but they can still sometimes get things spectacularly wrong. Iām also an advocate for maximum transparency where possible.
4 Likes
I think deepl adds the text automatically to the copy buffer, or st least it used to.
We donāt have any hard set policy. Certainly if a user thinks its helpful to add that note, then they should.
However I do not think our moderators should play translation police.
What if we just tried harder to take all messages in good faith, and if the tone sounds bad, assume the person is having a bad day or has misphrased instead of assuming the worst.
I think thatād smooth things over more than ātranslated by X softwareā tags.
Iām also a native speaker, and Iāve been told many times that my messages are āshortā and people assume I am mad. Iād rather that they appreciate the efficiency, compactness and sometimes brevity of my posts 
9 Likes
Yep, thatās the ideal, but of course we know it doesnāt always happen. Poor phrasing gets peopleās backs up. Anyway, no big deal. I certainly donāt want you being the translation police. I just meant as an āetiquetteā guideline to help make things more transparent and as a reminder that not everyone speaks at native level.
1 Like
Iāve just got back from a trip and only just caught up on this thread. Lots of interesting stuff and a new concept from @jandren. I had a little play with it, and itās another potentially great tool. Thanks for this!
I would be interested in knowing the advantages of working with colour channels rather than wavelets.
In terms of usability, I find it somewhat harder to grasp conceptually, and I found I was just using trial and error to get desired results.
3 Likes
Thanks for trying!
I do agree that its a bit cumbersome to use the three additive rgb sliders. But my point for posting it wasnāt first and foremost that I wanted merge that now but rather to give a reason to keep the initial version of this module simple and focus on merging a solid core method. I want to point out the the proposed āpyramidā and āscaledā versions are not wavelet based, not even implemented as band-pass filters, in other words they are basically hard coded internal copies of the original module even if you need it or not. And I get it, its quite convenient but is it worth it?
The price for that convenience is that it will be harder or even impossible to add future new functionality like my example with color based weights.
3 Likes