Darktable Tricks

Thank you! I love the softer vignetting. Here is my quickie attempt, before, after the technique, and fully edited (including applying Norman’s denoise trick above). IMG_3461IMG_3461_01IMG_3461_02IMG_3461.cr2.xmp (12.0 KB)

2 Likes

My current favorite trick to get postcard landscapes is to

  1. Use three exposures in base curve sparingly to get a bit more detail if there is a lot of dynamic range

  2. Apply a gentle base curve like the Leica preset because the next step adds contrast

  3. Use the filmic tonemap operator with maxed or high detail. Set blend mode to multiply and reduce opacity to taste.

  4. Enjoy very saturated postcard look.

1 Like

Thanks Cran!
Here is my attempt at using your technique:

IMG_3462.cr2.xmp (8.4 KB)

2 Likes

Here’s another one of Harry Durgin’s tricks I must have missed until the third viewing:

Using a parametric mask with spot removal module. With a high contrast border, create the spot removal area and adjust the parametric sliders to restore a portion of the underlying image. See Harry’s example in this vid at about 20:20 minute marker.

https://www.youtube.com/watch?v=JF5CFQPgidk

4 Likes

I’m not sure why the fence posts became green but the sky looks like it is supposed to.

I use a Mixture of Exposure and Color Zone to even out Fade in skies.

Settings (Masking should be identical for both the Exposure and Color Zone modules)
image
image

Use the “H” parametric mask to select only blue sky region, and not mountains, trees, etc…

Make sure to change the “L” parametric mask to be subtractive, then use the eye dropper tool to help you set up you 100% and 0% sliders as shown above.

For me, typical exposure adjustment is somewhere between -0.05 and -0.2
In the Color Zones module I just boost the blue’s saturation (with the same mask and parametric masks applied as were used in exposure module) until the regions appear to have similar hue/saturation.

Using “Small” view can help to see the Fade more pronounced.

Example below is a merge that came from enfuse.
Before:

After:

Mask:

This can really save images that get “Halos” from HDR or enfuse.

Hope this helps others.

17 Likes

Great trick — I played with this technique after shooting a Parade where I had no choice but to shoot into the sun. When highlight recontruction failed, when grad density filter failed, this worked very well. Thanks!

Keep em coming folks. I am loving this!

1 Like

Do you guys think we could have a common, shareable, easy to edit resource where we could put all these tricks? Something like a github repo (*), where we could organize all the tips/tricks in this thread, link to youtube tutorials for example, include screenshots etc?

I have a note on my computer with a number of these tips, recipes, notes from tutorials I watch, and I believe other people do the same. If we can collect all these bits of information together that would make a valuable addition to the excellent (but sometimes not up-to-date) online manual.

(*) I had this idea looking at this repo – the subject matter is another one altogether but it doesn’t matter, it is to show how it’s done, a simple readme.md file: https://github.com/softwareunderground/awesome-open-geoscience

5 Likes

I think it is a very good idea!

/Claes

We have several shared resources in our github organization, but we don’t get man commits at all.

You’re totally free to put this in an article on the main website (also a git repo) or I can start a separate repo for it. I’d prefer the former, though.

Maybe it could be under the “awesome” umbrella as well?

I like this idea alot as well, but am not versed in github. Looks like a wiki, kinda…

When you’re using github for editing a guide in markdown you might take a look at Grav CMS. It’s a open-source lightweight (no database) flatfile CMS that uses markdown files for its pages.

There is an git-sync plug-in available that can connect to git and make a website out of git’s markup files.

One example is the Grav CMS documentation at https://learn.getgrav.org/

On top of the page you see a “edit this page” link that opens the corresponding markdown file in git.

Grav just needs PHP, so you can put the files of one instance in a subdirectory of your website and place the documentation there

you can also just submit articles to the pixls.us page :smiley:

we do not have a CMS yet, but IIRC @patdavid was playing around with netlify for the RT page, which uses a similar processing as the pixls.us page.

For now we just need markdown files + images at GitHub - pixlsus/website: The PIXLS.US website

Or you add more documentation directly to darktable’s documentation documentation or the website

1 Like

Hey thanks everybody for the tips on how to make this markdown file a bit more easy to read.

The thing about adding it to DT’s official docs is that somebody need to approve it/publish it/etc… I mean I am trying to use github also for collaboration but it still work in progress for me and I’m not that good with pull requests etc.

Anyway, all of this will come later, first I need to at least clean up this first draft, add a few more tips etc.

And this is the draft I have compiled with the help of @Claes (we were debating about which format to write the tips, i.e. simple bullet points vs tables etc), I will gladly accept any sort of tips and meta-tips (=tip on how to write a tip)

2 Likes

Hey @pphoto thanks for this, now I wonder how is Grav different from pelican () that I use for my blog()?

(*) just realized that also DT’s website is built on Pelican!

After a first short view on pelican it seems that both do have a similar approach. File- and markdown based, both with a template system.
Pelican uses python, Grav uses PHP which makes it compatible with many webhosters.
Pelican seems to be more command-line oriented, while Grav does have own GUIs for editing pages and administration including updates, themes and plugins.

1 Like

Thanks, @darix. I did now know about these repositories and possibilities for everybody to contribute.

As I see we’ve got the following resources:

  • Darktable Documentation
  • Darktable Blog
  • Pixls.us Blog
  • Pixls.us Article
    All use gitlab repositories for storage. All use markdown format, except DT documentation which is in Docbook.

As @paperdigits mentioned, all of these resources could need more input. But for darktable-specific tricks or tutorials we would mix this content with other types. DT documentation contains explanations of the program, DT blog release notes and other info, Pixls.us got content for other programs than DT.

So I think that starting a new repo for DT tips and tricks would be the best option. For making usable output like a searchable website or PDF document there seem to be various possibilities.
As mentioned before, Grav CMS is able to do that. Moreover, I did a quick search for open-source Knowledge Base systems which use markdown and found:
https://ourcodeworld.com/articles/read/782/top-5-best-node-js-based-open-source-self-hosted-web-wiki-knowledgebase-applications

Moreover, we could need a repository of example RAW files for that. So everybody who reads a tutorial would be able to follow and reproduce. I suggest keeping the number of files as low as possible (preferably one image for multiple tutorials) and the image’s filesize too for easier handling (shoot the picture with lower resolution)

1 Like

I’ve also been trying to scrape tips and tricks for many software for a freely licensed photo book that I’ve yet to git push to any end point.

I’d suggest not using a table for this, as the information is not tabular in nature. Perhaps a list, or paragraphs…

I agree. There should be a defined content structure. An example:

  • Metadata
    Tag(s): exposure, brightness
    Used module(s): exposure
  • Content
    Title: Increasing exposure with the exposure module
    Description: This tutorial shows how to increase a photograph’s exposure using the exposure module
    Procedure:
    — Activate exposure module
    — Move the ‘exposure’ slider right to a positive value until the image is bright enough for your taste. Watch out for overexposure, for this you can check the histogram and the over/under exposed indication
    Result: (small image with before/after comparison)
1 Like