recover ooc jpeg detail/sharpness from raw with darktable

Yeah. An option with sharp mountains could fill a big gap in the market. :grin:

2 Likes

What I do use is Spencer Coxā€™s ā€œdouble the distanceā€ method:

https://photographylife.com/landscapes/double-the-distance-method-explained

I just pick the most prominent foreground object, eye-ball the double distance, and center the focus box on it. Works well.

I keep my Z 6 in AF-C, and itā€™s fun to watch it do focus-stuff in the EVF. For the landscapes, it just keeps plotting red boxes around things it thinks are interesting, until a person comes into the scene. Then, it first finds and yellow-boxes the head, then yellow-boxes an eye with little left-right arrows to tell you if there are other eyes from which to pick. In a choral lineup, I swear it knows who are my grandkidsā€¦ :laughing:

2 Likes

That sharpen demosaicing is like magic.

OP: This video from Boris shows a good trick to load your camera JPG into darktable and use it to help mimic the style with your raw image. You can then further develop to get even better the OOC Jpeg. https://youtu.be/8uXK5r8qhxE?t=789

1 Like

ART

1 Like

IMG_1946.CR2.xmp (8.7 KB)

Without trying to match the jpeg this is how I would edit this image
IMG_1946.CR2.pp3 (15.5 KB)

3 Likes

I know in his video that AP said only use the sharpness slider of the diffuse module for the most part to further tweak a blur but while experimenting I used just that slider alone and actually you can get some nice sharpening in my opinion by just doing thatā€¦in any case you can try and see what you thinkā€¦

Not true right ?
ā€˜local contrastā€™ bilateral mode comes with warnings, but the normal one is just fine. Itā€™s just meant to be used after filmic (where it goes by default ) so outside scene referred.

ā€˜contrast equalizerā€™ is fine and made for scene referred. Nothing wrong or deprecated about it.

The thing is, it got used a lot to enhance contrast in the smallest of details , as a detail enhancer and/or sharpener. And for that usage, diffuse can yield better results (although people still stick to something they know for speed other reasons, which is orcourse ok ).

Exactly. And because in the future filmic should be at the end of the pixelpipe, the question is how and whether this module will be modified or adapted.

Not yet. It still works in LAB color space and has to be adapted to Yxy space color space:

"Before filmic RGB, in the linear pipe, we still find some modules that work in Lab but perform linear operations that should (strictly speaking) be realized in linear RGB:

  1. contrast equalizer
  2. high pass
  3. low pass
  4. sharpen
  5. denoise (non-local means)

These modules need to be adapted in the future to be able to work on a linear Yxy space (derived from CIE XYZ) because it is a mistake to make them work in Lab (at least, as a default). Itā€™s a relatively easy job to do, because Yxy breaks down the luminance (Y channel) and chrominance (channels x and y) with a logic similar to Lab, minus the non-linear transformation. In the meantime, you can continue to use them, but with moderation. For the contrast equalizer, note that it uses an edge-sensitive wavelet separation, which makes it quite cumbersome to execute, but very effective at preventing halos, even considering that it works in Lab."

So, both modules can be used, but since diffuse and sharpen module is already designed for scene-reffered workflow, contains the functionality of both modules and is partly far superior in terms of results, I recommend to use this module.

I guess for low pass we can use the new blurs module now?

Unfortunately, no. Blur is limited to 128 pixels and it does not have a de-saturation function.

For this is better suited diffuse and sharpen module, since you can have much larger blur radii and on much better control over different aspects of it. Unfortunately, I also miss the de-saturation function in this module.

If the highlights are not clamped Lab is a superior color space for a lot of operations compared to linear rgb and linear xyy, and totally adeguate for the scene referred workflow

This is not my quote but from the text in the link I gave in the post.

If you want to discuss about it I recommend to create a new topic because we are slowly deviating from the main topic here, and to address it to AurƩlien,who wrote this article.

1 Like

But you have learned something wrong from that article

Done

1 Like


IMG_1946.CR2.xmp (12.3 KB)

My effort as I might do itā€¦

2 Likes

I have to go and search, but somehow I thought it was changed somewhere in the 3.6/3.8 cycle. Could be completely wrong about this.

Filmic is already at the end by default, doesnā€™t mean you canā€™t or shouldnā€™t put stuff after it.
ā€˜local contrastā€™ now has a function you use as a sweetener after filmic, and from what Iā€™ve read i was under the impression that the devs were fine with it as it is. Again, could have misunderstood. Forums here are not always a reliable source.

Still LAB according to darktable 3.8 user manual - local contrast


IMG_1946.CR2.xmp (15.7 KB)

IMG_1946_01.CR2.xmp (13.8 KB)

2 Likes

So it seemsā€¦ Weird I guess I must be completely wrong about this.

But that also means there is no scene referred way for local contrast , except diffuse now (although with the performance cost that seems not practical).

I would have sworn I had some edite where I slammed the exposure , then use contrast equalizer and then filmic and the higights were not lostā€¦ Maybe just a dream :pensive: .
Newborn baby here so my mind is a bit hazyā€¦