HCL Mask like selection for local adjustments

Hi,

I dared to create this feature request

Please share your opinions on that here or in RT’s github page … I believe that it would be very cool if local adjustments regions could be based on HCL mask that is already available in Color Toning tool (so hopefully re-using existing code is possible)

cheerds, ~dan

Is this not what you are looking for ?

http://rawpedia.rawtherapee.com/Local_Adjustments#Using_a_simple_mask_to_improve_color_selection

I’ve read that page and tried to use it a couple of times and I have to admit I couldn’t figure out the first thing about it. I don’t think it’s doing the simple thing OP is asking for.

There is a brief explanation in Rawpedia that may help:

"Some important points about the mask curves C(C), L(L), LC(H).

  • These curves are used to create the mask.
  • For each of the curves, the vertical dark-gray/light-gray separation line represents the 3 references of the RT-spot: luminance, chroma and hue.
  • For the first curve shown below (with the highest point of the curve on the selection – L in this case), the deltaE selection is improved.
  • For the second curve shown below, the hue used by the mask corresponds to the hue reference of the RT-spot (the peak of the curve is on the selection – H in this case). Pulling the curve downwards will progressively mask (or reduce the impact) of whatever adjustment has been applied to the selected hue (or L, or C depending on which of the curves you are using).
  • For the third curve shown below, the hue selection for the mask does not match the hue reference of the RT-spot. In this case, pulling the curve downwards will progressively mask whatever adjustment (luminance and chrominance) has been applied to that particular color.

Note that the effect of the combined LC (H) curve can be visualised by referring to the spatial representation of the Lch coordinates i.e. as you move up and down the vertical L axis, there will be a corresponding increase or decrease in the chroma values."

Mask selection

1 Like

guys, I am talking about to have a HCL Mask as an another method for a selection of a region for a local editing … It actually supports only one method which is just Spot (with some feathering on edges)

Spot is here https://infophagia.com/ntz/paste/Screenshot_20240920_101912.png
HCL Mask is here https://infophagia.com/ntz/paste/Screenshot_20240920_124232.png

I would love to have this Mask creation feature from Color Toning tool as a next method for region selection in the Local Adjustments tool …

I hope it’s understandable …

Point is that it already exists so it wouldn’t be so terribly difficult to re-use that feature in another tool (to re-use a Mask feature from Color Toning->Color correction regions within the Local Adjustmens tool)

What everybody is trying to tell you is exactly what you are looking for…

1 Like

An example would be helpful, for example how to increase & decrease the saturation of a color using only a mask in Color & Light.

the spot is set to the sky, so the reference is blue here:

when moving the spot ot the trees in the background, the the scope already is sufficient, for more, LC(h)-Curve easily can be adapted:

1 Like

btw, in most of the cases, you even don’t need the curves, just refine the scope:
(mask ΔE is the green area)


spot on sky


spot on trees


same spot on trees, but smaller scope

1 Like

Thanks, but these are with DeltaE activated right ? If I set the scope to 100 they seem to do nothing. As I understand OP wants masks as a method of selection instead of DeltaE-based spots, not as a refinement.

yes, and the delta E show you, in which area any chances will be applied. But anyway, this only was an additional example for easy use.

guys, dont take me wrong but I understand that I can use a mask additionally within tool (when in advanced mode) … but it is not a same !!!

You can add multiple tools to spot … each can have its own mask … YES I UNDERSTAND to this concept … But what seems that you do not understand is that I want to add another method for “Spot selection” that will HCL mask … Or at least that HCL Mask (as it is in Color Toning will be global (tool independent) region subselection within Spot \wo even adding a particular tool)

1 Like

What I understand is: now you create a spot and each tool you add works based on the structure of the spot. Moving the spot or changing its size will change the effect. The HCL mask would create a region where you can add tools. Effects will be independent to structures. E.g. dodge and burn would be possible (at least for me, as I haven’t found out how to do it in RT :thinking:).

1 Like

yes Kurt :slight_smile: … exactly … mask could be `another’ way to define a region for local edits independent on spot … also what could work is that mask can be bound to spot, I don’t care, but it should be independent on any tool that you possibly add so you could apply mask even before adding any tool into your spot …

Also current mask from particular tool that you add to spot doesn’t have a view mask button like this, which looks very handy

1 Like

Hello everyone

I’m only joining this discussion late, because I think it’s better, in terms of organizational learning, to let users explain and find the solution together.

The explanation you gave wasn’t clear, and I don’t think I’m the only one who misunderstood your request.

I’ll get straight to the point, and then explain why. I don’t intend to implement your request (at least for the time being), but of course I don’t forbid anyone to do so.

A brief historical reminder:

  • Most software (Gimp, Photoshop, Darktable…) has a selection system directly derived from that created by Adobe almost 30 years ago (layers and masks).
  • When the question arose in RT of what to do, I had a Nikon D200 at the time, and Capture NX2 came with it… another way of doing things, based on intuitive Spot (no masks, no layers…). I challenged myself to do something similar, if not better.
  • The system has been in development for 10 years now, and with the help of Pierre @Pandagrapher for GUI and Ingo for code optimization, and others…, we’re now at a stage where the system is working (although there’s always room for improvement, especially in terms of the GUI).
  • This system is basically “maskless”. In response to demand, I’ve reluctantly added masks, which in my view are of marginal use. Nevertheless, they do exist and can be improved. Almost all tools have a dedicated mask (more or less complex depending on user choice) and complement the “U point” selection system.

What you’re asking for (beyond its interest, which I don’t dispute), will increase the complexity of Rawtherapee - this is what many people blame RT for, often by associating it with vocabulary (Ciecam, Wavelets, etc.). Just one question: for a photographer, what’s more incomprehensible: the words Wavelets (RT) or Log encoding (Filmic), or Sigmoid, or the differences between Brightness and Lightness… present in other software? Another comment on the apparent complexity: it’s easier to hide an existing setting than to add one.

The solution you propose resembles those implemented in ART (with masks dedicated to tools) or Darktable, which don’t have the U-point solution.

For “dodge and burn”, you can use “Original Retinex” in Selective Editing (in Soft Light & Original Retinex - Standard or Advanced)… and here of course no mask, no layers, but Laplacian.

And excuse my bad english.

Jacques

3 Likes

Isn’t the DeltaE a mask already ? I.e. a black-and-white image that is used to apply the tool processing to the image, or it’s more complicated than that ?

Since the HCL stuff also produce a mask, wouldn’t just using the HLC mask instead of the DeltaE gets you almost there ? not really in term of interface, but at least in terms of processing it would be possible to do what OP wants.

Maybe I’m missing something but I really don’t understand what problem the OP is hoping to solve by having standalone LCH masks in addition to dE + LCH masks in Selective Editing. If it’s to make RawTherapee more like ART then why not simply use the latter rather than complicating the RT interface even further.

3 Likes

RTspots are not masks; the concept is completely different. In a nutshell, we use the color difference for each pixel (delta L, delta a, delta b - in Lab mode), in the area chosen by the user: “Local spot” or “full-image”.

In addition, transitions are applied from the center of the RTspot, to make the system progressive.

You can place many RTspots, one on top of the other or next to each other. Excluding spots can be used to cancel an action on a desired area.

In Rawpedia:
Introduction
General principes and settings

Added masks are only used in rare cases, where the deltaE system is not sufficient. They too work (in part) with deltaE. They are irrelevant in Global mode . I’m going to make a small GUI modification to make their interface disappear in Global mode (it’s an oversight).

As Wayne @Wayne_Sutton says, the solution you want seems to be ART’s - which works very well.
Adding something will make it more complex, and I don’t think it will bring much benefit. There’s a lot of work to be done at GUI and machine code level, and in this solution it’s one tool = one mask.

Nevertheless, I’m not totally closing the door - it’s only fools who don’t change their minds.

Jacques

3 Likes

Ha I see, it’s not so easy then indeed. In general I’m quite happy with how the deltaE works, although I can also see the appeal of masks in some cases. Thanks for the explanations.

thanks much to everybody for your input … and thanks for pointing me to ART … I know it exists but I never had any reason to use it over RT … Now I have reason (to try it out at least)