"Graduated Neutral Density." Gimp vs. RT

Resized from the Raw file’s embedded JPG. Thankfully, there is no clipping in the raw file:

I’m trying to figure how to treat the sky in this picture. RT has a Graduated Neutral Density (GND) effect that works well, but then it would be applied to the pole in the foreground as well and I don’t want that. My choices right now:

  1. Make 2 versions of the file in RT, then use layer masks in GIMP to combine the images on separate layers.
  2. Just do one pic in RT, then apply a GND effect in GIMP. Mask out the pole, maybe the trees on the left.

What would be the best overall way to approach this picture? I am still often torn about how much to do in RT before I move over to the gimp. Especially with Gimp 2.10.2 and its ability to handle 16 bit TIFFs, much of the reason I was using RT for is no longer relevant.

If I did this in GIMP, I think I could put a black layer on top of the image, then a layer mask with a gradient going down to the waterline. For the GND filter layer, would I use normal mode? Overlay? Soft light?

Or maybe there is a better way to do this entirely. I’m open to suggestions.

You could use darktable with a combination of drawn and parametric mask.

1 Like

You can use LH module in the LAB tab of RT. Choose the hue of the sky and darken it.

EDIT: Here is the output - did not do anything except using LH module.

1 Like

That looks full of blocky artifacts here…

JPEG artifacts?

1 Like

Here version with darktable.

2 Likes

@shreedhar Can you post a screen shot of your LH sliders? You say, “choose the hue of the sky, and darken it.” But the sky is a bunch of different hues. Blue up top, white in the middle, and pink on the bottom.

Did you just move the slider on the major colors, or did you use the picker tool that’s built into the LH graph?

@Thomas_Do your darktable version looks great. But I have my hands full. I absolutely don’t want to learn a new raw editor, unless someone convinces me that DT would be better for me than RT. It has been a nightmare learning RT, so I’m not eager to start all over again with new software. I actually hate post processing. It’s not fun for me, but I do it because I have to.

I understand that. I started learning darktable a year ago and later also installed RawTherapee. But the operating concept is quite different and I find it really hard to use both programs in parallel.

I am sorry if I was too brief in my first post. Also, my first attempt was too hasty (to only show proof of concepts) but it lead to artifacts as pointed out by @heckflosse . I became a bit more circumspect and here is a more rounded output using the LH, CH and A curve. In the first two modules, I used the dropper to select the hue.
Have a look:
44bfa0728eb7acb1844ac90443d996f7ff4858e3-1.jpg.out.pp3 (11.8 KB)

1 Like

Neutral Density filters are constricting light. That means the operation is a simple multiply by amount. 0.5 would be a stop, 0.25 would be two stops, etc.

Requires a scene referred linear reference.

I figured I’d check back in to show my final edit of the picture. I wound up starting it in RT, then too it to the gimp where I used a layer mask with a gradient and a solid black layer in soft light mode to get the sky to look nice.

Thanks for the suggestions. Thanks to @shreedhar’s edit, I started messing around with the LAB controls on the exposure tab. I never used them before, but after reading the RT documentation on them, I get what they are doing now, and I’m just starting to realize what powerful tools those can be. Anyways, thanks all, for the help.

1 Like