The problem is that with other software I can get beautifull results easily. I suppose due to being more used to them and not being used to the linear editing.
I don’t see any problem in the photo itself. A bit of the sky in the brightest parts are burned (not so much and not a great problem in other soft) and too much dynamic range, but was captured correctly by de camera.
Compressing dynamic range to the range of the screen is a task for tone mapping tools and for filmic itself as it is its main task. In other soft I use curves and masks.
The problem I find is that color balance cannot put enough light in the shadows, and curves here don’t give me good results. I can brighten it but parametric mask do not work well to separate the building and the sky and don’t get halos.
But you have demostrated it can be done. I have to study your edit too.
I like the the sky a lot.
I would have used less contrast and light for the building, but that is a matter of taste, and if you can force it to the point you have done, it would be easier to be not so agressive.
Great edit too, thank you very much. I will see what you have done.
You are right even though I did mask the sky a bit I did go over the top…in a way though I liked my edit …would have just been nice to have a global opacity slider and dial it back…
One thing with DT is often the best approach is several instances of the same module. In many software packages this is not possible. So use more than one if you need also you can push many modules more than the slider defaults…just right click and try a bigger number . I think color balance has some that are 25 or 50 percent…I have a preset that expands them all to 100…So you can bump the sliders often and you can just apply a second or third instance if you need more of an effect…ignore me if this is obvious…
I am a bit overloaded for so many and good responses. Thank you all.
I see one module I was not taking into account is tone equalizer. I have to give it a try. @s7habo: you seem to use it a lot, three times. What is strange to me is that it is used in opposite directions (first one darkens darks and brightens lights and other does the opposite) and you don’t seem to be using masks.
Previously I was using C1 and previously LR (some times a bit of PS).
But a year or so ago (may be version 3.0), I discovered DT and liked it its philosophy and use of masks and ability to use masked modules and use them several times (for that kind of things I had to use PS).
Funny enough one of the things I liked was it was based in LAB, with made sense for me as it was near to our perception.
Later I discovered that was not a good thing for many processing modules as it easily produced halos, and came the evolutions of DT to scene referred mode.
I am sure it is the correct way, but it is getting more difficult than expected with so many frecuent additions, modules that get obsolete…
I have used modules duplicated previously in DT and parameters forced beyond the limit.
But in many photos (specially backlighted ones, this is one example) there are two peaks in histogram, one in the darks and one in the lights.
I try to expand the peak in the lights and darks with more contrast and use masks and parametric masks to mix results and have a more natural look. But color balance many times is not enough to get the result.
I have not used other software in this photo, and I have been using DT too much.
But I can provide other photo which I processed in LR to get a dramatic effect and see what can be get in DT (of course the edition I liked may not be good for others, but the issue if to learn to process skies an other more or less difficult photos in several ways).
Tone eq has its own internal mask for tones so usually you use that and then maybe just use opacity to dial it back if needed but you could also mask it…it just might work against the masking in the tone eq
If you are tweaking with color balance rgb be sure to visit the master tab and define the what white fulcrum and the gray fulcrum as this will define the distribution of the global contrast. The picker will usually be a good value but like all things DT experiment with the slider extremes to gauge the impact
here is my edit.
I have used a second instance of the exposure module with a parametric mask selecting the highlights and I have decreased -1EV.
After that use the Tone Equalizer to adjust the lights and shadows in a “natural” way.
Smashing result! Would their be a way to talk us through your thought process?
I did have a look at your xmp file, but I’d really to know your reasonings behind some of them…
Tone equalizer is a very powerful tool to adjust the dynamic range of the photo. It has a wide range of functionality that can be combined in one or more instances.
No, the first instance refers to the shadows, but does not darken them, but controls the brightness of this area with the help of the mask exposure compensation function, which serves as a kind of brightness fulcrum. Note that the perserve details option is turned off, so you have a black and white version of the photo as a basis for the mask.
Now, with the movement of the slider you can influence the brightness of the area:
The slope of the curve increases the contrast a little bit, which has a positive effect on the flat looking area of the photo.
The second instance of the tone equalizer applies to the lighter areas of the photo while protecting the darker ones.
The third instance reduces the contrast caused by the preceding color calibration modules.
The whole procedure was necessary because the building, besides being the main subject, is the darkest element in the photo, and at the same time it is uniformly illuminated in a very flat-diffuse way (because it is a backlight shot), so it loses a lot of contrast/depth.
This combined with the fact that the photo is also a bit overexposed in the sky area, the need for highlight reconstruction additionally limits the handling of the dynamic range, because you always have to be careful about creating artifacts in this overexposed area.
Thank you! Normally I like to do this, but this has been a rather difficult and complex processing, which is rather useful to demonstrate what can be done with the darktable, but is not one of the usual (recommended) processing strategies. I think this will overwhelm many. But I am ready to answer the questions.
Yes, exactly my feeling. Every time I do some processing in DT I always have this “yes, I finally get it now”-feeling and then I see your work and this turns into “I know nothing”
I’ll have a look at your xmp file and get back to you
