recent macOS arm64 build:
darktable-5.1.0+955~ge36bbfa244_arm64.dmg
same comments as in Blender AgX in darktable (proof of concept) - #181
recent macOS arm64 build:
darktable-5.1.0+955~ge36bbfa244_arm64.dmg
same comments as in Blender AgX in darktable (proof of concept) - #181
is agx module available in the nightly build from OBS? I’m currently using darktable 5.1.0~git925.48486b7f-12869.1
but I cannot find it. I guess is only available in the appimage, am I right?
Nope.
Wouldn’t it make sense to avoid this small overexposed (and drowned) areas, that the autopicker is not exactly trying to match white and black but maybe some grey which is 0.2 EV darker than white (same with black). Therefore, we wouldn’t need to raise relative exposure after using autoclicker. Everything could then be steered with the power sliders.
Just a suggestion, knowing, that probably there are technical aspects which speak against.
I’ve considered it, it’s worth a try.
Wow, this module is really great. The level of control is so much more than competing modules and it’s also quick once you’ve figured out your workflow. I can now get the overall look I want with about four sliders.
I’ve been playing around with different approaches, and I can’t say I’ve found the “best” way yet, but I would agree with @s7habo that setting Exposure first for the area where you want the contrast pivot to be seems to work best with AgX. I know this is basically best practice for all image editing, but I think it’s worth emphasizing for AgX. Sigmoid seems to be more forgiving if your exposure is slightly off.
So far, my favoured workflow is this:
So far, I don’t expect to use the Look section much, but if I want to give a saturation boost, I need to expand this section. I find this to be a slightly clunky step to the workflow. I’m still conflicted whether this slider is needed (I might just use Color Balance RGB) and/or if it’s in the right place, but if others use it and are happy with its placement, it’s not something that bothers me too much.
Great work @kofa and to everyone else providing feedback!
I still use the look section. Even so I don’t change the sliders really often, they give me a base setting for my preferred look and I would miss zhis section badly, if it doesn’t make it to the final release. Without this section, you would have to use another module for pushing saturation, so it isn’t such a big difference. Or did I got something wrong?
There is no doubt and I’m waiting eagerly for the day kofa will say:
That’s one small step for man, one giant leap for darktable users
You can do it with the recover purity settings in the primaries tab I think? You can even separately adjust R, G, and B purity there (with the attenuation reversal
). The original AgX (which did not have the outset) has a punchy look that is the combination of a CDL power and saturation, but the CDL saturation was later replaced by the outset (the recover purity setting). So IMO, the punchy preset should not double the saturation with the outset in place.
Fair enough, and I don’t think there’s a right or wrong answer here. Do you feel the sliders in the Look section do anything unique/different to what other sliders in the module do? Or is it just more of a workflow preference?
I think you could achieve any look as well with the other sliders. But especially for the Offset Slider, I don’t see a direct replacement. For getting similar results you have to move more than just one other slider.
Offset is compressing the dynamic range quite evenly. That’s one of the sliders I move more often (besides saturation). The rest is more a workflow thing I think.
I don’t see how I can set with this slider the amount of saturation. Did I miss something?
It scales the color outwards away from the achromatic point, hence producing that effect.
Probably because you actually used the primaries tab. I usually don’t need it, so I left every slider in place and just used the promoted slider.
I use the primaries only when a colour freaks out.
It’s kind of weird, from your screenshot the slider isn’t even doing anything. Some kind of bug? Anyways, the ability to scale colors outwards is a part of AgX’s mechanism, whether you use it or not.
i think this is a bug. for me, these two sliders (recover purity and master rotation) have never had any effect in any version of darktable in windows.
I actually use that section a lot to refine the “look”… I think I don’t push the curve too hard. I get a global tonemapping and I further I refine things with the look sliders but maybe that is just me I find it fills in the gaps and the saturation slider is great to have on hand to bump or pull back saturation without going to another module in many cases…but each to their own…
If it was taken from the sigmoid code as it might have been I feel that the recover purity was only going to work if there were pre-existing changes to the channel sliders… but maybe I am not recalling that correctly…
EDIT:
But you can see the difference if you take an image and use the sigmoid defaults vs the smooth preset…using smooth you will see the recover purity slider impact…
This means… The default setting disables the core AgX mechanism (the matrix), and the default setting is not one of the presets… I think that’s problematic, the default should be one of the presets.
The attenuation to white is essential for AgX. With the current default, the matrix is turned off by default, that means the colors never attenuate to white by default. This is very problematic, considering many people don’t really touch the primaries tab, they might never know, that they have been using a simply curve without AgX mechanism involved at all.
Your individual reversal settings are all 0
The ‘master’ control just scales the individual ones.