This was the same argument I used against having a dedicated Vignette module and it seems like everyone dislikes it
I agree with you
The case of the vignette module is that the control of the mask is very suitable for a vignette and you cannot have the same amount of control with the standard masks. Especially, the ellipse and circle maximum radii are rather limited in comparison and the aspect ratio of the ellipse can only be changed after the fact. Plus, thereās no method to exactly place the mask, e.g., in the center.
The only problem with vignette is that the math for the vignette itself is not suitable.
Sorry for interrupting, though ā¦
Iāve love a vignette module that could be pre-initiated to replicate the vignette of the lens used in the image, so that the image could be corrected, cropped, then uncorrected. With a zoom lens, it would be nice to be able to correct the real vignette over a session shot at different focal lengths (or even lenses), but return a constant vignette to the whole batch for consistency.
Isnāt this what the lens correction module does?
Exactly. Just switch from ācorrectā mode to ādistortā and select only vignetting.

It has a distort mode? I guess Iām so used to using it on automatic that I missed that. Thanks.
I donāt find the Color Zones module complex at all. Itās just a dropdown to switch between the different HSL modes, and the UI remains essentially the same for each mode. I personally think that using masks is more complex, especially for those who are new to Darktable. I think many of us who have been using Darktable for many years forget how daunting the masking panel can be for newer users.
But I respect the decision to leave the Color Equalizer as is. Itās not a big deal for me, even though Iām one of those users that prefers specialized modules and prefers not to use masks unless necessary.
A question about how hues reposition.
In this image, with Harmonizer off, we can see the image has hues aligning with orange and red:
I select ātetradā harmony, with the goal of keeping the orange (skin tones) and red (shirt) in place, but you can see the orange has shifted towards red, instead of staying where it was within the harmony guides.
My expectation would be that the colours outside the harmony lines get shifted, not those inside. How does the math determine which hues stay the same and which get shifted?
all colours are moved to the target set for the harmony - depending on strength and width. The displayed sections in the scope doesnāt imply a āprotected zoneā
All hues are pulled towards the anchor and the harmony hues. The closer they are, the stronger the pull. You can fine tune the angles by switching to custom rule and moving the red hue a bit towards orange.
Hm ok. I would probably use color equaliser for custom tuning. Mathematically, is it possible to anchor each hue according to the harmony (eg. Four anchors for ātetradā), then move the other hues towards the nearest anchor? To me this would be more desirable behaviour.
This could be a nice extension for the next iteration. Something like āprotect neutralā, but for colors that are close to one of the harmony nodes, e.g., āmin pull distanceā.
It might function like a pull amount inverter. Currently hues close to the target hue are pulled more afaict, but the inverter would cause hues further from the target hue to be pulled more.
It is a very interesting way of keeping natural variations in the target hues while still unifying the image.
Though for this new inverted pull formula, protect neutrals will become an even more important control and might need an extra āboostā to keep the transition from neutrals to non neutrals from being too harsh, since those values nearest neutrals would be pulled the hardest.
I think Iāve found a bug in harmonizer - it seems to be using raster masks the wrong way round, i.e. itās acting on the part thatās not yellow.
This is with latest Master - 5.5.0+903~gae456d37b8
Ubuntu Studio 24.04 plasma X11, No GPU
Can anyone confirm pls?
I just saw this thread and came by to say thanks for this module! I had just upgraded my Darktable and saw it sitting there in the module list, and I was thrilled when I opened it and saw what it was for. Great idea! Excited to try it out in my photos.
Works fine in win11 recent buildā¦
Any chance you have a reverse blend mode selected or missed the mask polarity??
Defaults
Thatās almost impossible as the masking feature is not done in every module but generic to the pixelpipe and so shared with all other modules.
Yes, I think I must have messed up.
Last night, to check my sanity, I put a color balance instance next to the harmonizer using the same raster mask and both showed what I expected when I clicked the mask icon in the module header, i.e. not inverted. So I thought there was a bug.
This morning I accidentally went into 5.4.1, so no harmonizer of course. I went into Master but my harmonizer settings were gone, and I canāt reproduce the issue, so as I say, looks like I messed up.
As a side note, I was a bit confused this morning with raster masks. Suppose a module is using one, all fine, then you use the dropdown list to select āno mask usedā. Now the module has no effect at all, whereas it seems logical to apply the module across the whole image, yes? You have to deselect the raster mask icon, the football, for the module to take effect. Perhaps setting no mask should auto deselect the icon?


