I’m slightly puzzled at this as I find it very user friendly - I just hover the cursor over the image and it shows the line on the graph (in the advanced tab) so I know exactly what’s what.
Not meaning to be critical, just curious as to how it could be improved… I suppose it’s very much down to personal preference.
1 Like
I set both checkerboard colours to black, which gets rid of the chexkerboard in favour of a solid colour. Much better.
4 Likes
It was so long ago I’d forgotten you could do this. Thanks!
I guess that means no change to darktable required.
3 Likes
How do we change the checker board to solid black?
I’d forgotten too!
left-click → select from dialog
Unless you change defaults.
Remember that in this module the clearly-visible (and in their normal colour) parts of the image are those that are included in the chosen mask, and the (more or less) coloured areas are excluded.
For example, in the following image I have chosen to view the highlights mask with a yellow “checkerboard” colour, and it’s the non-highlights that are shown in yellow. The yellow areas are those that will not be affected (or will be less affected, since there are few pure yellow areas) when I change the highlights in the module.
This is the opposite to how normal masks are shown. For example, here I have selected the highlights using a parametric mask. Here, the yellow parts are those that will be affected by changes in the module:
I think this is the main thing that throws me when I use color balance rgb
3 Likes
Thanks for all the great answers here. I can certainly live with the current ability to change the mask color and I am sure I can adapt my brain to realise the mask is opposite to what I would have expected. A colour mask like yellow or red seems better than a grey or black mask to my way of thinking. So my problem is solved.
However, I like the idea of the heat map for some other modules, but I am not sure which?
1 Like
I’ve played around with different colours, single colours, different checkerbox sizes, and it still doesn’t give me a clear visualization, especially the shadows mask for some reason. I don’t use the masking feature that regularly, so it’s not a big issue for me, but I still don’t think the masking feature in this module is particularly clear.
Maybe there is room to improve the visualisation of the masks. But that would take one of the developers agreeing with this and proposing a solution. I like the idea that the three masks could have different colours because there is an overlap between masks which is important to the effective working of the module. But being able to change from the checker board as explained in this thread is a huge help to me. Thanks to everyone who indulged me and replied to this post.
Here’s an example of an all-black mask with the shadows fall-off mask enabled:
And here’s an example of an all-yellow mask with the shadows fall-off mask enabled:
I don’t know about anyone else, but I find it really hard to see what is being masked and what isn’t. And this is a properly exposed image.
@elstoc’s earlier examples were much clearer for some reason, maybe because there was clearer separation of the tonal ranges? My image is a typical landscape snapshot with a wide dynamic range and lots of different tonal values.
Seems like pretty much the whole image is considered to be in shadows. Try adjusting the mask middle-gray fulcrum.
Adjust white fulcrum first too. Use the auto picker.
That’s the crazy thing - it’s a properly exposed image of a hillside on a sunny day. Surely the default mask middle gray fulcum of 18.45% is ok? Even when I adjust that fulcrum and the white fulcrum, the whole image remains a shade of yellow with very little separation visible between the “clear” parts and the masked parts. The checkerboard actually creates better separation in this photo.
Well, just above those sliders and buttons is a graph, showing the influence of the different masks as a function of luminance.
Each of those curves is above zero for most of its range. That means that for each mask all tonalities are influenced in varying degrees. The “shadows fall off” and “highlights fall off” sliders limit that influence on the “other” tones.
But they both have a fixed point at the “mask middle gray fulcrum”. Thus, there will always be an important overlap between the different masks. And that’s a good thing: if you have a gradient that starts below the fulcrum and ends above is, you would get sharp edges if the masks did not overlap…
So there’s nothing “crazy” about the masks selecting a large part of the image, it’s designed to behave like that…
Also, note that the checkerboard colors show the areas where the mask is transparent, i.e. the yellow shows the pixels that are not influenced by (in this case) the ‘shadows’ sliders.
Yes, I just realised that I got it wrong in my previous reply. I should of course have said that the whole image is considered to be in the highlights 
I thought I didn’t know, then I thought I did, now I know I don’t
2 Likes
I just realised a neat little trick to make sure you have mid-grey where you want it, when adjusting exposure.
Enable mid-grey mask in color balance rgb before adjusting exposure.
Maybe, maybe not. That depends on if you have correctly set exposure in the first place (in DT, not the camera) and what you have done in the pixelpipe between exposure and color balance rgb. Tone eq, notably, can easily move mid-grey if you’re not careful.
What exactly is it you don’t understand?
4 Likes
Yes, nice trick!
@rvietor, I am aware of everything you pointed out in your last post. I wasn’t confused about how the masks work and why they select a large part of the image. Maybe my phrasing gave that impression. In fact, I don’t have a problem at all with how the masks function, just how they are displayed. In my earlier example of the all-yellow mask and all-black mask, it’s very hard to see which parts are masked and which parts are the unmasked image coming through, at least to my eyes.
I find the checkerboard works better than the single-colour masks, but I’ve always found the masks in this module hard to read. Part of that is because of the smooth gradients needed, but this thread is to discuss if there is a better (more obvious) way of displaying the masked areas. Maybe there isn’t…