Dear List,

I was hoping someone could help me with the new drawn mask behaviour
that I recently noticed in 3.x. I am using 3.1.0-git279.9ecbf1787

On earlier versions like 2.4, drawn masks would work like this:
1. click on drawn mask
2. For example, take a circle with radius of solid circle 1 and dotted
circle 2.
3. If I click display mask (yellow stuff) the inner circle would be
opaque yellow and between the solid circle and dotted circle there would
be a falloff of translucent yellow
4. If I invert the polarities, the places in the image with X yellow
(where X is between 0 for not there and 1 for opaque) would now have 1-X
yellow -- precisely what I would expect from inverting polarity

Now, in the latest version of darktable, I have this very weird behaviour:

1. click on drawn mask
2. Take a circle with radius of solid circle 1 and radius of dotted circle 2
3. If I click on display mask (yellow stuff), now the inner cirlce isn't
even opaque yellow -- I now have to adjust the mask opacity in a new
slider to about 0.70 to get it to look similarly to the old mask
4. If I invert the polarities, the X to 1-X doesn't even work any more.
I have to readjust the "mask opacity" to about -0.70 to get the effect
that would have happened in the old darktable

I truly do not understand the logic of having so many mask controls, or
why the default drawn mask is no longer 100% opaque in the solid circle.
Moreover, having 4 new sliders to control various aspects of the mask
seems like overkill -- everything that can be done with these new
controls can be pretty much done with the old darktable with fewer
sliders and less fiddling.

Please, is there any setting or way to get the old darktable behaviour
with the drawn masks?

Thanks!
Jason

On 2020-04-03 10:47 a.m., Romano Giannetti wrote:
> Greetings!
> 
> On 01/04/2020 04:08, Guillermo Rozas wrote:
>>
>> However, there are really few instances in which both would disagree.
>> If you keep all your edition on darktable, the most probable reason
>> for disagreement is xmp file corruption and non-updates because of
>> power failure, and then the database should be the most accurate
>> source.
>>
> 
> In my case, this is more or less the norm. I use darktable on three
> different computers; the photo repository is shared and kept in sync
> between the three (via unison, which is a kind of rsync). Keeping in
> sync the database is not viable (different computers, very difficult to
> avoid corruption if I have two darktable left opened on two computers,
> etc).
> 
> So I rely on the "xmp scan" features, and always give preference to the
> sidecar. It has worked ok the last couple of years, provided that the
> version of DT on the three machine is always the same.
> 
> Romano
> 
____________________________________________________________________________
darktable user mailing list
to unsubscribe send a mail to [email protected]

Reply via email to