The future of selections and masks in Krita

Boudewijn Rempt boud at valdyas.org
Tue Aug 1 01:00:03 CEST 2006


Okay... Some more thinking while I was trying to go to bed...

There are two issues:

* Global selection vs. per-layer selection. That's adequately
discussed by Bart's response. It's also not much related
to

* difference between masks and selections

Actually, there is a relation: both masks and selections (and
adjustment layers in my original conception and other property
layers such as wetness, height and gravity) are byte-masks
that convey a certain property to certain pixels of a layer.

- In the case of per-layer selections is the property of selectedness.
- In the case of wetness, it's, well, wetness. 
- In the case of masks, it's visibility (which should be just
another filter, or perhaps a fill with a certain composite-op. In
fact, and in a roundabout way, by creating a layer group consisting
of two paint layers combined with the COPY_ALPHA composite op, masks
are already achieved, or would be, if that composite op hadn't been
commented out in 1.5). And having just visibility, or alpha, is a silly 
limitation. There are more channels that could be independently masked.
- In the case of per-layer adjustment layers, it's the property of
that specific filter.

And just like real artists (and I'm old enough to have done that
myself) use different paper masks to mask bits of a painting
when airbrushing, we should allow more than one "mask" -- but
that holds for the other types, too -- per layer, so they can
be positioned and modified independently.

This is more or less my original conception: I think it's better than 
anything that's currently in existence, that it can be made to work,
and that it's possible to create a usable user interface for it.

Boudewijn



More information about the kimageshop mailing list