Support a special "Luminance mask" layer type #1183
Conversation
DJMcNab
added
enhancement
|
Would it make sense to generalise this new method to allow alpha masks too? I suspect that might be quite convenient for implementing https://developer.mozilla.org/en-US/docs/Web/CSS/mask-image... |
|
I suppose so; it's not entirely clear to me whether that is already covered by an existing blend mode. But this operation has been discussed as blocking for Graphite, whereas alpha masks have not been. Either way, that should be a separate PR to this (but before 0.6.0, probably). |
DJMcNab
force-pushed
the
luminance-mask-layer
branch
from
1903894 to
2a6e445
Compare
| if end_clip.blend == LUMINANCE_MASK_LAYER { | ||
| // TODO: Does this case apply more generally? | ||
| // See https://github.com/linebender/vello/issues/1061 | ||
| // TODO: How do we handle anti-aliased edges here? |
There was a problem hiding this comment.
Presumably they're accounted for in the multiplication of rgba[i] and area[i] above?
There was a problem hiding this comment.
Kind of; area[i] indicates "how much the blending should apply to this pixel".
However, we're instead using it as an alpha mask over the foreground. For Porter-Duff Over, this is equivalent. However, for the "destructive" blend modes, that doesn't give anywhere near the same results.
In the extreme case, consider the case where area[i] is zero; that is the cause of #1061.
There was a problem hiding this comment.
Makes sense. That seems unrelated to this PR and shouldn't block landing IMO. Let's try to resolve that separately.
There was a problem hiding this comment.
As an aside, that issue is interesting because I don't think we anticipated that someone would push/pop an empty layer simply for the side effects of the blend operation.
There was a problem hiding this comment.
Ahah:
Note: Shape is defined by the mathematical description of the shape. A particular point is either inside the shape or it is not. There are no gradations.
DJMcNab
force-pushed
the
luminance-mask-layer
branch
from
cd611c4 to
09eea7d
Compare
4 participants
Supersedes #1121
cc @Keavon does this address your need/use case?
cc @mTvare6, as this impacts your PRs