Vector masking broken with unfilled mask

[mqudsi]

It seems that masking is broken in the case of an unfilled mask (i.e. fill color: none) under 1.6.2

 

no fill masking bug.afdesign

[Sean P]

Hi mqudsi,

I think you've misunderstood how these features work and have incorrectly set up your document. Both of the objects on the left hand side are actually vector clipped. This means that any object placed inside the parent object are clipped to them. Removing the green fill from this will just show the fill of the parent object's fill.

The right hand column are actually vector cropped. As the bottom object has no fill in that cropping object it will not show the content of the parent object. However as the stroke does have a fill, it is showing that area. 

Take a look at these two pages:
https://affinity.help/designer/English.lproj/pages/Layers/layerClip.html

https://affinity.help/designer/English.lproj/pages/Tools/tools_crop.html

clippingcropping.afdesign

[mqudsi]

Thanks for the reply, Sean. The terminology is eluding me, but (referencing your images from here on out) I'm not understanding how the 4th image (bottom right) is correct. Each row has the same operation performed on it, each column has the same items/color.

 

The operation performed in the second row, as demonstrated in the first column, results in the intersection of the two shapes being retained. The green fill of the rectangle does not appear anywhere in that resulting image. When the only change that is made is that the green fill is changed to no fill, we go from "intersection of the two shapes retaining the fill of the red rectangle" to "stroke of second shape in the color of the first, and the intersection of the two shapes with the (no) fill of the second shape (rectangle)."

 

I am extremely confused how, if the only difference between operations 3 and 4 is the fill of the second shape which does not appear in the final result per image 3, do we end up with a completely different result.

[MEB]

Hi mqudsi,

You are confusing clipping with masking (which Sean is calling cropping). 

- In the first row you only see what's inside the triangle: on the left a green rectangle with a black stroke; on the right just the black stroke because the green fill was removed from the rectangle.

 

- on the second row the triangle is being masked by the rectangle - you are only able to see the triangle through the "filled/visible" areas of the rectangle (no matter if it's the stroke or fill): on the left both the fill and stroke of the rectangle allows you to see part of the triangle - because both stroke and fill have a solid colour; on the right just the stroke of the rectangle allows you to see part of the triangle because it's the only thing "filled with a colour" (it corresponds to the intersection of the stroke of the rectangle with the triangle area)

[Bri-Toon]

Clipping is putting objects inside one another. Vector masking (not to confuse with pixel masking), is cropping them. So let's say you took a square shape and masked it on the top of a triangle. The square is a different color than the triangle. After masking, you now only see the top of the triangle cropped down to that square shape. That shape, which was a different color, is now converted to a square shaped mask. Now if the square was just a stroke and no fill, then it is the stroke that gets converted. That is why the forth drawing in your example looks like that. It got cropped down to the part of the shape that was visible, the stroke.

[mqudsi]

Thank you both for explaining. I understand the difference between cropping and clipping, but it was @MEB's point about visibility's role in all this that cleared things up. I now understand (but still disagree with) the results that AD presents.

 

It is a difference in how AD treats objects vs how other software might. As a vector-based program, I would expect AD to treat a filled object and an unfilled object exactly the same, except in how they appear (i.e. unfilled is "filled but with transparent background"), whereas for a pixel-based program like AP or Photoshop (where masking comes from in the first place) it makes sense for masking to work on colored vs uncolored.

 

But when it comes to vector objects, an area "is present" when it has an enclosed path, not when it has color. For a rasterized/flattened pixel layer, AD should treat masking the way it does: by colored vs uncolored. But for vector objects, the definition of "existent" is fundamentally different, and imho only whether or not a closed shape exists and not whether it is filled with a color or has no fill should play a role.