I've found the
being very useful to mask things like lines on a map, and that kind of things.However, when I have a underlaying bitmap with high contrasts (like games from early 90's) and using the
, the resulting object won't be as sharp as the pixels in the image below (ie, the resulting objects node handles does have a little distance from the nodes).This means that when "Grow/shrink by" = 0,00 px, the pixels that the resulting figure was meant to cover isn't completely covered by the resulting object, just because of the rounded corners.
In practice, this is not a problem as I can increase the "Grow/shrink by" to more than 1 pixel. That way I'll make sure that the object is completely covered. In some rare cases I have this issue that I have to choose if the resulting object should cover too little or too much (ie it covers other pixels that it shouldn't when "Grown/shrink by" > 0,00 px)
Is there some settings that force the resulting object to replicate (sharp) pixel structure more accurate?
Attachment is just an example on this. When zoming in to the flooded object, you should clearly see that the flooded object isn't as sharp as the white pixel area on the background image.
Another observation I did is that using the
on very small group of pixels (like a smiley or a letter) will actually create a path that is nearly (or is) complete replica of the pixel group to fill onto. Look at the letter P on atached file.