On 01/17/2012 12:48 PM, William F. Maddock wrote: > I can picture in my mind what a drop shadow should do to a layout, and that > allows me to surmise certain things. I have seen folks here proposing the use > of bitmapped graphics as a workaround. Unless that bitmapped graphic is the > very bottom thing in the stack, that work around is not going work right. > Particularly, it won't work to fake a shadow for text. It's going to mask out > things that should show through the shadow?and how that thing shows through > the shadow will depend on what part of the shadow is above it, and how far > above it. Usually, transparency is a set value per object, but in the case of > a soft shadow, the transparency needs to vary *within* the shadow object. > Think about watching the sun through mottled cloud cover, and you'll see what > I mean. It's like applying a shape gradient, but instead of grading from one > color to another, you're grading from opaque (on the object's edge) to > transparent (away from the object's edge), and the opaque part of the shado w will actually be smaller than the object that casts it. The farther the casting object is supposed to be from the objects under the shadow, the smaller that opaque portion to the shadow will be and the larger the translucent outer shadow will be. Applying this effect to text is going to make it even more complex (you'll have shadows running into and over and through other parts of the shadow, and other shadows as well). In addition, the effect of the shadow can appear to be abruptly different across different objects, because of those different objects being at different distances from the casting object. The shadowing effect will have to be individually calculated for each object that is under the shadow, and that effect will have to be applied to those other objects?not to the object casting the shadow. In order to do the job properly, it is going to need to be done directly in Scribus, so that no matter where in the stack the shadow is, and no matter what is in the sta ck beneath the shadow, the effect of the shadow will be executed correctly every time. Once the layout is complete and ready for output, the script that Greg implied can be run, if needed, in order to prepare it for PDF/printing. >
Aside from this long-winded description, I don't think that Gimp and Inkscape do drop shadows this way. It seems to be more like a gradual opening of spaces between the pixels of the shadow than also working with transparency. This having been said, we already have gradients in which one end of the gradient may be opaque and the other transparent to some degree. One question is how many settings do you want or need for drop shadows? Greg
