On 12/3/06, Jasper van de Gronde <th.v.d.gronde@...528...> wrote:
I do have a question though. I found the relation between the radius entered in the fill & stroke dialog and the actual value of sigma used quite odd (sigma=radius*400/(width+height)), so I was wondering whether there was a reason for this, because it makes it quite difficult to predict the result of the specified blur.
This is the end user control, and therefore the formula is designed to make most sense for the user, not the programmer :)
When using blur, I'm rarely interested as a designer in setting a precise radius. Instead I tend to think in terms like "slightly blurred" or "heavily blurred", and these terms are clearly relative to the size of the selected object. If the blur slider was setting an absolute radius, it would be very hard to use for small obejcts (where even a slightest movement of the slider would cause great changes in visible image) and of limited use for large objects (where even the maximum value may be not enough). Therefore it uses percentage. Taking width+height as a base is an attempt to define blur as percentage of a vaguely understood "visual size" of the object - surely the "size" is not the same as as width+height, but it correlates better with that sum than with just width or just height, and slightly better than with sqrt(width^2 + height^2) because a square looks "bigger" than an oblong rectangle with the same diagonal. Finally, the 400 coefficient is there so that the maximum 100% corresponds to the maximum useful blur radius, which turns out to be approximately 1/4 of width+height; anything larger will not alter the image in any meaningful way, just make the already amorphous blob larger and lighter.