bulia byak wrote:
Actually it's hard to say how well your interpolation works, because it only produces one intermediate path. Can you take the number of steps as a parameter, and create the specified number of equidistant steps? Then it will be easy to see how good is your algorithm based on how smooth and natural is the transition.
Absolutely, that was one variation I thought of. I will do this right away. Interpolation Time will only mater if the number of steps is set to 1.
Still I noticed one problem: load the attached file, select the left and then right path, and interpolate at 0.9. The result is close to the left path but only half as long, the bottom half is missing.
Have you tried blend groups in Skencil? I played with it for a few minutes to see what it was you were talking about. Skencil truncates the interpolation at the shortest path. I don't like it either. But it made sense to me because of the requirements for animating "d" attributes of SVG paths. SVG requires that the source and destination path data contain the same number of segments and have the same commands in the same positions. (At least that is what I remember from when I last checked but that was a few years ago when I was editing SVG exclusively in notepad.exe.) This first draft interpolator one ups that requirement and works even if the segments aren't the same type (though it ignores A and Q nodes for the time being). I would be happy to make this work with paths of varrying lengths but I will need some help thinking up the best way to do it. Suggestions (and patches) welcome. :)
Aaron Spike