Lets say i have 2 curves. i define the first with @width = 1 and the second with @width = 0.5 and i want to use these attributes as the offset value in the Polyexpand.

How could this be done?

PolyExpand2D in H15 does not support varying widths. This is in the plan for the next release. This SOP works by computing the straight skeleton of the input graph and doing this for varying widths (in the general case) is quite a bit more complicated.

In the special case where your curves won't interact, you can use a for loop in SOPs to process the curves individually.

Ondrej
PolyExpand2D in H15 does not support varying widths. This is in the plan for the next release. This SOP works by computing the straight skeleton of the input graph and doing this for varying widths (in the general case) is quite a bit more complicated.

In the special case where your curves won't interact, you can use a for loop in SOPs to process the curves individually.

I assume PolyExpand3D is out of the question then?

pusat
I assume PolyExpand3D is out of the question then?

Bardia has some thoughts, but I temper optimism with an equal measure of skepticism.

Ondrej

pusat
I assume PolyExpand3D is out of the question then?

I temper optimism with an equal measure of skepticism.

That would make a nice quote

Ping… Has the “varying widths” issue on the polyexpand2d been solved?
I'm experiencing some weirdness.

In the sense that we added support for varying widths, yes, the issue has been resolved. To do this, the SOP computes the weighted straight skeleton, and so there are some limitations. The two big ones are that edges still propagate in a parallel manner, and that collinear edges are forced to propagate at the same speed.

Consequently, differing edge weights between adjacent edges affect the direction the bisector moves. It's much easier to see this when outputting offset curves. Without a hip file it's hard to say if there is something wrong from just the above screenshot, because high relative differences between adjacent edge weights will affect the bisectors dramatically. The top bit of the curve does look a bit odd though.