Force Surface UV to be non-uniform?

[d_b (David Burpee)]

Is it possible to adjust the underlying UV structure of a surface so that the spacing is non-linear / non-uniform? Illustration below:

I am flowing a complex pattern onto a surface using an unrolled surface UV, but I would like to control the density/spacing of one directional parameter (say, U) while the other remains constant.

Ideally this could be done using a distance attraction setup / graph mapper so that the density can be controlled easily and create a smooth gradient from high > low density for example.

The pattern I am flowing is non-periodic so not sure I can simply split the surface into non-uniform rectangles. Also the surface I am flowing onto is doubley-curved, these are just simplified for illustration purposes. Any thoughts?

[scottd (Scott Davidson)]

Using panelingtools is the correct way to approach this. Using UV comes with many problems when trying to be accurate.

Here are some threads:

Converting Polysurface to a Solid Surface

That surface could be a SUBD. but for patterns, especially accurate patterns that are use on Architecture, Using Surface UV for spacing is not a great technique. Part of the reason is not all shapes can be represented by one surface. Even if the surface might be the right shape the UV spacing is not too accurate. NURBS surfaces want to be 4 edges. The shape shown there would be difficult to find 4 edges. A much more flexible and accurate paneling technique would be to use paneling tools; …

And

Nurbs through points doesn't have even spacing

Paneling tools for grasshopper allows spacing beyond consistent surface parameterization. Here is a post about it:

[d_b (David Burpee)]

Hi Scott,

I am not mapping a repeating / tileable surface.

My assumption is that Flow Along Surface is doing some interpolation to the target surface in UV space. I would like to know if there is a way to control a bias with respect to the UV space mapping so that the pattern is, for example, tighter in the U direction on one end of the surface and wider on the other. This would allow one to compensate for a surface that dramatically changes dimensions from one end to the other and maintain a (mostly) square U to V spacing relationship as shown below:

Again, I am using a square grid to represent interpolated UV values (0.0, 0.1, 0.2, etc.) but the pattern I am mapping is non-repeating.