Sweep with UVs

[Kumar]

Hi everyone,

                  i was watching scott keating's GDC video on procedural modelling, and i have an issue with the UVs and the sweep node, that he shows working with polygons. If anyone can show me a basic sweep with proper uvs, it would be great.

 

[mestela]

funny you mention that:

 

 

[Kumar]

Thanks, for the quick reply. can you explain why the attribute wrangle node @uv[1] = @uv[0] ?

Scott has the arclen function his setup, which i was also trying but somehow never got it to work. He also has a attribute promote which also is confusing me.

Care to explain a little on this.

Anyway, i did a resample and arclen and found the uvs to be a bit better, attached is the updated file.

Sweep_uvs2.hiplc

Edited August 23, 2017 by Kumar
Added updated file

[mestela]

Like I mention in that wiki post, if sweep finds uv's on the input geo, it will copy uv.v from the cross section, and uv.u from the backbone. The uv texture sop on curves always puts the interesting value into uv.u, so you need to swap it to uv.v.

To be annoying, you can't actually refer to u and v with uv.u and uv.v. Vectors components in houdini are referred to by vec.x, vec.y, vec.z. I figured using that syntax would be confusing, so used the alternative array syntax, vec[0], vec[1]. Maybe it wasn't such a good idea.

 

[Kumar]

Thanks for the reply, its cleared up a lot of stuff for me. Now just wondering, how scott managed to pull it off, anyway never got a chance to analyze his entire graph so we pass on.