SubD Multipipe

I’ve been playing around with the Multipipe feature and finding it quite robust. Every now and then it fails but that’s only about 5%. Rendered out in Maverick.


Very nice!
and we’re keen to improve on those 5% of cases where it fails - it would be useful to have the curves that give bad/wrong results.

I’ll try to re-create.

Here are a few more which worked.


Here’s one which didn’t work. (49.8 KB)

1 Like

I split all the curves at their intersections-

got them all to go at .01" except the yellow ones.

@DanielPiker any ideas?

I was however able to do the flat curves, and then splop it onto the sphere-

Not sure what you men by ‘Splop’?

splop is a command that lets you “plop” geometry onto other geometry-

1 Like

Very cool. I did notice in using multipipe, that when the angle between legs gets small that there are these overlaps. I would think that SubD would just blend the bits in, but it looks like it isn’t Has anyone else seen this?

multipipe crossed example.3dm (420.2 KB)

my goofy file (attached). :slight_smile:

@DanielPiker has not implemented any complex joints to this tool so that is as expected st this stage of development. you can always bounce to box mode and pull that spot into shape as needed.

In this image, I deleted some faces, then stitched them closed and then removed some edges to improve the smoothness. (remember in SubD, less is more… remove edges till the shape falls apart, then put the last one back in)


I used “pipe” of T-spline a lot and there are much more options to control the multipipe. Wish SubD could provide the similar options in the near future.

agreed, remember, this is very early in the development of these tools… now that they are 1st class citizens, the development will be ongoing indefinitely!


Thanks Kyle. I figured that was the case. Just want to make sure McNeel was aware. I would imagine there will be some algorithm that dictates what happens when curves intersect at different angles. Looking forward to the change. Cheers, j

@DanielPiker is the wonderful mad scientist behind this, and I believe has further plans to add this type of functionality. (Dan, please correct me if i"m wrong)

1 Like

Pretty sure I found the cause. if the curves do not absolutely intersect, even if below the document tolerance the pipe will fail at those apparent intersections. I’ve managed to fix a few by zooming in extremely close (with 2 point perspective switched off) to the intersections after splitting the curves and manually co-locating all four end points…when I run multipipe again and it works. Does the command run on its own tolerance independent of the document tolerance?