emPolygonizer3 lagoa cloth meshing

Discussions about simulation, hair and cloth in SOFTIMAGE©
User avatar
mattmos
Posts: 437
Joined: 02 Dec 2009, 16:59

emPolygonizer3 lagoa cloth meshing

Post by mattmos » 26 Jul 2012, 14:16

I've got a tearable cloth lagoa sim which I'm using emPolygonizer to mesh with thickness, and passing UVs onto it. Ideally I'd like to be able to tell one side of the mesh to use a different material, but as its a changing topology I can't figure out how to set a cluster. Any suggestions welcome!

grahamef
Posts: 281
Joined: 23 Jun 2009, 21:01

Re: emPolygonizer3 lagoa cloth meshing

Post by grahamef » 26 Jul 2012, 18:13

Maybe you can figure out which way each polygon's normal is pointing, and use that to set MaterialID.

User avatar
mattmos
Posts: 437
Joined: 02 Dec 2009, 16:59

Re: emPolygonizer3 lagoa cloth meshing

Post by mattmos » 26 Jul 2012, 18:20

Relative to the underlying lagoa point structure? Its folding over on itself so no easy worldspace options.

grahamef
Posts: 281
Joined: 23 Jun 2009, 21:01

Re: emPolygonizer3 lagoa cloth meshing

Post by grahamef » 26 Jul 2012, 19:38

OK then that's trickier -- I was hoping it wasn't the case.

There might be a Lagoa attribute you can read off, otherwise maybe you can walk the mesh. I need to think about it a bit more...

grahamef
Posts: 281
Joined: 23 Jun 2009, 21:01

Re: emPolygonizer3 lagoa cloth meshing

Post by grahamef » 26 Jul 2012, 20:29

OK I presume that the original mesh used to generate the Lagoa is one-sided. In that case, maybe you can store the closest location's normal on the Lagoa particles on the first frame. After the sim you can read that from the polygonized mesh and compare with its own normal.

EDIT: Wait that's not going to work. You're going to need to update that vector using the point's orientation, and if Lagoa doesn't update orientation you'll need to deduce it yourself based on neighbors' relative positions.

User avatar
mattmos
Posts: 437
Joined: 02 Dec 2009, 16:59

Re: emPolygonizer3 lagoa cloth meshing

Post by mattmos » 27 Jul 2012, 17:58

Seem to be getting somewhere by keeping the polygonizer sim one sided and adding thickness with an ice compound, which (unintentionally) is only passing UVs to the front side. Not quite sure what it will do to motion user data I'm also passing through, but hopefully getting closer.