CP Deleter

[robcat2075]

Today at Live Answer Time we were looking at my attempt to write a program to help thin out meshes... for use with Transfer_AW

After some more monkeying I got it to survive a more significant test case.

• the left image is the original model
• the middle image is a copy of just the splines I want in my thinned mesh. However, this still retains many CPs along the splines that I don't want.
• the right image the result of processing by the program. All the CPs not at spline intersections have been deleted.

Still to do... getting it to complete the first spline loop.

@ChuckGram

@Roger

@Rodney

 

[Rodney]

Nice.

That is a very considerable (and successful) simplification!

 

It isn't clear to me what we need to do to allow the program to maintain curvature a little more strictly.

I suppose we could attempt a (user adjusted) tolerance value where if a CP is found within a certain distance from one marked to be retained that new set of CPs  is removed and if outside of that tolerance it is marked to be the next CP to measure tolerance from.

In a way this almost looks like a reverse 'loft' or sweep.

This particular kind of shape at least.

 

Aside:  I thought I got my attempt at that program to close the end loops but who knows where I put that code.    

[robcat2075]

I think I've got it! the program correctly re-establishes spline loops even if the CP that originally carried the isLoop flag has been deleted.

 

I need a nap before I test this anymore.

 

On 12/1/2024 at 4:06 PM, Rodney said:

It isn't clear to me what we need to do to allow the program to maintain curvature a little more strictly.

Somehow the magnitude of the remaining CPs would need to be increased to puff out the curvature. However, these donut shapes are a special case where the deformity is most obvious.

Quote

I suppose we could attempt a (user adjusted) tolerance value where if a CP is found within a certain distance from one marked to be retained that new set of CPs  is removed and if outside of that tolerance it is marked to be the next CP to measure tolerance from.

 

That sounds like the right idea.  Fortunately, for the purposes of creating meshes to use with Transfer_AW, the curvature of the splines won't matter. I don't think it matters for the interpolation of weights from one mesh to another and the thinned mesh is just in-the-studio stuff, the audience never sees it.