AWS Thinkbox Discussion Forums

Anisotropic Mesh Stabilization?

Is there anyway to stabilize this type of meshing based on a value? I really like the results but it is very difficult to use on any non-uniform systems.

For instance a straight particle stream with no speed variation or divergence in particles yields a really nice mesh. The moment any speed or direction variation is introduced the mesh destabilizes and can jitter quite bad. You will get gaps that only last a frame.

The idea being able to slow down such rapid changes in the mesh based on some value would be really useful for this mesh type.

Ludicrous idea? Not possible?

The sheeting Anisotropic creates is great just unstable. Can nothing be done to better this? Or is there a different/better way?

Yes, we have an idea for a better Anisotropic mode. However it would not be a tweak of the existing one, but a completely new mode.
Right now it is just an idea. Not sure when it might come to fruition, assuming it actually works.

So right now Anisotropic is slow due to the sampling of neighbors, and unstable over time due to the change of the neighbors and their influence. We can make the existing one faster, but not more stable. It really looks good only on static meshes. Then again, we couldn’t even get it to work with Hacksaw in Sequoia so we removed it from it, too…

It is the red-headed step-child in the Frost family :mrgreen:

Haha okay. Fingers crossed.

I kind of figured if it would have been possible in the current iteration you guys would have already accounted for it.

Red hair is unique and wonderful :smiley:

Privacy | Site terms | Cookie preferences