Thought I’d share this, today I discovered a filter in ParaView called “Tensor Glyphs” - which consume tensor variables at points (e.g. via Crunch’s PolyVertex output sampling) and makes ellipses representing the tensor:

How do you interpret this? Is this only for symmetric rank-2 tensors like stress? I assume they eigen decompose the tensor to find the principal axes and values. The axes then orient the ellipse while the values provide the sizing of the ellipse.

That’s weird that they have that toggle. I wonder how much the two would differ from each other. I don’t even know how I would interpret it without it toggled to use the eigen decomposition