TabletoPoints->Delauny3D vs TabletoStructuredGrid

Help! I am trying to convert a .csv to a 3D plot. The .csv contains x, y, and z voxel coordinates. If I use TabletoStructured Grid, I get:

But if I use TabletoPointsDelauny3D I get:

BUT. The TabletoPoints and Delauny3D filters takes about 10 minutes to load, and for my larger .csv files, it gets stuck after 5 HOURS. This is why I tried the TabletoStructured Grid filter, but it gives me these weird lines across the surface. Is there a way to make it smoother/no lines??


If you click the Reset icon reset_icon in the toolbar, will the artifacts be removed?

Hi Kyoshimi,

No, nothing happens when I click the Reset icon.

I suspect those “lines,” which look to me like small creases in the surface, are just a part of your data. It looks like the points on that face of your data do not all lie on the same plane. You wouldn’t see these creases using the Delaunay filter because that filter, which knows nothing about the structure of the points, fills a convex hull around the points. So, the Delaunay filter fills those creases so you don’t see them.

Hi Kenneth,

I see what you are saying. Is there a way to view the .csv in 3D that won’t take hours? The TabletoStructuredGrid is very fast, but gives me those small creases.


What I’m trying to say is that the creases are the correct 3D representation. Delaunay isn’t just slow, it’s wrong.

There might be some tricks you can pull to hide the fact that the face of your data isn’t flat (such turning off lighting or playing with the normals), but I suggest going back to the source of your .csv files and checking why the point coordinates you think should be on a flat plane are actually perturbed from that plane.

Hi Kenneth,

You are right, I went back to my .csv files and found an error. After correction, the TabletoStructuredGrid produced the same 3D plot as Delaunay. Thanks for your input!