# Smoothing the output of Rotational Extrusion
**URL:** https://discourse.paraview.org/t/smoothing-the-output-of-rotational-extrusion/6211
**Category:** ParaView Support
**Created:** [January 12, 2021, 5:30pm UTC](https://discourse.paraview.org/t/smoothing-the-output-of-rotational-extrusion/6211 "2021-01-12T17:30:26Z")
**Posts on this page:** 8
**Page:** 1
### Author:  [@woodscn](https://discourse.paraview.org/u/woodscn)
#### Post date: [January 12, 2021, 5:30pm UTC](https://discourse.paraview.org/t/smoothing-the-output-of-rotational-extrusion/6211/1 "2021-01-12T17:30:26Z")
I’m trying to make a visually impressive (as in, I’m willing to sacrifice _some_ technical accuracy) 3-D movie based on 2-D axisymmetric data. I’ll walk through the steps I’ve tried and the reasons why.
I have cell-centered, multimaterial data, where the material is identified by an integer scalar field:
I want to color materials differently, so I will focus on one of those first. I extract the specific material using two Clip filters:
The result is quite blocky, because it is Cell Data, but I get spurious artifacts if I try to use Point Data: (using Contours is similar)
I convert to Point Data and extract the surface in order to perform a rotational extrusion:
This looks good, other than the discrete steps. I try to smooth these using the Smooth filter:
This looks great, except for the seam where the extrusion “connects”.
Does anyone have any ideas on how I can improve this?
Thanks,
Nathan
---
### Author:  [@boonth](https://discourse.paraview.org/u/boonth)
#### Post date: [January 12, 2021, 5:57pm UTC](https://discourse.paraview.org/t/smoothing-the-output-of-rotational-extrusion/6211/2 "2021-01-12T17:57:16Z")
I’m wondering why you would get artifacts when converting to point data. Are you doing MergeBlocks and computing Ghost Cells before applying Cell Data To Point Data?
---
### Author:  [@woodscn](https://discourse.paraview.org/u/woodscn)
#### Post date: [January 12, 2021, 6:00pm UTC](https://discourse.paraview.org/t/smoothing-the-output-of-rotational-extrusion/6211/3 "2021-01-12T18:00:38Z")
Yes, I did both of those (2 levels of ghost cells). I assume that it’s because clipping around mat==2 captures some of the interpolation between materials 3 and 1.
---
### Author:  [@patchett2002](https://discourse.paraview.org/u/patchett2002)
#### Post date: [January 12, 2021, 9:35pm UTC](https://discourse.paraview.org/t/smoothing-the-output-of-rotational-extrusion/6211/4 "2021-01-12T21:35:44Z")
I wonder if you run clean to grid first. Ghost Cells wouldn’t get caught because the boundaries are likely global boundaries on the dataset. I believe Clean to grid would identify identical points and merge them.
---
### Author:  [@woodscn](https://discourse.paraview.org/u/woodscn)
#### Post date: [January 12, 2021, 9:40pm UTC](https://discourse.paraview.org/t/smoothing-the-output-of-rotational-extrusion/6211/5 "2021-01-12T21:40:30Z")
I’ve made some progress for what I need by only doing a 270 deg. rotation. It’s not terrible. Where would I apply the Clean to Grid filter?
---
### Author:  [@patchett2002](https://discourse.paraview.org/u/patchett2002)
#### Post date: [January 12, 2021, 9:42pm UTC](https://discourse.paraview.org/t/smoothing-the-output-of-rotational-extrusion/6211/6 "2021-01-12T21:42:52Z")
Clean to grid after the rotation, if the last bit of new geometry from the rotation matches the position of the initial geometry (nearly identically) it should work…
---
### Author:  [@mwestphal](https://discourse.paraview.org/u/mwestphal)
#### Post date: [January 13, 2021, 10:52am UTC](https://discourse.paraview.org/t/smoothing-the-output-of-rotational-extrusion/6211/7 "2021-01-13T10:52:25Z")
> [@woodscn](#):
>
> I’m trying to make a visually impressive
You may want to take a look into PBR or Raytracing rendering.
Supported by ParaView since 5.6.0.
---
### Author:  [@woodscn](https://discourse.paraview.org/u/woodscn)
#### Post date: [January 13, 2021, 5:35pm UTC](https://discourse.paraview.org/t/smoothing-the-output-of-rotational-extrusion/6211/8 "2021-01-13T17:35:32Z")
I’m using PBR for the metallic parts in the later image above. Will ray-tracing help me? Everything we’re looking at is extruded, hollow, data. I’d LOVE for a way to make the black portion look smoky, but so far the best I’ve come up with is to texture it.