Costa's minimal surface with PovRay

Costa's minimal surface is one of those amazing mathematical objects that look good when rendered with PovRay. While it is not too hard to find triangulations of Costa's surface suitable for the PovRay triangle element, triangulations with surface normals are not that easy to come by. But these are needed for the smooth_triangle element which makes Phong shading possible.

In the service of humanity, and especially my students who wanted to render the surface, I provide here the code to triangulate and compute the normals to Costa's minimal surface. The code is written in Mathematica, which is not freely accessible. Therefore, for convenience the repository also contains certain files generated by Mathematica:

• a CSV file costa.csv with ready-made triangulations and surface normals, with a Python script cvs-to-mesh2.py for converting the CSV file to a partial PovRay
  mesh2 element,

• a PDF file Costa.pdf, the PDF export of Costa.nb, so you can see how I did it.

How to use the files

To generate a picture using PovRay, proceed as follows:

1. Run Costa.nb with Mathematica and use the CostaCSV function to generate a CSV file which contains the triangulation and normals, say costa.csv.

2. Use csv-to-mesh2.py to convert the CSV file to a partial PovRay mesh2 object.

3. Include the partial mesh2 object into your PovRay file, as demonstrated in costa.pov.

4. Generate costa.png with something like

    povray costa.pov +A +J +W1024 +H1024
   

The format produced by Mathematica CostaCSV function

The Mathematica function CostaCSV outputs a CSV file in a format that is suitable for generating a PovRay mesh2 object:

1. The first line lists the parameters that were used with in the call to CostaCSV:

    r1,r2,k,l,m,n
   

Their meaning is as follows:

* `r1` and `r2` are the radii which determine the upper and middle radius of the surface
* `k`, `l`, `m`, `n` specify the number of vertices used in the triangulation, specifically:
  `k` for the inner points, `l` for the edges, `m` for one rim and `n` for the other rim.
  In general, larger numbers mean a finer mesh.

2. The triangulation points, where N is the number of points, followed by one point per line:

    N
    x1,y1,z1
    x2,y2,z2
    ...
    xN,yN,zN
   

3. The normalized normals to the surface at respective triangulation points, where N is the number of points (and equal to the N above):

    N
    u1,v1,w1
    u2,v2,w2
    ...
    uN,vN,wN  
   

4. The triangles that form the surface triangulation, where M is their numbers and the indices refer to the points above.

    M
    i1,j1,k1
    i2,j2,k2
    ...
    iM,jM,kM
   

The result

The file costa.pov generates the following picture: