To do this you basically need two steps:

1. Get the text in 2D (contour points and/or triangles)
2. Extrude the text to 3D

I recommend using Geomerative for step 1 (specifically RFont to Polygon to RMesh).

I recommend using Hemesh for step 2 (specifically the HEM_Extrude modifier).

Of course it can also be done manually without such libraries, but using them makes it so much simpler.

Here is an example I made with the described method last february. Don't mind the lines on the left.

Thanks - that looks like a very good starting point. I'll give it a try.

While I can follow the examples in both the Geometaritive and Hemish libraries, I can't see how to convert Geometerive's rmesh into Hemish's hmesh.

Do I have to break the rmesh down to an array of points and then reconstruct an hmesh?

Basically yes, you need some kind of solution like that. Which is why I prefer the RMesh because it is made of and able to return an array of RStrips which together form triangle strips that make up the 2D letter shape. If you are able to retreive the correct triangles you can make a Hemesh of it with the HEC_FromTriangles creator. Then you can move on to step 2.

Thanks, I will give that a try.

I have tried to start from a basic RPolygon, a rectangle.

Indeed the key is in the order. Both triangles should be created in clockwise order!

Adapted Code

1. import wblut.geom.core.*;
2. import wblut.hemesh.core.*;
3. import wblut.hemesh.creators.*;
4. import wblut.core.processing.*;
5. import geomerative.*;
6.  
7. WB_Render render;
8. HE_Mesh hMesh;
9. RMesh rMesh;
10.  
11. void setup() {
12.   size(400, 400, P3D);
13.   render = new WB_Render(this);
14.   RG.init(this);
15.   rMesh = createRMeshRectangle();
16.   hMesh = createHmeshFromRStrips(rMesh.strips);
17. }
18.  
19. void draw() {
20.   background(100);
21.   translate(width/2-50, height/2-70);
22.  
23.   fill(255, 0, 0);
24.   rMesh.draw();
25.  
26.   translate(0, 80);
27.  
28.   fill(0, 0, 255);
29.   render.drawFaces(hMesh);
30. }
31.  
32. RMesh createRMeshRectangle() {
33.   return RPolygon.createRectangle(0, 0, 100, 60).toMesh();
34. }
35.  
36. HE_Mesh createHmeshFromRStrips(RStrip[] _rStrips) {
37.  
38.   RPoint rA, rB, rC;
39.   WB_Point3d wbA, wbB, wbC;
40.   WB_ExplicitTriangle wExT;
41.  
42.   ArrayList triangleList = new ArrayList();
43.  
44.   for (int i=0; i<_rStrips.length; ++i) {
45.     RPoint[] rPoints = _rStrips[i].getPoints();
46.     for (int j=2; j<rPoints.length; ++j) {
47.      
48.       rA = rPoints[j-2];
49.       rB = rPoints[j-1];
50.       rC = rPoints[j];
51.      
52.       wbA = new WB_Point3d(rA.x, rA.y, 0);
53.       wbB = new WB_Point3d(rB.x, rB.y, 0);
54.       wbC = new WB_Point3d(rC.x, rC.y, 0);
55.      
56.       if (j % 2 == 0) { wExT = new WB_ExplicitTriangle(wbA, wbB, wbC); }
57.       else { wExT = new WB_ExplicitTriangle(wbC, wbB, wbA); }
58.      
59.       triangleList.add(wExT);
60.     }
61.   }
62.  
63.   HEC_FromTriangles creator = new HEC_FromTriangles();
64.   creator.setTriangles(triangleList);
65.   return new HE_Mesh(creator);
66. }

Ha - I knew it should be something like that!

this code runs fine for the triangle example but if i just punt a letter of a font it blocks just in

1. creator.setTriangles(triangleList);

i tried to keep my computer (ok it is not a superfast machine) for 10 minutes and i closed it because i consider the progran not work properly.

there are only 65 quads in the strips.. i don't think it is too much..

Yeah, there were some changes in Hemesh and they are causing this problem and also broke my original code. I don't know if this is due to stricter requirements for correct order, another method that needs to be called or a bug. And I also didn't have the time to look into it (when I wrote my code it worked for that version, don't know what has changed).

I was hoping if would either be fixed in an upcoming version of Hemesh or that c3notaph would give this a shot, but perhaps he has given up in disillusionment. The last thing I remember from this is that if you have different normals, this hang/crash problem appears. It's easy to write write a reproducible bug example. Change line 57 to abc and it hangs. You could submit to Frederik so he could take a look.

It would be nice if this was fixed or a solution would become known, because then I could revive my old code and perhaps take it to another level.

or that c3notaph would give this a shot, but perhaps he has given up in disillusionment

Sorry - it was a potential freelance job, and I invested too much time and effort, without success. I passed it onto someone else.

I've tried to pinpoint the problem. I feel there are at least two things that create a hang:

1. Vertices that are not in clockwise order (or counterwise normals between faces)
2. Overlapping faces

ah yeah i've just opened an issue on googlecode, good to know that you are testing around it! thanks!

Ok, i'm on it.

Thanks.

Found it. Something is hayware in the HE_Mesh.resolvePinchPoints(). In the latest release HEC_FromFaceList calls this function. It is meant to resolve issues with non-manifold meshes. But it f*x up with surfaces. Sorry guys, didn't test it enough.

thanks!

Hi,