Nice. Though I think that you need to add an operator function argument, because there's more than one way to combine those waves.  You show what appears to be a multiplicative surface, but an additive surface is another valid interpretation, yet has a rather different result:



include <BOSL2/std.scad>

function surface_from_2_waves(xwave, ywave, oper = function(x,y) x*y) = [
for(ypt = ywave) [
for(xpt = xwave)
[xpt.x, ypt.y, oper(xpt.z,ypt.z)]
]
];

function surf_extrude(surfpts, height) =
vnf_vertex_array(
[for (row=surfpts) concat(row, down(height, p=reverse(row)))],
caps=true, col_wrap=true, row_wrap=false
);

xywave = [for (i = [0:2:100]) [i,2sin(3i*360/100)]];
xzwave = xrot(90, p=path3d(xywave));
yzwave = zrot(90, p=xzwave);
surfpts = surface_from_2_waves(xzwave, yzwave, function(x,y) x+y);
vnf = surf_extrude(surfpts, 0.1);
vnf_polyhedron(vnf, convexity=10);

-Revar

Thanks
Will try this

On Sun, 29 Oct, 2023, 2:11 am Revar Desmera, revarbat@gmail.com wrote:

You can get interesting effects using min() or max() on combinations of sine and triangle wave inputs.



• Revar

very nice.
other interesting shapes
[image: Screenshot 2023-10-29 at 3.21.18 PM.png]

On Sun, 29 Oct 2023 at 11:08, Revar Desmera revarbat@gmail.com wrote:

Found a fun one!  You can do a quilted looking surface by getting the norm() of two sine waves:

• Revar

 It would be great to have a library of these shapes.  I'm sure they
will come in handy from time to time.

On 10/30/2023 4:59 PM, Revar Desmera wrote:

Technically that surface isn't quilted, exactly.  When you make a blanket
by securing the layers at single points you do that by tying.  It's a less
secure technique than real quilting, where the layers are sewn together
along lines.

It seems like the BOSL2 texture library does collect patterns like this.

https://github.com/BelfrySCAD/BOSL2/wiki/skin.scad#subsection-vnf-textures

On Mon, Oct 30, 2023 at 8:31 PM Sanjeev Prabhakar sprabhakar2006@gmail.com
wrote:

Wonderful textures, but how did you create these?
Is it with the 2 waves method?

On Tue, 31 Oct, 2023, 6:54 am Adrian Mariano, avm4@cornell.edu wrote:

Another question: I wonder if it is faster to do it with the 2 waves
method or SDF? E.g. with libfive integration in python.

On 31/10/2023 09:36, Sanjeev Prabhakar wrote:

Can you share some examples as to how this works?

On Tue, 31 Oct, 2023, 7:29 am Chun Kit LAM, john.lck40@gmail.com wrote:

Just something like abs(z - sin(x) - sin(y)) <= d/2, where d is the
desired thickness.

I don't have time to compile the python integration for now so I just
tried it with manifold sdf. The surface is not very smooth due to
marching cubes used. It is probably better to do it with the method in
BOSL2.

On 31/10/2023 10:24, Sanjeev Prabhakar wrote:

All the BOSL2 textures were created by hand, I'm afraid.  I had a need to do things like knurling on some parts, and that generalized into textures.  Since heightfields were insufficient to make some textures cleanly, I expanded the system out to allow more precise VNF based textures.  I just made a bunch of standard textures that seemed useful.  Textures can be applied to cyl()s, rotate_sweep()s, and linear_sweep()s.

You CAN use the two-wave technique to generate heightfields or VNFs that are useable as textures, but you have to beware how many vertices you introduce.
The fact that you can generate the diamonds I originally needed for knurling via the 2-wave method was actually an interesting surprise to me, though it's far more expensive in vertex counts.  The manually created heightfield texture for diamonds in BOSL2 is actually just [[0,1], [1,0]].

• Revar