When creating a heightfield from an image, is there any rule of thumb for what resolution to choose for the mesh? I know 300 DPI is the industry standard for printing photographs on paper, but am clueless when it comes to mesh density. If I intend to make a detailed print of the height field generated texture, at what point am I creating an excessivly detailed part? It’s very easy to create a 400 MG file, when converted to an STL for printing, which most printing services can’t even quote.
Hello- the hight uses the grayscale value of the pixel in the image that is at the vertex location. Ideally, you want one mesh vertex per pixel. Pixels that don’t hit a vertex will be ignored.
https://docs.mcneel.com/rhino/7/help/en-us/index.htm#commands/heightfield.htm?
-Pascal
Thanks for the reply Pascal.
That’s generally how I’ve been doing it, choosing the same number of vertexes as pixels. In a 400 x 200 mm part, with a 3500 x 1500 image, that creates a massive file size around 160 MG as a 3Dm, and 380 MG as an STL. Most printers just can’t deal with a file size that large. Is there a solution that will give me a smaller, printable file size but that still retains the resolution of the original image?
Hi Dregan - I suppose the only way out is to resample the image downward or use say half the number of vertices in each direction. I guess I’d be more inclined to down sample the image but you may want to smooth it as well.
-Pascal
Hi
after you make the hight field on the very dens mesh, you can use _ReduceMesh to bring down the size to a printable level…
Disclaimer, I haven’t use it in Rhino so no idea how well it works, But I do this regularly in other apps [Zbrush in my case] you should be able to lower the poly count down to 20 - 10% without losing any visible details.
I think you should be able to reduce your mash to around 1 million polys which all modern printers will be happy to print.
Akash
Thanks for the replies. I’ve been testing reducemesh out this morning. It seems to work pretty well but is kind of difficult to see the actual part in redered mode as the original image that is used to generate the heightfield is overlayed. Does anyone know how to remove the original image so view in redered mode just shows the texture of the part?
does it have to be mesh? if you use a nurbs surface instead you can also use FitSrf after the result to generate a surface that is detailed enough but with a reduced amount of spans. you can also experiment with quadremesh to bring the polygons count down
It seems like the heightfield from image command automatically creates a mesh, no nurbs option.
Heightfield can generate a Nurbs surface or a mesh, see options here Heightfield | Rhino 3-D modeling
Displacement generates a mesh, Displacement properties | Rhino 3-D modeling
You need to export it or Extract Render Mesh to see the results, although it can create a preview. Displacement gives a few additional option for control over the mesh, and may be preferable if you don’t need a Nurbs surface. (Nurbs will smooth out sharp edges.)