Julien Bayle

basically a lack of opengl knowledge I'm sure:

how to exclude (=hide) a part of a gridshape ?

For instance, I have an opencube, scaled like 1. 1. 100.

I have no other solutions than creating 2 separated by a distance, creating a node linking both of them and manage the node.

hi there,
basically a lack of opengl knowledge I'm sure:
how to exclude (=hide) a part of a gridshape ?

For instance, I have an opencube, scaled like 1. 1. 100.
I want to "remove" the center of it.
I have no other solutions than creating 2 separated by a distance, creating a node linking both of them and manage the node.

how-to-exclude-hide-a-part-of-a-gridshape

Ya, similar to the boole subtract in cinema4d, where you literally subtract an object from another one based on how they're intersecting.

This is fairly simple to do if you don't care about how the object's faces look:

With this one, play with the cutoff value for the [gen]. you can also get some visual trickery where you make the dimensions so high, that you can't notice the separation between faces (try dim 8 1000).

but then if you change the gl.mesh's draw_mode to tri_grid, you'll see how Max tries to connect all the faces together. Realistically you want "triangles", but as you can see when you select that, nothing appears. That's because of the index array, which is a whole nother beast to conquer.

here's a working example using an index array. you'd have to create this array beforehand. i've developed a way to programmatically create the index array with [jit.gl.model] but haven't currently solved it for gl.gridshape (shouldn't be too difficult, but i'll wait on that in case someone has a better solution to this). Also, i know index array probably isn't a familiar term, simply it's a way to tell gl.mesh which vertices to connect to form a face. Again play with the cutoff param for [gen] in this one.

then lighting this properly is yet another thing to conquer.

if anyone knows any better methods, or more efficient ways to do mine, i'd love to see them.

This is fairly simple to do if you don't care about how the object's faces look:
With this one, play with the cutoff value for the [gen]. you can also get some visual trickery where  you make the dimensions so high, that you can't notice the separation between faces (try dim 8 1000).

ok i looked into the jit.gl.gridshape matrixoutput, and i have to say, i have no idea how it's telling the gl.mesh what the index array is. It spits out all the coordinates in order of front to back position, rather than the gl.model method which spits out vertices that you can group every 3 of to form a face. Don't want to hijack this thread too much, especially if a better solution doesn't involve the index array, but I'd love to hear from Rob Ramirez about gridshape's matrixoutput and why its organization/arrangement of vertices is different than gl.model... (how does gl.mesh know which arrangement-type of vertices it is receiving?)

thanks Greg. hi-jacking is very okay and you (also) answered. yes I totally got it and I didn't have tought about cutting the matrix like that

you can do this another way simply using textures and alpha blending.

this patch creates a "cutout" shape (by capturing a gl.gridshape to texture) and adds it as the alpha channel of the main shape's texture (using a simple gl.pix). we use two gridshapes for the main shape, one for front-facing polys and one for back-facing.

you can do this another way simply using textures and alpha blending.
this patch creates a "cutout" shape (by capturing a gl.gridshape to texture) and adds it as the alpha channel of the main shape's texture (using a simple gl.pix). we use two gridshapes for the main shape, one for front-facing polys and one for back-facing.