Andro

Not sure how to pull this one off so I'm asking for the jitter wizards to help me out.

I'm experimenting with point cloud data with jit.gl.mesh.

I want to be able to change the scale of each point (circle) so obviously i have to step over to jit.gl.multiple @shape circle.

Heres the catch. I want to be able to turn jit.gl.multiple with jit.anim.drive and have the circles ALWAYS facing towards the camera.

Lookat and locklook with jit.anim.node don't work with jit.anim.drive as the shapes just stop moving.

Has anybody got any clue how I should approach this ??

I'm experimenting with point cloud data with jit.gl.mesh.
So far so good.
I want to be able to change the scale of each point (circle) so obviously i have to step over to jit.gl.multiple @shape circle.
Heres the catch. I want to be able to turn jit.gl.multiple with jit.anim.drive and have the circles ALWAYS facing towards the camera.
Lookat and locklook with jit.anim.node don't work with jit.anim.drive as the shapes just stop moving.
Has anybody got any clue how I should approach this ??


jit-gl-multiple-how-to-make-make-shapes-look-at-the-camera

hi andro. it might be time for you to utilize geometry shaders. there is a basic billboard geometry shader that comes with max that will do this from a simple jit.gl.mesh point cloud. you will have to make some modifications to incorporate the individual scaling (e.g., you could store the scale value in the alpha channel of the color value, and pass that on to the geometry shader, or you could create a vertex-attribute to store those values).

here's a basic example utilizing the billboard shader:

hi andro. it might be time for you to utilize geometry shaders. there is a basic billboard geometry shader that comes with max that will do this from a simple jit.gl.mesh point cloud.  you will have to make some modifications to incorporate the individual scaling (e.g., you could store the scale value in the alpha channel of the color value, and pass that on to the geometry shader, or you could create a vertex-attribute to store those values).

here's a basic example utilizing the billboard shader:


sending vertex attributes through the pipeline is handled via the "vertex_attr_matrix" message, followed by the name of a matrix containing the vertex attributes. search the forum for more info on this.

i've attached a basic glsl particle system patch that demonstrates:

The 1st patch demonstrates perfectly what I'd like to do.

Though to be honest I think this is pretty much beyond my abilities with shaders (i've only been getting a grasp on jit.gl.pix and and jit.gen the last few months)

I'll do my best to try and pull this apart to try and understand whats going on,.

Could you maybe just post a few lines of comments on main sections so I have more of an idea of what each section is doing (1st patch) ?

The 1st patch demonstrates perfectly what I'd like to do.
Though to be honest I think this is pretty much beyond my abilities with shaders (i've only been getting a grasp on jit.gl.pix and and jit.gen the last few months)
I'll do my best to try and pull this apart to try and understand whats going on,.
Could you maybe just post a few lines of comments on main sections so I have more of an idea of what each section is doing (1st patch) ?

I think the part I can't figure out is what piece of the shader makes the geometry not rotate when jit.anim.drive is turning the jit.gl.mesh.

prior to the shader we are only generating points. those points are rotated around the origin of the gl.mesh by the anim.drive. when we pass those point locations to the shader, the geometry program receives a point location and outputs 4 vertices for each. the 4 vertices are centered around the point, and create a quad shape based on the scale parameter. the billboarding magic happens by taking in the current gl_ModelViewMatrix, extracting the x and y vectors, and creating a quad using those values and the current point position. i barely understand much of this myself, so can't really offer more in the way of explanation.

prior to the shader we are only generating points. those points are rotated around the origin of the gl.mesh by the anim.drive.  when we pass those point locations to the shader, the geometry program receives a point location and outputs 4 vertices for each. the 4 vertices are centered around the point, and create a quad shape based on the scale parameter. the billboarding magic happens by taking in the current gl_ModelViewMatrix, extracting the x and y vectors, and creating a quad using those values and the current point position. i barely understand much of this myself, so can't really offer more in the way of explanation.


Jit.gl.multiple How to make make shapes look at the camera