attribute vec3 position;
|
|
uniform vec3 u_radii;
|
|
varying vec3 v_positionEC;
|
|
void main()
|
{
|
// In the vertex data, the cube goes from (-1.0, -1.0, -1.0) to (1.0, 1.0, 1.0) in model coordinates.
|
// Scale to consider the radii. We could also do this once on the CPU when using the BoxGeometry,
|
// but doing it here allows us to change the radii without rewriting the vertex data, and
|
// allows all ellipsoids to reuse the same vertex data.
|
vec4 p = vec4(u_radii * position, 1.0);
|
|
v_positionEC = (czm_modelView * p).xyz; // position in eye coordinates
|
gl_Position = czm_modelViewProjection * p; // position in clip coordinates
|
|
// With multi-frustum, when the ellipsoid primitive is positioned on the intersection of two frustums
|
// and close to terrain, the terrain (writes depth) in the closest frustum can overwrite part of the
|
// ellipsoid (does not write depth) that was rendered in the farther frustum.
|
//
|
// Here, we clamp the depth in the vertex shader to avoid being overwritten; however, this creates
|
// artifacts since some fragments can be alpha blended twice. This is solved by only rendering
|
// the ellipsoid in the closest frustum to the viewer.
|
gl_Position.z = clamp(gl_Position.z, czm_depthRange.near, czm_depthRange.far);
|
|
czm_vertexLogDepth();
|
}
|