I’m trying to load a mat4 from a texture in a vertex shader and am having problems when I load 4 components at a time from an RGBA texture.
To help figure out what is going on I’ve been trying to access one float at a time rather than load all 16. I’m using the NVIDIA linux drivers (v71.74) with a GeForce 6600 GT.
I can get it to work properly if I use a luminance texture as follows:
Specify the texture:
glTexImage2D(GL_TEXTURE_2D, 0, GL_LUMINANCE_FLOAT32_ATI, 4, 4, 0, GL_LUMINANCE, GL_FLOAT, h);
and access it in the vertex shader using:
vec2 v = vec2(gl_Vertex.x / 4.0, gl_Vertex.z / 4.0);
vec4 patch = texture2D(PatchesTexture, v);
gl_Vertex.y = patch[0];
h is defined as: float h[4][4] and the x and z vertex coordinates go from 0 to 3.95. PatchesTexture is a sampler2D.
When I change to using an RGBA texture with the following code for the texture:
glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA_FLOAT32_ATI, 1, 4, 0, GL_RGBA, GL_FLOAT, h);
and this for the vertex shader:
vec2 v = vec2(0.0, gl_Vertex.z / 4.0);
vec4 patch = texture2D(PatchesTexture, v);
gl_Vertex.y = patch[int(floor(gl_Vertex.x))];
I get invalid results. The height doesn’t change with x as it should. If I hard code the patch index to be from 0 to 3, it retrieves the values that would expected for that value of x.
Any suggestions to what I could try next?
dave j