Search code examples
opengltexturestexture2d

Mac - Linear interpolation not working

I searched a lot and just can't seem to find anything on this topic. I got a Macbook with OpenGL 4.1. I create a Texture with:

glGenTextures(1, &heightMapTexture);
glActiveTexture(GL_TEXTURE0);
glBindTexture(GL_TEXTURE_2D, heightMapTexture);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_BASE_LEVEL, 0);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAX_LEVEL, 13);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST_MIPMAP_NEAREST);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_MIRRORED_REPEAT);
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_MIRRORED_REPEAT);
glUniform1i(program1->uniformLocation("heightMap"), 0);

glTexImage2D(GL_TEXTURE_2D, 0, GL_R16UI, heightMap.getWidth(), heightMap.getHeight(), 0, GL_RED_INTEGER, GL_UNSIGNED_SHORT, heightMap.getData().data());
glGenerateMipmap(GL_TEXTURE_2D);

This works fine but I actually would like to use GL_LINEAR. But GL_LINEAR gives me a texture with all zeros / black.

Does Mac just not support GL_LINEAR or is there an extension for it? In the official documentation i found the following piece of code:

glTexParameteri(GL_TEXTURE_RECTANGLE_ARB,
                GL_TEXTURE_MIN_FILTER, GL_LINEAR);

(https://developer.apple.com/library/mac/documentation/GraphicsImaging/Conceptual/OpenGL-MacProgGuide/opengl_texturedata/opengl_texturedata.html)

But I can't get the extension to work either.

Solution

  • GL_R16UI is an integer texture format. OpenGL does not support linear filtering of integer formats.

    From section 8.17 "Texture Completeness" of the OpenGL 4.5 spec, page 252:

    Using the preceding definitions, a texture is complete unless any of the following conditions hold true:
    [..] The internal format of the texture is integer [..] and either the magnification filter is not NEAREST, or the minification filter is neither NEAREST nor NEAREST_MIPMAP_NEAREST.

    To use texures with 16-bit per component precision that you can sample with linear filtering, you'll need to use a normalized format like GL_R16, or go all the way to a float with GL_R32F.