Issue with using Sampler2D with a Shader

Question

! I am aware that there are easier ways to draw to the screen than this, but i need to do it this specific way !

I am drawing to a texture through a fbo. I am then using a shader to re-draw it onto the screen, doing so with a sampler2D (of the texture) and having the shader set gl_fragColor to the color of the sampler2D at the specific point the fragment shader is at.

The issue i am having is that the display buffer (which the shader is drawing the fbo to) is only a solid blue, even though i draw a white squares to the fbo with a blue background.

My Code:

Main Render Loop:

while(!Display.isCloseRequested()){
        //Drawing on the fbo a blue background and white square
        glBindFramebufferEXT(GL_FRAMEBUFFER_EXT, fboID);
        glColor3f(0,0,1);
        drawQuad(0,0,WIDTH,HEIGHT);
        glColor3f(1,1,1);
        drawQuad(0,0,50,50);

        //Trying to draw the texture from the previous fbo
        //on to the display buffer with a fragment shader,
        //however it only draws a blue background, no white square.

        glBindFramebufferEXT(GL_FRAMEBUFFER_EXT, 0);

        GL13.glActiveTexture(GL13.GL_TEXTURE0);
        glBindTexture(GL_TEXTURE_2D,texID);
        glUseProgram(shaderProgram);
        glUniform1i(glGetUniformLocation(shaderProgram, "tex"), 0);
        glUniform1f(glGetUniformLocation(shaderProgram, "width"), WIDTH);
        glUniform1f(glGetUniformLocation(shaderProgram, "height"), HEIGHT);

        glBegin(GL_QUADS); {
            glVertex2f(0, 0);
            glVertex2f(0, HEIGHT);
            glVertex2f(WIDTH, HEIGHT);
            glVertex2f(WIDTH, 0);
        } glEnd();
        glUseProgram(0);

        Display.update();
    }

My Shader:

uniform sampler2D tex;
uniform float width;
uniform float height;

void main() {
    vec4 color = texture2D( tex, gl_FragCoord.xy / vec2(width, height));
    gl_FragColor = color;
}

Init method for the fbo:

    texID=glGenTextures();
    fboID=glGenFramebuffersEXT();

    glBindFramebufferEXT(GL_FRAMEBUFFER_EXT, fboID);
    glBindTexture(GL_TEXTURE_2D, texID);
    glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
    glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA8, WIDTH, HEIGHT, 0,GL_RGBA, GL_INT, (java.nio.ByteBuffer) null);
    glFramebufferTexture2DEXT(GL_FRAMEBUFFER_EXT,GL_COLOR_ATTACHMENT0_EXT,GL_TEXTURE_2D, texID, 0);
    glBindTexture(GL_TEXTURE_2D, 0);
    glBindFramebufferEXT(GL_FRAMEBUFFER_EXT, 0);

I think one reason why this is happening could be becuase i am not correctly passing the Sampler2D to the shader but i not sure how to fix it.

If any could tell why this is happening or how to fix it (or even just a point in the right direction) it would be much appreciated!

(Sorry for any bad English!)

Answer 1

You have transposed your calls to glUniform1i (...) and glUseProgram (...).

At the time that glUniform1i(glGetUniformLocation(shaderProgram, "tex"), 0); is executed, the currently active program is 0 and that should in fact be generating the following error:

GL_INVALID_OPERATION is generated if there is no current program object.

If you swap the two lines mentioned above, that should fix your problem. In the future you should make a point of checking glGetError (...) when something does not work.

---

UPDATE: Now that the original problem is fixed, you are not setting the texture minification filter correctly.

To set enumerants, you must use glTexParameteri (...). glTexParameterf (...) is going to interpret the value passed as a floating-point number and GL_LINEAR (0x2601) has type GLenum (32-bit unsigned integer). Fortunately all core OpenGL enums only use the lower 16-bits so they can be expressed precisely as 32-bit floats (can express all integers up to 2²⁴), but you can see why you would not want to convert an integer constant to a float.

Therefore, the following line:

glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);

Needs to be:

glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);

---

UPDATE 2: You must use normalized texture coordinates when using texture2D (...).

gl_FragCoord.xy has the range [0, width] and [0, height], which is outside the normalized range for any framebuffer larger than 1x1. To fix that, you must divide by width and height.

texelFetch (...) allows you to use unnormalized coordinates, but it requires GLSL 1.30 and you would have to cast gl_FragCoord.xy to ivec2 to use it. It may also prove inconvenient if you ever draw into an FBO that has different dimensions than your window.