I am trying to wrap my head around the various types of GLSL shaders in OpenGL. At the moment I am struggling with a 2d layered-tile implementation. For some reason the int values that get passed into my shader are always 0 (or more likely, null).
I currently have a 2048x2048px 2d texture composed of 20x20 tiles. I am trying to texture one quad with it and change the index of the tile based upon the block of ints I pass into the vertex shader.
I am passing in a vec2 of floats for the position of the quad (really a TRIANGLE_STRIP). I am also attempting to pass in 6 ints that will represent the 6 layers of tiles.
My input:
// Build and compile our shader program
Shader ourShader("b_vertex.vertexShader", "b_fragment.fragmentShader");
const int floatsPerPosition = 2;
const int intsPerTriangle = 6;
const int numVertices = 4;
const int sizeOfPositions = sizeof(float) * numVertices * floatsPerPosition;
const int sizeOfColors = sizeof(int) * numVertices * intsPerTriangle;
const int numIndices = 4;
const int sizeOfIndices = sizeof(int) * numIndices;
float positions[numVertices][floatsPerPosition] =
{
{ -1, 1 },
{ -1, -1 },
{ 1, 1 },
{ 1, -1 },
};
// ints indicating Tile Index
int colors[numVertices][intsPerTriangle] =
{
{ 1, 2, 3, 4, 5, 6 },
{ 1, 2, 3, 4, 5, 6 },
{ 1, 2, 3, 4, 5, 6 },
{ 1, 2, 3, 4, 5, 6 },
};
// Indexes on CPU
int indices[numVertices] =
{
0, 1, 2, 3,
};
My setup:
GLuint vao, vbo1, vbo2, ebo; // Identifiers of OpenGL objects
glGenVertexArrays(1, &vao); // Create new VAO
// Binded VAO will store connections between VBOs and attributes
glBindVertexArray(vao);
glGenBuffers(1, &vbo1); // Create new VBO
glBindBuffer(GL_ARRAY_BUFFER, vbo1); // Bind vbo1 as current vertex buffer
// initialize vertex buffer, allocate memory, fill it with data
glBufferData(GL_ARRAY_BUFFER, sizeOfPositions, positions, GL_STATIC_DRAW);
// indicate that current VBO should be used with vertex attribute with index 0
glEnableVertexAttribArray(0);
// indicate how vertex attribute 0 should interpret data in connected VBO
glVertexAttribPointer(0, floatsPerPosition, GL_FLOAT, GL_FALSE, 0, 0);
glGenBuffers(1, &vbo2); // Create new VBO
glBindBuffer(GL_ARRAY_BUFFER, vbo2); // Bind vbo2 as current vertex buffer
// initialize vertex buffer, allocate memory, fill it with data
glBufferData(GL_ARRAY_BUFFER, sizeOfColors, colors, GL_STATIC_DRAW);
// indicate that current VBO should be used with vertex attribute with index 1
glEnableVertexAttribArray(1);
// indicate how vertex attribute 1 should interpret data in connected VBO
glVertexAttribPointer(1, intsPerTriangle, GL_INT, GL_FALSE, 0, 0);
// Create new buffer that will be used to store indices
glGenBuffers(1, &ebo);
// Bind index buffer to corresponding target
glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, ebo);
// ititialize index buffer, allocate memory, fill it with data
glBufferData(GL_ELEMENT_ARRAY_BUFFER, sizeOfIndices, indices, GL_STATIC_DRAW);
// reset bindings for VAO, VBO and EBO
glBindVertexArray(0);
glBindBuffer(GL_ARRAY_BUFFER, 0);
glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, 0);
// Load and create a texture
GLuint texture1 = loadBMP_custom("uvtemplate3.bmp");
GLuint texture2 = loadBMP_custom("texture1.bmp");
My draw:
// Game loop
while (!glfwWindowShouldClose(window))
{
// Check if any events have been activiated (key pressed, mouse moved etc.) and call corresponding response functions
glfwPollEvents();
// Render
// Clear the colorbuffer
glClearColor(1.f, 0.0f, 1.0f, 1.0f);
glClear(GL_COLOR_BUFFER_BIT);
// Activate shader
ourShader.Use();
// Bind Textures using texture units
glActiveTexture(GL_TEXTURE0);
glBindTexture(GL_TEXTURE_2D, texture1);
//add some cool params
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_BORDER);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_BORDER);
float borderColor[] = { 0.45f, 0.25f, 0.25f, 0.25f };
glTexParameterfv(GL_TEXTURE_2D, GL_TEXTURE_BORDER_COLOR, borderColor);
glUniform1i(glGetUniformLocation(ourShader.Program, "ourTexture1"), 0);
glActiveTexture(GL_TEXTURE1);
glBindTexture(GL_TEXTURE_2D, texture2);
glUniform1i(glGetUniformLocation(ourShader.Program, "ourTexture2"), 1);
// Draw container
//glBindVertexArray(VAO);
//glDrawElements(GL_TRIANGLES, 6, GL_UNSIGNED_INT, 0);
glBindVertexArray(vao);
//glDrawElements(GL_TRIANGLES, 6, GL_UNSIGNED_INT, 0);
glDrawElements(GL_TRIANGLE_STRIP, numIndices, GL_UNSIGNED_INT, NULL);
glBindVertexArray(0);
// Swap the screen buffers
glfwSwapBuffers(window);
}
My shader most definitely works, as I can adjust the output by hard-coding the values from within the vertexShader. My suspicion is I am not passing the values correctly/ in the correct format or not indicating somewhere that the int[6] needs to be included per vertex.
I cannot read anything from my layout (location = 1) in int Base[6]; I've tried just about everything I can think of. Declaring each int individually, trying to read two ivec3's, uint and what ever else I could think of but everything comes back with 0.
The following are my vertex and fragment shader for completeness:
#version 330 core
layout (location = 0) in vec2 position;
layout (location = 1) in int Base[6];
out vec2 TexCoord;
out vec2 TexCoord2;
out vec2 TexCoord3;
out vec2 TexCoord4;
out vec2 TexCoord5;
out vec2 TexCoord6;
// 0.5f, 0.5f,// 0.0f, 118.0f, 0.0f, 0.0f, 0.0f, 0.0f, // Top Right
// 0.5f, -0.5f,// 0.0f, 118.0f, 1.0f, 0.0f, 0.0f,0.009765625f, // Bottom Right
// -0.5f, -0.5f,// 0.0f, 118.0f, 0.0f, 1.0f, 0.009765625f, 0.009765625f, // Bottom Left
// -0.5f, 0.5f//, 0.0f, 118.0f, 1.0f, 0.0f, 0.009765625f, 0.0f // Top Left
void main()
{
int curBase = Base[5];
int curVertex = gl_VertexID % 4;
vec2 texCoord = (curVertex == 0?
vec2(0.0,0.0):(
curVertex == 1?
vec2(0.0,0.009765625):(
curVertex == 2?
vec2(0.009765625,0.0):(
curVertex == 3?
vec2(0.009765625,0.009765625):(
vec2(0.0,0.0)))))
);
gl_Position = vec4(position, 0.0f, 1.0f);
TexCoord = vec2(texCoord.x + ((int(curBase)%102)*0.009765625f)
, (1.0 - texCoord.y) - ((int(curBase)/102)*0.009765625f));
//curBase = Base+1;
TexCoord2 = vec2(texCoord.x + ((int(curBase)%102)*0.009765625f)
, (1.0 - texCoord.y) - ((int(curBase)/102)*0.009765625f));
//curBase = Base+2;
TexCoord3 = vec2(texCoord.x + ((int(curBase)%102)*0.009765625f)
, (1.0 - texCoord.y) - ((int(curBase)/102)*0.009765625f));
}
Fragment:
#version 330 core
//in vec3 ourColor;
in vec2 TexCoord;
in vec2 TexCoord2;
in vec2 TexCoord3;
in vec2 TexCoord4;
in vec2 TexCoord5;
in vec2 TexCoord6;
out vec4 color;
// Texture samplers
uniform sampler2D ourTexture1;
uniform sampler2D ourTexture2;
void main()
{
color = (texture(ourTexture2, TexCoord )== vec4(1.0,0.0,1.0,1.0)?
(texture(ourTexture2, TexCoord2 )== vec4(1.0,0.0,1.0,1.0)?
(texture(ourTexture2, TexCoord3 )== vec4(1.0,0.0,1.0,1.0)?
(texture(ourTexture2, TexCoord4 )== vec4(1.0,0.0,1.0,1.0)?
(texture(ourTexture2, TexCoord5 )== vec4(1.0,0.0,1.0,1.0)?
(texture(ourTexture2, TexCoord6 )== vec4(1.0,0.0,1.0,1.0)?
vec4(0.0f,0.0f,0.0f,0.0f)
:texture(ourTexture2, TexCoord6 ))
:texture(ourTexture2, TexCoord5 ))
:texture(ourTexture2, TexCoord4 ))
:texture(ourTexture2, TexCoord3 ))
:texture(ourTexture2, TexCoord2 ))
:texture(ourTexture2, TexCoord ));
}
This is wrong in two different ways:
glVertexAttribPointer(1, intsPerTriangle, GL_INT, GL_FALSE, 0, 0);
Vertex attributes in the GL can be scalars or vectors of 2 to 4 components. Hence, the size parameter of glVertexAttribPointer can take the values of 1, 2, 3 or 4. Using a different value (intsPerTriangle == 6) means that the call will just generate an GL_INVALID_VALUE error and has no ther effect, so you don't even set a pointer.
If you you want to pass 6 values per vertex, you can either use 6 different scalr attributes (consuming 6 attribute slots), or pack this into some vectors, like 2 3d vectors (consuming only 2 slots). No matter what packing you chose, you'll need a proper attrib pointer setup for each attribute slot in use.
However, glVertexAttribPointer is also the wrong function for your use case. It is defining floating-point attributes, which musthave matching declarations as float/vec* in the shader. The fact that you can input GL_INT just means that the GPU can do the conversion to floating-point on the fly for you.
If you want to use an int or ivec (or their unsigned counterparts) attribute, you have to use glVertexAttribIPointer (note the I in that function name) when setting up the attribute.