Can't spot the issue with my GLSL/OpenGL code - c

I wrote a little program to display a 32bit float texture in a simple quad. When displaying the quad, the texture color is always black. I experimented with a lot of things, but I couldn't make it work. I'm really at loss what the problem with it.
The code of creating the OpenGL texture goes like this
glEnable(GL_TEXTURE_2D);
glGenTextures(1, &textureID);
glBindTexture(GL_TEXTURE_2D, textureID);
glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA32F, width, height, 0, GL_RGBA, GL_FLOAT, textureData);
Using the debugger, there's no error in any of these calls. I also examined the textureData pointer, and got the expected results (in my simplified program, it is just a gradient texture).
This is the vertex shader code in GLSL:
#version 400
in vec4 vertexPosition;
out vec2 uv;
void main() {
gl_Position = vertexPosition;
uv.x = (vertexPosition.x + 1.0) / 2;
uv.y = (vertexPosition.y + 1.0) / 2;
}
It's kind of a simple generation of the UV coordinates without taking them as vertex attributes. The corresponding vertex buffer object is really simple:
GLfloat vertices[4][4] = {
{ -1.0, 1.0, 0.0, 1.0 },
{ -1.0, -1.0, 0.0, 1.0 },
{ 1.0, 1.0, 0.0, 1.0 },
{ 1.0, -1.0, 0.0, 1.0 },
};
I've tested the solution, and it displays the quad covering the entire window as I wanted to. Displaying the UV coordinates in the fragment shader reproduce the gradient that I expected to get. Now here's the fragment shader:
#version 400
uniform sampler2D myTex;
in vec2 uv;
out vec4 fragColor;
void main() {
fragColor = texture(myTex, uv);
// fragColor += vec4(uv.x, uv.y, 0, 1);
}
The commented out line displays the UV coordinates as color for debugging purposes. What do I do wrong here? I just can't see why the texture() call returns 0 where the texture seems completely right, and the uv coordinates are also proper. I link here the full code if there's something else I do wrong: gl-view.c
EDIT: This is how I set up the myTex sampler:
glEnable(GL_TEXTURE_2D);
glActiveTexture(GL_TEXTURE0);
glBindTexture(GL_TEXTURE_2D, textureID);
glUniform1i(glGetUniformLocation(shaderProgram, "myTex"), 0);
glDrawArrays(GL_TRIANGLE_STRIP, 0, 4);
EDIT: Cleared up the vertex shader code.

I've found the issue: I didn't set any MAG or MIN filter on the texture. Setting the MIN filter to GL_NEAREST solved the problem.

Related

GLSL Shader going black when I try to sample a texture

I render a triangle strip this way, and with basic bypass shaders all is working fine:
EDIT:
I added TextCoords and modified the shaders , I keep getting the same result, my 3d objects are going black!
UPDATED CODE:
// Dibuixem tots els prismes
glBegin(GL_TRIANGLE_STRIP);
for(i=0;i<num_elems;i++) {
for(j=0;j<num_vertices;j++) {
glNormal3fv((GLfloat *)(a+j*2));
glTexCoord2f(0.0f, 0.0f);
glVertex3fv((GLfloat *)(a+j*2+1));
glTexCoord2f(1.0f, 0.0f);
glNormal3fv((GLfloat *)(b+j*2));
glTexCoord2f(1.0f, 1.0f);
glVertex3fv((GLfloat *)(b+j*2+1));
}
glNormal3fv((GLfloat *)(a));
glTexCoord2f(0.0f, 1.0f);
glVertex3fv((GLfloat *)(a+1));
glNormal3fv((GLfloat *)(b));
glTexCoord2f(0.0f, 0.0f);
glVertex3fv((GLfloat *)(b+1));
a+=face_size;
b+=face_size;
}
glEnd();
And I am trying to attach a texture to my shaders, but I can't figure out how to pass the texture.
I create and add the texture to my program this way. Texture data is verified
array with format unsigned char data[imageSize];:
GLuint textureID;
glGenTextures(1, &textureID);
glBindTexture(GL_TEXTURE_2D, textureID);
glActiveTexture(GL_TEXTURE0); // Texture unit 0
glPixelStorei(GL_UNPACK_ALIGNMENT, 1);
glTexImage2D(GL_TEXTURE_2D, 0,GL_BGR, width, height, 0, GL_BGR, GL_UNSIGNED_BYTE, data);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
GLuint VertexShaderID = glCreateShader(GL_VERTEX_SHADER);
This is what I read in another posts with the same issue and I added to my code after compiling my shaders and generating my program without errors.
Tutorials tend to dismiss this information (how you say to your shader the name and location of your binded texture).
GLuint t1Location = glGetUniformLocation(programID, "tex1");
glUniform1i(t1Location, 0);
And my shaders UPDATED CODE:
#define GLSL(version, shader) "#version " #version "\n" #shader
const char* vert = GLSL
(
110,
varying vec4 position;
varying vec3 normal;
varying out vec4 texCoord;
varying vec2 coord;
void main()
{
position = gl_ModelViewMatrix * gl_Vertex;
normal = normalize( gl_NormalMatrix * gl_Normal.xyz );
gl_Position = gl_ModelViewProjectionMatrix * gl_Vertex;
coord = vec2(gl_MultiTexCoord0);
}
);
const char* frag = GLSL
(
110,
uniform sampler2D tex1;
varying vec4 position;
varying vec3 normal;
varying vec2 coord;
void main()
{
gl_FragColor = texture2D(tex1, coord);
}
);
EDIT2:
I am setting up gl this way:(maybe something is conflicting with my texture shader, but I don't think so!
/* set up depth-buffering */
glEnable(GL_DEPTH_TEST);
glEnable(GL_POLYGON_SMOOTH);
glHint(GL_POLYGON_SMOOTH_HINT, GL_FASTEST);
/* set up lights */
glMatrixMode(GL_MODELVIEW);
glLoadIdentity();
glPolygonMode(GL_FRONT_AND_BACK,GL_FILL);
glShadeModel(GL_SMOOTH);
GLfloat lightpos[] = { 3.0, 0.0, 1.0, 0.0 };
GLfloat lightcolor[] = { 0.5, 0.5, 0.5, 1.0 };
GLfloat ambcolor[] = { 0.5, 0.5, 0.5, 1.0 };
glLightModelfv(GL_LIGHT_MODEL_AMBIENT,ambcolor);
glEnable(GL_LIGHTING);
glColorMaterial(GL_FRONT_AND_BACK,GL_AMBIENT_AND_DIFFUSE);
glEnable (GL_COLOR_MATERIAL);
glLightfv (GL_LIGHT0,GL_POSITION,lightpos);
glLightfv (GL_LIGHT0,GL_AMBIENT,ambcolor);
glLightfv (GL_LIGHT0,GL_DIFFUSE,lightcolor);
glLightfv (GL_LIGHT0,GL_SPECULAR,lightcolor);
glLightf (GL_LIGHT0,GL_CONSTANT_ATTENUATION,0.2);
glLightf (GL_LIGHT0,GL_LINEAR_ATTENUATION,0.0);
glLightf (GL_LIGHT0,GL_QUADRATIC_ATTENUATION,1.0);
glEnable (GL_LIGHT0);
glEnable(GL_TEXTURE_2D);
Replacing gl_FragColor by a flat color is working fine.
I know maybe is related to coord parameter but I am trying all the stuff I found and nothing is working for me.
The internal texture format GL_BGR is not valid. GL_BGR is a valid for the format of the source texture, but the internal representation has to be GL_RGB.
See glTexImage2D.
Adapt your code like this:
glTexImage2D(GL_TEXTURE_2D, 0, GL_RGB, width, height, 0, GL_BGR, GL_UNSIGNED_BYTE, data);
See the Khronos reference page GLAPI/glTexImage2D which says:
To define texture images, call glTexImage2D. The arguments describe the parameters of the texture image, such as height, width, width of the border, level-of-detail number (see glTexParameter), and number of color components provided. The last three arguments describe how the image is represented in memory.
format​ determines the composition of each element in data​. It can assume one of these symbolic values:
GL_BGR:
Each element is an RGB triple. The GL converts it to floating point and assembles it into an RGBA element by attaching 1 for alpha. Each component is clamped to the range [0,1].

Simple primitive rotation Opengl c

Im trying to do a simple rotation in opengl of my primitive object in the projection plane. I want to rotate the object like a propeller but i cant seem to get it going right. When i run the code my object looks like it shrinks into itself (i know its not that, but its rotating funny)
void rotateStuff()
{
spin = spin - .5; // inc for spin
if(spin < 360)
{
spin = spin + 360;
}
glPushMatrix();
glTranslatef(150, 95, 0.0);
glRotatef(spin, 1.0, 0.0, 0.0);
glTranslatef(-150, -95, 0);
displayStuff();
glPopMatrix();
drawButton();
glutSwapBuffers();
}
Heres a snippet of my object
glBegin(GL_POLYGON);
glVertex2i(50, 0);
glVertex2i(50, 75);
glVertex2i(150, 75);
glVertex2i(150, 0);
glEnd(); // end current shape
I think something is wrong with the setting of my origin but what exaclty? am i translating to a wrong origin?
This is a rotation around the x-axis: glRotatef(spin, 1.0, 0.0, 0.0).
Presumably you want things in the x-y plane to stay in the x-y plane,
so you want rotation around the z-axis: glRotatef(spin, 0.0, 0.0, 1.0).

Trouble getting view (lookat) and projection (perspective) matrices to work properly

I've been following the open.gl tutorials without using the the GLM library because reasons (stubbornness and C).
I can't get the view and projection matrices to work properly.
Here's the relevant vertex shader code,
#version 150 core
in vec3 size;
in vec3 color;
in vec2 texcoord;
out vec3 Color;
out vec2 Texcoord;
uniform vec3 pos;
uniform float angle;
uniform vec3 camPos;
uniform vec3 camTarget;
const float fov=90, ratio=4.0/3.0, near=1.0, far=10.0;
mat4 projection ()
{
float t = tan(radians(fov)),
l = ratio * t;
return mat4(
vec4(near/l, 0.0, 0.0, 0.0),
vec4(0.0, near/t, 0.0, 0.0),
vec4(0.0, 0.0, -(far+near)/(far-near), -(2*far*near)/(far-near)),
vec4(0.0, 0.0, -1.0, 0.0)
);
}
mat4 rotZ(float theta)
{
return mat4(
vec4(cos(theta), -sin(theta), 0.0, 0.0),
vec4(sin(theta), cos(theta), 0.0, 0.0),
vec4(0.0, 0.0, 1.0, 0.0),
vec4(0.0, 0.0, 0.0, 1.0)
);
}
mat4 translate(vec3 translation)
{
return mat4(
vec4(1.0, 0.0, 0.0, 0.0),
vec4(0.0, 1.0, 0.0, 0.0),
vec4(0.0, 0.0, 1.0, 0.0),
vec4(translation.x, translation.y, translation.z, 1.0)
);
}
mat4 lookAtRH(vec3 eye, vec3 target)
{
vec3 zaxis = normalize(target - eye); // The "forward" vector.
vec3 xaxis = normalize(cross(vec3(0.0,0.0,1.0), zaxis));// The "right" vector.
vec3 yaxis = normalize(cross(zaxis, xaxis)); // The "up" vector.
mat4 axis = {
vec4(xaxis.x, yaxis.x, zaxis.x, 0),
vec4(xaxis.y, yaxis.y, zaxis.y, 0),
vec4(xaxis.z, yaxis.z, zaxis.z, 0),
vec4(dot(xaxis,-eye), dot(yaxis,-eye), dot(zaxis,-eye), 1)
};
return axis;
}
void main()
{
Color = color;
Texcoord = texcoord;
mat4 model = translate(pos) * rotZ(angle);
mat4 view = lookAtRH(camPos, camTarget);
gl_Position = projection() * view * model * vec4(size, 1.0);
}
From tweaking things around it seems as if the view matrix is correct, but the projection matrix is causing the dodgyness.
First I must remark that it is a very bad idea to do this directly in the shaders.
However, if you really want to, you can do this. You should be aware that the GLSL matrix constructors work with column vectors. Your projection matrix is thuse specified transposed (however, your translation matrix is correct).
EDIT: If you want pure C, here is nice lib for math (+ you can check the code :) ) https://github.com/datenwolf/linmath.h
Never do something like that :) Creating matrices in shader is very bad idea...
Vertex shader is executed for each vertex. So if you pass to shader thousand vertices you calculate new matrices thousand times. I think there's nothing more to explain :)
#include <glm/glm.hpp>
#include <glm/gtc/matrix_transform.hpp>
... // somewhere
glm::mat4 projection = glm::perspective(45.0f, float(window_width) / window_height, 0.1f, 100.0f);
glm::mat4 world(1.0f); // world/model matrix
glm::mat4 view(1.0f); // view/camera matrix
glm::mat4 mvp = projection * view * model;
... // inside the main loop
glUniformMatrix4fv(glGetUniformLocation(program, "mvpMatrix"), 1, GL_FALSE, &mvp[0][0]);
draw_mesh();
It's really cool and optimised :)

OpenGL Lighting Failing when Scaling

I have to read a 3D object from an ASE file. This object turns to be too big for the world I have to create, therefore, I must scale it down.
With its original size, it is properly lighted up.
However, once I scale it down, it becomes oversaturated.
The world is centered around (0, 0, 0) and it is 100 meters long (y axis) and 50 meters wide (x axis), my upVector is (0, 0, 1). There are two lights, light0 in (20, 35, 750) and light1 in (-20, -35, 750).
Relevant parts of the code:
void init(void){
glClearColor(0.827, 0.925, 0.949, 0.0);
glEnable(GL_DEPTH_TEST);
glEnable(GL_COLOR_MATERIAL);
glColorMaterial(GL_FRONT, GL_DIFFUSE);
glEnable(GL_LIGHT0);
glEnable(GL_LIGHT1);
glEnable(GL_LIGHTING);
glShadeModel(GL_SMOOTH);
GLfloat difusa[] = { 1.0f, 1.0f, 1.0f, 1.0f}; // white light
glLightfv(GL_LIGHT0, GL_DIFFUSE, difusa);
glLightfv(GL_LIGHT1, GL_DIFFUSE, difusa);
loadObjectFromFile("objeto.ASE");
}
void display ( void ) {
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
glMatrixMode(GL_MODELVIEW);
glLoadIdentity();
gluLookAt(eyeX, eyeY, eyeZ, atX, atY, atZ, 0.0, 0.0, 1.0);
GLfloat posicion0[] = { 20.0f, 35.0f, 750.0f, 1.0f};
glLightfv(GL_LIGHT0, GL_POSITION, posicion0);
GLfloat posicion1[] = { -20.0f, -35.0f, 750.0f, 1.0f};
glLightfv(GL_LIGHT1, GL_POSITION, posicion1);
glColor3f(0.749, 0.918, 0.278);
glPushMatrix();
glTranslatef(0.0, 0.0, 1.5);
//Here comes the problem
glScalef(0.08, 0.08, 0.08);
glBegin(GL_TRIANGLES);
for(int i = 0; i < numFaces; i++){
glNormal3d(faces3D[i].n.nx, faces3D[i].n.ny, faces3D[i].n.nz);
glVertex3d(vertex[faces3D[i].s.A].x, vertex[faces3D[i].s.A].y, vertex[faces3D[i].s.A].z);
glVertex3d(vertex[faces3D[i].s.B].x, vertex[faces3D[i].s.B].y, vertex[faces3D[i].s.B].z);
glVertex3d(vertex[faces3D[i].s.C].x, vertex[faces3D[i].s.C].y, vertex[faces3D[i].s.C].z);
}
glEnd();
glPopMatrix();
glutSwapBuffers();
}
Why does lighting fail when the object is scaled down?
The problem you're running into is, that scaling the modelview matrix also influences the "normal matrix" normals are transformed with. The "normal matrix" is actually the transpose of the inverse of the modelview matrix. So by scaling down the modelview matrix, you're scaling up the normal matrix (because of the modelview inversion step used to obtain it).
Because of that the transformed normals must be rescaled, or normalized if the scale of the modelview matrix is not unitary. In fixed function OpenGL there are two methods to do this: Normal normalization (sounds funny, I know) and normal rescaling. You can enable either with
glEnable(GL_NORMALIZE);
glEnable(GL_RESCALE_NORMALS);
In a shader you'd simply normalize the transformed normal
#version ...
uniform mat3 mat_normal;
in vec3 vertex_normal;
void main()
{
...
vec3 view_normal = normalize( mat_normal * vertex_normal );
...
}
Depending on the setting of GL_NORMALIZE and GL_RESCALE_NORMALS, your normals can be transformed by the OpenGL-Pipeline.
Start with glEnable(GL_NORMALIZE) and see if that solves your problem

3D pyramid appears scattered, with mixed up sides

First of all I defined a structure to express the coordinated of a pyramid:
typedef struct
{
GLfloat xUp;
GLfloat yUp;
GLfloat zUp;
GLfloat base;
GLfloat height;
}pyramid;
Pretty self-explanatory here : I store the coordinates of the uppest point, the base and the height.
The I wrote a function to draw a pyramid:
void drawPyramid(pyramid pyr)
{
GLfloat p1[]= {pyr.xUp+pyr.base/2.0, pyr.yUp-pyr.height, pyr.zUp-pyr.base/2.0};
GLfloat p2[]= {pyr.xUp+pyr.base/2.0, pyr.yUp-pyr.height, pyr.zUp+pyr.base/2.0};
GLfloat p3[]= {pyr.xUp-pyr.base/2.0, pyr.yUp-pyr.height, pyr.zUp+pyr.base/2.0};
GLfloat p4[]= {pyr.xUp-pyr.base/2.0, pyr.yUp-pyr.height, pyr.zUp-pyr.base/2.0};
GLfloat up[]= {pyr.xUp, pyr.yUp, pyr.zUp};
glBegin(GL_TRIANGLES);
glColor4f(1.0, 0.0, 0.0, 0.0);
glVertex3fv(up);
glVertex3fv(p1);
glVertex3fv(p2);
glColor4f(0.0, 1.0, 0.0, 0.0);
glVertex3fv(up);
glVertex3fv(p2);
glVertex3fv(p3);
glColor4f(0.0, 0.0, 1.0, 0.0);
glVertex3fv(up);
glVertex3fv(p3);
glVertex3fv(p4);
glColor4f(1.0, 1.0, 0.0, 0.0);
glVertex3fv(up);
glVertex3fv(p4);
glVertex3fv(p1);
glEnd();
glColor4f(0.0, 1.0, 1.0, 0.0);
glBegin(GL_QUADS);
glVertex3fv(p1);
glVertex3fv(p2);
glVertex3fv(p3);
glVertex3fv(p4);
glEnd();
}
I struggled to draw all the vertices in anti-clockwise order, but probably I messed up something.
This is how I display the pyramid in my rendering function:
void display()
{
glClearColor(0.0, 0.0, 0.0, 0.0);
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
glTranslatef(0.0, -25.0, 50.0);
glRotatef(-angle, 0.0, 1.0, 0.0);
glTranslatef(0.0, 25.0, -50.0);
pyramid pyr;
pyr.xUp=0.0;
pyr.yUp=10.0;
pyr.zUp=50.0;
pyr.base=10.0;
pyr.height=18.0;
glColor4f(1.0, 0.0, 0.0, 0.0);
drawPyramid(pyr);
glutSwapBuffers();
}
I also use an init method called before the glut main loop:
void init()
{
glEnable(GL_DEPTH);
glViewport(-1.0, 1.0, -1.0, 1.0);
glMatrixMode(GL_PROJECTION);
glLoadIdentity();
gluPerspective(35.0, 1.0, 1.0, 100.0);
glMatrixMode(GL_MODELVIEW);
glLoadIdentity();
gluLookAt(0.0,1.0,0.0, 0.0,1.0,30.0, 0.0,1.0,0.0);
}
angle is just a double that I use to rotate the pyramid, changeable by pressing 'r', but this is not relevant.It appears that the real problem is how I draw the vertices.
The problem is that the faces of the pyramid appear scattered, messed up.I would better describe this situation with an image:
There's a face that is too small, that is displayed and I don't know why.
If I rotate the pyramid it appears messed up, I even recored a video to describe this.
Later I could upload it if the problem is not totally clear.
PS: Many people have noticed that I am using outdated techniques.But unfortunately this is what my university offers.
EDIT
I forgot to say about the main function:
int main(int argc, char** argv)
{
glutInit(&argc, argv);
glutInitWindowPosition(100, 100);
glutInitWindowSize(500, 500);
glutInitDisplayMode(GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH);
glutCreateWindow("Sierpinsky Pyramid");
glutDisplayFunc(display);
glutKeyboardFunc(keyboard);
init();
glutMainLoop();
return 0;
}
It looks like depth buffer isn't initialzied.
Calling glEnable(GL_DEPTH_TEST) is not enough. You must correctly initialize glut and specify that you want depth buffer support, otherwise you won't get a depth buffer. If I remember correctly, this is done using glutInitDisplayMode(GLUT_DEPTH|...). See documentation here and introduction here. Additional info can be found using google.
--EDIT--
You're passing invalid parameter to glEnable. call glEnable(GL_DEPTH_TEST) instead of glEnable(GL_DEPTH).
Also:
Matrix code in display function isn't protected by glPushMatrix/glPopMatrix. Which means that every time you rotate pyramid, rotation is applied to previous transform. I.e. calling display function will rotate the pyramid.
glViewport is called with invalid parameters. glViewport takes 4 integer arguments, but you're trying to pass floats. Also, what's "width of -1.0" supposed to mean?
You have not checked any error codes (glGetError). If you tried to call glGetError after glEnable call, then you'd see that it returns GL_INVALID_ENUM.
OpenGL has documentation. Documentation is available on opengl.org. Use it and read it. Also, I'd recommend reading "OpenGL red book".

Resources