As you can see the back face is lighting as well

I have a problem with lighting calculations. I am rotating the camera around a cube which was supposed to be illuminated by the light but the resulting lighting is very weird. Check out the image above, fragment shader and vertex shader below.

//Vertex shader

#version 330 core

layout (location = 0) in vec3 pos;
layout (location = 1) in vec3 norm;

uniform mat4 model;
uniform mat4 view;
uniform mat4 projection;
uniform vec3 lightPos;

out vec3 ourNormal;
out vec3 fragPos;
out vec3 ourLightPos;

void main()
    gl_Position = projection*view*model*vec4(pos,1.0);
    fragPos = vec3(view * model * vec4(pos, 1.0f));
    ourNormal = mat3(transpose(inverse(view * model))) * norm;
    ourLightPos = vec3(view * vec4(lightPos, 1.0));

//fragment shader

#version 330 core

//Output variables
out vec4 finalColor;

//Input variables
in vec3 ourNormal;
in vec3 fragPos;
in vec3 ourLightPos;

//Uniform variables

struct Material
    vec3 ambient;
    vec3 diffuse;
    vec3 specular;
    float shininess;

struct Light
    vec3 ambient;
    vec3 diffuse;
    vec3 specular;

uniform Light light;
uniform Material material;
uniform vec3 viewPos;

void main()
    vec3 ambient = light.ambient*material.ambient;

    vec3 normal = normalize(ourNormal);
    vec3 lightDir = normalize(ourLightPos - fragPos);
    float diffuseStrength = max(dot(lightDir,normal),0.0);
    vec3 diffuse = light.diffuse*(diffuseStrength*material.diffuse);

    vec3 viewDir = normalize(viewPos - fragPos);
    vec3 reflectDir = reflect(-lightDir,normal);
    float specularStrength =         pow(max(dot(viewDir,reflectDir),0.0),material.shininess);
    vec3 specular = light.specular * (specularStrength*material.specular);

    //Last step
    vec3 result = (ambient+diffuse+specular);
    finalColor = vec4(result,1.0);

As you can see, I transformed all the positions, normals and light source position in view space, i.e camera space, but I don't know why it's not working.

  • 2
    $\begingroup$ Maybe your model of the box has a wrong normal. $\endgroup$ – user1888 Feb 7 '16 at 10:05
  • 2
    $\begingroup$ This might not be the error, but you have a uniform "viewPos" which is, I assume, the camera position. But you do all calculations in Camera-Space, so this should always be $(0.0, 0.0, 0.0)$ and thereby the viewDir becomes $-1 * fragPos$. $\endgroup$ – Dragonseel Feb 7 '16 at 14:30
  • $\begingroup$ @Dragonseel Yup!! That was the problem! Thank you for taking out some time for this question! $\endgroup$ – Ankit singh kushwah Feb 8 '16 at 17:41
  • 3
    $\begingroup$ @Dragonseel: since it was the correct observation, could you turn the comment into an answer? $\endgroup$ – Julien Guertault Apr 7 '16 at 2:53
  • $\begingroup$ @JulienGuertault yes. Done $\endgroup$ – Dragonseel Apr 7 '16 at 14:59

The problem with your calculations is that you use a uniform for viewPos which is the camera position. The rest of your calculations is in Camera-Space, so the viewPos should always be $(0.0, 0.0, 0.0)$ and thereby the $viewDir$ becomes $-1*\text{fragPos}$.

In general: Always be careful about the different reference frames your calculations can be done in and aim to keep this consistent.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.