Hi everybody. I am new to OpenGL and I am also new here and this is my first post. I am from Slovakia and my English is not that great, but I hope you will understand what I am writing (sorry for mistakes) I have got a problem with shadow mapping implementation using GLSL. Everything works fine using glTexGeni with GL_EYE_LINEAR. I generate my shadow map, transform it and bind to texture. Problem is multitexturing because I have nvidia card which supports only 4 active texture units (also another pass for rendering shadowed area - on ATI I used GL_TEXTURE_COMPARE_FAIL_VALUE_ARB ) and I want to use diffuse map and normal map on model so only 2 remain for shadows and of course I want to use more shadows for some scenes. So my way leads to GLSL. I read a lot of tutorials and forums but I did not find good explanation for this. So far I was able to texture objects using GLSL and combine it with another textures. My real problem is projective shadow texture. In my current code I am using: glTexGenfv(GL_S, GL_EYE_PLANE, &(textureMatrix[0])); glTexGenfv(GL_T, GL_EYE_PLANE, &(textureMatrix[4])); glTexGenfv(GL_R, GL_EYE_PLANE, &(textureMatrix[8])); glTexGenfv(GL_Q, GL_EYE_PLANE, &(textureMatrix[12])); and as you can see my textureMatrix is matrix generated by translating->multiplying->transposing etc model view and projection matrix of my active light. Before I render my scene I activate texture unit, use glTexGenfv for shadow texture and disable mapping(I will enable it later in rendering) for active lights and everything works great. If I try to send this texture unit to my shader, my texture projection is not working well. What I need to know is how to perform this texgen operation for my projective texture for GLSL, if there is any equivalent for this function or some matrix operations. I already found some tuts for EYE_LINEAR mapping but I was unable to successfully implement it. I will really appreciate any hint. Thanks for replies Stanley [Edited by - SaTANO on January 8, 2010 6:06:25 PM]

Thanks V-man
I was out for a while (NEW YEAR and some semestral exams) and after I read the post I find out that there must be problem somewhere else and tried different approach. For many people works simple formula bias*lightProjection*lightModelView. I think the problem is somewhere in my matrixe. I am changing modelview matrix every time i am rendering object (cameraModelView translated rotated scaled to rendered object position). In Fabien Sanglard tutorial of shadow mapping I saw that when he is translating modelview matrix to object he is also translating texture matrix with same values. So first problem was here

As you said I need to transform each vertex from camera space to eye space. So its matrix only operations and I think this should works but its not

load bias=> multiply light projection => multiply light modelview => Scale Translate Rotate

In shader I should be able to use

gl_TextureMatrix[2] * gl_Vertex; // I stored light matrix to texture2

What I think is correct. I tried to project shadow map from camera to see if my shader works correctly so I used gl_ModelViewProjectionMatrix * gl_Vertex and switch my camera to position of light and it works.
There must be problem in my matrix.

I would like to know if there is some additional operation for shadowtexture matrix.

[Edited by - SaTANO on January 8, 2010 6:11:31 PM]

Punch me
Problem solved
I was making 3 different algorithms to calculate matrix and every one was right
After 2 weeks of researching, reading books/forums/tutorials, spamming my friend I found problem
Guess what
I was sending matrix to my function and store it (I used only pointer) and do matrix magic on them (for texgen operation). But what I forget is after I make this, I changed my matrix so I was multiplying already multiplied matrix. It was really strange when I send same matrix to shader but I have got different results so I decide to print this matrix and saw that "What Am I stupid?"
And this stupid thing cost me 2 weeks

So it works like charm

[Edited by - SaTANO on January 8, 2010 6:59:13 PM]