• FEATURED

View more

View more

View more

Image of the Day Submit

IOTD | Top Screenshots

The latest, straight to your Inbox.

Subscribe to GameDev.net Direct to receive the latest updates and exclusive content.

Orthographic matrix not working

Old topic!

Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

13 replies to this topic

#1sufimaster  Members

Posted 21 February 2013 - 10:34 PM

Hi all,

I am having trouble creating an orthographic projection in opengl (without using glOrtho).  And before anyone tells me to use glOrtho, lets just pretend I'm using opengl 3.0 as an exercise.  Here is the way I have set up my matrix:

projectionMatrix.m00 = 2/WIDTH;
projectionMatrix.m11 = 2/HEIGHT;
projectionMatrix.m22 = -2/(far_plane - near_plane);
projectionMatrix.m32 = -((far_plane + near_plane)/(far_plane - near_plane));
projectionMatrix.m33 = 1;


This is your typical ortho projection matrix, taken straight from the opengl spec def. for glOrtho.

When I set up a frustrum and use a perspective matrix, this works fine, but I can't use my regular screen coordinates properly.  But when I use the glOrtho matrix above, I see nothing but my background color. Anyone know what's going on? Is my matrix wrong?

My shader looks like this:

#version 150 core
uniform mat4 projectionMatrix;
uniform mat4 viewMatrix;
uniform mat4 modelMatrix;

in vec4 in_Position;
in vec4 in_Color;
in vec2 in_TextureCoord;

out vec4 pass_Color;
out vec2 pass_TextureCoord;

void main(void){
gl_Position = in_Position;

//override gl position with new calculated position
gl_Position = projectionMatrix * viewMatrix * modelMatrix * in_Position;

pass_Color = in_Color;
pass_TextureCoord = in_TextureCoord;

}

Any help appreciated.

Thanks!

#2blueshogun96  Members

Posted 22 February 2013 - 02:25 AM

Try this:

void glOrtho(float* out, float left, float right,float bottom, float top,float near, float far)
{

float a = 2.0f / (right - left);
float b = 2.0f / (top - bottom);
float c = -2.0f / (far - near);

float tx = - (right + left)/(right - left);
float ty = - (top + bottom)/(top - bottom);
float tz = - (far + near)/(far - near);

float ortho[16] = {
a, 0, 0, 0,
0, b, 0, 0,
0, 0, c, 0,
tx, ty, tz, 1
};

memcpy(out, ortho, sizeof(float)*16);
}


Your vertex program should be fine if it works with your perspective matrix.  Let me know if this fixes your problem, this has been confirmed to work for OpenGL ES 2.0.

Shogun.

Follow Shogun3D on the official website: http://shogun3d.net

"Yo mama so fat, she can't be frustum culled." - yoshi_lol

"One objection to a “critique of C#” would be that you can’t talk about C# without talking about the whole “.Net experience”. However, one can approach the topic of Hitler without a complete discussion of Nationalist Socialism, so I feel justified." - Steve White.

#3sufimaster  Members

Posted 22 February 2013 - 11:03 AM

Thanks Shogun - but isn't that matrix the same as the one I have, if I have 0 as left and 680 as right for example (and same for top/bottom/height)?

#4FLeBlanc  Members

Posted 22 February 2013 - 11:12 AM

What geometry are you feeding it?

A mistake I see sometimes is that people will be drawing, say, a quad: (-1,-1,0) -> (1,1,0) or some other geometry on a very small scale. Which will work fine with a perspective matrix, since you can always move your camera to near enough the object that it appears as large in the viewport as you need it to be. But in a screen-size orthographic projection, that 2x2 size quad will only ever be 2x2 pixels on screen, regardless of where you put your camera. This makes switching from perspective to ortho tricky, in that you have to think about the scale of your objects. If you make your base unit to be a meter for example (ie, 1 unit=1 meter) then you might have a character model that is 2 units tall. In an orthographic projection, this object will only be 2 pixels tall.

So if your geometry is small, you either need to scale the geometry or include a scaling factor when constructing your matrix.

#5sufimaster  Members

Posted 22 February 2013 - 11:34 AM

Well right now all I am trying to do is get a 40 pixel by 40 sprite displayed on the screen.  I have tried messing around with the size of the sprite to no avail.  Once I get home I will take a look at the code and give you some more details about it, but I have a feeling you ar right about the scaling/size/etc.

#6sufimaster  Members

Posted 23 February 2013 - 02:14 PM

Ok, so I have a sprite with x,y,z (10,10,0)  with w/h 10fx10f

I have left as 0 and right as 680

top as as 480, bottom as 0

Still I see nothing.

Also, I am using lwjgl, in which if you have a matrix4f, I think m23 means 2nd column, 3rd row.

              Sprite sprite = new Sprite("assets/texture/redman.png", 50f, 50f, 10f, 10f );

projectionMatrix.m00 = 2/WIDTH;
projectionMatrix.m11 = 2/HEIGHT;
projectionMatrix.m22 = -2/(far_plane - near_plane);
projectionMatrix.m30 = -1;
projectionMatrix.m31 = -1;
projectionMatrix.m32 = -((far_plane + near_plane)/(far_plane - near_plane));
projectionMatrix.m33 = 1;

cameraPos = new Vector3f(0,0,-1); ...

//Translate camera
Matrix4f.translate(cameraPos, viewMatrix, viewMatrix);

//Scale, translate and rotate model
Matrix4f.scale(modelScale, modelMatrix, modelMatrix);
Matrix4f.translate(modelPos, modelMatrix, modelMatrix);
Matrix4f.rotate((float)Math.toRadians(modelAngle.z), new Vector3f(0,0,1), modelMatrix, modelMatrix);
Matrix4f.rotate((float)Math.toRadians(modelAngle.y), new Vector3f(0,1,0), modelMatrix, modelMatrix);
Matrix4f.rotate((float)Math.toRadians(modelAngle.x), new Vector3f(1,0,0), modelMatrix, modelMatrix);


..

Edited by sufimaster, 24 February 2013 - 12:16 AM.

#7__SKYe  Members

Posted 24 February 2013 - 07:39 PM

Have you tried disabling back face culling?

If your sprite doesn't have the correct winding, and back face culling is enabled , then it will not be drawn.

Also, in OpenGL the Y coordinate goes from bottom to top, so make sure that you are positioning the sprite inside the view (unlike the projection matrix where point 0.0, 0.0, 0.0, is the center of the screen, on an orthographic projection, it is the left bottom corner).

Although you can change this when you build the matrix.

You can also try a bigger sprite, like with a size of (100, 100).

One last note. You should try the function Shogun posted to calculate your matrix. Somehow, and i might be wrong, but it doesn't seem like your matrix is being calculated like it should (again, it may be just me).

Don't know if this is your problem, but try it and see.

Edited by __SKYe, 24 February 2013 - 07:45 PM.

#8sufimaster  Members

Posted 25 February 2013 - 11:44 PM

Ok- so i used Shogun's method, I messed around with the camera position and near and far plane.

Now:

near plane is 1f, far plane is 100f (though I dont see why this matters in an ortho projection)

Camera is located at 0,0,-1

Sprite is at 0,0,0 dim: 10x10 pixels.

I can actually see it on the screen now, but it seems like everything is vanishing at the origin point, as attached:

This is supposed to be a quad, but the perspective is all wonky. Any idea what's going on now?

#9sufimaster  Members

Posted 28 February 2013 - 02:41 PM

Bump - anyone? This is super frustrating.  Is my ortho matrix still messed up?

#10mhagain  Members

Posted 28 February 2013 - 03:02 PM

How are you sending this to your shader?  Are you multiplying it by your previous projection matrix before doing so?  These are two other areas where things can go wrong, and you haven't provided sufficient info to diagnose.

It appears that the gentleman thought C++ was extremely difficult and he was overjoyed that the machine was absorbing it; he understood that good C++ is difficult but the best C++ is well-nigh unintelligible.

#11sufimaster  Members

Posted 01 March 2013 - 08:52 AM

This is how I send it to the shader: I know this part works since the perspective projection matrix works, but the ortho produces wierd results.

As for my previous projection matrix - this is my first projection matrix - so I am not multiplying it by anything before sending it tothe shader. Should I be?

///upload matrices to uniform varis in shader

projectionMatrix.store(matrix44Buffer);matrix44Buffer.flip();
GL20.glUniformMatrix4(projectionMatrixLocation, false, matrix44Buffer);

viewMatrix.store(matrix44Buffer);matrix44Buffer.flip();
GL20.glUniformMatrix4(viewMatrixLocation, false, matrix44Buffer);

modelMatrix.store(matrix44Buffer);matrix44Buffer.flip();
GL20.glUniformMatrix4(modelMatrixLocation, false, matrix44Buffer);

GL20.glUseProgram(0);


#12sufimaster  Members

Posted 01 March 2013 - 11:00 PM

I managed to figure this out - apparently I was not accessing the matrix elements correctly - transposing their accessors = m03 instead of m30 for example for the translation matrix.  All is well now! Thanks everyone.

#13Vincent_M  Members

Posted 04 March 2013 - 10:58 AM

I noticed you're transforming your matrices as so:

projectionMatrix * viewMatrix * modelMatrix

This is also how I do it, but a lot of people do it opposite (and I think glOrtho does too). If this is the case, than the glOrtho() code provided above may be correct, but it just has to be transposed. Here's my camera's ortho code:

void Camera::Ortho(float width, float height, float zNear, float zFar)
{
Ortho(0.0f, width, 0.0f, height, zNear, zFar);
}

void Camera::Ortho(float left, float right, float top, float bottom, float zNear, float zFar)
{
// find the translation vector
const float tx = - (right + left)/(right - left);
const float ty = - (top + bottom)/(top - bottom);
const float tz = - (zFar + zNear)/(zFar - zNear);

// column 1
projMat.m[ 0] = 2.0f / (right - left);
projMat.m[ 1] = 0;
projMat.m[ 2] = 0;
projMat.m[ 3] = 0;

// column 2
projMat.m[ 4] = 0;
projMat.m[ 5] = 2.0f / (top - bottom);
projMat.m[ 6] = 0;
projMat.m[ 7] = 0;

// column 3
projMat.m[ 8] = 0;
projMat.m[ 9] = 0;
projMat.m[10] = -2.0f / (zFar - zNear);
projMat.m[11] = 0;

// column 4
projMat.m[12] = tx;
projMat.m[13] = ty;
projMat.m[14] = tz;
projMat.m[15] = 1;

mode = PROJECTION_MODE_ORTHOGONAL;
Update(0.0f); // update the camera
}


The first method is just a shortcut I created where the origin is in the upper-left corner of the screen is treated as the origin. The second method is what would emulate glOrtho() from the old OpenGL days, only it should be multiplied first like you have it.

Also, keep in mind that all matrix elements are an array of 16 floats stored on a column-by-column basis:

m[ 0] m[ 4] m[ 8] m[12]
m[ 1] m[ 5] m[ 9] m[13]
m[ 2] m[ 6] m[10] m[14]
m[ 3] m[ 7] m[11] m[15]


When you feed the matrix as a 4x4 to OpenGL, make sure that the transpose argument is GL_FALSE. That's very important.

#14mhagain  Members

Posted 04 March 2013 - 11:42 AM

When you feed the matrix as a 4x4 to OpenGL, make sure that the transpose argument is GL_FALSE. That's very important.

Not really; please see http://www.opengl.org/archives/resources/faq/technical/transformations.htm

Column-major versus row-major is purely a notational convention. Note that post-multiplying with column-major matrices produces the same result as pre-multiplying with row-major matrices. The OpenGL Specification and the OpenGL Reference Manual both use column-major notation. You can use any notation, as long as it's clearly stated.

Also

Sadly, the use of column-major format in the spec and blue book has resulted in endless confusion in the OpenGL programming community. Column-major notation suggests that matrices are not laid out in memory as a programmer would expect.

OpenGL is perfectly capable of using either row-major or column-major matrices (and all the more so with the programmable pipeline); the only important thing is that you be consistent in your code.  That just means getting the multiplication orders correct in both your C/C++ code and your shader code.  You don't have to use the same major-ness in your shaders as you use in your C/C++: either transpose before sending (either manually or by using GL_TRUE in your glUniformMatrix call) or flip the multiplication order in your shader code.

So what's actually important is that you know which major-ness each element of your code uses and that you set things up appropriately for that; otherwise the importance of row-major versus column-major is hugely overstated.

It appears that the gentleman thought C++ was extremely difficult and he was overjoyed that the machine was absorbing it; he understood that good C++ is difficult but the best C++ is well-nigh unintelligible.

Old topic!

Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.