Orthographic Projection Issue

Started by
7 comments, last by nickyc95 9 years, 8 months ago

Hi there guys,

I have a problem with my Ortho Matrix. The engine uses the perspective projection fine but for some reason the Ortho matrix is messed up. (See screenshots below).

Can anyone understand what is happening here?

VIDEO Shows the same scene, rotating on the Y axis.

Ortho Projection code


void Matrix4f::InitOrthoProjTransform(float left, float right, float top, float bottom, float zNear, float zFar)
{
m[0][0] = 2 / (right - left);
m[0][1] = 0;
m[0][2] = 0;
m[0][3] = 0;

m[1][0] = 0;
m[1][1] = 2 / (top - bottom);
m[1][2] = 0;
m[1][3] = 0;

m[2][0] = 0;
m[2][1] = 0;
m[2][2] = -1 / (zFar - zNear);
m[2][3] = 0;

m[3][0] = -(right + left) / (right - left);
m[3][1] = -(top + bottom) / (top - bottom);
m[3][2] = -zNear / (zFar - zNear);
m[3][3] = 1;
}

Then using projectionMatrix.InitOrthoProjTransform(0.0f, 800.0f, 0.0f, 600.0f, 0.1f, 1000.0f);

OK so the first screenshot that is attached shows the Ortho, second one shows perspective (rotated slightly to show persp).

Advertisement

you should use these parameters (-1.0, 1.0, -1.0*height/width, 1.0*height/width, 0.1, 1000) for ortho

in order (bottom , top , left , right , near , far)

you should use these parameters (-1.0, 1.0, -1.0*height/width, 1.0*height/width, 0.1, 1000) for ortho

in order (bottom , top , left , right , near , far)

Why?

I thought that when you create an Ortho Matrix you use the screen dimensions to get a "pixel perfect" area to draw in (e.g. 800*600)

Thanks

no actually there is no pixels there so the point is a ratio between height and width and you should take the origin in center

no actually there is no pixels there so the point is a ratio between height and width and you should take the origin in center


Can you expand further?

Thanks

check this : http://www.glprogramming.com/red/chapter03.html#name3

and its mostly the idea of perspective in ortho u don't really nead to define top left and etc. its actually just used for things like scaling.

ortho matrix is easy to master in the manner of width and height. but you can easiily screw up depth! resulting in paranormal oclusion- or, objects being further ocluding objects being closer. Watch out. (Ortho normal projection is an observation phenomena that is being rather paranormal- as it is)

you should use these parameters (-1.0, 1.0, -1.0*height/width, 1.0*height/width, 0.1, 1000) for ortho
in order (bottom , top , left , right , near , far)

This is not correct.
http://msdn.microsoft.com/en-us/library/windows/desktop/dd373965(v=vs.85).aspx

You are already passing the correct values, which should be actual pixel sizes.

Your Z values are both incorrect.

m[2][2] = -1 / (zFar - zNear); should be m[2][2] = -2 / (zFar - zNear);.

m[3][2] = -zNear / (zFar - zNear); should be m[3][2] = -(zFar + zNear) / (zFar - zNear);.

This is OpenGL, not Direct3D.

Also, try transposing the matrix if these changes do not work.

L. Spiro

I restore Nintendo 64 video-game OST’s into HD! https://www.youtube.com/channel/UCCtX_wedtZ5BoyQBXEhnVZw/playlists?view=1&sort=lad&flow=grid

you should use these parameters (-1.0, 1.0, -1.0*height/width, 1.0*height/width, 0.1, 1000) for ortho
in order (bottom , top , left , right , near , far)

This is not correct.
http://msdn.microsoft.com/en-us/library/windows/desktop/dd373965(v=vs.85).aspx

You are already passing the correct values, which should be actual pixel sizes.

Your Z values are both incorrect.

m[2][2] = -1 / (zFar - zNear); should be m[2][2] = -2 / (zFar - zNear);.

m[3][2] = -zNear / (zFar - zNear); should be m[3][2] = -(zFar + zNear) / (zFar - zNear);.

This is OpenGL, not Direct3D.

Also, try transposing the matrix if these changes do not work.

L. Spiro

Hi,

thanks for "backing me up", haha.

Yeah turns out my matrices weren't consistent with each other, thus caused this issue.

Thanks

This topic is closed to new replies.

Advertisement