hello, I encounter a problem when getting the viewport: first I used the code glGetIntegerv (GL_VIEWPORT, viewport); to get the width and height of the current window. But according to the returned value, I found that the height is a bit larger than the actual window height. Let vheight be the height value return by the above function, and let wheight be the acctual window''s height. Their subtraction (vheight - wheight) will be useful later. What is more, I typed the following two lines: GLfloat* zbuffer = (GLfloat*)calloc(viewport[2] * viewport[3], sizeof(GLfloat)); glReadPixels(0, 0, viewport[2], viewport[3], GL_DEPTH_COMPONENT, GL_FLOAT, zbuffer); I am sure that the depth component has been successfully read into the zbuffer. However, I think there is something wrong in connection between the screen coordinate and the viewport coord. The problem appear like this: when I use my mouse to click a pixel on the screen and get its screen coord(x,y), I can''t find its depth component in zbuffer[y* width +x], but could get it in zbuffer[ (y + (vheight-wheight) * width +x]. It seem to me that the content in the zbuffer does a translation vertically. I also want to say that similar code generates successful result in other environment. I don''t know if the above problem is common or there is something I need to correct. I''d like to listen to your opinion.

first :
height = viewport[ 3 ] - viewport[ 1 ]
width = viewport[ 2 ] - viewport[ 0 ]

when you read back form the buffer coordinat system is changed. It''s origin is not at top left corner going right&down like in ortho mode but in lower left corner going right&up

pos = x + (height-y) * width or something like that...

There are more worlds than the one that you hold in your hand...