Setting ZBuffer

[Powerino]

i have big problem to find pixelformat where depth buffer works correctly.

my code is:

memset(&m_pfd,0,sizeof(m_pfd));

m_pfd.nSize = sizeof(m_pfd);
m_pfd.nVersion = 1;
m_pfd.dwFlags = PFD_DRAW_TO_WINDOW | PFD_SUPPORT_OPENGL | PFD_DOUBLEBUFFER | PFD_GENERIC_ACCELERATED;
m_pfd.iPixelType = PFD_TYPE_RGBA;
m_pfd.cColorBits = 32;
m_pfd.cDepthBits = 32;
m_pfd.cStencilBits = 8;
m_pfd.iLayerType = PFD_MAIN_PLANE;

//
// Get the pixelformat of this setup.
//
int pf;
PIXELFORMATDESCRIPTOR pfdCheck;

if (!(pf = ::ChoosePixelFormat(m_hDC,&m_pfd)))
{
GiottoPrintf(GIOTTO_MSG_INITIALIZATION,"ChoosePixelFormat() failed!");
return false;
}

//
// Get a description of the pixelformat chosen
///
if (!(DescribePixelFormat(m_hDC,pf,sizeof(PIXELFORMAT DESCRIPTOR),&pfdCheck)))
{
GiottoPrintf(GIOTTO_MSG_INITIALIZATION,"GLView: ChoosePixelFormat() The OpenGL Pixeldescription couldn't be retrieved!");
return false;
}

//
// Is hardware accelerated ?
//
if (((pfdCheck.dwFlags & PFD_GENERIC_ACCELERATED) == 0) && ((pfdCheck.dwFlags & PFD_GENERIC_FORMAT) > 0))
GiottoPrintf(GIOTTO_MSG_INITIALIZATION,"The OpenGL driver that's currently active doesn't support hardware acceleration or there is no decent OpenGL driver installed.");

if (SetPixelFormat(m_hDC,pf,&pfdCheck)!=TRUE)
{
GiottoPrintf(GIOTTO_MSG_INITIALIZATION,"SetPixelFormat() failed!");
return false;
}

if ((m_hRC = wglCreateContext(m_hDC)) == NULL)
{
GiottoPrintf(GIOTTO_MSG_INITIALIZATION,"Cannot create rendering context");
return false;
}

if (!wglMakeCurrent(m_hDC,m_hRC))
{
GiottoPrintf(GIOTTO_MSG_INITIALIZATION,"wglMakeCurrent() failed");;
return false;
}


but in windowed mode with G400 depth buffers doesn't work......anyone knows why!?!!?

[roffe]

Originally posted by Powerino:
i have big problem to find pixelformat where depth buffer works correctly.

m_pfd.cColorBits = 32;
m_pfd.cDepthBits = 32;


The pfd entry cColorBits for RGBA buffers is excluding alpha. You should specify 24.

Are you sure that the G400 supports a 32bit depth buffer? Few consumer level cards do(if any). See other thread in here regarding just that.

[Powerino]

Yes G400 works with 32 bit depth buffer!!!

i have the same problem with gforce4mx!!!

with Color depth 32, Depth 32 stencil 0

WHY?!?!?!

[roffe]

Found this article (http://www.tomshardware.com/graphic/19990628/g400-01.html) which might interest you. According to it, 32bit z-buffer support isn't enabled in GL drivers. This was however, 4 YEARS ago. Maybe Matrox has fixed it and there is some another problem, or maybe not. Matrox isn't really famous for their awesome GL support. Maybe you can whip up a simple Direct3D app and see if it works under DX(they claim it does). And I'm assuming you're using the latest drivers.

[Powerino]

Thanks very much...i have tried all pixelformat and i have found that all pixelformat with doublebuffer in windowed mode doesn't work with my engine.......
disabling doublebuffer it works fine!!!
WHY?!?!?
WHY if the DescribePixelFormat tells me there a a valid pfmt with depth buffer after in opengl it doesn't work!!!!?!?!

[Powerino]

Any idea how to find a valid pixelformat for my application.....using old opengl driver i can't use new wglARB functions for pixelformat!!!!

[roffe]

Originally posted by Mazy:
I dont really think you have 32bit zbuffer on g400.

I highly doubted this too, so I checked the specs at Matrox. And in fact, they do say they support a 32bit depth buffer, including an 8bit stencil. http://www.matrox.com/mga/products/tech_info/pdfs/g400/chip_specs.pdf

I also looked for something like "32bit depth only supported if: ..." but couldn't find any.

[Mazy]

you have 2 choices
either try to send different configs in pfd and try choosepixelformat until you get a good result ( in windows you often have the limitation that if the desktop is 16bit the opengl must be 16bit, and the zbuffer cannot exceed the color in bits. )

or number 2, loop through all pixelformats and run describepixelformats on them and make your own search between the formats.

[zeckensack]

You're requesting a stencil buffer. You can't get more than a 24 bit z buffer then.
I don't know whether 32bit z only is possible.