You create a cubic environment map texture by calling the IDirect3DDevice8::CreateCubeTexture method. Cubic environment map textures must be square, with dimensions that are a power of two.

The following code example shows how your C++ application might create a simple cubic environment map.

/*
 * For this example, m_d3dDevice is a valid
 * pointer to an IDirect3DDevice8 interface.
 */

LPDIRECT3DCUBETEXTURE8 m_pCubeMap;

m_d3dDevice->CreateCubeTexture( 256, 1, D3DUSAGE_RENDERTARGET, D3DFMT_R8G8B8,
                                D3DPOOL_DEFAULT, &m_pCubeMap );

You can navigate between faces of a cubic environment map by using the IDirect3DCubeTexture8::GetCubeMapSurface method.

The following code example uses GetCubeMapSurface to retrieve the cube-map surface used for the positive-y face (face 2).

/*
 * For this example, m_pCubeMap is a valid
 * pointer to a IDirect3DCubeTexture8 interface.
 */
LPDIRECT3DSURFACE8 pFace2;

m_pCubeMap->GetCubeMapSurface( D3DCUBEMAP_FACE_POSITIVE_Y, 0, &pFace2);

The first parameter that GetCubeMapSurface accepts is a D3DCUBEMAP_FACES enumerated value that describes the attached surface that the method should retrieve. The second parameter tells Microsoft® Direct3D® which level of a mipmapped cube texture to retrieve. The third parameter accepted is the address of the IDirect3DSurface8 interface, representing the returned cube texture surface. Because this cube-map is not mipmapped, 0 is used here.

Note  After calling this method, the internal reference count on the IDirect3DSurface8 interface is increased. When you are done using this surface, be sure to call the IUnknown::Release method on this IDirect3DSurface8 interface or you will have a memory leak.

The following C++ code example prepares and sets a view matrix according to the face being rendered.

/*
 * For this example, pCubeMap is a valid pointer to a 
 * IDirect3DCubeTexture8 interface and d3dDevice is a
 * valid pointer to a IDirect3DDevice8 interface.
 */
void RenderFaces()
{
    // Save transformation matrices of the device.
    D3DXMATRIX matProjSave, matViewSave;
    d3dDevice->GetTransform( D3DTS_VIEW,       &matViewSave );
    d3dDevice->GetTransform( D3DTS_PROJECTION, &matProjSave );

    // Store the current back buffer and z-buffer.
    LPDIRECT3DSURFACE8 pBackBuffer, pZBuffer;
    d3dDevice->GetRenderTarget( &pBackBuffer );
    d3dDevice->GetDepthStencilSurface( &pZBuffer );

Remember, each face of a cubic environment map represents a 90-degree field of view. Unless your application requires a different field of view angle—for special effects, for example—take care to set the projection matrix accordingly.

This code example creates and sets a projection matrix for the most common case.

    // Use 90-degree field of view in the projection.
    D3DMATRIX matProj;
    D3DXMatrixPerspectiveFovLH( matProj, D3DX_PI/2, 1.0f, 0.5f, 1000.0f );
    d3dDevice->SetTransform( D3DTS_PROJECTION, &matProj );

    // Loop through the six faces of the cube map.
    for( DWORD i=0; i<6; i++ )
    {
        // Standard view that will be overridden below.
        D3DVECTOR vEnvEyePt = D3DVECTOR( 0.0f, 0.0f, 0.0f );
        D3DVECTOR vLookatPt, vUpVec;

        switch( i )
        {
            case D3DCUBEMAP_FACE_POSITIVE_X:
                vLookatPt = D3DVECTOR( 1.0f, 0.0f, 0.0f );
                vUpVec    = D3DVECTOR( 0.0f, 1.0f, 0.0f );
                break;
            case D3DCUBEMAP_FACE_NEGATIVE_X:
                vLookatPt = D3DVECTOR(-1.0f, 0.0f, 0.0f );
                vUpVec    = D3DVECTOR( 0.0f, 1.0f, 0.0f );
                break;
            case D3DCUBEMAP_FACE_POSITIVE_Y:
                vLookatPt = D3DVECTOR( 0.0f, 1.0f, 0.0f );
                vUpVec    = D3DVECTOR( 0.0f, 0.0f,-1.0f );
                break;
            case D3DCUBEMAP_FACE_NEGATIVE_Y:
                vLookatPt = D3DVECTOR( 0.0f,-1.0f, 0.0f );
                vUpVec    = D3DVECTOR( 0.0f, 0.0f, 1.0f );
                break;
            case D3DCUBEMAP_FACE_POSITIVE_Z:
                vLookatPt = D3DVECTOR( 0.0f, 0.0f, 1.0f );
                vUpVec    = D3DVECTOR( 0.0f, 1.0f, 0.0f );
                break;
            case D3DCUBEMAP_FACE_NEGATIVE_Z:
                vLookatPt = D3DVECTOR( 0.0f, 0.0f,-1.0f );
                vUpVec    = D3DVECTOR( 0.0f, 1.0f, 0.0f );
                break;
        }

        D3DMATRIX matView;
        D3DXMatrixLookAtLH( matView, vEnvEyePt, vLookatPt, vUpVec );
        d3dDevice->SetTransform( D3DTS_VIEW, &matView );

Once the camera is in position and the projection matrix set, you can render the scene. Each object in the scene should be positioned as you would normally position them. The following code example, provided for completeness, outlines this task.

        //Get pointer to surface in order to render to it.
        LPDIRECT3DSURFACE8 pFace;
        pCubeMap->GetCubeMapSurface( (D3DCUBEMAP_FACES)i, 0, &pFace );
        d3dDevice->SetRenderTarget ( pFace , pZBuffer );
        pFace->Release();

        d3dDevice->BeginScene();
        // Render scene here.
        d3dDevice->EndScene();
    }

    // Change the render target back to the main back buffer.
    d3dDevice->SetRenderTarget( pBackBuffer, pZBuffer );
    pBackBuffer->Release();
    pZBuffer->Release();

    // Restore the original transformation matrices.
    d3dDevice->SetTransform( D3DTS_VIEW,       &matViewSave );
    d3dDevice->SetTransform( D3DTS_PROJECTION, &matProjSave );
}

Note the call to the IDirect3DDevice8::SetRenderTarget method. When rendering to the cube map faces, you must assign the face as the current render-target surface. Applications that use depth buffers can explicitly create a depth-buffer for the render-target, or reassign an existing depth-buffer to the render-target surface. The code sample above uses the latter approach.