Direct3D顶点结构使用总结

D3D里面最基本的就是顶点了，虽说一直在用，可是却也是自己比较模糊的一个点，知道其中的意思，却不是很清楚，今天就总结一下，扫一下这个盲区：

D3D中的顶点缓冲区的声明：

LPDIRECT3DVERTEXBUFFER9 g_pVB        = NULL;    //顶点缓冲区对象

通常都是用LPDIRECT3DVERTEXBUFFER9 来声明顶点缓冲区，它其实就是IDirect3DVertexBuffer9的指针类型，这两个起到的效果是一样的。用LPDIRECT3DVERTEXBUFFER9 声明之后，只是镇定了一个缓冲区的指针，下面还需要开辟一个缓冲区给这个指针。

在开辟真正的内存之前，我们先看一下顶点格式的定义，D3D里面是采用的灵活顶点格式，这点大家应该都是知道的，下面就来总结一下这些灵活顶点格式都具体有哪些，有什么用处。

一般定义顶点结构的时候都是用一个结构体，当然用类去定义也可以，但是一般没有那个必要。

struct CUSTOMVERTEX

{

    FLOAT x, y, z, rhw;

    DWORD color;

};

在还需要定义一个宏，来向D3D说明一下，自己定义的顶点的格式到底有哪些。

#define D3DFVF_CUSTOMVERTEX (D3DFVF_XYZRHW|D3DFVF_DIFFUSE)  //顶点格式

上面这一句话的意思就是，定义的顶点结构包含：位置变换信息（D3DFVF_XYZRHW）和漫反射颜色信息（D3DFVF_DIFFUSE）；

那么，一共都有哪些类型可以定义呢，都有什么样的用呢。

 

---------------------------------------------------------------------------------------------------------------------------------------

 

Vertex Data Flags

#define	Description	Data order and type
D3DFVF_DIFFUSE	Vertex format includes a diffuse color component.	DWORD in ARGB order. See D3DCOLOR_ARGB.
D3DFVF_NORMAL	Vertex format includes a vertex normal vector. This flag cannot be used with the D3DFVF_XYZRHW flag.	float, float, float
D3DFVF_PSIZE	Vertex format specified in point size. This size is expressed in camera space units for vertices that are not transformed and lit, and in device-space units for transformed and lit vertices.	float
D3DFVF_SPECULAR	Vertex format includes a specular color component.	DWORD in ARGB order. See D3DCOLOR_ARGB.
D3DFVF_XYZ	Vertex format includes the position of an untransformed vertex. This flag cannot be used with the D3DFVF_XYZRHW flag.	float, float, float.
D3DFVF_XYZRHW	Vertex format includes the position of a transformed vertex. This flag cannot be used with the D3DFVF_XYZ or D3DFVF_NORMAL flags.	float, float, float, float.
D3DFVF_XYZB1 through D3DFVF_XYZB5	Vertex format contains position data, and a corresponding number of weighting (beta) values to use for multimatrix vertex blending operations. Currently, Direct3D can blend with up to three weighting values and four blending matrices. For more information about using blending matrices, see Indexed Vertex Blending (Direct3D 9).	1, 2, or 3 floats. When D3DFVF_LASTBETA_UBYTE4 is used, the last blending weight is treated as a DWORD.
D3DFVF_XYZW	Vertex format contains transformed and clipped (x, y, z, w) data. ProcessVertices does not invoke the clipper, instead outputting data in clip coordinates. This constant is designed for, and can only be used with, the programmable vertex pipeline.	float, float, float, float

 

Texture Flags

 

#define	Description
D3DFVF_TEX0 - D3DFVF_TEX8	Number of texture coordinate sets for this vertex. The actual values for these flags are not sequential.
D3DFVF_TEXCOORDSIZEN(coordIndex)	Define a texture coordinate data set. n indicates the dimension of the texture coordinates. coordIndex indicates texture coordinate index number. See D3DFVF_TEXCOORDSIZEN and Texture coordinates and Texture Stages.

 

Mask Flags

 

#define	Description
D3DFVF_POSITION_MASK	Mask for position bits.
D3DFVF_RESERVED0, D3DFVF_RESERVED2	Mask values for reserved bits in the FVF. Do not use.
D3DFVF_TEXCOUNT_MASK	Mask value for texture flag bits.

 

Miscellaneous Flags

 

#define	Description
D3DFVF_LASTBETA_D3DCOLOR	The last beta field in the vertex position data will be of type D3DCOLOR. The data in the beta fields are used with matrix palette skinning to specify matrix indices.
D3DFVF_LASTBETA_UBYTE4	The last beta field in the vertex position data will be of type UBYTE4. The data in the beta fields are used with matrix palette skinning to specify matrix indices. // Given the following vertex data definition: struct VERTEXPOSITION { float pos[3]; union { float beta[5]; struct { float weights[4]; DWORD MatrixIndices; // Used as UBYTEs } } }; Given the FVF is declared as: D3DFVF_XYZB5 | D3DFVF_LASTBETA_UBYTE4. Weight and MatrixIndices are included in beta[5], where D3DFVF_LASTBETA_UBYTE4 says to interpret the last DWORD in beta[5] as type UBYTE4.
D3DFVF_TEXCOUNT_SHIFT	The number of bits by which to shift an integer value that identifies the number of texture coordinates for a vertex. This value might be used as shown below. DWORD dwNumTextures = 1; // Vertex has only one set of coordinates. // Shift the value for use when creating a // flexible vertex format (FVF) combination. dwFVF = dwNumTextures << D3DFVF_TEXCOUNT_SHIFT; // Now, create an FVF combination using the shifted value.

 

Examples

The following examples show other common flag combinations.

// Untransformed vertex for lit, untextured, Gouraud-shaded content.
dwFVF = ( D3DFVF_XYZ | D3DFVF_DIFFUSE );

// Untransformed vertex for unlit, untextured, Gouraud-shaded 
//   content with diffuse material color specified per vertex.
dwFVF = ( D3DFVF_XYZ | D3DFVF_NORMAL | D3DFVF_DIFFUSE );

// Untransformed vertex for light-map-based lighting.
dwFVF = ( D3DFVF_XYZ | D3DFVF_TEX2 );

// Transformed vertex for light-map-based lighting with shared rhw.
dwFVF = ( D3DFVF_XYZRHW | D3DFVF_TEX2 );

// Heavyweight vertex for unlit, colored content with two 
//   sets of texture coordinates.
dwFVF = ( D3DFVF_XYZ | D3DFVF_NORMAL | D3DFVF_DIFFUSE | 
          D3DFVF_SPECULAR | D3DFVF_TEX2 );

 

------------------------------------------------------------------------------------------------------------------

 

 

在顶点结构体中没有RHW时，Direct3D将执行视、投影、世界等变换以及进行光线计算，之后你才能在窗口中得到你所绘制的物体。当顶点结构体中有RHW时，就像上面那段英文所述，告知Direct3D使用的顶点已经在屏幕坐标系中了，不再执行视图、投影、世界等变换和光线计算，因为D3DFVF_XYZRHW标志告诉它顶点已经经过了这些处理，并直接将顶点进行光栅操作，任何用SetTransform进行的转换都对其无效。不过这时的原点就在客户区的左上角了，其中x向右为正，y向下为正，而z的意义已经变为z-buffer的象素深度。

    值得注意的是，D3DFVF_XYZRHW和D3DFVF_XYZ、D3DFVF_NORMAL不能共存，因为后两个标志与前一个矛盾。在使用这种顶点时，系统需要顶点的位置已经经过变换了，也就是说x、y必须在屏幕坐标系中，z必须是z-buffer中的象素深度，取值范围：0.0－1.0，离观察者最近的地方为0.0，观察范围内最远可见的地方为1.0。(不过我测试的时候似乎z值不起作用。)引自：http://www.cppblog.com/lovedday/archive/2009/03/22/48507.html

在定义完顶点格式以后，就要开辟一块顶点缓冲区：

g_pd3dDevice->CreateVertexBuffer( 3*sizeof(CUSTOMVERTEX),

                                                  0, D3DFVF_CUSTOMVERTEX,

                                                  D3DPOOL_DEFAULT, &g_pVB, NULL )

开辟缓冲区后，就需要对这个缓冲区进行填写，那么填写的数据呢，也需要先指定出来：

CUSTOMVERTEX vertices[] =

    {

{ 100.0f, 400.0f, 0.5f, 1.0f, 0xffff0000, },

        { 300.0f,  50.0f, 0.5f, 1.0f, 0xff00ff00, },

        { 500.0f, 400.0f, 0.5f, 1.0f, 0xff0000ff, },

    };

然后将数据写入缓冲区：

VOID* pVertices;

    if( FAILED( g_pVB->Lock( 0, sizeof(vertices), (void**)&pVertices, 0 ) ) )

        return E_FAIL;

    memcpy( pVertices, vertices, sizeof(vertices) );

    g_pVB->Unlock();

这里写入的过程用的是Lock函数得到的缓冲区的地址，然后用memcpy函数将自己写好的数据写进去。到这里，顶点就算是创建好了。