飘零风雨亭: 3d编程

maxsdk图解

Tue, 18 Dec 2007 08:14:12 GMT

参见:

3DSMAX_SDK_DavidLanier.pdf

游戏中换装处理（备忘）

Tue, 20 Nov 2007 04:31:25 GMT

游戏中换装处理（备忘）

一.cha 结构:
cha中可有几个model(对应为部位,可切换)
每个model可有几个mtl(即subset),
每个mtl可作为一个pass, 该mtl中可有几层贴图
但是由于硬件多层贴图的限制可考虑只充许2层, 否则可能需要多次pass

二.cha update:
updatebone(计算当前骨骼运行时矩阵) -- 注意,骨骼的骨架数据及骨骼的动画数据可共享
updatemodel(计算模型中动画) -- 注意,模型的网格数据及模型的动画数据可共享

注意:
需要将每model中引用到的骨骼运行时矩阵(在updatebone中计算好了)拷贝到该model中,
以便在后面渲染每个模型时能用上.

优化:
人物不可见时可不更新动画,但要更新动作(即当前播放第几个动作的第几帧),
故上述两函数可再追加一个标志说明是更新动作还是动画

三.cha render:
遍历渲染每一个model:
for(i loop models)
{
1. 设置每个模型的vb和ib: (vb和ib是静态创建的)
    setstream(model's vb && ib)

2. 渲染每一个模型
   a. 把model引用到的运行时矩阵乘上MVP设置给vs
   b. 分pass画:
  for(i loop mtls){
     把每个mtl设置给vs
     drawpass(i)
  }
}
优化:
   在第２部时要把model引用到的运行时矩阵乘上MVP设置给vs中的常量寄存器，
然后用每个顶点的矩阵索引及权重进行混合（矩阵索引需要vs1.1中的寻址寄存器a0.x）

另外，只有整体transform变换的模型（没有顶点或骨骼动画, 如地面上不同位置不同面向的不动的树）要与人物模型分开处理，
这类物件可用类似上述的常量矩阵数组的技术进行instancing批次渲染(带skin的instancing批次渲染由于常量寄存器已变骨骼矩阵使用，
因此一般需要硬件支持，如使用顶点纹理来保存骨骼矩阵，参见前面文章instancing批次渲染)

疑问：
目前有一个疑问，上述的渲染需不需要换成按贴图排序再渲染.
还有怎么在只支持vs1.1的显卡下实现skin instancing，或者有其它什么优化的方案？

instancing批次画

Mon, 19 Nov 2007 02:58:09 GMT

instancing批次画:

具体实现参见dxsdk9.0c中的instancing_2005(未包括最后一种需要sm3.0支持的技术)
一.单个shade const:
    V( g_pEffect->Begin( &cPasses, 0 ) );
    for( iPass = 0; iPass < cPasses; iPass++ )
    {
        V( g_pEffect->BeginPass( iPass ) );

        // Render the boxes with the applied technique
        V( g_pEffect->SetTexture( g_HandleTexture, g_pBoxTexture ) );

        for( int nRemainingBoxes = 0; nRemainingBoxes < g_NumBoxes; nRemainingBoxes++ )
        {
            // set the box instancing array
            V( g_pEffect->SetVector( g_HandleBoxInstance_Position, &g_vBoxInstance_Position[nRemainingBoxes] ) );
            V( g_pEffect->SetVector( g_HandleBoxInstance_Color, (D3DXVECTOR4*)&g_vBoxInstance_Color[nRemainingBoxes] ) );

            // The effect interface queues up the changes and performs them
            // with the CommitChanges call. You do not need to call CommitChanges if
            // you are not setting any parameters between the BeginPass and EndPass.
            V( g_pEffect->CommitChanges() );

            V( pd3dDevice->DrawIndexedPrimitive( D3DPT_TRIANGLELIST, 0, 0, 4 * 6, 0, 6 * 2 ) );
        }

        V( g_pEffect->EndPass() );
    }
    V( g_pEffect->End() );

二.多个shade const(因为使用类似于矩阵板实现skin bone的方法,所以存在冲突不能同时使用在骨骼动画中)
    V( g_pEffect->Begin( &cPasses, 0 ) );
    for( iPass = 0; iPass < cPasses; iPass++ )
    {
        V( g_pEffect->BeginPass( iPass ) );

        // Render the boxes with the applied technique
        V( g_pEffect->SetTexture( g_HandleTexture, g_pBoxTexture ) );

        int nRemainingBoxes = g_NumBoxes;
        while( nRemainingBoxes > 0 )
        {
            // determine how many instances are in this batch (up to g_nNumBatchInstance)
            int nRenderBoxes = min( nRemainingBoxes, g_nNumBatchInstance );

            // set the box instancing array
            V( g_pEffect->SetVectorArray( g_HandleBoxInstance_Position, g_vBoxInstance_Position + g_NumBoxes - nRemainingBoxes, nRenderBoxes) );
            V( g_pEffect->SetVectorArray( g_HandleBoxInstance_Color, (D3DXVECTOR4*)(g_vBoxInstance_Color + g_NumBoxes - nRemainingBoxes), nRenderBoxes) );

            // The effect interface queues up the changes and performs them
            // with the CommitChanges call. You do not need to call CommitChanges if
            // you are not setting any parameters between the BeginPass and EndPass.
            V( g_pEffect->CommitChanges() );

            V( pd3dDevice->DrawIndexedPrimitive( D3DPT_TRIANGLELIST, 0, 0, nRenderBoxes * 4 * 6, 0, nRenderBoxes * 6 * 2 ) );

            // subtract the rendered boxes from the remaining boxes
            nRemainingBoxes -= nRenderBoxes;
        }

        V( g_pEffect->EndPass() );
    }
    V( g_pEffect->End() );

三. vs3.0 的hwinstance(需要多流INSTANCEDATA支持):
这种方法是GPU通过虚拟复制（virtually duplicating）把顶点从第一个流打包到第二个流中。(也就是硬件使用多流和自动打包的方式帮忙实际了类似上述第二种的instance批次画)

    // Stream zero is our model, and its frequency is how we communicate the number of instances required,
    // which in this case is the total number of boxes
    V( pd3dDevice->SetStreamSource( 0, g_pVBBox, 0, sizeof(BOX_VERTEX)) );
    V( pd3dDevice->SetStreamSourceFreq( 0, D3DSTREAMSOURCE_INDEXEDDATA | g_NumBoxes ) );

    // Stream one is the instancing buffer, so this advances to the next value
    // after each box instance has been drawn, so the divider is 1.
    V( pd3dDevice->SetStreamSource( 1, g_pVBInstanceData, 0, sizeof( BOX_INSTANCEDATA_POS ) ) );
    V( pd3dDevice->SetStreamSourceFreq( 1, D3DSTREAMSOURCE_INSTANCEDATA | 1ul ) );

    V( pd3dDevice->SetIndices( g_pIBBox ) );

    // Render the scene with this technique
    // as defined in the .fx file
    V( g_pEffect->SetTechnique( g_HandleTechnique ) );

    V( g_pEffect->Begin( &cPasses, 0 ) );
    for( iPass = 0; iPass < cPasses; iPass++ )
    {
        V( g_pEffect->BeginPass( iPass ) );

        // Render the boxes with the applied technique
        V( g_pEffect->SetTexture( g_HandleTexture, g_pBoxTexture ) );

        // The effect interface queues up the changes and performs them
        // with the CommitChanges call. You do not need to call CommitChanges if
        // you are not setting any parameters between the BeginPass and EndPass.
        V( g_pEffect->CommitChanges() );

        V( pd3dDevice->DrawIndexedPrimitive( D3DPT_TRIANGLELIST, 0, 0, 4 * 6, 0, 6 * 2 ) );

        V( g_pEffect->EndPass() );
    }
    V( g_pEffect->End() );

    V( pd3dDevice->SetStreamSourceFreq( 0, 1 ) );
    V( pd3dDevice->SetStreamSourceFreq( 1, 1 ) );

四. 带骨骼实现的批次(顶点纹理方式):

上述方法不好同时实现skinning, 如果需要实现skinning，可以尝试把所有实体的所有骨骼信息储存为一张纹理，之后为相应的实体选择正确的骨骼，这需要用到Shader Model3.0中的顶点纹理访问功能。如果使用这种技术，那么访问顶点纹理带来的性能消耗是不确定的，应该实现进行测试。

具体例子也可以参考:

http://www.gamedev.net/community/forums/topic.asp?topic_id=366247

http://creators.xna.com/Headlines/developmentaspx/archive/2007/01/01/Mesh-Instancing.aspx

NoInstancing:

ShaderInstancing:

HardwareInstancing:

VFetchInstancing:

WOW的地形shader祥细解释

Wed, 14 Nov 2007 10:37:22 GMT

WOW的地形shader祥细解释

flipcode@msn.com

SPXG // 标识
// 用到5张图片:
{
   blendTexture = 0
   layer0Texture = 1
   layer1Texture = 2
   layer2Texture = 3
   layer3Texture = 4
}
!!ARBfp1.0

// 定义常量:
PARAM c[1] = { { 1, 0.30000001, 0.69999999 } };

// 声明3个寄存器:
TEMP R0;
TEMP R1;
TEMP R2;

// 开始混合:
// 一.第0层和第1层使用第4张图的x通道作为alpha进行混合:
TEX R1, fragment.texcoord[0], texture[0], 2D;
TEX R0, fragment.texcoord[1], texture[1], 2D;
ADD R2, R0, -R1;
// r2=Tex1-Tex0

TEX R0, fragment.texcoord[4], texture[4], 2D;
MAD R2, R0.x, R2, R1;
// r2 = Tex4.x*r2+Tex0

// 说明:
// 其中的Tex4.x是对应第1层alpha值, 下面把Tex4.x当a1看
// 即 r2 = a1 * (Tex1-Tex0) + Tex0
// 转换一下即是: Tex0*(1-a1)+a1*Tex1, 呵呵,看到了吧,这就是混合公式!

// 二.第2层和前面结果使用第4张图的y通道作为alpha进行混合:
TEX R1, fragment.texcoord[2], texture[2], 2D;
ADD R1, R1, -R2;
// r1=Tex2-r2

MAD R2, R0.y, R1, R2;
// r2 = Tex4.y*r1+r2, 即 a2*r1+r2 = a2*(Tex2-r2)+r2, 即r2*(1-a2)+Tex2*a2, 其中r2即是上次0和1层混合后的结果

// 三.第3层和前面结果使用第4张图的z通道作为alpha进行混合:
TEX R1, fragment.texcoord[3], texture[3], 2D;
ADD R1, R1, -R2;
// r1=Tex3-r2

MAD R1, R0.z, R1, R2;
// r1=a3*r1+r2 , 即 a3*r1+r2 = a3*(Tex3-r2)+r2, 即r2*(1-a3)+Tex3*a3
// 这样r1就保存了最终的混合结果

// 四.让阴影地表光泽系数为0(Tex4.w即a通道代表地形的阴影,0为阴影,1为正常.而每层贴图中的a通道是光泽通道,所以R1.w保存的是最终的光泽通道值):
MUL R0.x, R1.w, R0.w;
// r0.x = r1.w(光泽)*Tex4.w(阴影值, 是阴影则=0,否则=1)

// 计算削减系数?
MAD R0.w, R0, c[0].y, c[0].z; // r0乘上0.3(削减了30％)再加上一个常量0.69999999

// 反射高光 = secondary_color(反射光)*光泽:
MUL R0.xyz, R0.x, fragment.color.secondary;

// 贴图最终color = 最终贴图混出的color*削减系数:
MUL R1.xyz, R1, R0.w;

// 贴图最终color * primary_color(光照色或是顶点色) + 反射高光:
MAD result.color.xyz, R1, fragment.color.primary, R0;

// alpha:
MOV result.color.w, c[0].x;

END



参考资料:
<WOW的地形渲染>:
http://dev.csdn.net/author/ZERO2046/1b8df53b7f034e7999d70e0cac5665f6.html
魔兽世界的地形渲染，基本上有三种渲染路径：固定渲染管线（其中是不是又分几种就不清楚了）；shader（带高光）；shader（不带高光）
　　用到shader的渲染路径又分别针对1层，2层，3层，4层（最多允许每个chunk使用4层纹理）专门写了shader代码。
　　用MyWarCraftStudio打开WOW的misc.mpq包，shader \ pixel \ 目录下以"terrain"打头的bls文件就是地形渲染使用的shader，带有"_s"后缀的是带高光的渲染，否则就是不带高光的。
　　我仔细看了其中的terrain4_s.bls（用UltraEdit之类的工具可以直接当成文本文件打开）。是汇编形式的ps代码，由于之前俺只用过c形式的HLSL所以看起来有点吃力，好在还是看明白了。
　　texture0~3就是待混合的4层纹理，每层纹理的a通道是该层纹理对应的高光通道；texture4是一张用来控制混合权重的alpha纹理。texture4的r,g,b通道分别对应texture1~3的alpha值，而texture4.a则代表地形的阴影，0为阴影，1为正常。
　　混合的公式为res_n = res_n-1 * ( 1 - alpha_n ) + texture_n * alpha_n。其中n代表第n层纹理。res_n代表第n层混合后的结果。
　　从代码可以看出，最终混合结果的a通道（高光通道）被乘以阴影值，也就是说阴影中没有高光。同时，阴影中的diffuse光照被削减了30％。
　　1，2，3层纹理混合，以及不带高光的情况想必也没有什么特别之处，也不去细看了。
　　值得一提的是我原来一直以为WOW的地形渲染只用了ps_1_1，但从这个shader看用了5张纹理，超出了ps_1_1的4个纹理采样的限制，所以至少使用了ps_1_4（允许6个纹理采样）。而1～3层纹理混合的情况使用ps_1_1就够了。我不知道在渲染时频繁切换ps版本会不会收到性能上的惩罚，但有些显卡（比如NV的5200）同样的代码在ps_1_4上运行要比在ps_1_1上运行慢很多。
　　至于固定渲染管线的渲染路径，由于不可能看到代码，具体如何实现多层纹理混合无法揣测，集合了各层纹理混合系数的alpha图应该不能直接用于渲染，因为固定渲染管线似乎只能用a通道来控制混合系数，可能需要将这张纹理拆成4张才行。。。

地形上网格投射阴影(备忘)

Wed, 14 Nov 2007 09:30:35 GMT

地形上网格投射阴影:
目前有很多游戏及引擎用到了这种方法,
如开源的glest即时战略游戏和OGRE3D,
方法是:
照相到对到光线方向,渲染地表到深度缓冲,再渲染要投影的物件到贴图
然后将照相机设到顶视图,设置投影映射纹理把贴图贴到地面.

具本参见:
Generating Mesh Shadows On Terrain Using OpenGL
http://gpwiki.org/index.php/Generating_Mesh_Shadows_On_Terrain_Using_OpenGL

地形pvs

Wed, 14 Nov 2007 09:24:43 GMT

地形pvs
将地形分成n*n个区域,分成32层
然后,遍历每个区域, 在每个区域上的每层上向其它区域拉一直接,测试它们之间是否有障碍(不用测试三角形只用比较经过的区域内每个格子高度),是则表示当前所在区域看不见所测区域,保存此PVS供用.
注意PVS要用bit来保存,否则很大

参见:
GameDev.net -- Terrain Geomorphing in the Vertex Shader
http://www.gamedev.net/columns/hardcore/geomorph/

geomorph可将格子进行4次lod, 并在不同lod的格子边界进行插入点形成新的三角形(以免出现T形裂缝)
由于lod形成的高度差,所以还可以设置一个error充许值.具体参见原文(该地形也实现了类似目前一些网游所用的alpha混合多层贴图)

地形碰撞高度计算(备忘)

Wed, 14 Nov 2007 09:10:16 GMT

地形碰撞计算:
一.如果是鼠标点击地面,
可将射线固定长度并分段,然后用射线与每一个分段上的点所落在的地面tile(两个三角形)进行求交,
相交则返回对应位置,否则继续与下一个段上的点所落在的地面tile求交
/*
      0      1
       ---->
      | \ |
      | \ |
    2 V --- 3
*/
这里不讨论点击屏幕求射线方法与射线分段处理,只算求交
如下d3d求交得出uv,再求位置:
VECTOR3 vPickPos;
// 右边:
if( D3DXIntersectTri(&v0, &v1, &v3, &vOrig, &vDir, &u, &v, NULL) == TRUE)
{
  vPickPos = v0 + u * (v1 - v0) + v * (v3 - v0);
  return vPickPos.z;
}
// 左边:
if( D3DXIntersectTri(&v0, &v3, &v2, &vOrig, &vDir, &u, &v, NULL) == TRUE)
{
  vPickPos = v0 + u * (v3 - v0) + v * (v2 - v0);
  return vPickPos.z;
}

二.如果是仅是求地表某点高度
则可用uv重心求法(注意上述第一点的d3d的那个uv重心求法不太一样,他的uv跟他的向量走?)
float u = (fX-(int)fX);
float v = (fY-(int)fY);

float p0 = fHeight[0];
float p1 = fHeight[1];
float p2 = fHeight[2];
float p3 = fHeight[3];

if(u>v){ // 右边
return p0 + u * (p1 - p0) + v * (p3 - p1);;
}
else{ // 左边
return p0 + u * (p3 - p2) + v * (p2 - p0);
}

d3d中实现简单的水

Mon, 12 Nov 2007 08:14:32 GMT

1. 最简单的一张带alpha的贴图直接uv动画
2. 另一种是使用序列贴图, 即水的动画贴图,可在每个自定义大小的格子播放相同的一帧图片来实现.

3. 反射水面:
上述其中一种方法再加上带bump map效果的反射图

下面给出最后一种具体例子:

用d3d固定管道实现bump mapping振动/移动的水:

(注意:该例子只用D3DTOP_BUMPENVMAP没用D3DTOP_BUMPENVMAPLUMINANCE):
  // 混合状态:
  m_pD3DDevice->SetRenderState(D3DRS_SRCBLEND, D3DBLEND_SRCALPHA);
  m_pD3DDevice->SetRenderState(D3DRS_DESTBLEND, D3DBLEND_INVSRCALPHA);
  m_pD3DDevice->SetRenderState(D3DRS_ALPHABLENDENABLE, TRUE);

// 贴图基台0:
  // 贴图因子(见后面的alpha参数D3DTA_TFACTOR):
  m_pD3DDevice->SetRenderState(D3DRS_TEXTUREFACTOR, 0x90000000);
  // 设置Bump map: 就是一些uv偏移值,保存在bump贴图中(还可以保存一个附加的光照值)
  m_pD3DDevice->SetTexture(0, m_pTexBump);
  m_pD3DDevice->SetTextureStageState(0, D3DTSS_COLOROP, D3DTOP_BUMPENVMAP);
  m_pD3DDevice->SetTextureStageState(0, D3DTSS_COLORARG1, D3DTA_TEXTURE);
  m_pD3DDevice->SetTextureStageState(0, D3DTSS_ALPHAOP, D3DTA_DIFFUSE);//D3DTOP_DISABLE);
                // bump矩阵(参见上面msdn链接说明):
  m_pD3DDevice->SetTextureStageState(0, D3DTSS_BUMPENVMAT00, F2DW(0.05f));
  m_pD3DDevice->SetTextureStageState(0, D3DTSS_BUMPENVMAT01, F2DW(0.0f));
  m_pD3DDevice->SetTextureStageState(0, D3DTSS_BUMPENVMAT10, F2DW(0.0f));
  m_pD3DDevice->SetTextureStageState(0, D3DTSS_BUMPENVMAT11, F2DW(0.05f));
  m_pD3DDevice->SetTextureStageState(0, D3DTSS_BUMPENVLSCALE, F2DW(1.0f));
  m_pD3DDevice->SetTextureStageState(0, D3DTSS_BUMPENVLOFFSET, F2DW(0.0f));
                // uv矩阵(uv动画用):
  m_pD3DDevice->SetTransform(D3DTS_TEXTURE0, &m_MatTransEau);
  m_pD3DDevice->SetTextureStageState(0, D3DTSS_TEXTURETRANSFORMFLAGS, D3DTTFF_COUNT2);

                // 贴图基台1:
                // 反射图(可实时计算):
  // Etage 1 : affichage de la texture de r閒lexion
  m_pD3DDevice->SetTexture(1, m_pEnvMiroir->GetMiroir());
                //
  m_pD3DDevice->SetTextureStageState (1, D3DTSS_COLOROP,   D3DTOP_SELECTARG1);
  m_pD3DDevice->SetTextureStageState (1, D3DTSS_COLORARG1, D3DTA_TEXTURE);
  m_pD3DDevice->SetTextureStageState (1, D3DTSS_ALPHAOP,   D3DTOP_SELECTARG1);
  m_pD3DDevice->SetTextureStageState (1, D3DTSS_ALPHAARG1, D3DTA_TFACTOR);
                // uv矩阵(投影矩阵,将贴图映射到水面用):
  m_pD3DDevice->SetTransform(D3DTS_TEXTURE1, &m_pScene->GetCameraCourante()->GetMatriceProjectionTexture());
  m_pD3DDevice->SetTextureStageState(1, D3DTSS_TEXCOORDINDEX, D3DTSS_TCI_CAMERASPACEPOSITION);
  m_pD3DDevice->SetTextureStageState(1, D3DTSS_TEXTURETRANSFORMFLAGS, D3DTTFF_PROJECTED |D3DTTFF_COUNT3);
  m_pD3DDevice->SetSamplerState(1, D3DSAMP_ADDRESSU, D3DTADDRESS_CLAMP);
  m_pD3DDevice->SetSamplerState(1, D3DSAMP_ADDRESSV, D3DTADDRESS_CLAMP);

                // set fvf, vb 和 ib:
  m_pD3DDevice->SetFVF(D3DFVF_CUSTOMVERTEX);
  m_pD3DDevice->SetStreamSource(0, m_pVB, 0, sizeof(CUSTOMVERTEX));
  m_pD3DDevice->SetIndices(m_pIB);

// draw:
m_pD3DDevice->DrawIndexedPrimitive(D3DPT_TRIANGLELIST, 0, 0, NB_VERTEX_EAU, 0, NB_TRIANGLE_EAU);

提示:

上述的bump矩阵只是简单的偏移,实际的uv动画使用了另一个贴图uv矩阵(每帧更新,使得uv移动)来设置的,

另外一个更好技巧是可以使用bump矩阵直接应用sin函数,这样会使用得水波有一定的振动,效果挺好:

D3DXMATRIX matBumpMat;

   const static DWORD startick(GetTickCount());
   DWORD dwtime=GetTickCount()-startick;
   float time=(float)dwtime/1000;
   FLOAT r = 0.005f;
   matBumpMat._11 = r*cosf(time*8.0f);
   matBumpMat._12 = r*sinf(time*8.0f);
   matBumpMat._21 = r*sinf(time*8.0f);
   matBumpMat._22 =-r*cosf(time*8.0f);

   m_pD3DDevice->SetTextureStageState(0,D3DTSS_BUMPENVMAT00,F2DW(matBumpMat._11));
   m_pD3DDevice->SetTextureStageState(0,D3DTSS_BUMPENVMAT01,F2DW(matBumpMat._12));
   m_pD3DDevice->SetTextureStageState(0,D3DTSS_BUMPENVMAT10,F2DW(matBumpMat._21));
   m_pD3DDevice->SetTextureStageState(0,D3DTSS_BUMPENVMAT11,F2DW(matBumpMat._22));

上述方法需要检测是否支持 D3DTEXOPCAPS_BUMPENVMAP

关于bump mapping参见msdn:

1. Bump Mapping (Direct3D 9)
http://msdn2.microsoft.com/en-us/library/bb172379.aspx

2. Bump Mapping Formulas (Direct3D 9)
http://msdn2.microsoft.com/en-us/library/bb172380.aspx

3. Bump Map Pixel Formats (Direct3D 9)
http://msdn2.microsoft.com/en-us/library/bb172381.aspx

4.Using Bump Mapping (Direct3D 9)
http://msdn2.microsoft.com/en-us/library/bb206304.aspx

原理参见第2点,使用方法参见第4点

附:

关于bumpmap图是可以直接程序创建的,下面给出该方法的代码(这是从盖小房（BYOH）source code中看到的,直接贴这里可参考):

LPDIRECT3DTEXTURE9 CreateBumpMap(LPDIRECT3DDEVICE9 device, DWORD dwWidth, DWORD dwHeight,D3DFORMAT d3dBumpFormat )
{
LPDIRECT3DTEXTURE9 psBumpMap;

// Create the bump map texture
if( FAILED( device->CreateTexture( dwWidth, dwHeight, 1, 0/* Usage */,
d3dBumpFormat, D3DPOOL_MANAGED,&psBumpMap,NULL ) ) )
return NULL;

// Lock the surface and write in some bumps for the waves
D3DLOCKED_RECT d3dlr;
psBumpMap->LockRect( 0, &d3dlr, 0, 0 );
CHAR* pDst = (CHAR*)d3dlr.pBits;
CHAR iDu=0, iDv=0;

for( DWORD y=0; y<dwHeight; y++ )
{
  CHAR* pPixel = pDst;
  for( DWORD x=0; x<dwWidth; x++ )
  {
   FLOAT fx = x/(FLOAT)dwWidth - 0.5f;
   FLOAT fy = y/(FLOAT)dwHeight - 0.5f;
   FLOAT r = sqrtf( fx*fx + fy*fy );
//  iDu = (CHAR)( 32 * cosf( 300.0f * r ) * expf( -r * 5.0f ) );
   iDu += (CHAR)( 128 * cosf( 100.0f * ( fx + fy ) ) );
//  iDu += (CHAR)( 16 * cosf( 140.0f * ( fx * 0.85f - fy ) ) );
//  iDv = (CHAR)( 32 * sinf( 300.0f * r ) * expf( -r * 5.0f ) );
   iDv += (CHAR)( 128 * sinf( 100.0f * ( fx + fy ) ) );
//  iDv += (CHAR)( 16 * sinf( 140.0f * ( fx * 0.85f - fy ) ) );
   *pPixel++ = iDu;
   *pPixel++ = iDv;
  }
  pDst += d3dlr.Pitch;
}
psBumpMap->UnlockRect(0);

return psBumpMap;
}

区域选择显示图

Thu, 08 Nov 2007 01:38:22 GMT

最近做的魔兽里区域选择显示图, 这个与地面贴花不一样,要求边框不会因为区域放大而跟着

缩放(与魔兽不一样的是中间不用小格子省点顶点), 而关于地面贴花可以用匹配地形网格再放绽放uv进行调整

并且设置边框模式或clamp模式(图片边框需透明)而简单得到,当然复杂的也可以用投影贴图来作,(类似于实时影子的作法:

先渲染场景深度,再调相机到灯光处渲染需需投影的物件到贴图上,最后将贴图以顶视图方式投影到地形上)

real time 3d scene post processing

Thu, 01 Nov 2007 03:20:55 GMT

具体请参见<<real time 3d scene post processing>>:
http://ati.amd.com/developer/gdce/Oat-ScenePostprocessing.pdf

下面说说简单的hdr原理,
读取hdr文件
将颜色转到64位浮点原贴图中
将原贴图以1/4大小渲染到downsample贴图中（只取高光，即亮度低于１的被过滤掉）
对downsample进行横向和纵向blur

最后将blur与原贴图合成tone map, 即按比例混合在一起（并调整uv越近图中心的越亮，最后还进行高光指数pow）
最后的tone map就是最终的贴图了

具体参见hdr教程参见(非常简单):
http://www.gamedev.net/columns/hardcore/hdrrendering/

另外bloom作法一般是类似的(只是可以不用精度浮点贴图)：

渲染场景到原贴图

取出高亮部分（只取高光，即亮度在一定值以上）

blur 一下

最后与原贴图混合一下得出最终的tone map

附
如下是azure的一个bloom处理方法:
http://www.azure.com.cn/article.asp?id=195

uniform sampler2D texture;
uniform float blurfactor;
uniform float expfactor;
const float cdelp = 0.001953125;
uniform float sharpness;

float4 xposure(float4 cl,float e)
{
return ( exp(expfactor) - exp(expfactor-e)) * cl;
}

void main( float2 texcoord : TEXCOORD0,
out float4 finalColor : COLOR)
{
float delp = blurfactor* cdelp;
float4 texColor = tex2D(texture, texcoord);
float4 color = texColor*2.0 / 22.0;
color += tex2D(texture,float2(texcoord.x+delp, texcoord.y+delp))*5.0 /22.0;
color += tex2D(texture,float2(texcoord.x+delp, texcoord.y- delp))*5.0 /22.0;
color += tex2D(texture,float2(texcoord.x- delp, texcoord.y- delp))*5.0 /22.0;
color += tex2D(texture,float2(texcoord.x- delp, texcoord.y+delp))*5.0 /22.0;
float e = color.r * 0.3 + color.g * 0.59 + color.b * 11;
finalColor = xposure(texColor*sharpness + color*(1-sharpness),e);
}

设备获取

Tue, 23 Oct 2007 07:11:51 GMT

设备获取

#include <stdio.h>
#include <windows.h>
#include <setupapi.h>
#include <devguid.h>
#include <regstr.h>

#pragma comment(lib, "setupapi.lib")

int main( int argc, char *argv[ ], char *envp[ ] )
{
    HDEVINFO hDevInfo;
    SP_DEVINFO_DATA DeviceInfoData;
    DWORD i;

    // Create a HDEVINFO with all present devices.
    hDevInfo = SetupDiGetClassDevs(NULL,
        0, // Enumerator
        0,
        DIGCF_PRESENT | DIGCF_ALLCLASSES );

    if (hDevInfo == INVALID_HANDLE_VALUE)
    {
        // Insert error handling here.
        return 1;
    }

    // Enumerate through all devices in Set.

    DeviceInfoData.cbSize = sizeof(SP_DEVINFO_DATA);
    for (i=0;SetupDiEnumDeviceInfo(hDevInfo,i,
        &DeviceInfoData);i++)
    {
        DWORD DataT;
        LPTSTR buffer = NULL;
        DWORD buffersize = 0;

        //
        // Call function with null to begin with,
        // then use the returned buffer size
        // to Alloc the buffer. Keep calling until
        // success or an unknown failure.
        //
        while (!SetupDiGetDeviceRegistryProperty(
            hDevInfo,
            &DeviceInfoData,
            SPDRP_DEVICEDESC,
            &DataT,
            (PBYTE)buffer,
            buffersize,
            &buffersize))
        {
            if (GetLastError() ==
                ERROR_INSUFFICIENT_BUFFER)
            {
                // Change the buffer size.
                if (buffer) LocalFree(buffer);
                buffer = (char*)LocalAlloc(LPTR,buffersize);
            }
            else
            {
                // Insert error handling here.
                break;
            }
        }

        printf("Result:[%s]\n",buffer);

        if (buffer) LocalFree(buffer);
    }


    if ( GetLastError()!=NO_ERROR &&
         GetLastError()!=ERROR_NO_MORE_ITEMS )
    {
        // Insert error handling here.
        return 1;
    }

    // Cleanup
    SetupDiDestroyDeviceInfoList(hDevInfo);

    return 0;
}

获取显存的方法

Thu, 27 Sep 2007 06:35:04 GMT

获取显存的方法:
1. dx7 ddraw方法:
   GetAvailableVidMem( DWORD &dwTotal, DWORD &dwFree );
   主要问题：
      发现在一台机器的vista系统下创建ddraw时crash!
2. dx9下设备查询方法:
   GetAvailableTextureMem();
   主要问题:
      只获得剩余显存大小（包括AGP），不知道总显存大小.
此函数msdn说明:
IDirect3DDevice9::GetAvailableTextureMem() returns the total available memory, including AGP. Allocating resources based on an assumption of how much video memory you have is not a great idea. For example, what if the card is running under a Unified Memory Architecture (UMA) or is able to compress the textures? There might be more space available than you might have thought. You should create resources and check for 'out of memory' errors, then scale back on the textures. For example, you could remove the top mip-levels of your textures.

3. dxdiag的方法:
需要创建dxidag相关的com对象来查询，只知道总的，不知道当前所剩的.

4. 其它（16位系统的中断获取方法不在考虑范围内）:
通过WMI获取系统信息(包括显卡及显存信息)

我们的解决方案：
因为第３和第４方法相对麻烦，而且不知道会不会因为写多这些代码在不同的玩家机器上引起其它的crash问题
所以不考虑第３和第４方法.

目前MPRender提供第１和第２种方法，第１种ddraw方法用在初始init.log中并且判断系统如果是vista的系统第一种方法将不起作用.　当crash发现需输出显存信息时先尝试使用第１种ddraw方法, 不行（vista系统不使用）则使用第２种方法仅打印所剩显存大小.

第４种方式csdn论坛上一位网友给出的代码:

// videmem.cpp : Defines the entry point for the console application.
//

#include "stdafx.h"
// Get video card name and memory size.
//tested by onega VC20003,windows 2003

#include   "stdafx.h"
#define   _WIN32_DCOM
#include   <iostream>
using   namespace   std;
#include   <comdef.h>
#include   <Wbemidl.h>

# pragma comment(lib, "wbemuuid.lib")

int   main(int   argc,   char   **argv)
{
HRESULT   hres;

//   Step   1:   --------------------------------------------------
//   Initialize   COM.   ------------------------------------------

hres   =     CoInitializeEx(0,   COINIT_MULTITHREADED);
if   (FAILED(hres))
{
  cout   <<   "Failed   to   initialize   COM   library.   Error   code   =   0x"
   <<   hex   <<   hres   <<   endl;
  return   1;                                     //   Program   has   failed.
}

//   Step   2:   --------------------------------------------------
//   Set   general   COM   security   levels   --------------------------
//   Note:   If   you   are   using   Windows   2000,   you   need   to   specify   -
//   the   default   authentication   credentials   for   a   user   by   using
//   a   SOLE_AUTHENTICATION_LIST   structure   in   the   pAuthList   ----
//   parameter   of   CoInitializeSecurity   ------------------------

hres   =     CoInitializeSecurity(
  NULL,
  -1,                                                     //   COM   authentication
  NULL,                                                 //   Authentication   services
  NULL,                                                 //   Reserved
  RPC_C_AUTHN_LEVEL_DEFAULT,       //   Default   authentication
  RPC_C_IMP_LEVEL_IMPERSONATE,   //   Default   Impersonation
  NULL,                                                 //   Authentication   info
  EOAC_NONE,                                       //   Additional   capabilities
  NULL                                                   //   Reserved
  );


if   (FAILED(hres))
{
  cout   <<   "Failed   to   initialize   security.   Error   code   =   0x"
   <<   hex   <<   hres   <<   endl;
  CoUninitialize();
  return   1;                                         //   Program   has   failed.
}

//   Step   3:   ---------------------------------------------------
//   Obtain   the   initial   locator   to   WMI   -------------------------

IWbemLocator   *pLoc   =   NULL;

hres   =   CoCreateInstance(
  CLSID_WbemLocator,
  0,
  CLSCTX_INPROC_SERVER,
  IID_IWbemLocator,   (LPVOID   *)   &pLoc);

if   (FAILED(hres))
{
  cout   <<   "Failed   to   create   IWbemLocator   object."
   <<   "   Err   code   =   0x"
   <<   hex   <<   hres   <<   endl;
  CoUninitialize();
  return   1;                                   //   Program   has   failed.
}

//   Step   4:   -----------------------------------------------------
//   Connect   to   WMI   through   the   IWbemLocator::ConnectServer   method

IWbemServices   *pSvc   =   NULL;

//   Connect   to   the   root\cimv2   namespace   with
//   the   current   user   and   obtain   pointer   pSvc
//   to   make   IWbemServices   calls.
hres   =   pLoc->ConnectServer(
  _bstr_t(L"ROOT\\CIMV2"),   //   Object   path   of   WMI   namespace
  NULL,                                         //   User   name.   NULL   =   current   user
  NULL,                                         //   User   password.   NULL   =   current
  0,                                               //   Locale.   NULL   indicates   current
  NULL,                                         //   Security   flags.
  0,                                               //   Authority   (e.g.   Kerberos)
  0,                                               //   Context   object
  &pSvc                                         //   pointer   to   IWbemServices   proxy
  );

if   (FAILED(hres))
{
  cout   <<   "Could   not   connect.   Error   code   =   0x"
   <<   hex   <<   hres   <<   endl;
  pLoc->Release();
  CoUninitialize();
  return   1;                                 //   Program   has   failed.
}

cout   <<   "Connected   to   ROOT\\CIMV2   WMI   namespace"   <<   endl;


//   Step   5:   --------------------------------------------------
//   Set   security   levels   on   the   proxy   -------------------------

hres   =   CoSetProxyBlanket(
  pSvc,                                                 //   Indicates   the   proxy   to   set
  RPC_C_AUTHN_WINNT,                       //   RPC_C_AUTHN_xxx
  RPC_C_AUTHZ_NONE,                         //   RPC_C_AUTHZ_xxx
  NULL,                                                 //   Server   principal   name
  RPC_C_AUTHN_LEVEL_CALL,             //   RPC_C_AUTHN_LEVEL_xxx
  RPC_C_IMP_LEVEL_IMPERSONATE,   //   RPC_C_IMP_LEVEL_xxx
  NULL,                                                 //   client   identity
  EOAC_NONE                                         //   proxy   capabilities
  );

if   (FAILED(hres))
{
  cout   <<   "Could   not   set   proxy   blanket.   Error   code   =   0x"
   <<   hex   <<   hres   <<   endl;
  pSvc->Release();
  pLoc->Release();
  CoUninitialize();
  return   1;                               //   Program   has   failed.
}

//   Step   6:   --------------------------------------------------
//   Use   the   IWbemServices   pointer   to   make   requests   of   WMI   ----

//   For   example,   get   the   name   of   the   operating   system
_bstr_t   wmi_query_str("SELECT   *   FROM   Win32_VideoController");
//bstr_t("SELECT   *   FROM   Win32_OperatingSystem")
IEnumWbemClassObject*   pEnumerator   =   NULL;
hres   =   pSvc->ExecQuery(
  bstr_t("WQL"),
  wmi_query_str,
  WBEM_FLAG_FORWARD_ONLY   |   WBEM_FLAG_RETURN_IMMEDIATELY,
  NULL,
  &pEnumerator);

if   (FAILED(hres))
{
  cout   <<   "Query   for   operating   system   name   failed."
   <<   "   Error   code   =   0x"
   <<   hex   <<   hres   <<   endl;
  pSvc->Release();
  pLoc->Release();
  CoUninitialize();
  return   1;                               //   Program   has   failed.
}

//   Step   7:   -------------------------------------------------
//   Get   the   data   from   the   query   in   step   6   -------------------

IWbemClassObject   *pclsObj;
ULONG   uReturn   =   0;

while   (pEnumerator)
{
  HRESULT   hr   =   pEnumerator->Next(WBEM_INFINITE,   1,
   &pclsObj,   &uReturn);

  if(0   ==   uReturn)
  {
   break;
  }

  VARIANT   vtProp;
  VariantInit(&vtProp);

  //   Get   the   value   of   the   Name   property
  hr   =   pclsObj->Get(L"Caption",   0,   &vtProp,   0,   0);
  wcout   <<   "   Caption   :   "   <<   vtProp.bstrVal   <<   endl;
  VariantClear(&vtProp);

  hr   =   pclsObj->Get(L"AdapterRAM",   0,   &vtProp,   0,   0);
  wcout   <<   "   AdapterRAM   :   "<<   vtProp.lVal/(1024*1024)<<"   MB"   <<   endl;
  VariantClear(&vtProp);

}

//   Cleanup
//   ========

pSvc->Release();
pLoc->Release();
pEnumerator->Release();
pclsObj->Release();
CoUninitialize();

return   0;       //   Program   successfully   completed.

}

ogl绘制两圆(转)

Thu, 06 Sep 2007 01:20:26 GMT

ogl绘制两圆 (转)

#include <math.h>
#include <vector>
#include <gl/glut.h>

#pragma comment(lib, "opengl32.lib")
#pragma comment(lib, "glu32.lib")
#pragma comment(lib, "glut32.lib")

const float PI = 3.14159265358979323846f;

struct MyVector
{
float x;
float y;
float z;
};

class StackedSphere
{
public:

StackedSphere(const float radius = 1.0f, const unsigned int stacks = 8, const unsigned int slices = 16);
void render() const;

private:
std::vector<MyVector> geometryData_;
std::vector<unsigned short> indexData_;

};

StackedSphere::StackedSphere(const float radius, const unsigned int stacks, const unsigned int slices)
{
unsigned int stackNumber = 0;
for (stackNumber = 0; stackNumber <= stacks; ++stackNumber)
{
for (unsigned int sliceNumber = 0; sliceNumber < slices; ++sliceNumber)
{
float theta = stackNumber * PI / stacks;
float phi = sliceNumber * 2 * PI / slices;
float sinTheta = sin(theta);
float cosTheta = cos(theta);

float sinPhi = sin(phi);
float cosPhi = cos(phi);

MyVector ge = {radius * cosPhi * sinTheta, radius * sinPhi * sinTheta, radius * cosTheta};
geometryData_.push_back(ge);
}
}
for (stackNumber = 0; stackNumber < stacks; ++stackNumber)
{
for (unsigned int sliceNumber = 0; sliceNumber <= slices; ++sliceNumber)
{
indexData_.push_back((stackNumber * slices) + (sliceNumber % slices));
indexData_.push_back(((stackNumber + 1) * slices) + (sliceNumber % slices));
}
}
}

void StackedSphere::render() const
{
glVertexPointer(3, GL_FLOAT, 0, &geometryData_[0]);
glEnableClientState(GL_VERTEX_ARRAY);
glDrawElements(GL_TRIANGLE_STRIP, indexData_.size(), GL_UNSIGNED_SHORT, &indexData_[0]);
}

class SpiralSphere
{

public:

SpiralSphere(const float radius = 1.0f, const unsigned int loops = 8, const unsigned int segmentsPerLoop = 16);
void render() const;

private:

std::vector<MyVector > geometryData_;
std::vector<unsigned short> indexData_;

};

SpiralSphere::SpiralSphere(const float radius, const unsigned int loops, const unsigned int segmentsPerLoop)
{
unsigned int loopSegmentNumber = 0;
for (loopSegmentNumber = 0; loopSegmentNumber < segmentsPerLoop; ++loopSegmentNumber)
{
float theta = 0;
float phi = loopSegmentNumber * 2 * PI / segmentsPerLoop;
float sinTheta = sin(theta);
float sinPhi = sin(phi);
float cosTheta = cos(theta);
float cosPhi = cos(phi);

MyVector ge = {radius * cosPhi * sinTheta, radius * sinPhi * sinTheta, radius * cosTheta};
geometryData_.push_back(ge);

}

unsigned int loopNumber = 0;
for (loopNumber = 0; loopNumber <= loops; ++loopNumber)
{
for (unsigned int loopSegmentNumber = 0; loopSegmentNumber < segmentsPerLoop; ++loopSegmentNumber)
{
float theta = (loopNumber * PI / loops) + ((PI * loopSegmentNumber) / (segmentsPerLoop * loops));
if (loopNumber == loops)
{
theta = PI;
}
float phi = loopSegmentNumber * 2 * PI / segmentsPerLoop;
float sinTheta = sin(theta);
float sinPhi = sin(phi);
float cosTheta = cos(theta);
float cosPhi = cos(phi);

MyVector ge = {radius * cosPhi * sinTheta, radius * sinPhi * sinTheta, radius * cosTheta};
geometryData_.push_back(ge);

}
}

for (loopSegmentNumber = 0; loopSegmentNumber < segmentsPerLoop; ++loopSegmentNumber)
{
indexData_.push_back(loopSegmentNumber);
indexData_.push_back(segmentsPerLoop + loopSegmentNumber);
}

for (loopNumber = 0; loopNumber < loops; ++loopNumber)
{
for (unsigned int loopSegmentNumber = 0; loopSegmentNumber < segmentsPerLoop; ++loopSegmentNumber)
{
indexData_.push_back(((loopNumber + 1) * segmentsPerLoop) + loopSegmentNumber);
indexData_.push_back(((loopNumber + 2) * segmentsPerLoop) + loopSegmentNumber);
}
}
}

void SpiralSphere::render() const
{
glVertexPointer(3, GL_FLOAT, 0, &geometryData_[0]);
glEnableClientState(GL_VERTEX_ARRAY);
glDrawElements(GL_TRIANGLE_STRIP, indexData_.size(), GL_UNSIGNED_SHORT, &indexData_[0]);
}

StackedSphere sphere1(4, 8, 16);
SpiralSphere sphere2(4, 8, 16);
int r = 0;

void reshape(int width, int height){
if (height == 0)
{
height = 1;
}
glMatrixMode(GL_PROJECTION);
glLoadIdentity();
glViewport(0, 0, width, height);
gluPerspective(45.0f, GLfloat(width) / GLfloat(height), 0.1f, 50.0f);
glMatrixMode(GL_MODELVIEW);
glLoadIdentity();
}

void display(){
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
glLoadIdentity();
glTranslatef(0, 0, -25);
glPushMatrix();
glTranslatef(-5, 0, 0);
glRotatef(r, 0, 1, 0);
sphere1.render();
glPopMatrix();
glPushMatrix();
glTranslatef(5, 0, 0);
glRotatef(r, 0, 1, 0);
sphere2.render();
glPopMatrix();
glutSwapBuffers();
r++;
}

int main(int argc, char** argv)
{
glutInit(&argc, argv);
glutInitDisplayMode(GLUT_DEPTH | GLUT_DOUBLE | GLUT_RGBA);
glutInitWindowPosition(0, 0);
glutInitWindowSize(512,512);
glutCreateWindow("Sphere Test");
glutReshapeFunc(reshape);
glutDisplayFunc(display);
glutIdleFunc(display);
glPolygonMode(GL_FRONT, GL_LINE);
glPolygonMode(GL_BACK, GL_LINE);
glutMainLoop();
return 0;
}

常用链接(转)

Wed, 04 Jul 2007 05:33:17 GMT

常用链接(转)
来自: http://www.hays.com.cn/viewthread.php?tid=256

Graphics Engines
Ogre 3D (http://www.ogre3d.org/)
Nebula Device (http://www.radonlabs.de/nebula.html)
Haddd (http://www.haddd.com/)
Irrlicht (http://irrlicht.sourceforge.net/)
Haaf's Game Engine (http://hge.relishgames.com/) (hardware accelerated 2D games engine)

Axiom 3D (http://www.axiom3d.org/) (OGRE for .NET)
TrueVision (http://truevision.sourceforge.net/)
Revolution3D (http://www.revolution3d.net/)
Genesis3D (http://www.genesis3d.com/)
Crystal Space 3D (http://www.crystalspace3d.org/tikiwiki/tiki-view_articles.php)
Panda3D (http://panda3d.etc.cmu.edu/)
Quake 2 Source Code (http://www.fileshack.com/file.x?fid=7547)
Quake 2 (http://www.fileplanet.com/6621/0 ... -Source-Code-v3.21) (complete sources)

Tenebrae (http://tenebrae.sourceforge.net/) (Quake with per pixel lights and shaders)
Wild Magic (http://www.geometrictools.com/index.html) (David Eberly's engine)
Apocalyx (http://apocalyx.sourceforge.net/)
ClanLib Game SDK (http://www.clanlib.org/)
DevLib (http://www.devlib-central.org/mambo/)
kjAPI (http://www.kjapi.com/)
Portable Game Library (PLIB) (http://plib.sourceforge.net/)

RealmForge GDK (http://realmforge.com/)
Delta3D (http://delta3d.org/)
G3D (http://g3d-cpp.sourceforge.net/)
NeoEngine (http://www.neoengine.org/)
QuakeForge (http://www.quakeforge.net/)
XEngine (http://xengine.sourceforge.net/)
Yake (http://www.yake.org/)
Kyra Sprite Engine (http://www.grinninglizard.com/kyra/)

Libraries
Free Game Development Libraries (http://abattoir.wolfpaw.net/personal/gamelibs.php) (external list)
Graphics:
Crazy Eddie's GUI system (http://www.cegui.org.uk/modules/news/)
Cal3D (http://cal3d.sourceforge.net/) (character animation library)
GLFW (http://glfw.sourceforge.net/) (OpenGL Framework)

GLee (http://elf-stone.com/glee.php) (OpenGL easy extension library)
DevIL (http://openil.sourceforge.net/)
Game Texture Loader (http://www.gamedev.net/community ... sp?topic_id=334348)
SDL (http://www.libsdl.org)
Cpw (http://www.mathies.com/cpw/about.html) (another OGL framework)
Allegro (http://www.talula.demon.co.uk/allegro/)
Open Scene Graph (http://www.openscenegraph.org/)

Corona (http://corona.sourceforge.net/)
Titan (http://titan.sourceforge.net/)
SFont (http://www.linux-games.com/sfont/)
CxImage (http://www.xdp.it/cximage.htm)
CImg (http://cimg.sourceforge.net/)
FreeImage (http://freeimage.sourceforge.net/)
BMF_Font (http://trenki.gippsbiz.com/bmf/)
BFont (http://www.cs.unibo.it/%7Edbilli/bfont/bfont.html)
FreeType (freetype.sf.net) (font rendering library)

PaintLib (http://www.paintlib.de/paintlib/)
GLEW (http://glew.sourceforge.net/) (OpenGL extension wrangler library)
gluX
(http://www-evasion.imag.fr/Membres/Sylvain.Lefebvre/glux/)GLUT (http://www.opengl.org/resources/ ... lut_downloads.html)
FreeGLUT (http://freeglut.sourceforge.net/)
Open Inventor (http://oss.sgi.com/projects/inventor/)
FTGL (http://homepages.paradise.net.nz/henryj/code/index.html#FTGL) (OpenGL Font Library

nVidia Scene Graph SDK (http://developer.nvidia.com/object/nvsg_home.html)
OpenSG (http://www.opensg.org/)
ParaGUI (http://www.paragui.org/)
LibUFO (http://libufo.sourceforge.net/index.html) (OpenGL GUI Toolkit)
GPUmesh (http://www-evasion.imag.fr/Membres/Sylvain.Lefebvre/GPUmesh/)
GNU Triangulated Surface Library (http://gts.sourceforge.net/)
libASE (http://interreality.org/projects/libase)
lib3DS (http://lib3ds.sourceforge.net/)

GLOD (http://www.cs.jhu.edu/%7Egraphics/GLOD/) (Geometric Level of Detail for OpenGL)
Video:
FFmpeg (http://ffmpeg.sourceforge.net/index.php)
Dirac (http://sourceforge.net/projects/dirac)
Theora (http://www.theora.org/)
XviD (http://www.xvid.org/)
Networking:
RakNet (http://www.rakkarsoft.com/)

OpenTNL (http://www.opentnl.org/) (Torque network library)
HawkNL (http://www.hawksoft.com/hawknl/)
OpenPlay (http://developer.apple.com/darwin/projects/openplay/)
ReplicaNet (http://www.replicanet.com/)
Game Networking Engine (http://www.rit.edu/%7Ejpw9607/gne/)
Sound:
FMOD (http://www.fmod.org/)
OpenAL (http://www.openal.org/)

Audiere (http://audiere.sourceforge.net/)
BASS (http://www.un4seen.com/)
libsndfile (http://www.mega-nerd.com/libsndfile/)
Ogg Vorbis (http://www.vorbis.com/)
Physics and Collision:
Open Dynamics Engine (ODE) (http://ode.org/)
Tokamak (http://www.tokamakphysics.com/)
Newton Game Dynamics (http://www.physicsengine.com/)
NovodeX SDK (http://www.ageia.com/novodex.html)

Physical (http://physical.alecrivers.com/)
True Axis (http://www.trueaxis.com/)
Open Dynamics Framework (http://www.physicstools.org/)
ColDet (http://photoneffect.com/coldet/) (Free 3D Collision Detection Library)
Opcode (http://www.codercorner.com/Opcode.htm)
RAPID (http://www.cs.unc.edu/%7Egeom/OBB/OBBT.html)
FreeSOLID (http://www.win.tue.nl/%7Egino/solid/index.html)
V-Collide (http://www.cs.unc.edu/%7Egeom/V_COLLIDE/)

Scripting and XML:
Lua (http://www.lua.org/)
AngelScript (http://www.angelcode.com/angelscript/)
CInt (http://root.cern.ch/root/Cint.html)
EOS (http://eosscript.sourceforge.net/)
GameMonkey Script (http://www.somedude.net/gamemonkey/)
Guile (http://www.gnu.org/software/guile/guile.html)
IO (http://www.iolanguage.com/about/)
javascript (http://www.mozilla.org/js/)

Squirrel (http://squirrel.sourceforge.net/)
Small (http://www.compuphase.com/small.htm)
Ruby (http://www.ruby-lang.org/en/)
Python (http://www.python.org/)
Tcl (http://tcl.sourceforge.net/)
Perl (http://www.perl.com/)
LibXML (http://xmlsoft.org/)
TinyXML (http://www.grinninglizard.com/tinyxml/)
Expat (http://expat.sourceforge.net/)

Apache XML libraries (http://xml.apache.org/)
ezXML (http://ezxml.sourceforge.net/)
Compression and File Systems:
zlib (http://www.zlib.net/)
bzip2 (http://www.bzip.org/)
LZMA SDK (http://www.7-zip.org/sdk.html)
LZO (http://www.oberhumer.com/opensource/lzo/)
PhysicsFS (http://icculus.org/physfs/)
#ZipLib (http://www.icsharpcode.net/OpenSource/SharpZipLib/) (native C# archive library)

SharpFS (http://www.gamedev.net/community ... sp?topic_id=391656) (native C# VFS)
unrarlib (http://www.unrarlib.org/)
ZipArchive (http://www.artpol-software.com/index_zip.html)

Note: Much of the above lists (but not all) were taken from this page (http://abattoir.wolfpaw.net/personal/gamelibs.php), maintained by Kylotan here (http://www.gamedev.net/community ... p?topic_id=291432).

Applications and Utilities:

Development Environments:
Code::Blocks IDE (http://www.codeblocks.org/)
Dev-C++ (http://www.bloodshed.net/devcpp.html)
Visual Studio 2005 betas (http://lab.msdn.microsoft.com/vs2005/)
SharpDevelop (http://www.icsharpcode.net/OpenSource/SD/)
ActivePython (http://www.activestate.com/Products/ActivePython/)
Graphics Development:
gDEBugger (http://www.gremedy.com/)
FX Composer (http://developer.nvidia.com/object/fx_composer_home.html)

RenderMonkey (http://www.ati.com/developer/rendermonkey/)
NVIDIA SDK (http://developer.nvidia.com/object/sdk_home.html)
ATI Radeon SDK (http://www.ati.com/developer/radeonSDK.html)
DirectX Developer Center (http://msdn.microsoft.com/directx/)
NVIDIA Cg (http://developer.nvidia.com/page/cg_main.html)
Sh (http://libsh.org/) (shader language)
Modeling:
SoftImage|XSI Mod Tool (http://www.softimage.com/community/xsi_mod_tool/default.aspx)

Milkshape 3D (http://www.swissquake.ch/chumbalum-soft/) (not free, $25)
Wings 3D (http://www.wings3d.com/)
Blender (http://www.blender3d.com/cms/Home.2.0.html)
Anim8or (http://www.anim8or.com/)
DeleD (http://www.delgine.com/index.php ... eled&menus=off)
World Building:
Quake Army Knife (QuArK) (http://dynamic.gamespy.com/%7Equark/)
GtkRadiant (http://www.qeradiant.com/)

Getic3D (http://www.zalsoft.com/index.php)
Audio:
ModPlug (http://www.modplug.com/)
Audacity (http://audacity.sourceforge.net/)
SynthEdit (http://www.synthedit.com/)
Buzz Machines (http://www.buzzmachines.com/)
KVR Audio (http://www.kvraudio.com/) (not actually software, but has tons of links)
Images:

Paint.NET (excellent free image editor)
(http://www.eecs.wsu.edu/paint.net/)The GIMP (http://www.gimp.org/)
BMFont (http://www.angelcode.com/products/bmfont/) (font texture generator)

Media
Textures:
Transmogrifying Textures Vol. I (NVIDIA) (http://developer.nvidia.com/object/IO_TTVol_01.html)
Mayang's Free Textures (http://www.mayang.com/textures/)

3D Cafe's Free Textures (http://www.3dcafe.com/asp/textures.asp)
Absolute Background Textures Archive (http://www.grsites.com/textures/)
Image * After (http://www.imageafter.com/)
Media Link's Free Backgrounds and Textures (http://my.erinet.com/%7Ecunning1/textures.htm)
Free Game Textures (http://www.vb3d.com/Textures.html)
Seamless Textures -- Absolute Cross (http://www.absolutecross.com/graphics/textures/)
3D Millenium Free Textures (http://www.m3corp.com/a/download/3d_textures/pages/index.htm)
Pixel Poke (http://www.pixelpoke.com/FREE%20PIXELS.html)
The Texture Studio (http://www.planethalflife.com/studio/)

Textures (http://astronomy.swin.edu.au/%7Epbourke/texture/)
Marlin Studios (http://www.marlinstudios.com/samples/samples.htm)
Free Foto (http://www.freefoto.com/browse.jsp?id=33-00-0)
Free Textures (http://www.noctua-graphics.de/english/freetex_e.htm)
Free Textures (http://www.indiaeye.com/free.htm)
MorgueFile (http://www.morguefile.com/)
Psionic3D (http://www.psionic3d.co.uk/)
Models:
3D Millenium Free Models (http://www.m3corp.com/a/download/models/index.htm)

Polycount (http://www.planetquake.com/polycount/)
Marlin Studios (http://www.marlinstudios.com/samples/samples.htm)
Free 3D Model Collections (http://www.amazing3d.com/modfree.html)
Free 3D Models (http://www.rocky3d.com/free3d.htm)
Digital Animators Models (http://www.digitalanimators.com/ ... s/3dModelsMain.htm)
Psionic3D (http://www.psionic3d.co.uk/)
Sprites
This thread has lots...and lots...and lots... (http://www.gamedev.net/community ... sp?topic_id=272386)
Sounds:

SFX and Music Resources (http://www.gamedev.net/community ... sp?topic_id=279518) (another long GDNet post)
Jamendo (http://www.jamendo.com/us/) (Creative Commons licensed music)
A1 Free Sound Effects (http://www.a1freesoundeffects.com/noflash.htm)
Sound Effects (http://www.pacdv.com/sounds/index.html)
Tintagel's Free Sound File Archive (http://freesoundfiles.tintagel.net/Audio/)
Free Sound Effects (http://www.partnersinrhyme.com/pir/PIRsfx.html)
Sound Hunter (http://www.soundhunter.com/)

Wav Sound Complete Archive (http://www.members.tripod.com/%7Ebuggerluggs/ie/wav-dir184.htm)
deusX Sound Design (http://www.deusx.com/studio.html)
FindSounds (http://www.findsounds.com/) (sound search engine)
Absolute Sound Effects Archive (http://www.grsites.com/sounds/)

飘零风雨亭: 3d编程

maxsdk图解

游戏中换装处理（备忘）

instancing批次画

WOW的地形shader祥细解释

地形上网格投射阴影(备忘)

地形pvs

地形碰撞高度计算(备忘)

d3d中实现简单的水

1. 最简单的一张带alpha的贴图直接uv动画2. 另一种是使用序列贴图, 即水的动画贴图,可在每个自定义大小的格子播放相同的一帧图片来实现.

3. 反射水面: 上述其中一种方法再加上带bump map效果的反射图

用d3d固定管道实现bump mapping振动/移动的水:

区域选择显示图

real time 3d scene post processing

设备获取

获取显存的方法

ogl绘制两圆(转)

常用链接(转)

1. 最简单的一张带alpha的贴图直接uv动画
2. 另一种是使用序列贴图, 即水的动画贴图,可在每个自定义大小的格子播放相同的一帧图片来实现.

3. 反射水面:
上述其中一种方法再加上带bump map效果的反射图