Yaker's ideal

2007年1月25日

从Google Reader上看到了Rss的必然趋势

因为对网络了解不多，其实rss是什么这个问题我到现在还是没搞清楚，但是我知道它很方便
像团队的blog聚合，非常方便
现在用google-reader了，把感兴趣的网站（有rss的）都放进来，隔一段时间reader会帮你去查看是否有更新，更新的站点会标记出来，再也不需要逐个网站去翻了。
更好的user experience,更快捷，方便。

2007年1月20日

I didn't do very well in the exam.I didn't finish both Physics and discrete mathematics exam.
I should pay more attention to my course next year. In short, i overestimated myself, i have thought that i could do very well even if i just spend little time on it.In fact, i didn't spend a lot of time when i am in grade one,but i still get a satisfactory score,after all a high score is not my ideal.The key is that i didn't take the course seriously.
But i don't feel very bad, on the contrary , i could see a passionate yaker, who will struggle for his future and ideal. Now i have a friend, this is the most exciting thing for me.I have a insight to myself now.During the exam, i read milan kundera's book <<The Unbearable lightness of Being>> , a external book.It tell me a lot about the meaning of the life.
I should make plan for the next year now.
Firstly, take my course seriously.
Secondly, commit to enhance my English ability(especially spoken english )and my programming ability.Seeking the meaning of my life.
Finally,spend more time on basketball, less time on WarCraft.
Do it.

2006年12月23日

Shadow Volume详解(3)

事隔几个月之后，终于有时间把代码贴上来了。代码下载
今天在搞Hardware shadow map，有一点小进展，但是还是有不少距离。期待写shadow map系列

1.先说一下3ds文件的结构吧，只说跟投影有关的东西，就是它是包含多个object的，读进来的模型也是包含多个object的。
// 模型信息结构体

struct t3DModel
{
UINT    texture[MAX_TEXTURES];
    int     numOfObjects;               // 模型中对象的数目
    int     numOfMaterials;             // 模型中材质的数目
std::vector <tMaterialInfo> pMaterials;// 材质链表信息
std::vector <t3DObject> pObject;       // 模型中对象链表信息
};

2.首先执行阴影的初始化工作

//载入模型
g_model.Import3DS(fileName);  
// 初始化阴影
g_modelShadow.Init(g_model.GetModel(), g_lightPos);

以下是ModelShadow::Init的定义


bool ModelShadow::Init(const t3DModel &model,const float *lightPos)
{
// 获取object的数目
objNum_         =   model.numOfObjects;
objInfoTable_.resize(objNum_);
for(int objIndex = 0 ; objIndex < objNum_ ; objIndex++ )
{
....
}
glClearStencil(0);
....
SetConnectivity(); // 设置三角形之间的链接关系
CalcPlane();         // 计算每个三角形的平面方程
....
}

3.初始化之后就可以渲染了

// 绘制模型，阴影
// 重设光源位置，如果光源位置不便可以忽略
g_modelShadow.ResetLightPos(g_lightPos);    
// 这里获取当前的model view matrix(MVM)
g_modelShadow.GetCurMatrix();                    
glPushMatrix();
// 放置物体
glTranslatef(g_modelPos[0], g_modelPos[1], g_modelPos[2]);       
// 这里又获取了一次model view matrix
g_model.Draw();                                       
g_modelShadow.Draw();
glPopMatrix();

通过两次获得的矩阵的差异就可以算出中间做了那些变换，比如(glTranslatef(g_modelPos[0], g_modelPos[1], g_modelPos[2]); )
现在看一下ModelShadow::Draw的实现

void ModelShadow::Draw() const
{
glClear(GL_STENCIL_BUFFER_BIT);

GLfloat objMatrix[16];
glGetFloatv(GL_MODELVIEW_MATRIX, objMatrix);

// 求逆矩阵，这四条语句就是在求物体的局部坐标系里灯的坐标
InverseMatrix16(objMatrix);  
MultiplyMatrix16(objMatrix, curMatrix_);
memcpy(lightPos_, orignLightPos_, 4 * sizeof(float));
Mtx16MultVect4(objMatrix, lightPos_);

//Cast the shadow
CastShadow();
}

再看ModelShadow::castShadow


void ModelShadow::CastShadow() const
{
SetVisibility();  // 设置每个三角形的可见性
....
// 第一趟渲染，通过模板测试的话模版值增一
glFrontFace(GL_CCW);
glStencilOp(GL_KEEP, GL_KEEP, GL_INCR);
DoShadowPass();

// 第二趟渲染，通过模板测试的话模版值减一
glFrontFace(GL_CW);
glStencilOp(GL_KEEP, GL_KEEP, GL_DECR);
DoShadowPass();
....
glStencilFunc(GL_NOTEQUAL, 0, 1);
glStencilOp(GL_KEEP, GL_KEEP, GL_KEEP);

//在全屏幕混合一张图
glEnable(GL_BLEND);
glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);
glColor4f(0, 0, 0, 0.3f);
....
}

有一些问题，我在（2）里说错了的

glLoadIdentity();
glBegin(GL_TRIANGLE_STRIP);
glVertex3f(-0.1f,  0.1f, -0.10f);
glVertex3f(-0.1f, -0.1f, -0.10f);
glVertex3f( 0.1f,  0.1f, -0.10f);
glVertex3f( 0.1f, -0.1f, -0.10f);
glEnd();

这个覆盖全屏真是投机取巧
gluPerspective(45.0, (GLfloat)width/(GLfloat)height, 0.1, 100.0);
如果那个角度比45.0f大就无法覆盖了
正确的做法是改用正投影。