Quadtree/terrain frustum culling help

1,953

Author

January 11, 2014 02:09 PM

I tried to implement frustum culling of terrain by my own but unfortunately, I failed.

I also searched the forums but didn't found any good tutorials on frustum culling on terrain, so can anyone help me out with this (by just linking/giving a good tutorial on terrain frustum cull.)

ericrrichards22

2,441

January 11, 2014 06:09 PM

There are a few different ways you could do this.

If you are using DX11, you can do frustum culling in the hull shader. I've got a post on how I implemented the terrain example from Luna's DX11 book in SlimDX (the relevant parts are HLSL, so that shouldn't be a problem if you're using C++ rather than C#).
If you are using DX10, I would imagine you could do something similar using the geometry shader, although I don't have any examples off the shelf to point you to.
You could also do the culling on the CPU, by dividing your terrain mesh up into patches, and culling based on the bounding box of the patch. O have another article using this technique, although I imagine there are a fair number of other resources, since this is a fairly established way of doing terrain culling.

Eric Richards

SlimDX tutorials - http://www.richardssoftware.net/

Twitter - @EricRichards22

215648

1,953

Author

January 12, 2014 11:02 AM

I'm using third method, here is my code, I don't know whats going wrong.

I get vector subscript out of range error.

using namespace std;
 
class Terrain
{
private:
struct SubGrid
{
SubGrid()
{
VertexBuffer = 0;
IndexBuffer = 0;
}
 
GeometryGenerator::MeshData mesh;
XNA::AxisAlignedBox box;
 
ID3D11Buffer* VertexBuffer;
ID3D11Buffer* IndexBuffer;
 
bool operator < (const SubGrid& rhs)const;
 
const static int NUM_ROWS  = 33;
const static int NUM_COLS  = 33;
const static int NUM_TRIS  = (NUM_ROWS-1)*(NUM_COLS-1)*2;
const static int NUM_VERTS = NUM_ROWS*NUM_COLS;
};
public:
struct InitInfo
{
wstring HeightmapFileName,
LayerMapFileName0,
LayerMapFileName1,
LayerMapFileName2,
LayerMapFileName3,
LayerMapFileName4,
BlendMapFileName;
 
float HeightScale;
float HeightOffset;
UINT NumRows;
UINT NumCols;
float CellSpacing;
};
 
public:
Terrain();
~Terrain();
 
void Init(const InitInfo& info);
 
float width()const;
float depth()const;
 
float getHeight(float x, float y)const;
void SetDirectionToSun(const XMFLOAT3& v);
void draw(CXMMATRIX world, Camera& cam);
 
UINT GetNumVertices()const { return mNumVertices; }
private:
void BuildFX();
void loadHeightMap();
void Smooth();
bool inBounds(UINT i, UINT j);
float average(UINT i, UINT j);
void BuildGeometry();
void BuildSubGrid(RECT &R, Vertex::Basic32* gridVerts);
private:
InitInfo mInfo;
UINT mNumVertices;
UINT mNumFaces;
 
vector<float> mHeightmap;
vector<SubGrid> mSubGrids;
 
ID3DX11Effect* mFX;
 
 
ID3D11Buffer* mVB;
ID3D11Buffer* mIB;
 
ID3D11ShaderResourceView* mLayer0;
ID3D11ShaderResourceView* mLayer1;
ID3D11ShaderResourceView* mLayer2;
ID3D11ShaderResourceView* mLayer3;
ID3D11ShaderResourceView* mLayer4;
ID3D11ShaderResourceView* mBlendMap;
 
ID3DX11EffectTechnique* mTech;
ID3DX11EffectMatrixVariable* mfxWVPVar;
ID3DX11EffectMatrixVariable* mfxWorldVar;
ID3DX11EffectVectorVariable* mfxDirToSunVar;
ID3DX11EffectShaderResourceVariable* mfxLayer0Var;
ID3DX11EffectShaderResourceVariable* mfxLayer1Var;
ID3DX11EffectShaderResourceVariable* mfxLayer2Var;
ID3DX11EffectShaderResourceVariable* mfxLayer3Var;
ID3DX11EffectShaderResourceVariable* mfxLayer4Var;
ID3DX11EffectShaderResourceVariable* mfxBlendMapVar;
 
};
//====================================================================================================================================
#include "stdafx.h"
#include <list>
 
Terrain::Terrain()
{
 
mVB = nullptr;
mIB = nullptr;
 
 
mLayer0 = nullptr;
mLayer1 = nullptr;
mLayer2 = nullptr;
mLayer3 = nullptr;
mLayer4 = nullptr;
mBlendMap = nullptr;
}
 
Terrain::~Terrain()
{
ReleaseCOM(mVB);
ReleaseCOM(mIB);
ReleaseCOM(mFX);
 
 
ReleaseCOM(mBlendMap);
ReleaseCOM(mLayer0);
ReleaseCOM(mLayer1);
ReleaseCOM(mLayer2);
ReleaseCOM(mLayer3);
ReleaseCOM(mLayer4);
 
for (UINT i = 0; i < mSubGrids.size(); ++i)
{
ReleaseCOM(mSubGrids[i].VertexBuffer);
ReleaseCOM(mSubGrids[i].IndexBuffer);
}
}
 
float Terrain::width()const
{
return (mInfo.NumCols-1)*mInfo.CellSpacing;
}
 
float Terrain::depth()const
{
return (mInfo.NumRows-1)*mInfo.CellSpacing;
}
 
float Terrain::getHeight(float x, float z)const
{
// Transform from terrain local space to "cell" space.
float c = (x + 0.5f*width()) /  mInfo.CellSpacing;
float d = (z - 0.5f*depth()) / -mInfo.CellSpacing;
 
// Get the row and column we are in.
int row = (int)floorf(d);
int col = (int)floorf(c);
 
// Grab the heights of the cell we are in.
// A*--*B
//  | /|
//  |/ |
// C*--*D
float A = mHeightmap[row*mInfo.NumCols + col];
float B = mHeightmap[row*mInfo.NumCols + col + 1];
float C = mHeightmap[(row+1)*mInfo.NumCols + col];
float D = mHeightmap[(row+1)*mInfo.NumCols + col + 1];
 
// Where we are relative to the cell.
float s = c - (float)col;
float t = d - (float)row;
 
// If upper triangle ABC.
if( s + t <= 1.0f)
{
float uy = B - A;
float vy = C - A;
return A + s*uy + t*vy;
}
else // lower triangle DCB.
{
float uy = C - D;
float vy = B - D;
return D + (1.0f-s)*uy + (1.0f-t)*vy;
}
}
 
void Terrain::BuildFX()
{
std::ifstream fin("Resources\\Shaders\\Terrain.fxo", std::ios::binary);
 
fin.seekg(0, std::ios_base::end);
int size = (int)fin.tellg();
fin.seekg(0, std::ios_base::beg);
std::vector<char> compiledShader(size);
 
fin.read(&compiledShader[0], size);
fin.close();
 
HR(D3DX11CreateEffectFromMemory(&compiledShader[0], size, 
0, pDevice, &mFX));
}
 
void Terrain::Init(const InitInfo& initInfo)
{
BuildFX();
 
mTech          = mFX->GetTechniqueByName("TerrainTech");
mfxWVPVar      = mFX->GetVariableByName("gWVP")->AsMatrix();
mfxWorldVar    = mFX->GetVariableByName("gWorld")->AsMatrix();
mfxDirToSunVar = mFX->GetVariableByName("gDirToSunW")->AsVector();
mfxLayer0Var   = mFX->GetVariableByName("gLayer0")->AsShaderResource();
mfxLayer1Var   = mFX->GetVariableByName("gLayer1")->AsShaderResource();
mfxLayer2Var   = mFX->GetVariableByName("gLayer2")->AsShaderResource();
mfxLayer3Var   = mFX->GetVariableByName("gLayer3")->AsShaderResource();
mfxLayer4Var   = mFX->GetVariableByName("gLayer4")->AsShaderResource();
mfxBlendMapVar = mFX->GetVariableByName("gBlendMap")->AsShaderResource();
 
mInfo = initInfo;
 
mNumVertices = mInfo.NumRows*mInfo.NumCols;
mNumFaces    = (mInfo.NumRows-1)*(mInfo.NumCols-1)*2;
 
loadHeightMap();
Smooth();
 
BuildGeometry();
 
HR(D3DX11CreateShaderResourceViewFromFile(pDevice, initInfo.LayerMapFileName0.c_str(), 0, 0, &mLayer0, 0));
HR(D3DX11CreateShaderResourceViewFromFile(pDevice, initInfo.LayerMapFileName1.c_str(), 0, 0, &mLayer1, 0));
HR(D3DX11CreateShaderResourceViewFromFile(pDevice, initInfo.LayerMapFileName2.c_str(), 0, 0, &mLayer2, 0));
HR(D3DX11CreateShaderResourceViewFromFile(pDevice, initInfo.LayerMapFileName3.c_str(), 0, 0, &mLayer3, 0));
HR(D3DX11CreateShaderResourceViewFromFile(pDevice, initInfo.LayerMapFileName4.c_str(), 0, 0, &mLayer4, 0));
HR(D3DX11CreateShaderResourceViewFromFile(pDevice, initInfo.BlendMapFileName.c_str(), 0, 0, &mBlendMap, 0));
}
 
void Terrain::SetDirectionToSun(const XMFLOAT3& v)
{
XMVECTOR temp = XMVectorSet(v.x, v.y, v.z, 0.0f);
mfxDirToSunVar->SetFloatVector((float*)&temp);
}
 
void Terrain::draw(CXMMATRIX world, Camera& cam)
{
pDeviceContext->IASetInputLayout(InputLayouts::Basic32);
 
UINT stride = sizeof(Vertex::Basic32);
    UINT offset = 0;
 
 
std::list<SubGrid> visibleSubGrids;
for(UINT i = 0; i < mSubGrids.size(); ++i)
{
if( XNA::IntersectAxisAlignedBoxFrustum(&mSubGrids[i].box, &d3d->GetFrustum()))
visibleSubGrids.push_back(mSubGrids[i]);
}
 
// Sort front-to-back from camera.
visibleSubGrids.sort();
 
 
 
for(std::list<SubGrid>::iterator iter = visibleSubGrids.begin(); iter != visibleSubGrids.end(); ++iter)
{
pDeviceContext->IASetVertexBuffers(0, 1, &iter->VertexBuffer, &stride, &offset);
pDeviceContext->IASetIndexBuffer(iter->IndexBuffer, DXGI_FORMAT_R32_UINT, 0);
 
}
 
 
XMMATRIX view = cam.View();
XMMATRIX proj = cam.Proj();
 
XMMATRIX WVP = world * view * proj;
 
 
mfxWVPVar->SetMatrix((float*)&WVP);
mfxWorldVar->SetMatrix((float*)&world);
 
mfxLayer0Var->SetResource(mLayer0);
mfxLayer1Var->SetResource(mLayer1);
mfxLayer2Var->SetResource(mLayer2);
mfxLayer3Var->SetResource(mLayer3);
mfxLayer4Var->SetResource(mLayer4);
mfxBlendMapVar->SetResource(mBlendMap);
 
    D3DX11_TECHNIQUE_DESC techDesc;
    mTech->GetDesc(&techDesc);
 
    for(UINT i = 0; i < techDesc.Passes; ++i)
    {
        ID3DX11EffectPass* pass = mTech->GetPassByIndex(i);
pass->Apply(0, pDeviceContext);
 
for(std::list<SubGrid>::iterator iter = visibleSubGrids.begin(); iter != visibleSubGrids.end(); ++iter)
pDeviceContext->DrawIndexed(iter->mesh.Indices.size(), 0, 0);
} 
}
 
void Terrain::loadHeightMap()
{
// A height for each vertex
std::vector<unsigned char> in( mInfo.NumRows * mInfo.NumCols );
 
// Open the file.
std::ifstream inFile;
inFile.open(mInfo.HeightmapFileName.c_str(), std::ios_base::binary);
 
if(inFile)
{
// Read the RAW bytes.
inFile.read((char*)&in[0], (std::streamsize)in.size());
 
// Done with file.
inFile.close();
}
 
// Copy the array data into a float array, and scale and offset the heights.
mHeightmap.resize(mInfo.NumRows * mInfo.NumCols, 0);
for(UINT i = 0; i < mInfo.NumRows * mInfo.NumCols; ++i)
{
mHeightmap[i] = (float)in[i] * mInfo.HeightScale + mInfo.HeightOffset;
}
}
 
 
void Terrain::Smooth()
{
std::vector<float> dest( mHeightmap.size() );
 
for(UINT i = 0; i < mInfo.NumRows; ++i)
{
for(UINT j = 0; j < mInfo.NumCols; ++j)
{
dest[i*mInfo.NumCols+j] = average(i,j);
}
}
 
// Replace the old heightmap with the filtered one.
mHeightmap = dest;
}
 
bool Terrain::SubGrid::operator<(const SubGrid& rhs) const
{
 
XMFLOAT3 camerapos = d3d->GetCameraPosition();
 
XMVECTOR d1 = XMLoadFloat3(&box.Center) - XMLoadFloat3(&camerapos);
XMVECTOR d2 = XMLoadFloat3(&rhs.box.Center) - XMLoadFloat3(&camerapos);
 
XMVECTOR v1, v2;
    v1 = XMVector3LengthSq(d1);
v2 = XMVector3LengthSq(d2);
 
XMFLOAT2 vv1, vv2;
 
    XMStoreFloat2(&vv1, v1);
XMStoreFloat2(&vv2, v2);
 
FLOAT Scalar1, Scalar2;
 
Scalar1 = vv1.x;
Scalar2 = vv2.y;
 
return Scalar1 < Scalar2;
}
 
bool Terrain::inBounds(UINT i, UINT j)
{
// True if ij are valid indices; false otherwise.
return 
i >= 0 && i < mInfo.NumRows && 
j >= 0 && j < mInfo.NumCols;
}
 
float Terrain::average(UINT i, UINT j)
{
// Function computes the average height of the ij element.
// It averages itself with its eight neighbor pixels.  Note
// that if a pixel is missing neighbor, we just don't include it
// in the average--that is, edge pixels don't have a neighbor pixel.
//
// ----------
// | 1| 2| 3|
// ----------
// |4 |ij| 6|
// ----------
// | 7| 8| 9|
// ----------
 
float avg = 0.0f;
float num = 0.0f;
 
for(UINT m = i-1; m <= i+1; ++m)
{
for(UINT n = j-1; n <= j+1; ++n)
{
if( inBounds(m,n) )
{
avg += mHeightmap[m*mInfo.NumCols + n];
num += 1.0f;
}
}
}
 
return avg / num;
}
 
 
 
void Terrain::BuildGeometry()
{
//=========================================
//               Build Vertices
//=========================================
 
std::vector<Vertex::Basic32> vertices(mNumVertices);
 
float halfWidth = (mInfo.NumCols-1)*mInfo.CellSpacing*0.5f;
float halfDepth = (mInfo.NumRows-1)*mInfo.CellSpacing*0.5f;
 
float du = 1.0f / (mInfo.NumCols-1);
float dv = 1.0f / (mInfo.NumRows-1);
for(UINT i = 0; i < mInfo.NumRows; ++i)
{
float z = halfDepth - i*mInfo.CellSpacing;
for(UINT j = 0; j < mInfo.NumCols; ++j)
{
float x = -halfWidth + j*mInfo.CellSpacing;
 
float y = mHeightmap[i*mInfo.NumCols+j];
vertices[i*mInfo.NumCols+j].Pos    = XMFLOAT3(x, y, z);
vertices[i*mInfo.NumCols+j].Normal = XMFLOAT3(0.0f, 1.0f, 0.0f);
 
// Stretch texture over grid.
vertices[i*mInfo.NumCols+j].Tex.x = j*du;
vertices[i*mInfo.NumCols+j].Tex.y = i*dv;
}
}
 
// Estimate normals for interior nodes using central difference.
float invTwoDX = 1.0f / (2.0f*mInfo.CellSpacing);
float invTwoDZ = 1.0f / (2.0f*mInfo.CellSpacing);
for(UINT i = 2; i < mInfo.NumRows-1; ++i)
{
for(UINT j = 2; j < mInfo.NumCols-1; ++j)
{
float t = mHeightmap[(i-1)*mInfo.NumCols + j];
float b = mHeightmap[(i+1)*mInfo.NumCols + j];
float l = mHeightmap[i*mInfo.NumCols + j - 1];
float r = mHeightmap[i*mInfo.NumCols + j + 1];
 
XMFLOAT3 tanZ(0.0f, (t-b)*invTwoDZ, 1.0f);
XMFLOAT3 tanX(1.0f, (r-l)*invTwoDX, 0.0f);
 
XMVECTOR N;
N = XMVector3Cross(XMLoadFloat3(&tanZ), XMLoadFloat3(&tanX));
N = XMVector3Normalize(N);
 
XMStoreFloat3(&vertices[i * mInfo.NumCols + j].Normal, N);
 
}
}
 
//=======================================================
//                     Build Indices
//=======================================================
 
std::vector<DWORD> indices(mNumFaces*3); // 3 indices per face
 
// Iterate over each quad and compute indices.
int k = 0;
for(UINT i = 0; i < mInfo.NumRows-1; ++i)
{
for(UINT j = 0; j < mInfo.NumCols-1; ++j)
{
indices[k]   = i*mInfo.NumCols+j;
indices[k+1] = i*mInfo.NumCols+j+1;
indices[k+2] = (i+1)*mInfo.NumCols+j;
 
indices[k+3] = (i+1)*mInfo.NumCols+j;
indices[k+4] = i*mInfo.NumCols+j+1;
indices[k+5] = (i+1)*mInfo.NumCols+j+1;
 
k += 6; // next quad
}
}
 
 
//===============================================================
// Now break the grid up into subgrid meshes.
 
// Find out the number of subgrids we'll have.  For example, if
// m = 513, n = 257, SUBGRID_VERT_ROWS = SUBGRID_VERT_COLS = 33,
// then subGridRows = 512/32 = 16 and sibGridCols = 256/32 = 8.
int subGridRows = (mInfo.NumRows-1) / (SubGrid::NUM_ROWS-1);
int subGridCols = (mInfo.NumCols-1) / (SubGrid::NUM_COLS-1);
 
for(int r = 0; r < subGridRows; ++r)
{
for(int c = 0; c < subGridCols; ++c)
{
// Rectangle that indicates (via matrix indices ij) the
// portion of grid vertices to use for this subgrid.
RECT R = 
{
c * (SubGrid::NUM_COLS - 1),
r * (SubGrid::NUM_ROWS - 1),
(c+1) * (SubGrid::NUM_COLS - 1),
(r+1) * (SubGrid::NUM_ROWS - 1)
};
 
BuildSubGrid(R, &vertices[0]); 
}
}
 
    D3D11_BUFFER_DESC vbd;
    vbd.Usage = D3D11_USAGE_IMMUTABLE;
vbd.ByteWidth = sizeof(Vertex::Basic32) * mNumVertices;
    vbd.BindFlags = D3D11_BIND_VERTEX_BUFFER;
    vbd.CPUAccessFlags = 0;
    vbd.MiscFlags = 0;
 
    
for (int i = 0; i < mSubGrids.size(); i++)
{
   D3D11_SUBRESOURCE_DATA vinitData;
vinitData.pSysMem = mSubGrids[i].VertexBuffer;  
 
HR(pDevice->CreateBuffer(&vbd, &vinitData, &mSubGrids[i].VertexBuffer));
}
 
D3D11_BUFFER_DESC ibd;
    ibd.Usage = D3D11_USAGE_IMMUTABLE;
    ibd.ByteWidth = sizeof(DWORD) * mNumFaces * 3;
    ibd.BindFlags = D3D11_BIND_INDEX_BUFFER;
    ibd.CPUAccessFlags = 0;
    ibd.MiscFlags = 0;
 
    
for (int i = 0; i < mSubGrids.size(); ++i)
{    
D3D11_SUBRESOURCE_DATA iinitData;
iinitData.pSysMem = mSubGrids[i].IndexBuffer;
 
HR(pDevice->CreateBuffer(&ibd, &iinitData, &mSubGrids[i].IndexBuffer)); 
}
}
 
 
void Terrain::BuildSubGrid(RECT &R, Vertex::Basic32* gridVerts)
{
 
 
GeometryGenerator::MeshData temp;
GeometryGenerator gen;
 
gen.CreateGrid(SubGrid::NUM_ROWS, SubGrid::NUM_COLS,
SubGrid::NUM_ROWS * mInfo.CellSpacing, SubGrid::NUM_COLS * mInfo.CellSpacing, temp);
 
GeometryGenerator::MeshData subMesh;
 
int k = 0;
for(int i = R.top; i <= R.bottom; ++i)
{
for(int j = R.left; j <= R.right; ++j)
{
 
subMesh.Vertices[k++].Position = gridVerts[i*mInfo.NumCols+j].Pos;
}
}
 
XNA::AxisAlignedBox box;
XNA::ComputeBoundingAxisAlignedBoxFromPoints(&box, subMesh.Vertices.size(), &subMesh.Vertices[0].Position, sizeof(Vertex::Basic32));
  
 
for (int i = 0; i < SubGrid::NUM_TRIS; i++)
{
subMesh.Indices[i*3+0] = temp.Indices[i*3+0];
subMesh.Indices[i*3+1] = temp.Indices[i*3+1];
subMesh.Indices[i*3+2] = temp.Indices[i*3+2];
}
 
SubGrid g;
 
g.mesh = subMesh;
g.box = box;
 
mSubGrids.push_back(g);
}

greenpig83

330

January 12, 2014 06:07 PM

This is the tut i used for quadtree : http://rastertek.com/tertut05.html
And this is for frustum culling : http://www.flipcode.com/archives/Frustum_Culling.shtml
Here also : http://www.chadvernon.com/blog/resources/directx9/frustum-culling/
I think combine this 2, can help u !

215648

1,953

Author

January 13, 2014 02:28 PM

I can't because I use XNAMath.

Anyways, I changed the code but still not getting it to work. (getting vector subscript out of range error.)

http://pastebin.com/kSibZFEf

mister345

1

November 06, 2017 01:16 PM

Hi, there's another great tutorial on frustum culling, but impossible to compile because it uses old DirectX 11 types (Direct3DXPlane instead of XMVECTOR, etc). Can someone please help me update this one class - frustumClass - to the new DirectX11 types (XMMATRIX, XMVECTOR, etc)?

http://www.rastertek.com/dx11tut16.html

Furthermore, can anyone please explain how he gets the minimum Z distance from the projection matrix by dividing one element by another? He leaves no explanation for this math and it's extremely frustrating.

// Calculate the minimum Z distance in the frustum.
zMinimum = -projectionMatrix._43 / projectionMatrix._33;
r = screenDepth / (screenDepth - zMinimum);
projectionMatrix._33 = r;
projectionMatrix._43 = -r * zMinimum;

Also not sure how to use an XMVECTOR instead of the old Plane class that he uses. Confused as to where all the m12, m13, etc correspond to the elements of an XMVECTOR. I thought you're not supposed to even access the elements of an XMVECTOR directly! So incredibly frustrating.

Please help, thanks.

_WeirdCat_

1,512

November 07, 2017 07:45 AM

Heres quadtree pvs system


#ifndef ocean_pvs_systemH
#define ocean_pvs_systemH

#include "DxcMath.h"
#include "DxcMatrix.h"
#include "frustum.h"
#include "glwrapper.h"
#include "shaderclass.h"


struct TOceanPVSCell
{
Pushort * indices; //each row corresponds to each row in the ocean
int len; //both dimensions of an array
int ilen;
unsigned int * elementbuff;
bool add2;
int x;
int z;
//points of grid (adding translation will give correct worldspace pos)
vec3 A;
vec3 B;
vec3 C;
vec3 D;
TOceanPVSCell()
{
	add2		= false;
	indices 	= NULL;
	elementbuff = NULL;
	len		= 0;
}

~TOceanPVSCell()
{
if (indices != NULL)
{
	delete [] indices;

	for (int i=0; i < len; i++)
		glDeleteBuffers(1, &elementbuff[i]);

	delete [] elementbuff;
}
}

};

struct TOceanPVS
{


float pvs_cell_size;

TFrustum EYE;

TOceanPVSCell * cells;
int sectors;

int sim_grid_size;

int * vis[9]; //each direction, that stores visible pvs cells, because i first determine which i see then i draw them
int vislen[9];
bool visvis[9];

float ocean_grid_size;
int cells_in_sector;

vec3 A;
vec3 B;
vec3 C;
vec3 D;
TOceanPVS()
{

	cells 	 = NULL;
}

~TOceanPVS()
{
if (cells != NULL)
{
delete [] cells;

for (int i=0; i < 9; i++) delete [] vis[i];
}
}

bool TGridVisible(int translation_index, int X, int Z)
{
    float DirX[9] = {0.0, -1.0, 1.0,  0.0, 0.0, -1.0, 1.0, -1.0,  1.0};
    float DirY[9] = {0.0,  0.0, 0.0, -1.0, 1.0,  1.0, 1.0, -1.0, -1.0};

    vec3 t = vec3(ocean_grid_size*float(X), 0.0, ocean_grid_size*float(Z));
    vec3 v[4];
    v[0] = A + vec3(ocean_grid_size*DirX[translation_index], 0.0, ocean_grid_size*DirY[translation_index]) + t;
    v[1] = B + vec3(ocean_grid_size*DirX[translation_index], 0.0, ocean_grid_size*DirY[translation_index]) + t;
    v[2] = C + vec3(ocean_grid_size*DirX[translation_index], 0.0, ocean_grid_size*DirY[translation_index]) + t;
    v[3] = D + vec3(ocean_grid_size*DirX[translation_index], 0.0, ocean_grid_size*DirY[translation_index]) + t;

return EYE.BoxIsVisible(v, 4);
}



bool PVSCellVisible(int index, int translation_index, int X, int Z)
{
	  	  float DirX[9] = {0.0, -1.0, 1.0,  0.0, 0.0, -1.0, 1.0, -1.0,  1.0};
	  	  float DirY[9] = {0.0,  0.0, 0.0, -1.0, 1.0,  1.0, 1.0, -1.0, -1.0};
	  	  vec3 t = vec3(ocean_grid_size*float(X), 0.0, ocean_grid_size*float(Z));
	  	  vec3 v[4];

	      v[0] = cells[index].A + vec3(ocean_grid_size*DirX[translation_index], 0.0, ocean_grid_size*DirY[translation_index]) + t;
	      v[1] = cells[index].B + vec3(ocean_grid_size*DirX[translation_index], 0.0, ocean_grid_size*DirY[translation_index]) + t;
	      v[2] = cells[index].C + vec3(ocean_grid_size*DirX[translation_index], 0.0, ocean_grid_size*DirY[translation_index]) + t;
	      v[3] = cells[index].D + vec3(ocean_grid_size*DirX[translation_index], 0.0, ocean_grid_size*DirY[translation_index]) + t;


return EYE.BoxIsVisible(v, 4);
}


void CalculatePVS(int X, int Z, TSpecCamera * FPP_CAM, mat4 CAM_VIEW, int sw, int sh, float fovy, float z_near, float z_far, float aspect)
{

EYE.CalcFrustum(FPP_CAM,CAM_VIEW, sw, sh, fovy, z_near, z_far, aspect);
vec3 v[4];
v[0] = vec3(0.0, 0.0, 0.0);
v[1] = vec3(400.0, 0.0, 0.0);
v[2] = vec3(400.0, 0.0, 400.0);
v[3] = vec3(0.0, 0.0, 400.0);
int h = 0;
//ALOG("NEW PVS FRAME");
for (int i=0; i < 9; i++)
{
	visvis[i] = TGridVisible(i, X, Z);
//	if (visvis) h = h + 1;
	vislen[i] = -1;
}


for (int i=0; i < 9; i++)
if (visvis[i])
{

for (int n=0; n < sectors*sectors; n++)
if (PVSCellVisible(n, i, X, Z))
{
	vislen[i] = vislen[i] + 1;
	vis[i][vislen[i]] = n;
}

}

for (int i=0; i < 9; i++) vislen[i] = vislen[i] + 1;

//for (int i=0; i < 9; i++)
//ALOG("VIS: "+IntToStr(i) + "    "+IntToStr(vislen)+ " pvs cells visible");

}


void DrawOceanPVS(TShaderObject * shader, int X, int Z)
{
	  float DirX[9] = {0.0, -1.0, 1.0,  0.0, 0.0, -1.0, 1.0, -1.0,  1.0};
	  float DirY[9] = {0.0,  0.0, 0.0, -1.0, 1.0,  1.0, 1.0, -1.0, -1.0};

	for (int i=0; i < 9; i++)
		if (visvis[i])
		{
			vec3 t = vec3(DirX[i]*float(ocean_grid_size), 0.0,  DirY[i]*float(ocean_grid_size)) +
					 vec3(float(X)*ocean_grid_size, 0.0, float(Z)*ocean_grid_size);
mat4 wrld;
wrld.LoadIdentity();
wrld.TranslateP(t);
shader->SendWorldMatrixtoShader(wrld);
mat4 MVP = (wrld * ACTUAL_VIEW) * ACTUAL_PROJECTION;
shader->SendMVPtoShader(MVP);
			for (int n=0; n < vislen[i]; n++)
			{
				int celli = vis[i][n];

			int len = cells[celli].len * 2;
			if (cells[celli].add2) len = len + 2;


for (int k=0; k < cells[celli].len; k++)
{
glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, cells[celli].elementbuff[k]);



	 glDrawElements(GL_TRIANGLE_STRIP, GLsizei(len), GL_UNSIGNED_SHORT,
			 (void*)( 0 ) );
}

		}
}
	  glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, 0);
}


int ArrI(int x, int y)
{
return y*sim_grid_size + x;
}

int pvsI(int x, int y)
{
return y*sectors + x;
}



void Initialize(int sections, float oc_grid_size, int sim_grd_size)
{
	CAN_LOG = false;
	sim_grid_size = sim_grd_size;
	ocean_grid_size = oc_grid_size;
//	OCEAN_TEX_GRID_SIZE
	for (int i=0; i < 9; i++)
	{
		vislen[i] 	= 0;
		visvis[i]   = false;
		vis[i] 		= new int[sections*sections];
	}

sectors = sections;
cells_in_sector = sim_grid_size / sections; //64 / 8 = 8
cells = new TOceanPVSCell[sections * sections]; //8 * 8 = 64

B.x = oc_grid_size;

C.x = oc_grid_size;
C.z = oc_grid_size;

D.z = oc_grid_size;

ALOG("1");


pvs_cell_size = float(oc_grid_size) / sectors;
for (int y = 0; y < sim_grid_size; y++)
for (int x = 0; x < sim_grid_size; x++)
{

int sectionX = x / cells_in_sector;
int sectionY = y / cells_in_sector;

int row = y % cells_in_sector;
int col = x % cells_in_sector;
int pvsi = pvsI(sectionX, sectionY);

cells[pvsi].x = sectionX;
cells[pvsi].z = sectionY;
}
//ALOG("2");
for (int y = 0; y < sections; y++)
for (int x = 0; x < sections; x++)
{



float xper = float(x) / float(sections);
float yper = float(y) / float(sections);

float x2per = float(x+1) / float(sections);
float y2per = float(y+1) / float(sections);


int pvsi = pvsI(x, y);

//ALOG("VXX: "+IntToStr(cells[pvsi].x)+ "  y:  "+IntToStr(cells[pvsi].z));

cells[pvsi].A.x = xper * ocean_grid_size;
cells[pvsi].A.z = yper * ocean_grid_size;

cells[pvsi].B.x = x2per * ocean_grid_size;
cells[pvsi].B.z = yper  * ocean_grid_size;

cells[pvsi].C.x = x2per * ocean_grid_size;
cells[pvsi].C.z = y2per * ocean_grid_size;

cells[pvsi].D.x = xper  * ocean_grid_size;
cells[pvsi].D.z = y2per * ocean_grid_size;

}

//ALOG("3");
for (int i=0; i < sections*sections; i++)
{
	int alen = cells_in_sector;
//	if (cells.z == sections-1) alen = alen - 1;
cells[i].len = alen;
cells[i].indices 	 = new Pushort[cells[i].len];
cells[i].elementbuff = new unsigned int[cells[i].len];

bool add2 = true;
if (cells[i].x == sections-1) add2 = false;
int k = 2;
if (!add2) k = 0;

cells[i].ilen = cells_in_sector*2+k;

for (int n=0; n < cells[i].len; n++)
{
	cells[i].indices[n] = new unsigned short[cells[i].ilen];
	cells[i].add2 = add2;
}

}


for (int n=0; n < sectors*sectors; n++)
	for (int i=0; i < cells[n].len; i++)
		for (int j=0; j < cells[n].ilen; j++)
		cells[n].indices[i][j] = 0;

//ALOG("4");

for (int y = 0; y < sim_grid_size-1; y++)
for (int x = 0; x < sim_grid_size; x++)
{

int sectionX = x / cells_in_sector;
int sectionY = y / cells_in_sector;

int row = y % cells_in_sector;
int col = x % cells_in_sector;
int pvsi = pvsI(sectionX, sectionY);

cells[pvsi].x = sectionX;
cells[pvsi].z = sectionY;

cells[pvsi].indices[row][col*2] 	= (unsigned short)ArrI(x,y+1);
cells[pvsi].indices[row][col*2+1] 	= (unsigned short)ArrI(x,y);

//ArrI(x,y);
//return y*sim_grid_size + x;
//ALOG("CXX x: "+IntToStr(x)+ "  y: "+IntToStr(y));

//last element needs additional 2 indices from next first, do not do that for last column
if  ( (col == cells_in_sector-1) && (sectionX != sectors - 1) )
{

	if (!cells[pvsi].add2) ALOG("SOMETHING BAD IS HAPPENING");
	cells[pvsi].indices[row][col*2 + 2] 	= (unsigned short)ArrI(x+1,y+1);
	cells[pvsi].indices[row][col*2+1 + 2] 	= (unsigned short)ArrI(x+1,y);

//	ALOG("BXX: x2y2"+IntToStr(ArrI(x+1,y+1)) + "   x2y1 :"+IntToStr(ArrI(x+1,y)) +" x: "+IntToStr(x+1)+ "  y: "+IntToStr(y+1));
}

}

//ALOG("5");
//flush to gpu
for (int n=0; n < sectors*sectors; n++)
{
	int k = 0;
	if (cells[n].add2) k = 2;

	for (int i=0; i < cells[n].len; i++)
	{

    glGenBuffers(1, &cells[n].elementbuff[i]);
    glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, cells[n].elementbuff[i]);
    glBufferData(GL_ELEMENT_ARRAY_BUFFER, cells[n].ilen * sizeof(unsigned short), cells[n].indices[i], GL_STATIC_DRAW);
	}
    glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, 0);
}

glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, 0);
//ALOG("6");

for (int n=0; n < sectors*sectors; n++)
{

//	ALOG("cell len: "+IntToStr(cells[n].len) + " additional 2: "+booltostring(cells[n].add2));

	AnsiString indi;
	for (int i=0; i < cells[n].len; i++)
	{
		indi = "";
		int ylen = cells[n].len * 2;
		if (cells[n].add2) ylen = ylen + 2;
		for (int j=0; j < ylen; j++)
		indi = indi + " " + IntToStr(int(cells[n].indices[i][j]));
//ALOG("cx: "+IntToStr(cells[n].x)+ " cy: "+IntToStr(cells[n].z)+ "  Indices: "+indi);
	}
}



}

};

#endif

Master & Mentor
https://sites.google.com/site/customprog/

DerekB

7

November 07, 2017 09:00 PM

On 14/01/2014 at 1:28 AM, newtechnology said:

I can't because I use XNAMath.

Anyways, I changed the code but still not getting it to work. (getting vector subscript out of range error.)

http://pastebin.com/kSibZFEf

You don't say where you are are getting the exception?

This line in Terrain::draw is wrong, I don't know if it would cause your exception, but it's definitely going to run off the end of your mSubMesh collection:

pDeviceContext->IASetVertexBuffers(0, mSubMesh.size(), &mSubMesh[a].VertexBuffer, &stride, &offset);

Edit: actually it's more wrong than I first thought, your SubMesh type has multiple members, that call will bind mSubMesh[a].VertexBuffer and then attempt to bind everything in memory following the vertex buffer as vertex buffers.

Quadtree/terrain frustum culling help

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Quadtree/terrain frustum culling help

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Reticulating splines