Jump to content
  • Advertisement

Warwick Michael Dumas

  • Content Count

  • Joined

  • Last visited

Everything posted by Warwick Michael Dumas

  1. Hi. N00b here. I wrote a class to create a graph of a surface and this class includes a shadow map. With one object it works fine. With multiple objects, strange things are happening. See picture - for all six objects the same calls are made, only the viewport coordinates are different. Only one of them displays the shadow correctly. The rendering sequence of the objects makes a difference to the behaviour, so it seems to me it is a problem with the way the shadow maps get initialised.   For the attached picture, this was the way the objects were treated:   surfacegraph Graph3, Graph4, Graph2, Graph5, Graph6,Graph; ..... Graph.Initialise(perRow,0,0,600,520,nscale,Eye); Graph2.Initialise(perRow,0,525,600,520,nscale,Eye); Graph3.Initialise(perRow,580,0,480,520,nscale,Eye); Graph4.Initialise(perRow,580,525,480,520,nscale,Eye); Graph5.Initialise(perRow,1050,0,600,520,nscale,Eye); Graph6.Initialise(perRow,1050,525,600,520,nscale,Eye); ..... Graph2.Render(); // the rendering sequence appears to make a big difference Graph4.Render(); // to the shadow maps that result Graph6.Render(); Graph.Render(); Graph3.Render(); Graph5.Render();     This creates the picture where Graph2 displays correctly, Graph displays with no shadow and all the rest display completely in shadow.   However I don't see why the objects are affecting each others' shadow maps.   This is some of the class definition:   class surfacegraph { public: static const int SHADOWMAPRESOLUTION = 4096; int perRow; int texWidth,texHeight; int numVertices, numTriangles; IDirect3DVertexBuffer9* TerrainVertexBuffer; IDirect3DIndexBuffer9* TerrainIndexBuffer; IDirect3DTexture9* texture_grid; D3DVIEWPORT9 vp; D3DVIEWPORT9 shadow_vp; ID3DXEffect* mFX; D3DXMATRIXA16 matView; D3DXMATRIXA16 matProj; DirLight mLight; Mtrl mWhiteMtrl; D3DXVECTOR3 Eye; // eye position // for shadow map: IDirect3DSurface9* pSurface; // never gets used for anything though - not sure why this was here D3DXMATRIX mLightVP; DrawableTex2D* mShadowMap; D3DXHANDLE mhTech; D3DXHANDLE mhWVP; D3DXHANDLE mhTex; D3DXHANDLE mhEyePos; // for specular D3DXHANDLE mhDiffuseMtrl; D3DXHANDLE mhAmbientMtrl; D3DXHANDLE mhLightVector; D3DXHANDLE mhLightDiffuseColour; D3DXHANDLE mhLightAmbientColour; D3DXHANDLE mhLightSpecularColour; D3DXHANDLE mhSpecularMtrl; D3DXHANDLE mhSpecularPower; D3DXHANDLE mhBuildShadowMapTech; D3DXHANDLE mhLightWVP; D3DXHANDLE mhShadowMap; D3DXHANDLE mhSwitch; float xyscale,nscale; surfacegraph(); HRESULT Initialise( parameters ); ... HRESULT SetxzDataAutomatic (float minimum, float maximum, float in_xyscale); HRESULT SetDataFromDoubleScaled(double **, double); HRESULT SetDataAgain(float **); HRESULT SetColourData(float **, float **, float **); VOID surfacegraph::Render(void); ~surfacegraph(); };   The following is the start of the Render method for the class. See code comments.   VOID surfacegraph::Render() { .... // DRAW SHADOW MAP: Direct3D.pd3dDevice->SetViewport(&vp); // without this^^^, 5 graphs, all except the last rendered, appear as small shadows on white mShadowMap->beginScene(); //Direct3D.pd3dDevice->SetViewport(&shadow_vp); // completely whites out all but last-to-render graph. No idea why. // That may suggest that Device->Clear is affecting the back buffer but we also know it is affecting the shadow map! Direct3D.pd3dDevice->Clear( 0, 0, D3DCLEAR_TARGET | D3DCLEAR_ZBUFFER, D3DCOLOR_XRGB( 255, 255, 255 ), 1.0f, 0 ); // this clearly does affect shadow map successfully... // 0x00000000, 1.0f, 0 ); // because this^^^^^ leads to everything being in shadow DXChk(mFX->SetTechnique(mhBuildShadowMapTech)); UINT numPasses = 0; DXChk(mFX->Begin(&numPasses, 0)); DXChk(mFX->SetMatrix(mhLightWVP, &(mLightVP))); DXChk(mFX->CommitChanges()); Direct3D.pd3dDevice->SetStreamSource(0, TerrainVertexBuffer, 0, sizeof(VertexPNT)); Direct3D.pd3dDevice->SetIndices(TerrainIndexBuffer); Direct3D.pd3dDevice->SetVertexDeclaration(VertexPNT::Decl); for (UINT ii = 0; ii < numPasses; ++ii) { mFX->BeginPass(ii); Direct3D.pd3dDevice->DrawIndexedPrimitive( D3DPT_TRIANGLELIST, 0, 0, numVertices, 0, numTriangles); mFX->EndPass(); }; DXChk(mFX->End()); mShadowMap->endScene(); // axis drawing in fixed function pipeline: Direct3D.pd3dDevice->SetViewport(&vp); ...   In case it is necessary background information, the "DrawableTex2D" class is    class DrawableTex2D { public: IDirect3DTexture9* mTex; ID3DXRenderToSurface* mRTS; IDirect3DSurface9* mTopSurf; UINT mWidth; UINT mHeight; UINT mMipLevels; D3DFORMAT mTexFormat; bool mUseDepthBuffer; D3DFORMAT mDepthFormat; D3DVIEWPORT9 mViewPort; bool mAutoGenMips; DrawableTex2D(UINT width, UINT height, UINT mipLevels, D3DFORMAT texFormat, bool useDepthBuffer, D3DFORMAT depthFormat, D3DVIEWPORT9& viewport, bool autoGenMips); ~DrawableTex2D(); IDirect3DTexture9* d3dTex(); void beginScene(); void endScene(); void onLostDevice(); void onResetDevice(); DrawableTex2D(const DrawableTex2D& rhs); DrawableTex2D& operator=(const DrawableTex2D& rhs); }; IDirect3DTexture9* DrawableTex2D::d3dTex() { return mTex; } void DrawableTex2D::onResetDevice() { UINT usage = D3DUSAGE_RENDERTARGET; if(mAutoGenMips) usage |= D3DUSAGE_AUTOGENMIPMAP; DXChk(D3DXCreateTexture(Direct3D.pd3dDevice, mWidth, mHeight, mMipLevels, usage, mTexFormat, D3DPOOL_DEFAULT, &mTex)); DXChk(D3DXCreateRenderToSurface(Direct3D.pd3dDevice, mWidth, mHeight, mTexFormat, mUseDepthBuffer, mDepthFormat, &mRTS)); DXChk(mTex->GetSurfaceLevel(0, &mTopSurf)); } void DrawableTex2D::beginScene() { mRTS->BeginScene(mTopSurf, &mViewPort); } void DrawableTex2D::endScene() { mRTS->EndScene(D3DX_FILTER_NONE); }   The following method is called, once only, after the graph object has been declared.   HRESULT surfacegraph::Initialise( parameters ) { ... shadow_vp.X = 0; shadow_vp.Y = 0; shadow_vp.Width = SHADOWMAPRESOLUTION; shadow_vp.Height = SHADOWMAPRESOLUTION; shadow_vp.MinZ = 0.0f; shadow_vp.MaxZ = 1.0f; mShadowMap = new DrawableTex2D(SHADOWMAPRESOLUTION, SHADOWMAPRESOLUTION, 1, D3DFMT_R32F, true, D3DFMT_D24X8, shadow_vp, false); vp.X = left; vp.Y = top; vp.Width = width; vp.Height = height; vp.MinZ = 0.0f; vp.MaxZ = 0.0f; float GlobalAspectRatio = ((float)vp.Width)/(float)vp.Height; // Create the FX from a .fx file. ID3DXBuffer* errors = 0; D3DXCreateEffectFromFile(Direct3D.pd3dDevice, "test1.fx", 0, 0, D3DXSHADER_DEBUG, 0, &mFX, &errors); if( errors ) MessageBox(0, (char*)errors->GetBufferPointer(), 0, 0); // Obtain handles. mhTech = mFX->GetTechniqueByName("WarwickTech"); ........ mShadowMap->onResetDevice(); static D3DXVECTOR3 vLookatPt( 0.0f, 1.2f, 0.0f ); static D3DXVECTOR3 vUpVec( 0.0f, 1.0f, 0.0f ); Eye = in_Eye; D3DXMatrixLookAtLH( &matView, &Eye, &vLookatPt, &vUpVec ); D3DXMatrixPerspectiveFovLH( &matProj, D3DX_PI / 4.0f, GlobalAspectRatio, 1.0f, 12.0f); mLight.dirW = D3DXVECTOR3(-1.0f, -0.5f, 0.5f);// remember y is height D3DXVec3Normalize(&mLight.dirW, &mLight.dirW); mLight.ambient = D3DXCOLOR(0.4f, 0.4f, 0.4f, 1.0f); mLight.diffuse = D3DXCOLOR(0.6f, 0.6f, 0.6f, 1.0f); mLight.spec = D3DXCOLOR(0.5f, 0.5f, 0.5f, 1.0f); mWhiteMtrl.ambient = WHITE*1.0f; mWhiteMtrl.diffuse = WHITE*1.0f; mWhiteMtrl.spec = WHITE*1.0f; mWhiteMtrl.specPower = 8.0f; // Shadow mapping: D3DXMATRIX lightView; D3DXVECTOR3 lightPosW(20.0f, 10.0f, -10.0f); D3DXVECTOR3 lightTargetW(0.0f, 0.0f, 0.0f); D3DXVECTOR3 lightUpW(0.0f, 1.0f, 0.0f); D3DXMatrixLookAtLH(&lightView, &lightPosW, &lightTargetW, &lightUpW); D3DXMATRIX lightLens; float lightFOV = D3DX_PI*0.25f; D3DXMatrixPerspectiveFovLH(&lightLens, lightFOV, 1.0f, 1.0f, 100.0f); mLightVP = lightView*lightLens; // graph data: perRow = nperRow; numVertices = perRow*perRow; numTriangles = 2*(perRow-1)*(perRow-1); texWidth = perRow; texHeight = perRow; definedindices = 0; if ( DXChk( Direct3D.pd3dDevice->CreateVertexBuffer(numVertices*sizeof(VertexPNT),D3DUSAGE_WRITEONLY,0,D3DPOOL_MANAGED,&TerrainVertexBuffer,0),1) + DXChk( Direct3D.pd3dDevice->CreateIndexBuffer(numTriangles*3*sizeof(DWORD),D3DUSAGE_WRITEONLY, D3DFMT_INDEX32, D3DPOOL_MANAGED,&TerrainIndexBuffer,0),2) + DXChk( D3DXCreateTexture( Direct3D.pd3dDevice, texWidth,texHeight, 0,0,D3DFMT_X8R8G8B8,D3DPOOL_MANAGED, &texture_grid),3) ) return E_FAIL; D3DSURFACE_DESC textureDesc; texture_grid->GetLevelDesc(0, &textureDesc); if ( textureDesc.Format != D3DFMT_X8R8G8B8 ) return E_FAIL; return S_OK; }   Sorry to provide so much code, I'm not all that sure what is going on here. I have tried tinkering with what happens at the start of the Render method but I don't seem to know what I'm doing, the results (as described in the code comments) are hard for me to interpret.   Grateful for any suggestions.   I am happy to dropbox the whole source if anyone wishes to tinker with it.
  2. Sure here are the arguments HRESULT surfacegraph::Initialise(int nperRow, // data points per row if square grid data to be supplied int left, int top, int width, int height, float fscale, D3DXVECTOR3 in_Eye, bool Polar = 0, int number_in_outermost = 0, // data points in outermost circle if radially distributed data int number_circles = 0, // number of circles if radially distributed data float radius_outermost = 0.0f // max radius if radially distributed data )      The viewport coordinates are left,top,width,height. (And it's not radially distributed data). The calls to Initialise are in the top code box above. So yes, they are a bit overlapping. I think that should be OK though... the rendering of each object is meant to be independent.  
  • Advertisement

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

GameDev.net is your game development community. Create an account for your GameDev Portfolio and participate in the largest developer community in the games industry.

Sign me up!