Big Commit this will break a lot, fix a lot, but i thing is a good step:

Implemented all the correct format conversions in efb to texture copy.
replaced all the stretcrect calls with quad draws this must improve speed a bit.
A BIGGGGGGGGGG cleanup in the code and reorganization.
reimplemented zpeek using a secondary render target ( this still is buggy so many issues left)
please a lot off feedback.

git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@4520 8ced0084-cf51-0410-be5f-012b33b47a6e

Source/Core/VideoCommon/Src/PixelShaderGen.cpp

@@ -451,7 +451,9 @@ const char *GeneratePixelShader(u32 texture_mask, bool dstAlphaEnable, bool HLSL
 
     WRITE(p, "void main(\n");
 
-    WRITE(p, "  out half4 ocol0 : COLOR0,\n");
+    WRITE(p, "  out float4 ocol0 : COLOR0,\n");
+	if (HLSL)
+		WRITE(p, "  out float4 ocol1 : COLOR1,\n");
     WRITE(p, "  out float depth : DEPTH,\n");
 
     // compute window position if needed because binding semantic WPOS is not widely supported
@@ -522,12 +524,20 @@ const char *GeneratePixelShader(u32 texture_mask, bool dstAlphaEnable, bool HLSL
 
     // use the texture input of the last texture stage (textemp), hopefully this has been read and is in correct format...
     if (bpmem.ztex2.op == ZTEXTURE_ADD)
-        WRITE(p, "depth = frac(dot("I_ZBIAS"[0].xyzw, textemp.xyzw) + "I_ZBIAS"[1].w + zCoord);\n");
+        WRITE(p, "zCoord = frac(dot("I_ZBIAS"[0].xyzw, textemp.xyzw) + "I_ZBIAS"[1].w + zCoord);\n");
     else if (bpmem.ztex2.op == ZTEXTURE_REPLACE)
-        WRITE(p, "depth = frac(dot("I_ZBIAS"[0].xyzw, textemp.xyzw) + "I_ZBIAS"[1].w);\n");
+        WRITE(p, "zCoord = frac(dot("I_ZBIAS"[0].xyzw, textemp.xyzw) + "I_ZBIAS"[1].w);\n");
-    else
+    
     WRITE(p, "depth = zCoord;\n");
 
+	if(HLSL)
+	{
+		//WRITE(p, "ocol1 = float4(1.0f/255.0f,2.0f/255.0f,3.0f/255.0f,0.0f);\n");
+		WRITE(p, "float4 EncodedDepth = frac(zCoord * float4(254.0f*255.0f,255.0f,254.0f/255.0f,254.0f*255.0f*255.0f));\n");
+		//WRITE(p, "EncodedDepth -= EncodedDepth.aarg * float4(1.0f/255.0f,1.0f/255.0f,1.0f/255.0f,0.0f);\n");
+		WRITE(p, "ocol1 = EncodedDepth;\n");
+		
+	}
     //if (bpmem.genMode.numindstages ) WRITE(p, "prev.rg = indtex0.xy;\nprev.b = 0;\n");
 
     if (!WriteAlphaTest(p, HLSL))

Source/Plugins/Plugin_VideoDX9/Src/D3DBase.cpp

@@ -485,6 +485,14 @@ void SetSamplerState(DWORD Sampler, D3DSAMPLERSTATETYPE Type, DWORD Value)
 	}
 }
 
+void RefreshVertexDeclaration()
+{
+	if (m_VtxDecl)
+	{
+		D3D::dev->SetVertexDeclaration(m_VtxDecl);		
+	}
+}
+
 void SetVertexDeclaration(LPDIRECT3DVERTEXDECLARATION9 decl)
 {
 	if (!decl) {

Source/Plugins/Plugin_VideoDX9/Src/D3DBase.h

@@ -63,7 +63,8 @@ int GetBackBufferWidth();
 int GetBackBufferHeight();
 LPDIRECT3DSURFACE9 GetBackBufferSurface();
 LPDIRECT3DSURFACE9 GetBackBufferDepthSurface();
-
+LPDIRECT3DVERTEXBUFFER9  GetquadVB();
+LPDIRECT3DVERTEXDECLARATION9 GetBasicvertexDecl();
 const D3DCAPS9 &GetCaps();
 const char *PixelShaderVersionString();
 const char *VertexShaderVersionString();
@@ -74,6 +75,7 @@ void SetTexture(DWORD Stage, IDirect3DBaseTexture9 *pTexture);
 void SetRenderState(D3DRENDERSTATETYPE State, DWORD Value);
 void SetTextureStageState(DWORD Stage, D3DTEXTURESTAGESTATETYPE Type, DWORD Value);
 void SetSamplerState(DWORD Sampler, D3DSAMPLERSTATETYPE Type, DWORD Value);
+void RefreshVertexDeclaration();
 void SetVertexDeclaration(LPDIRECT3DVERTEXDECLARATION9 decl);
 void ApplyCachedState();
 

Source/Plugins/Plugin_VideoDX9/Src/D3DUtil.cpp

@@ -208,10 +208,24 @@ void SaveRenderStates()
 
 void RestoreRenderStates()
 {
+	if(texture_old)
+	{
 		D3D::SetTexture(0, texture_old);
-
+		texture_old->Release();
+		texture_old = NULL;
+	}
+	if(ps_old)
+	{
 		dev->SetPixelShader(ps_old);
+		ps_old->Release();
+		ps_old = NULL;
+	}
+	if(vs_old)
+	{
 		dev->SetVertexShader(vs_old);
+		vs_old->Release();
+		vs_old = NULL;
+	}
 	
 	for (int i = 0; i < 6; i++)
 	{
@@ -374,4 +388,37 @@ void quad2d(float x1, float y1, float x2, float y2, u32 color, float u1, float v
 	RestoreRenderStates();
 }
 
+void drawShadedTexQuad(IDirect3DTexture9 *texture,
+					   const RECT *rSource,
+					   int SourceWidth,
+					   int SourceHeight,
+					   const RECT *rDest,
+					   IDirect3DPixelShader9 *PShader,
+					   IDirect3DVertexShader9 *Vshader)
+{
+	SaveRenderStates();
+	
+	float span = ((rSource->right-rSource->left - 1.0f) * (rDest->right - rDest->left))/(SourceWidth*((rDest->right - rDest->left)-1.0f));
+	float u1=((0.5f+rSource->left)/(float) SourceWidth)-(span*0.5f/(float)(rDest->right - rDest->left));
+	float u2=u1+span;  
+	span = ((rSource->bottom-rSource->top - 1.0f) * (rDest->bottom - rDest->top))/(SourceHeight*((rDest->bottom - rDest->top)-1.0f));
+	float v1=((0.5f+rSource->top)/(float) SourceHeight)-(span*0.5f/(float)(rDest->bottom - rDest->top));
+	float v2=v1+span; 
+
+	struct Q2DVertex { float x,y,z,rhw,u,v; } coords[4] = {
+		{(float)rDest->left-1.0f, (float)rDest->top-1.0f, 0.0f, 1.0f, u1, v1},
+		{(float)rDest->right, (float)rDest->top-1.0f, 0.0f,1.0f, u2, v1},
+		{(float)rDest->right, (float)rDest->bottom, 0.0f,1.0f, u2, v2},
+		{(float)rDest->left-1.0f, (float)rDest->bottom, 0.0f,1.0f, u1, v2}
+	};
+	dev->SetFVF(D3DFVF_XYZRHW | D3DFVF_TEX1);
+	dev->SetVertexShader(Vshader);
+	dev->SetPixelShader(PShader);	
+	dev->SetTexture(0, texture);
+	dev->DrawPrimitiveUP(D3DPT_TRIANGLEFAN, 2, coords, sizeof(Q2DVertex));	
+	D3D::RefreshVertexDeclaration();
+	RestoreRenderStates();
+
+}	
+
 }  // namespace

Source/Plugins/Plugin_VideoDX9/Src/D3DUtil.h

@@ -58,4 +58,13 @@ namespace D3D
 	extern CD3DFont font;
 
 	void quad2d(float x1, float y1, float x2, float y2, u32 color, float u1=0, float v1=0, float u2=1, float v2=1);
+	void drawShadedTexQuad(IDirect3DTexture9 *texture,
+					   const RECT *rSource,
+					   int SourceWidth,
+					   int SourceHeight,
+					   const RECT *rDest,
+					   IDirect3DPixelShader9 *PShader,
+					   IDirect3DVertexShader9 *Vshader);
+	void SaveRenderStates();
+	void RestoreRenderStates();
 }

Source/Plugins/Plugin_VideoDX9/Src/FramebufferManager.cpp

@@ -23,34 +23,62 @@
 namespace FBManager
 {
 
-static LPDIRECT3DTEXTURE9 s_efb_color_texture;
+static LPDIRECT3DTEXTURE9 s_efb_color_texture;//Texture thats contains the color data of the render target
-static LPDIRECT3DTEXTURE9 s_efb_colorBuffer_texture;
+static LPDIRECT3DTEXTURE9 s_efb_colorRead_texture;//1 pixel texture for temporal data store
-static LPDIRECT3DTEXTURE9 s_efb_depth_texture;
+static LPDIRECT3DTEXTURE9 s_efb_depth_texture;//Texture thats contains the depth data of the render target
-static LPDIRECT3DTEXTURE9 s_efb_depthBuffer_texture;
+static LPDIRECT3DTEXTURE9 s_efb_depthRead_texture;//1 pixel texture for temporal data store
-static LPDIRECT3DSURFACE9 s_efb_color_surface;
-static LPDIRECT3DSURFACE9 s_efb_depth_surface;
 
-static LPDIRECT3DSURFACE9 s_efb_color_ReadBuffer;
+static LPDIRECT3DSURFACE9 s_efb_depth_surface;//Depth Surface
-static LPDIRECT3DSURFACE9 s_efb_depth_ReadBuffer;
+static LPDIRECT3DSURFACE9 s_efb_depthColor_surface;//Depth, color encoded Surface
+static LPDIRECT3DSURFACE9 s_efb_color_surface;//Color Surface
+static LPDIRECT3DSURFACE9 s_efb_color_ReadBuffer;//Surface 0 of s_efb_colorRead_texture
+static LPDIRECT3DSURFACE9 s_efb_depth_ReadBuffer;//Surface 0 of s_efb_depthRead_texture
 
-static LPDIRECT3DSURFACE9 s_efb_color_OffScreenReadBuffer;
+static LPDIRECT3DSURFACE9 s_efb_color_OffScreenReadBuffer;//System memory Surface that can be locked to retriebe the data
-static LPDIRECT3DSURFACE9 s_efb_depth_OffScreenReadBuffer;
+static LPDIRECT3DSURFACE9 s_efb_depth_OffScreenReadBuffer;//System memory Surface that can be locked to retriebe the data
 
 
-static D3DFORMAT s_efb_color_surface_Format;
+static D3DFORMAT s_efb_color_surface_Format;//Format of the color Surface
-static D3DFORMAT s_efb_depth_surface_Format;
+static D3DFORMAT s_efb_depth_surface_Format;//Format of the Depth color encoded Surface
 #undef CHECK
 #define CHECK(hr,Message) if (FAILED(hr)) { PanicAlert(__FUNCTION__ " FAIL: %s" ,Message); }
 
 
 
-LPDIRECT3DSURFACE9 GetEFBColorRTSurface() { return s_efb_color_surface; }
+LPDIRECT3DSURFACE9 GetEFBColorRTSurface() 
-LPDIRECT3DSURFACE9 GetEFBDepthRTSurface() { return s_efb_depth_surface; }
+{ 	
-LPDIRECT3DSURFACE9 GetEFBColorOffScreenRTSurface() { return s_efb_color_OffScreenReadBuffer; }
+	return s_efb_color_surface; 
-LPDIRECT3DSURFACE9 GetEFBDepthOffScreenRTSurface() { return s_efb_depth_OffScreenReadBuffer; }
+}
+LPDIRECT3DSURFACE9 GetEFBDepthRTSurface() 
+{ 
+	return s_efb_depth_surface; 
+}
+
+LPDIRECT3DSURFACE9 GetEFBDepthEncodedSurface() 
+{ 	
+	return s_efb_depthColor_surface; 
+}
+
+LPDIRECT3DSURFACE9 GetEFBColorOffScreenRTSurface() 
+{ 
+	return s_efb_color_OffScreenReadBuffer; 
+}
+LPDIRECT3DSURFACE9 GetEFBDepthOffScreenRTSurface() 
+{ 
+	return s_efb_depth_OffScreenReadBuffer; 
+}
+
+LPDIRECT3DSURFACE9 GetEFBColorReadSurface() 
+{ 
+	
+	return s_efb_color_ReadBuffer; 
+}
+LPDIRECT3DSURFACE9 GetEFBDepthReadSurface() 
+{
+	
+	return s_efb_depth_ReadBuffer; 
+}
 
-LPDIRECT3DSURFACE9 GetEFBColorReadSurface() { return s_efb_color_ReadBuffer; }
-LPDIRECT3DSURFACE9 GetEFBDepthReadSurface() { return s_efb_depth_ReadBuffer; }
 
 D3DFORMAT GetEFBDepthRTSurfaceFormat(){return s_efb_depth_surface_Format;}
 D3DFORMAT GetEFBColorRTSurfaceFormat(){return s_efb_color_surface_Format;}
@@ -63,14 +91,9 @@ LPDIRECT3DTEXTURE9 GetEFBColorTexture(const EFBRectangle& sourceRc)
 
 LPDIRECT3DTEXTURE9 GetEFBDepthTexture(const EFBRectangle &sourceRc)
 {
-	// Depth textures not supported under DX9. We're gonna fake this
-	// with a secondary render target later.
 	return s_efb_depth_texture;
 }
 
-
-
-
 void Create()
 {
 	
@@ -79,29 +102,55 @@ void Create()
 	int target_height = Renderer::GetTargetHeight();
 	s_efb_color_surface_Format = D3DFMT_A8R8G8B8;
 	//get the framebuffer texture
-	HRESULT hr = D3D::dev->CreateTexture(target_width, target_height, 1, D3DUSAGE_RENDERTARGET, D3DFMT_A8R8G8B8,
+	HRESULT hr = D3D::dev->CreateTexture(target_width, target_height, 1, D3DUSAGE_RENDERTARGET, s_efb_color_surface_Format,
 		                                 D3DPOOL_DEFAULT, &s_efb_color_texture, NULL);
-	CHECK(hr,"Create Color Texture");
+	if(s_efb_color_texture)
-	//get the Surface
-  	hr = s_efb_color_texture->GetSurfaceLevel(0, &s_efb_color_surface);
-	CHECK(hr,"Get Color Surface");
-	//create a one pixel texture to work as a buffer for peeking
-	hr = D3D::dev->CreateTexture(1, 1, 1, D3DUSAGE_RENDERTARGET, D3DFMT_A8R8G8B8,
-		                                  D3DPOOL_DEFAULT, &s_efb_colorBuffer_texture, NULL);
-	if (!FAILED(hr))
 	{
-		//get the surface for the peeking texture
+		s_efb_color_texture->GetSurfaceLevel(0,&s_efb_color_surface);
-		hr = s_efb_colorBuffer_texture->GetSurfaceLevel(0, &s_efb_color_ReadBuffer);
-		CHECK(hr,"Get Color Pixel Surface");
-		//create an offscreen surface that we can lock to retrieve the data
-		hr = D3D::dev->CreateOffscreenPlainSurface(1, 1, D3DFMT_A8R8G8B8, D3DPOOL_SYSTEMMEM, &s_efb_color_OffScreenReadBuffer, NULL );
-		CHECK(hr,"Create Color offScreen Surface");
 	}	
+	CHECK(hr,"Create Color Texture");
+	hr = D3D::dev->CreateTexture(1, 1, 1, D3DUSAGE_RENDERTARGET, s_efb_color_surface_Format,
+		                                  D3DPOOL_DEFAULT, &s_efb_colorRead_texture, NULL);
+	CHECK(hr,"Create Color Read Texture");
+	if(s_efb_colorRead_texture)
+	{
+		s_efb_colorRead_texture->GetSurfaceLevel(0,&s_efb_color_ReadBuffer);
+	}
+	//create an offscreen surface that we can lock to retrieve the data
+	hr = D3D::dev->CreateOffscreenPlainSurface(1, 1, s_efb_color_surface_Format, D3DPOOL_SYSTEMMEM, &s_efb_color_OffScreenReadBuffer, NULL );
+	CHECK(hr,"Create Color offScreen Surface");	
 	
 	//Select Zbuffer format supported by hadware.
 	if (g_ActiveConfig.bEFBAccessEnable)
 	{
-		//depth format in prefered order
+		
+		hr = D3D::dev->CreateDepthStencilSurface(target_width, target_height, D3DFMT_D24X8,
+											 D3DMULTISAMPLE_NONE, 0, FALSE, &s_efb_depth_surface, NULL);
+		CHECK(hr,"CreateDepthStencilSurface");
+
+		s_efb_depth_surface_Format = D3DFMT_A8R8G8B8;
+		//get the framebuffer Depth texture
+		HRESULT hr = D3D::dev->CreateTexture(target_width, target_height, 1, D3DUSAGE_RENDERTARGET, s_efb_depth_surface_Format,
+											 D3DPOOL_DEFAULT, &s_efb_depth_texture, NULL);
+		CHECK(hr,"Depth Color Texture");
+		//get the Surface
+		if(s_efb_depth_texture)
+		{
+			s_efb_depth_texture->GetSurfaceLevel(0,&s_efb_depthColor_surface);
+		}
+  		//create a one pixel texture to work as a buffer for peeking
+		hr = D3D::dev->CreateTexture(1, 1, 1, D3DUSAGE_RENDERTARGET, s_efb_depth_surface_Format,
+											  D3DPOOL_DEFAULT, &s_efb_depthRead_texture, NULL);
+		CHECK(hr,"Create Depth Read texture");
+		if(s_efb_depthRead_texture)
+		{
+			s_efb_depthRead_texture->GetSurfaceLevel(0,&s_efb_depth_ReadBuffer);
+		}
+		//create an offscreen surface that we can lock to retrieve the data
+		hr = D3D::dev->CreateOffscreenPlainSurface(1, 1, s_efb_depth_surface_Format, D3DPOOL_SYSTEMMEM, &s_efb_depth_OffScreenReadBuffer, NULL );
+		CHECK(hr,"Create Depth offScreen Surface");
+		
+		/*//depth format in prefered order
 		D3DFORMAT *DepthTexFormats = new D3DFORMAT[3];
 		DepthTexFormats[0] = D3DFMT_D32F_LOCKABLE;		
 		DepthTexFormats[1] = D3DFMT_D16_LOCKABLE;
@@ -116,7 +165,7 @@ void Create()
 		s_efb_depth_ReadBuffer = s_efb_depth_surface;
 		s_efb_depth_OffScreenReadBuffer = s_efb_depth_surface;		
 		CHECK(hr,"CreateDepthStencilSurface");
-		delete [] DepthTexFormats;
+		delete [] DepthTexFormats;*/
 	}
 	else
 	{
@@ -129,44 +178,51 @@ void Create()
 
 void Destroy()
 {
-	if(s_efb_depth_ReadBuffer)
-		s_efb_depth_ReadBuffer->Release();
-	s_efb_depth_ReadBuffer = NULL;
 	
-	if(s_efb_depth_OffScreenReadBuffer)
+if(s_efb_depth_surface)
-		s_efb_depth_OffScreenReadBuffer->Release();
+s_efb_depth_surface->Release();
+s_efb_depth_surface=NULL;
 
-	if(s_efb_depth_surface)
+if(s_efb_depthColor_surface)
-		s_efb_depth_surface->Release();
+s_efb_depthColor_surface->Release();
-	s_efb_depth_surface = NULL;
+s_efb_depthColor_surface=NULL;
 
-	if(s_efb_depthBuffer_texture)
+if(s_efb_color_surface)
-		s_efb_depthBuffer_texture->Release();
+s_efb_color_surface->Release();
-	s_efb_depthBuffer_texture=NULL;
+s_efb_color_surface=NULL;
 
-	if(s_efb_depth_texture)
+if(s_efb_color_ReadBuffer)
-		s_efb_depth_texture->Release();
+s_efb_color_ReadBuffer->Release();
-	s_efb_depth_texture = NULL;
+s_efb_color_ReadBuffer=NULL;
 
-	if(s_efb_color_OffScreenReadBuffer )
+if(s_efb_depth_ReadBuffer)
-		s_efb_color_OffScreenReadBuffer->Release();
+s_efb_depth_ReadBuffer->Release();
-	s_efb_color_OffScreenReadBuffer  = NULL;
+s_efb_depth_ReadBuffer=NULL;
 
-	if(s_efb_color_ReadBuffer )
+if(s_efb_color_OffScreenReadBuffer)
-		s_efb_color_ReadBuffer->Release();
+s_efb_color_OffScreenReadBuffer->Release();
-	s_efb_color_ReadBuffer  = NULL;
+s_efb_color_OffScreenReadBuffer=NULL;
 
-	if(s_efb_color_surface)
+if(s_efb_depth_OffScreenReadBuffer)
-		s_efb_color_surface->Release();
+s_efb_depth_OffScreenReadBuffer->Release();
-	s_efb_color_surface = NULL;	
+s_efb_depth_OffScreenReadBuffer=NULL;
 
-	if(s_efb_colorBuffer_texture)
+if(s_efb_color_texture)
-		s_efb_colorBuffer_texture->Release();
-	s_efb_colorBuffer_texture = NULL;
-	
-	if(s_efb_color_texture)
 	s_efb_color_texture->Release();
-	s_efb_color_texture = NULL;
+s_efb_color_texture=NULL;
+
+if(s_efb_colorRead_texture)
+s_efb_colorRead_texture->Release();
+s_efb_colorRead_texture=NULL;
+
+if(s_efb_depth_texture)
+s_efb_depth_texture->Release();
+s_efb_depth_texture=NULL;
+
+if(s_efb_depthRead_texture)
+s_efb_depthRead_texture->Release();
+s_efb_depthRead_texture=NULL;
+
 }
 
 }  // namespace

Source/Plugins/Plugin_VideoDX9/Src/FramebufferManager.h

@@ -40,6 +40,7 @@ D3DFORMAT GetEFBDepthRTSurfaceFormat();
 D3DFORMAT GetEFBColorRTSurfaceFormat();
 LPDIRECT3DSURFACE9 GetEFBColorReadSurface();
 LPDIRECT3DSURFACE9 GetEFBDepthReadSurface();
+LPDIRECT3DSURFACE9 GetEFBDepthEncodedSurface();
 
 
 

Source/Plugins/Plugin_VideoDX9/Src/PixelShaderCache.cpp

@@ -34,6 +34,19 @@ PixelShaderCache::PSCache PixelShaderCache::PixelShaders;
 const PixelShaderCache::PSCacheEntry *PixelShaderCache::last_entry;
 static float lastPSconstants[C_COLORMATRIX+16][4];
 
+static LPDIRECT3DPIXELSHADER9 s_ColorMatrixProgram = 0;
+static LPDIRECT3DPIXELSHADER9 s_ColorCopyProgram = 0;
+
+LPDIRECT3DPIXELSHADER9 PixelShaderCache::GetColorMatrixProgram()
+{
+	return s_ColorMatrixProgram;
+}
+
+LPDIRECT3DPIXELSHADER9 PixelShaderCache::GetColorCopyProgram()
+{
+	return s_ColorCopyProgram;
+}
+
 void SetPSConstant4f(int const_number, float f1, float f2, float f3, float f4)
 {
 	if( lastPSconstants[const_number][0] != f1 || lastPSconstants[const_number][1] != f2 ||
@@ -63,6 +76,25 @@ void SetPSConstant4fv(int const_number, const float *f)
 
 void PixelShaderCache::Init()
 {
+	char pmatrixprog[1024];
+	sprintf(pmatrixprog,"uniform sampler samp0 : register(s0);\n"
+						"uniform float4 cColMatrix[5] : register(c%d);\n"
+						"void main(\n"
+						"out float4 ocol0 : COLOR0,\n"
+						" in float3 uv0 : TEXCOORD0){\n"
+						"float4 texcol = tex2D(samp0,uv0.xy);\n"
+						"ocol0 = float4(dot(texcol,cColMatrix[0]),dot(texcol,cColMatrix[1]),dot(texcol,cColMatrix[2]),dot(texcol,cColMatrix[3])) + cColMatrix[4];\n"						
+						"}\n",C_COLORMATRIX);
+	char pcopyprog[1024];
+	sprintf(pcopyprog,"uniform sampler samp0 : register(s0);\n"
+						"void main(\n"
+						"out float4 ocol0 : COLOR0,\n"
+						" in float3 uv0 : TEXCOORD0){\n"
+						"ocol0 = tex2D(samp0,uv0.xy);\n"						
+						"}\n");
+
+	s_ColorMatrixProgram = D3D::CompilePixelShader(pmatrixprog, (int)strlen(pmatrixprog));
+	s_ColorCopyProgram = D3D::CompilePixelShader(pcopyprog, (int)strlen(pcopyprog));
 	Clear();
 }
 
@@ -80,6 +112,12 @@ void PixelShaderCache::Clear()
 
 void PixelShaderCache::Shutdown()
 {
+	if(s_ColorMatrixProgram)
+		s_ColorMatrixProgram->Release();
+	s_ColorMatrixProgram = NULL;
+	if(s_ColorCopyProgram)
+		s_ColorCopyProgram->Release();
+	s_ColorCopyProgram=NULL;
 	Clear();
 }
 

Source/Plugins/Plugin_VideoDX9/Src/PixelShaderCache.h

@@ -59,6 +59,8 @@ public:
 	static void Clear();
 	static void Shutdown();
 	static bool SetShader(bool dstAlpha);
+	static LPDIRECT3DPIXELSHADER9 GetColorMatrixProgram();
+	static LPDIRECT3DPIXELSHADER9 GetColorCopyProgram();
 #if defined(_DEBUG) || defined(DEBUGFAST)
 	static std::string GetCurrentShaderCode();
 #endif

Source/Plugins/Plugin_VideoDX9/Src/Render.cpp

@@ -63,6 +63,7 @@ static u32 s_blendMode;
 
 char st[32768];
 
+
 // State translation lookup tables
 static const D3DBLEND d3dSrcFactors[8] =
 {
@@ -98,6 +99,7 @@ void SetupDeviceObjects()
 	PixelShaderManager::Dirty();
 
 	// Tex and shader caches will recreate themselves over time.
+
 }
 
 // Kill off all POOL_DEFAULT device objects.
@@ -159,15 +161,25 @@ bool Renderer::Init()
 
 	for (int stage = 0; stage < 8; stage++)
 		D3D::SetSamplerState(stage, D3DSAMP_MAXANISOTROPY, g_ActiveConfig.iMaxAnisotropy);
-
+	D3DVIEWPORT9 vp;
-	D3D::dev->Clear(0, NULL, D3DCLEAR_ZBUFFER,D3DCOLOR_XRGB(255,255,255),1.0f,0);
+	vp.X = 0;
+	vp.Y = 0;
+	vp.Width  = s_backbuffer_width;
+	vp.Height = s_backbuffer_height;
+	vp.MinZ = 0.0f;
+	vp.MaxZ = 0.0f;
+	D3D::dev->SetViewport(&vp);
 	D3D::dev->Clear(0, NULL, D3DCLEAR_TARGET, 0x0, 0, 0);
 	
 	D3D::dev->SetRenderTarget(0, FBManager::GetEFBColorRTSurface());
+	D3D::dev->SetRenderTarget(1, FBManager::GetEFBDepthEncodedSurface());
 	D3D::dev->SetDepthStencilSurface(FBManager::GetEFBDepthRTSurface());
-
+	vp.Width  = s_target_width;
+	vp.Height = s_target_height;
+	vp.MinZ = 0.0f;
+	vp.MaxZ = 0.0f;
+	D3D::dev->SetViewport(&vp);
 	D3D::dev->Clear(0, NULL, D3DCLEAR_TARGET | D3DCLEAR_ZBUFFER, 0x0, 1.0f, 0);
-
 	D3D::BeginFrame();
 	D3D::SetRenderState(D3DRS_SCISSORTESTENABLE, true);
 	return true;
@@ -209,9 +221,9 @@ void dumpMatrix(D3DXMATRIX &mtx)
 TargetRectangle Renderer::ConvertEFBRectangle(const EFBRectangle& rc)
 {
 	TargetRectangle result;
-	result.left   = (rc.left   * s_target_width)  / EFB_WIDTH;
+	result.left   = (rc.left   * s_target_width)  / EFB_WIDTH ;
 	result.top    = (rc.top    * s_target_height) / EFB_HEIGHT;
-	result.right  = (rc.right  * s_target_width)  / EFB_WIDTH;
+	result.right  = (rc.right  * s_target_width)  / EFB_WIDTH ;
 	result.bottom = (rc.bottom * s_target_height) / EFB_HEIGHT;
 	return result;
 }
@@ -261,31 +273,48 @@ void CheckForResize()
 static void EFBTextureToD3DBackBuffer(const EFBRectangle& sourceRc)
 {
 	// Set the backbuffer as the rendering target
-	D3D::dev->SetRenderTarget(0, D3D::GetBackBufferSurface());
 	D3D::dev->SetDepthStencilSurface(NULL);
+	D3D::dev->SetRenderTarget(1, NULL);
+	D3D::dev->SetRenderTarget(0, D3D::GetBackBufferSurface());	
 	
-	// Blit our render target onto the backbuffer.
-	// TODO: Change to a quad so we can do post processing.
 	TargetRectangle src_rect, dst_rect;
 	src_rect = Renderer::ConvertEFBRectangle(sourceRc);
 	ComputeDrawRectangle(s_backbuffer_width, s_backbuffer_height, false, &dst_rect);
-
+	D3DVIEWPORT9 vp;
-	//LPD3DXSPRITE pSprite=NULL;
+	vp.X = 0;
-	//D3DXCreateSprite(D3D::dev, &pSprite);
+	vp.Y = 0;
-	//D3DXVECTOR3 pos(0,0,0);
+	vp.Width  = s_backbuffer_width;
-	//EFBRectangle efbRect;
+	vp.Height = s_backbuffer_height;
-	//
+	vp.MinZ = 0.0f;
-	//pSprite->Begin(D3DXSPRITE_ALPHABLEND);
+	vp.MaxZ = 0.0f;
-	//pSprite->Draw(FBManager::GetEFBColorTexture(efbRect),NULL, NULL, &pos, 0xFFFFFFFF);
+	D3D::dev->SetViewport(&vp);
-	//pSprite->End();
-	//pSprite->Release();
 
 	D3D::dev->Clear(0,NULL, D3DCLEAR_TARGET,D3DCOLOR_XRGB(0,0,0),1.0f,0);
+	vp.X = dst_rect.left;
+	vp.Y = dst_rect.top;
+	vp.Width  = dst_rect.right - dst_rect.left;
+	vp.Height = dst_rect.bottom - dst_rect.top;
+	vp.MinZ = 0.0f;
+	vp.MaxZ = 0.0f;
+	D3D::dev->SetViewport(&vp);
 
-	// todo, to draw the EFB texture to the backbuffer instead of StretchRect
+	EFBRectangle efbRect;
-	D3D::dev->StretchRect(FBManager::GetEFBColorRTSurface(), src_rect.AsRECT(),
+	
-		D3D::GetBackBufferSurface(), dst_rect.AsRECT(),
+	
-		D3DTEXF_LINEAR);
+	
+	LPDIRECT3DTEXTURE9 read_texture = FBManager::GetEFBColorTexture(efbRect);
+	RECT destinationrect;
+	destinationrect.bottom = dst_rect.bottom;
+	destinationrect.left = dst_rect.left;
+	destinationrect.right = dst_rect.right;
+	destinationrect.top = dst_rect.top;
+	RECT sourcerect;
+	sourcerect.bottom = src_rect.bottom;
+	sourcerect.left = src_rect.left;
+	sourcerect.right = src_rect.right;
+	sourcerect.top = src_rect.top;
+
+	D3D::drawShadedTexQuad(read_texture,&sourcerect,Renderer::GetTargetWidth(),Renderer::GetTargetHeight(),&destinationrect,PixelShaderCache::GetColorCopyProgram(),VertexShaderCache::GetSimpleVertexSahder());	
 	
 	// Finish up the current frame, print some stats
 	if (g_ActiveConfig.bOverlayStats)
@@ -301,14 +330,11 @@ static void EFBTextureToD3DBackBuffer(const EFBRectangle& sourceRc)
 
 	OSD::DrawMessages();
 
-	// u32 clearColor = (bpmem.clearcolorAR << 16) | bpmem.clearcolorGB;
-
-	// Clear the render target. We probably don't need to do this every frame.
-	//D3D::dev->Clear(0, NULL, D3DCLEAR_TARGET, 0x0, 1.0f, 0);
-
-	// Set rendering target back to the EFB rendering texture
 	D3D::dev->SetRenderTarget(0, FBManager::GetEFBColorRTSurface());
+	D3D::dev->SetRenderTarget(1, FBManager::GetEFBDepthEncodedSurface());
 	D3D::dev->SetDepthStencilSurface(FBManager::GetEFBDepthRTSurface());
+	
+    VertexShaderManager::SetViewportChanged();
 }
 
 
@@ -386,10 +412,11 @@ void Renderer::RenderToXFB(u32 xfbAddr, u32 fbWidth, u32 fbHeight, const EFBRect
 		return;
 	}
 
-	D3D::EndFrame();
+	Renderer::ResetAPIState();
 	D3DDumpFrame();
 	EFBTextureToD3DBackBuffer(sourceRc);
-	D3D::BeginFrame();
+	D3D::EndFrame();
+	
 
 	DEBUGGER_LOG_AT((NEXT_XFB_CMD|NEXT_EFB_CMD|NEXT_FRAME),
 		{printf("StretchRect, EFB->XFB\n");});
@@ -399,21 +426,12 @@ void Renderer::RenderToXFB(u32 xfbAddr, u32 fbWidth, u32 fbHeight, const EFBRect
 		xfbAddr, fbWidth, fbHeight, 
 		sourceRc.left, sourceRc.top, sourceRc.right, sourceRc.bottom);}
 	);
-
-
-	RECT rc;
-	rc.left   = 0; 
-	rc.top    = 0;
-	rc.right  = (LONG)s_target_width;
-	rc.bottom = (LONG)s_target_height;
-	D3D::dev->SetScissorRect(&rc);
-	D3D::SetRenderState(D3DRS_SCISSORTESTENABLE, false);
-
-	UpdateViewport();
-
 	Swap(0,FIELD_PROGRESSIVE,0,0);	// we used to swap the buffer here, now we will wait
 										// until the XFB pointer is updated by VI	
-	D3D::SetRenderState(D3DRS_SCISSORTESTENABLE, true);
+	D3D::BeginFrame();
+	Renderer::RestoreAPIState();
+	UpdateViewport();
+	
 }
 
 bool Renderer::SetScissorRect()
@@ -432,10 +450,27 @@ bool Renderer::SetScissorRect()
 	rc.bottom = (int)(rc.bottom * yScale);
 
 	if (rc.left < 0) rc.left = 0;
+	if (rc.right < 0) rc.right = 0;
+	if (rc.left > s_target_width) rc.left = s_target_width;
 	if (rc.right > s_target_width) rc.right = s_target_width;	
 	if (rc.top < 0) rc.top = 0;
+	if (rc.bottom < 0) rc.bottom = 0;
+	if (rc.top > s_target_height) rc.top = s_target_height;
 	if (rc.bottom > s_target_height) rc.bottom = s_target_height;
-
+	/*LONG temprc = 0;
+	if(rc.right < rc.left)
+	{
+		temprc = rc.right;
+		rc.right = rc.left;
+		rc.left = temprc;
+	}
+	if(rc.bottom < rc.top)
+	{
+		temprc = rc.bottom;
+		rc.bottom = rc.top;
+		rc.top = temprc;
+	}
+	D3D::dev->SetScissorRect(&rc);*/
 	if (rc.right >= rc.left && rc.bottom >= rc.top)
 	{
 		D3D::dev->SetScissorRect(&rc);
@@ -446,6 +481,7 @@ bool Renderer::SetScissorRect()
 		WARN_LOG(VIDEO, "Bad scissor rectangle: %i %i %i %i", rc.left, rc.top, rc.right, rc.bottom);		
 		return false;
 	}
+	return true;
 }
 
 void Renderer::SetColorMask() 
@@ -460,10 +496,9 @@ void Renderer::SetColorMask()
 
 u32 Renderer::AccessEFB(EFBAccessType type, int x, int y)
 {
-	
 	//Get the working buffer
 	LPDIRECT3DSURFACE9 pBuffer = (type == PEEK_Z || type == POKE_Z) ? 
-		FBManager::GetEFBDepthRTSurface() : FBManager::GetEFBColorRTSurface();
+		FBManager::GetEFBDepthEncodedSurface() : FBManager::GetEFBColorRTSurface();
 	//get the temporal buffer to move 1pixel data
 	LPDIRECT3DSURFACE9 RBuffer = (type == PEEK_Z || type == POKE_Z) ? 
 		FBManager::GetEFBDepthReadSurface() : FBManager::GetEFBColorReadSurface();
@@ -482,11 +517,6 @@ u32 Renderer::AccessEFB(EFBAccessType type, int x, int y)
 		PanicAlert("No %s!", (type == PEEK_Z || type == POKE_Z) ? "Z-Buffer" : "Color EFB");
 		return 0;
 	}
-	// Z buffer lock not suported: returning
-	if((type == PEEK_Z || type == POKE_Z) && BufferFormat == D3DFMT_D24X8)
-	{
-		return 0;
-	}
 	// Get the rectangular target region covered by the EFB pixel.
 	
 	EFBRectangle efbPixelRc;
@@ -506,11 +536,6 @@ u32 Renderer::AccessEFB(EFBAccessType type, int x, int y)
 	RectToLock.right = targetPixelRc.right;
 	RectToLock.top = targetPixelRc.top;
 	
-	//lock the buffer
-
-	if(!(BufferFormat == D3DFMT_D32F_LOCKABLE || BufferFormat == D3DFMT_D16_LOCKABLE))
-	{
-		//the hard support stretchrect in both color and z so use it
 	hr = D3D::dev->StretchRect(pBuffer,&RectToLock,RBuffer,NULL, D3DTEXF_NONE);
 	if(FAILED(hr))
 	{
@@ -529,7 +554,7 @@ u32 Renderer::AccessEFB(EFBAccessType type, int x, int y)
 	RectToLock.left = 0;
 	RectToLock.right = 1;
 	RectToLock.top = 0;
-	}	
+		
 	//the surface is good.. lock it
 	if((hr = pOffScreenBuffer->LockRect(&drect, &RectToLock, D3DLOCK_READONLY)) != D3D_OK)
 	{
@@ -541,24 +566,17 @@ u32 Renderer::AccessEFB(EFBAccessType type, int x, int y)
 	switch(type) {
 		case PEEK_Z:
 			{
-			switch (BufferFormat)
+				static float ffrac = 255.0f/254.0f;
-			{
+				z = ((u32 *)drect.pBits)[0];
-			case D3DFMT_D32F_LOCKABLE:
+				float fvalue = (((float)(z & 0xFF)) / 255.0f) * ffrac;
-				val = ((float *)drect.pBits)[0];
+				fvalue += (((float)((z>>8) & 0xFF)) / 255.0f) * (ffrac/255.0f);
-				z = ((u32)(val * 0xffffff));// 0xFFFFFFFF;
+				fvalue += (((float)((z>>16) & 0xFF)) / 255.0f) * (ffrac/(255.0f*255.0f));
-				break;
+				fvalue += (((float)((z>>24) & 0xFF)) / 255.0f) * (ffrac/(255.0f*255.0f*255.0f));
-			case D3DFMT_D16_LOCKABLE:
+				if(fvalue>1.0f)fvalue=1.0f;
-				val = ((float)((u16 *)drect.pBits)[0])/((float)0xFFFF);
+				if(fvalue<0.0f)fvalue=0.0f;
-				z = ((u32)(val * 0xffffff));
+				z = ((u32)(fvalue * 0xffffff));			
-				break;
-			default:
-				z = ((u32 *)drect.pBits)[0] >> 8;
-				break;
-			};			
-			// [0.0, 1.0] ==> [0, 0xFFFFFFFF]
-			
-			break;
 			}
+			break;			
 		case POKE_Z:
 			// TODO: Get that Z value to poke from somewhere
 			//((float *)drect.pBits)[0] = val;
@@ -568,7 +586,6 @@ u32 Renderer::AccessEFB(EFBAccessType type, int x, int y)
 		case PEEK_COLOR:
 			z = ((u32 *)drect.pBits)[0];
 			break;
-
 		case POKE_COLOR:
 			// TODO: Get that ARGB value to poke from somewhere
 			//((float *)drect.pBits)[0] = val;
@@ -578,8 +595,6 @@ u32 Renderer::AccessEFB(EFBAccessType type, int x, int y)
 
 
 	pOffScreenBuffer->UnlockRect();	
-
-
 	// TODO: in RE0 this value is often off by one, which causes lighting to disappear
 	return z;
 	
@@ -605,6 +620,14 @@ void UpdateViewport()
 	vp.Y = (int)(ceil(xfregs.rawViewport[4] + xfregs.rawViewport[1] - (scissorYOff)) * MValueY);
 	vp.Width  = (int)ceil(abs((int)(2 * xfregs.rawViewport[0])) * MValueX);
 	vp.Height = (int)ceil(abs((int)(2 * xfregs.rawViewport[1])) * MValueY);
+	if(vp.X < 0) vp.X = 0;
+	if(vp.Y < 0) vp.Y = 0;
+	if(vp.X > s_target_width) vp.X = s_target_width;
+	if(vp.Y > s_target_height) vp.Y = s_target_height;
+	if(vp.Width < 0) vp.Width = 0;
+	if(vp.Height < 0) vp.Height = 0;
+	if(vp.Width > (s_target_width - vp.X)) vp.Width = s_target_width - vp.X;
+	if(vp.Height > (s_target_height - vp.Y)) vp.Height = s_target_height - vp.Y;
 	//some games set invalids values for z min and z max so fix them to the max an min alowed and let the shaders do this work
 	vp.MinZ = 0.0f;//(xfregs.rawViewport[5] - xfregs.rawViewport[2]) / 16777216.0f;
 	vp.MaxZ = 1.0f;//xfregs.rawViewport[5] / 16777216.0f;
@@ -719,3 +742,29 @@ void Renderer::Swap(u32 xfbAddr, FieldType field, u32 fbWidth, u32 fbHeight)
 	// Flip/present backbuffer to frontbuffer here
 	D3D::Present();
 }
+
+void Renderer::ResetAPIState()
+{
+	D3D::SetRenderState(D3DRS_SCISSORTESTENABLE, FALSE);
+    D3D::SetRenderState(D3DRS_CULLMODE, D3DCULL_NONE);
+    D3D::SetRenderState(D3DRS_ALPHABLENDENABLE, FALSE);
+    D3D::SetRenderState(D3DRS_ZENABLE, FALSE);
+	D3D::SetRenderState(D3DRS_ZWRITEENABLE, FALSE);
+    DWORD color_mask = D3DCOLORWRITEENABLE_ALPHA| D3DCOLORWRITEENABLE_RED | D3DCOLORWRITEENABLE_GREEN | D3DCOLORWRITEENABLE_BLUE;
+	D3D::SetRenderState(D3DRS_COLORWRITEENABLE, color_mask);
+}
+
+void Renderer::RestoreAPIState()
+{
+	// Gets us back into a more game-like state.
+
+	UpdateViewport();
+
+    if (bpmem.zmode.testenable) D3D::SetRenderState(D3DRS_ZENABLE, TRUE);
+    if (bpmem.zmode.updateenable)   D3D::SetRenderState(D3DRS_ZWRITEENABLE, TRUE);
+
+    D3D::SetRenderState(D3DRS_SCISSORTESTENABLE, TRUE);
+	SetScissorRect();
+    SetColorMask();
+	SetBlendMode(true);	
+}

Source/Plugins/Plugin_VideoDX9/Src/TextureCache.cpp

@@ -27,7 +27,12 @@
 
 #include "D3DBase.h"
 #include "D3DTexture.h"
+#include "D3DUtil.h"
 #include "FramebufferManager.h"
+#include "PixelShaderCache.h"
+#include "PixelShaderManager.h"
+#include "VertexShaderManager.h"
+#include "VertexShaderCache.h"
 
 #include "Render.h"
 
@@ -257,9 +262,12 @@ TextureCache::TCacheEntry *TextureCache::Load(int stage, u32 address, int width,
 	return &entry;
 }
  
+#undef CHECK
+#define CHECK(hr) if (FAILED(hr)) { PanicAlert(__FUNCTION__ " FAIL"); }
 // EXTREMELY incomplete.
 void TextureCache::CopyRenderTargetToTexture(u32 address, bool bFromZBuffer, bool bIsIntensityFmt, u32 copyfmt, int bScaleByHalf, const EFBRectangle &source_rect)
 {
+	HRESULT hr = S_OK;
 	int efb_w = source_rect.GetWidth();
 	int efb_h = source_rect.GetHeight();
 
@@ -271,19 +279,20 @@ void TextureCache::CopyRenderTargetToTexture(u32 address, bool bFromZBuffer, boo
 	iter = textures.find(address);
 	if (iter != textures.end())
 	{
-		if (!iter->second.isRenderTarget)
+		if (iter->second.isRenderTarget && iter->second.w == tex_w && iter->second.h == tex_h)
+		{
+			
+			tex = iter->second.texture;
+			iter->second.frameCount = frameCount;
+			goto have_texture;
+		}
+		else
 		{
 			// Remove it and recreate it as a render target
 			iter->second.texture->Release();
 			iter->second.texture = 0;
 			textures.erase(iter);
 		}
-		else
-		{
-			tex = iter->second.texture;
-			iter->second.frameCount = frameCount;
-			goto have_texture;
-		}
 	}
 
 	{
@@ -293,14 +302,16 @@ void TextureCache::CopyRenderTargetToTexture(u32 address, bool bFromZBuffer, boo
 		entry.frameCount = frameCount;
 		entry.w = tex_w;
 		entry.h = tex_h;
+		entry.fmt = copyfmt;
 		
-		D3D::dev->CreateTexture(tex_w, tex_h, 1, D3DUSAGE_RENDERTARGET, D3DFMT_A8R8G8B8, D3DPOOL_DEFAULT, &entry.texture, 0);
+		hr = D3D::dev->CreateTexture(tex_w, tex_h, 1, D3DUSAGE_RENDERTARGET, D3DFMT_A8R8G8B8, D3DPOOL_DEFAULT, &entry.texture, 0);
+		CHECK(hr);
 		textures[address] = entry;
 		tex = entry.texture;
 	}
 
 have_texture:
-	TargetRectangle targetSource = Renderer::ConvertEFBRectangle(source_rect);
+	/*TargetRectangle targetSource = Renderer::ConvertEFBRectangle(source_rect);
 	RECT source_rc;
 	source_rc.left = targetSource.left;
 	source_rc.top = targetSource.top;
@@ -317,8 +328,162 @@ have_texture:
 	srcSurface = FBManager::GetEFBColorRTSurface();
 	D3D::dev->StretchRect(srcSurface, &source_rc, destSurface, &dest_rc, D3DTEXF_LINEAR);
 	destSurface->Release();
+	return;*/
+	float colmat[16]= {0.0f};
+	float fConstAdd[4] = {0.0f};	
 
-	DEBUGGER_LOG_AT((NEXT_XFB_CMD|NEXT_EFB_CMD|NEXT_FRAME|NEXT_NEW_TEXTURE|NEXT_FLUSH),
+    if (bFromZBuffer) 
-	{printf("StretchRect, EFB (%d,%d) -> Texture at 0x%08X (%d,%d)\n",efb_w,efb_h,address,tex_w,tex_h);});
+	{
-	DEBUGGER_PAUSE_AT((NEXT_EFB_CMD|NEXT_NEW_TEXTURE),false);
+        switch(copyfmt) 
+		{
+            case 0: // Z4
+            case 1: // Z8
+                colmat[2] = colmat[6] = colmat[10] = colmat[14] = 1;
+                break;
+            
+            case 3: // Z16 //?
+                colmat[1] = colmat[5] = colmat[9] = colmat[14] = 1;
+            case 11: // Z16 (reverse order)
+                colmat[2] = colmat[6] = colmat[10] = colmat[13] = 1;
+                break;
+            case 6: // Z24X8
+                colmat[0] = 1;
+                colmat[5] = 1;
+                colmat[10] = 1;
+				colmat[15] = 1;
+                break;
+            case 9: // Z8M
+                colmat[1] = colmat[5] = colmat[9] = colmat[13] = 1;
+                break;
+            case 10: // Z8L
+                colmat[0] = colmat[4] = colmat[8] = colmat[12] = 1;
+                break;
+            case 12: // Z16L
+                colmat[0] = colmat[4] = colmat[8] = colmat[13] = 1;
+                break;
+            default:
+                ERROR_LOG(VIDEO, "Unknown copy zbuf format: 0x%x", copyfmt);
+                colmat[0] = colmat[5] = colmat[10] = colmat[15] = 1;
+                break;
+        }
+    }
+    else if (bIsIntensityFmt) 
+	{
+        fConstAdd[0] = fConstAdd[1] = fConstAdd[2] = 16.0f/255.0f;
+        switch (copyfmt) 
+		{
+            case 0: // I4
+            case 1: // I8
+            case 2: // IA4
+            case 3: // IA8
+				// TODO - verify these coefficients
+                colmat[0] = 0.257f; colmat[1] = 0.504f; colmat[2] = 0.098f;
+                colmat[4] = 0.257f; colmat[5] = 0.504f; colmat[6] = 0.098f;
+                colmat[8] = 0.257f; colmat[9] = 0.504f; colmat[10] = 0.098f;
+
+                if (copyfmt < 2) 
+				{
+                    fConstAdd[3] = 16.0f / 255.0f;
+                    colmat[12] = 0.257f; colmat[13] = 0.504f; colmat[14] = 0.098f;
+                }
+                else// alpha
+                    colmat[15] = 1;
+
+                break;
+            default:
+                ERROR_LOG(VIDEO, "Unknown copy intensity format: 0x%x", copyfmt);
+                colmat[0] = colmat[5] = colmat[10] = colmat[15] = 1;
+                break;
+        }
+    }
+    else 
+	{
+        switch (copyfmt) 
+		{
+            case 0: // R4
+            case 8: // R8
+                colmat[0] = colmat[4] = colmat[8] = colmat[12] = 1;
+                break;
+            case 2: // RA4
+            case 3: // RA8
+                colmat[0] = colmat[4] = colmat[8] = colmat[15] = 1;
+                break;
+
+            case 7: // A8
+                colmat[3] = colmat[7] = colmat[11] = colmat[15] = 1; 
+                break;
+            case 9: // G8
+                colmat[1] = colmat[5] = colmat[9] = colmat[13] = 1;
+                break;
+            case 10: // B8
+                colmat[2] = colmat[6] = colmat[10] = colmat[14] = 1;
+                break;
+            case 11: // RG8
+                colmat[0] = colmat[4] = colmat[8] = colmat[13] = 1;
+                break;
+            case 12: // GB8
+                colmat[1] = colmat[5] = colmat[9] = colmat[14] = 1;
+                break;
+
+            case 4: // RGB565
+                colmat[0] = colmat[5] = colmat[10] = 1;
+                fConstAdd[3] = 1; // set alpha to 1
+                break;
+            case 5: // RGB5A3
+            case 6: // RGBA8
+                colmat[0] = colmat[5] = colmat[10] = colmat[15] = 1;
+                break;
+
+            default:
+                ERROR_LOG(VIDEO, "Unknown copy color format: 0x%x", copyfmt);
+                colmat[0] = colmat[5] = colmat[10] = colmat[15] = 1;
+                break;
+        }
+    }
+	// Make sure to resolve anything we need to read from.
+	LPDIRECT3DTEXTURE9 read_texture = bFromZBuffer ? FBManager::GetEFBDepthTexture(source_rect) : FBManager::GetEFBColorTexture(source_rect);
+    
+    // We have to run a pixel shader, for color conversion.
+    Renderer::ResetAPIState(); // reset any game specific settings
+	LPDIRECT3DSURFACE9 Rendersurf = NULL;
+	hr = tex->GetSurfaceLevel(0,&Rendersurf);
+	CHECK(hr);
+	D3D::dev->SetDepthStencilSurface(NULL);
+	D3D::dev->SetRenderTarget(1, NULL);
+	D3D::dev->SetRenderTarget(0, Rendersurf);		
+    
+	D3DVIEWPORT9 vp;
+
+	// Stretch picture with increased internal resolution
+	vp.X = 0;
+	vp.Y = 0;
+	vp.Width  = tex_w;
+	vp.Height = tex_h;
+	vp.MinZ = 0.0f;
+	vp.MaxZ = 1.0f;
+	hr = D3D::dev->SetViewport(&vp);
+	CHECK(hr);
+	RECT destrect;
+	destrect.bottom = tex_h;
+	destrect.left = 0;
+	destrect.right = tex_w;
+	destrect.top = 0;
+	
+
+	PixelShaderManager::SetColorMatrix(colmat, fConstAdd); // set transformation
+	//TargetRectangle targetSource = Renderer::ConvertEFBRectangle(source_rect);
+	RECT sourcerect;
+	sourcerect.bottom = source_rect.bottom;
+	sourcerect.left = source_rect.left;
+	sourcerect.right = source_rect.right;
+	sourcerect.top = source_rect.top;
+
+	D3D::drawShadedTexQuad(read_texture,&sourcerect, EFB_WIDTH , EFB_HEIGHT,&destrect,PixelShaderCache::GetColorMatrixProgram(),NULL);	
+
+	D3D::dev->SetRenderTarget(0, FBManager::GetEFBColorRTSurface());
+	D3D::dev->SetRenderTarget(1, FBManager::GetEFBDepthEncodedSurface());
+	D3D::dev->SetDepthStencilSurface(FBManager::GetEFBDepthRTSurface());
+	VertexShaderManager::SetViewportChanged();
+	Renderer::RestoreAPIState();	
+	Rendersurf->Release();
 }

Source/Plugins/Plugin_VideoDX9/Src/VertexShaderCache.cpp

@@ -35,6 +35,13 @@ const VertexShaderCache::VSCacheEntry *VertexShaderCache::last_entry;
 
 static float GC_ALIGNED16(lastVSconstants[C_FOGPARAMS+8][4]);
 
+static LPDIRECT3DVERTEXSHADER9 SimpleVertexSahder;
+
+LPDIRECT3DVERTEXSHADER9 VertexShaderCache::GetSimpleVertexSahder()
+{
+	return SimpleVertexSahder;
+}
+
 void SetVSConstant4f(int const_number, float f1, float f2, float f3, float f4)
 {
 	if (lastVSconstants[const_number][0] != f1 || 
@@ -121,6 +128,21 @@ void SetMultiVSConstant4fv(int const_number, int count, const float *f)
 
 void VertexShaderCache::Init()
 {
+	char vSimpleProg[1024];
+	sprintf(vSimpleProg,"struct VSOUTPUT\n"
+						"{\n"
+						   "float4 vPosition   : POSITION;\n"
+						   "float4 vTexCoord   : TEXCOORD0;\n"
+						"};\n"
+						"VSOUTPUT main( float4 inPosition : POSITION, float4 inUV : TEXCOORD0)\n"
+						"{\n"
+						   "VSOUTPUT OUT = (VSOUTPUT)0;\n"
+						   "OUT.vPosition = inPosition;\n"
+						   "OUT.vTexCoord = inUV;\n"
+						   "return OUT;\n"
+						"}\n");
+
+	SimpleVertexSahder = D3D::CompileVertexShader(vSimpleProg, (int)strlen(vSimpleProg));
 	Clear();
 }
 
@@ -138,6 +160,8 @@ void VertexShaderCache::Clear()
 
 void VertexShaderCache::Shutdown()
 {
+	if(SimpleVertexSahder)
+		SimpleVertexSahder->Release();
 	Clear();
 }
 

Source/Plugins/Plugin_VideoDX9/Src/VertexShaderCache.h

@@ -56,6 +56,7 @@ public:
 	static void Cleanup();
 	static void Shutdown();
 	static bool SetShader(u32 components);
+	static LPDIRECT3DVERTEXSHADER9 GetSimpleVertexSahder();
 #if defined(_DEBUG) || defined(DEBUGFAST)
 	static std::string GetCurrentShaderCode();
 #endif