initial implementation of fog range adjust, I don't think is correct or work right but is a start.

some tweaks and fixes.

git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6957 8ced0084-cf51-0410-be5f-012b33b47a6e

7 files changed, 83 insertions, 26 deletions

diff --git a/Source/Core/VideoCommon/Src/BPMemory.h b/Source/Core/VideoCommon/Src/BPMemory.h
index abafe201d2..422b515dcc 100644
--- a/Source/Core/VideoCommon/Src/BPMemory.h
+++ b/Source/Core/VideoCommon/Src/BPMemory.h
@@ -678,6 +678,33 @@ union FogParam3
     u32 hex;
 };
 
+union FogRangeKElement
+{
+	struct
+	{
+		u32 HI : 12;
+		u32 LO : 12;
+		u32 regid : 8;
+	};
+	u32 HEX;
+};
+
+struct FogRangeParams
+{
+	union RangeBase
+	{
+		struct
+		{
+			u32 Center : 10;
+			u32 Enabled : 1;
+			u32 unused : 13;
+			u32 regid : 8;
+		};
+		u32 hex;
+	};
+	RangeBase Base;
+	FogRangeKElement K[5];
+};
 // final eq: ze = A/(B_MAG - (Zs>>B_SHF));
 struct FogParams
 {
@@ -902,10 +929,7 @@ struct BPMemory
     FourTexUnits tex[2]; //80-bf
     TevStageCombiner combiners[16]; //0xC0-0xDF
     TevReg tevregs[4];  //0xE0
-    u32 fogRangeAdj;   //0xE8
-    u32 unknown15[3];  //0xe9,0xea,0xeb - fog related
-    u32 tev_range_adj_c; //0xec - screenx center for range adjustment, range adjustment enable
-    u32 tev_range_adj_k; //0xed - specifies range adjustment function = sqrt(x*x+k*k)/k
+    FogRangeParams fogRange;
     FogParams fog; //0xEE,0xEF,0xF0,0xF1,0xF2
     AlphaFunc alphaFunc; //0xF3
     ZTex1 ztex1; //0xf4,0xf5
diff --git a/Source/Core/VideoCommon/Src/BPStructs.cpp b/Source/Core/VideoCommon/Src/BPStructs.cpp
index 9a8e5af10c..17ca146653 100644
--- a/Source/Core/VideoCommon/Src/BPStructs.cpp
+++ b/Source/Core/VideoCommon/Src/BPStructs.cpp
@@ -320,6 +320,9 @@ void BPWritten(const BPCmd& bp)
 	case BPMEM_FOGRANGE+3:
 	case BPMEM_FOGRANGE+4:
 	case BPMEM_FOGRANGE+5:
+		if (!GetConfig(CONFIG_DISABLEFOG))
+			PixelShaderManager::SetFogRangeAdjustChanged();
+		break;
 	case BPMEM_FOGPARAM0:
 	case BPMEM_FOGBMAGNITUDE:
 	case BPMEM_FOGBEXPONENT:
diff --git a/Source/Core/VideoCommon/Src/PixelShaderGen.cpp b/Source/Core/VideoCommon/Src/PixelShaderGen.cpp
index 8c3805df9a..f5f9c84f46 100644
--- a/Source/Core/VideoCommon/Src/PixelShaderGen.cpp
+++ b/Source/Core/VideoCommon/Src/PixelShaderGen.cpp
@@ -61,7 +61,7 @@ void GetPixelShaderId(PIXELSHADERUID *uid, DSTALPHA_MODE dstAlphaMode)
 
 	uid->values[2] = (u32)bpmem.fog.c_proj_fsel.fsel |
 				   ((u32)bpmem.fog.c_proj_fsel.proj << 3) |
-				   ((u32)enableZTexture << 4);
+				   ((u32)enableZTexture << 4) | ((u32)bpmem.fogRange.Base.Enabled << 5);
 
 	if(g_ActiveConfig.bEnablePixelLigting && g_ActiveConfig.backend_info.bSupportsPixelLighting)
 	{
@@ -504,7 +504,7 @@ const char *GeneratePixelShaderCode(DSTALPHA_MODE dstAlphaMode, API_TYPE ApiType
 	WRITE(p, "uniform float4 "I_ZBIAS"[2] : register(c%d);\n", C_ZBIAS);
 	WRITE(p, "uniform float4 "I_INDTEXSCALE"[2] : register(c%d);\n", C_INDTEXSCALE);
 	WRITE(p, "uniform float4 "I_INDTEXMTX"[6] : register(c%d);\n", C_INDTEXMTX);
-	WRITE(p, "uniform float4 "I_FOG"[2] : register(c%d);\n", C_FOG);
+	WRITE(p, "uniform float4 "I_FOG"[3] : register(c%d);\n", C_FOG);
 
 	if(g_ActiveConfig.bEnablePixelLigting && g_ActiveConfig.backend_info.bSupportsPixelLighting)
 	{
@@ -521,7 +521,7 @@ const char *GeneratePixelShaderCode(DSTALPHA_MODE dstAlphaMode, API_TYPE ApiType
 		WRITE(p, "  out float4 ocol0 : COLOR0,%s%s\n  in float4 rawpos : %s,\n",
 			dstAlphaMode == DSTALPHA_DUAL_SOURCE_BLEND ? "\n  out float4 ocol1 : COLOR1," : "",
 			DepthTextureEnable ? "\n  out float depth : DEPTH," : "",
-			ApiType == API_OPENGL ? "WPOS" : "POSITION");
+			ApiType == API_OPENGL ? "WPOS" : "VPOS");
 	}
 	else
 	{
@@ -792,7 +792,7 @@ const char *GeneratePixelShaderCode(DSTALPHA_MODE dstAlphaMode, API_TYPE ApiType
 	}
 	// emulation of unisgned 8 overflow when casting if needed
 	if(RegisterStates[0].AlphaNeedOverflowControl || RegisterStates[0].ColorNeedOverflowControl)
-		WRITE(p, "prev = frac(4.0f + prev * (255.0f/256.0f)) * (256.0f/255.0f);\n");
+		WRITE(p, "prev = frac(prev * (255.0f/256.0f)) * (256.0f/255.0f);\n");
 	
 	if (!WriteAlphaTest(p, ApiType, dstAlphaMode))
 	{
@@ -997,7 +997,7 @@ static void WriteStage(char *&p, int n, API_TYPE ApiType)
 	if(bCRas || bARas)
 	{
 		WRITE(p, "rastemp = %s.%s;\n", tevRasTable[bpmem.tevorders[n / 2].getColorChan(n & 1)], rasswap);
-		WRITE(p, "crastemp = frac(4.0f + rastemp * (255.0f/256.0f)) * (256.0f/255.0f);\n");
+		WRITE(p, "crastemp = frac(rastemp * (255.0f/256.0f)) * (256.0f/255.0f);\n");
 	}
 
 
@@ -1030,7 +1030,7 @@ static void WriteStage(char *&p, int n, API_TYPE ApiType)
 		WRITE(p, "konsttemp = float4(%s, %s);\n", tevKSelTableC[kc], tevKSelTableA[ka]);
 		if(kc > 7 || ka > 7)
 		{
-			WRITE(p, "ckonsttemp = frac(4.0f + konsttemp * (255.0f/256.0f)) * (256.0f/255.0f);\n");
+			WRITE(p, "ckonsttemp = frac(konsttemp * (255.0f/256.0f)) * (256.0f/255.0f);\n");
 		}
 		else
 		{
@@ -1050,7 +1050,7 @@ static void WriteStage(char *&p, int n, API_TYPE ApiType)
 	{
 		if(RegisterStates[0].AlphaNeedOverflowControl || RegisterStates[0].ColorNeedOverflowControl)
 		{
-			WRITE(p, "cprev = frac(4.0f + prev * (255.0f/256.0f)) * (256.0f/255.0f);\n");
+			WRITE(p, "cprev = frac(prev * (255.0f/256.0f)) * (256.0f/255.0f);\n");
 			RegisterStates[0].AlphaNeedOverflowControl = false;
 			RegisterStates[0].ColorNeedOverflowControl = false;
 		}
@@ -1073,7 +1073,7 @@ static void WriteStage(char *&p, int n, API_TYPE ApiType)
 	{
 		if(RegisterStates[1].AlphaNeedOverflowControl || RegisterStates[1].ColorNeedOverflowControl)
 		{
-			WRITE(p, "cc0 = frac(4.0f + c0 * (255.0f/256.0f)) * (256.0f/255.0f);\n");
+			WRITE(p, "cc0 = frac(c0 * (255.0f/256.0f)) * (256.0f/255.0f);\n");
 			RegisterStates[1].AlphaNeedOverflowControl = false;
 			RegisterStates[1].ColorNeedOverflowControl = false;
 		}
@@ -1096,7 +1096,7 @@ static void WriteStage(char *&p, int n, API_TYPE ApiType)
 	{
 		if(RegisterStates[2].AlphaNeedOverflowControl || RegisterStates[2].ColorNeedOverflowControl)
 		{
-			WRITE(p, "cc1 = frac(4.0f + c1 * (255.0f/256.0f)) * (256.0f/255.0f);\n");
+			WRITE(p, "cc1 = frac(c1 * (255.0f/256.0f)) * (256.0f/255.0f);\n");
 			RegisterStates[2].AlphaNeedOverflowControl = false;
 			RegisterStates[2].ColorNeedOverflowControl = false;
 		}
@@ -1119,7 +1119,7 @@ static void WriteStage(char *&p, int n, API_TYPE ApiType)
 	{
 		if(RegisterStates[3].AlphaNeedOverflowControl || RegisterStates[3].ColorNeedOverflowControl)
 		{
-			WRITE(p, "cc2 = frac(4.0f + c2 * (255.0f/256.0f)) * (256.0f/255.0f);\n");
+			WRITE(p, "cc2 = frac(c2 * (255.0f/256.0f)) * (256.0f/255.0f);\n");
 			RegisterStates[3].AlphaNeedOverflowControl = false;
 			RegisterStates[3].ColorNeedOverflowControl = false;
 		}
@@ -1340,10 +1340,15 @@ static void WriteFog(char *&p)
 		WRITE (p, "  float ze = "I_FOG"[1].x * zCoord;\n");
 	}
 	
-	// stuff to do!
-	// here, where we'll have to add/handle x range adjustment (if related BP register it's enabled)
 	// x_adjust = sqrt((x-center)^2 + k^2)/k
 	// ze *= x_adjust
+	//this is complitly teorical as the real hard seems to use a table intead of calculate the values.
+	if(bpmem.fogRange.Base.Enabled)
+	{
+		WRITE (p, "  float x_adjust = (2.0f * (clipPos.x / "I_FOG"[2].y)) - 1.0f - "I_FOG"[2].x;\n");
+		WRITE (p, "  x_adjust = sqrt(x_adjust * x_adjust + "I_FOG"[2].z * "I_FOG"[2].z) / "I_FOG"[2].z;\n");
+		WRITE (p, "  ze *= x_adjust;\n");
+	}
 
 	WRITE (p, "  float fog = saturate(ze - "I_FOG"[1].z);\n");
 
@@ -1359,4 +1364,5 @@ static void WriteFog(char *&p)
 
 	WRITE(p, "  prev.rgb = lerp(prev.rgb,"I_FOG"[0].rgb,fog);\n");
 	
+	
 }
diff --git a/Source/Core/VideoCommon/Src/PixelShaderGen.h b/Source/Core/VideoCommon/Src/PixelShaderGen.h
index e2d8ca847e..10c39ad34e 100644
--- a/Source/Core/VideoCommon/Src/PixelShaderGen.h
+++ b/Source/Core/VideoCommon/Src/PixelShaderGen.h
@@ -41,7 +41,7 @@
 #define C_INDTEXMTX		(C_INDTEXSCALE + 2)		//21
 #define C_FOG			(C_INDTEXMTX + 6)		//27
 
-#define C_PLIGHTS		(C_FOG + 2)
+#define C_PLIGHTS		(C_FOG + 3)
 #define C_PMATERIALS	(C_PLIGHTS + 40)
 #define C_PENVCONST_END (C_PMATERIALS + 4)
 #define PIXELSHADERUID_MAX_VALUES (5 + 32 + 6 + 11 + 2)
diff --git a/Source/Core/VideoCommon/Src/PixelShaderManager.cpp b/Source/Core/VideoCommon/Src/PixelShaderManager.cpp
index fb484ca0c0..96ba2cb3f2 100644
--- a/Source/Core/VideoCommon/Src/PixelShaderManager.cpp
+++ b/Source/Core/VideoCommon/Src/PixelShaderManager.cpp
@@ -22,6 +22,7 @@
 #include "PixelShaderManager.h"
 #include "VideoCommon.h"
 #include "VideoConfig.h"
+#include "RenderBase.h"
 static float GC_ALIGNED16(s_fMaterials[16]);
 static int s_nColorsChanged[2]; // 0 - regular colors, 1 - k colors
 static int s_nIndTexMtxChanged;
@@ -31,6 +32,7 @@ static bool s_bZTextureTypeChanged;
 static bool s_bDepthRangeChanged;
 static bool s_bFogColorChanged;
 static bool s_bFogParamChanged;
+static bool s_bFogRangeAdjustChanged;
 static int nLightsChanged[2]; // min,max
 static float lastDepthRange[2]; // 0 = far z, 1 = far - near
 static float lastRGBAfull[2][4][4];
@@ -57,7 +59,7 @@ void PixelShaderManager::Dirty()
 	s_nIndTexScaleChanged = 0xFF;
 	s_nIndTexMtxChanged = 15;
 	s_bAlphaChanged = s_bZBiasChanged = s_bZTextureTypeChanged = s_bDepthRangeChanged = true;
-	s_bFogColorChanged = s_bFogParamChanged = true;
+	s_bFogRangeAdjustChanged = s_bFogColorChanged = s_bFogParamChanged = true;
 	nLightsChanged[0] = 0; nLightsChanged[1] = 0x80;
 	nMaterialsChanged = 15;
 }
@@ -214,6 +216,25 @@ void PixelShaderManager::SetConstants()
         s_bFogParamChanged = false;
     }
 
+	if (s_bFogRangeAdjustChanged)
+	{
+		if(!g_ActiveConfig.bDisableFog && bpmem.fogRange.Base.Enabled == 1)
+		{
+			//bpmem.fogRange.Base.Center : center of the viewport in x axis. observation: bpmem.fogRange.Base.Center = realcenter + 342;
+			int center = ((u32)bpmem.fogRange.Base.Center) - 342;
+			// normalice center to make calculations easy
+			float ScreenSpaceCenter = center / (2.0f * xfregs.rawViewport[0]);
+			ScreenSpaceCenter = (ScreenSpaceCenter * 2.0f) - 1.0f;
+			//bpmem.fogRange.K seems to be  a table of precalculated coeficients for the adjust factor
+			//observations: bpmem.fogRange.K[0].LO apears to be the lowest value and bpmem.fogRange.K[4].HI the largest
+			// they always seems to be larger than 256 so my teory is :
+			// they are the coeficients from the center to th e border of the screen
+			// so to simplify i use the hi coeficient as K in the shader taking 256 as the scale
+			SetPSConstant4f(C_FOG + 2, ScreenSpaceCenter, (float)Renderer::EFBToScaledX((int)(2.0f * xfregs.rawViewport[0])), bpmem.fogRange.K[4].HI / 256.0f,0.0f);
+		}		
+		s_bFogRangeAdjustChanged = false;
+	}
+
 	if (g_ActiveConfig.bEnablePixelLigting && g_ActiveConfig.backend_info.bSupportsPixelLighting && nLightsChanged[0] >= 0) // config check added because the code in here was crashing for me inside SetPSConstant4f
 	{
 		// lights don't have a 1 to 1 mapping, the color component needs to be converted to 4 floats
@@ -394,6 +415,11 @@ void PixelShaderManager::SetFogParamChanged()
     s_bFogParamChanged = true;
 }
 
+void PixelShaderManager::SetFogRangeAdjustChanged()
+{
+    s_bFogRangeAdjustChanged = true;
+}
+
 void PixelShaderManager::SetColorMatrix(const float* pmatrix)
 {
 	SetMultiPSConstant4fv(C_COLORMATRIX,7,pmatrix);
diff --git a/Source/Core/VideoCommon/Src/PixelShaderManager.h b/Source/Core/VideoCommon/Src/PixelShaderManager.h
index de46341441..f3428eda6f 100644
--- a/Source/Core/VideoCommon/Src/PixelShaderManager.h
+++ b/Source/Core/VideoCommon/Src/PixelShaderManager.h
@@ -51,9 +51,11 @@ public:
 	static void SetTexCoordChanged(u8 texmapid);
 	static void SetFogColorChanged();
 	static void SetFogParamChanged();
+	static void SetFogRangeAdjustChanged();
 	static void SetColorMatrix(const float* pmatrix);
 	static void InvalidateXFRange(int start, int end);
 	static void SetMaterialColor(int index, u32 data);
+	
 };
 
 
diff --git a/Source/Core/VideoCommon/Src/VideoCommon.h b/Source/Core/VideoCommon/Src/VideoCommon.h
index 5a573369cd..0b29874178 100644
--- a/Source/Core/VideoCommon/Src/VideoCommon.h
+++ b/Source/Core/VideoCommon/Src/VideoCommon.h
@@ -157,14 +157,10 @@ inline u32 RGBA8ToRGBA6ToRGBA8(u32 src)
 
 inline u32 RGBA8ToRGB565ToRGBA8(u32 src)
 {
-	u32 color = src;
-	u32 dstr5 = (color & 0xFF0000) >> 19;
-	u32 dstg6 = (color & 0xFF00) >> 10;
-	u32 dstb5 = (color & 0xFF) >> 3;
-	u32 dstr8 = (dstr5 << 3) | (dstr5 >> 2);
-	u32 dstg8 = (dstg6 << 2) | (dstg6 >> 4);
-	u32 dstb8 = (dstb5 << 3) | (dstb5 >> 2);
-	color = 0xFF000000 | (dstr8 << 16) | (dstg8 << 8) | dstb8;
+	u32 color = (src & 0xF8FCF8);
+	color |= (color >> 5) & 0x070007;
+	color |= (color >> 6) & 0x000300;
+	color |= 0xFF000000;
 	return color;
 }