q3aql/Serious-Engine
q3aql/Serious-Engine
1
0
Fork 0
mirror of https://github.com/ptitSeb/Serious-Engine synced 2024-10-05 23:44:51 +02:00
Code Issues Packages Projects Releases Wiki Activity
88a0ec269b
Serious-Engine/Sources/Engine/Light/LayerMixer.cpp
notaz fd045e5025 fix LayerMixer asm 
- clobber the whole x87 state for mmx (emms alone requires this)
- add all modified registers to clobber list
  (in some cases use dummy output vars instead)
- use symbolic names
- use more relaxed constraints where possible
- allow gcc to allocate ebx replacement reg
2016-04-22 02:28:42 +03:00

2101 lines
72 KiB
C++
Executable File
Raw Blame History

/* Copyright (c) 2002-2012 Croteam Ltd.
This program is free software; you can redistribute it and/or modify
it under the terms of version 2 of the GNU General Public License as published by
the Free Software Foundation
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License along
with this program; if not, write to the Free Software Foundation, Inc.,
51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. */
#include "Engine/StdH.h"
#include <Engine/Brushes/Brush.h>
#include <Engine/Brushes/BrushTransformed.h>
#include <Engine/Light/LightSource.h>
#include <Engine/Light/Gradient.h>
#include <Engine/Base/ListIterator.inl>
#include <Engine/Base/Statistics_Internal.h>
#include <Engine/Graphics/Color.h>
#include <Engine/Math/FixInt.h>
#include <Engine/Entities/Entity.h>
#include <Engine/Graphics/GfxLibrary.h>
#include <Engine/Math/Clipping.inl>
#include <Engine/Light/Shadows_internal.h>
#include <Engine/World/WorldEditingProfile.h>
#include <Engine/Templates/StaticArray.cpp>
#include <Engine/Templates/DynamicArray.cpp>
// asm shortcuts
#define O offset
#define Q qword ptr
#define D dword ptr
#define W word ptr
#define B byte ptr
#if (defined USE_PORTABLE_C)
#define ASMOPT 0
#elif (defined __MSVC_INLINE__)
#define ASMOPT 1
#elif (defined __GNU_INLINE__)
#define ASMOPT 1
#else
#define ASMOPT 0
#endif
extern INDEX shd_bFineQuality;
extern INDEX shd_iFiltering;
extern INDEX shd_iDithering;
extern const UBYTE *pubClipByte;
extern UBYTE aubSqrt[ SQRTTABLESIZE];
extern UWORD auw1oSqrt[SQRTTABLESIZE];
// static FLOAT3D _v00;
// internal class for layer mixing
class CLayerMixer
{
public:
CBrushShadowMap *lm_pbsmShadowMap; // shadow map whose layers are mixed
CBrushPolygon *lm_pbpoPolygon; // polygon of the shadow map
BOOL lm_bDynamic; // set while doing dynamic light mixing
// dimensions of currently processed shadow map
MEX lm_mexOffsetU; // offsets in mex
MEX lm_mexOffsetV;
INDEX lm_iFirstLevel; // mip level of first possible mip-map
INDEX lm_iMipLevel; // mip level
INDEX lm_iMipShift; // relative mip level (miplevel-firstlevel)
PIX lm_pixCanvasSizeU; // shadowmap canvas size in pixels
PIX lm_pixCanvasSizeV;
PIX lm_pixPolygonSizeU; // polygon (used part of shadowmap size in pixels
PIX lm_pixPolygonSizeV;
ULONG*lm_pulShadowMap; // buffer for final shadow map
ULONG*lm_pulStaticShadowMap; // precached static shadow map
// gradients for shadow map walking
FLOAT3D lm_vO; // upper left corner of shadow map in 3D
FLOAT3D lm_vStepU; // step between pixels in same row
FLOAT3D lm_vStepV; // step between rows
FLOAT3D lm_vLightDirection; // light direction for directional light sources
CLightSource *lm_plsLight; // current light source
// color components of current light
COLOR lm_colLight;
COLOR lm_colAmbient;
// constructor
CLayerMixer( CBrushShadowMap *pbsm, INDEX iFirstMip, INDEX iLastMip, BOOL bDynamic);
// remember general data
void CalculateData( CBrushShadowMap *pbsm, INDEX iMipmap);
// mix one mip-map
void MixOneMipmap( CBrushShadowMap *pbsm, INDEX iMipmap);
// mix dynamic lights
void MixOneMipmapDynamic(CBrushShadowMap *pbsm, INDEX iMipmap);
// find start of a mip-map inside a layer
void FindLayerMipmap( CBrushShadowLayer *pbsl, UBYTE *&pub, UBYTE &ubMask);
// add one point layer to the shadow map
void AddAmbientPoint(void);
void AddAmbientMaskPoint( UBYTE *pubMask, UBYTE ubMask);
void AddDiffusionPoint(void);
void AddDiffusionMaskPoint( UBYTE *pubMask, UBYTE ubMask);
BOOL PrepareOneLayerPoint( CBrushShadowLayer *pbsl, BOOL bNoMask);
void AddOneLayerPoint( CBrushShadowLayer *pbsl, UBYTE *pub, UBYTE ubMask=0);
// add one directional layer to the shadow map
void AddDirectional(void);
void AddMaskDirectional( UBYTE *pubMask, UBYTE ubMask);
void AddOneLayerDirectional( CBrushShadowLayer *pbsl, UBYTE *pub, UBYTE ubMask=0);
// add one gradient layer to the shadow map
void AddOneLayerGradient( CGradientParameters &gp);
// add the intensity to the pixel
inline void AddToCluster( UBYTE *pub);
inline void AddAmbientToCluster( UBYTE *pub);
inline void AddToCluster( UBYTE *pub, SLONG slIntensity);
// additional functions
__forceinline void CopyShadowLayer(void);
__forceinline void FillShadowLayer( COLOR col);
// inline FLOAT IntensityAtDistance( FLOAT fDistance, FLOAT fMinDistance);
// FLOAT lm_fLightMax, lm_fLightStep;
};
// increment a byte without overflowing it
static inline void IncrementByteWithClip( UBYTE &ub, SLONG slAdd)
{
ub = pubClipByte[(SLONG)ub+slAdd];
}
// increment a color without overflowing it
static inline void IncrementColorWithClip( UBYTE &ubR, UBYTE &ubG, UBYTE &ubB,
SLONG slR, SLONG slG, SLONG slB)
{
IncrementByteWithClip( ubR, slR);
IncrementByteWithClip( ubG, slG);
IncrementByteWithClip( ubB, slB);
}
// add the intensity to the pixel
inline void CLayerMixer::AddToCluster( UBYTE *pub)
{
IncrementByteWithClip(pub[0], ((UBYTE*)&lm_colLight)[3]);
IncrementByteWithClip(pub[1], ((UBYTE*)&lm_colLight)[2]);
IncrementByteWithClip(pub[2], ((UBYTE*)&lm_colLight)[1]);
}
inline void CLayerMixer::AddAmbientToCluster( UBYTE *pub)
{
IncrementByteWithClip(pub[0], ((UBYTE*)&lm_colAmbient)[3]);
IncrementByteWithClip(pub[1], ((UBYTE*)&lm_colAmbient)[2]);
IncrementByteWithClip(pub[2], ((UBYTE*)&lm_colAmbient)[1]);
}
inline void CLayerMixer::AddToCluster( UBYTE *pub, SLONG slIntensity)
{
IncrementByteWithClip(pub[0], (long) (((UBYTE*)&lm_colLight)[3] *slIntensity)>>16);
IncrementByteWithClip(pub[1], (long) (((UBYTE*)&lm_colLight)[2] *slIntensity)>>16);
IncrementByteWithClip(pub[2], (long) (((UBYTE*)&lm_colLight)[1] *slIntensity)>>16);
}
// remember general data
void CLayerMixer::CalculateData( CBrushShadowMap *pbsm, INDEX iMipmap)
{
_pfWorldEditingProfile.StartTimer(CWorldEditingProfile::PTI_CALCULATEDATA);
// cache class vars
lm_pbsmShadowMap = pbsm;
lm_pbpoPolygon = pbsm->GetBrushPolygon();
lm_mexOffsetU = pbsm->sm_mexOffsetX;
lm_mexOffsetV = pbsm->sm_mexOffsetY;
lm_iFirstLevel = pbsm->sm_iFirstMipLevel;
lm_iMipLevel = iMipmap;
lm_iMipShift = lm_iMipLevel - lm_iFirstLevel;
lm_pixCanvasSizeU = pbsm->sm_mexWidth >>lm_iMipLevel;
lm_pixCanvasSizeV = pbsm->sm_mexHeight>>lm_iMipLevel;
lm_pixPolygonSizeU = Min( lm_pixCanvasSizeU, (PIX)(lm_pbsmShadowMap->sm_pixPolygonSizeU >>lm_iMipShift)+1);
lm_pixPolygonSizeV = Min( lm_pixCanvasSizeV, (PIX)(lm_pbsmShadowMap->sm_pixPolygonSizeV >>lm_iMipShift)+1);
// determine where this mip-map is relative to the allocated shadow map memory
PIX pixOffset = pbsm->sm_slMemoryUsed/BYTES_PER_TEXEL
- GetMipmapOffset( 15, lm_pixCanvasSizeU, lm_pixCanvasSizeV);
// get right pointers to the shadow mipmap
if( lm_bDynamic) {
lm_pulShadowMap = pbsm->sm_pulDynamicShadowMap + pixOffset;
lm_pulStaticShadowMap = pbsm->sm_pulCachedShadowMap + pixOffset;
} else {
lm_pulShadowMap = pbsm->sm_pulCachedShadowMap + pixOffset;
lm_pulStaticShadowMap = NULL;
}
// prepare 3D positions
CEntity *penWithPolygon = lm_pbpoPolygon->bpo_pbscSector->bsc_pbmBrushMip->bm_pbrBrush->br_penEntity;
ASSERT(penWithPolygon!=NULL);
const FLOATmatrix3D &mPolygonRotation = penWithPolygon->en_mRotation;
const FLOAT3D &vPolygonTranslation = penWithPolygon->GetPlacement().pl_PositionVector;
// get first pixel in texture in 3D
Vector<MEX, 2> vmex0;
vmex0(1) = -lm_mexOffsetU+(1<<(lm_iMipLevel-1));
vmex0(2) = -lm_mexOffsetV+(1<<(lm_iMipLevel-1));
lm_pbpoPolygon->bpo_mdShadow.GetSpaceCoordinates(
lm_pbpoPolygon->bpo_pbplPlane->bpl_pwplWorking->wpl_mvRelative, vmex0, lm_vO);
lm_vO = lm_vO*mPolygonRotation+vPolygonTranslation;
// get steps for walking in texture in 3D
Vector<MEX, 2> vmexU, vmexV;
vmexU(1) = (1<<lm_iMipLevel)-lm_mexOffsetU+(1<<(lm_iMipLevel-1));
vmexU(2) = (0<<lm_iMipLevel)-lm_mexOffsetV+(1<<(lm_iMipLevel-1));
vmexV(1) = (0<<lm_iMipLevel)-lm_mexOffsetU+(1<<(lm_iMipLevel-1));
vmexV(2) = (1<<lm_iMipLevel)-lm_mexOffsetV+(1<<(lm_iMipLevel-1));
lm_pbpoPolygon->bpo_mdShadow.GetSpaceCoordinates(
lm_pbpoPolygon->bpo_pbplPlane->bpl_pwplWorking->wpl_mvRelative, vmexU, lm_vStepU);
lm_vStepU = lm_vStepU*mPolygonRotation+vPolygonTranslation;
lm_pbpoPolygon->bpo_mdShadow.GetSpaceCoordinates(
lm_pbpoPolygon->bpo_pbplPlane->bpl_pwplWorking->wpl_mvRelative, vmexV, lm_vStepV);
lm_vStepV = lm_vStepV*mPolygonRotation+vPolygonTranslation;
lm_vStepU-= lm_vO;
lm_vStepV-= lm_vO;
ASSERT( lm_pixPolygonSizeU>0 && lm_pixPolygonSizeV>0);
_pfWorldEditingProfile.StopTimer(CWorldEditingProfile::PTI_CALCULATEDATA);
}
// find start of a mip-map inside a layer
void CLayerMixer::FindLayerMipmap( CBrushShadowLayer *pbsl, UBYTE *&pub, UBYTE &ubMask)
{
struct MipmapTable mmtLayer;
// find mip-mapping information for the layer
MakeMipmapTable(pbsl->bsl_pixSizeU, pbsl->bsl_pixSizeV, mmtLayer);
// get pixel offset of the mipmap
SLONG slPixOffset = mmtLayer.mmt_aslOffsets[lm_iMipLevel-lm_iFirstLevel];
// convert offset to bits
pub = pbsl->bsl_pubLayer + (slPixOffset>>3);
ubMask = 1<<(slPixOffset&7);
}
// BEWARE: Changing these #defines WILL screw up the GNU C inline asm...
#define FTOX 0x10000000
#define SHIFTX (28-SQRTTABLESIZELOG2)
// variables for easier transfers
const FLOAT3D *_vLight;
FLOAT _fMinLightDistance;
FLOAT _f1oFallOff;
INDEX _iPixCt;
INDEX _iRowCt;
SLONG _slModulo;
ULONG _ulLightFlags;
ULONG _ulPolyFlags;
SLONG _slL2Row;
SLONG _slDDL2oDU;
SLONG _slDDL2oDV;
SLONG _slDDL2oDUoDV;
SLONG _slDL2oDURow;
SLONG _slDL2oDV;
SLONG _slLightMax;
SLONG _slHotSpot;
SLONG _slLightStep;
ULONG *_pulLayer;
// !!! FIXME : rcg01072001 These statics are a pain in the ass.
extern "C" {
__int64 mmDDL2oDU_AddAmbientPoint;
__int64 mmDDL2oDV_AddAmbientPoint;
}
// add one layer point light without diffusion and mask
void CLayerMixer::AddAmbientPoint(void)
{
// prepare some local variables
mmDDL2oDU_AddAmbientPoint = _slDDL2oDU;
mmDDL2oDV_AddAmbientPoint = _slDDL2oDV;
ULONG ulLightRGB = ByteSwap(lm_colLight);
_slLightMax<<=7;
_slLightStep>>=1;
#if (ASMOPT == 1)
#if (defined __MSVC_INLINE__)
__asm {
// prepare interpolants
movd mm0,D [_slL2Row]
movd mm1,D [_slDL2oDURow]
psllq mm1,32
por mm1,mm0 // MM1 = slDL2oDURow | slL2Row
movd mm0,D [_slDL2oDV]
movd mm2,D [_slDDL2oDUoDV]
psllq mm2,32
por mm2,mm0 // MM2 = slDDL2oDUoDV | slDL2oDV
// prepare color
pxor mm0,mm0
movd mm7,D [ulLightRGB]
punpcklbw mm7,mm0
psllw mm7,1
// loop thru rows
mov edi,D [_pulLayer]
mov ebx,D [_iRowCt]
rowLoop:
push ebx
movd ebx,mm1 // EBX = slL2Point
movq mm3,mm1
psrlq mm3,32 // MM3 = 0 | slDL2oDU
// loop thru pixels in current row
mov ecx,D [_iPixCt]
pixLoop:
// check if pixel need to be drawn
cmp ebx,FTOX
jge skipPixel
// calculate intensities and do actual drawing of shadow pixel ARGB
movd mm4,ecx
mov eax,ebx
sar eax,SHIFTX
and eax,(SQRTTABLESIZE-1)
movzx eax,B aubSqrt[eax]
mov ecx,D [_slLightMax]
cmp eax,D [_slHotSpot]
jle skipInterpolation
mov ecx,255
sub ecx,eax
imul ecx,D [_slLightStep]
skipInterpolation:
// calculate rgb pixel to add
movd mm6,ecx
punpcklwd mm6,mm6
punpckldq mm6,mm6
pmulhw mm6,mm7
// add dynamic light pixel to underlying pixel
movd mm5,D [edi]
punpcklbw mm5,mm0
paddw mm5,mm6
packuswb mm5,mm0
movd D [edi],mm5
movd ecx,mm4
skipPixel:
// advance to next pixel
add edi,4
movd eax,mm3
add ebx,eax
paddd mm3,Q [mmDDL2oDU_AddAmbientPoint]
dec ecx
jnz pixLoop
// advance to the next row
pop ebx
add edi,D [_slModulo]
paddd mm1,mm2
paddd mm2,Q [mmDDL2oDV_AddAmbientPoint]
dec ebx
jnz rowLoop
emms
}
#elif (defined __GNU_INLINE__)
ULONG tmp1, tmp2;
__asm__ __volatile__ (
// prepare interpolants
"movd (" ASMSYM(_slL2Row) "), %%mm0 \n\t"
"movd (" ASMSYM(_slDL2oDURow) "), %%mm1 \n\t"
"psllq $32, %%mm1 \n\t"
"por %%mm0, %%mm1 \n\t" // MM1 = slDL2oDURow | slL2Row
"movd (" ASMSYM(_slDL2oDV) "), %%mm0 \n\t"
"movd (" ASMSYM(_slDDL2oDUoDV) "), %%mm2 \n\t"
"psllq $32, %%mm2 \n\t"
"por %%mm0, %%mm2 \n\t" // MM2 = slDDL2oDUoDV | slDL2oDV
// prepare color
"pxor %%mm0, %%mm0 \n\t"
"movd %[ulLightRGB], %%mm7 \n\t"
"punpcklbw %%mm0, %%mm7 \n\t"
"psllw $1, %%mm7 \n\t"
// loop thru rows
"movl (" ASMSYM(_pulLayer) "), %%edi \n\t"
"movl (" ASMSYM(_iRowCt) "), %[xbx] \n\t"
"0: \n\t" // rowLoop
"movd %%mm1, %[slL2Point] \n\t"
"movq %%mm1, %%mm3 \n\t"
"psrlq $32, %%mm3 \n\t" // MM3 = 0 | slDL2oDU
// loop thru pixels in current row
"movl (" ASMSYM(_iPixCt) "), %%ecx \n\t"
"1: \n\t" // pixLoop
// check if pixel need to be drawn
"cmpl $0x10000000, %[slL2Point] \n\t"
"jge 3f \n\t" // skipPixel
// calculate intensities and do actual drawing of shadow pixel ARGB
"movd %%ecx, %%mm4 \n\t"
"movl %[slL2Point], %%eax \n\t"
"sarl $15, %%eax \n\t"
"andl $8191, %%eax \n\t"
"movzbl " ASMSYM(aubSqrt) "(%%eax), %%eax \n\t"
"movl (" ASMSYM(_slLightMax) "), %%ecx \n\t"
"cmpl (" ASMSYM(_slHotSpot) "), %%eax \n\t"
"jle 2f \n\t" // skipInterpolation
"movl $255, %%ecx \n\t"
"subl %%eax, %%ecx \n\t"
"imull (" ASMSYM(_slLightStep) "), %%ecx \n\t"
"2: \n\t" // skipInterpolation
// calculate rgb pixel to add
"movd %%ecx, %%mm6 \n\t"
"punpcklwd %%mm6, %%mm6 \n\t"
"punpckldq %%mm6, %%mm6 \n\t"
"pmulhw %%mm7, %%mm6 \n\t"
// add dynamic light pixel to underlying pixel
"movd (%%edi), %%mm5 \n\t"
"punpcklbw %%mm0, %%mm5 \n\t"
"paddw %%mm6, %%mm5 \n\t"
"packuswb %%mm0, %%mm5 \n\t"
"movd %%mm5, (%%edi) \n\t"
"movd %%mm4, %%ecx \n\t"
"3: \n\t" // skipPixel
// advance to next pixel
"addl $4, %%edi \n\t"
"movd %%mm3, %%eax \n\t"
"addl %%eax, %[slL2Point] \n\t"
"paddd (" ASMSYM(mmDDL2oDU_AddAmbientPoint) "), %%mm3 \n\t"
"decl %%ecx \n\t"
"jnz 1b \n\t" // pixLoop
// advance to the next row
"addl (" ASMSYM(_slModulo) "), %%edi \n\t"
"paddd %%mm2, %%mm1 \n\t"
"paddd (" ASMSYM(mmDDL2oDV_AddAmbientPoint) "), %%mm2 \n\t"
"decl %[xbx] \n\t"
"jnz 0b \n\t" // rowLoop
"emms \n\t"
: [xbx] "=&r" (tmp1), [slL2Point] "=&g" (tmp2)
: [ulLightRGB] "g" (ulLightRGB)
: FPU_REGS, MMX_REGS, "eax", "ecx", "edi", "cc", "memory"
);
#else
#error Write inline asm for your platform.
#endif
#else
// !!! FIXME WARNING: I have not checked this code, and it could be
// !!! FIXME totally and utterly wrong. --ryan.
// STUBBED("may not work");
UBYTE* pubLayer = (UBYTE*)_pulLayer;
for( PIX pixV=0; pixV<_iRowCt; pixV++)
{
SLONG slL2Point = _slL2Row;
SLONG slDL2oDU = _slDL2oDURow;
for( PIX pixU=0; pixU<_iPixCt; pixU++)
{
// if the point is not masked
if( slL2Point < FTOX ) {
SLONG slL = (slL2Point>>SHIFTX)&(SQRTTABLESIZE-1); // and is just for degenerate cases
SLONG slIntensity = _slLightMax;
slL = aubSqrt[slL];
if( slL>_slHotSpot) slIntensity = ((255-slL)*_slLightStep);
// add the intensity to the pixel
AddToCluster( pubLayer, slIntensity);
}
// go to the next pixel
pubLayer+=4;
slL2Point += slDL2oDU;
slDL2oDU += _slDDL2oDU;
}
// go to the next row
pubLayer += _slModulo;
_slL2Row += _slDL2oDV;
_slDL2oDV += _slDDL2oDV;
_slDL2oDURow += _slDDL2oDUoDV;
}
#endif
}
extern "C" {
__int64 mmDDL2oDU_addAmbientMaskPoint;
__int64 mmDDL2oDV_addAmbientMaskPoint;
}
// add one layer point light without diffusion and with mask
void CLayerMixer::AddAmbientMaskPoint( UBYTE *pubMask, UBYTE ubMask)
{
// prepare some local variables
mmDDL2oDU_addAmbientMaskPoint = _slDDL2oDU;
mmDDL2oDV_addAmbientMaskPoint = _slDDL2oDV;
ULONG ulLightRGB = ByteSwap(lm_colLight);
_slLightMax<<=7;
_slLightStep>>=1;
#if (ASMOPT == 1)
#if (defined __MSVC_INLINE__)
__asm {
// prepare interpolants
movd mm0,D [_slL2Row]
movd mm1,D [_slDL2oDURow]
psllq mm1,32
por mm1,mm0 // MM1 = slDL2oDURow | slL2Row
movd mm0,D [_slDL2oDV]
movd mm2,D [_slDDL2oDUoDV]
psllq mm2,32
por mm2,mm0 // MM2 = slDDL2oDUoDV | slDL2oDV
// prepare color
pxor mm0,mm0 // MM0 = 0 | 0 (for unpacking purposes)
movd mm7,D [ulLightRGB]
punpcklbw mm7,mm0
psllw mm7,1
// loop thru rows
mov esi,D [pubMask]
mov edi,D [_pulLayer]
movzx edx,B [ubMask]
mov ebx,D [_iRowCt]
rowLoop:
push ebx
movd ebx,mm1 // EBX = slL2Point
movq mm3,mm1
psrlq mm3,32 // MM3 = 0 | slDL2oDU
// loop thru pixels in current row
mov ecx,D [_iPixCt]
pixLoop:
// check if pixel need to be drawn; i.e. draw if( [esi] & ubMask && (slL2Point<FTOX))
cmp ebx,FTOX
jge skipPixel
test dl,B [esi]
je skipPixel
// calculate intensities and do actual drawing of shadow pixel ARGB
movd mm4,ecx
mov eax,ebx
sar eax,SHIFTX
and eax,(SQRTTABLESIZE-1)
movzx eax,B aubSqrt[eax]
mov ecx,D [_slLightMax]
cmp eax,D [_slHotSpot]
jle skipInterpolation
mov ecx,255
sub ecx,eax
imul ecx,D [_slLightStep]
skipInterpolation:
// mix underlaying pixels with the calculated one
movd mm6,ecx
punpcklwd mm6,mm6
punpckldq mm6,mm6
pmulhw mm6,mm7
// add light pixel to underlying pixel
movd mm5,D [edi]
punpcklbw mm5,mm0
paddw mm5,mm6
packuswb mm5,mm0
movd D [edi],mm5
movd ecx,mm4
skipPixel:
// advance to next pixel
add edi,4
movd eax,mm3
add ebx,eax
paddd mm3,Q [mmDDL2oDU_addAmbientMaskPoint]
rol dl,1
adc esi,0
dec ecx
jnz pixLoop
// advance to the next row
pop ebx
add edi,D [_slModulo]
paddd mm1,mm2
paddd mm2,Q [mmDDL2oDV_addAmbientMaskPoint]
dec ebx
jnz rowLoop
emms
}
#elif (defined __GNU_INLINE__)
ULONG tmp1, tmp2;
__asm__ __volatile__ (
// prepare interpolants
"movd (" ASMSYM(_slL2Row) "), %%mm0 \n\t"
"movd (" ASMSYM(_slDL2oDURow) "), %%mm1 \n\t"
"psllq $32, %%mm1 \n\t"
"por %%mm0, %%mm1 \n\t" // MM1 = slDL2oDURow | slL2Row
"movd (" ASMSYM(_slDL2oDV) "), %%mm0 \n\t"
"movd (" ASMSYM(_slDDL2oDUoDV) "), %%mm2 \n\t"
"psllq $32, %%mm2 \n\t"
"por %%mm0, %%mm2 \n\t" // MM2 = slDDL2oDUoDV | slDL2oDV
// prepare color
"pxor %%mm0, %%mm0 \n\t" // MM0 = 0 | 0 (for unpacking purposes)
"movd %[ulLightRGB], %%mm7 \n\t"
"punpcklbw %%mm0, %%mm7 \n\t"
"psllw $1, %%mm7 \n\t"
// loop thru rows
"movl %[pubMask], %%esi \n\t"
"movl (" ASMSYM(_pulLayer) "), %%edi \n\t"
"movzbl %[ubMask], %%edx \n\t"
"movl (" ASMSYM(_iRowCt) "), %%eax \n\t"
"movl %%eax, %[xbx] \n\t"
"0: \n\t" // rowLoop
"movd %%mm1, %[slL2Point] \n\t"
"movq %%mm1, %%mm3 \n\t"
"psrlq $32, %%mm3 \n\t" // MM3 = 0 | slDL2oDU
// loop thru pixels in current row
"movl (" ASMSYM(_iPixCt) "), %%ecx \n\t"
"1: \n\t" // pixLoop
// check if pixel need to be drawn; i.e. draw if( [esi] & ubMask && (slL2Point<FTOX))
"cmpl $0x10000000, %[slL2Point] \n\t"
"jge 3f \n\t" // skipPixel
"testb (%%esi), %%dl \n\t"
"je 3f \n\t" // skipPixel
// calculate intensities and do actual drawing of shadow pixel ARGB
"movd %%ecx, %%mm4 \n\t"
"movl %[slL2Point], %%eax \n\t"
"sarl $15, %%eax \n\t"
"andl $8191, %%eax \n\t"
"movzbl " ASMSYM(aubSqrt) "(%%eax), %%eax \n\t"
"movl (" ASMSYM(_slLightMax) "), %%ecx \n\t"
"cmpl (" ASMSYM(_slHotSpot) "), %%eax \n\t"
"jle 2f \n\t" // skipInterpolation
"movl $255, %%ecx \n\t"
"subl %%eax, %%ecx \n\t"
"imull (" ASMSYM(_slLightStep) "), %%ecx \n\t"
"2: \n\t" // skipInterpolation
// mix underlaying pixels with the calculated one
"movd %%ecx, %%mm6 \n\t"
"punpcklwd %%mm6, %%mm6 \n\t"
"punpckldq %%mm6, %%mm6 \n\t"
"pmulhw %%mm7, %%mm6 \n\t"
// add light pixel to underlying pixel
"movd (%%edi), %%mm5 \n\t"
"punpcklbw %%mm0, %%mm5 \n\t"
"paddw %%mm6, %%mm5 \n\t"
"packuswb %%mm0, %%mm5 \n\t"
"movd %%mm5, (%%edi) \n\t"
"movd %%mm4, %%ecx \n\t"
"3: \n\t" // skipPixel
// advance to next pixel
"addl $4, %%edi \n\t"
"movd %%mm3, %%eax \n\t"
"addl %%eax, %[slL2Point] \n\t"
"paddd (" ASMSYM(mmDDL2oDU_addAmbientMaskPoint) "), %%mm3 \n\t"
"rolb $1, %%dl \n\t"
"adcl $0, %%esi \n\t"
"decl %%ecx \n\t"
"jnz 1b \n\t" // pixLoop
// advance to the next row
"addl (" ASMSYM(_slModulo) "), %%edi \n\t"
"paddd %%mm2, %%mm1 \n\t"
"paddd (" ASMSYM(mmDDL2oDV_addAmbientMaskPoint) "), %%mm2 \n\t"
"decl %[xbx] \n\t"
"jnz 0b \n\t" // rowLoop
"emms \n\t"
: [xbx] "=&g" (tmp1), [slL2Point] "=&g" (tmp2)
: [ulLightRGB] "g" (ulLightRGB), [pubMask] "g" (pubMask),
[ubMask] "m" (ubMask)
: FPU_REGS, MMX_REGS, "eax", "ecx", "edx", "esi", "edi",
"cc", "memory"
);
#else
#error Please write inline assembly for your platform.
#endif
#else // Portable C version...
UBYTE* pubLayer = (UBYTE*)_pulLayer;
for( PIX pixV=0; pixV<_iRowCt; pixV++)
{
SLONG slL2Point = _slL2Row;
SLONG slDL2oDU = _slDL2oDURow;
for( PIX pixU=0; pixU<_iPixCt; pixU++)
{
// if the point is not masked
if( *pubMask & ubMask && (slL2Point<FTOX)) {
SLONG slL = (slL2Point>>SHIFTX)&(SQRTTABLESIZE-1); // and is just for degenerate cases
SLONG slIntensity = _slLightMax;
slL = aubSqrt[slL];
if( slL>_slHotSpot) slIntensity = ((255-slL)*_slLightStep);
// add the intensity to the pixel
AddToCluster( pubLayer, slIntensity);
}
// go to the next pixel
pubLayer+=4;
slL2Point += slDL2oDU;
slDL2oDU += _slDDL2oDU;
ubMask<<=1;
if( ubMask==0) {
pubMask++;
ubMask = 1;
}
}
// go to the next row
pubLayer += _slModulo;
_slL2Row += _slDL2oDV;
_slDL2oDV += _slDDL2oDV;
_slDL2oDURow += _slDDL2oDUoDV;
}
#endif
}
extern "C" {
__int64 mmDDL2oDU_AddDiffusionPoint;
__int64 mmDDL2oDV_AddDiffusionPoint;
}
// add one layer point light with diffusion and without mask
void CLayerMixer::AddDiffusionPoint(void)
{
// adjust params for diffusion lighting
SLONG slMax1oL = MAX_SLONG;
_slLightStep = FloatToInt(_slLightStep * _fMinLightDistance * _f1oFallOff);
if( _slLightStep!=0) slMax1oL = (256<<8) / _slLightStep +256;
// prepare some local variables
mmDDL2oDU_AddDiffusionPoint = _slDDL2oDU;
mmDDL2oDV_AddDiffusionPoint = _slDDL2oDV;
ULONG ulLightRGB = ByteSwap(lm_colLight);
_slLightMax<<=7;
_slLightStep>>=1;
#if ASMOPT == 1
#if (defined __MSVC_INLINE__)
__asm {
// prepare interpolants
movd mm0,D [_slL2Row]
movd mm1,D [_slDL2oDURow]
psllq mm1,32
por mm1,mm0 // MM1 = slDL2oDURow | slL2Row
movd mm0,D [_slDL2oDV]
movd mm2,D [_slDDL2oDUoDV]
psllq mm2,32
por mm2,mm0 // MM2 = slDDL2oDUoDV | slDL2oDV
// prepare color
pxor mm0,mm0
movd mm7,D [ulLightRGB]
punpcklbw mm7,mm0
psllw mm7,1
// loop thru rows
mov edi,D [_pulLayer]
mov ebx,D [_iRowCt]
rowLoop:
push ebx
movd ebx,mm1 // EBX = slL2Point
movq mm3,mm1
psrlq mm3,32 // MM3 = 0 | slDL2oDU
// loop thru pixels in current row
mov ecx,D [_iPixCt]
pixLoop:
// check if pixel need to be drawn
cmp ebx,FTOX
jge skipPixel
// calculate intensities and do actual drawing of shadow pixel ARGB
movd mm4,ecx
mov eax,ebx
sar eax,SHIFTX
and eax,(SQRTTABLESIZE-1)
movzx eax,W auw1oSqrt[eax*2]
mov ecx,D [_slLightMax]
cmp eax,D [slMax1oL]
jge skipInterpolation
lea ecx,[eax-256]
imul ecx,D [_slLightStep]
skipInterpolation:
// calculate rgb pixel to add
movd mm6,ecx
punpcklwd mm6,mm6
punpckldq mm6,mm6
pmulhw mm6,mm7
// add dynamic light pixel to underlying pixel
movd mm5,D [edi]
punpcklbw mm5,mm0
paddw mm5,mm6
packuswb mm5,mm0
movd D [edi],mm5
movd ecx,mm4
skipPixel:
// advance to next pixel
add edi,4
movd eax,mm3
add ebx,eax
paddd mm3,Q [mmDDL2oDU_AddDiffusionPoint]
dec ecx
jnz pixLoop
// advance to the next row
pop ebx
add edi,D [_slModulo]
paddd mm1,mm2
paddd mm2,Q [mmDDL2oDV_AddDiffusionPoint]
dec ebx
jnz rowLoop
emms
}
#elif (defined __GNU_INLINE__)
ULONG tmp1, tmp2;
__asm__ __volatile__ (
// prepare interpolants
"movd (" ASMSYM(_slL2Row) "), %%mm0 \n\t"
"movd (" ASMSYM(_slDL2oDURow) "), %%mm1 \n\t"
"psllq $32, %%mm1 \n\t"
"por %%mm0, %%mm1 \n\t" // MM1 = slDL2oDURow | slL2Row
"movd (" ASMSYM(_slDL2oDV) "), %%mm0 \n\t"
"movd (" ASMSYM(_slDDL2oDUoDV) "), %%mm2 \n\t"
"psllq $32, %%mm2 \n\t"
"por %%mm0, %%mm2 \n\t" // MM2 = slDDL2oDUoDV | slDL2oDV
// prepare color
"pxor %%mm0, %%mm0 \n\t"
"movd %[ulLightRGB], %%mm7 \n\t"
"punpcklbw %%mm0, %%mm7 \n\t"
"psllw $1, %%mm7 \n\t"
// loop thru rows
"movl (" ASMSYM(_pulLayer) "), %%edi \n\t"
"movl (" ASMSYM(_iRowCt) "), %[xbx] \n\t"
"0: \n\t" // rowLoop
"movd %%mm1, %[slL2Point] \n\t"
"movq %%mm1, %%mm3 \n\t"
"psrlq $32, %%mm3 \n\t" // MM3 = 0 | slDL2oDU
// loop thru pixels in current row
"movl (" ASMSYM(_iPixCt) "), %%ecx \n\t"
"1: \n\t" // pixLoop
// check if pixel need to be drawn
"cmpl $0x10000000, %[slL2Point] \n\t"
"jge 3f \n\t" // skipPixel
// calculate intensities and do actual drawing of shadow pixel ARGB
"movd %%ecx, %%mm4 \n\t"
"movl %[slL2Point], %%eax \n\t"
"sarl $15, %%eax \n\t"
"andl $8191, %%eax \n\t"
"movzwl " ASMSYM(auw1oSqrt) "(, %%eax, 2), %%eax \n\t"
"movl (" ASMSYM(_slLightMax) "), %%ecx \n\t"
"cmpl %[slMax1oL], %%eax \n\t"
"jge 2f \n\t" // skipInterpolation
"leal -256(%%eax), %%ecx \n\t"
"imull (" ASMSYM(_slLightStep) "), %%ecx \n\t"
"2: \n\t" // skipInterpolation
// calculate rgb pixel to add
"movd %%ecx, %%mm6 \n\t"
"punpcklwd %%mm6, %%mm6 \n\t"
"punpckldq %%mm6, %%mm6 \n\t"
"pmulhw %%mm7, %%mm6 \n\t"
// add dynamic light pixel to underlying pixel
"movd (%%edi), %%mm5 \n\t"
"punpcklbw %%mm0, %%mm5 \n\t"
"paddw %%mm6, %%mm5 \n\t"
"packuswb %%mm0, %%mm5 \n\t"
"movd %%mm5, (%%edi) \n\t"
"movd %%mm4, %%ecx \n\t"
"3: \n\t" // skipPixel
// advance to next pixel
"addl $4, %%edi \n\t"
"movd %%mm3, %%eax \n\t"
"addl %%eax, %[slL2Point] \n\t"
"paddd (" ASMSYM(mmDDL2oDU_AddDiffusionPoint) "), %%mm3 \n\t"
"decl %%ecx \n\t"
"jnz 1b \n\t" // pixLoop
// advance to the next row
"addl (" ASMSYM(_slModulo) "), %%edi \n\t"
"paddd %%mm2, %%mm1 \n\t"
"paddd (" ASMSYM(mmDDL2oDV_AddDiffusionPoint) "), %%mm2 \n\t"
"decl %[xbx] \n\t"
"jnz 0b \n\t" // rowLoop
"emms \n\t"
: [xbx] "=&r" (tmp1), [slL2Point] "=&g" (tmp2)
: [ulLightRGB] "g" (ulLightRGB), [slMax1oL] "g" (slMax1oL)
: FPU_REGS, MMX_REGS, "eax", "ecx", "edi", "cc", "memory"
);
#else
#error Write inline assembly for your platform.
#endif
#else
// for each pixel in the shadow map
UBYTE* pubLayer = (UBYTE*)_pulLayer;
for( PIX pixV=0; pixV<_iRowCt; pixV++)
{
SLONG slL2Point = _slL2Row;
SLONG slDL2oDU = _slDL2oDURow;
for( PIX pixU=0; pixU<_iPixCt; pixU++)
{
// if the point is not masked
if(slL2Point<FTOX) {
SLONG sl1oL = (slL2Point>>SHIFTX)&(SQRTTABLESIZE-1); // and is just for degenerate cases
sl1oL = auw1oSqrt[sl1oL];
SLONG slIntensity = _slLightMax;
if( sl1oL<256) slIntensity = 0;
else if( sl1oL<slMax1oL) slIntensity = ((sl1oL-256)*_slLightStep)/*>>16*/;
// add the intensity to the pixel
AddToCluster( pubLayer, slIntensity);
}
// advance to next pixel
pubLayer+=4;
slL2Point += slDL2oDU;
slDL2oDU += _slDDL2oDU;
}
// advance to next row
pubLayer += _slModulo;
_slL2Row += _slDL2oDV;
_slDL2oDV += _slDDL2oDV;
_slDL2oDURow += _slDDL2oDUoDV;
}
#endif
}
extern "C" {
__int64 mmDDL2oDU_AddDiffusionMaskPoint;
__int64 mmDDL2oDV_AddDiffusionMaskPoint;
}
// add one layer point light with diffusion and mask
void CLayerMixer::AddDiffusionMaskPoint( UBYTE *pubMask, UBYTE ubMask)
{
// adjust params for diffusion lighting
SLONG slMax1oL = MAX_SLONG;
_slLightStep = FloatToInt(_slLightStep * _fMinLightDistance * _f1oFallOff);
if( _slLightStep!=0) slMax1oL = (256<<8) / _slLightStep +256;
// prepare some local variables
mmDDL2oDU_AddDiffusionMaskPoint = _slDDL2oDU;
mmDDL2oDV_AddDiffusionMaskPoint = _slDDL2oDV;
ULONG ulLightRGB = ByteSwap(lm_colLight);
_slLightMax<<=7;
_slLightStep>>=1;
#if (ASMOPT == 1)
#if (defined __MSVC_INLINE__)
__asm {
// prepare interpolants
movd mm0,D [_slL2Row]
movd mm1,D [_slDL2oDURow]
psllq mm1,32
por mm1,mm0 // MM1 = slDL2oDURow | slL2Row
movd mm0,D [_slDL2oDV]
movd mm2,D [_slDDL2oDUoDV]
psllq mm2,32
por mm2,mm0 // MM2 = slDDL2oDUoDV | slDL2oDV
// prepare color
pxor mm0,mm0 // MM0 = 0 | 0 (for unpacking purposes)
movd mm7,D [ulLightRGB]
punpcklbw mm7,mm0
psllw mm7,1
// loop thru rows
mov esi,D [pubMask]
mov edi,D [_pulLayer]
movzx edx,B [ubMask]
mov ebx,D [_iRowCt]
rowLoop:
push ebx
movd ebx,mm1 // EBX = slL2Point
movq mm3,mm1
psrlq mm3,32 // MM3 = 0 | slDL2oDU
// loop thru pixels in current row
mov ecx,D [_iPixCt]
pixLoop:
// check if pixel need to be drawn; i.e. draw if( [esi] & ubMask && (slL2Point<FTOX))
cmp ebx,FTOX
jge skipPixel
test dl,B [esi]
je skipPixel
// calculate intensities and do actual drawing of shadow pixel ARGB
movd mm4,ecx
mov eax,ebx
sar eax,SHIFTX
and eax,(SQRTTABLESIZE-1)
movzx eax,W auw1oSqrt[eax*2]
mov ecx,D [_slLightMax]
cmp eax,D [slMax1oL]
jge skipInterpolation
lea ecx,[eax-256]
imul ecx,D [_slLightStep]
skipInterpolation:
// mix underlaying pixels with the calculated one
movd mm6,ecx
punpcklwd mm6,mm6
punpckldq mm6,mm6
pmulhw mm6,mm7
// add light pixel to underlying pixel
movd mm5,D [edi]
punpcklbw mm5,mm0
paddw mm5,mm6
packuswb mm5,mm0
movd D [edi],mm5
movd ecx,mm4
skipPixel:
// advance to next pixel
add edi,4
movd eax,mm3
add ebx,eax
paddd mm3,Q [mmDDL2oDU_AddDiffusionMaskPoint]
rol dl,1
adc esi,0
dec ecx
jnz pixLoop
// advance to the next row
pop ebx
add edi,D [_slModulo]
paddd mm1,mm2
paddd mm2,Q [mmDDL2oDV_AddDiffusionMaskPoint]
dec ebx
jnz rowLoop
emms
}
#elif (defined __GNU_INLINE__)
ULONG tmp1, tmp2;
__asm__ __volatile__ (
// prepare interpolants
"movd (" ASMSYM(_slL2Row) "), %%mm0 \n\t"
"movd (" ASMSYM(_slDL2oDURow) "), %%mm1 \n\t"
"psllq $32, %%mm1 \n\t"
"por %%mm0, %%mm1 \n\t" // MM1 = slDL2oDURow | slL2Row
"movd (" ASMSYM(_slDL2oDV) "), %%mm0 \n\t"
"movd (" ASMSYM(_slDDL2oDUoDV) "), %%mm2 \n\t"
"psllq $32, %%mm2 \n\t"
"por %%mm0, %%mm2 \n\t" // MM2 = slDDL2oDUoDV | slDL2oDV
// prepare color
"pxor %%mm0, %%mm0 \n\t" // MM0 = 0 | 0 (for unpacking purposes)
"movd %[ulLightRGB], %%mm7 \n\t"
"punpcklbw %%mm0, %%mm7 \n\t"
"psllw $1, %%mm7 \n\t"
// loop thru rows
"movl %[pubMask], %%esi \n\t"
"movl (" ASMSYM(_pulLayer) "), %%edi \n\t"
"movzbl %[ubMask], %%edx \n\t"
"movl (" ASMSYM(_iRowCt) "), %%eax \n\t"
"movl %%eax, %[xbx] \n\t"
"0: \n\t" // rowLoop
"movd %%mm1, %[slL2Point] \n\t"
"movq %%mm1, %%mm3 \n\t"
"psrlq $32, %%mm3 \n\t" // MM3 = 0 | slDL2oDU
// loop thru pixels in current row
"movl (" ASMSYM(_iPixCt) "), %%ecx \n\t"
"1: \n\t" // pixLoop
// check if pixel need to be drawn; i.e. draw if( [esi] & ubMask && (slL2Point<FTOX))
"cmpl $0x10000000, %[slL2Point] \n\t"
"jge 3f \n\t" // skipPixel
"testb (%%esi), %%dl \n\t"
"je 3f \n\t" // skipPixel
// calculate intensities and do actual drawing of shadow pixel ARGB
"movd %%ecx, %%mm4 \n\t"
"movl %[slL2Point], %%eax \n\t"
"sarl $15, %%eax \n\t"
"andl $8191, %%eax \n\t"
"movzwl " ASMSYM(auw1oSqrt) "(, %%eax, 2), %%eax \n\t"
"movl (" ASMSYM(_slLightMax) "), %%ecx \n\t"
"cmpl %[slMax1oL], %%eax \n\t"
"jge 2f \n\t" // skipInterpolation
"leal -256(%%eax), %%ecx \n\t"
"imull (" ASMSYM(_slLightStep) "), %%ecx \n\t"
"2: \n\t" // skipInterpolation
// mix underlaying pixels with the calculated one
"movd %%ecx, %%mm6 \n\t"
"punpcklwd %%mm6, %%mm6 \n\t"
"punpckldq %%mm6, %%mm6 \n\t"
"pmulhw %%mm7, %%mm6 \n\t"
// add light pixel to underlying pixel
"movd (%%edi), %%mm5 \n\t"
"punpcklbw %%mm0, %%mm5 \n\t"
"paddw %%mm6, %%mm5 \n\t"
"packuswb %%mm0, %%mm5 \n\t"
"movd %%mm5, (%%edi) \n\t"
"movd %%mm4, %%ecx \n\t"
"3: \n\t" // skipPixel
// advance to next pixel
"addl $4, %%edi \n\t"
"movd %%mm3, %%eax \n\t"
"addl %%eax, %[slL2Point] \n\t"
"paddd (" ASMSYM(mmDDL2oDU_AddDiffusionMaskPoint) "), %%mm3 \n\t"
"rolb $1, %%dl \n\t"
"adcl $0, %%esi \n\t"
"decl %%ecx \n\t"
"jnz 1b \n\t" // pixLoop
// advance to the next row
"addl (" ASMSYM(_slModulo) "), %%edi \n\t"
"paddd %%mm2, %%mm1 \n\t"
"paddd (" ASMSYM(mmDDL2oDV_AddDiffusionMaskPoint) "), %%mm2 \n\t"
"decl %[xbx] \n\t"
"jnz 0b \n\t" // rowLoop
"emms \n\t"
: [xbx] "=&g" (tmp1), [slL2Point] "=&g" (tmp2)
: [ulLightRGB] "g" (ulLightRGB), [pubMask] "g" (pubMask),
[ubMask] "m" (ubMask), [slMax1oL] "g" (slMax1oL)
: FPU_REGS, MMX_REGS, "eax", "ecx", "edx", "esi", "edi",
"cc", "memory"
);
#else
#error Write inline ASM for your platform.
#endif
#else
// for each pixel in the shadow map
UBYTE* pubLayer = (UBYTE*)_pulLayer;
for( PIX pixV=0; pixV<_iRowCt; pixV++)
{
SLONG slL2Point = _slL2Row;
SLONG slDL2oDU = _slDL2oDURow;
for( PIX pixU=0; pixU<_iPixCt; pixU++)
{
// if the point is not masked
if( *pubMask&ubMask && (slL2Point<FTOX)) {
SLONG sl1oL = (slL2Point>>SHIFTX)&(SQRTTABLESIZE-1); // and is just for degenerate cases
sl1oL = auw1oSqrt[sl1oL];
SLONG slIntensity = _slLightMax;
if( sl1oL<256) slIntensity = 0;
else if( sl1oL<slMax1oL) slIntensity = ((sl1oL-256)*_slLightStep)/*>>16*/;
// add the intensity to the pixel
AddToCluster( pubLayer, slIntensity);
}
// advance to next pixel
pubLayer+=4;
slL2Point += slDL2oDU;
slDL2oDU += _slDDL2oDU;
ubMask<<=1;
if( ubMask==0) {
pubMask++;
ubMask = 1;
}
}
// advance to next row
pubLayer += _slModulo;
_slL2Row += _slDL2oDV;
_slDL2oDV += _slDDL2oDV;
_slDL2oDURow += _slDDL2oDUoDV;
}
#endif
}
// prepares point light that creates layer (returns TRUE if there is infulence)
BOOL CLayerMixer::PrepareOneLayerPoint( CBrushShadowLayer *pbsl, BOOL bNoMask)
{
// determine light infulence dimensions
_iPixCt = pbsl->bsl_pixSizeU >>lm_iMipShift;
_iRowCt = pbsl->bsl_pixSizeV >>lm_iMipShift;
PIX pixMinU = pbsl->bsl_pixMinU >>lm_iMipShift;
PIX pixMinV = pbsl->bsl_pixMinV >>lm_iMipShift;
// clamp influence to polygon size
if( (pixMinU+_iPixCt) > lm_pixPolygonSizeU && bNoMask) _iPixCt = lm_pixPolygonSizeU-pixMinU;
if( (pixMinV+_iRowCt) > lm_pixPolygonSizeV) _iRowCt = lm_pixPolygonSizeV-pixMinV;
_slModulo = (lm_pixCanvasSizeU-_iPixCt) *BYTES_PER_TEXEL;
_pulLayer = lm_pulShadowMap + (pixMinV*lm_pixCanvasSizeU)+pixMinU;
ASSERT( pixMinU>=0 && pixMinU<lm_pixCanvasSizeU && pixMinV>=0 && pixMinV<lm_pixCanvasSizeV);
// get the light source properties of the layer
lm_plsLight = pbsl->bsl_plsLightSource;
_vLight = &lm_plsLight->ls_penEntity->GetPlacement().pl_PositionVector;
_fMinLightDistance = lm_pbpoPolygon->bpo_pbplPlane->bpl_plAbsolute.PointDistance(*_vLight);
_f1oFallOff = 1.0f / lm_plsLight->ls_rFallOff;
_ulLightFlags = lm_plsLight->ls_ulFlags;
_ulPolyFlags = lm_pbpoPolygon->bpo_ulFlags;
lm_colLight = lm_plsLight->GetLightColor();
pbsl->bsl_colLastAnim = lm_colLight;
// if there is no influence, do nothing
if( (pbsl->bsl_pixSizeU>>lm_iMipShift)==0 || (pbsl->bsl_pixSizeV>>lm_iMipShift)==0
|| _iPixCt<=0 || _iRowCt<=0) return FALSE;
// adjust for sector ambient
if( _ulLightFlags&LSF_SUBSTRACTSECTORAMBIENT) {
COLOR colAmbient = lm_pbpoPolygon->bpo_pbscSector->bsc_colAmbient;
IncrementByteWithClip( ((UBYTE*)&lm_colLight)[1], -((UBYTE*)&colAmbient)[1]);
IncrementByteWithClip( ((UBYTE*)&lm_colLight)[2], -((UBYTE*)&colAmbient)[2]);
IncrementByteWithClip( ((UBYTE*)&lm_colLight)[3], -((UBYTE*)&colAmbient)[3]);
if( _ulPolyFlags&BPOF_HASDIRECTIONALAMBIENT)
{ // find directional layers for each shadow layer
FOREACHINLIST( CBrushShadowLayer, bsl_lnInShadowMap, lm_pbsmShadowMap->bsm_lhLayers, itbsl)
{ // loop thru layers
CBrushShadowLayer &bsl = *itbsl;
if( bsl.bsl_plsLightSource->ls_ulFlags&LSF_DIRECTIONAL)
{ // skip if no ambient color
colAmbient = bsl.bsl_plsLightSource->ls_colAmbient & 0xFFFFFF00;
if( IsBlack(colAmbient)) continue;
// substract ambient
IncrementByteWithClip( ((UBYTE*)&lm_colLight)[1], -((UBYTE*)&colAmbient)[1]);
IncrementByteWithClip( ((UBYTE*)&lm_colLight)[2], -((UBYTE*)&colAmbient)[2]);
IncrementByteWithClip( ((UBYTE*)&lm_colLight)[3], -((UBYTE*)&colAmbient)[3]);
}
}
}
}
// prepare intermediate light interpolants
FLOAT3D v00 = (lm_vO+lm_vStepU*pixMinU + lm_vStepV*pixMinV) - *_vLight;
FLOAT fFactor = FTOX * _f1oFallOff*_f1oFallOff;
FLOAT fL2Row = v00%v00;
FLOAT fDDL2oDU = lm_vStepU%lm_vStepU;
FLOAT fDDL2oDV = lm_vStepV%lm_vStepV;
FLOAT fDDL2oDUoDV = lm_vStepU%lm_vStepV;
FLOAT fDL2oDURow = fDDL2oDU + 2*(lm_vStepU%v00);
FLOAT fDL2oDV = fDDL2oDV + 2*(lm_vStepV%v00);
//_v00 = v00;
#if ((ASMOPT == 1) && (!defined __GNU_INLINE__))
#if (defined __MSVC_INLINE__)
__asm {
fld D [fDDL2oDU]
fadd D [fDDL2oDU]
fld D [fDDL2oDV]
fadd D [fDDL2oDV]
fld D [fDDL2oDUoDV]
fadd D [fDDL2oDUoDV]
// st0=2*fDDL2oDUoDV, st1=2*fDDL2oDV, st2=2*fDDL2oDU
fld D [fL2Row]
fmul D [fFactor]
fld D [fDL2oDURow]
fmul D [fFactor]
fld D [fDL2oDV]
fmul D [fFactor]
// st0=fDL2oDV*fFactor, st1=fDL2oDURow*fFactor, st2=fL2Row*fFactor,
// st3=2*fDDL2oDUoDV, st4=2*fDDL2oDV, st5=2*fDDL2oDU
fld D [fFactor]
fmul st(4),st(0)
fmul st(5),st(0)
fmulp st(6),st(0)
fistp D [_slDL2oDV]
fistp D [_slDL2oDURow]
fistp D [_slL2Row]
fistp D [_slDDL2oDUoDV]
fistp D [_slDDL2oDV]
fistp D [_slDDL2oDU]
}
#elif (defined __GNU_INLINE__)
STUBBED("inline asm.");
#else
#error Please write inline assembly for your platform.
#endif
#else
fDDL2oDU *= 2;
fDDL2oDV *= 2;
fDDL2oDUoDV *= 2;
_slL2Row = FloatToInt( fL2Row * fFactor);
_slDDL2oDU = FloatToInt( fDDL2oDU * fFactor);
_slDDL2oDV = FloatToInt( fDDL2oDV * fFactor);
_slDDL2oDUoDV = FloatToInt( fDDL2oDUoDV * fFactor);
_slDL2oDURow = FloatToInt( fDL2oDURow * fFactor);
_slDL2oDV = FloatToInt( fDL2oDV * fFactor);
#endif
// prepare final light interpolants
_slLightMax = 255;
_slHotSpot = FloatToInt( 255.0f * lm_plsLight->ls_rHotSpot * _f1oFallOff);
_slLightStep = FloatToInt( 65535.0f / (255.0f - _slHotSpot));
// dark light inverts parameters
if( _ulLightFlags & LSF_DARKLIGHT) {
_slLightMax = -_slLightMax;
_slLightStep = -_slLightStep;
}
// saturate light color
lm_colLight = AdjustColor( lm_colLight, _slShdHueShift, _slShdSaturation);
// all done
return TRUE;
}
// add one layer to the shadow map (pubMask=NULL for no mask)
void CLayerMixer::AddOneLayerPoint( CBrushShadowLayer *pbsl, UBYTE *pubMask, UBYTE ubMask)
{
// try to prepare layer for this point light
_pfWorldEditingProfile.StartTimer( CWorldEditingProfile::PTI_ADDONELAYERPOINT);
if( !PrepareOneLayerPoint( pbsl, pubMask==NULL)) {
_pfWorldEditingProfile.StopTimer( CWorldEditingProfile::PTI_ADDONELAYERPOINT);
return;
}
// determine diffusion presence and corresponding routine
BOOL bDiffusion = (_ulLightFlags&LSF_DIFFUSION) && !(_ulPolyFlags&BPOF_NOPLANEDIFFUSION);
// masked or non-masked?
if( pubMask==NULL) {
// non-masked
if( !lm_bDynamic && bDiffusion) {
// non-masked diffusion
AddDiffusionPoint();
} else {
// non-masked ambient
AddAmbientPoint();
}
} else {
// masked
if( bDiffusion) {
// masked diffusion
AddDiffusionMaskPoint( pubMask, ubMask);
} else {
AddAmbientMaskPoint( pubMask, ubMask);
}
}
// all done
_pfWorldEditingProfile.StopTimer(CWorldEditingProfile::PTI_ADDONELAYERPOINT);
}
// apply gradient to layer
void CLayerMixer::AddOneLayerGradient( CGradientParameters &gp)
{
// convert gradient parameters for plane
ASSERT( Abs(gp.gp_fH1-gp.gp_fH0)>0.0001f);
FLOAT f1oDH = 1.0f / (gp.gp_fH1-gp.gp_fH0);
FLOAT fGr00 = (lm_vO % gp.gp_vGradientDir - gp.gp_fH0) *f1oDH;
FLOAT fDGroDI = (lm_vStepU % gp.gp_vGradientDir) *f1oDH;
FLOAT fDGroDJ = (lm_vStepV % gp.gp_vGradientDir) *f1oDH;
fDGroDI += fDGroDI/lm_pixPolygonSizeU;
fDGroDJ += fDGroDJ/lm_pixPolygonSizeV;
SLONG fixDGroDI = FloatToInt(fDGroDI*32767.0f); // 16:15
SLONG fixDGroDJ = FloatToInt(fDGroDJ*32767.0f); // 16:15
COLOR col0 = gp.gp_col0;
COLOR col1 = gp.gp_col1;
_pulLayer = lm_pulShadowMap;
FLOAT fStart = Clamp( fGr00-(fDGroDJ+fDGroDI)*0.5f, 0.0f, 1.0f);
#if ((ASMOPT == 1) && (!defined __GNU_INLINE__))
#if (defined __MSVC_INLINE__)
__int64 mmRowAdv;
SLONG fixGRow = (fGr00-(fDGroDJ+fDGroDI)*0.5f)*32767.0f; // 16:15
SLONG slModulo = (lm_pixCanvasSizeU-lm_pixPolygonSizeU) *BYTES_PER_TEXEL;
COLOR colStart = LerpColor( col0, col1, fStart);
INDEX ctCols = lm_pixPolygonSizeU;
INDEX ctRows = lm_pixPolygonSizeV;
BOOL bDarkLight = gp.gp_bDark;
__asm {
mov eax,D [colStart]
mov ebx,D [col0]
mov edx,D [col1]
bswap eax
bswap ebx
bswap edx
movd mm1,eax
movd mm2,ebx
movd mm3,edx
punpcklbw mm1,mm1
punpcklbw mm2,mm2
punpcklbw mm3,mm3
psrlw mm1,2
psrlw mm2,1
psrlw mm3,1
// eventually adjust for dark light
cmp D [bDarkLight],0
je skipDark
pcmpeqd mm0,mm0
pxor mm1,mm0
pxor mm2,mm0
pxor mm3,mm0
psubw mm1,mm0
psubw mm2,mm0
psubw mm3,mm0
skipDark:
// keep colors
movq mm0,mm2
movq mm7,mm3
// find row and column advancers
psubw mm3,mm2
movq mm2,mm3
movd mm5,D [fixDGroDI]
movd mm6,D [fixDGroDJ]
punpcklwd mm5,mm5
punpcklwd mm6,mm6
punpckldq mm5,mm5
punpckldq mm6,mm6
pmulhw mm2,mm5 // column color advancer (8:6)
pmulhw mm3,mm6 // row color advancer (8:6)
movq Q [mmRowAdv],mm3
// prepare starting variables
movq mm5,mm0
movq mm6,mm7
psraw mm5,1 // starting color
psraw mm6,1 // ending color
pxor mm0,mm0
mov esi,D [fixGRow]
mov edi,D [_pulLayer]
mov edx,D [ctRows]
rowLoop:
mov ebx,esi
movq mm4,mm1
mov ecx,D [ctCols]
pixLoop:
// add or substract light pixel to underlying pixel
movq mm7,mm4
psraw mm7,6
movd mm3,D [edi]
punpcklbw mm3,mm0
paddw mm3,mm7
packuswb mm3,mm0
movd D [edi],mm3
// advance to next pixel
add ebx,D [fixDGroDI]
cmp ebx,0x8000
ja pixClamp
paddw mm4,mm2
add edi,4
dec ecx
jnz pixLoop
jmp pixDone
pixClamp:
movq mm4,mm6
jg pixNext
movq mm4,mm5
pixNext:
add edi,4
dec ecx
jnz pixLoop
pixDone:
// advance to next row
add esi,D [fixDGroDJ]
cmp esi,0x8000
ja rowClamp
paddw mm1,Q [mmRowAdv]
add edi,D [slModulo]
dec edx
jnz rowLoop
jmp rowDone
rowClamp:
movq mm1,mm6
jg rowNext
movq mm1,mm5
rowNext:
add edi,D [slModulo]
dec edx
jnz rowLoop
rowDone:
emms
}
#elif (defined __GNU_INLINE__)
STUBBED("WRITE ME. Argh.");
#else
#error Need inline assembly for your platform.
#endif
#else
// well, make gradient ...
SLONG slR0=0,slG0=0,slB0=0;
SLONG slR1=0,slG1=0,slB1=0;
ColorToRGB( col0, (UBYTE&)slR0,(UBYTE&)slG0,(UBYTE&)slB0);
ColorToRGB( col1, (UBYTE&)slR1,(UBYTE&)slG1,(UBYTE&)slB1);
if( gp.gp_bDark) {
slR0 = -slR0; slG0 = -slG0; slB0 = -slB0;
slR1 = -slR1; slG1 = -slG1; slB1 = -slB1;
}
fixDGroDI >>= 1; // 16:14
fixDGroDJ >>= 1; // 16:14
SWORD fixRrow = Lerp( slR0,slR1,fStart) <<6; // 8:6
SWORD fixGrow = Lerp( slG0,slG1,fStart) <<6; // 8:6
SWORD fixBrow = Lerp( slB0,slB1,fStart) <<6; // 8:6
SWORD fixDRoDI = ((slR1-slR0)*fixDGroDI)>>8;
SWORD fixDGoDI = ((slG1-slG0)*fixDGroDI)>>8;
SWORD fixDBoDI = ((slB1-slB0)*fixDGroDI)>>8;
SWORD fixDRoDJ = ((slR1-slR0)*fixDGroDJ)>>8;
SWORD fixDGoDJ = ((slG1-slG0)*fixDGroDJ)>>8;
SWORD fixDBoDJ = ((slB1-slB0)*fixDGroDJ)>>8;
// loop it, baby
FLOAT fGrRow = fGr00 - (fDGroDJ+fDGroDI)*0.5f;
PIX pixOffset = 0;
PIX pixModulo = lm_pixCanvasSizeU-lm_pixPolygonSizeU;
for( INDEX j=0; j<lm_pixPolygonSizeV; j++)
{ // prepare row
FLOAT fGrCol = fGrRow;
SWORD fixRcol = fixRrow;
SWORD fixGcol = fixGrow;
SWORD fixBcol = fixBrow;
for( INDEX i=0; i<lm_pixPolygonSizeU; i++)
{ // loop pixels
SLONG slR = Clamp( fixRcol>>6, -255, +255);
SLONG slG = Clamp( fixGcol>>6, -255, +255);
SLONG slB = Clamp( fixBcol>>6, -255, +255);
IncrementByteWithClip( ((UBYTE*)&_pulLayer[pixOffset])[0], slR);
IncrementByteWithClip( ((UBYTE*)&_pulLayer[pixOffset])[1], slG);
IncrementByteWithClip( ((UBYTE*)&_pulLayer[pixOffset])[2], slB);
// advance to next pixel
fGrCol += fDGroDI;
pixOffset++;
if( fGrCol<0) {
fixRcol = slR0<<6;
fixGcol = slG0<<6;
fixBcol = slB0<<6;
} else if( fGrCol>1) {
fixRcol = slR1<<6;
fixGcol = slG1<<6;
fixBcol = slB1<<6;
} else {
fixRcol += fixDRoDI;
fixGcol += fixDGoDI;
fixBcol += fixDBoDI;
}
}
// advance to next row
fGrRow += fDGroDJ;
pixOffset += pixModulo;
if( fGrRow<0) {
fixRrow = slR0<<6;
fixGrow = slG0<<6;
fixBrow = slB0<<6;
} else if( fGrRow>1) {
fixRrow = slR1<<6;
fixGrow = slG1<<6;
fixBrow = slB1<<6;
} else {
fixRrow += fixDRoDJ;
fixGrow += fixDGoDJ;
fixBrow += fixDBoDJ;
}
}
#endif
}
// apply directional light or ambient to layer
void CLayerMixer::AddDirectional(void)
{
#if ASMOPT == 1
ULONG ulLight = ByteSwap( lm_colLight);
#if (defined __MSVC_INLINE__)
__asm {
// prepare pointers and variables
mov edi,D [_pulLayer]
mov ebx,D [_iRowCt]
movd mm6,D [ulLight]
punpckldq mm6,mm6
rowLoop:
mov ecx,D [_iPixCt]
shr ecx,1
jz pixRest
pixLoop:
// mix underlaying pixels with the constant color pixel
movq mm5,Q [edi]
paddusb mm5,mm6
movq Q [edi],mm5
// advance to next pixel
add edi,8
dec ecx
jnz pixLoop
pixRest:
test D [_iPixCt],1
jz rowNext
movd mm5,D [edi]
paddusb mm5,mm6
movd D [edi],mm5
add edi,4
rowNext:
// advance to the next row
add edi,D [_slModulo]
dec ebx
jnz rowLoop
emms
}
#elif (defined __GNU_INLINE__)
ULONG tmp;
__asm__ __volatile__ (
// prepare pointers and variables
"movl (" ASMSYM(_pulLayer) "), %%edi \n\t"
"movl (" ASMSYM(_iRowCt) "), %[xbx] \n\t"
"movd %[ulLight], %%mm6 \n\t"
"punpckldq %%mm6, %%mm6 \n\t"
"0: \n\t" // rowLoop
"movl (" ASMSYM(_iPixCt) "), %%ecx \n\t"
"shrl $1, %%ecx \n\t"
"jz 2f \n\t" // pixRest
"1: \n\t" // pixLoop
// mix underlaying pixels with the constant color pixel
"movq (%%edi), %%mm5 \n\t"
"paddusb %%mm6, %%mm5 \n\t"
"movq %%mm5, (%%edi) \n\t"
// advance to next pixel
"addl $8, %%edi \n\t"
"decl %%ecx \n\t"
"jnz 1b \n\t" // pixLoop
"2: \n\t" // pixRest
"testl $1, (" ASMSYM(_iPixCt) ") \n\t"
"jz 3f \n\t" // rowNext
"movd (%%edi), %%mm5 \n\t"
"paddusb %%mm6, %%mm5 \n\t"
"movd %%mm5, (%%edi) \n\t"
"addl $4, %%edi \n\t"
"3: \n\t" // rowNext
// advance to the next row
"addl (" ASMSYM(_slModulo) "), %%edi \n\t"
"decl %[xbx] \n\t"
"jnz 0b \n\t" // rowLoop
"emms \n\t"
: [xbx] "=&r" (tmp)
: [ulLight] "g" (ulLight)
: FPU_REGS, "mm5", "mm6", "ecx", "edi", "cc", "memory"
);
#else
#error Write inline assembly for your platform.
#endif
#else
UBYTE* pubLayer = (UBYTE*)_pulLayer;
// for each pixel in the shadow map
for( PIX pixV=0; pixV<_iRowCt; pixV++) {
for( PIX pixU=0; pixU<_iPixCt; pixU++) {
// add the intensity to the pixel
AddToCluster( pubLayer );
pubLayer+=4; // go to the next pixel
} // go to the next row
pubLayer += _slModulo;
}
#endif
}
// apply directional light thru mask to layer
void CLayerMixer::AddMaskDirectional( UBYTE *pubMask, UBYTE ubMask)
{
#if ASMOPT == 1
ULONG ulLight = ByteSwap( lm_colLight);
#if (defined __MSVC_INLINE__)
// prepare some local variables
__asm {
// prepare pointers and variables
movzx edx,B [ubMask]
mov esi,D [pubMask]
mov edi,D [_pulLayer]
mov ebx,D [_iRowCt]
movd mm6,D [ulLight]
rowLoop:
mov ecx,D [_iPixCt]
pixLoop:
// mix underlaying pixels with the constant light color if not shaded
test dl,B [esi]
jz skipLight
movd mm5,D [edi]
paddusb mm5,mm6
movd D [edi],mm5
skipLight:
// advance to next pixel
add edi,4
rol dl,1
adc esi,0
dec ecx
jnz pixLoop
// advance to the next row
add edi,D [_slModulo]
dec ebx
jnz rowLoop
emms
}
#elif (defined __GNU_INLINE__)
ULONG tmp;
__asm__ __volatile__ (
// prepare pointers and variables
"movzbl %[ubMask], %%edx \n\t"
"movl %[pubMask], %%esi \n\t"
"movl (" ASMSYM(_pulLayer) "), %%edi \n\t"
"movl (" ASMSYM(_iRowCt) "), %[xbx]\n\t"
"movd %[ulLight], %%mm6 \n\t"
"0: \n\t" // rowLoop
"movl (" ASMSYM(_iPixCt) "), %%ecx \n\t"
"1: \n\t" // pixLoop
// mix underlaying pixels with the constant light color if not shaded
"testb (%%esi), %%dl \n\t"
"jz 2f \n\t" // skipLight
"movd (%%edi), %%mm5 \n\t"
"paddusb %%mm6, %%mm5 \n\t"
"movd %%mm5, (%%edi) \n\t"
"2: \n\t" // skipLight
// advance to next pixel
"addl $4, %%edi \n\t"
"rolb $1, %%dl \n\t"
"adcl $0, %%esi \n\t"
"decl %%ecx \n\t"
"jnz 1b \n\t" // pixLoop
// advance to the next row
"addl (" ASMSYM(_slModulo) "), %%edi \n\t"
"decl %[xbx] \n\t"
"jnz 0b \n\t" // rowLoop
"emms \n\t"
: [xbx] "=&r" (tmp)
: [ubMask] "m" (ubMask), [pubMask] "g" (pubMask),
[ulLight] "g" (ulLight)
: FPU_REGS, "mm5", "mm6", "ecx", "edx", "esi", "edi",
"cc", "memory"
);
#else
#error Please write inline assembly for your platform.
#endif
#else
UBYTE* pubLayer = (UBYTE*)_pulLayer;
// for each pixel in the shadow map
for( PIX pixV=0; pixV<_iRowCt; pixV++) {
for( PIX pixU=0; pixU<_iPixCt; pixU++) {
// if the point is not masked
if( *pubMask&ubMask) {
// add the intensity to the pixel
AddToCluster( pubLayer);
} // go to the next pixel
pubLayer+=4;
ubMask<<=1;
if( ubMask==0) {
pubMask ++;
ubMask = 1;
}
} // go to the next row
pubLayer += _slModulo;
}
#endif
}
// apply directional light to layer
// (pubMask=NULL for no mask, ubMask = 0xFF for full mask)
void CLayerMixer::AddOneLayerDirectional( CBrushShadowLayer *pbsl, UBYTE *pubMask, UBYTE ubMask)
{
// only if there is color light (ambient is added at initial fill)
if( !(lm_pbpoPolygon->bpo_ulFlags&BPOF_HASDIRECTIONALLIGHT)) return;
_pfWorldEditingProfile.StartTimer(CWorldEditingProfile::PTI_ADDONELAYERDIRECTIONAL);
// determine light influence dimensions
_iPixCt = pbsl->bsl_pixSizeU >>lm_iMipShift;
_iRowCt = pbsl->bsl_pixSizeV >>lm_iMipShift;
PIX pixMinU = pbsl->bsl_pixMinU >>lm_iMipShift;
PIX pixMinV = pbsl->bsl_pixMinV >>lm_iMipShift;
ASSERT( pixMinU==0 && pixMinV==0);
// clamp influence to polygon size
if( _iPixCt > lm_pixPolygonSizeU && pubMask==NULL) _iPixCt = lm_pixPolygonSizeU;
if( _iRowCt > lm_pixPolygonSizeV) _iRowCt = lm_pixPolygonSizeV;
_slModulo = (lm_pixCanvasSizeU-_iPixCt) *BYTES_PER_TEXEL;
_pulLayer = lm_pulShadowMap;
// if there is no influence, do nothing
if( (pbsl->bsl_pixSizeU>>lm_iMipShift)==0 || (pbsl->bsl_pixSizeV>>lm_iMipShift)==0
|| _iPixCt<=0 || _iRowCt<=0) {
_pfWorldEditingProfile.StopTimer(CWorldEditingProfile::PTI_ADDONELAYERDIRECTIONAL);
return;
}
// get the light source of the layer
lm_plsLight = pbsl->bsl_plsLightSource;
const FLOAT3D &vLight = lm_plsLight->ls_penEntity->GetPlacement().pl_PositionVector;
AnglesToDirectionVector( lm_plsLight->ls_penEntity->GetPlacement().pl_OrientationAngle,
lm_vLightDirection);
// calculate intensity
FLOAT fIntensity = 1.0f;
if( !(lm_pbpoPolygon->bpo_ulFlags&BPOF_NOPLANEDIFFUSION)) {
fIntensity = -((lm_pbpoPolygon->bpo_pbplPlane->bpl_plAbsolute)%lm_vLightDirection);
fIntensity = ClampDn( fIntensity, 0.0f);
}
// calculate light color and ambient
lm_colLight = lm_plsLight->GetLightColor();
pbsl->bsl_colLastAnim = lm_colLight;
ULONG ulIntensity = NormFloatToByte(fIntensity);
ulIntensity = (ulIntensity<<CT_RSHIFT)|(ulIntensity<<CT_GSHIFT)|(ulIntensity<<CT_BSHIFT);
lm_colLight = MulColors( lm_colLight, ulIntensity);
lm_colLight = AdjustColor( lm_colLight, _slShdHueShift, _slShdSaturation);
// masked or non-masked?
if( pubMask==NULL) {
// non-masked
AddDirectional();
} else {
// masked
AddMaskDirectional( pubMask, ubMask);
}
// all done
_pfWorldEditingProfile.StopTimer(CWorldEditingProfile::PTI_ADDONELAYERDIRECTIONAL);
}
// clamper helper
static INDEX GetDither(void)
{
shd_iDithering = Clamp( shd_iDithering, 0, 5);
INDEX iDither = shd_iDithering;
if( iDither>2) iDither++;
return iDither;
}
// mix one mip-map
void CLayerMixer::MixOneMipmap(CBrushShadowMap *pbsm, INDEX iMipmap)
{
// remember general data
CalculateData( pbsm, iMipmap);
const BOOL bDynamicOnly = lm_pbpoPolygon->bpo_ulFlags&BPOF_DYNAMICLIGHTSONLY;
// fill with sector ambient
_pfWorldEditingProfile.StartTimer(CWorldEditingProfile::PTI_AMBIENTFILL);
// eventually add ambient component of all directional layers that might contribute
COLOR colAmbient = 0x80808000UL; // overide ambient light color for dynamic lights only
if( !bDynamicOnly) {
colAmbient = AdjustColor( lm_pbpoPolygon->bpo_pbscSector->bsc_colAmbient, _slShdHueShift, _slShdSaturation);
if( lm_pbpoPolygon->bpo_ulFlags&BPOF_HASDIRECTIONALAMBIENT) {
{FOREACHINLIST( CBrushShadowLayer, bsl_lnInShadowMap, lm_pbsmShadowMap->bsm_lhLayers, itbsl) {
CBrushShadowLayer &bsl = *itbsl;
CLightSource &ls = *bsl.bsl_plsLightSource;
ASSERT( &ls!=NULL); if( &ls==NULL) continue; // safety check
if( !(ls.ls_ulFlags&LSF_DIRECTIONAL)) continue; // skip non-directional layers
COLOR col = AdjustColor( ls.GetLightAmbient(), _slShdHueShift, _slShdSaturation);
colAmbient = AddColors( colAmbient, col);
}}
}
} // set initial color
#if (defined USE_PORTABLE_C)
register ULONG count = this->lm_pixCanvasSizeU * this->lm_pixCanvasSizeV;
#if PLATFORM_LITTLEENDIAN
// Forces C fallback; BYTESWAP itself is a no-op on little endian.
register ULONG swapped = BYTESWAP32_unsigned(colAmbient);
#else
STUBBED("actually need byteswap?");
// (uses inline asm on MacOS PowerPC)
register ULONG swapped = colAmbient;
BYTESWAP(swapped);
#endif
for (ULONG *ptr = this->lm_pulShadowMap; count; count--)
{
*ptr = swapped;
ptr++;
}
#elif (defined __MSVC_INLINE__)
__asm {
cld
mov ebx,D [this]
mov ecx,D [ebx].lm_pixCanvasSizeU
imul ecx,D [ebx].lm_pixCanvasSizeV
mov edi,D [ebx].lm_pulShadowMap
mov eax,D [colAmbient]
bswap eax
rep stosd
}
#elif (defined __GNU_INLINE__)
ULONG clob1, clob2, clob3;
__asm__ __volatile__ (
"cld \n\t"
"imull %%esi, %%ecx \n\t"
"bswapl %%eax \n\t"
"rep \n\t"
"stosl \n\t"
: "=a" (clob1), "=c" (clob2), "=D" (clob3)
: "c" (this->lm_pixCanvasSizeU), "S" (this->lm_pixCanvasSizeV),
"a" (colAmbient), "D" (this->lm_pulShadowMap)
: "cc", "memory"
);
#else
#error Please write inline assembly for your platform.
#endif
_pfWorldEditingProfile.StopTimer(CWorldEditingProfile::PTI_AMBIENTFILL);
// find gradient layer
CGradientParameters gpGradient;
BOOL bHasGradient = FALSE;
ULONG ulGradientType = lm_pbpoPolygon->bpo_bppProperties.bpp_ubGradientType;
if( ulGradientType>0) {
CEntity *pen = lm_pbpoPolygon->bpo_pbscSector->bsc_pbmBrushMip->bm_pbrBrush->br_penEntity;
if( pen!=NULL) bHasGradient = pen->GetGradient( ulGradientType, gpGradient);
}
// add gradient if gradient is light
if( bHasGradient && !gpGradient.gp_bDark) AddOneLayerGradient( gpGradient);
// for each shadow layer
lm_pbsmShadowMap->sm_ulFlags &= ~SMF_ANIMATINGLIGHTS;
{FORDELETELIST( CBrushShadowLayer, bsl_lnInShadowMap, lm_pbsmShadowMap->bsm_lhLayers, itbsl)
{
CBrushShadowLayer &bsl = *itbsl;
CLightSource &ls = *bsl.bsl_plsLightSource;
ASSERT( &ls!=NULL); if( &ls==NULL) continue; // safety check
// skip if should not be applied
if( (bDynamicOnly && !(ls.ls_ulFlags&LSF_NONPERSISTENT)) || ls.ls_ulFlags&LSF_DYNAMIC) continue;
// set corresponding shadowmap flag if this is an animating light
if( ls.ls_paoLightAnimation!=NULL) lm_pbsmShadowMap->sm_ulFlags |= SMF_ANIMATINGLIGHTS;
// if the layer is calculated
if( bsl.bsl_pubLayer!=NULL)
{
UBYTE *pub;
UBYTE ubMask;
FindLayerMipmap( itbsl, pub, ubMask);
// add the layer to the shadow map with masking
if( ls.ls_ulFlags&LSF_DIRECTIONAL) {
AddOneLayerDirectional( itbsl, pub, ubMask);
} else {
AddOneLayerPoint( itbsl, pub, ubMask);
}
}
// if the layer is all light
else if( !(bsl.bsl_ulFlags&BSLF_CALCULATED) || (bsl.bsl_ulFlags&BSLF_ALLLIGHT))
{
// add the layer to the shadow map without masking
if( ls.ls_ulFlags&LSF_DIRECTIONAL) {
AddOneLayerDirectional( itbsl, NULL);
} else {
AddOneLayerPoint( itbsl, NULL);
}
}
}}
// if gradient is dark, substract gradient
if( bHasGradient && gpGradient.gp_bDark) AddOneLayerGradient( gpGradient);
// do eventual filtering of shadow layer
shd_iFiltering = Clamp( shd_iFiltering, 0, 6);
if( shd_iFiltering>0) {
FilterBitmap( shd_iFiltering, lm_pulShadowMap, lm_pulShadowMap,
lm_pixPolygonSizeU, lm_pixPolygonSizeV, lm_pixCanvasSizeU, lm_pixCanvasSizeV);
}
// do eventual dithering of shadow layer
const INDEX iDither = GetDither();
if( !(_pGfx->gl_ulFlags&GLF_32BITTEXTURES)) shd_bFineQuality = FALSE;
if( iDither && !(shd_bFineQuality)) {
DitherBitmap( iDither, lm_pulShadowMap, lm_pulShadowMap,
lm_pixPolygonSizeU, lm_pixPolygonSizeV, lm_pixCanvasSizeU, lm_pixCanvasSizeV);
}
}
// copy from static shadow map to dynamic layer
__forceinline void CLayerMixer::CopyShadowLayer(void)
{
#if (defined USE_PORTABLE_C)
memcpy(lm_pulShadowMap, lm_pulStaticShadowMap, lm_pixCanvasSizeU*lm_pixCanvasSizeV*4);
#elif (defined __MSVC_INLINE__)
__asm {
cld
mov ebx,D [this]
mov ecx,D [ebx].lm_pixCanvasSizeU
imul ecx,D [ebx].lm_pixCanvasSizeV
mov esi,D [ebx].lm_pulStaticShadowMap
mov edi,D [ebx].lm_pulShadowMap
rep movsd
}
#elif (defined __GNU_INLINE__)
ULONG clob1, clob2, clob3;
__asm__ __volatile__ (
"cld \n\t"
"imull %%eax, %%ecx \n\t"
"rep \n\t"
"movsl \n\t"
: "=c" (clob1), "=S" (clob2), "=D" (clob3)
: "c" (this->lm_pixCanvasSizeU), "a" (this->lm_pixCanvasSizeV),
"S" (this->lm_pulStaticShadowMap), "D" (this->lm_pulShadowMap)
: "cc", "memory"
);
#else
#error Please write inline assembly for your platform.
#endif
}
// copy from static shadow map to dynamic layer
__forceinline void CLayerMixer::FillShadowLayer( COLOR col)
{
#if (defined USE_PORTABLE_C)
DWORD* dst = (DWORD*)lm_pulShadowMap;
int n = lm_pixCanvasSizeU*lm_pixCanvasSizeV;
DWORD color = __builtin_bswap32(col);
while(n--) {*(dst++)=color;}
#elif (defined __MSVC_INLINE__)
__asm {
cld
mov ebx,D [this]
mov ecx,D [ebx].lm_pixCanvasSizeU
imul ecx,D [ebx].lm_pixCanvasSizeV
mov edi,D [ebx].lm_pulShadowMap
mov eax,D [col]
bswap eax // convert to R,G,B,A memory format!
rep stosd
}
#elif (defined __GNU_INLINE__)
ULONG clob1, clob2, clob3;
__asm__ __volatile__ (
"cld \n\t"
"imull %%edx, %%ecx \n\t"
"bswapl %%eax \n\t" // convert to R,G,B,A memory format!
"rep \n\t"
"stosl \n\t"
: "=a" (clob1), "=c" (clob2), "=D" (clob3)
: "c" (this->lm_pixCanvasSizeU), "d" (this->lm_pixCanvasSizeV),
"a" (col), "D" (this->lm_pulShadowMap)
: "cc", "memory"
);
#else
#error Please write inline assembly for your platform.
#endif
}
// mix dynamic lights
void CLayerMixer::MixOneMipmapDynamic( CBrushShadowMap *pbsm, INDEX iMipmap)
{
// remember general data
CalculateData( pbsm, iMipmap);
// if static shadow map is all flat
if( pbsm->sm_pulCachedShadowMap==&pbsm->sm_colFlat) {
// just fill dynamic shadow map with flat color
FillShadowLayer( pbsm->sm_colFlat);
} // if not flat
else {
// copy static layer
CopyShadowLayer();
}
// for each shadow layer
{FORDELETELIST( CBrushShadowLayer, bsl_lnInShadowMap, lm_pbsmShadowMap->bsm_lhLayers, itbsl)
{ // the layer's light source must be valid
CBrushShadowLayer &bsl = *itbsl;
CLightSource &ls = *bsl.bsl_plsLightSource;
ASSERT( &ls!=NULL);
if( !(ls.ls_ulFlags&LSF_DYNAMIC)) continue;
COLOR colLight = ls.GetLightColor() & ~CT_AMASK;
if( IsBlack(colLight)) continue;
// apply one layer
colLight = AdjustColor( colLight, _slShdHueShift, _slShdSaturation);
AddOneLayerPoint( itbsl, NULL);
}}
}
// constructor
CLayerMixer::CLayerMixer( CBrushShadowMap *pbsm, INDEX iFirstMip, INDEX iLastMip, BOOL bDynamic)
{
lm_bDynamic = bDynamic;
if( bDynamic) {
// check dynamic layers for complete blackness
BOOL bAllBlack = TRUE;
pbsm->sm_ulFlags &= ~SMF_DYNAMICBLACK;
{FORDELETELIST( CBrushShadowLayer, bsl_lnInShadowMap, pbsm->bsm_lhLayers, itbsl) {
CLightSource &ls = *itbsl->bsl_plsLightSource;
ASSERT( &ls!=NULL);
if( !(ls.ls_ulFlags&LSF_DYNAMIC)) continue;
COLOR colLight = ls.GetLightColor() & ~CT_AMASK;
itbsl->bsl_colLastAnim = colLight;
if( !IsBlack(colLight)) bAllBlack = FALSE; // must continue because of layer info update (light anim and such stuff)
}}
// skip mixing if dynamic layers were all black
if( bAllBlack) {
pbsm->sm_ulFlags |= SMF_DYNAMICBLACK;
return;
}
// need to mix in
for( INDEX iMipmap=iFirstMip; iMipmap<=iLastMip; iMipmap++) MixOneMipmapDynamic( pbsm, iMipmap);
}
// mix static layers
else {
for( INDEX iMipmap=iFirstMip; iMipmap<=iLastMip; iMipmap++) MixOneMipmap( pbsm, iMipmap);
}
}
// mix all layers into cached shadow map
void CBrushShadowMap::MixLayers( INDEX iFirstMip, INDEX iLastMip, BOOL bDynamic/*=FALSE*/)
{
_sfStats.StartTimer( CStatForm::STI_SHADOWUPDATE);
_pfWorldEditingProfile.StartTimer( CWorldEditingProfile::PTI_MIXLAYERS);
// mix the layers with a shadow mixer
CLayerMixer lmMixer( this, iFirstMip, iLastMip, bDynamic);
_pfWorldEditingProfile.StopTimer( CWorldEditingProfile::PTI_MIXLAYERS);
_sfStats.StopTimer( CStatForm::STI_SHADOWUPDATE);
}
Reference in New Issue View Git Blame Copy Permalink