COSMETICS: tabs --> spaces, some prettyprinting

Originally committed as revision 4764 to svn://svn.ffmpeg.org/ffmpeg/trunk

5 files changed, 4334 insertions, 4334 deletions

diff --git a/libavcodec/libpostproc/postprocess.c b/libavcodec/libpostproc/postprocess.c
index 21ee6f6a6e..141e5e2b89 100644
--- a/libavcodec/libpostproc/postprocess.c
+++ b/libavcodec/libpostproc/postprocess.c
@@ -24,26 +24,26 @@
  */
 
 /*
-			C	MMX	MMX2	3DNow	AltiVec
-isVertDC		Ec	Ec			Ec
-isVertMinMaxOk		Ec	Ec			Ec
-doVertLowPass		E		e	e	Ec
-doVertDefFilter		Ec	Ec	e	e	Ec
-isHorizDC		Ec	Ec			Ec
-isHorizMinMaxOk		a	E			Ec
-doHorizLowPass		E		e	e	Ec
-doHorizDefFilter	Ec	Ec	e	e	Ec
-do_a_deblock		Ec	E	Ec	E
-deRing			E		e	e*	Ecp
-Vertical RKAlgo1	E		a	a
-Horizontal RKAlgo1			a	a
-Vertical X1#		a		E	E
-Horizontal X1#		a		E	E
-LinIpolDeinterlace	e		E	E*
-CubicIpolDeinterlace	a		e	e*
-LinBlendDeinterlace	e		E	E*
-MedianDeinterlace#	E	Ec	Ec
-TempDeNoiser#		E		e	e	Ec
+                        C       MMX     MMX2    3DNow   AltiVec
+isVertDC                Ec      Ec                      Ec
+isVertMinMaxOk          Ec      Ec                      Ec
+doVertLowPass           E               e       e       Ec
+doVertDefFilter         Ec      Ec      e       e       Ec
+isHorizDC               Ec      Ec                      Ec
+isHorizMinMaxOk         a       E                       Ec
+doHorizLowPass          E               e       e       Ec
+doHorizDefFilter        Ec      Ec      e       e       Ec
+do_a_deblock            Ec      E       Ec      E
+deRing                  E               e       e*      Ecp
+Vertical RKAlgo1        E               a       a
+Horizontal RKAlgo1                      a       a
+Vertical X1#            a               E       E
+Horizontal X1#          a               E       E
+LinIpolDeinterlace      e               E       E*
+CubicIpolDeinterlace    a               e       e*
+LinBlendDeinterlace     e               E       E*
+MedianDeinterlace#      E       Ec      Ec
+TempDeNoiser#           E               e       e       Ec
 
 * i dont have a 3dnow CPU -> its untested, but noone said it doesnt work so it seems to work
 # more or less selfinvented filters so the exactness isnt too meaningfull
@@ -61,7 +61,7 @@ unroll stuff if instructions depend too much on the prior one
 move YScale thing to the end instead of fixing QP
 write a faster and higher quality deblocking filter :)
 make the mainloop more flexible (variable number of blocks at once
-	(the if/else stuff per block is slowing things down)
+        (the if/else stuff per block is slowing things down)
 compare the quality & speed of all filters
 split this huge file
 optimize c versions
@@ -120,14 +120,14 @@ try to unroll inner for(x=0 ... loop to avoid these damn if(x ... checks
 #endif
 
 #if defined(ARCH_X86) || defined(ARCH_X86_64)
-static uint64_t __attribute__((aligned(8))) attribute_used w05=		0x0005000500050005LL;
-static uint64_t __attribute__((aligned(8))) attribute_used w04=		0x0004000400040004LL;
-static uint64_t __attribute__((aligned(8))) attribute_used w20=		0x0020002000200020LL;
-static uint64_t __attribute__((aligned(8))) attribute_used b00= 		0x0000000000000000LL;
-static uint64_t __attribute__((aligned(8))) attribute_used b01= 		0x0101010101010101LL;
-static uint64_t __attribute__((aligned(8))) attribute_used b02= 		0x0202020202020202LL;
-static uint64_t __attribute__((aligned(8))) attribute_used b08= 		0x0808080808080808LL;
-static uint64_t __attribute__((aligned(8))) attribute_used b80= 		0x8080808080808080LL;
+static uint64_t __attribute__((aligned(8))) attribute_used w05= 0x0005000500050005LL;
+static uint64_t __attribute__((aligned(8))) attribute_used w04= 0x0004000400040004LL;
+static uint64_t __attribute__((aligned(8))) attribute_used w20= 0x0020002000200020LL;
+static uint64_t __attribute__((aligned(8))) attribute_used b00= 0x0000000000000000LL;
+static uint64_t __attribute__((aligned(8))) attribute_used b01= 0x0101010101010101LL;
+static uint64_t __attribute__((aligned(8))) attribute_used b02= 0x0202020202020202LL;
+static uint64_t __attribute__((aligned(8))) attribute_used b08= 0x0808080808080808LL;
+static uint64_t __attribute__((aligned(8))) attribute_used b80= 0x8080808080808080LL;
 #endif
 
 static uint8_t clip_table[3*256];
@@ -140,65 +140,65 @@ static const int attribute_used deringThreshold= 20;
 
 static struct PPFilter filters[]=
 {
-	{"hb", "hdeblock", 		1, 1, 3, H_DEBLOCK},
-	{"vb", "vdeblock", 		1, 2, 4, V_DEBLOCK},
-/*	{"hr", "rkhdeblock", 		1, 1, 3, H_RK1_FILTER},
-	{"vr", "rkvdeblock", 		1, 2, 4, V_RK1_FILTER},*/
-	{"h1", "x1hdeblock", 		1, 1, 3, H_X1_FILTER},
-	{"v1", "x1vdeblock", 		1, 2, 4, V_X1_FILTER},
-	{"ha", "ahdeblock", 		1, 1, 3, H_A_DEBLOCK},
-	{"va", "avdeblock", 		1, 2, 4, V_A_DEBLOCK},
-	{"dr", "dering", 		1, 5, 6, DERING},
-	{"al", "autolevels", 		0, 1, 2, LEVEL_FIX},
-	{"lb", "linblenddeint", 	1, 1, 4, LINEAR_BLEND_DEINT_FILTER},
-	{"li", "linipoldeint", 		1, 1, 4, LINEAR_IPOL_DEINT_FILTER},
-	{"ci", "cubicipoldeint",	1, 1, 4, CUBIC_IPOL_DEINT_FILTER},
-	{"md", "mediandeint", 		1, 1, 4, MEDIAN_DEINT_FILTER},
-	{"fd", "ffmpegdeint", 		1, 1, 4, FFMPEG_DEINT_FILTER},
-	{"l5", "lowpass5", 		1, 1, 4, LOWPASS5_DEINT_FILTER},
-	{"tn", "tmpnoise", 		1, 7, 8, TEMP_NOISE_FILTER},
-	{"fq", "forcequant", 		1, 0, 0, FORCE_QUANT},
-	{NULL, NULL,0,0,0,0} //End Marker
+        {"hb", "hdeblock",              1, 1, 3, H_DEBLOCK},
+        {"vb", "vdeblock",              1, 2, 4, V_DEBLOCK},
+/*      {"hr", "rkhdeblock",            1, 1, 3, H_RK1_FILTER},
+        {"vr", "rkvdeblock",            1, 2, 4, V_RK1_FILTER},*/
+        {"h1", "x1hdeblock",            1, 1, 3, H_X1_FILTER},
+        {"v1", "x1vdeblock",            1, 2, 4, V_X1_FILTER},
+        {"ha", "ahdeblock",             1, 1, 3, H_A_DEBLOCK},
+        {"va", "avdeblock",             1, 2, 4, V_A_DEBLOCK},
+        {"dr", "dering",                1, 5, 6, DERING},
+        {"al", "autolevels",            0, 1, 2, LEVEL_FIX},
+        {"lb", "linblenddeint",         1, 1, 4, LINEAR_BLEND_DEINT_FILTER},
+        {"li", "linipoldeint",          1, 1, 4, LINEAR_IPOL_DEINT_FILTER},
+        {"ci", "cubicipoldeint",        1, 1, 4, CUBIC_IPOL_DEINT_FILTER},
+        {"md", "mediandeint",           1, 1, 4, MEDIAN_DEINT_FILTER},
+        {"fd", "ffmpegdeint",           1, 1, 4, FFMPEG_DEINT_FILTER},
+        {"l5", "lowpass5",              1, 1, 4, LOWPASS5_DEINT_FILTER},
+        {"tn", "tmpnoise",              1, 7, 8, TEMP_NOISE_FILTER},
+        {"fq", "forcequant",            1, 0, 0, FORCE_QUANT},
+        {NULL, NULL,0,0,0,0} //End Marker
 };
 
 static char *replaceTable[]=
 {
-	"default", 	"hdeblock:a,vdeblock:a,dering:a",
-	"de", 		"hdeblock:a,vdeblock:a,dering:a",
-	"fast", 	"x1hdeblock:a,x1vdeblock:a,dering:a",
-	"fa", 		"x1hdeblock:a,x1vdeblock:a,dering:a",
-	"ac", 		"ha:a:128:7,va:a,dering:a",
-	NULL //End Marker
+        "default",      "hdeblock:a,vdeblock:a,dering:a",
+        "de",           "hdeblock:a,vdeblock:a,dering:a",
+        "fast",         "x1hdeblock:a,x1vdeblock:a,dering:a",
+        "fa",           "x1hdeblock:a,x1vdeblock:a,dering:a",
+        "ac",           "ha:a:128:7,va:a,dering:a",
+        NULL //End Marker
 };
 
 
 #if defined(ARCH_X86) || defined(ARCH_X86_64)
 static inline void prefetchnta(void *p)
 {
-	asm volatile(	"prefetchnta (%0)\n\t"
-		: : "r" (p)
-	);
+        asm volatile(   "prefetchnta (%0)\n\t"
+                : : "r" (p)
+        );
 }
 
 static inline void prefetcht0(void *p)
 {
-	asm volatile(	"prefetcht0 (%0)\n\t"
-		: : "r" (p)
-	);
+        asm volatile(   "prefetcht0 (%0)\n\t"
+                : : "r" (p)
+        );
 }
 
 static inline void prefetcht1(void *p)
 {
-	asm volatile(	"prefetcht1 (%0)\n\t"
-		: : "r" (p)
-	);
+        asm volatile(   "prefetcht1 (%0)\n\t"
+                : : "r" (p)
+        );
 }
 
 static inline void prefetcht2(void *p)
 {
-	asm volatile(	"prefetcht2 (%0)\n\t"
-		: : "r" (p)
-	);
+        asm volatile(   "prefetcht2 (%0)\n\t"
+                : : "r" (p)
+        );
 }
 #endif
 
@@ -209,171 +209,171 @@ static inline void prefetcht2(void *p)
  */
 static inline int isHorizDC_C(uint8_t src[], int stride, PPContext *c)
 {
-	int numEq= 0;
-	int y;
-	const int dcOffset= ((c->nonBQP*c->ppMode.baseDcDiff)>>8) + 1;
-	const int dcThreshold= dcOffset*2 + 1;
-
-	for(y=0; y<BLOCK_SIZE; y++)
-	{
-		if(((unsigned)(src[0] - src[1] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[1] - src[2] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[2] - src[3] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[3] - src[4] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[4] - src[5] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[5] - src[6] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[6] - src[7] + dcOffset)) < dcThreshold) numEq++;
-		src+= stride;
-	}
-	return numEq > c->ppMode.flatnessThreshold;
+        int numEq= 0;
+        int y;
+        const int dcOffset= ((c->nonBQP*c->ppMode.baseDcDiff)>>8) + 1;
+        const int dcThreshold= dcOffset*2 + 1;
+
+        for(y=0; y<BLOCK_SIZE; y++)
+        {
+                if(((unsigned)(src[0] - src[1] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[1] - src[2] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[2] - src[3] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[3] - src[4] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[4] - src[5] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[5] - src[6] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[6] - src[7] + dcOffset)) < dcThreshold) numEq++;
+                src+= stride;
+        }
+        return numEq > c->ppMode.flatnessThreshold;
 }
 
 /**
  * Check if the middle 8x8 Block in the given 8x16 block is flat
  */
 static inline int isVertDC_C(uint8_t src[], int stride, PPContext *c){
-	int numEq= 0;
-	int y;
-	const int dcOffset= ((c->nonBQP*c->ppMode.baseDcDiff)>>8) + 1;
-	const int dcThreshold= dcOffset*2 + 1;
-
-	src+= stride*4; // src points to begin of the 8x8 Block
-	for(y=0; y<BLOCK_SIZE-1; y++)
-	{
-		if(((unsigned)(src[0] - src[0+stride] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[1] - src[1+stride] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[2] - src[2+stride] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[3] - src[3+stride] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[4] - src[4+stride] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[5] - src[5+stride] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[6] - src[6+stride] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[7] - src[7+stride] + dcOffset)) < dcThreshold) numEq++;
-		src+= stride;
-	}
-	return numEq > c->ppMode.flatnessThreshold;
+        int numEq= 0;
+        int y;
+        const int dcOffset= ((c->nonBQP*c->ppMode.baseDcDiff)>>8) + 1;
+        const int dcThreshold= dcOffset*2 + 1;
+
+        src+= stride*4; // src points to begin of the 8x8 Block
+        for(y=0; y<BLOCK_SIZE-1; y++)
+        {
+                if(((unsigned)(src[0] - src[0+stride] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[1] - src[1+stride] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[2] - src[2+stride] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[3] - src[3+stride] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[4] - src[4+stride] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[5] - src[5+stride] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[6] - src[6+stride] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[7] - src[7+stride] + dcOffset)) < dcThreshold) numEq++;
+                src+= stride;
+        }
+        return numEq > c->ppMode.flatnessThreshold;
 }
 
 static inline int isHorizMinMaxOk_C(uint8_t src[], int stride, int QP)
 {
-	int i;
+        int i;
 #if 1
-	for(i=0; i<2; i++){
-		if((unsigned)(src[0] - src[5] + 2*QP) > 4*QP) return 0;
-		src += stride;
-		if((unsigned)(src[2] - src[7] + 2*QP) > 4*QP) return 0;
-		src += stride;
-		if((unsigned)(src[4] - src[1] + 2*QP) > 4*QP) return 0;
-		src += stride;
-		if((unsigned)(src[6] - src[3] + 2*QP) > 4*QP) return 0;
-		src += stride;
-	}
+        for(i=0; i<2; i++){
+                if((unsigned)(src[0] - src[5] + 2*QP) > 4*QP) return 0;
+                src += stride;
+                if((unsigned)(src[2] - src[7] + 2*QP) > 4*QP) return 0;
+                src += stride;
+                if((unsigned)(src[4] - src[1] + 2*QP) > 4*QP) return 0;
+                src += stride;
+                if((unsigned)(src[6] - src[3] + 2*QP) > 4*QP) return 0;
+                src += stride;
+        }
 #else
-	for(i=0; i<8; i++){
-		if((unsigned)(src[0] - src[7] + 2*QP) > 4*QP) return 0;
-		src += stride;
-	}
+        for(i=0; i<8; i++){
+                if((unsigned)(src[0] - src[7] + 2*QP) > 4*QP) return 0;
+                src += stride;
+        }
 #endif
-	return 1;
+        return 1;
 }
 
 static inline int isVertMinMaxOk_C(uint8_t src[], int stride, int QP)
 {
 #if 1
 #if 1
-	int x;
-	src+= stride*4;
-	for(x=0; x<BLOCK_SIZE; x+=4)
-	{
-		if((unsigned)(src[  x + 0*stride] - src[  x + 5*stride] + 2*QP) > 4*QP) return 0;
-		if((unsigned)(src[1+x + 2*stride] - src[1+x + 7*stride] + 2*QP) > 4*QP) return 0;
-		if((unsigned)(src[2+x + 4*stride] - src[2+x + 1*stride] + 2*QP) > 4*QP) return 0;
-		if((unsigned)(src[3+x + 6*stride] - src[3+x + 3*stride] + 2*QP) > 4*QP) return 0;
-	}
+        int x;
+        src+= stride*4;
+        for(x=0; x<BLOCK_SIZE; x+=4)
+        {
+                if((unsigned)(src[  x + 0*stride] - src[  x + 5*stride] + 2*QP) > 4*QP) return 0;
+                if((unsigned)(src[1+x + 2*stride] - src[1+x + 7*stride] + 2*QP) > 4*QP) return 0;
+                if((unsigned)(src[2+x + 4*stride] - src[2+x + 1*stride] + 2*QP) > 4*QP) return 0;
+                if((unsigned)(src[3+x + 6*stride] - src[3+x + 3*stride] + 2*QP) > 4*QP) return 0;
+        }
 #else
-	int x;
-	src+= stride*3;
-	for(x=0; x<BLOCK_SIZE; x++)
-	{
-		if((unsigned)(src[x + stride] - src[x + (stride<<3)] + 2*QP) > 4*QP) return 0;
-	}
+        int x;
+        src+= stride*3;
+        for(x=0; x<BLOCK_SIZE; x++)
+        {
+                if((unsigned)(src[x + stride] - src[x + (stride<<3)] + 2*QP) > 4*QP) return 0;
+        }
 #endif
-	return 1;
+        return 1;
 #else
-	int x;
-	src+= stride*4;
-	for(x=0; x<BLOCK_SIZE; x++)
-	{
-		int min=255;
-		int max=0;
-		int y;
-		for(y=0; y<8; y++){
-			int v= src[x + y*stride];
-			if(v>max) max=v;
-			if(v<min) min=v;
-		}
-		if(max-min > 2*QP) return 0;
-	}
-	return 1;
+        int x;
+        src+= stride*4;
+        for(x=0; x<BLOCK_SIZE; x++)
+        {
+                int min=255;
+                int max=0;
+                int y;
+                for(y=0; y<8; y++){
+                        int v= src[x + y*stride];
+                        if(v>max) max=v;
+                        if(v<min) min=v;
+                }
+                if(max-min > 2*QP) return 0;
+        }
+        return 1;
 #endif
 }
 
 static inline int horizClassify_C(uint8_t src[], int stride, PPContext *c){
-	if( isHorizDC_C(src, stride, c) ){
-		if( isHorizMinMaxOk_C(src, stride, c->QP) )
-			return 1;
-		else
-			return 0;
-	}else{
-		return 2;
-	}
+        if( isHorizDC_C(src, stride, c) ){
+                if( isHorizMinMaxOk_C(src, stride, c->QP) )
+                        return 1;
+                else
+                        return 0;
+        }else{
+                return 2;
+        }
 }
 
 static inline int vertClassify_C(uint8_t src[], int stride, PPContext *c){
-	if( isVertDC_C(src, stride, c) ){
-		if( isVertMinMaxOk_C(src, stride, c->QP) )
-			return 1;
-		else
-			return 0;
-	}else{
-		return 2;
-	}
+        if( isVertDC_C(src, stride, c) ){
+                if( isVertMinMaxOk_C(src, stride, c->QP) )
+                        return 1;
+                else
+                        return 0;
+        }else{
+                return 2;
+        }
 }
 
 static inline void doHorizDefFilter_C(uint8_t dst[], int stride, PPContext *c)
 {
-	int y;
-	for(y=0; y<BLOCK_SIZE; y++)
-	{
-		const int middleEnergy= 5*(dst[4] - dst[3]) + 2*(dst[2] - dst[5]);
-
-		if(ABS(middleEnergy) < 8*c->QP)
-		{
-			const int q=(dst[3] - dst[4])/2;
-			const int leftEnergy=  5*(dst[2] - dst[1]) + 2*(dst[0] - dst[3]);
-			const int rightEnergy= 5*(dst[6] - dst[5]) + 2*(dst[4] - dst[7]);
-
-			int d= ABS(middleEnergy) - MIN( ABS(leftEnergy), ABS(rightEnergy) );
-			d= MAX(d, 0);
-
-			d= (5*d + 32) >> 6;
-			d*= SIGN(-middleEnergy);
-
-			if(q>0)
-			{
-				d= d<0 ? 0 : d;
-				d= d>q ? q : d;
-			}
-			else
-			{
-				d= d>0 ? 0 : d;
-				d= d<q ? q : d;
-			}
-
-        		dst[3]-= d;
-	        	dst[4]+= d;
-		}
-		dst+= stride;
-	}
+        int y;
+        for(y=0; y<BLOCK_SIZE; y++)
+        {
+                const int middleEnergy= 5*(dst[4] - dst[3]) + 2*(dst[2] - dst[5]);
+
+                if(ABS(middleEnergy) < 8*c->QP)
+                {
+                        const int q=(dst[3] - dst[4])/2;
+                        const int leftEnergy=  5*(dst[2] - dst[1]) + 2*(dst[0] - dst[3]);
+                        const int rightEnergy= 5*(dst[6] - dst[5]) + 2*(dst[4] - dst[7]);
+
+                        int d= ABS(middleEnergy) - MIN( ABS(leftEnergy), ABS(rightEnergy) );
+                        d= MAX(d, 0);
+
+                        d= (5*d + 32) >> 6;
+                        d*= SIGN(-middleEnergy);
+
+                        if(q>0)
+                        {
+                                d= d<0 ? 0 : d;
+                                d= d>q ? q : d;
+                        }
+                        else
+                        {
+                                d= d>0 ? 0 : d;
+                                d= d<q ? q : d;
+                        }
+
+                        dst[3]-= d;
+                        dst[4]+= d;
+                }
+                dst+= stride;
+        }
 }
 
 /**
@@ -382,35 +382,35 @@ static inline void doHorizDefFilter_C(uint8_t dst[], int stride, PPContext *c)
  */
 static inline void doHorizLowPass_C(uint8_t dst[], int stride, PPContext *c)
 {
-	int y;
-	for(y=0; y<BLOCK_SIZE; y++)
-	{
-		const int first= ABS(dst[-1] - dst[0]) < c->QP ? dst[-1] : dst[0];
-		const int last= ABS(dst[8] - dst[7]) < c->QP ? dst[8] : dst[7];
-
-		int sums[10];
-		sums[0] = 4*first + dst[0] + dst[1] + dst[2] + 4;
-		sums[1] = sums[0] - first  + dst[3];
-		sums[2] = sums[1] - first  + dst[4];
-		sums[3] = sums[2] - first  + dst[5];
-		sums[4] = sums[3] - first  + dst[6];
-		sums[5] = sums[4] - dst[0] + dst[7];
-		sums[6] = sums[5] - dst[1] + last;
-		sums[7] = sums[6] - dst[2] + last;
-		sums[8] = sums[7] - dst[3] + last;
-		sums[9] = sums[8] - dst[4] + last;
-
-		dst[0]= (sums[0] + sums[2] + 2*dst[0])>>4;
-		dst[1]= (sums[1] + sums[3] + 2*dst[1])>>4;
-		dst[2]= (sums[2] + sums[4] + 2*dst[2])>>4;
-		dst[3]= (sums[3] + sums[5] + 2*dst[3])>>4;
-		dst[4]= (sums[4] + sums[6] + 2*dst[4])>>4;
-		dst[5]= (sums[5] + sums[7] + 2*dst[5])>>4;
-		dst[6]= (sums[6] + sums[8] + 2*dst[6])>>4;
-		dst[7]= (sums[7] + sums[9] + 2*dst[7])>>4;
-
-		dst+= stride;
-	}
+        int y;
+        for(y=0; y<BLOCK_SIZE; y++)
+        {
+                const int first= ABS(dst[-1] - dst[0]) < c->QP ? dst[-1] : dst[0];
+                const int last= ABS(dst[8] - dst[7]) < c->QP ? dst[8] : dst[7];
+
+                int sums[10];
+                sums[0] = 4*first + dst[0] + dst[1] + dst[2] + 4;
+                sums[1] = sums[0] - first  + dst[3];
+                sums[2] = sums[1] - first  + dst[4];
+                sums[3] = sums[2] - first  + dst[5];
+                sums[4] = sums[3] - first  + dst[6];
+                sums[5] = sums[4] - dst[0] + dst[7];
+                sums[6] = sums[5] - dst[1] + last;
+                sums[7] = sums[6] - dst[2] + last;
+                sums[8] = sums[7] - dst[3] + last;
+                sums[9] = sums[8] - dst[4] + last;
+
+                dst[0]= (sums[0] + sums[2] + 2*dst[0])>>4;
+                dst[1]= (sums[1] + sums[3] + 2*dst[1])>>4;
+                dst[2]= (sums[2] + sums[4] + 2*dst[2])>>4;
+                dst[3]= (sums[3] + sums[5] + 2*dst[3])>>4;
+                dst[4]= (sums[4] + sums[6] + 2*dst[4])>>4;
+                dst[5]= (sums[5] + sums[7] + 2*dst[5])>>4;
+                dst[6]= (sums[6] + sums[8] + 2*dst[6])>>4;
+                dst[7]= (sums[7] + sums[9] + 2*dst[7])>>4;
+
+                dst+= stride;
+        }
 }
 
 /**
@@ -423,161 +423,161 @@ static inline void doHorizLowPass_C(uint8_t dst[], int stride, PPContext *c)
  */
 static inline void horizX1Filter(uint8_t *src, int stride, int QP)
 {
-	int y;
-	static uint64_t *lut= NULL;
-	if(lut==NULL)
-	{
-		int i;
-		lut= (uint64_t*)memalign(8, 256*8);
-		for(i=0; i<256; i++)
-		{
-			int v= i < 128 ? 2*i : 2*(i-256);
+        int y;
+        static uint64_t *lut= NULL;
+        if(lut==NULL)
+        {
+                int i;
+                lut= (uint64_t*)memalign(8, 256*8);
+                for(i=0; i<256; i++)
+                {
+                        int v= i < 128 ? 2*i : 2*(i-256);
 /*
 //Simulate 112242211 9-Tap filter
-			uint64_t a= (v/16) & 0xFF;
-			uint64_t b= (v/8) & 0xFF;
-			uint64_t c= (v/4) & 0xFF;
-			uint64_t d= (3*v/8) & 0xFF;
+                        uint64_t a= (v/16) & 0xFF;
+                        uint64_t b= (v/8) & 0xFF;
+                        uint64_t c= (v/4) & 0xFF;
+                        uint64_t d= (3*v/8) & 0xFF;
 */
 //Simulate piecewise linear interpolation
-			uint64_t a= (v/16) & 0xFF;
-			uint64_t b= (v*3/16) & 0xFF;
-			uint64_t c= (v*5/16) & 0xFF;
-			uint64_t d= (7*v/16) & 0xFF;
-			uint64_t A= (0x100 - a)&0xFF;
-			uint64_t B= (0x100 - b)&0xFF;
-			uint64_t C= (0x100 - c)&0xFF;
-			uint64_t D= (0x100 - c)&0xFF;
-
-			lut[i]   = (a<<56) | (b<<48) | (c<<40) | (d<<32) |
-				(D<<24) | (C<<16) | (B<<8) | (A);
-			//lut[i] = (v<<32) | (v<<24);
-		}
-	}
-
-	for(y=0; y<BLOCK_SIZE; y++)
-	{
-		int a= src[1] - src[2];
-		int b= src[3] - src[4];
-		int c= src[5] - src[6];
-
-		int d= MAX(ABS(b) - (ABS(a) + ABS(c))/2, 0);
-
-		if(d < QP)
-		{
-			int v = d * SIGN(-b);
-
-			src[1] +=v/8;
-			src[2] +=v/4;
-			src[3] +=3*v/8;
-			src[4] -=3*v/8;
-			src[5] -=v/4;
-			src[6] -=v/8;
-
-		}
-		src+=stride;
-	}
+                        uint64_t a= (v/16) & 0xFF;
+                        uint64_t b= (v*3/16) & 0xFF;
+                        uint64_t c= (v*5/16) & 0xFF;
+                        uint64_t d= (7*v/16) & 0xFF;
+                        uint64_t A= (0x100 - a)&0xFF;
+                        uint64_t B= (0x100 - b)&0xFF;
+                        uint64_t C= (0x100 - c)&0xFF;
+                        uint64_t D= (0x100 - c)&0xFF;
+
+                        lut[i]   = (a<<56) | (b<<48) | (c<<40) | (d<<32) |
+                                (D<<24) | (C<<16) | (B<<8) | (A);
+                        //lut[i] = (v<<32) | (v<<24);
+                }
+        }
+
+        for(y=0; y<BLOCK_SIZE; y++)
+        {
+                int a= src[1] - src[2];
+                int b= src[3] - src[4];
+                int c= src[5] - src[6];
+
+                int d= MAX(ABS(b) - (ABS(a) + ABS(c))/2, 0);
+
+                if(d < QP)
+                {
+                        int v = d * SIGN(-b);
+
+                        src[1] +=v/8;
+                        src[2] +=v/4;
+                        src[3] +=3*v/8;
+                        src[4] -=3*v/8;
+                        src[5] -=v/4;
+                        src[6] -=v/8;
+
+                }
+                src+=stride;
+        }
 }
 
 /**
  * accurate deblock filter
  */
 static always_inline void do_a_deblock_C(uint8_t *src, int step, int stride, PPContext *c){
-	int y;
-	const int QP= c->QP;
-	const int dcOffset= ((c->nonBQP*c->ppMode.baseDcDiff)>>8) + 1;
-	const int dcThreshold= dcOffset*2 + 1;
+        int y;
+        const int QP= c->QP;
+        const int dcOffset= ((c->nonBQP*c->ppMode.baseDcDiff)>>8) + 1;
+        const int dcThreshold= dcOffset*2 + 1;
 //START_TIMER
-	src+= step*4; // src points to begin of the 8x8 Block
-	for(y=0; y<8; y++){
-		int numEq= 0;
-
-		if(((unsigned)(src[-1*step] - src[0*step] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[ 0*step] - src[1*step] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[ 1*step] - src[2*step] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[ 2*step] - src[3*step] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[ 3*step] - src[4*step] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[ 4*step] - src[5*step] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[ 5*step] - src[6*step] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[ 6*step] - src[7*step] + dcOffset)) < dcThreshold) numEq++;
-		if(((unsigned)(src[ 7*step] - src[8*step] + dcOffset)) < dcThreshold) numEq++;
-		if(numEq > c->ppMode.flatnessThreshold){
-			int min, max, x;
-
-			if(src[0] > src[step]){
-			    max= src[0];
-			    min= src[step];
-			}else{
-			    max= src[step];
-			    min= src[0];
-			}
-			for(x=2; x<8; x+=2){
-				if(src[x*step] > src[(x+1)*step]){
-					if(src[x    *step] > max) max= src[ x   *step];
-					if(src[(x+1)*step] < min) min= src[(x+1)*step];
-				}else{
-					if(src[(x+1)*step] > max) max= src[(x+1)*step];
-					if(src[ x   *step] < min) min= src[ x   *step];
-				}
-			}
-			if(max-min < 2*QP){
-				const int first= ABS(src[-1*step] - src[0]) < QP ? src[-1*step] : src[0];
-				const int last= ABS(src[8*step] - src[7*step]) < QP ? src[8*step] : src[7*step];
-
-				int sums[10];
-				sums[0] = 4*first + src[0*step] + src[1*step] + src[2*step] + 4;
-				sums[1] = sums[0] - first       + src[3*step];
-				sums[2] = sums[1] - first       + src[4*step];
-				sums[3] = sums[2] - first       + src[5*step];
-				sums[4] = sums[3] - first       + src[6*step];
-				sums[5] = sums[4] - src[0*step] + src[7*step];
-				sums[6] = sums[5] - src[1*step] + last;
-				sums[7] = sums[6] - src[2*step] + last;
-				sums[8] = sums[7] - src[3*step] + last;
-				sums[9] = sums[8] - src[4*step] + last;
-
-				src[0*step]= (sums[0] + sums[2] + 2*src[0*step])>>4;
-				src[1*step]= (sums[1] + sums[3] + 2*src[1*step])>>4;
-				src[2*step]= (sums[2] + sums[4] + 2*src[2*step])>>4;
-				src[3*step]= (sums[3] + sums[5] + 2*src[3*step])>>4;
-				src[4*step]= (sums[4] + sums[6] + 2*src[4*step])>>4;
-				src[5*step]= (sums[5] + sums[7] + 2*src[5*step])>>4;
-				src[6*step]= (sums[6] + sums[8] + 2*src[6*step])>>4;
-				src[7*step]= (sums[7] + sums[9] + 2*src[7*step])>>4;
-			}
-		}else{
-			const int middleEnergy= 5*(src[4*step] - src[3*step]) + 2*(src[2*step] - src[5*step]);
-
-			if(ABS(middleEnergy) < 8*QP)
-			{
-				const int q=(src[3*step] - src[4*step])/2;
-				const int leftEnergy=  5*(src[2*step] - src[1*step]) + 2*(src[0*step] - src[3*step]);
-				const int rightEnergy= 5*(src[6*step] - src[5*step]) + 2*(src[4*step] - src[7*step]);
-
-				int d= ABS(middleEnergy) - MIN( ABS(leftEnergy), ABS(rightEnergy) );
-				d= MAX(d, 0);
-
-				d= (5*d + 32) >> 6;
-				d*= SIGN(-middleEnergy);
-
-				if(q>0)
-				{
-					d= d<0 ? 0 : d;
-					d= d>q ? q : d;
-				}
-				else
-				{
-					d= d>0 ? 0 : d;
-					d= d<q ? q : d;
-				}
-
-				src[3*step]-= d;
-				src[4*step]+= d;
-			}
-		}
-
-		src += stride;
-	}
+        src+= step*4; // src points to begin of the 8x8 Block
+        for(y=0; y<8; y++){
+                int numEq= 0;
+
+                if(((unsigned)(src[-1*step] - src[0*step] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[ 0*step] - src[1*step] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[ 1*step] - src[2*step] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[ 2*step] - src[3*step] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[ 3*step] - src[4*step] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[ 4*step] - src[5*step] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[ 5*step] - src[6*step] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[ 6*step] - src[7*step] + dcOffset)) < dcThreshold) numEq++;
+                if(((unsigned)(src[ 7*step] - src[8*step] + dcOffset)) < dcThreshold) numEq++;
+                if(numEq > c->ppMode.flatnessThreshold){
+                        int min, max, x;
+
+                        if(src[0] > src[step]){
+                            max= src[0];
+                            min= src[step];
+                        }else{
+                            max= src[step];
+                            min= src[0];
+                        }
+                        for(x=2; x<8; x+=2){
+                                if(src[x*step] > src[(x+1)*step]){
+                                        if(src[x    *step] > max) max= src[ x   *step];
+                                        if(src[(x+1)*step] < min) min= src[(x+1)*step];
+                                }else{
+                                        if(src[(x+1)*step] > max) max= src[(x+1)*step];
+                                        if(src[ x   *step] < min) min= src[ x   *step];
+                                }
+                        }
+                        if(max-min < 2*QP){
+                                const int first= ABS(src[-1*step] - src[0]) < QP ? src[-1*step] : src[0];
+                                const int last= ABS(src[8*step] - src[7*step]) < QP ? src[8*step] : src[7*step];
+
+                                int sums[10];
+                                sums[0] = 4*first + src[0*step] + src[1*step] + src[2*step] + 4;
+                                sums[1] = sums[0] - first       + src[3*step];
+                                sums[2] = sums[1] - first       + src[4*step];
+                                sums[3] = sums[2] - first       + src[5*step];
+                                sums[4] = sums[3] - first       + src[6*step];
+                                sums[5] = sums[4] - src[0*step] + src[7*step];
+                                sums[6] = sums[5] - src[1*step] + last;
+                                sums[7] = sums[6] - src[2*step] + last;
+                                sums[8] = sums[7] - src[3*step] + last;
+                                sums[9] = sums[8] - src[4*step] + last;
+
+                                src[0*step]= (sums[0] + sums[2] + 2*src[0*step])>>4;
+                                src[1*step]= (sums[1] + sums[3] + 2*src[1*step])>>4;
+                                src[2*step]= (sums[2] + sums[4] + 2*src[2*step])>>4;
+                                src[3*step]= (sums[3] + sums[5] + 2*src[3*step])>>4;
+                                src[4*step]= (sums[4] + sums[6] + 2*src[4*step])>>4;
+                                src[5*step]= (sums[5] + sums[7] + 2*src[5*step])>>4;
+                                src[6*step]= (sums[6] + sums[8] + 2*src[6*step])>>4;
+                                src[7*step]= (sums[7] + sums[9] + 2*src[7*step])>>4;
+                        }
+                }else{
+                        const int middleEnergy= 5*(src[4*step] - src[3*step]) + 2*(src[2*step] - src[5*step]);
+
+                        if(ABS(middleEnergy) < 8*QP)
+                        {
+                                const int q=(src[3*step] - src[4*step])/2;
+                                const int leftEnergy=  5*(src[2*step] - src[1*step]) + 2*(src[0*step] - src[3*step]);
+                                const int rightEnergy= 5*(src[6*step] - src[5*step]) + 2*(src[4*step] - src[7*step]);
+
+                                int d= ABS(middleEnergy) - MIN( ABS(leftEnergy), ABS(rightEnergy) );
+                                d= MAX(d, 0);
+
+                                d= (5*d + 32) >> 6;
+                                d*= SIGN(-middleEnergy);
+
+                                if(q>0)
+                                {
+                                        d= d<0 ? 0 : d;
+                                        d= d>q ? q : d;
+                                }
+                                else
+                                {
+                                        d= d>0 ? 0 : d;
+                                        d= d<q ? q : d;
+                                }
+
+                                src[3*step]-= d;
+                                src[4*step]+= d;
+                        }
+                }
+
+                src += stride;
+        }
 /*if(step==16){
     STOP_TIMER("step16")
 }else{
@@ -668,290 +668,290 @@ static always_inline void do_a_deblock_C(uint8_t *src, int step, int stride, PPC
 // minor note: the HAVE_xyz is messed up after that line so dont use it
 
 static inline void postProcess(uint8_t src[], int srcStride, uint8_t dst[], int dstStride, int width, int height,
-	QP_STORE_T QPs[], int QPStride, int isColor, pp_mode_t *vm, pp_context_t *vc)
+        QP_STORE_T QPs[], int QPStride, int isColor, pp_mode_t *vm, pp_context_t *vc)
 {
-	PPContext *c= (PPContext *)vc;
-	PPMode *ppMode= (PPMode *)vm;
-	c->ppMode= *ppMode; //FIXME
+        PPContext *c= (PPContext *)vc;
+        PPMode *ppMode= (PPMode *)vm;
+        c->ppMode= *ppMode; //FIXME
 
-	// useing ifs here as they are faster than function pointers allthough the
-	// difference wouldnt be messureable here but its much better because
-	// someone might exchange the cpu whithout restarting mplayer ;)
+        // useing ifs here as they are faster than function pointers allthough the
+        // difference wouldnt be messureable here but its much better because
+        // someone might exchange the cpu whithout restarting mplayer ;)
 #ifdef RUNTIME_CPUDETECT
 #if defined(ARCH_X86) || defined(ARCH_X86_64)
-	// ordered per speed fasterst first
-	if(c->cpuCaps & PP_CPU_CAPS_MMX2)
-		postProcess_MMX2(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
-	else if(c->cpuCaps & PP_CPU_CAPS_3DNOW)
-		postProcess_3DNow(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
-	else if(c->cpuCaps & PP_CPU_CAPS_MMX)
-		postProcess_MMX(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
-	else
-		postProcess_C(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
+        // ordered per speed fasterst first
+        if(c->cpuCaps & PP_CPU_CAPS_MMX2)
+                postProcess_MMX2(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
+        else if(c->cpuCaps & PP_CPU_CAPS_3DNOW)
+                postProcess_3DNow(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
+        else if(c->cpuCaps & PP_CPU_CAPS_MMX)
+                postProcess_MMX(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
+        else
+                postProcess_C(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
 #else
 #ifdef ARCH_POWERPC
 #ifdef HAVE_ALTIVEC
         if(c->cpuCaps & PP_CPU_CAPS_ALTIVEC)
-		postProcess_altivec(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
+                postProcess_altivec(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
         else
 #endif
 #endif
-		postProcess_C(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
+                postProcess_C(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
 #endif
 #else //RUNTIME_CPUDETECT
 #ifdef HAVE_MMX2
-		postProcess_MMX2(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
+                postProcess_MMX2(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
 #elif defined (HAVE_3DNOW)
-		postProcess_3DNow(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
+                postProcess_3DNow(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
 #elif defined (HAVE_MMX)
-		postProcess_MMX(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
+                postProcess_MMX(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
 #elif defined (HAVE_ALTIVEC)
-		postProcess_altivec(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
+                postProcess_altivec(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
 #else
-		postProcess_C(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
+                postProcess_C(src, srcStride, dst, dstStride, width, height, QPs, QPStride, isColor, c);
 #endif
 #endif //!RUNTIME_CPUDETECT
 }
 
 //static void postProcess(uint8_t src[], int srcStride, uint8_t dst[], int dstStride, int width, int height,
-//	QP_STORE_T QPs[], int QPStride, int isColor, struct PPMode *ppMode);
+//        QP_STORE_T QPs[], int QPStride, int isColor, struct PPMode *ppMode);
 
 /* -pp Command line Help
 */
 char *pp_help=
 "Available postprocessing filters:\n"
-"Filters			Options\n"
-"short	long name	short	long option	Description\n"
-"*	*		a	autoq		CPU power dependent enabler\n"
-"			c	chrom		chrominance filtering enabled\n"
-"			y	nochrom		chrominance filtering disabled\n"
-"			n	noluma		luma filtering disabled\n"
-"hb	hdeblock	(2 threshold)		horizontal deblocking filter\n"
-"	1. difference factor: default=32, higher -> more deblocking\n"
-"	2. flatness threshold: default=39, lower -> more deblocking\n"
-"			the h & v deblocking filters share these\n"
-"			so you can't set different thresholds for h / v\n"
-"vb	vdeblock	(2 threshold)		vertical deblocking filter\n"
-"ha	hadeblock	(2 threshold)		horizontal deblocking filter\n"
-"va	vadeblock	(2 threshold)		vertical deblocking filter\n"
-"h1	x1hdeblock				experimental h deblock filter 1\n"
-"v1	x1vdeblock				experimental v deblock filter 1\n"
-"dr	dering					deringing filter\n"
-"al	autolevels				automatic brightness / contrast\n"
-"			f	fullyrange	stretch luminance to (0..255)\n"
-"lb	linblenddeint				linear blend deinterlacer\n"
-"li	linipoldeint				linear interpolating deinterlace\n"
-"ci	cubicipoldeint				cubic interpolating deinterlacer\n"
-"md	mediandeint				median deinterlacer\n"
-"fd	ffmpegdeint				ffmpeg deinterlacer\n"
-"l5	lowpass5				FIR lowpass deinterlacer\n"
-"de	default					hb:a,vb:a,dr:a\n"
-"fa	fast					h1:a,v1:a,dr:a\n"
-"ac						ha:a:128:7,va:a,dr:a\n"
-"tn	tmpnoise	(3 threshold)		temporal noise reducer\n"
-"			1. <= 2. <= 3.		larger -> stronger filtering\n"
-"fq	forceQuant	<quantizer>		force quantizer\n"
+"Filters                        Options\n"
+"short  long name       short   long option     Description\n"
+"*      *               a       autoq           CPU power dependent enabler\n"
+"                       c       chrom           chrominance filtering enabled\n"
+"                       y       nochrom         chrominance filtering disabled\n"
+"                       n       noluma          luma filtering disabled\n"
+"hb     hdeblock        (2 threshold)           horizontal deblocking filter\n"
+"       1. difference factor: default=32, higher -> more deblocking\n"
+"       2. flatness threshold: default=39, lower -> more deblocking\n"
+"                       the h & v deblocking filters share these\n"
+"                       so you can't set different thresholds for h / v\n"
+"vb     vdeblock        (2 threshold)           vertical deblocking filter\n"
+"ha     hadeblock       (2 threshold)           horizontal deblocking filter\n"
+"va     vadeblock       (2 threshold)           vertical deblocking filter\n"
+"h1     x1hdeblock                              experimental h deblock filter 1\n"
+"v1     x1vdeblock                              experimental v deblock filter 1\n"
+"dr     dering                                  deringing filter\n"
+"al     autolevels                              automatic brightness / contrast\n"
+"                       f        fullyrange     stretch luminance to (0..255)\n"
+"lb     linblenddeint                           linear blend deinterlacer\n"
+"li     linipoldeint                            linear interpolating deinterlace\n"
+"ci     cubicipoldeint                          cubic interpolating deinterlacer\n"
+"md     mediandeint                             median deinterlacer\n"
+"fd     ffmpegdeint                             ffmpeg deinterlacer\n"
+"l5     lowpass5                                FIR lowpass deinterlacer\n"
+"de     default                                 hb:a,vb:a,dr:a\n"
+"fa     fast                                    h1:a,v1:a,dr:a\n"
+"ac                                             ha:a:128:7,va:a,dr:a\n"
+"tn     tmpnoise        (3 threshold)           temporal noise reducer\n"
+"                     1. <= 2. <= 3.            larger -> stronger filtering\n"
+"fq     forceQuant      <quantizer>             force quantizer\n"
 "Usage:\n"
 "<filterName>[:<option>[:<option>...]][[,|/][-]<filterName>[:<option>...]]...\n"
 "long form example:\n"
-"vdeblock:autoq/hdeblock:autoq/linblenddeint	default,-vdeblock\n"
+"vdeblock:autoq/hdeblock:autoq/linblenddeint    default,-vdeblock\n"
 "short form example:\n"
-"vb:a/hb:a/lb					de,-vb\n"
+"vb:a/hb:a/lb                                   de,-vb\n"
 "more examples:\n"
 "tn:64:128:256\n"
 ;
 
 pp_mode_t *pp_get_mode_by_name_and_quality(char *name, int quality)
 {
-	char temp[GET_MODE_BUFFER_SIZE];
-	char *p= temp;
-	char *filterDelimiters= ",/";
-	char *optionDelimiters= ":";
-	struct PPMode *ppMode;
-	char *filterToken;
-
-	ppMode= memalign(8, sizeof(PPMode));
-
-	ppMode->lumMode= 0;
-	ppMode->chromMode= 0;
-	ppMode->maxTmpNoise[0]= 700;
-	ppMode->maxTmpNoise[1]= 1500;
-	ppMode->maxTmpNoise[2]= 3000;
-	ppMode->maxAllowedY= 234;
-	ppMode->minAllowedY= 16;
-	ppMode->baseDcDiff= 256/8;
-	ppMode->flatnessThreshold= 56-16-1;
-	ppMode->maxClippedThreshold= 0.01;
-	ppMode->error=0;
-
-	strncpy(temp, name, GET_MODE_BUFFER_SIZE);
-
-	if(verbose>1) printf("pp: %s\n", name);
-
-	for(;;){
-		char *filterName;
-		int q= 1000000; //PP_QUALITY_MAX;
-		int chrom=-1;
-		int luma=-1;
-		char *option;
-		char *options[OPTIONS_ARRAY_SIZE];
-		int i;
-		int filterNameOk=0;
-		int numOfUnknownOptions=0;
-		int enable=1; //does the user want us to enabled or disabled the filter
-
-		filterToken= strtok(p, filterDelimiters);
-		if(filterToken == NULL) break;
-		p+= strlen(filterToken) + 1; // p points to next filterToken
-		filterName= strtok(filterToken, optionDelimiters);
-		if(verbose>1) printf("pp: %s::%s\n", filterToken, filterName);
-
-		if(*filterName == '-')
-		{
-			enable=0;
-			filterName++;
-		}
-
-		for(;;){ //for all options
-			option= strtok(NULL, optionDelimiters);
-			if(option == NULL) break;
-
-			if(verbose>1) printf("pp: option: %s\n", option);
-			if(!strcmp("autoq", option) || !strcmp("a", option)) q= quality;
-			else if(!strcmp("nochrom", option) || !strcmp("y", option)) chrom=0;
-			else if(!strcmp("chrom", option) || !strcmp("c", option)) chrom=1;
-			else if(!strcmp("noluma", option) || !strcmp("n", option)) luma=0;
-			else
-			{
-				options[numOfUnknownOptions] = option;
-				numOfUnknownOptions++;
-			}
-			if(numOfUnknownOptions >= OPTIONS_ARRAY_SIZE-1) break;
-		}
-		options[numOfUnknownOptions] = NULL;
-
-		/* replace stuff from the replace Table */
-		for(i=0; replaceTable[2*i]!=NULL; i++)
-		{
-			if(!strcmp(replaceTable[2*i], filterName))
-			{
-				int newlen= strlen(replaceTable[2*i + 1]);
-				int plen;
-				int spaceLeft;
-
-				if(p==NULL) p= temp, *p=0; 	//last filter
-				else p--, *p=',';		//not last filter
-
-				plen= strlen(p);
-				spaceLeft= p - temp + plen;
-				if(spaceLeft + newlen  >= GET_MODE_BUFFER_SIZE)
-				{
-					ppMode->error++;
-					break;
-				}
-				memmove(p + newlen, p, plen+1);
-				memcpy(p, replaceTable[2*i + 1], newlen);
-				filterNameOk=1;
-			}
-		}
-
-		for(i=0; filters[i].shortName!=NULL; i++)
-		{
-//			printf("Compareing %s, %s, %s\n", filters[i].shortName,filters[i].longName, filterName);
-			if(   !strcmp(filters[i].longName, filterName)
-			   || !strcmp(filters[i].shortName, filterName))
-			{
-				ppMode->lumMode &= ~filters[i].mask;
-				ppMode->chromMode &= ~filters[i].mask;
-
-				filterNameOk=1;
-				if(!enable) break; // user wants to disable it
-
-				if(q >= filters[i].minLumQuality && luma)
-					ppMode->lumMode|= filters[i].mask;
-				if(chrom==1 || (chrom==-1 && filters[i].chromDefault))
-					if(q >= filters[i].minChromQuality)
-						ppMode->chromMode|= filters[i].mask;
-
-				if(filters[i].mask == LEVEL_FIX)
-				{
-					int o;
-					ppMode->minAllowedY= 16;
-					ppMode->maxAllowedY= 234;
-					for(o=0; options[o]!=NULL; o++)
-					{
-						if(  !strcmp(options[o],"fullyrange")
-						   ||!strcmp(options[o],"f"))
-						{
-							ppMode->minAllowedY= 0;
-							ppMode->maxAllowedY= 255;
-							numOfUnknownOptions--;
-						}
-					}
-				}
-				else if(filters[i].mask == TEMP_NOISE_FILTER)
-				{
-					int o;
-					int numOfNoises=0;
-
-					for(o=0; options[o]!=NULL; o++)
-					{
-						char *tail;
-						ppMode->maxTmpNoise[numOfNoises]=
-							strtol(options[o], &tail, 0);
-						if(tail!=options[o])
-						{
-							numOfNoises++;
-							numOfUnknownOptions--;
-							if(numOfNoises >= 3) break;
-						}
-					}
-				}
-				else if(filters[i].mask == V_DEBLOCK   || filters[i].mask == H_DEBLOCK
-				     || filters[i].mask == V_A_DEBLOCK || filters[i].mask == H_A_DEBLOCK)
-				{
-					int o;
-
-					for(o=0; options[o]!=NULL && o<2; o++)
-					{
-						char *tail;
-						int val= strtol(options[o], &tail, 0);
-						if(tail==options[o]) break;
-
-						numOfUnknownOptions--;
-						if(o==0) ppMode->baseDcDiff= val;
-						else ppMode->flatnessThreshold= val;
-					}
-				}
-				else if(filters[i].mask == FORCE_QUANT)
-				{
-					int o;
-					ppMode->forcedQuant= 15;
-
-					for(o=0; options[o]!=NULL && o<1; o++)
-					{
-						char *tail;
-						int val= strtol(options[o], &tail, 0);
-						if(tail==options[o]) break;
-
-						numOfUnknownOptions--;
-						ppMode->forcedQuant= val;
-					}
-				}
-			}
-		}
-		if(!filterNameOk) ppMode->error++;
-		ppMode->error += numOfUnknownOptions;
-	}
-
-	if(verbose>1) printf("pp: lumMode=%X, chromMode=%X\n", ppMode->lumMode, ppMode->chromMode);
-	if(ppMode->error)
-	{
-		fprintf(stderr, "%d errors in postprocess string \"%s\"\n", ppMode->error, name);
-		free(ppMode);
-		return NULL;
-	}
-	return ppMode;
+        char temp[GET_MODE_BUFFER_SIZE];
+        char *p= temp;
+        char *filterDelimiters= ",/";
+        char *optionDelimiters= ":";
+        struct PPMode *ppMode;
+        char *filterToken;
+
+        ppMode= memalign(8, sizeof(PPMode));
+
+        ppMode->lumMode= 0;
+        ppMode->chromMode= 0;
+        ppMode->maxTmpNoise[0]= 700;
+        ppMode->maxTmpNoise[1]= 1500;
+        ppMode->maxTmpNoise[2]= 3000;
+        ppMode->maxAllowedY= 234;
+        ppMode->minAllowedY= 16;
+        ppMode->baseDcDiff= 256/8;
+        ppMode->flatnessThreshold= 56-16-1;
+        ppMode->maxClippedThreshold= 0.01;
+        ppMode->error=0;
+
+        strncpy(temp, name, GET_MODE_BUFFER_SIZE);
+
+        if(verbose>1) printf("pp: %s\n", name);
+
+        for(;;){
+                char *filterName;
+                int q= 1000000; //PP_QUALITY_MAX;
+                int chrom=-1;
+                int luma=-1;
+                char *option;
+                char *options[OPTIONS_ARRAY_SIZE];
+                int i;
+                int filterNameOk=0;
+                int numOfUnknownOptions=0;
+                int enable=1; //does the user want us to enabled or disabled the filter
+
+                filterToken= strtok(p, filterDelimiters);
+                if(filterToken == NULL) break;
+                p+= strlen(filterToken) + 1; // p points to next filterToken
+                filterName= strtok(filterToken, optionDelimiters);
+                if(verbose>1) printf("pp: %s::%s\n", filterToken, filterName);
+
+                if(*filterName == '-')
+                {
+                        enable=0;
+                        filterName++;
+                }
+
+                for(;;){ //for all options
+                        option= strtok(NULL, optionDelimiters);
+                        if(option == NULL) break;
+
+                        if(verbose>1) printf("pp: option: %s\n", option);
+                        if(!strcmp("autoq", option) || !strcmp("a", option)) q= quality;
+                        else if(!strcmp("nochrom", option) || !strcmp("y", option)) chrom=0;
+                        else if(!strcmp("chrom", option) || !strcmp("c", option)) chrom=1;
+                        else if(!strcmp("noluma", option) || !strcmp("n", option)) luma=0;
+                        else
+                        {
+                                options[numOfUnknownOptions] = option;
+                                numOfUnknownOptions++;
+                        }
+                        if(numOfUnknownOptions >= OPTIONS_ARRAY_SIZE-1) break;
+                }
+                options[numOfUnknownOptions] = NULL;
+
+                /* replace stuff from the replace Table */
+                for(i=0; replaceTable[2*i]!=NULL; i++)
+                {
+                        if(!strcmp(replaceTable[2*i], filterName))
+                        {
+                                int newlen= strlen(replaceTable[2*i + 1]);
+                                int plen;
+                                int spaceLeft;
+
+                                if(p==NULL) p= temp, *p=0;      //last filter
+                                else p--, *p=',';               //not last filter
+
+                                plen= strlen(p);
+                                spaceLeft= p - temp + plen;
+                                if(spaceLeft + newlen  >= GET_MODE_BUFFER_SIZE)
+                                {
+                                        ppMode->error++;
+                                        break;
+                                }
+                                memmove(p + newlen, p, plen+1);
+                                memcpy(p, replaceTable[2*i + 1], newlen);
+                                filterNameOk=1;
+                        }
+                }
+
+                for(i=0; filters[i].shortName!=NULL; i++)
+                {
+//                        printf("Compareing %s, %s, %s\n", filters[i].shortName,filters[i].longName, filterName);
+                        if(   !strcmp(filters[i].longName, filterName)
+                           || !strcmp(filters[i].shortName, filterName))
+                        {
+                                ppMode->lumMode &= ~filters[i].mask;
+                                ppMode->chromMode &= ~filters[i].mask;
+
+                                filterNameOk=1;
+                                if(!enable) break; // user wants to disable it
+
+                                if(q >= filters[i].minLumQuality && luma)
+                                        ppMode->lumMode|= filters[i].mask;
+                                if(chrom==1 || (chrom==-1 && filters[i].chromDefault))
+                                        if(q >= filters[i].minChromQuality)
+                                                ppMode->chromMode|= filters[i].mask;
+
+                                if(filters[i].mask == LEVEL_FIX)
+                                {
+                                        int o;
+                                        ppMode->minAllowedY= 16;
+                                        ppMode->maxAllowedY= 234;
+                                        for(o=0; options[o]!=NULL; o++)
+                                        {
+                                                if(  !strcmp(options[o],"fullyrange")
+                                                   ||!strcmp(options[o],"f"))
+                                                {
+                                                        ppMode->minAllowedY= 0;
+                                                        ppMode->maxAllowedY= 255;
+                                                        numOfUnknownOptions--;
+                                                }
+                                        }
+                                }
+                                else if(filters[i].mask == TEMP_NOISE_FILTER)
+                                {
+                                        int o;
+                                        int numOfNoises=0;
+
+                                        for(o=0; options[o]!=NULL; o++)
+                                        {
+                                                char *tail;
+                                                ppMode->maxTmpNoise[numOfNoises]=
+                                                        strtol(options[o], &tail, 0);
+                                                if(tail!=options[o])
+                                                {
+                                                        numOfNoises++;
+                                                        numOfUnknownOptions--;
+                                                        if(numOfNoises >= 3) break;
+                                                }
+                                        }
+                                }
+                                else if(filters[i].mask == V_DEBLOCK   || filters[i].mask == H_DEBLOCK
+                                     || filters[i].mask == V_A_DEBLOCK || filters[i].mask == H_A_DEBLOCK)
+                                {
+                                        int o;
+
+                                        for(o=0; options[o]!=NULL && o<2; o++)
+                                        {
+                                                char *tail;
+                                                int val= strtol(options[o], &tail, 0);
+                                                if(tail==options[o]) break;
+
+                                                numOfUnknownOptions--;
+                                                if(o==0) ppMode->baseDcDiff= val;
+                                                else ppMode->flatnessThreshold= val;
+                                        }
+                                }
+                                else if(filters[i].mask == FORCE_QUANT)
+                                {
+                                        int o;
+                                        ppMode->forcedQuant= 15;
+
+                                        for(o=0; options[o]!=NULL && o<1; o++)
+                                        {
+                                                char *tail;
+                                                int val= strtol(options[o], &tail, 0);
+                                                if(tail==options[o]) break;
+
+                                                numOfUnknownOptions--;
+                                                ppMode->forcedQuant= val;
+                                        }
+                                }
+                        }
+                }
+                if(!filterNameOk) ppMode->error++;
+                ppMode->error += numOfUnknownOptions;
+        }
+
+        if(verbose>1) printf("pp: lumMode=%X, chromMode=%X\n", ppMode->lumMode, ppMode->chromMode);
+        if(ppMode->error)
+        {
+                fprintf(stderr, "%d errors in postprocess string \"%s\"\n", ppMode->error, name);
+                free(ppMode);
+                return NULL;
+        }
+        return ppMode;
 }
 
 void pp_free_mode(pp_mode_t *mode){
@@ -959,199 +959,199 @@ void pp_free_mode(pp_mode_t *mode){
 }
 
 static void reallocAlign(void **p, int alignment, int size){
-	if(*p) free(*p);
-	*p= memalign(alignment, size);
-	memset(*p, 0, size);
+        if(*p) free(*p);
+        *p= memalign(alignment, size);
+        memset(*p, 0, size);
 }
 
 static void reallocBuffers(PPContext *c, int width, int height, int stride, int qpStride){
-	int mbWidth = (width+15)>>4;
-	int mbHeight= (height+15)>>4;
-	int i;
-
-	c->stride= stride;
-	c->qpStride= qpStride;
-
-	reallocAlign((void **)&c->tempDst, 8, stride*24);
-	reallocAlign((void **)&c->tempSrc, 8, stride*24);
-	reallocAlign((void **)&c->tempBlocks, 8, 2*16*8);
-	reallocAlign((void **)&c->yHistogram, 8, 256*sizeof(uint64_t));
-	for(i=0; i<256; i++)
-		c->yHistogram[i]= width*height/64*15/256;
-
-	for(i=0; i<3; i++)
-	{
-		//Note:the +17*1024 is just there so i dont have to worry about r/w over te end
-		reallocAlign((void **)&c->tempBlured[i], 8, stride*mbHeight*16 + 17*1024);
-		reallocAlign((void **)&c->tempBluredPast[i], 8, 256*((height+7)&(~7))/2 + 17*1024);//FIXME size
-	}
-
-	reallocAlign((void **)&c->deintTemp, 8, 2*width+32);
-	reallocAlign((void **)&c->nonBQPTable, 8, qpStride*mbHeight*sizeof(QP_STORE_T));
-	reallocAlign((void **)&c->stdQPTable, 8, qpStride*mbHeight*sizeof(QP_STORE_T));
-	reallocAlign((void **)&c->forcedQPTable, 8, mbWidth*sizeof(QP_STORE_T));
+        int mbWidth = (width+15)>>4;
+        int mbHeight= (height+15)>>4;
+        int i;
+
+        c->stride= stride;
+        c->qpStride= qpStride;
+
+        reallocAlign((void **)&c->tempDst, 8, stride*24);
+        reallocAlign((void **)&c->tempSrc, 8, stride*24);
+        reallocAlign((void **)&c->tempBlocks, 8, 2*16*8);
+        reallocAlign((void **)&c->yHistogram, 8, 256*sizeof(uint64_t));
+        for(i=0; i<256; i++)
+                c->yHistogram[i]= width*height/64*15/256;
+
+        for(i=0; i<3; i++)
+        {
+                //Note:the +17*1024 is just there so i dont have to worry about r/w over te end
+                reallocAlign((void **)&c->tempBlured[i], 8, stride*mbHeight*16 + 17*1024);
+                reallocAlign((void **)&c->tempBluredPast[i], 8, 256*((height+7)&(~7))/2 + 17*1024);//FIXME size
+        }
+
+        reallocAlign((void **)&c->deintTemp, 8, 2*width+32);
+        reallocAlign((void **)&c->nonBQPTable, 8, qpStride*mbHeight*sizeof(QP_STORE_T));
+        reallocAlign((void **)&c->stdQPTable, 8, qpStride*mbHeight*sizeof(QP_STORE_T));
+        reallocAlign((void **)&c->forcedQPTable, 8, mbWidth*sizeof(QP_STORE_T));
 }
 
 static void global_init(void){
-	int i;
-	memset(clip_table, 0, 256);
-	for(i=256; i<512; i++)
-		clip_table[i]= i;
-	memset(clip_table+512, 0, 256);
+        int i;
+        memset(clip_table, 0, 256);
+        for(i=256; i<512; i++)
+                clip_table[i]= i;
+        memset(clip_table+512, 0, 256);
 }
 
 pp_context_t *pp_get_context(int width, int height, int cpuCaps){
-	PPContext *c= memalign(32, sizeof(PPContext));
-	int stride= (width+15)&(~15); //assumed / will realloc if needed
-	int qpStride= (width+15)/16 + 2; //assumed / will realloc if needed
+        PPContext *c= memalign(32, sizeof(PPContext));
+        int stride= (width+15)&(~15);    //assumed / will realloc if needed
+        int qpStride= (width+15)/16 + 2; //assumed / will realloc if needed
 
-	global_init();
+        global_init();
 
-	memset(c, 0, sizeof(PPContext));
-	c->cpuCaps= cpuCaps;
-	if(cpuCaps&PP_FORMAT){
-		c->hChromaSubSample= cpuCaps&0x3;
-		c->vChromaSubSample= (cpuCaps>>4)&0x3;
-	}else{
-		c->hChromaSubSample= 1;
-		c->vChromaSubSample= 1;
-	}
+        memset(c, 0, sizeof(PPContext));
+        c->cpuCaps= cpuCaps;
+        if(cpuCaps&PP_FORMAT){
+                c->hChromaSubSample= cpuCaps&0x3;
+                c->vChromaSubSample= (cpuCaps>>4)&0x3;
+        }else{
+                c->hChromaSubSample= 1;
+                c->vChromaSubSample= 1;
+        }
 
-	reallocBuffers(c, width, height, stride, qpStride);
+        reallocBuffers(c, width, height, stride, qpStride);
 
-	c->frameNum=-1;
+        c->frameNum=-1;
 
-	return c;
+        return c;
 }
 
 void pp_free_context(void *vc){
-	PPContext *c = (PPContext*)vc;
-	int i;
+        PPContext *c = (PPContext*)vc;
+        int i;
 
-	for(i=0; i<3; i++) free(c->tempBlured[i]);
-	for(i=0; i<3; i++) free(c->tempBluredPast[i]);
+        for(i=0; i<3; i++) free(c->tempBlured[i]);
+        for(i=0; i<3; i++) free(c->tempBluredPast[i]);
 
-	free(c->tempBlocks);
-	free(c->yHistogram);
-	free(c->tempDst);
-	free(c->tempSrc);
-	free(c->deintTemp);
-	free(c->stdQPTable);
-	free(c->nonBQPTable);
-	free(c->forcedQPTable);
+        free(c->tempBlocks);
+        free(c->yHistogram);
+        free(c->tempDst);
+        free(c->tempSrc);
+        free(c->deintTemp);
+        free(c->stdQPTable);
+        free(c->nonBQPTable);
+        free(c->forcedQPTable);
 
-	memset(c, 0, sizeof(PPContext));
+        memset(c, 0, sizeof(PPContext));
 
-	free(c);
+        free(c);
 }
 
 void  pp_postprocess(uint8_t * src[3], int srcStride[3],
                  uint8_t * dst[3], int dstStride[3],
                  int width, int height,
                  QP_STORE_T *QP_store,  int QPStride,
-		 pp_mode_t *vm,  void *vc, int pict_type)
+                 pp_mode_t *vm,  void *vc, int pict_type)
 {
-	int mbWidth = (width+15)>>4;
-	int mbHeight= (height+15)>>4;
-	PPMode *mode = (PPMode*)vm;
-	PPContext *c = (PPContext*)vc;
-	int minStride= MAX(ABS(srcStride[0]), ABS(dstStride[0]));
-	int absQPStride = ABS(QPStride);
-
-	// c->stride and c->QPStride are always positive
-	if(c->stride < minStride || c->qpStride < absQPStride)
-		reallocBuffers(c, width, height,
-				MAX(minStride, c->stride),
-				MAX(c->qpStride, absQPStride));
-
-	if(QP_store==NULL || (mode->lumMode & FORCE_QUANT))
-	{
-		int i;
-		QP_store= c->forcedQPTable;
-		absQPStride = QPStride = 0;
-		if(mode->lumMode & FORCE_QUANT)
-			for(i=0; i<mbWidth; i++) QP_store[i]= mode->forcedQuant;
-		else
-			for(i=0; i<mbWidth; i++) QP_store[i]= 1;
-	}
+        int mbWidth = (width+15)>>4;
+        int mbHeight= (height+15)>>4;
+        PPMode *mode = (PPMode*)vm;
+        PPContext *c = (PPContext*)vc;
+        int minStride= MAX(ABS(srcStride[0]), ABS(dstStride[0]));
+        int absQPStride = ABS(QPStride);
+
+        // c->stride and c->QPStride are always positive
+        if(c->stride < minStride || c->qpStride < absQPStride)
+                reallocBuffers(c, width, height,
+                                MAX(minStride, c->stride),
+                                MAX(c->qpStride, absQPStride));
+
+        if(QP_store==NULL || (mode->lumMode & FORCE_QUANT))
+        {
+                int i;
+                QP_store= c->forcedQPTable;
+                absQPStride = QPStride = 0;
+                if(mode->lumMode & FORCE_QUANT)
+                        for(i=0; i<mbWidth; i++) QP_store[i]= mode->forcedQuant;
+                else
+                        for(i=0; i<mbWidth; i++) QP_store[i]= 1;
+        }
 //printf("pict_type:%d\n", pict_type);
 
-	if(pict_type & PP_PICT_TYPE_QP2){
-		int i;
-		const int count= mbHeight * absQPStride;
-		for(i=0; i<(count>>2); i++){
-			((uint32_t*)c->stdQPTable)[i] = (((uint32_t*)QP_store)[i]>>1) & 0x7F7F7F7F;
-		}
-		for(i<<=2; i<count; i++){
-			c->stdQPTable[i] = QP_store[i]>>1;
-		}
+        if(pict_type & PP_PICT_TYPE_QP2){
+                int i;
+                const int count= mbHeight * absQPStride;
+                for(i=0; i<(count>>2); i++){
+                        ((uint32_t*)c->stdQPTable)[i] = (((uint32_t*)QP_store)[i]>>1) & 0x7F7F7F7F;
+                }
+                for(i<<=2; i<count; i++){
+                        c->stdQPTable[i] = QP_store[i]>>1;
+                }
                 QP_store= c->stdQPTable;
-		QPStride= absQPStride;
-	}
+                QPStride= absQPStride;
+        }
 
 if(0){
 int x,y;
 for(y=0; y<mbHeight; y++){
-	for(x=0; x<mbWidth; x++){
-		printf("%2d ", QP_store[x + y*QPStride]);
-	}
-	printf("\n");
+        for(x=0; x<mbWidth; x++){
+                printf("%2d ", QP_store[x + y*QPStride]);
+        }
+        printf("\n");
 }
-	printf("\n");
+        printf("\n");
 }
 
-	if((pict_type&7)!=3)
-	{
-		if (QPStride >= 0) {
-			int i;
-			const int count= mbHeight * QPStride;
-			for(i=0; i<(count>>2); i++){
-				((uint32_t*)c->nonBQPTable)[i] = ((uint32_t*)QP_store)[i] & 0x3F3F3F3F;
-			}
-			for(i<<=2; i<count; i++){
-				c->nonBQPTable[i] = QP_store[i] & 0x3F;
-			}
-		} else {
-			int i,j;
-			for(i=0; i<mbHeight; i++) {
-		    		for(j=0; j<absQPStride; j++) {
-					c->nonBQPTable[i*absQPStride+j] = QP_store[i*QPStride+j] & 0x3F;
-				}
-			}
-		}
-	}
-
-	if(verbose>2)
-	{
-		printf("using npp filters 0x%X/0x%X\n", mode->lumMode, mode->chromMode);
-	}
-
-	postProcess(src[0], srcStride[0], dst[0], dstStride[0],
-		width, height, QP_store, QPStride, 0, mode, c);
-
-	width  = (width )>>c->hChromaSubSample;
-	height = (height)>>c->vChromaSubSample;
-
-	if(mode->chromMode)
-	{
-		postProcess(src[1], srcStride[1], dst[1], dstStride[1],
-			width, height, QP_store, QPStride, 1, mode, c);
-		postProcess(src[2], srcStride[2], dst[2], dstStride[2],
-			width, height, QP_store, QPStride, 2, mode, c);
-	}
-	else if(srcStride[1] == dstStride[1] && srcStride[2] == dstStride[2])
-	{
-		linecpy(dst[1], src[1], height, srcStride[1]);
-		linecpy(dst[2], src[2], height, srcStride[2]);
-	}
-	else
-	{
-		int y;
-		for(y=0; y<height; y++)
-		{
-			memcpy(&(dst[1][y*dstStride[1]]), &(src[1][y*srcStride[1]]), width);
-			memcpy(&(dst[2][y*dstStride[2]]), &(src[2][y*srcStride[2]]), width);
-		}
-	}
+        if((pict_type&7)!=3)
+        {
+                if (QPStride >= 0) {
+                        int i;
+                        const int count= mbHeight * QPStride;
+                        for(i=0; i<(count>>2); i++){
+                                ((uint32_t*)c->nonBQPTable)[i] = ((uint32_t*)QP_store)[i] & 0x3F3F3F3F;
+                        }
+                        for(i<<=2; i<count; i++){
+                                c->nonBQPTable[i] = QP_store[i] & 0x3F;
+                        }
+                } else {
+                        int i,j;
+                        for(i=0; i<mbHeight; i++) {
+                                    for(j=0; j<absQPStride; j++) {
+                                        c->nonBQPTable[i*absQPStride+j] = QP_store[i*QPStride+j] & 0x3F;
+                                }
+                        }
+                }
+        }
+
+        if(verbose>2)
+        {
+                printf("using npp filters 0x%X/0x%X\n", mode->lumMode, mode->chromMode);
+        }
+
+        postProcess(src[0], srcStride[0], dst[0], dstStride[0],
+                width, height, QP_store, QPStride, 0, mode, c);
+
+        width  = (width )>>c->hChromaSubSample;
+        height = (height)>>c->vChromaSubSample;
+
+        if(mode->chromMode)
+        {
+                postProcess(src[1], srcStride[1], dst[1], dstStride[1],
+                        width, height, QP_store, QPStride, 1, mode, c);
+                postProcess(src[2], srcStride[2], dst[2], dstStride[2],
+                        width, height, QP_store, QPStride, 2, mode, c);
+        }
+        else if(srcStride[1] == dstStride[1] && srcStride[2] == dstStride[2])
+        {
+                linecpy(dst[1], src[1], height, srcStride[1]);
+                linecpy(dst[2], src[2], height, srcStride[2]);
+        }
+        else
+        {
+                int y;
+                for(y=0; y<height; y++)
+                {
+                        memcpy(&(dst[1][y*dstStride[1]]), &(src[1][y*srcStride[1]]), width);
+                        memcpy(&(dst[2][y*dstStride[2]]), &(src[2][y*srcStride[2]]), width);
+                }
+        }
 }
 
diff --git a/libavcodec/libpostproc/postprocess.h b/libavcodec/libpostproc/postprocess.h
index 539ea1e181..dc1487622b 100644
--- a/libavcodec/libpostproc/postprocess.h
+++ b/libavcodec/libpostproc/postprocess.h
@@ -42,7 +42,7 @@ void  pp_postprocess(uint8_t * src[3], int srcStride[3],
                  uint8_t * dst[3], int dstStride[3],
                  int horizontalSize, int verticalSize,
                  QP_STORE_T *QP_store,  int QP_stride,
-		 pp_mode_t *mode, pp_context_t *ppContext, int pict_type);
+                 pp_mode_t *mode, pp_context_t *ppContext, int pict_type);
 
 
 /**
diff --git a/libavcodec/libpostproc/postprocess_altivec_template.c b/libavcodec/libpostproc/postprocess_altivec_template.c
index 5c431c89d9..828f9e0462 100644
--- a/libavcodec/libpostproc/postprocess_altivec_template.c
+++ b/libavcodec/libpostproc/postprocess_altivec_template.c
@@ -26,35 +26,35 @@
 #endif
 
 #define ALTIVEC_TRANSPOSE_8x8_SHORT(src_a,src_b,src_c,src_d,src_e,src_f,src_g,src_h) \
-  do {									\
-    __typeof__(src_a) tempA1, tempB1, tempC1, tempD1;			\
-    __typeof__(src_a) tempE1, tempF1, tempG1, tempH1;			\
-    __typeof__(src_a) tempA2, tempB2, tempC2, tempD2;			\
-    __typeof__(src_a) tempE2, tempF2, tempG2, tempH2;			\
-    tempA1 = vec_mergeh (src_a, src_e);					\
-    tempB1 = vec_mergel (src_a, src_e);					\
-    tempC1 = vec_mergeh (src_b, src_f);					\
-    tempD1 = vec_mergel (src_b, src_f);					\
-    tempE1 = vec_mergeh (src_c, src_g);					\
-    tempF1 = vec_mergel (src_c, src_g);					\
-    tempG1 = vec_mergeh (src_d, src_h);					\
-    tempH1 = vec_mergel (src_d, src_h);					\
-    tempA2 = vec_mergeh (tempA1, tempE1);				\
-    tempB2 = vec_mergel (tempA1, tempE1);				\
-    tempC2 = vec_mergeh (tempB1, tempF1);				\
-    tempD2 = vec_mergel (tempB1, tempF1);				\
-    tempE2 = vec_mergeh (tempC1, tempG1);				\
-    tempF2 = vec_mergel (tempC1, tempG1);				\
-    tempG2 = vec_mergeh (tempD1, tempH1);				\
-    tempH2 = vec_mergel (tempD1, tempH1);				\
-    src_a = vec_mergeh (tempA2, tempE2);				\
-    src_b = vec_mergel (tempA2, tempE2);				\
-    src_c = vec_mergeh (tempB2, tempF2);				\
-    src_d = vec_mergel (tempB2, tempF2);				\
-    src_e = vec_mergeh (tempC2, tempG2);				\
-    src_f = vec_mergel (tempC2, tempG2);				\
-    src_g = vec_mergeh (tempD2, tempH2);				\
-    src_h = vec_mergel (tempD2, tempH2);				\
+  do {                                                                  \
+    __typeof__(src_a) tempA1, tempB1, tempC1, tempD1;                   \
+    __typeof__(src_a) tempE1, tempF1, tempG1, tempH1;                   \
+    __typeof__(src_a) tempA2, tempB2, tempC2, tempD2;                   \
+    __typeof__(src_a) tempE2, tempF2, tempG2, tempH2;                   \
+    tempA1 = vec_mergeh (src_a, src_e);                                 \
+    tempB1 = vec_mergel (src_a, src_e);                                 \
+    tempC1 = vec_mergeh (src_b, src_f);                                 \
+    tempD1 = vec_mergel (src_b, src_f);                                 \
+    tempE1 = vec_mergeh (src_c, src_g);                                 \
+    tempF1 = vec_mergel (src_c, src_g);                                 \
+    tempG1 = vec_mergeh (src_d, src_h);                                 \
+    tempH1 = vec_mergel (src_d, src_h);                                 \
+    tempA2 = vec_mergeh (tempA1, tempE1);                               \
+    tempB2 = vec_mergel (tempA1, tempE1);                               \
+    tempC2 = vec_mergeh (tempB1, tempF1);                               \
+    tempD2 = vec_mergel (tempB1, tempF1);                               \
+    tempE2 = vec_mergeh (tempC1, tempG1);                               \
+    tempF2 = vec_mergel (tempC1, tempG1);                               \
+    tempG2 = vec_mergeh (tempD1, tempH1);                               \
+    tempH2 = vec_mergel (tempD1, tempH1);                               \
+    src_a = vec_mergeh (tempA2, tempE2);                                \
+    src_b = vec_mergel (tempA2, tempE2);                                \
+    src_c = vec_mergeh (tempB2, tempF2);                                \
+    src_d = vec_mergel (tempB2, tempF2);                                \
+    src_e = vec_mergeh (tempC2, tempG2);                                \
+    src_f = vec_mergel (tempC2, tempG2);                                \
+    src_g = vec_mergeh (tempD2, tempH2);                                \
+    src_h = vec_mergel (tempD2, tempH2);                                \
   } while (0)
 
 
@@ -94,25 +94,25 @@ static inline int vertClassify_altivec(uint8_t src[], int stride, PPContext *c)
 
   vector signed short v_srcAss0, v_srcAss1, v_srcAss2, v_srcAss3, v_srcAss4, v_srcAss5, v_srcAss6, v_srcAss7;
 
-#define LOAD_LINE(i)							\
-  register int j##i = i * stride;					\
-  vector unsigned char perm##i = vec_lvsl(j##i, src2);			\
-  const vector unsigned char v_srcA1##i = vec_ld(j##i, src2);		\
-  vector unsigned char v_srcA2##i;					\
-  if (two_vectors)							\
-    v_srcA2##i = vec_ld(j##i + 16, src2);				\
-  const vector unsigned char v_srcA##i =				\
-    vec_perm(v_srcA1##i, v_srcA2##i, perm##i);				\
+#define LOAD_LINE(i)                                                    \
+  register int j##i = i * stride;                                       \
+  vector unsigned char perm##i = vec_lvsl(j##i, src2);                  \
+  const vector unsigned char v_srcA1##i = vec_ld(j##i, src2);           \
+  vector unsigned char v_srcA2##i;                                      \
+  if (two_vectors)                                                      \
+    v_srcA2##i = vec_ld(j##i + 16, src2);                               \
+  const vector unsigned char v_srcA##i =                                \
+    vec_perm(v_srcA1##i, v_srcA2##i, perm##i);                          \
   v_srcAss##i =                                                         \
-    (vector signed short)vec_mergeh((vector signed char)zero,		\
-				    (vector signed char)v_srcA##i)
+    (vector signed short)vec_mergeh((vector signed char)zero,           \
+                                    (vector signed char)v_srcA##i)
 
 #define LOAD_LINE_ALIGNED(i)                                            \
   register int j##i = i * stride;                                       \
   const vector unsigned char v_srcA##i = vec_ld(j##i, src2);            \
   v_srcAss##i =                                                         \
-    (vector signed short)vec_mergeh((vector signed char)zero,		\
-				    (vector signed char)v_srcA##i)
+    (vector signed short)vec_mergeh((vector signed char)zero,           \
+                                    (vector signed char)v_srcA##i)
 
     // special casing the aligned case is worthwhile, as all call from
     // the (transposed) horizontable deblocks will be aligned, i naddition
@@ -139,15 +139,15 @@ static inline int vertClassify_altivec(uint8_t src[], int stride, PPContext *c)
 #undef LOAD_LINE
 #undef LOAD_LINE_ALIGNED
 
-#define ITER(i, j)							\
-  const vector signed short v_diff##i =					\
-    vec_sub(v_srcAss##i, v_srcAss##j);					\
-  const vector signed short v_sum##i =					\
-    vec_add(v_diff##i, v_dcOffset);					\
-  const vector signed short v_comp##i =					\
-    (vector signed short)vec_cmplt((vector unsigned short)v_sum##i,	\
-				   v_dcThreshold);			\
-  const vector signed short v_part##i = vec_and(mask, v_comp##i);	\
+#define ITER(i, j)                                                      \
+  const vector signed short v_diff##i =                                 \
+    vec_sub(v_srcAss##i, v_srcAss##j);                                  \
+  const vector signed short v_sum##i =                                  \
+    vec_add(v_diff##i, v_dcOffset);                                     \
+  const vector signed short v_comp##i =                                 \
+    (vector signed short)vec_cmplt((vector unsigned short)v_sum##i,     \
+                                   v_dcThreshold);                      \
+  const vector signed short v_part##i = vec_and(mask, v_comp##i);       \
   v_numEq = vec_sum4s(v_part##i, v_numEq);
 
   ITER(0, 1);
@@ -167,13 +167,13 @@ static inline int vertClassify_altivec(uint8_t src[], int stride, PPContext *c)
   if (numEq > c->ppMode.flatnessThreshold)
     {
       const vector unsigned char mmoP1 = (const vector unsigned char)
-	AVV(0x1f, 0x1f, 0x1f, 0x1f, 0x1f, 0x1f, 0x1f, 0x1f,
-	    0x00, 0x01, 0x12, 0x13, 0x08, 0x09, 0x1A, 0x1B);
+        AVV(0x1f, 0x1f, 0x1f, 0x1f, 0x1f, 0x1f, 0x1f, 0x1f,
+            0x00, 0x01, 0x12, 0x13, 0x08, 0x09, 0x1A, 0x1B);
       const vector unsigned char mmoP2 = (const vector unsigned char)
-	AVV(0x04, 0x05, 0x16, 0x17, 0x0C, 0x0D, 0x1E, 0x1F,
-	    0x1f, 0x1f, 0x1f, 0x1f, 0x1f, 0x1f, 0x1f, 0x1f);
+        AVV(0x04, 0x05, 0x16, 0x17, 0x0C, 0x0D, 0x1E, 0x1F,
+            0x1f, 0x1f, 0x1f, 0x1f, 0x1f, 0x1f, 0x1f, 0x1f);
       const vector unsigned char mmoP = (const vector unsigned char)
-	vec_lvsl(8, (unsigned char*)0);
+        vec_lvsl(8, (unsigned char*)0);
 
       vector signed short mmoL1 = vec_perm(v_srcAss0, v_srcAss2, mmoP1);
       vector signed short mmoL2 = vec_perm(v_srcAss4, v_srcAss6, mmoP2);
@@ -185,9 +185,9 @@ static inline int vertClassify_altivec(uint8_t src[], int stride, PPContext *c)
       vector unsigned short mmoSum = (vector unsigned short)vec_add(mmoDiff, v2QP);
 
       if (vec_any_gt(mmoSum, v4QP))
-	return 0;
+        return 0;
       else
-	return 1;
+        return 1;
     }
   else return 2;
 }
@@ -218,21 +218,21 @@ static inline void doVertLowPass_altivec(uint8_t *src, int stride, PPContext *c)
   vector unsigned char vbT0, vbT1, vbT2, vbT3, vbT4, vbT5, vbT6, vbT7, vbT8, vbT9;
 
 #define LOAD_LINE(i)                                                    \
-  const vector unsigned char perml##i =					\
-    vec_lvsl(i * stride, src2);						\
+  const vector unsigned char perml##i =                                 \
+    vec_lvsl(i * stride, src2);                                         \
   vbA##i = vec_ld(i * stride, src2);                                    \
   vbB##i = vec_ld(i * stride + 16, src2);                               \
   vbT##i = vec_perm(vbA##i, vbB##i, perml##i);                          \
   vb##i =                                                               \
-    (vector signed short)vec_mergeh((vector unsigned char)zero,		\
-				    (vector unsigned char)vbT##i)
+    (vector signed short)vec_mergeh((vector unsigned char)zero,         \
+                                    (vector unsigned char)vbT##i)
 
 #define LOAD_LINE_ALIGNED(i)                                            \
   register int j##i = i * stride;                                       \
   vbT##i = vec_ld(j##i, src2);                                          \
   vb##i =                                                               \
-    (vector signed short)vec_mergeh((vector signed char)zero,		\
-				    (vector signed char)vbT##i)
+    (vector signed short)vec_mergeh((vector signed char)zero,           \
+                                    (vector signed char)vbT##i)
 
     // special casing the aligned case is worthwhile, as all call from
     // the (transposed) horizontable deblocks will be aligned, in addition
@@ -308,11 +308,11 @@ static inline void doVertLowPass_altivec(uint8_t *src, int stride, PPContext *c)
   const vector signed short temp91 = vec_sub(v_sumsB8, vb5);
   const vector signed short v_sumsB9 = vec_add(temp91, v_last);
 
-#define COMPUTE_VR(i, j, k)						\
-  const vector signed short temps1##i =					\
-    vec_add(v_sumsB##i, v_sumsB##k);					\
-  const vector signed short temps2##i =					\
-    vec_mladd(vb##j, (vector signed short)v_2, temps1##i);		\
+#define COMPUTE_VR(i, j, k)                                             \
+  const vector signed short temps1##i =                                 \
+    vec_add(v_sumsB##i, v_sumsB##k);                                    \
+  const vector signed short temps2##i =                                 \
+    vec_mladd(vb##j, (vector signed short)v_2, temps1##i);              \
   const vector signed short  vr##j = vec_sra(temps2##i, v_4)
 
   COMPUTE_VR(0, 1, 2);
@@ -326,31 +326,31 @@ static inline void doVertLowPass_altivec(uint8_t *src, int stride, PPContext *c)
 
   const vector signed char neg1 = vec_splat_s8(-1);
   const vector unsigned char permHH = (const vector unsigned char)AVV(0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
-								      0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F);
-
-#define PACK_AND_STORE(i)					\
-  const vector unsigned char perms##i =				\
-    vec_lvsr(i * stride, src2);					\
-  const vector unsigned char vf##i =				\
-    vec_packsu(vr##i, (vector signed short)zero);		\
-  const vector unsigned char vg##i =				\
-    vec_perm(vf##i, vbT##i, permHH);				\
-  const vector unsigned char mask##i =				\
-    vec_perm((vector unsigned char)zero, (vector unsigned char)neg1, perms##i);	\
-  const vector unsigned char vg2##i =				\
-    vec_perm(vg##i, vg##i, perms##i);				\
-  const vector unsigned char svA##i =				\
-    vec_sel(vbA##i, vg2##i, mask##i);				\
-  const vector unsigned char svB##i =				\
-    vec_sel(vg2##i, vbB##i, mask##i);				\
-  vec_st(svA##i, i * stride, src2);				\
+                                                                      0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F);
+
+#define PACK_AND_STORE(i)                                       \
+  const vector unsigned char perms##i =                         \
+    vec_lvsr(i * stride, src2);                                 \
+  const vector unsigned char vf##i =                            \
+    vec_packsu(vr##i, (vector signed short)zero);               \
+  const vector unsigned char vg##i =                            \
+    vec_perm(vf##i, vbT##i, permHH);                            \
+  const vector unsigned char mask##i =                          \
+    vec_perm((vector unsigned char)zero, (vector unsigned char)neg1, perms##i); \
+  const vector unsigned char vg2##i =                           \
+    vec_perm(vg##i, vg##i, perms##i);                           \
+  const vector unsigned char svA##i =                           \
+    vec_sel(vbA##i, vg2##i, mask##i);                           \
+  const vector unsigned char svB##i =                           \
+    vec_sel(vg2##i, vbB##i, mask##i);                           \
+  vec_st(svA##i, i * stride, src2);                             \
   vec_st(svB##i, i * stride + 16, src2)
 
-#define PACK_AND_STORE_ALIGNED(i)				\
-  const vector unsigned char vf##i =				\
-    vec_packsu(vr##i, (vector signed short)zero);		\
-  const vector unsigned char vg##i =				\
-    vec_perm(vf##i, vbT##i, permHH);				\
+#define PACK_AND_STORE_ALIGNED(i)                               \
+  const vector unsigned char vf##i =                            \
+    vec_packsu(vr##i, (vector signed short)zero);               \
+  const vector unsigned char vg##i =                            \
+    vec_perm(vf##i, vbT##i, permHH);                            \
   vec_st(vg##i, i * stride, src2)
 
   // special casing the aligned case is worthwhile, as all call from
@@ -398,17 +398,17 @@ static inline void doVertDefFilter_altivec(uint8_t src[], int stride, PPContext
   vqp = vec_splat(vqp, 0);
 
 #define LOAD_LINE(i)                                                    \
-  const vector unsigned char perm##i =					\
-    vec_lvsl(i * stride, src2);						\
-  const vector unsigned char vbA##i =					\
-    vec_ld(i * stride, src2);						\
-  const vector unsigned char vbB##i =					\
-    vec_ld(i * stride + 16, src2);					\
-  const vector unsigned char vbT##i =					\
-    vec_perm(vbA##i, vbB##i, perm##i);					\
-  const vector signed short vb##i =					\
-    (vector signed short)vec_mergeh((vector unsigned char)zero,		\
-				    (vector unsigned char)vbT##i)
+  const vector unsigned char perm##i =                                  \
+    vec_lvsl(i * stride, src2);                                         \
+  const vector unsigned char vbA##i =                                   \
+    vec_ld(i * stride, src2);                                           \
+  const vector unsigned char vbB##i =                                   \
+    vec_ld(i * stride + 16, src2);                                      \
+  const vector unsigned char vbT##i =                                   \
+    vec_perm(vbA##i, vbB##i, perm##i);                                  \
+  const vector signed short vb##i =                                     \
+    (vector signed short)vec_mergeh((vector unsigned char)zero,         \
+                                    (vector unsigned char)vbT##i)
 
   src2 += stride*3;
 
@@ -426,7 +426,7 @@ static inline void doVertDefFilter_altivec(uint8_t src[], int stride, PPContext
   const vector signed short v_2 = vec_splat_s16(2);
   const vector signed short v_5 = vec_splat_s16(5);
   const vector signed short v_32 = vec_sl(v_1,
-					  (vector unsigned short)v_5);
+                                          (vector unsigned short)v_5);
   /* middle energy */
   const vector signed short l3minusl6 = vec_sub(vb3, vb6);
   const vector signed short l5minusl4 = vec_sub(vb5, vb4);
@@ -483,22 +483,22 @@ static inline void doVertDefFilter_altivec(uint8_t src[], int stride, PPContext
 
   const vector signed char neg1 = vec_splat_s8(-1);
   const vector unsigned char permHH = (const vector unsigned char)AVV(0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
-								      0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F);
-
-#define STORE(i)						\
-  const vector unsigned char perms##i =				\
-    vec_lvsr(i * stride, src2);					\
-  const vector unsigned char vg##i =				\
-    vec_perm(st##i, vbT##i, permHH);				\
-  const vector unsigned char mask##i =				\
-    vec_perm((vector unsigned char)zero, (vector unsigned char)neg1, perms##i);	\
-  const vector unsigned char vg2##i =				\
-    vec_perm(vg##i, vg##i, perms##i);				\
-  const vector unsigned char svA##i =				\
-    vec_sel(vbA##i, vg2##i, mask##i);				\
-  const vector unsigned char svB##i =				\
-    vec_sel(vg2##i, vbB##i, mask##i);				\
-  vec_st(svA##i, i * stride, src2);				\
+                                                                      0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F);
+
+#define STORE(i)                                                \
+  const vector unsigned char perms##i =                         \
+    vec_lvsr(i * stride, src2);                                 \
+  const vector unsigned char vg##i =                            \
+    vec_perm(st##i, vbT##i, permHH);                            \
+  const vector unsigned char mask##i =                          \
+    vec_perm((vector unsigned char)zero, (vector unsigned char)neg1, perms##i); \
+  const vector unsigned char vg2##i =                           \
+    vec_perm(vg##i, vg##i, perms##i);                           \
+  const vector unsigned char svA##i =                           \
+    vec_sel(vbA##i, vg2##i, mask##i);                           \
+  const vector unsigned char svB##i =                           \
+    vec_sel(vg2##i, vbB##i, mask##i);                           \
+  vec_st(svA##i, i * stride, src2);                             \
   vec_st(svB##i, i * stride + 16, src2)
 
   STORE(4);
@@ -522,11 +522,11 @@ static inline void dering_altivec(uint8_t src[], int stride, PPContext *c) {
   dt[0] = deringThreshold;
   v_dt = vec_splat(vec_ld(0, dt), 0);
 
-#define LOAD_LINE(i)							\
-  const vector unsigned char perm##i =					\
-    vec_lvsl(i * stride, srcCopy);					\
-  vector unsigned char sA##i = vec_ld(i * stride, srcCopy);		\
-  vector unsigned char sB##i = vec_ld(i * stride + 16, srcCopy);	\
+#define LOAD_LINE(i)                                                    \
+  const vector unsigned char perm##i =                                  \
+    vec_lvsl(i * stride, srcCopy);                                      \
+  vector unsigned char sA##i = vec_ld(i * stride, srcCopy);             \
+  vector unsigned char sB##i = vec_ld(i * stride + 16, srcCopy);        \
   vector unsigned char src##i = vec_perm(sA##i, sB##i, perm##i)
 
   LOAD_LINE(0);
@@ -545,13 +545,13 @@ static inline void dering_altivec(uint8_t src[], int stride, PPContext *c) {
   {
     const vector unsigned char trunc_perm = (vector unsigned char)
       AVV(0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08,
-	  0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, 0x18);
+          0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, 0x18);
     const vector unsigned char trunc_src12 = vec_perm(src1, src2, trunc_perm);
     const vector unsigned char trunc_src34 = vec_perm(src3, src4, trunc_perm);
     const vector unsigned char trunc_src56 = vec_perm(src5, src6, trunc_perm);
     const vector unsigned char trunc_src78 = vec_perm(src7, src8, trunc_perm);
 
-#define EXTRACT(op) do {						\
+#define EXTRACT(op) do {                                                \
       const vector unsigned char s##op##_1 = vec_##op(trunc_src12, trunc_src34); \
       const vector unsigned char s##op##_2 = vec_##op(trunc_src56, trunc_src78); \
       const vector unsigned char s##op##_6 = vec_##op(s##op##_1, s##op##_2); \
@@ -584,29 +584,29 @@ static inline void dering_altivec(uint8_t src[], int stride, PPContext *c) {
   {
     const vector unsigned short mask1 = (vector unsigned short)
       AVV(0x0001, 0x0002, 0x0004, 0x0008,
-	  0x0010, 0x0020, 0x0040, 0x0080);
+          0x0010, 0x0020, 0x0040, 0x0080);
     const vector unsigned short mask2 = (vector unsigned short)
       AVV(0x0100, 0x0200, 0x0000, 0x0000,
-	  0x0000, 0x0000, 0x0000, 0x0000);
+          0x0000, 0x0000, 0x0000, 0x0000);
 
     const vector unsigned int vuint32_16 = vec_sl(vec_splat_u32(1), vec_splat_u32(4));
     const vector unsigned int vuint32_1 = vec_splat_u32(1);
 
-#define COMPARE(i)							\
-    vector signed int sum##i;						\
-    do {								\
-      const vector unsigned char cmp##i =				\
-	(vector unsigned char)vec_cmpgt(src##i, v_avg);			\
-      const vector unsigned short cmpHi##i =				\
-	(vector unsigned short)vec_mergeh(cmp##i, cmp##i);		\
-      const vector unsigned short cmpLi##i =				\
-	(vector unsigned short)vec_mergel(cmp##i, cmp##i);		\
-      const vector signed short cmpHf##i =				\
-	(vector signed short)vec_and(cmpHi##i, mask1);			\
-      const vector signed short cmpLf##i =				\
-	(vector signed short)vec_and(cmpLi##i, mask2);			\
-      const vector signed int sump##i = vec_sum4s(cmpHf##i, zero);	\
-      const vector signed int sumq##i = vec_sum4s(cmpLf##i, sump##i);	\
+#define COMPARE(i)                                                      \
+    vector signed int sum##i;                                           \
+    do {                                                                \
+      const vector unsigned char cmp##i =                               \
+        (vector unsigned char)vec_cmpgt(src##i, v_avg);                 \
+      const vector unsigned short cmpHi##i =                            \
+        (vector unsigned short)vec_mergeh(cmp##i, cmp##i);              \
+      const vector unsigned short cmpLi##i =                            \
+        (vector unsigned short)vec_mergel(cmp##i, cmp##i);              \
+      const vector signed short cmpHf##i =                              \
+        (vector signed short)vec_and(cmpHi##i, mask1);                  \
+      const vector signed short cmpLf##i =                              \
+        (vector signed short)vec_and(cmpLi##i, mask2);                  \
+      const vector signed int sump##i = vec_sum4s(cmpHf##i, zero);      \
+      const vector signed int sumq##i = vec_sum4s(cmpLf##i, sump##i);   \
       sum##i  = vec_sums(sumq##i, zero); } while (0)
 
     COMPARE(0);
@@ -643,11 +643,11 @@ static inline void dering_altivec(uint8_t src[], int stride, PPContext *c) {
       const vector signed int t2B = vec_or(sumB, tB);
       const vector signed int t2C = vec_or(sumC, tC);
       const vector signed int t3A = vec_and(vec_sra(t2A, vuint32_1),
-					    vec_sl(t2A, vuint32_1));
+                                            vec_sl(t2A, vuint32_1));
       const vector signed int t3B = vec_and(vec_sra(t2B, vuint32_1),
-					    vec_sl(t2B, vuint32_1));
+                                            vec_sl(t2B, vuint32_1));
       const vector signed int t3C = vec_and(vec_sra(t2C, vuint32_1),
-					    vec_sl(t2C, vuint32_1));
+                                            vec_sl(t2C, vuint32_1));
       const vector signed int yA = vec_and(t2A, t3A);
       const vector signed int yB = vec_and(t2B, t3B);
       const vector signed int yC = vec_and(t2C, t3C);
@@ -659,15 +659,15 @@ static inline void dering_altivec(uint8_t src[], int stride, PPContext *c) {
       const vector signed int sumBd4 = vec_perm(yB, yC, strangeperm1);
       const vector signed int sumBd8 = vec_perm(yB, yC, strangeperm2);
       const vector signed int sumAp = vec_and(yA,
-					      vec_and(sumAd4,sumAd8));
+                                              vec_and(sumAd4,sumAd8));
       const vector signed int sumBp = vec_and(yB,
-					      vec_and(sumBd4,sumBd8));
+                                              vec_and(sumBd4,sumBd8));
       sumA2 = vec_or(sumAp,
-		     vec_sra(sumAp,
-			     vuint32_16));
+                     vec_sra(sumAp,
+                             vuint32_16));
       sumB2  = vec_or(sumBp,
-		      vec_sra(sumBp,
-			      vuint32_16));
+                      vec_sra(sumBp,
+                              vuint32_16));
     }
     vec_st(sumA2, 0, S);
     vec_st(sumB2, 16, S);
@@ -686,84 +686,84 @@ static inline void dering_altivec(uint8_t src[], int stride, PPContext *c) {
 
   const vector unsigned char permA1 = (vector unsigned char)
     AVV(0x00, 0x01, 0x02, 0x10, 0x11, 0x12, 0x1F, 0x1F,
-	0x1F, 0x1F, 0x1F, 0x1F, 0x1F, 0x1F, 0x1F, 0x1F);
+        0x1F, 0x1F, 0x1F, 0x1F, 0x1F, 0x1F, 0x1F, 0x1F);
   const vector unsigned char permA2 = (vector unsigned char)
     AVV(0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x10, 0x11,
-	0x12, 0x1F, 0x1F, 0x1F, 0x1F, 0x1F, 0x1F, 0x1F);
+        0x12, 0x1F, 0x1F, 0x1F, 0x1F, 0x1F, 0x1F, 0x1F);
   const vector unsigned char permA1inc = (vector unsigned char)
     AVV(0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x00, 0x00,
-	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00);
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00);
   const vector unsigned char permA2inc = (vector unsigned char)
     AVV(0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01,
-	0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00);
+        0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00);
   const vector unsigned char magic = (vector unsigned char)
     AVV(0x01, 0x02, 0x01, 0x02, 0x04, 0x02, 0x01, 0x02,
-	0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00);
+        0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00);
   const vector unsigned char extractPerm = (vector unsigned char)
     AVV(0x10, 0x10, 0x10, 0x01, 0x10, 0x10, 0x10, 0x01,
-	0x10, 0x10, 0x10, 0x01, 0x10, 0x10, 0x10, 0x01);
+        0x10, 0x10, 0x10, 0x01, 0x10, 0x10, 0x10, 0x01);
   const vector unsigned char extractPermInc = (vector unsigned char)
     AVV(0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01,
-	0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01);
+        0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01);
   const vector unsigned char identity = vec_lvsl(0,(unsigned char *)0);
   const vector unsigned char tenRight = (vector unsigned char)
     AVV(0x00, 0x10, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00);
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00);
   const vector unsigned char eightLeft = (vector unsigned char)
     AVV(0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x08);
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x08);
 
 
-#define F_INIT(i)					\
-  vector unsigned char tenRightM##i = tenRight;		\
-  vector unsigned char permA1M##i = permA1;		\
-  vector unsigned char permA2M##i = permA2;		\
+#define F_INIT(i)                                       \
+  vector unsigned char tenRightM##i = tenRight;         \
+  vector unsigned char permA1M##i = permA1;             \
+  vector unsigned char permA2M##i = permA2;             \
   vector unsigned char extractPermM##i = extractPerm
 
-#define F2(i, j, k, l)							\
-  if (S[i] & (1 << (l+1))) {						\
-    const vector unsigned char a_##j##_A##l =				\
-      vec_perm(src##i, src##j, permA1M##i);				\
-    const vector unsigned char a_##j##_B##l =				\
-      vec_perm(a_##j##_A##l, src##k, permA2M##i);			\
-    const vector signed int a_##j##_sump##l =				\
-      (vector signed int)vec_msum(a_##j##_B##l, magic,			\
-				  (vector unsigned int)zero);		\
-    vector signed int F_##j##_##l =					\
-      vec_sr(vec_sums(a_##j##_sump##l, vsint32_8), vuint32_4);		\
-    F_##j##_##l = vec_splat(F_##j##_##l, 3);				\
-    const vector signed int p_##j##_##l =				\
-      (vector signed int)vec_perm(src##j,				\
-				  (vector unsigned char)zero,		\
-				  extractPermM##i);			\
-    const vector signed int sum_##j##_##l = vec_add( p_##j##_##l, vQP2); \
-    const vector signed int diff_##j##_##l = vec_sub( p_##j##_##l, vQP2); \
-    vector signed int newpm_##j##_##l;					\
-    if (vec_all_lt(sum_##j##_##l, F_##j##_##l))				\
-      newpm_##j##_##l = sum_##j##_##l;					\
-    else if (vec_all_gt(diff_##j##_##l, F_##j##_##l))			\
-      newpm_##j##_##l = diff_##j##_##l;					\
-    else newpm_##j##_##l = F_##j##_##l;					\
-    const vector unsigned char newpm2_##j##_##l =			\
-      vec_splat((vector unsigned char)newpm_##j##_##l, 15);		\
-    const vector unsigned char mask##j##l = vec_add(identity,		\
-						    tenRightM##i);	\
-    src##j = vec_perm(src##j, newpm2_##j##_##l, mask##j##l);		\
-  }									\
-  permA1M##i = vec_add(permA1M##i, permA1inc);				\
-  permA2M##i = vec_add(permA2M##i, permA2inc);				\
-  tenRightM##i = vec_sro(tenRightM##i, eightLeft);			\
+#define F2(i, j, k, l)                                                  \
+  if (S[i] & (1 << (l+1))) {                                            \
+    const vector unsigned char a_##j##_A##l =                           \
+      vec_perm(src##i, src##j, permA1M##i);                             \
+    const vector unsigned char a_##j##_B##l =                           \
+      vec_perm(a_##j##_A##l, src##k, permA2M##i);                       \
+    const vector signed int a_##j##_sump##l =                           \
+      (vector signed int)vec_msum(a_##j##_B##l, magic,                  \
+                                  (vector unsigned int)zero);           \
+    vector signed int F_##j##_##l =                                     \
+      vec_sr(vec_sums(a_##j##_sump##l, vsint32_8), vuint32_4);          \
+    F_##j##_##l = vec_splat(F_##j##_##l, 3);                            \
+    const vector signed int p_##j##_##l =                               \
+      (vector signed int)vec_perm(src##j,                               \
+                                  (vector unsigned char)zero,           \
+                                  extractPermM##i);                     \
+    const vector signed int sum_##j##_##l = vec_add( p_##j##_##l, vQP2);\
+    const vector signed int diff_##j##_##l = vec_sub( p_##j##_##l, vQP2);\
+    vector signed int newpm_##j##_##l;                                  \
+    if (vec_all_lt(sum_##j##_##l, F_##j##_##l))                         \
+      newpm_##j##_##l = sum_##j##_##l;                                  \
+    else if (vec_all_gt(diff_##j##_##l, F_##j##_##l))                   \
+      newpm_##j##_##l = diff_##j##_##l;                                 \
+    else newpm_##j##_##l = F_##j##_##l;                                 \
+    const vector unsigned char newpm2_##j##_##l =                       \
+      vec_splat((vector unsigned char)newpm_##j##_##l, 15);             \
+    const vector unsigned char mask##j##l = vec_add(identity,           \
+                                                    tenRightM##i);      \
+    src##j = vec_perm(src##j, newpm2_##j##_##l, mask##j##l);            \
+  }                                                                     \
+  permA1M##i = vec_add(permA1M##i, permA1inc);                          \
+  permA2M##i = vec_add(permA2M##i, permA2inc);                          \
+  tenRightM##i = vec_sro(tenRightM##i, eightLeft);                      \
   extractPermM##i = vec_add(extractPermM##i, extractPermInc)
 
-#define ITER(i, j, k)				\
-  F_INIT(i);					\
-  F2(i, j, k, 0);				\
-  F2(i, j, k, 1);				\
-  F2(i, j, k, 2);				\
-  F2(i, j, k, 3);				\
-  F2(i, j, k, 4);				\
-  F2(i, j, k, 5);				\
-  F2(i, j, k, 6);				\
+#define ITER(i, j, k)                           \
+  F_INIT(i);                                    \
+  F2(i, j, k, 0);                               \
+  F2(i, j, k, 1);                               \
+  F2(i, j, k, 2);                               \
+  F2(i, j, k, 3);                               \
+  F2(i, j, k, 4);                               \
+  F2(i, j, k, 5);                               \
+  F2(i, j, k, 6);                               \
   F2(i, j, k, 7)
 
   ITER(0, 1, 2);
@@ -777,16 +777,16 @@ static inline void dering_altivec(uint8_t src[], int stride, PPContext *c) {
 
   const vector signed char neg1 = vec_splat_s8(-1);
 
-#define STORE_LINE(i)					\
-  const vector unsigned char permST##i =		\
-    vec_lvsr(i * stride, srcCopy);			\
-  const vector unsigned char maskST##i =		\
-    vec_perm((vector unsigned char)zero,		\
-	     (vector unsigned char)neg1, permST##i);	\
-  src##i = vec_perm(src##i ,src##i, permST##i);		\
-  sA##i= vec_sel(sA##i, src##i, maskST##i);		\
-  sB##i= vec_sel(src##i, sB##i, maskST##i);		\
-  vec_st(sA##i, i * stride, srcCopy);			\
+#define STORE_LINE(i)                                   \
+  const vector unsigned char permST##i =                \
+    vec_lvsr(i * stride, srcCopy);                      \
+  const vector unsigned char maskST##i =                \
+    vec_perm((vector unsigned char)zero,                \
+             (vector unsigned char)neg1, permST##i);    \
+  src##i = vec_perm(src##i ,src##i, permST##i);         \
+  sA##i= vec_sel(sA##i, src##i, maskST##i);             \
+  sB##i= vec_sel(src##i, sB##i, maskST##i);             \
+  vec_st(sA##i, i * stride, srcCopy);                   \
   vec_st(sB##i, i * stride + 16, srcCopy)
 
   STORE_LINE(1);
@@ -808,7 +808,7 @@ static inline void dering_altivec(uint8_t src[], int stride, PPContext *c) {
 #define do_a_deblock_altivec(a...) do_a_deblock_C(a)
 
 static inline void RENAME(tempNoiseReducer)(uint8_t *src, int stride,
-				    uint8_t *tempBlured, uint32_t *tempBluredPast, int *maxNoise)
+                                    uint8_t *tempBlured, uint32_t *tempBluredPast, int *maxNoise)
 {
   const vector signed int zero = vec_splat_s32(0);
   const vector signed short vsint16_1 = vec_splat_s16(1);
@@ -820,16 +820,16 @@ static inline void RENAME(tempNoiseReducer)(uint8_t *src, int stride,
   tempBluredPast[128]= maxNoise[1];
   tempBluredPast[129]= maxNoise[2];
 
-#define LOAD_LINE(src, i)						\
-  register int j##src##i = i * stride;					\
-  vector unsigned char perm##src##i = vec_lvsl(j##src##i, src);		\
-  const vector unsigned char v_##src##A1##i = vec_ld(j##src##i, src);	\
+#define LOAD_LINE(src, i)                                               \
+  register int j##src##i = i * stride;                                  \
+  vector unsigned char perm##src##i = vec_lvsl(j##src##i, src);         \
+  const vector unsigned char v_##src##A1##i = vec_ld(j##src##i, src);   \
   const vector unsigned char v_##src##A2##i = vec_ld(j##src##i + 16, src); \
-  const vector unsigned char v_##src##A##i =				\
-    vec_perm(v_##src##A1##i, v_##src##A2##i, perm##src##i);		\
-  vector signed short v_##src##Ass##i =					\
-    (vector signed short)vec_mergeh((vector signed char)zero,		\
-				    (vector signed char)v_##src##A##i)
+  const vector unsigned char v_##src##A##i =                            \
+    vec_perm(v_##src##A1##i, v_##src##A2##i, perm##src##i);             \
+  vector signed short v_##src##Ass##i =                                 \
+    (vector signed short)vec_mergeh((vector signed char)zero,           \
+                                    (vector signed char)v_##src##A##i)
 
   LOAD_LINE(src, 0);
   LOAD_LINE(src, 1);
@@ -850,10 +850,10 @@ static inline void RENAME(tempNoiseReducer)(uint8_t *src, int stride,
   LOAD_LINE(tempBlured, 7);
 #undef LOAD_LINE
 
-#define ACCUMULATE_DIFFS(i)					\
-  vector signed short v_d##i = vec_sub(v_tempBluredAss##i,	\
-				       v_srcAss##i);		\
-  v_dp = vec_msums(v_d##i, v_d##i, v_dp);			\
+#define ACCUMULATE_DIFFS(i)                                     \
+  vector signed short v_d##i = vec_sub(v_tempBluredAss##i,      \
+                                       v_srcAss##i);            \
+  v_dp = vec_msums(v_d##i, v_d##i, v_dp);                       \
   v_sysdp = vec_msums(v_d##i, vsint16_1, v_sysdp)
 
   ACCUMULATE_DIFFS(0);
@@ -916,12 +916,12 @@ static inline void RENAME(tempNoiseReducer)(uint8_t *src, int stride,
       const vector signed short vsint16_4 = vec_splat_s16(4);
       const vector unsigned short vuint16_3 = vec_splat_u16(3);
 
-#define OP(i)								\
-      const vector signed short v_temp##i =				\
-	vec_mladd(v_tempBluredAss##i,					\
-		  vsint16_7, v_srcAss##i);				\
-      const vector signed short v_temp2##i =				\
-	vec_add(v_temp##i, vsint16_4);					\
+#define OP(i)                                                   \
+      const vector signed short v_temp##i =                     \
+        vec_mladd(v_tempBluredAss##i,                           \
+                  vsint16_7, v_srcAss##i);                      \
+      const vector signed short v_temp2##i =                    \
+        vec_add(v_temp##i, vsint16_4);                          \
       v_tempBluredAss##i = vec_sr(v_temp2##i, vuint16_3)
 
       OP(0);
@@ -937,12 +937,12 @@ static inline void RENAME(tempNoiseReducer)(uint8_t *src, int stride,
       const vector signed short vsint16_3 = vec_splat_s16(3);
       const vector signed short vsint16_2 = vec_splat_s16(2);
 
-#define OP(i)								\
-      const vector signed short v_temp##i =				\
-	vec_mladd(v_tempBluredAss##i,					\
-		  vsint16_3, v_srcAss##i);				\
-      const vector signed short v_temp2##i =				\
-	vec_add(v_temp##i, vsint16_2);					\
+#define OP(i)                                                   \
+      const vector signed short v_temp##i =                     \
+        vec_mladd(v_tempBluredAss##i,                           \
+                  vsint16_3, v_srcAss##i);                      \
+      const vector signed short v_temp2##i =                    \
+        vec_add(v_temp##i, vsint16_2);                          \
       v_tempBluredAss##i = vec_sr(v_temp2##i, (vector unsigned short)vsint16_2)
 
       OP(0);
@@ -959,24 +959,24 @@ static inline void RENAME(tempNoiseReducer)(uint8_t *src, int stride,
 
   const vector signed char neg1 = vec_splat_s8(-1);
   const vector unsigned char permHH = (const vector unsigned char)AVV(0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
-								      0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F);
-
-#define PACK_AND_STORE(src, i)						\
-  const vector unsigned char perms##src##i =				\
-    vec_lvsr(i * stride, src);						\
-  const vector unsigned char vf##src##i =				\
-    vec_packsu(v_tempBluredAss##i, (vector signed short)zero);		\
-  const vector unsigned char vg##src##i =				\
-    vec_perm(vf##src##i, v_##src##A##i, permHH);			\
-  const vector unsigned char mask##src##i =				\
+                                                                      0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F);
+
+#define PACK_AND_STORE(src, i)                                  \
+  const vector unsigned char perms##src##i =                    \
+    vec_lvsr(i * stride, src);                                  \
+  const vector unsigned char vf##src##i =                       \
+    vec_packsu(v_tempBluredAss##i, (vector signed short)zero);  \
+  const vector unsigned char vg##src##i =                       \
+    vec_perm(vf##src##i, v_##src##A##i, permHH);                \
+  const vector unsigned char mask##src##i =                     \
     vec_perm((vector unsigned char)zero, (vector unsigned char)neg1, perms##src##i); \
-  const vector unsigned char vg2##src##i =				\
-    vec_perm(vg##src##i, vg##src##i, perms##src##i);			\
-  const vector unsigned char svA##src##i =				\
-    vec_sel(v_##src##A1##i, vg2##src##i, mask##src##i);			\
-  const vector unsigned char svB##src##i =				\
-    vec_sel(vg2##src##i, v_##src##A2##i, mask##src##i);			\
-  vec_st(svA##src##i, i * stride, src);					\
+  const vector unsigned char vg2##src##i =                      \
+    vec_perm(vg##src##i, vg##src##i, perms##src##i);            \
+  const vector unsigned char svA##src##i =                      \
+    vec_sel(v_##src##A1##i, vg2##src##i, mask##src##i);         \
+  const vector unsigned char svB##src##i =                      \
+    vec_sel(vg2##src##i, v_##src##A2##i, mask##src##i);         \
+  vec_st(svA##src##i, i * stride, src);                         \
   vec_st(svB##src##i, i * stride + 16, src)
 
   PACK_AND_STORE(src, 0);
@@ -1001,14 +1001,14 @@ static inline void RENAME(tempNoiseReducer)(uint8_t *src, int stride,
 static inline void transpose_16x8_char_toPackedAlign_altivec(unsigned char* dst, unsigned char* src, int stride) {
   const vector unsigned char zero = vec_splat_u8(0);
 
-#define LOAD_DOUBLE_LINE(i, j)						\
-  vector unsigned char perm1##i = vec_lvsl(i * stride, src);		\
-  vector unsigned char perm2##i = vec_lvsl(j * stride, src);		\
-  vector unsigned char srcA##i = vec_ld(i * stride, src);		\
+#define LOAD_DOUBLE_LINE(i, j)                                          \
+  vector unsigned char perm1##i = vec_lvsl(i * stride, src);            \
+  vector unsigned char perm2##i = vec_lvsl(j * stride, src);            \
+  vector unsigned char srcA##i = vec_ld(i * stride, src);               \
   vector unsigned char srcB##i = vec_ld(i * stride + 16, src);          \
-  vector unsigned char srcC##i = vec_ld(j * stride, src);		\
+  vector unsigned char srcC##i = vec_ld(j * stride, src);               \
   vector unsigned char srcD##i = vec_ld(j * stride+ 16, src);           \
-  vector unsigned char src##i = vec_perm(srcA##i, srcB##i, perm1##i);	\
+  vector unsigned char src##i = vec_perm(srcA##i, srcB##i, perm1##i);   \
   vector unsigned char src##j = vec_perm(srcC##i, srcD##i, perm2##i)
 
   LOAD_DOUBLE_LINE(0, 1);
@@ -1107,10 +1107,10 @@ static inline void transpose_8x16_char_fromPackedAlign_altivec(unsigned char* ds
   const vector unsigned char zero = vec_splat_u8(0);
   const vector unsigned char magic_perm = (const vector unsigned char)
     AVV(0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
-	0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F);
+        0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F);
 
-#define LOAD_DOUBLE_LINE(i, j)			    		\
-  vector unsigned char src##i = vec_ld(i * 16, src);		\
+#define LOAD_DOUBLE_LINE(i, j)                                  \
+  vector unsigned char src##i = vec_ld(i * 16, src);            \
   vector unsigned char src##j = vec_ld(j * 16, src)
 
   LOAD_DOUBLE_LINE(0, 1);
@@ -1169,24 +1169,24 @@ static inline void transpose_8x16_char_fromPackedAlign_altivec(unsigned char* ds
 
 
   const vector signed char neg1 = vec_splat_s8(-1);
-#define STORE_DOUBLE_LINE(i, j)						\
-  vector unsigned char dstA##i = vec_ld(i * stride, dst);		\
-  vector unsigned char dstB##i = vec_ld(i * stride + 16, dst);		\
-  vector unsigned char dstA##j = vec_ld(j * stride, dst);		\
-  vector unsigned char dstB##j = vec_ld(j * stride+ 16, dst);		\
-  vector unsigned char align##i = vec_lvsr(i * stride, dst);		\
-  vector unsigned char align##j = vec_lvsr(j * stride, dst);		\
-  vector unsigned char mask##i = vec_perm(zero, (vector unsigned char)neg1, align##i);	\
-  vector unsigned char mask##j = vec_perm(zero, (vector unsigned char)neg1, align##j);	\
-  vector unsigned char dstR##i = vec_perm(temp##i, temp##i, align##i);	\
-  vector unsigned char dstR##j = vec_perm(temp##j, temp##j, align##j);	\
-  vector unsigned char dstAF##i = vec_sel(dstA##i, dstR##i, mask##i);	\
-  vector unsigned char dstBF##i = vec_sel(dstR##i, dstB##i, mask##i);	\
-  vector unsigned char dstAF##j = vec_sel(dstA##j, dstR##j, mask##j);	\
-  vector unsigned char dstBF##j = vec_sel(dstR##j, dstB##j, mask##j);	\
-  vec_st(dstAF##i, i * stride, dst);					\
-  vec_st(dstBF##i, i * stride + 16, dst);				\
-  vec_st(dstAF##j, j * stride, dst);					\
+#define STORE_DOUBLE_LINE(i, j)                                         \
+  vector unsigned char dstA##i = vec_ld(i * stride, dst);               \
+  vector unsigned char dstB##i = vec_ld(i * stride + 16, dst);          \
+  vector unsigned char dstA##j = vec_ld(j * stride, dst);               \
+  vector unsigned char dstB##j = vec_ld(j * stride+ 16, dst);           \
+  vector unsigned char align##i = vec_lvsr(i * stride, dst);            \
+  vector unsigned char align##j = vec_lvsr(j * stride, dst);            \
+  vector unsigned char mask##i = vec_perm(zero, (vector unsigned char)neg1, align##i); \
+  vector unsigned char mask##j = vec_perm(zero, (vector unsigned char)neg1, align##j); \
+  vector unsigned char dstR##i = vec_perm(temp##i, temp##i, align##i);  \
+  vector unsigned char dstR##j = vec_perm(temp##j, temp##j, align##j);  \
+  vector unsigned char dstAF##i = vec_sel(dstA##i, dstR##i, mask##i);   \
+  vector unsigned char dstBF##i = vec_sel(dstR##i, dstB##i, mask##i);   \
+  vector unsigned char dstAF##j = vec_sel(dstA##j, dstR##j, mask##j);   \
+  vector unsigned char dstBF##j = vec_sel(dstR##j, dstB##j, mask##j);   \
+  vec_st(dstAF##i, i * stride, dst);                                    \
+  vec_st(dstBF##i, i * stride + 16, dst);                               \
+  vec_st(dstAF##j, j * stride, dst);                                    \
   vec_st(dstBF##j, j * stride + 16, dst)
 
   STORE_DOUBLE_LINE(0,1);
diff --git a/libavcodec/libpostproc/postprocess_internal.h b/libavcodec/libpostproc/postprocess_internal.h
index f72dc05cd4..9b50c13c97 100644
--- a/libavcodec/libpostproc/postprocess_internal.h
+++ b/libavcodec/libpostproc/postprocess_internal.h
@@ -21,42 +21,42 @@
  * internal api header.
  */
 
-#define V_DEBLOCK	0x01
-#define H_DEBLOCK	0x02
-#define DERING		0x04
-#define LEVEL_FIX	0x08 ///< Brightness & Contrast
-
-#define LUM_V_DEBLOCK	V_DEBLOCK		//   1
-#define LUM_H_DEBLOCK	H_DEBLOCK		//   2
-#define CHROM_V_DEBLOCK	(V_DEBLOCK<<4)		//  16
-#define CHROM_H_DEBLOCK	(H_DEBLOCK<<4)		//  32
-#define LUM_DERING	DERING			//   4
-#define CHROM_DERING	(DERING<<4)		//  64
-#define LUM_LEVEL_FIX	LEVEL_FIX		//   8
-#define CHROM_LEVEL_FIX	(LEVEL_FIX<<4)		// 128 (not implemented yet)
+#define V_DEBLOCK       0x01
+#define H_DEBLOCK       0x02
+#define DERING          0x04
+#define LEVEL_FIX       0x08 ///< Brightness & Contrast
+
+#define LUM_V_DEBLOCK   V_DEBLOCK               //   1
+#define LUM_H_DEBLOCK   H_DEBLOCK               //   2
+#define CHROM_V_DEBLOCK (V_DEBLOCK<<4)          //  16
+#define CHROM_H_DEBLOCK (H_DEBLOCK<<4)          //  32
+#define LUM_DERING      DERING                  //   4
+#define CHROM_DERING    (DERING<<4)             //  64
+#define LUM_LEVEL_FIX   LEVEL_FIX               //   8
+#define CHROM_LEVEL_FIX (LEVEL_FIX<<4)          // 128 (not implemented yet)
 
 // Experimental vertical filters
-#define V_X1_FILTER	0x0200			// 512
-#define V_A_DEBLOCK	0x0400
+#define V_X1_FILTER     0x0200                  // 512
+#define V_A_DEBLOCK     0x0400
 
 // Experimental horizontal filters
-#define H_X1_FILTER	0x2000			// 8192
-#define H_A_DEBLOCK	0x4000
+#define H_X1_FILTER     0x2000                  // 8192
+#define H_A_DEBLOCK     0x4000
 
 /// select between full y range (255-0) or standart one (234-16)
-#define FULL_Y_RANGE	0x8000			// 32768
+#define FULL_Y_RANGE    0x8000                  // 32768
 
 //Deinterlacing Filters
-#define	LINEAR_IPOL_DEINT_FILTER	0x10000	// 65536
-#define	LINEAR_BLEND_DEINT_FILTER	0x20000	// 131072
-#define	CUBIC_BLEND_DEINT_FILTER	0x8000	// (not implemented yet)
-#define	CUBIC_IPOL_DEINT_FILTER		0x40000	// 262144
-#define	MEDIAN_DEINT_FILTER		0x80000	// 524288
-#define	FFMPEG_DEINT_FILTER		0x400000
-#define	LOWPASS5_DEINT_FILTER		0x800000
+#define        LINEAR_IPOL_DEINT_FILTER         0x10000 // 65536
+#define        LINEAR_BLEND_DEINT_FILTER        0x20000 // 131072
+#define        CUBIC_BLEND_DEINT_FILTER         0x8000  // (not implemented yet)
+#define        CUBIC_IPOL_DEINT_FILTER          0x40000 // 262144
+#define        MEDIAN_DEINT_FILTER              0x80000 // 524288
+#define        FFMPEG_DEINT_FILTER              0x400000
+#define        LOWPASS5_DEINT_FILTER            0x800000
 
-#define TEMP_NOISE_FILTER		0x100000
-#define FORCE_QUANT			0x200000
+#define TEMP_NOISE_FILTER               0x100000
+#define FORCE_QUANT                     0x200000
 
 //use if u want a faster postprocessing code
 //cant differentiate between chroma & luma filters (both on or both off)
@@ -66,8 +66,8 @@
 
 #if 1
 static inline int CLIP(int a){
-	if(a&256) return ((a)>>31)^(-1);
-	else      return a;
+        if(a&256) return ((a)>>31)^(-1);
+        else      return a;
 }
 //#define CLIP(a) (((a)&256) ? ((a)>>31)^(-1) : (a))
 #elif 0
@@ -79,92 +79,92 @@ static inline int CLIP(int a){
  * Postprocessng filter.
  */
 struct PPFilter{
-	char *shortName;
-	char *longName;
-	int chromDefault; 	///< is chrominance filtering on by default if this filter is manually activated
-	int minLumQuality; 	///< minimum quality to turn luminance filtering on
-	int minChromQuality;	///< minimum quality to turn chrominance filtering on
-	int mask; 		///< Bitmask to turn this filter on
+        char *shortName;
+        char *longName;
+        int chromDefault;       ///< is chrominance filtering on by default if this filter is manually activated
+        int minLumQuality;      ///< minimum quality to turn luminance filtering on
+        int minChromQuality;    ///< minimum quality to turn chrominance filtering on
+        int mask;               ///< Bitmask to turn this filter on
 };
 
 /**
  * Postprocessng mode.
  */
 typedef struct PPMode{
-	int lumMode; 			///< acivates filters for luminance
-	int chromMode; 			///< acivates filters for chrominance
-	int error; 			///< non zero on error
+        int lumMode;                    ///< acivates filters for luminance
+        int chromMode;                  ///< acivates filters for chrominance
+        int error;                      ///< non zero on error
 
-	int minAllowedY; 		///< for brigtness correction
-	int maxAllowedY; 		///< for brihtness correction
-	float maxClippedThreshold;	///< amount of "black" u r willing to loose to get a brightness corrected picture
+        int minAllowedY;                ///< for brigtness correction
+        int maxAllowedY;                ///< for brihtness correction
+        float maxClippedThreshold;      ///< amount of "black" u r willing to loose to get a brightness corrected picture
 
-	int maxTmpNoise[3]; 		///< for Temporal Noise Reducing filter (Maximal sum of abs differences)
+        int maxTmpNoise[3];             ///< for Temporal Noise Reducing filter (Maximal sum of abs differences)
 
-	int baseDcDiff;
-	int flatnessThreshold;
+        int baseDcDiff;
+        int flatnessThreshold;
 
-	int forcedQuant; 		///< quantizer if FORCE_QUANT is used
+        int forcedQuant;                ///< quantizer if FORCE_QUANT is used
 } PPMode;
 
 /**
  * postprocess context.
  */
 typedef struct PPContext{
-	uint8_t *tempBlocks; ///<used for the horizontal code
+        uint8_t *tempBlocks; ///<used for the horizontal code
 
-	/**
-	 * luma histogram.
-	 * we need 64bit here otherwise we'll going to have a problem
-	 * after watching a black picture for 5 hours
-	 */
-	uint64_t *yHistogram;
+        /**
+         * luma histogram.
+         * we need 64bit here otherwise we'll going to have a problem
+         * after watching a black picture for 5 hours
+         */
+        uint64_t *yHistogram;
 
-	uint64_t __attribute__((aligned(8))) packedYOffset;
-	uint64_t __attribute__((aligned(8))) packedYScale;
+        uint64_t __attribute__((aligned(8))) packedYOffset;
+        uint64_t __attribute__((aligned(8))) packedYScale;
 
-	/** Temporal noise reducing buffers */
-	uint8_t *tempBlured[3];
-	int32_t *tempBluredPast[3];
+        /** Temporal noise reducing buffers */
+        uint8_t *tempBlured[3];
+        int32_t *tempBluredPast[3];
 
-	/** Temporary buffers for handling the last row(s) */
-	uint8_t *tempDst;
-	uint8_t *tempSrc;
+        /** Temporary buffers for handling the last row(s) */
+        uint8_t *tempDst;
+        uint8_t *tempSrc;
 
-	uint8_t *deintTemp;
+        uint8_t *deintTemp;
 
-	uint64_t __attribute__((aligned(8))) pQPb;
-	uint64_t __attribute__((aligned(8))) pQPb2;
+        uint64_t __attribute__((aligned(8))) pQPb;
+        uint64_t __attribute__((aligned(8))) pQPb2;
 
-	uint64_t __attribute__((aligned(8))) mmxDcOffset[64];
-	uint64_t __attribute__((aligned(8))) mmxDcThreshold[64];
+        uint64_t __attribute__((aligned(8))) mmxDcOffset[64];
+        uint64_t __attribute__((aligned(8))) mmxDcThreshold[64];
 
-	QP_STORE_T *stdQPTable;       ///< used to fix MPEG2 style qscale
-	QP_STORE_T *nonBQPTable;
-	QP_STORE_T *forcedQPTable;
+        QP_STORE_T *stdQPTable;       ///< used to fix MPEG2 style qscale
+        QP_STORE_T *nonBQPTable;
+        QP_STORE_T *forcedQPTable;
 
-	int QP;
-	int nonBQP;
+        int QP;
+        int nonBQP;
 
-	int frameNum;
+        int frameNum;
 
-	int cpuCaps;
+        int cpuCaps;
 
-	int qpStride; ///<size of qp buffers (needed to realloc them if needed)
-	int stride;   ///<size of some buffers (needed to realloc them if needed)
+        int qpStride; ///<size of qp buffers (needed to realloc them if needed)
+        int stride;   ///<size of some buffers (needed to realloc them if needed)
 
-	int hChromaSubSample;
-	int vChromaSubSample;
+        int hChromaSubSample;
+        int vChromaSubSample;
 
-	PPMode ppMode;
+        PPMode ppMode;
 } PPContext;
 
 
 static inline void linecpy(void *dest, void *src, int lines, int stride)
 {
-	if (stride > 0) {
-		memcpy(dest, src, lines*stride);
-	} else {
-		memcpy(dest+(lines-1)*stride, src+(lines-1)*stride, -lines*stride);
-	}
+        if (stride > 0) {
+                memcpy(dest, src, lines*stride);
+        } else {
+                memcpy(dest+(lines-1)*stride, src+(lines-1)*stride, -lines*stride);
+        }
 }
diff --git a/libavcodec/libpostproc/postprocess_template.c b/libavcodec/libpostproc/postprocess_template.c
index 7f83cf7260..a60010b515 100644
--- a/libavcodec/libpostproc/postprocess_template.c
+++ b/libavcodec/libpostproc/postprocess_template.c
@@ -58,17 +58,17 @@
 #define PMINUB(a,b,t) "pminub " #a ", " #b " \n\t"
 #elif defined (HAVE_MMX)
 #define PMINUB(b,a,t) \
-	"movq " #a ", " #t " \n\t"\
-	"psubusb " #b ", " #t " \n\t"\
-	"psubb " #t ", " #a " \n\t"
+        "movq " #a ", " #t " \n\t"\
+        "psubusb " #b ", " #t " \n\t"\
+        "psubb " #t ", " #a " \n\t"
 #endif
 
 #ifdef HAVE_MMX2
 #define PMAXUB(a,b) "pmaxub " #a ", " #b " \n\t"
 #elif defined (HAVE_MMX)
 #define PMAXUB(a,b) \
-	"psubusb " #a ", " #b " \n\t"\
-	"paddb " #a ", " #b " \n\t"
+        "psubusb " #a ", " #b " \n\t"\
+        "paddb " #a ", " #b " \n\t"
 #endif
 
 //FIXME? |255-0| = 1 (shouldnt be a problem ...)
@@ -77,109 +77,109 @@
  * Check if the middle 8x8 Block in the given 8x16 block is flat
  */
 static inline int RENAME(vertClassify)(uint8_t src[], int stride, PPContext *c){
-	int numEq= 0, dcOk;
-	src+= stride*4; // src points to begin of the 8x8 Block
+        int numEq= 0, dcOk;
+        src+= stride*4; // src points to begin of the 8x8 Block
 asm volatile(
-		"movq %0, %%mm7					\n\t"
-		"movq %1, %%mm6					\n\t"
+                "movq %0, %%mm7                         \n\t"
+                "movq %1, %%mm6                         \n\t"
                 : : "m" (c->mmxDcOffset[c->nonBQP]),  "m" (c->mmxDcThreshold[c->nonBQP])
                 );
 
 asm volatile(
-		"lea (%2, %3), %%"REG_a"			\n\t"
-//	0	1	2	3	4	5	6	7	8	9
-//	%1	eax	eax+%2	eax+2%2	%1+4%2	ecx	ecx+%2	ecx+2%2	%1+8%2	ecx+4%2
-
-		"movq (%2), %%mm0				\n\t"
-		"movq (%%"REG_a"), %%mm1			\n\t"
-                "movq %%mm0, %%mm3				\n\t"
-                "movq %%mm0, %%mm4				\n\t"
+                "lea (%2, %3), %%"REG_a"                \n\t"
+//      0       1       2       3       4       5       6       7       8       9
+//      %1      eax     eax+%2  eax+2%2 %1+4%2  ecx     ecx+%2  ecx+2%2 %1+8%2  ecx+4%2
+
+                "movq (%2), %%mm0                       \n\t"
+                "movq (%%"REG_a"), %%mm1                \n\t"
+                "movq %%mm0, %%mm3                      \n\t"
+                "movq %%mm0, %%mm4                      \n\t"
                 PMAXUB(%%mm1, %%mm4)
                 PMINUB(%%mm1, %%mm3, %%mm5)
-		"psubb %%mm1, %%mm0				\n\t" // mm0 = differnece
-		"paddb %%mm7, %%mm0				\n\t"
-		"pcmpgtb %%mm6, %%mm0				\n\t"
+                "psubb %%mm1, %%mm0                     \n\t" // mm0 = differnece
+                "paddb %%mm7, %%mm0                     \n\t"
+                "pcmpgtb %%mm6, %%mm0                   \n\t"
 
-		"movq (%%"REG_a",%3), %%mm2			\n\t"
+                "movq (%%"REG_a",%3), %%mm2             \n\t"
                 PMAXUB(%%mm2, %%mm4)
                 PMINUB(%%mm2, %%mm3, %%mm5)
-		"psubb %%mm2, %%mm1				\n\t"
-		"paddb %%mm7, %%mm1				\n\t"
-		"pcmpgtb %%mm6, %%mm1				\n\t"
-		"paddb %%mm1, %%mm0				\n\t"
+                "psubb %%mm2, %%mm1                     \n\t"
+                "paddb %%mm7, %%mm1                     \n\t"
+                "pcmpgtb %%mm6, %%mm1                   \n\t"
+                "paddb %%mm1, %%mm0                     \n\t"
 
-		"movq (%%"REG_a", %3, 2), %%mm1			\n\t"
+                "movq (%%"REG_a", %3, 2), %%mm1         \n\t"
                 PMAXUB(%%mm1, %%mm4)
                 PMINUB(%%mm1, %%mm3, %%mm5)
-		"psubb %%mm1, %%mm2				\n\t"
-		"paddb %%mm7, %%mm2				\n\t"
-		"pcmpgtb %%mm6, %%mm2				\n\t"
-		"paddb %%mm2, %%mm0				\n\t"
+                "psubb %%mm1, %%mm2                     \n\t"
+                "paddb %%mm7, %%mm2                     \n\t"
+                "pcmpgtb %%mm6, %%mm2                   \n\t"
+                "paddb %%mm2, %%mm0                     \n\t"
 
-		"lea (%%"REG_a", %3, 4), %%"REG_a"		\n\t"
+                "lea (%%"REG_a", %3, 4), %%"REG_a"      \n\t"
 
-		"movq (%2, %3, 4), %%mm2			\n\t"
+                "movq (%2, %3, 4), %%mm2                \n\t"
                 PMAXUB(%%mm2, %%mm4)
                 PMINUB(%%mm2, %%mm3, %%mm5)
-		"psubb %%mm2, %%mm1				\n\t"
-		"paddb %%mm7, %%mm1				\n\t"
-		"pcmpgtb %%mm6, %%mm1				\n\t"
-		"paddb %%mm1, %%mm0				\n\t"
+                "psubb %%mm2, %%mm1                     \n\t"
+                "paddb %%mm7, %%mm1                     \n\t"
+                "pcmpgtb %%mm6, %%mm1                   \n\t"
+                "paddb %%mm1, %%mm0                     \n\t"
 
-		"movq (%%"REG_a"), %%mm1			\n\t"
+                "movq (%%"REG_a"), %%mm1                \n\t"
                 PMAXUB(%%mm1, %%mm4)
                 PMINUB(%%mm1, %%mm3, %%mm5)
-		"psubb %%mm1, %%mm2				\n\t"
-		"paddb %%mm7, %%mm2				\n\t"
-		"pcmpgtb %%mm6, %%mm2				\n\t"
-		"paddb %%mm2, %%mm0				\n\t"
+                "psubb %%mm1, %%mm2                     \n\t"
+                "paddb %%mm7, %%mm2                     \n\t"
+                "pcmpgtb %%mm6, %%mm2                   \n\t"
+                "paddb %%mm2, %%mm0                     \n\t"
 
-		"movq (%%"REG_a", %3), %%mm2			\n\t"
+                "movq (%%"REG_a", %3), %%mm2            \n\t"
                 PMAXUB(%%mm2, %%mm4)
                 PMINUB(%%mm2, %%mm3, %%mm5)
-		"psubb %%mm2, %%mm1				\n\t"
-		"paddb %%mm7, %%mm1				\n\t"
-		"pcmpgtb %%mm6, %%mm1				\n\t"
-		"paddb %%mm1, %%mm0				\n\t"
+                "psubb %%mm2, %%mm1                     \n\t"
+                "paddb %%mm7, %%mm1                     \n\t"
+                "pcmpgtb %%mm6, %%mm1                   \n\t"
+                "paddb %%mm1, %%mm0                     \n\t"
 
-		"movq (%%"REG_a", %3, 2), %%mm1			\n\t"
+                "movq (%%"REG_a", %3, 2), %%mm1         \n\t"
                 PMAXUB(%%mm1, %%mm4)
                 PMINUB(%%mm1, %%mm3, %%mm5)
-		"psubb %%mm1, %%mm2				\n\t"
-		"paddb %%mm7, %%mm2				\n\t"
-		"pcmpgtb %%mm6, %%mm2				\n\t"
-		"paddb %%mm2, %%mm0				\n\t"
-		"psubusb %%mm3, %%mm4				\n\t"
+                "psubb %%mm1, %%mm2                     \n\t"
+                "paddb %%mm7, %%mm2                     \n\t"
+                "pcmpgtb %%mm6, %%mm2                   \n\t"
+                "paddb %%mm2, %%mm0                     \n\t"
+                "psubusb %%mm3, %%mm4                   \n\t"
 
-		"						\n\t"
+                "                                       \n\t"
 #ifdef HAVE_MMX2
-		"pxor %%mm7, %%mm7				\n\t"
-		"psadbw %%mm7, %%mm0				\n\t"
+                "pxor %%mm7, %%mm7                      \n\t"
+                "psadbw %%mm7, %%mm0                    \n\t"
 #else
-		"movq %%mm0, %%mm1				\n\t"
-		"psrlw $8, %%mm0				\n\t"
-		"paddb %%mm1, %%mm0				\n\t"
-		"movq %%mm0, %%mm1				\n\t"
-		"psrlq $16, %%mm0				\n\t"
-		"paddb %%mm1, %%mm0				\n\t"
-		"movq %%mm0, %%mm1				\n\t"
-		"psrlq $32, %%mm0				\n\t"
-		"paddb %%mm1, %%mm0				\n\t"
+                "movq %%mm0, %%mm1                      \n\t"
+                "psrlw $8, %%mm0                        \n\t"
+                "paddb %%mm1, %%mm0                     \n\t"
+                "movq %%mm0, %%mm1                      \n\t"
+                "psrlq $16, %%mm0                       \n\t"
+                "paddb %%mm1, %%mm0                     \n\t"
+                "movq %%mm0, %%mm1                      \n\t"
+                "psrlq $32, %%mm0                       \n\t"
+                "paddb %%mm1, %%mm0                     \n\t"
 #endif
-                "movq %4, %%mm7					\n\t" // QP,..., QP
-		"paddusb %%mm7, %%mm7				\n\t" // 2QP ... 2QP
-		"psubusb %%mm7, %%mm4				\n\t" // Diff <= 2QP -> 0
-		"packssdw %%mm4, %%mm4				\n\t"
-		"movd %%mm0, %0					\n\t"
-		"movd %%mm4, %1					\n\t"
-
-		: "=r" (numEq), "=r" (dcOk)
-		: "r" (src), "r" ((long)stride), "m" (c->pQPb)
-		: "%"REG_a
-		);
-
-	numEq= (-numEq) &0xFF;
-	if(numEq > c->ppMode.flatnessThreshold){
+                "movq %4, %%mm7                         \n\t" // QP,..., QP
+                "paddusb %%mm7, %%mm7                   \n\t" // 2QP ... 2QP
+                "psubusb %%mm7, %%mm4                   \n\t" // Diff <= 2QP -> 0
+                "packssdw %%mm4, %%mm4                  \n\t"
+                "movd %%mm0, %0                         \n\t"
+                "movd %%mm4, %1                         \n\t"
+
+                : "=r" (numEq), "=r" (dcOk)
+                : "r" (src), "r" ((long)stride), "m" (c->pQPb)
+                : "%"REG_a
+                );
+
+        numEq= (-numEq) &0xFF;
+        if(numEq > c->ppMode.flatnessThreshold){
             if(dcOk) return 0;
             else     return 1;
         }else{
@@ -196,172 +196,172 @@ asm volatile(
 static inline void RENAME(doVertLowPass)(uint8_t *src, int stride, PPContext *c)
 {
 #if defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	src+= stride*3;
-	asm volatile(	//"movv %0 %1 %2\n\t"
-		"movq %2, %%mm0			\n\t"  // QP,..., QP
-		"pxor %%mm4, %%mm4				\n\t"
-
-		"movq (%0), %%mm6				\n\t"
-		"movq (%0, %1), %%mm5				\n\t"
-		"movq %%mm5, %%mm1				\n\t"
-		"movq %%mm6, %%mm2				\n\t"
-		"psubusb %%mm6, %%mm5				\n\t"
-		"psubusb %%mm1, %%mm2				\n\t"
-		"por %%mm5, %%mm2				\n\t" // ABS Diff of lines
-		"psubusb %%mm0, %%mm2				\n\t" // diff <= QP -> 0
-		"pcmpeqb %%mm4, %%mm2			\n\t" // diff <= QP -> FF
-
-		"pand %%mm2, %%mm6				\n\t"
-		"pandn %%mm1, %%mm2				\n\t"
-		"por %%mm2, %%mm6				\n\t"// First Line to Filter
-
-		"movq (%0, %1, 8), %%mm5			\n\t"
-		"lea (%0, %1, 4), %%"REG_a"			\n\t"
-		"lea (%0, %1, 8), %%"REG_c"			\n\t"
-		"sub %1, %%"REG_c"				\n\t"
-		"add %1, %0					\n\t" // %0 points to line 1 not 0
-		"movq (%0, %1, 8), %%mm7			\n\t"
-		"movq %%mm5, %%mm1				\n\t"
-		"movq %%mm7, %%mm2				\n\t"
-		"psubusb %%mm7, %%mm5				\n\t"
-		"psubusb %%mm1, %%mm2				\n\t"
-		"por %%mm5, %%mm2				\n\t" // ABS Diff of lines
-		"psubusb %%mm0, %%mm2				\n\t" // diff <= QP -> 0
-		"pcmpeqb %%mm4, %%mm2			\n\t" // diff <= QP -> FF
-
-		"pand %%mm2, %%mm7				\n\t"
-		"pandn %%mm1, %%mm2				\n\t"
-		"por %%mm2, %%mm7				\n\t" // First Line to Filter
-
-
-		// 	1	2	3	4	5	6	7	8
-		//	%0	%0+%1	%0+2%1	eax	%0+4%1	eax+2%1	ecx	eax+4%1
-		// 6 4 2 2 1 1
-		// 6 4 4 2
-		// 6 8 2
-
-		"movq (%0, %1), %%mm0				\n\t" //  1
-		"movq %%mm0, %%mm1				\n\t" //  1
-		PAVGB(%%mm6, %%mm0)				      //1 1	/2
-		PAVGB(%%mm6, %%mm0)				      //3 1	/4
-
-		"movq (%0, %1, 4), %%mm2			\n\t" //     1
-		"movq %%mm2, %%mm5				\n\t" //     1
-		PAVGB((%%REGa), %%mm2)				      //    11	/2
-		PAVGB((%0, %1, 2), %%mm2)			      //   211	/4
-		"movq %%mm2, %%mm3				\n\t" //   211	/4
-		"movq (%0), %%mm4				\n\t" // 1
-		PAVGB(%%mm4, %%mm3)				      // 4 211	/8
-		PAVGB(%%mm0, %%mm3)				      //642211	/16
-		"movq %%mm3, (%0)				\n\t" // X
-		// mm1=2 mm2=3(211) mm4=1 mm5=5 mm6=0 mm7=9
-		"movq %%mm1, %%mm0				\n\t" //  1
-		PAVGB(%%mm6, %%mm0)				      //1 1	/2
-		"movq %%mm4, %%mm3				\n\t" // 1
-		PAVGB((%0,%1,2), %%mm3)				      // 1 1	/2
-		PAVGB((%%REGa,%1,2), %%mm5)			      //     11	/2
-		PAVGB((%%REGa), %%mm5)				      //    211 /4
-		PAVGB(%%mm5, %%mm3)				      // 2 2211 /8
-		PAVGB(%%mm0, %%mm3)				      //4242211 /16
-		"movq %%mm3, (%0,%1)				\n\t" //  X
-		// mm1=2 mm2=3(211) mm4=1 mm5=4(211) mm6=0 mm7=9
-		PAVGB(%%mm4, %%mm6)				      //11	/2
-		"movq (%%"REG_c"), %%mm0			\n\t" //       1
-		PAVGB((%%REGa, %1, 2), %%mm0)			      //      11/2
-		"movq %%mm0, %%mm3				\n\t" //      11/2
-		PAVGB(%%mm1, %%mm0)				      //  2   11/4
-		PAVGB(%%mm6, %%mm0)				      //222   11/8
-		PAVGB(%%mm2, %%mm0)				      //22242211/16
-		"movq (%0, %1, 2), %%mm2			\n\t" //   1
-		"movq %%mm0, (%0, %1, 2)			\n\t" //   X
-		// mm1=2 mm2=3 mm3=6(11) mm4=1 mm5=4(211) mm6=0(11) mm7=9
-		"movq (%%"REG_a", %1, 4), %%mm0			\n\t" //        1
-		PAVGB((%%REGc), %%mm0)				      //       11	/2
-		PAVGB(%%mm0, %%mm6)				      //11     11	/4
-		PAVGB(%%mm1, %%mm4)				      // 11		/2
-		PAVGB(%%mm2, %%mm1)				      //  11		/2
-		PAVGB(%%mm1, %%mm6)				      //1122   11	/8
-		PAVGB(%%mm5, %%mm6)				      //112242211	/16
-		"movq (%%"REG_a"), %%mm5			\n\t" //    1
-		"movq %%mm6, (%%"REG_a")			\n\t" //    X
-		// mm0=7(11) mm1=2(11) mm2=3 mm3=6(11) mm4=1(11) mm5=4 mm7=9
-		"movq (%%"REG_a", %1, 4), %%mm6			\n\t" //        1
-		PAVGB(%%mm7, %%mm6)				      //        11	/2
-		PAVGB(%%mm4, %%mm6)				      // 11     11	/4
-		PAVGB(%%mm3, %%mm6)				      // 11   2211	/8
-		PAVGB(%%mm5, %%mm2)				      //   11		/2
-		"movq (%0, %1, 4), %%mm4			\n\t" //     1
-		PAVGB(%%mm4, %%mm2)				      //   112		/4
-		PAVGB(%%mm2, %%mm6)				      // 112242211	/16
-		"movq %%mm6, (%0, %1, 4)			\n\t" //     X
-		// mm0=7(11) mm1=2(11) mm2=3(112) mm3=6(11) mm4=5 mm5=4 mm7=9
-		PAVGB(%%mm7, %%mm1)				      //  11     2	/4
-		PAVGB(%%mm4, %%mm5)				      //    11		/2
-		PAVGB(%%mm5, %%mm0)				      //    11 11	/4
-		"movq (%%"REG_a", %1, 2), %%mm6			\n\t" //      1
-		PAVGB(%%mm6, %%mm1)				      //  11  4  2	/8
-		PAVGB(%%mm0, %%mm1)				      //  11224222	/16
-		"movq %%mm1, (%%"REG_a", %1, 2)			\n\t" //      X
-		// mm2=3(112) mm3=6(11) mm4=5 mm5=4(11) mm6=6 mm7=9
-		PAVGB((%%REGc), %%mm2)				      //   112 4	/8
-		"movq (%%"REG_a", %1, 4), %%mm0			\n\t" //        1
-		PAVGB(%%mm0, %%mm6)				      //      1 1	/2
-		PAVGB(%%mm7, %%mm6)				      //      1 12	/4
-		PAVGB(%%mm2, %%mm6)				      //   1122424	/4
-		"movq %%mm6, (%%"REG_c")			\n\t" //       X
-		// mm0=8 mm3=6(11) mm4=5 mm5=4(11) mm7=9
-		PAVGB(%%mm7, %%mm5)				      //    11   2	/4
-		PAVGB(%%mm7, %%mm5)				      //    11   6	/8
-
-		PAVGB(%%mm3, %%mm0)				      //      112	/4
-		PAVGB(%%mm0, %%mm5)				      //    112246	/16
-		"movq %%mm5, (%%"REG_a", %1, 4)			\n\t" //        X
-		"sub %1, %0					\n\t"
-
-		:
-		: "r" (src), "r" ((long)stride), "m" (c->pQPb)
-		: "%"REG_a, "%"REG_c
-	);
+        src+= stride*3;
+        asm volatile(        //"movv %0 %1 %2\n\t"
+                "movq %2, %%mm0                         \n\t"  // QP,..., QP
+                "pxor %%mm4, %%mm4                      \n\t"
+
+                "movq (%0), %%mm6                       \n\t"
+                "movq (%0, %1), %%mm5                   \n\t"
+                "movq %%mm5, %%mm1                      \n\t"
+                "movq %%mm6, %%mm2                      \n\t"
+                "psubusb %%mm6, %%mm5                   \n\t"
+                "psubusb %%mm1, %%mm2                   \n\t"
+                "por %%mm5, %%mm2                       \n\t" // ABS Diff of lines
+                "psubusb %%mm0, %%mm2                   \n\t" // diff <= QP -> 0
+                "pcmpeqb %%mm4, %%mm2                   \n\t" // diff <= QP -> FF
+
+                "pand %%mm2, %%mm6                      \n\t"
+                "pandn %%mm1, %%mm2                     \n\t"
+                "por %%mm2, %%mm6                       \n\t"// First Line to Filter
+
+                "movq (%0, %1, 8), %%mm5                \n\t"
+                "lea (%0, %1, 4), %%"REG_a"             \n\t"
+                "lea (%0, %1, 8), %%"REG_c"             \n\t"
+                "sub %1, %%"REG_c"                      \n\t"
+                "add %1, %0                             \n\t" // %0 points to line 1 not 0
+                "movq (%0, %1, 8), %%mm7                \n\t"
+                "movq %%mm5, %%mm1                      \n\t"
+                "movq %%mm7, %%mm2                      \n\t"
+                "psubusb %%mm7, %%mm5                   \n\t"
+                "psubusb %%mm1, %%mm2                   \n\t"
+                "por %%mm5, %%mm2                       \n\t" // ABS Diff of lines
+                "psubusb %%mm0, %%mm2                   \n\t" // diff <= QP -> 0
+                "pcmpeqb %%mm4, %%mm2                   \n\t" // diff <= QP -> FF
+
+                "pand %%mm2, %%mm7                      \n\t"
+                "pandn %%mm1, %%mm2                     \n\t"
+                "por %%mm2, %%mm7                       \n\t" // First Line to Filter
+
+
+                //      1       2       3       4       5       6       7       8
+                //      %0      %0+%1   %0+2%1  eax     %0+4%1  eax+2%1 ecx     eax+4%1
+                // 6 4 2 2 1 1
+                // 6 4 4 2
+                // 6 8 2
+
+                "movq (%0, %1), %%mm0                   \n\t" //  1
+                "movq %%mm0, %%mm1                      \n\t" //  1
+                PAVGB(%%mm6, %%mm0)                           //1 1        /2
+                PAVGB(%%mm6, %%mm0)                           //3 1        /4
+
+                "movq (%0, %1, 4), %%mm2                \n\t" //     1
+                "movq %%mm2, %%mm5                      \n\t" //     1
+                PAVGB((%%REGa), %%mm2)                        //    11        /2
+                PAVGB((%0, %1, 2), %%mm2)                     //   211        /4
+                "movq %%mm2, %%mm3                      \n\t" //   211        /4
+                "movq (%0), %%mm4                       \n\t" // 1
+                PAVGB(%%mm4, %%mm3)                           // 4 211        /8
+                PAVGB(%%mm0, %%mm3)                           //642211        /16
+                "movq %%mm3, (%0)                       \n\t" // X
+                // mm1=2 mm2=3(211) mm4=1 mm5=5 mm6=0 mm7=9
+                "movq %%mm1, %%mm0                      \n\t" //  1
+                PAVGB(%%mm6, %%mm0)                           //1 1        /2
+                "movq %%mm4, %%mm3                      \n\t" // 1
+                PAVGB((%0,%1,2), %%mm3)                       // 1 1        /2
+                PAVGB((%%REGa,%1,2), %%mm5)                   //     11        /2
+                PAVGB((%%REGa), %%mm5)                        //    211 /4
+                PAVGB(%%mm5, %%mm3)                           // 2 2211 /8
+                PAVGB(%%mm0, %%mm3)                           //4242211 /16
+                "movq %%mm3, (%0,%1)                    \n\t" //  X
+                // mm1=2 mm2=3(211) mm4=1 mm5=4(211) mm6=0 mm7=9
+                PAVGB(%%mm4, %%mm6)                                   //11        /2
+                "movq (%%"REG_c"), %%mm0                \n\t" //       1
+                PAVGB((%%REGa, %1, 2), %%mm0)                 //      11/2
+                "movq %%mm0, %%mm3                      \n\t" //      11/2
+                PAVGB(%%mm1, %%mm0)                           //  2   11/4
+                PAVGB(%%mm6, %%mm0)                           //222   11/8
+                PAVGB(%%mm2, %%mm0)                           //22242211/16
+                "movq (%0, %1, 2), %%mm2                \n\t" //   1
+                "movq %%mm0, (%0, %1, 2)                \n\t" //   X
+                // mm1=2 mm2=3 mm3=6(11) mm4=1 mm5=4(211) mm6=0(11) mm7=9
+                "movq (%%"REG_a", %1, 4), %%mm0         \n\t" //        1
+                PAVGB((%%REGc), %%mm0)                        //       11        /2
+                PAVGB(%%mm0, %%mm6)                           //11     11        /4
+                PAVGB(%%mm1, %%mm4)                           // 11                /2
+                PAVGB(%%mm2, %%mm1)                           //  11                /2
+                PAVGB(%%mm1, %%mm6)                           //1122   11        /8
+                PAVGB(%%mm5, %%mm6)                           //112242211        /16
+                "movq (%%"REG_a"), %%mm5                \n\t" //    1
+                "movq %%mm6, (%%"REG_a")                \n\t" //    X
+                // mm0=7(11) mm1=2(11) mm2=3 mm3=6(11) mm4=1(11) mm5=4 mm7=9
+                "movq (%%"REG_a", %1, 4), %%mm6         \n\t" //        1
+                PAVGB(%%mm7, %%mm6)                           //        11        /2
+                PAVGB(%%mm4, %%mm6)                           // 11     11        /4
+                PAVGB(%%mm3, %%mm6)                           // 11   2211        /8
+                PAVGB(%%mm5, %%mm2)                           //   11                /2
+                "movq (%0, %1, 4), %%mm4                \n\t" //     1
+                PAVGB(%%mm4, %%mm2)                           //   112                /4
+                PAVGB(%%mm2, %%mm6)                           // 112242211        /16
+                "movq %%mm6, (%0, %1, 4)                \n\t" //     X
+                // mm0=7(11) mm1=2(11) mm2=3(112) mm3=6(11) mm4=5 mm5=4 mm7=9
+                PAVGB(%%mm7, %%mm1)                           //  11     2        /4
+                PAVGB(%%mm4, %%mm5)                           //    11                /2
+                PAVGB(%%mm5, %%mm0)                           //    11 11        /4
+                "movq (%%"REG_a", %1, 2), %%mm6         \n\t" //      1
+                PAVGB(%%mm6, %%mm1)                           //  11  4  2        /8
+                PAVGB(%%mm0, %%mm1)                           //  11224222        /16
+                "movq %%mm1, (%%"REG_a", %1, 2)         \n\t" //      X
+                // mm2=3(112) mm3=6(11) mm4=5 mm5=4(11) mm6=6 mm7=9
+                PAVGB((%%REGc), %%mm2)                        //   112 4        /8
+                "movq (%%"REG_a", %1, 4), %%mm0         \n\t" //        1
+                PAVGB(%%mm0, %%mm6)                           //      1 1        /2
+                PAVGB(%%mm7, %%mm6)                           //      1 12        /4
+                PAVGB(%%mm2, %%mm6)                           //   1122424        /4
+                "movq %%mm6, (%%"REG_c")                \n\t" //       X
+                // mm0=8 mm3=6(11) mm4=5 mm5=4(11) mm7=9
+                PAVGB(%%mm7, %%mm5)                           //    11   2        /4
+                PAVGB(%%mm7, %%mm5)                           //    11   6        /8
+
+                PAVGB(%%mm3, %%mm0)                           //      112        /4
+                PAVGB(%%mm0, %%mm5)                           //    112246        /16
+                "movq %%mm5, (%%"REG_a", %1, 4)         \n\t" //        X
+                "sub %1, %0                             \n\t"
+
+                :
+                : "r" (src), "r" ((long)stride), "m" (c->pQPb)
+                : "%"REG_a, "%"REG_c
+        );
 #else //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	const int l1= stride;
-	const int l2= stride + l1;
-	const int l3= stride + l2;
-	const int l4= stride + l3;
-	const int l5= stride + l4;
-	const int l6= stride + l5;
-	const int l7= stride + l6;
-	const int l8= stride + l7;
-	const int l9= stride + l8;
-	int x;
-	src+= stride*3;
-	for(x=0; x<BLOCK_SIZE; x++)
-	{
-		const int first= ABS(src[0] - src[l1]) < c->QP ? src[0] : src[l1];
-		const int last= ABS(src[l8] - src[l9]) < c->QP ? src[l9] : src[l8];
-
-		int sums[10];
-		sums[0] = 4*first + src[l1] + src[l2] + src[l3] + 4;
-		sums[1] = sums[0] - first  + src[l4];
-		sums[2] = sums[1] - first  + src[l5];
-		sums[3] = sums[2] - first  + src[l6];
-		sums[4] = sums[3] - first  + src[l7];
-		sums[5] = sums[4] - src[l1] + src[l8];
-		sums[6] = sums[5] - src[l2] + last;
-		sums[7] = sums[6] - src[l3] + last;
-		sums[8] = sums[7] - src[l4] + last;
-		sums[9] = sums[8] - src[l5] + last;
-
-		src[l1]= (sums[0] + sums[2] + 2*src[l1])>>4;
-		src[l2]= (sums[1] + sums[3] + 2*src[l2])>>4;
-		src[l3]= (sums[2] + sums[4] + 2*src[l3])>>4;
-		src[l4]= (sums[3] + sums[5] + 2*src[l4])>>4;
-		src[l5]= (sums[4] + sums[6] + 2*src[l5])>>4;
-		src[l6]= (sums[5] + sums[7] + 2*src[l6])>>4;
-		src[l7]= (sums[6] + sums[8] + 2*src[l7])>>4;
-		src[l8]= (sums[7] + sums[9] + 2*src[l8])>>4;
-
-		src++;
-	}
+        const int l1= stride;
+        const int l2= stride + l1;
+        const int l3= stride + l2;
+        const int l4= stride + l3;
+        const int l5= stride + l4;
+        const int l6= stride + l5;
+        const int l7= stride + l6;
+        const int l8= stride + l7;
+        const int l9= stride + l8;
+        int x;
+        src+= stride*3;
+        for(x=0; x<BLOCK_SIZE; x++)
+        {
+                const int first= ABS(src[0] - src[l1]) < c->QP ? src[0] : src[l1];
+                const int last= ABS(src[l8] - src[l9]) < c->QP ? src[l9] : src[l8];
+
+                int sums[10];
+                sums[0] = 4*first + src[l1] + src[l2] + src[l3] + 4;
+                sums[1] = sums[0] - first  + src[l4];
+                sums[2] = sums[1] - first  + src[l5];
+                sums[3] = sums[2] - first  + src[l6];
+                sums[4] = sums[3] - first  + src[l7];
+                sums[5] = sums[4] - src[l1] + src[l8];
+                sums[6] = sums[5] - src[l2] + last;
+                sums[7] = sums[6] - src[l3] + last;
+                sums[8] = sums[7] - src[l4] + last;
+                sums[9] = sums[8] - src[l5] + last;
+
+                src[l1]= (sums[0] + sums[2] + 2*src[l1])>>4;
+                src[l2]= (sums[1] + sums[3] + 2*src[l2])>>4;
+                src[l3]= (sums[2] + sums[4] + 2*src[l3])>>4;
+                src[l4]= (sums[3] + sums[5] + 2*src[l4])>>4;
+                src[l5]= (sums[4] + sums[6] + 2*src[l5])>>4;
+                src[l6]= (sums[5] + sums[7] + 2*src[l6])>>4;
+                src[l7]= (sums[6] + sums[8] + 2*src[l7])>>4;
+                src[l8]= (sums[7] + sums[9] + 2*src[l8])>>4;
+
+                src++;
+        }
 #endif //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 }
 #endif //HAVE_ALTIVEC
@@ -372,101 +372,101 @@ static inline void RENAME(doVertLowPass)(uint8_t *src, int stride, PPContext *c)
  * values are correctly clipped (MMX2)
  * values are wraparound (C)
  * conclusion: its fast, but introduces ugly horizontal patterns if there is a continious gradient
-	0 8 16 24
-	x = 8
-	x/2 = 4
-	x/8 = 1
-	1 12 12 23
+        0 8 16 24
+        x = 8
+        x/2 = 4
+        x/8 = 1
+        1 12 12 23
  */
 static inline void RENAME(vertRK1Filter)(uint8_t *src, int stride, int QP)
 {
 #if defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	src+= stride*3;
+        src+= stride*3;
 // FIXME rounding
-	asm volatile(
-		"pxor %%mm7, %%mm7				\n\t" // 0
-		"movq "MANGLE(b80)", %%mm6			\n\t" // MIN_SIGNED_BYTE
-		"leal (%0, %1), %%"REG_a"			\n\t"
-		"leal (%%"REG_a", %1, 4), %%"REG_c"		\n\t"
-//	0	1	2	3	4	5	6	7	8	9
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	ecx	ecx+%1	ecx+2%1	%0+8%1	ecx+4%1
-		"movq "MANGLE(pQPb)", %%mm0			\n\t" // QP,..., QP
-		"movq %%mm0, %%mm1				\n\t" // QP,..., QP
-		"paddusb "MANGLE(b02)", %%mm0			\n\t"
-		"psrlw $2, %%mm0				\n\t"
-		"pand "MANGLE(b3F)", %%mm0			\n\t" // QP/4,..., QP/4
-		"paddusb %%mm1, %%mm0				\n\t" // QP*1.25 ...
-		"movq (%0, %1, 4), %%mm2			\n\t" // line 4
-		"movq (%%"REG_c"), %%mm3				\n\t" // line 5
-		"movq %%mm2, %%mm4				\n\t" // line 4
-		"pcmpeqb %%mm5, %%mm5				\n\t" // -1
-		"pxor %%mm2, %%mm5				\n\t" // -line 4 - 1
-		PAVGB(%%mm3, %%mm5)
-		"paddb %%mm6, %%mm5				\n\t" // (l5-l4)/2
-		"psubusb %%mm3, %%mm4				\n\t"
-		"psubusb %%mm2, %%mm3				\n\t"
-		"por %%mm3, %%mm4				\n\t" // |l4 - l5|
-		"psubusb %%mm0, %%mm4				\n\t"
-		"pcmpeqb %%mm7, %%mm4				\n\t"
-		"pand %%mm4, %%mm5				\n\t" // d/2
-
-//		"paddb %%mm6, %%mm2				\n\t" // line 4 + 0x80
-		"paddb %%mm5, %%mm2				\n\t"
-//		"psubb %%mm6, %%mm2				\n\t"
-		"movq %%mm2, (%0,%1, 4)				\n\t"
-
-		"movq (%%"REG_c"), %%mm2				\n\t"
-//		"paddb %%mm6, %%mm2				\n\t" // line 5 + 0x80
-		"psubb %%mm5, %%mm2				\n\t"
-//		"psubb %%mm6, %%mm2				\n\t"
-		"movq %%mm2, (%%"REG_c")				\n\t"
-
-		"paddb %%mm6, %%mm5				\n\t"
-		"psrlw $2, %%mm5				\n\t"
-		"pand "MANGLE(b3F)", %%mm5			\n\t"
-		"psubb "MANGLE(b20)", %%mm5			\n\t" // (l5-l4)/8
-
-		"movq (%%"REG_a", %1, 2), %%mm2			\n\t"
-		"paddb %%mm6, %%mm2				\n\t" // line 3 + 0x80
-		"paddsb %%mm5, %%mm2				\n\t"
-		"psubb %%mm6, %%mm2				\n\t"
-		"movq %%mm2, (%%"REG_a", %1, 2)			\n\t"
-
-		"movq (%%"REG_c", %1), %%mm2			\n\t"
-		"paddb %%mm6, %%mm2				\n\t" // line 6 + 0x80
-		"psubsb %%mm5, %%mm2				\n\t"
-		"psubb %%mm6, %%mm2				\n\t"
-		"movq %%mm2, (%%"REG_c", %1)			\n\t"
-
-		:
-		: "r" (src), "r" ((long)stride)
-		: "%"REG_a, "%"REG_c
-	);
+        asm volatile(
+                "pxor %%mm7, %%mm7                      \n\t" // 0
+                "movq "MANGLE(b80)", %%mm6              \n\t" // MIN_SIGNED_BYTE
+                "leal (%0, %1), %%"REG_a"               \n\t"
+                "leal (%%"REG_a", %1, 4), %%"REG_c"     \n\t"
+//      0       1       2       3       4       5       6       7       8       9
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  ecx     ecx+%1  ecx+2%1 %0+8%1  ecx+4%1
+                "movq "MANGLE(pQPb)", %%mm0             \n\t" // QP,..., QP
+                "movq %%mm0, %%mm1                      \n\t" // QP,..., QP
+                "paddusb "MANGLE(b02)", %%mm0           \n\t"
+                "psrlw $2, %%mm0                        \n\t"
+                "pand "MANGLE(b3F)", %%mm0              \n\t" // QP/4,..., QP/4
+                "paddusb %%mm1, %%mm0                   \n\t" // QP*1.25 ...
+                "movq (%0, %1, 4), %%mm2                \n\t" // line 4
+                "movq (%%"REG_c"), %%mm3                \n\t" // line 5
+                "movq %%mm2, %%mm4                      \n\t" // line 4
+                "pcmpeqb %%mm5, %%mm5                   \n\t" // -1
+                "pxor %%mm2, %%mm5                      \n\t" // -line 4 - 1
+                PAVGB(%%mm3, %%mm5)
+                "paddb %%mm6, %%mm5                     \n\t" // (l5-l4)/2
+                "psubusb %%mm3, %%mm4                   \n\t"
+                "psubusb %%mm2, %%mm3                   \n\t"
+                "por %%mm3, %%mm4                       \n\t" // |l4 - l5|
+                "psubusb %%mm0, %%mm4                   \n\t"
+                "pcmpeqb %%mm7, %%mm4                   \n\t"
+                "pand %%mm4, %%mm5                      \n\t" // d/2
+
+//                "paddb %%mm6, %%mm2                     \n\t" // line 4 + 0x80
+                "paddb %%mm5, %%mm2                     \n\t"
+//                "psubb %%mm6, %%mm2                     \n\t"
+                "movq %%mm2, (%0,%1, 4)                 \n\t"
+
+                "movq (%%"REG_c"), %%mm2                \n\t"
+//                "paddb %%mm6, %%mm2                     \n\t" // line 5 + 0x80
+                "psubb %%mm5, %%mm2                     \n\t"
+//                "psubb %%mm6, %%mm2                     \n\t"
+                "movq %%mm2, (%%"REG_c")                \n\t"
+
+                "paddb %%mm6, %%mm5                     \n\t"
+                "psrlw $2, %%mm5                        \n\t"
+                "pand "MANGLE(b3F)", %%mm5              \n\t"
+                "psubb "MANGLE(b20)", %%mm5             \n\t" // (l5-l4)/8
+
+                "movq (%%"REG_a", %1, 2), %%mm2         \n\t"
+                "paddb %%mm6, %%mm2                     \n\t" // line 3 + 0x80
+                "paddsb %%mm5, %%mm2                    \n\t"
+                "psubb %%mm6, %%mm2                     \n\t"
+                "movq %%mm2, (%%"REG_a", %1, 2)         \n\t"
+
+                "movq (%%"REG_c", %1), %%mm2            \n\t"
+                "paddb %%mm6, %%mm2                     \n\t" // line 6 + 0x80
+                "psubsb %%mm5, %%mm2                    \n\t"
+                "psubb %%mm6, %%mm2                     \n\t"
+                "movq %%mm2, (%%"REG_c", %1)            \n\t"
+
+                :
+                : "r" (src), "r" ((long)stride)
+                : "%"REG_a, "%"REG_c
+        );
 #else //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
- 	const int l1= stride;
-	const int l2= stride + l1;
-	const int l3= stride + l2;
-	const int l4= stride + l3;
-	const int l5= stride + l4;
-	const int l6= stride + l5;
-//	const int l7= stride + l6;
-//	const int l8= stride + l7;
-//	const int l9= stride + l8;
-	int x;
-	const int QP15= QP + (QP>>2);
-	src+= stride*3;
-	for(x=0; x<BLOCK_SIZE; x++)
-	{
-		const int v = (src[x+l5] - src[x+l4]);
-		if(ABS(v) < QP15)
-		{
-			src[x+l3] +=v>>3;
-			src[x+l4] +=v>>1;
-			src[x+l5] -=v>>1;
-			src[x+l6] -=v>>3;
-
-		}
-	}
+         const int l1= stride;
+        const int l2= stride + l1;
+        const int l3= stride + l2;
+        const int l4= stride + l3;
+        const int l5= stride + l4;
+        const int l6= stride + l5;
+//        const int l7= stride + l6;
+//        const int l8= stride + l7;
+//        const int l9= stride + l8;
+        int x;
+        const int QP15= QP + (QP>>2);
+        src+= stride*3;
+        for(x=0; x<BLOCK_SIZE; x++)
+        {
+                const int v = (src[x+l5] - src[x+l4]);
+                if(ABS(v) < QP15)
+                {
+                        src[x+l3] +=v>>3;
+                        src[x+l4] +=v>>1;
+                        src[x+l5] -=v>>1;
+                        src[x+l6] -=v>>3;
+
+                }
+        }
 
 #endif //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 }
@@ -482,128 +482,128 @@ static inline void RENAME(vertRK1Filter)(uint8_t *src, int stride, int QP)
 static inline void RENAME(vertX1Filter)(uint8_t *src, int stride, PPContext *co)
 {
 #if defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	src+= stride*3;
-
-	asm volatile(
-		"pxor %%mm7, %%mm7				\n\t" // 0
-		"lea (%0, %1), %%"REG_a"			\n\t"
-		"lea (%%"REG_a", %1, 4), %%"REG_c"		\n\t"
-//	0	1	2	3	4	5	6	7	8	9
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	ecx	ecx+%1	ecx+2%1	%0+8%1	ecx+4%1
-		"movq (%%"REG_a", %1, 2), %%mm0			\n\t" // line 3
-		"movq (%0, %1, 4), %%mm1			\n\t" // line 4
-		"movq %%mm1, %%mm2				\n\t" // line 4
-		"psubusb %%mm0, %%mm1				\n\t"
-		"psubusb %%mm2, %%mm0				\n\t"
-		"por %%mm1, %%mm0				\n\t" // |l2 - l3|
-		"movq (%%"REG_c"), %%mm3				\n\t" // line 5
-		"movq (%%"REG_c", %1), %%mm4			\n\t" // line 6
-		"movq %%mm3, %%mm5				\n\t" // line 5
-		"psubusb %%mm4, %%mm3				\n\t"
-		"psubusb %%mm5, %%mm4				\n\t"
-		"por %%mm4, %%mm3				\n\t" // |l5 - l6|
-		PAVGB(%%mm3, %%mm0)				      // (|l2 - l3| + |l5 - l6|)/2
-		"movq %%mm2, %%mm1				\n\t" // line 4
-		"psubusb %%mm5, %%mm2				\n\t"
-		"movq %%mm2, %%mm4				\n\t"
-		"pcmpeqb %%mm7, %%mm2				\n\t" // (l4 - l5) <= 0 ? -1 : 0
-		"psubusb %%mm1, %%mm5				\n\t"
-		"por %%mm5, %%mm4				\n\t" // |l4 - l5|
-		"psubusb %%mm0, %%mm4		\n\t" //d = MAX(0, |l4-l5| - (|l2-l3| + |l5-l6|)/2)
-		"movq %%mm4, %%mm3				\n\t" // d
-		"movq %2, %%mm0			\n\t"
-                "paddusb %%mm0, %%mm0				\n\t"
-		"psubusb %%mm0, %%mm4				\n\t"
-		"pcmpeqb %%mm7, %%mm4				\n\t" // d <= QP ? -1 : 0
-		"psubusb "MANGLE(b01)", %%mm3			\n\t"
-		"pand %%mm4, %%mm3				\n\t" // d <= QP ? d : 0
-
-		PAVGB(%%mm7, %%mm3)				      // d/2
-		"movq %%mm3, %%mm1				\n\t" // d/2
-		PAVGB(%%mm7, %%mm3)				      // d/4
-		PAVGB(%%mm1, %%mm3)				      // 3*d/8
-
-		"movq (%0, %1, 4), %%mm0			\n\t" // line 4
-		"pxor %%mm2, %%mm0				\n\t" //(l4 - l5) <= 0 ? -l4-1 : l4
-		"psubusb %%mm3, %%mm0				\n\t"
-		"pxor %%mm2, %%mm0				\n\t"
-		"movq %%mm0, (%0, %1, 4)			\n\t" // line 4
-
-		"movq (%%"REG_c"), %%mm0			\n\t" // line 5
-		"pxor %%mm2, %%mm0				\n\t" //(l4 - l5) <= 0 ? -l5-1 : l5
-		"paddusb %%mm3, %%mm0				\n\t"
-		"pxor %%mm2, %%mm0				\n\t"
-		"movq %%mm0, (%%"REG_c")			\n\t" // line 5
-
-		PAVGB(%%mm7, %%mm1)				      // d/4
-
-		"movq (%%"REG_a", %1, 2), %%mm0			\n\t" // line 3
-		"pxor %%mm2, %%mm0				\n\t" //(l4 - l5) <= 0 ? -l4-1 : l4
-		"psubusb %%mm1, %%mm0				\n\t"
-		"pxor %%mm2, %%mm0				\n\t"
-		"movq %%mm0, (%%"REG_a", %1, 2)			\n\t" // line 3
-
-		"movq (%%"REG_c", %1), %%mm0			\n\t" // line 6
-		"pxor %%mm2, %%mm0				\n\t" //(l4 - l5) <= 0 ? -l5-1 : l5
-		"paddusb %%mm1, %%mm0				\n\t"
-		"pxor %%mm2, %%mm0				\n\t"
-		"movq %%mm0, (%%"REG_c", %1)			\n\t" // line 6
-
-		PAVGB(%%mm7, %%mm1)				      // d/8
-
-		"movq (%%"REG_a", %1), %%mm0			\n\t" // line 2
-		"pxor %%mm2, %%mm0				\n\t" //(l4 - l5) <= 0 ? -l2-1 : l2
-		"psubusb %%mm1, %%mm0				\n\t"
-		"pxor %%mm2, %%mm0				\n\t"
-		"movq %%mm0, (%%"REG_a", %1)			\n\t" // line 2
-
-		"movq (%%"REG_c", %1, 2), %%mm0			\n\t" // line 7
-		"pxor %%mm2, %%mm0				\n\t" //(l4 - l5) <= 0 ? -l7-1 : l7
-		"paddusb %%mm1, %%mm0				\n\t"
-		"pxor %%mm2, %%mm0				\n\t"
-		"movq %%mm0, (%%"REG_c", %1, 2)			\n\t" // line 7
-
-		:
-		: "r" (src), "r" ((long)stride), "m" (co->pQPb)
-		: "%"REG_a, "%"REG_c
-	);
+        src+= stride*3;
+
+        asm volatile(
+                "pxor %%mm7, %%mm7                      \n\t" // 0
+                "lea (%0, %1), %%"REG_a"                \n\t"
+                "lea (%%"REG_a", %1, 4), %%"REG_c"      \n\t"
+//      0       1       2       3       4       5       6       7       8       9
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  ecx     ecx+%1  ecx+2%1 %0+8%1  ecx+4%1
+                "movq (%%"REG_a", %1, 2), %%mm0         \n\t" // line 3
+                "movq (%0, %1, 4), %%mm1                \n\t" // line 4
+                "movq %%mm1, %%mm2                      \n\t" // line 4
+                "psubusb %%mm0, %%mm1                   \n\t"
+                "psubusb %%mm2, %%mm0                   \n\t"
+                "por %%mm1, %%mm0                       \n\t" // |l2 - l3|
+                "movq (%%"REG_c"), %%mm3                \n\t" // line 5
+                "movq (%%"REG_c", %1), %%mm4            \n\t" // line 6
+                "movq %%mm3, %%mm5                      \n\t" // line 5
+                "psubusb %%mm4, %%mm3                   \n\t"
+                "psubusb %%mm5, %%mm4                   \n\t"
+                "por %%mm4, %%mm3                       \n\t" // |l5 - l6|
+                PAVGB(%%mm3, %%mm0)                           // (|l2 - l3| + |l5 - l6|)/2
+                "movq %%mm2, %%mm1                      \n\t" // line 4
+                "psubusb %%mm5, %%mm2                   \n\t"
+                "movq %%mm2, %%mm4                      \n\t"
+                "pcmpeqb %%mm7, %%mm2                   \n\t" // (l4 - l5) <= 0 ? -1 : 0
+                "psubusb %%mm1, %%mm5                   \n\t"
+                "por %%mm5, %%mm4                       \n\t" // |l4 - l5|
+                "psubusb %%mm0, %%mm4                   \n\t" //d = MAX(0, |l4-l5| - (|l2-l3| + |l5-l6|)/2)
+                "movq %%mm4, %%mm3                      \n\t" // d
+                "movq %2, %%mm0                         \n\t"
+                "paddusb %%mm0, %%mm0                   \n\t"
+                "psubusb %%mm0, %%mm4                   \n\t"
+                "pcmpeqb %%mm7, %%mm4                   \n\t" // d <= QP ? -1 : 0
+                "psubusb "MANGLE(b01)", %%mm3           \n\t"
+                "pand %%mm4, %%mm3                      \n\t" // d <= QP ? d : 0
+
+                PAVGB(%%mm7, %%mm3)                           // d/2
+                "movq %%mm3, %%mm1                      \n\t" // d/2
+                PAVGB(%%mm7, %%mm3)                           // d/4
+                PAVGB(%%mm1, %%mm3)                           // 3*d/8
+
+                "movq (%0, %1, 4), %%mm0                \n\t" // line 4
+                "pxor %%mm2, %%mm0                      \n\t" //(l4 - l5) <= 0 ? -l4-1 : l4
+                "psubusb %%mm3, %%mm0                   \n\t"
+                "pxor %%mm2, %%mm0                      \n\t"
+                "movq %%mm0, (%0, %1, 4)                \n\t" // line 4
+
+                "movq (%%"REG_c"), %%mm0                \n\t" // line 5
+                "pxor %%mm2, %%mm0                      \n\t" //(l4 - l5) <= 0 ? -l5-1 : l5
+                "paddusb %%mm3, %%mm0                   \n\t"
+                "pxor %%mm2, %%mm0                      \n\t"
+                "movq %%mm0, (%%"REG_c")                \n\t" // line 5
+
+                PAVGB(%%mm7, %%mm1)                           // d/4
+
+                "movq (%%"REG_a", %1, 2), %%mm0         \n\t" // line 3
+                "pxor %%mm2, %%mm0                      \n\t" //(l4 - l5) <= 0 ? -l4-1 : l4
+                "psubusb %%mm1, %%mm0                   \n\t"
+                "pxor %%mm2, %%mm0                      \n\t"
+                "movq %%mm0, (%%"REG_a", %1, 2)         \n\t" // line 3
+
+                "movq (%%"REG_c", %1), %%mm0            \n\t" // line 6
+                "pxor %%mm2, %%mm0                      \n\t" //(l4 - l5) <= 0 ? -l5-1 : l5
+                "paddusb %%mm1, %%mm0                   \n\t"
+                "pxor %%mm2, %%mm0                      \n\t"
+                "movq %%mm0, (%%"REG_c", %1)            \n\t" // line 6
+
+                PAVGB(%%mm7, %%mm1)                           // d/8
+
+                "movq (%%"REG_a", %1), %%mm0            \n\t" // line 2
+                "pxor %%mm2, %%mm0                      \n\t" //(l4 - l5) <= 0 ? -l2-1 : l2
+                "psubusb %%mm1, %%mm0                   \n\t"
+                "pxor %%mm2, %%mm0                      \n\t"
+                "movq %%mm0, (%%"REG_a", %1)            \n\t" // line 2
+
+                "movq (%%"REG_c", %1, 2), %%mm0         \n\t" // line 7
+                "pxor %%mm2, %%mm0                      \n\t" //(l4 - l5) <= 0 ? -l7-1 : l7
+                "paddusb %%mm1, %%mm0                   \n\t"
+                "pxor %%mm2, %%mm0                      \n\t"
+                "movq %%mm0, (%%"REG_c", %1, 2)         \n\t" // line 7
+
+                :
+                : "r" (src), "r" ((long)stride), "m" (co->pQPb)
+                : "%"REG_a, "%"REG_c
+        );
 #else //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 
- 	const int l1= stride;
-	const int l2= stride + l1;
-	const int l3= stride + l2;
-	const int l4= stride + l3;
-	const int l5= stride + l4;
-	const int l6= stride + l5;
-	const int l7= stride + l6;
-//	const int l8= stride + l7;
-//	const int l9= stride + l8;
-	int x;
-
-	src+= stride*3;
-	for(x=0; x<BLOCK_SIZE; x++)
-	{
-		int a= src[l3] - src[l4];
-		int b= src[l4] - src[l5];
-		int c= src[l5] - src[l6];
-
-		int d= ABS(b) - ((ABS(a) + ABS(c))>>1);
-		d= MAX(d, 0);
-
-		if(d < co->QP*2)
-		{
-			int v = d * SIGN(-b);
-
-			src[l2] +=v>>3;
-			src[l3] +=v>>2;
-			src[l4] +=(3*v)>>3;
-			src[l5] -=(3*v)>>3;
-			src[l6] -=v>>2;
-			src[l7] -=v>>3;
-
-		}
-		src++;
-	}
+        const int l1= stride;
+        const int l2= stride + l1;
+        const int l3= stride + l2;
+        const int l4= stride + l3;
+        const int l5= stride + l4;
+        const int l6= stride + l5;
+        const int l7= stride + l6;
+//        const int l8= stride + l7;
+//        const int l9= stride + l8;
+        int x;
+
+        src+= stride*3;
+        for(x=0; x<BLOCK_SIZE; x++)
+        {
+                int a= src[l3] - src[l4];
+                int b= src[l4] - src[l5];
+                int c= src[l5] - src[l6];
+
+                int d= ABS(b) - ((ABS(a) + ABS(c))>>1);
+                d= MAX(d, 0);
+
+                if(d < co->QP*2)
+                {
+                        int v = d * SIGN(-b);
+
+                        src[l2] +=v>>3;
+                        src[l3] +=v>>2;
+                        src[l4] +=(3*v)>>3;
+                        src[l5] -=(3*v)>>3;
+                        src[l6] -=v>>2;
+                        src[l7] -=v>>3;
+
+                }
+                src++;
+        }
 #endif //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 }
 
@@ -612,569 +612,569 @@ static inline void RENAME(doVertDefFilter)(uint8_t src[], int stride, PPContext
 {
 #if defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 /*
-	uint8_t tmp[16];
-	const int l1= stride;
-	const int l2= stride + l1;
-	const int l3= stride + l2;
-	const int l4= (int)tmp - (int)src - stride*3;
-	const int l5= (int)tmp - (int)src - stride*3 + 8;
-	const int l6= stride*3 + l3;
-	const int l7= stride + l6;
-	const int l8= stride + l7;
-
-	memcpy(tmp, src+stride*7, 8);
-	memcpy(tmp+8, src+stride*8, 8);
+        uint8_t tmp[16];
+        const int l1= stride;
+        const int l2= stride + l1;
+        const int l3= stride + l2;
+        const int l4= (int)tmp - (int)src - stride*3;
+        const int l5= (int)tmp - (int)src - stride*3 + 8;
+        const int l6= stride*3 + l3;
+        const int l7= stride + l6;
+        const int l8= stride + l7;
+
+        memcpy(tmp, src+stride*7, 8);
+        memcpy(tmp+8, src+stride*8, 8);
 */
-	src+= stride*4;
-	asm volatile(
+        src+= stride*4;
+        asm volatile(
 
 #if 0 //sligtly more accurate and slightly slower
-		"pxor %%mm7, %%mm7				\n\t" // 0
-		"lea (%0, %1), %%"REG_a"			\n\t"
-		"lea (%%"REG_a", %1, 4), %%"REG_c"		\n\t"
-//	0	1	2	3	4	5	6	7
-//	%0	%0+%1	%0+2%1	eax+2%1	%0+4%1	eax+4%1	ecx+%1	ecx+2%1
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	ecx	ecx+%1	ecx+2%1
-
-
-		"movq (%0, %1, 2), %%mm0			\n\t" // l2
-		"movq (%0), %%mm1				\n\t" // l0
-		"movq %%mm0, %%mm2				\n\t" // l2
-		PAVGB(%%mm7, %%mm0)				      // ~l2/2
-		PAVGB(%%mm1, %%mm0)				      // ~(l2 + 2l0)/4
-		PAVGB(%%mm2, %%mm0)				      // ~(5l2 + 2l0)/8
-
-		"movq (%%"REG_a"), %%mm1			\n\t" // l1
-		"movq (%%"REG_a", %1, 2), %%mm3			\n\t" // l3
-		"movq %%mm1, %%mm4				\n\t" // l1
-		PAVGB(%%mm7, %%mm1)				      // ~l1/2
-		PAVGB(%%mm3, %%mm1)				      // ~(l1 + 2l3)/4
-		PAVGB(%%mm4, %%mm1)				      // ~(5l1 + 2l3)/8
-
-		"movq %%mm0, %%mm4				\n\t" // ~(5l2 + 2l0)/8
-		"psubusb %%mm1, %%mm0				\n\t"
-		"psubusb %%mm4, %%mm1				\n\t"
-		"por %%mm0, %%mm1				\n\t" // ~|2l0 - 5l1 + 5l2 - 2l3|/8
+                "pxor %%mm7, %%mm7                      \n\t" // 0
+                "lea (%0, %1), %%"REG_a"                \n\t"
+                "lea (%%"REG_a", %1, 4), %%"REG_c"      \n\t"
+//      0       1       2       3       4       5       6       7
+//      %0      %0+%1   %0+2%1  eax+2%1 %0+4%1  eax+4%1 ecx+%1  ecx+2%1
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  ecx     ecx+%1  ecx+2%1
+
+
+                "movq (%0, %1, 2), %%mm0                \n\t" // l2
+                "movq (%0), %%mm1                       \n\t" // l0
+                "movq %%mm0, %%mm2                      \n\t" // l2
+                PAVGB(%%mm7, %%mm0)                           // ~l2/2
+                PAVGB(%%mm1, %%mm0)                           // ~(l2 + 2l0)/4
+                PAVGB(%%mm2, %%mm0)                           // ~(5l2 + 2l0)/8
+
+                "movq (%%"REG_a"), %%mm1                \n\t" // l1
+                "movq (%%"REG_a", %1, 2), %%mm3         \n\t" // l3
+                "movq %%mm1, %%mm4                      \n\t" // l1
+                PAVGB(%%mm7, %%mm1)                           // ~l1/2
+                PAVGB(%%mm3, %%mm1)                           // ~(l1 + 2l3)/4
+                PAVGB(%%mm4, %%mm1)                           // ~(5l1 + 2l3)/8
+
+                "movq %%mm0, %%mm4                      \n\t" // ~(5l2 + 2l0)/8
+                "psubusb %%mm1, %%mm0                   \n\t"
+                "psubusb %%mm4, %%mm1                   \n\t"
+                "por %%mm0, %%mm1                       \n\t" // ~|2l0 - 5l1 + 5l2 - 2l3|/8
 // mm1= |lenergy|, mm2= l2, mm3= l3, mm7=0
 
-		"movq (%0, %1, 4), %%mm0			\n\t" // l4
-		"movq %%mm0, %%mm4				\n\t" // l4
-		PAVGB(%%mm7, %%mm0)				      // ~l4/2
-		PAVGB(%%mm2, %%mm0)				      // ~(l4 + 2l2)/4
-		PAVGB(%%mm4, %%mm0)				      // ~(5l4 + 2l2)/8
-
-		"movq (%%"REG_c"), %%mm2			\n\t" // l5
-		"movq %%mm3, %%mm5				\n\t" // l3
-		PAVGB(%%mm7, %%mm3)				      // ~l3/2
-		PAVGB(%%mm2, %%mm3)				      // ~(l3 + 2l5)/4
-		PAVGB(%%mm5, %%mm3)				      // ~(5l3 + 2l5)/8
-
-		"movq %%mm0, %%mm6				\n\t" // ~(5l4 + 2l2)/8
-		"psubusb %%mm3, %%mm0				\n\t"
-		"psubusb %%mm6, %%mm3				\n\t"
-		"por %%mm0, %%mm3				\n\t" // ~|2l2 - 5l3 + 5l4 - 2l5|/8
-		"pcmpeqb %%mm7, %%mm0				\n\t" // SIGN(2l2 - 5l3 + 5l4 - 2l5)
+                "movq (%0, %1, 4), %%mm0                \n\t" // l4
+                "movq %%mm0, %%mm4                      \n\t" // l4
+                PAVGB(%%mm7, %%mm0)                           // ~l4/2
+                PAVGB(%%mm2, %%mm0)                           // ~(l4 + 2l2)/4
+                PAVGB(%%mm4, %%mm0)                           // ~(5l4 + 2l2)/8
+
+                "movq (%%"REG_c"), %%mm2                \n\t" // l5
+                "movq %%mm3, %%mm5                      \n\t" // l3
+                PAVGB(%%mm7, %%mm3)                           // ~l3/2
+                PAVGB(%%mm2, %%mm3)                           // ~(l3 + 2l5)/4
+                PAVGB(%%mm5, %%mm3)                           // ~(5l3 + 2l5)/8
+
+                "movq %%mm0, %%mm6                      \n\t" // ~(5l4 + 2l2)/8
+                "psubusb %%mm3, %%mm0                   \n\t"
+                "psubusb %%mm6, %%mm3                   \n\t"
+                "por %%mm0, %%mm3                       \n\t" // ~|2l2 - 5l3 + 5l4 - 2l5|/8
+                "pcmpeqb %%mm7, %%mm0                   \n\t" // SIGN(2l2 - 5l3 + 5l4 - 2l5)
 // mm0= SIGN(menergy), mm1= |lenergy|, mm2= l5, mm3= |menergy|, mm4=l4, mm5= l3, mm7=0
 
-		"movq (%%"REG_c", %1), %%mm6			\n\t" // l6
-		"movq %%mm6, %%mm5				\n\t" // l6
-		PAVGB(%%mm7, %%mm6)				      // ~l6/2
-		PAVGB(%%mm4, %%mm6)				      // ~(l6 + 2l4)/4
-		PAVGB(%%mm5, %%mm6)				      // ~(5l6 + 2l4)/8
-
-		"movq (%%"REG_c", %1, 2), %%mm5			\n\t" // l7
-		"movq %%mm2, %%mm4				\n\t" // l5
-		PAVGB(%%mm7, %%mm2)				      // ~l5/2
-		PAVGB(%%mm5, %%mm2)				      // ~(l5 + 2l7)/4
-		PAVGB(%%mm4, %%mm2)				      // ~(5l5 + 2l7)/8
-
-		"movq %%mm6, %%mm4				\n\t" // ~(5l6 + 2l4)/8
-		"psubusb %%mm2, %%mm6				\n\t"
-		"psubusb %%mm4, %%mm2				\n\t"
-		"por %%mm6, %%mm2				\n\t" // ~|2l4 - 5l5 + 5l6 - 2l7|/8
+                "movq (%%"REG_c", %1), %%mm6            \n\t" // l6
+                "movq %%mm6, %%mm5                      \n\t" // l6
+                PAVGB(%%mm7, %%mm6)                           // ~l6/2
+                PAVGB(%%mm4, %%mm6)                           // ~(l6 + 2l4)/4
+                PAVGB(%%mm5, %%mm6)                           // ~(5l6 + 2l4)/8
+
+                "movq (%%"REG_c", %1, 2), %%mm5         \n\t" // l7
+                "movq %%mm2, %%mm4                      \n\t" // l5
+                PAVGB(%%mm7, %%mm2)                           // ~l5/2
+                PAVGB(%%mm5, %%mm2)                           // ~(l5 + 2l7)/4
+                PAVGB(%%mm4, %%mm2)                           // ~(5l5 + 2l7)/8
+
+                "movq %%mm6, %%mm4                      \n\t" // ~(5l6 + 2l4)/8
+                "psubusb %%mm2, %%mm6                   \n\t"
+                "psubusb %%mm4, %%mm2                   \n\t"
+                "por %%mm6, %%mm2                       \n\t" // ~|2l4 - 5l5 + 5l6 - 2l7|/8
 // mm0= SIGN(menergy), mm1= |lenergy|/8, mm2= |renergy|/8, mm3= |menergy|/8, mm7=0
 
 
-		PMINUB(%%mm2, %%mm1, %%mm4)			      // MIN(|lenergy|,|renergy|)/8
-		"movq %2, %%mm4					\n\t" // QP //FIXME QP+1 ?
-		"paddusb "MANGLE(b01)", %%mm4			\n\t"
-		"pcmpgtb %%mm3, %%mm4				\n\t" // |menergy|/8 < QP
-		"psubusb %%mm1, %%mm3				\n\t" // d=|menergy|/8-MIN(|lenergy|,|renergy|)/8
-		"pand %%mm4, %%mm3				\n\t"
-
-		"movq %%mm3, %%mm1				\n\t"
-//		"psubusb "MANGLE(b01)", %%mm3			\n\t"
-		PAVGB(%%mm7, %%mm3)
-		PAVGB(%%mm7, %%mm3)
-		"paddusb %%mm1, %%mm3				\n\t"
-//		"paddusb "MANGLE(b01)", %%mm3			\n\t"
-
-		"movq (%%"REG_a", %1, 2), %%mm6			\n\t" //l3
-		"movq (%0, %1, 4), %%mm5			\n\t" //l4
-		"movq (%0, %1, 4), %%mm4			\n\t" //l4
-		"psubusb %%mm6, %%mm5				\n\t"
-		"psubusb %%mm4, %%mm6				\n\t"
-		"por %%mm6, %%mm5				\n\t" // |l3-l4|
-		"pcmpeqb %%mm7, %%mm6				\n\t" // SIGN(l3-l4)
-		"pxor %%mm6, %%mm0				\n\t"
-		"pand %%mm0, %%mm3				\n\t"
-		PMINUB(%%mm5, %%mm3, %%mm0)
-
-		"psubusb "MANGLE(b01)", %%mm3			\n\t"
-		PAVGB(%%mm7, %%mm3)
-
-		"movq (%%"REG_a", %1, 2), %%mm0			\n\t"
-		"movq (%0, %1, 4), %%mm2			\n\t"
-		"pxor %%mm6, %%mm0				\n\t"
-		"pxor %%mm6, %%mm2				\n\t"
-		"psubb %%mm3, %%mm0				\n\t"
-		"paddb %%mm3, %%mm2				\n\t"
-		"pxor %%mm6, %%mm0				\n\t"
-		"pxor %%mm6, %%mm2				\n\t"
-		"movq %%mm0, (%%"REG_a", %1, 2)			\n\t"
-		"movq %%mm2, (%0, %1, 4)			\n\t"
+                PMINUB(%%mm2, %%mm1, %%mm4)                   // MIN(|lenergy|,|renergy|)/8
+                "movq %2, %%mm4                         \n\t" // QP //FIXME QP+1 ?
+                "paddusb "MANGLE(b01)", %%mm4           \n\t"
+                "pcmpgtb %%mm3, %%mm4                   \n\t" // |menergy|/8 < QP
+                "psubusb %%mm1, %%mm3                   \n\t" // d=|menergy|/8-MIN(|lenergy|,|renergy|)/8
+                "pand %%mm4, %%mm3                      \n\t"
+
+                "movq %%mm3, %%mm1                      \n\t"
+//                "psubusb "MANGLE(b01)", %%mm3           \n\t"
+                PAVGB(%%mm7, %%mm3)
+                PAVGB(%%mm7, %%mm3)
+                "paddusb %%mm1, %%mm3                   \n\t"
+//                "paddusb "MANGLE(b01)", %%mm3           \n\t"
+
+                "movq (%%"REG_a", %1, 2), %%mm6         \n\t" //l3
+                "movq (%0, %1, 4), %%mm5                \n\t" //l4
+                "movq (%0, %1, 4), %%mm4                \n\t" //l4
+                "psubusb %%mm6, %%mm5                   \n\t"
+                "psubusb %%mm4, %%mm6                   \n\t"
+                "por %%mm6, %%mm5                       \n\t" // |l3-l4|
+                "pcmpeqb %%mm7, %%mm6                   \n\t" // SIGN(l3-l4)
+                "pxor %%mm6, %%mm0                      \n\t"
+                "pand %%mm0, %%mm3                      \n\t"
+                PMINUB(%%mm5, %%mm3, %%mm0)
+
+                "psubusb "MANGLE(b01)", %%mm3           \n\t"
+                PAVGB(%%mm7, %%mm3)
+
+                "movq (%%"REG_a", %1, 2), %%mm0         \n\t"
+                "movq (%0, %1, 4), %%mm2                \n\t"
+                "pxor %%mm6, %%mm0                      \n\t"
+                "pxor %%mm6, %%mm2                      \n\t"
+                "psubb %%mm3, %%mm0                     \n\t"
+                "paddb %%mm3, %%mm2                     \n\t"
+                "pxor %%mm6, %%mm0                      \n\t"
+                "pxor %%mm6, %%mm2                      \n\t"
+                "movq %%mm0, (%%"REG_a", %1, 2)         \n\t"
+                "movq %%mm2, (%0, %1, 4)                \n\t"
 #endif //0
 
-		"lea (%0, %1), %%"REG_a"			\n\t"
-		"pcmpeqb %%mm6, %%mm6				\n\t" // -1
-//	0	1	2	3	4	5	6	7
-//	%0	%0+%1	%0+2%1	eax+2%1	%0+4%1	eax+4%1	ecx+%1	ecx+2%1
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	ecx	ecx+%1	ecx+2%1
+                "lea (%0, %1), %%"REG_a"                \n\t"
+                "pcmpeqb %%mm6, %%mm6                   \n\t" // -1
+//      0       1       2       3       4       5       6       7
+//      %0      %0+%1   %0+2%1  eax+2%1 %0+4%1  eax+4%1 ecx+%1  ecx+2%1
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  ecx     ecx+%1  ecx+2%1
 
 
-		"movq (%%"REG_a", %1, 2), %%mm1			\n\t" // l3
-		"movq (%0, %1, 4), %%mm0			\n\t" // l4
-		"pxor %%mm6, %%mm1				\n\t" // -l3-1
-		PAVGB(%%mm1, %%mm0)				      // -q+128 = (l4-l3+256)/2
+                "movq (%%"REG_a", %1, 2), %%mm1         \n\t" // l3
+                "movq (%0, %1, 4), %%mm0                \n\t" // l4
+                "pxor %%mm6, %%mm1                      \n\t" // -l3-1
+                PAVGB(%%mm1, %%mm0)                           // -q+128 = (l4-l3+256)/2
 // mm1=-l3-1, mm0=128-q
 
-		"movq (%%"REG_a", %1, 4), %%mm2			\n\t" // l5
-		"movq (%%"REG_a", %1), %%mm3			\n\t" // l2
-		"pxor %%mm6, %%mm2				\n\t" // -l5-1
-		"movq %%mm2, %%mm5				\n\t" // -l5-1
-		"movq "MANGLE(b80)", %%mm4			\n\t" // 128
-		"lea (%%"REG_a", %1, 4), %%"REG_c"		\n\t"
-		PAVGB(%%mm3, %%mm2)				      // (l2-l5+256)/2
-		PAVGB(%%mm0, %%mm4)				      // ~(l4-l3)/4 + 128
-		PAVGB(%%mm2, %%mm4)				      // ~(l2-l5)/4 +(l4-l3)/8 + 128
-		PAVGB(%%mm0, %%mm4)				      // ~(l2-l5)/8 +5(l4-l3)/16 + 128
+                "movq (%%"REG_a", %1, 4), %%mm2         \n\t" // l5
+                "movq (%%"REG_a", %1), %%mm3            \n\t" // l2
+                "pxor %%mm6, %%mm2                      \n\t" // -l5-1
+                "movq %%mm2, %%mm5                      \n\t" // -l5-1
+                "movq "MANGLE(b80)", %%mm4              \n\t" // 128
+                "lea (%%"REG_a", %1, 4), %%"REG_c"      \n\t"
+                PAVGB(%%mm3, %%mm2)                           // (l2-l5+256)/2
+                PAVGB(%%mm0, %%mm4)                           // ~(l4-l3)/4 + 128
+                PAVGB(%%mm2, %%mm4)                           // ~(l2-l5)/4 +(l4-l3)/8 + 128
+                PAVGB(%%mm0, %%mm4)                           // ~(l2-l5)/8 +5(l4-l3)/16 + 128
 // mm1=-l3-1, mm0=128-q, mm3=l2, mm4=menergy/16 + 128, mm5= -l5-1
 
-		"movq (%%"REG_a"), %%mm2			\n\t" // l1
-		"pxor %%mm6, %%mm2				\n\t" // -l1-1
-		PAVGB(%%mm3, %%mm2)				      // (l2-l1+256)/2
-		PAVGB((%0), %%mm1)				      // (l0-l3+256)/2
-		"movq "MANGLE(b80)", %%mm3			\n\t" // 128
-		PAVGB(%%mm2, %%mm3)				      // ~(l2-l1)/4 + 128
-		PAVGB(%%mm1, %%mm3)				      // ~(l0-l3)/4 +(l2-l1)/8 + 128
-		PAVGB(%%mm2, %%mm3)				      // ~(l0-l3)/8 +5(l2-l1)/16 + 128
+                "movq (%%"REG_a"), %%mm2                \n\t" // l1
+                "pxor %%mm6, %%mm2                      \n\t" // -l1-1
+                PAVGB(%%mm3, %%mm2)                           // (l2-l1+256)/2
+                PAVGB((%0), %%mm1)                            // (l0-l3+256)/2
+                "movq "MANGLE(b80)", %%mm3              \n\t" // 128
+                PAVGB(%%mm2, %%mm3)                           // ~(l2-l1)/4 + 128
+                PAVGB(%%mm1, %%mm3)                           // ~(l0-l3)/4 +(l2-l1)/8 + 128
+                PAVGB(%%mm2, %%mm3)                           // ~(l0-l3)/8 +5(l2-l1)/16 + 128
 // mm0=128-q, mm3=lenergy/16 + 128, mm4= menergy/16 + 128, mm5= -l5-1
 
-		PAVGB((%%REGc, %1), %%mm5)			      // (l6-l5+256)/2
-		"movq (%%"REG_c", %1, 2), %%mm1			\n\t" // l7
-		"pxor %%mm6, %%mm1				\n\t" // -l7-1
-		PAVGB((%0, %1, 4), %%mm1)			      // (l4-l7+256)/2
-		"movq "MANGLE(b80)", %%mm2			\n\t" // 128
-		PAVGB(%%mm5, %%mm2)				      // ~(l6-l5)/4 + 128
-		PAVGB(%%mm1, %%mm2)				      // ~(l4-l7)/4 +(l6-l5)/8 + 128
-		PAVGB(%%mm5, %%mm2)				      // ~(l4-l7)/8 +5(l6-l5)/16 + 128
+                PAVGB((%%REGc, %1), %%mm5)                    // (l6-l5+256)/2
+                "movq (%%"REG_c", %1, 2), %%mm1         \n\t" // l7
+                "pxor %%mm6, %%mm1                      \n\t" // -l7-1
+                PAVGB((%0, %1, 4), %%mm1)                     // (l4-l7+256)/2
+                "movq "MANGLE(b80)", %%mm2              \n\t" // 128
+                PAVGB(%%mm5, %%mm2)                           // ~(l6-l5)/4 + 128
+                PAVGB(%%mm1, %%mm2)                           // ~(l4-l7)/4 +(l6-l5)/8 + 128
+                PAVGB(%%mm5, %%mm2)                           // ~(l4-l7)/8 +5(l6-l5)/16 + 128
 // mm0=128-q, mm2=renergy/16 + 128, mm3=lenergy/16 + 128, mm4= menergy/16 + 128
 
-		"movq "MANGLE(b00)", %%mm1			\n\t" // 0
-		"movq "MANGLE(b00)", %%mm5			\n\t" // 0
-		"psubb %%mm2, %%mm1				\n\t" // 128 - renergy/16
-		"psubb %%mm3, %%mm5				\n\t" // 128 - lenergy/16
-		PMAXUB(%%mm1, %%mm2)				      // 128 + |renergy/16|
- 		PMAXUB(%%mm5, %%mm3)				      // 128 + |lenergy/16|
-		PMINUB(%%mm2, %%mm3, %%mm1)			      // 128 + MIN(|lenergy|,|renergy|)/16
+                "movq "MANGLE(b00)", %%mm1              \n\t" // 0
+                "movq "MANGLE(b00)", %%mm5              \n\t" // 0
+                "psubb %%mm2, %%mm1                     \n\t" // 128 - renergy/16
+                "psubb %%mm3, %%mm5                     \n\t" // 128 - lenergy/16
+                PMAXUB(%%mm1, %%mm2)                          // 128 + |renergy/16|
+                 PMAXUB(%%mm5, %%mm3)                         // 128 + |lenergy/16|
+                PMINUB(%%mm2, %%mm3, %%mm1)                   // 128 + MIN(|lenergy|,|renergy|)/16
 
 // mm0=128-q, mm3=128 + MIN(|lenergy|,|renergy|)/16, mm4= menergy/16 + 128
 
-		"movq "MANGLE(b00)", %%mm7			\n\t" // 0
-		"movq %2, %%mm2					\n\t" // QP
-		PAVGB(%%mm6, %%mm2)				      // 128 + QP/2
-		"psubb %%mm6, %%mm2				\n\t"
-
-		"movq %%mm4, %%mm1				\n\t"
-		"pcmpgtb %%mm7, %%mm1				\n\t" // SIGN(menergy)
-		"pxor %%mm1, %%mm4				\n\t"
-		"psubb %%mm1, %%mm4				\n\t" // 128 + |menergy|/16
-		"pcmpgtb %%mm4, %%mm2				\n\t" // |menergy|/16 < QP/2
-		"psubusb %%mm3, %%mm4				\n\t" //d=|menergy|/16 - MIN(|lenergy|,|renergy|)/16
+                "movq "MANGLE(b00)", %%mm7              \n\t" // 0
+                "movq %2, %%mm2                         \n\t" // QP
+                PAVGB(%%mm6, %%mm2)                           // 128 + QP/2
+                "psubb %%mm6, %%mm2                     \n\t"
+
+                "movq %%mm4, %%mm1                      \n\t"
+                "pcmpgtb %%mm7, %%mm1                   \n\t" // SIGN(menergy)
+                "pxor %%mm1, %%mm4                      \n\t"
+                "psubb %%mm1, %%mm4                     \n\t" // 128 + |menergy|/16
+                "pcmpgtb %%mm4, %%mm2                   \n\t" // |menergy|/16 < QP/2
+                "psubusb %%mm3, %%mm4                   \n\t" //d=|menergy|/16 - MIN(|lenergy|,|renergy|)/16
 // mm0=128-q, mm1= SIGN(menergy), mm2= |menergy|/16 < QP/2, mm4= d/16
 
-		"movq %%mm4, %%mm3				\n\t" // d
-		"psubusb "MANGLE(b01)", %%mm4			\n\t"
-		PAVGB(%%mm7, %%mm4)				      // d/32
-		PAVGB(%%mm7, %%mm4)				      // (d + 32)/64
-		"paddb %%mm3, %%mm4				\n\t" // 5d/64
-		"pand %%mm2, %%mm4				\n\t"
-
-		"movq "MANGLE(b80)", %%mm5			\n\t" // 128
-		"psubb %%mm0, %%mm5				\n\t" // q
-		"paddsb %%mm6, %%mm5				\n\t" // fix bad rounding
-		"pcmpgtb %%mm5, %%mm7				\n\t" // SIGN(q)
-		"pxor %%mm7, %%mm5				\n\t"
-
-		PMINUB(%%mm5, %%mm4, %%mm3)			      // MIN(|q|, 5d/64)
-		"pxor %%mm1, %%mm7				\n\t" // SIGN(d*q)
-
-		"pand %%mm7, %%mm4				\n\t"
-		"movq (%%"REG_a", %1, 2), %%mm0			\n\t"
-		"movq (%0, %1, 4), %%mm2			\n\t"
-		"pxor %%mm1, %%mm0				\n\t"
-		"pxor %%mm1, %%mm2				\n\t"
-		"paddb %%mm4, %%mm0				\n\t"
-		"psubb %%mm4, %%mm2				\n\t"
-		"pxor %%mm1, %%mm0				\n\t"
-		"pxor %%mm1, %%mm2				\n\t"
-		"movq %%mm0, (%%"REG_a", %1, 2)			\n\t"
-		"movq %%mm2, (%0, %1, 4)			\n\t"
-
-		:
-		: "r" (src), "r" ((long)stride), "m" (c->pQPb)
-		: "%"REG_a, "%"REG_c
-	);
+                "movq %%mm4, %%mm3                      \n\t" // d
+                "psubusb "MANGLE(b01)", %%mm4           \n\t"
+                PAVGB(%%mm7, %%mm4)                           // d/32
+                PAVGB(%%mm7, %%mm4)                           // (d + 32)/64
+                "paddb %%mm3, %%mm4                     \n\t" // 5d/64
+                "pand %%mm2, %%mm4                      \n\t"
+
+                "movq "MANGLE(b80)", %%mm5              \n\t" // 128
+                "psubb %%mm0, %%mm5                     \n\t" // q
+                "paddsb %%mm6, %%mm5                    \n\t" // fix bad rounding
+                "pcmpgtb %%mm5, %%mm7                   \n\t" // SIGN(q)
+                "pxor %%mm7, %%mm5                      \n\t"
+
+                PMINUB(%%mm5, %%mm4, %%mm3)                   // MIN(|q|, 5d/64)
+                "pxor %%mm1, %%mm7                      \n\t" // SIGN(d*q)
+
+                "pand %%mm7, %%mm4                      \n\t"
+                "movq (%%"REG_a", %1, 2), %%mm0         \n\t"
+                "movq (%0, %1, 4), %%mm2                \n\t"
+                "pxor %%mm1, %%mm0                      \n\t"
+                "pxor %%mm1, %%mm2                      \n\t"
+                "paddb %%mm4, %%mm0                     \n\t"
+                "psubb %%mm4, %%mm2                     \n\t"
+                "pxor %%mm1, %%mm0                      \n\t"
+                "pxor %%mm1, %%mm2                      \n\t"
+                "movq %%mm0, (%%"REG_a", %1, 2)         \n\t"
+                "movq %%mm2, (%0, %1, 4)                \n\t"
+
+                :
+                : "r" (src), "r" ((long)stride), "m" (c->pQPb)
+                : "%"REG_a, "%"REG_c
+        );
 
 /*
-	{
-	int x;
-	src-= stride;
-	for(x=0; x<BLOCK_SIZE; x++)
-	{
-		const int middleEnergy= 5*(src[l5] - src[l4]) + 2*(src[l3] - src[l6]);
-		if(ABS(middleEnergy)< 8*QP)
-		{
-			const int q=(src[l4] - src[l5])/2;
-			const int leftEnergy=  5*(src[l3] - src[l2]) + 2*(src[l1] - src[l4]);
-			const int rightEnergy= 5*(src[l7] - src[l6]) + 2*(src[l5] - src[l8]);
-
-			int d= ABS(middleEnergy) - MIN( ABS(leftEnergy), ABS(rightEnergy) );
-			d= MAX(d, 0);
-
-			d= (5*d + 32) >> 6;
-			d*= SIGN(-middleEnergy);
-
-			if(q>0)
-			{
-				d= d<0 ? 0 : d;
-				d= d>q ? q : d;
-			}
-			else
-			{
-				d= d>0 ? 0 : d;
-				d= d<q ? q : d;
-			}
-
-        		src[l4]-= d;
-	        	src[l5]+= d;
-		}
-		src++;
-	}
+        {
+        int x;
+        src-= stride;
+        for(x=0; x<BLOCK_SIZE; x++)
+        {
+                const int middleEnergy= 5*(src[l5] - src[l4]) + 2*(src[l3] - src[l6]);
+                if(ABS(middleEnergy)< 8*QP)
+                {
+                        const int q=(src[l4] - src[l5])/2;
+                        const int leftEnergy=  5*(src[l3] - src[l2]) + 2*(src[l1] - src[l4]);
+                        const int rightEnergy= 5*(src[l7] - src[l6]) + 2*(src[l5] - src[l8]);
+
+                        int d= ABS(middleEnergy) - MIN( ABS(leftEnergy), ABS(rightEnergy) );
+                        d= MAX(d, 0);
+
+                        d= (5*d + 32) >> 6;
+                        d*= SIGN(-middleEnergy);
+
+                        if(q>0)
+                        {
+                                d= d<0 ? 0 : d;
+                                d= d>q ? q : d;
+                        }
+                        else
+                        {
+                                d= d>0 ? 0 : d;
+                                d= d<q ? q : d;
+                        }
+
+                        src[l4]-= d;
+                        src[l5]+= d;
+                }
+                src++;
+        }
 src-=8;
-	for(x=0; x<8; x++)
-	{
-		int y;
-		for(y=4; y<6; y++)
-		{
-			int d= src[x+y*stride] - tmp[x+(y-4)*8];
-			int ad= ABS(d);
-			static int max=0;
-			static int sum=0;
-			static int num=0;
-			static int bias=0;
-
-			if(max<ad) max=ad;
-			sum+= ad>3 ? 1 : 0;
-			if(ad>3)
-			{
-				src[0] = src[7] = src[stride*7] = src[(stride+1)*7]=255;
-			}
-			if(y==4) bias+=d;
-			num++;
-			if(num%1000000 == 0)
-			{
-				printf(" %d %d %d %d\n", num, sum, max, bias);
-			}
-		}
-	}
+        for(x=0; x<8; x++)
+        {
+                int y;
+                for(y=4; y<6; y++)
+                {
+                        int d= src[x+y*stride] - tmp[x+(y-4)*8];
+                        int ad= ABS(d);
+                        static int max=0;
+                        static int sum=0;
+                        static int num=0;
+                        static int bias=0;
+
+                        if(max<ad) max=ad;
+                        sum+= ad>3 ? 1 : 0;
+                        if(ad>3)
+                        {
+                                src[0] = src[7] = src[stride*7] = src[(stride+1)*7]=255;
+                        }
+                        if(y==4) bias+=d;
+                        num++;
+                        if(num%1000000 == 0)
+                        {
+                                printf(" %d %d %d %d\n", num, sum, max, bias);
+                        }
+                }
+        }
 }
 */
 #elif defined (HAVE_MMX)
-	src+= stride*4;
-	asm volatile(
-		"pxor %%mm7, %%mm7				\n\t"
-		"lea -40(%%"REG_SP"), %%"REG_c"			\n\t" // make space for 4 8-byte vars
-		"and "ALIGN_MASK", %%"REG_c"			\n\t" // align
-//	0	1	2	3	4	5	6	7
-//	%0	%0+%1	%0+2%1	eax+2%1	%0+4%1	eax+4%1	edx+%1	edx+2%1
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	edx	edx+%1	edx+2%1
-
-		"movq (%0), %%mm0				\n\t"
-		"movq %%mm0, %%mm1				\n\t"
-		"punpcklbw %%mm7, %%mm0				\n\t" // low part of line 0
-		"punpckhbw %%mm7, %%mm1				\n\t" // high part of line 0
-
-		"movq (%0, %1), %%mm2				\n\t"
-		"lea (%0, %1, 2), %%"REG_a"			\n\t"
-		"movq %%mm2, %%mm3				\n\t"
-		"punpcklbw %%mm7, %%mm2				\n\t" // low part of line 1
-		"punpckhbw %%mm7, %%mm3				\n\t" // high part of line 1
-
-		"movq (%%"REG_a"), %%mm4			\n\t"
-		"movq %%mm4, %%mm5				\n\t"
-		"punpcklbw %%mm7, %%mm4				\n\t" // low part of line 2
-		"punpckhbw %%mm7, %%mm5				\n\t" // high part of line 2
-
-		"paddw %%mm0, %%mm0				\n\t" // 2L0
-		"paddw %%mm1, %%mm1				\n\t" // 2H0
-		"psubw %%mm4, %%mm2				\n\t" // L1 - L2
-		"psubw %%mm5, %%mm3				\n\t" // H1 - H2
-		"psubw %%mm2, %%mm0				\n\t" // 2L0 - L1 + L2
-		"psubw %%mm3, %%mm1				\n\t" // 2H0 - H1 + H2
-
-		"psllw $2, %%mm2				\n\t" // 4L1 - 4L2
-		"psllw $2, %%mm3				\n\t" // 4H1 - 4H2
-		"psubw %%mm2, %%mm0				\n\t" // 2L0 - 5L1 + 5L2
-		"psubw %%mm3, %%mm1				\n\t" // 2H0 - 5H1 + 5H2
-
-		"movq (%%"REG_a", %1), %%mm2			\n\t"
-		"movq %%mm2, %%mm3				\n\t"
-		"punpcklbw %%mm7, %%mm2				\n\t" // L3
-		"punpckhbw %%mm7, %%mm3				\n\t" // H3
-
-		"psubw %%mm2, %%mm0				\n\t" // 2L0 - 5L1 + 5L2 - L3
-		"psubw %%mm3, %%mm1				\n\t" // 2H0 - 5H1 + 5H2 - H3
-		"psubw %%mm2, %%mm0				\n\t" // 2L0 - 5L1 + 5L2 - 2L3
-		"psubw %%mm3, %%mm1				\n\t" // 2H0 - 5H1 + 5H2 - 2H3
-		"movq %%mm0, (%%"REG_c")			\n\t" // 2L0 - 5L1 + 5L2 - 2L3
-		"movq %%mm1, 8(%%"REG_c")			\n\t" // 2H0 - 5H1 + 5H2 - 2H3
-
-		"movq (%%"REG_a", %1, 2), %%mm0			\n\t"
-		"movq %%mm0, %%mm1				\n\t"
-		"punpcklbw %%mm7, %%mm0				\n\t" // L4
-		"punpckhbw %%mm7, %%mm1				\n\t" // H4
-
-		"psubw %%mm0, %%mm2				\n\t" // L3 - L4
-		"psubw %%mm1, %%mm3				\n\t" // H3 - H4
-		"movq %%mm2, 16(%%"REG_c")			\n\t" // L3 - L4
-		"movq %%mm3, 24(%%"REG_c")			\n\t" // H3 - H4
-		"paddw %%mm4, %%mm4				\n\t" // 2L2
-		"paddw %%mm5, %%mm5				\n\t" // 2H2
-		"psubw %%mm2, %%mm4				\n\t" // 2L2 - L3 + L4
-		"psubw %%mm3, %%mm5				\n\t" // 2H2 - H3 + H4
-
-		"lea (%%"REG_a", %1), %0			\n\t"
-		"psllw $2, %%mm2				\n\t" // 4L3 - 4L4
-		"psllw $2, %%mm3				\n\t" // 4H3 - 4H4
-		"psubw %%mm2, %%mm4				\n\t" // 2L2 - 5L3 + 5L4
-		"psubw %%mm3, %%mm5				\n\t" // 2H2 - 5H3 + 5H4
+        src+= stride*4;
+        asm volatile(
+                "pxor %%mm7, %%mm7                      \n\t"
+                "lea -40(%%"REG_SP"), %%"REG_c"         \n\t" // make space for 4 8-byte vars
+                "and "ALIGN_MASK", %%"REG_c"            \n\t" // align
+//      0       1       2       3       4       5       6       7
+//      %0      %0+%1   %0+2%1  eax+2%1 %0+4%1  eax+4%1 edx+%1  edx+2%1
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  edx     edx+%1  edx+2%1
+
+                "movq (%0), %%mm0                       \n\t"
+                "movq %%mm0, %%mm1                      \n\t"
+                "punpcklbw %%mm7, %%mm0                 \n\t" // low part of line 0
+                "punpckhbw %%mm7, %%mm1                 \n\t" // high part of line 0
+
+                "movq (%0, %1), %%mm2                   \n\t"
+                "lea (%0, %1, 2), %%"REG_a"             \n\t"
+                "movq %%mm2, %%mm3                      \n\t"
+                "punpcklbw %%mm7, %%mm2                 \n\t" // low part of line 1
+                "punpckhbw %%mm7, %%mm3                 \n\t" // high part of line 1
+
+                "movq (%%"REG_a"), %%mm4                \n\t"
+                "movq %%mm4, %%mm5                      \n\t"
+                "punpcklbw %%mm7, %%mm4                 \n\t" // low part of line 2
+                "punpckhbw %%mm7, %%mm5                 \n\t" // high part of line 2
+
+                "paddw %%mm0, %%mm0                     \n\t" // 2L0
+                "paddw %%mm1, %%mm1                     \n\t" // 2H0
+                "psubw %%mm4, %%mm2                     \n\t" // L1 - L2
+                "psubw %%mm5, %%mm3                     \n\t" // H1 - H2
+                "psubw %%mm2, %%mm0                     \n\t" // 2L0 - L1 + L2
+                "psubw %%mm3, %%mm1                     \n\t" // 2H0 - H1 + H2
+
+                "psllw $2, %%mm2                        \n\t" // 4L1 - 4L2
+                "psllw $2, %%mm3                        \n\t" // 4H1 - 4H2
+                "psubw %%mm2, %%mm0                     \n\t" // 2L0 - 5L1 + 5L2
+                "psubw %%mm3, %%mm1                     \n\t" // 2H0 - 5H1 + 5H2
+
+                "movq (%%"REG_a", %1), %%mm2            \n\t"
+                "movq %%mm2, %%mm3                      \n\t"
+                "punpcklbw %%mm7, %%mm2                 \n\t" // L3
+                "punpckhbw %%mm7, %%mm3                 \n\t" // H3
+
+                "psubw %%mm2, %%mm0                     \n\t" // 2L0 - 5L1 + 5L2 - L3
+                "psubw %%mm3, %%mm1                     \n\t" // 2H0 - 5H1 + 5H2 - H3
+                "psubw %%mm2, %%mm0                     \n\t" // 2L0 - 5L1 + 5L2 - 2L3
+                "psubw %%mm3, %%mm1                     \n\t" // 2H0 - 5H1 + 5H2 - 2H3
+                "movq %%mm0, (%%"REG_c")                \n\t" // 2L0 - 5L1 + 5L2 - 2L3
+                "movq %%mm1, 8(%%"REG_c")               \n\t" // 2H0 - 5H1 + 5H2 - 2H3
+
+                "movq (%%"REG_a", %1, 2), %%mm0         \n\t"
+                "movq %%mm0, %%mm1                      \n\t"
+                "punpcklbw %%mm7, %%mm0                 \n\t" // L4
+                "punpckhbw %%mm7, %%mm1                 \n\t" // H4
+
+                "psubw %%mm0, %%mm2                     \n\t" // L3 - L4
+                "psubw %%mm1, %%mm3                     \n\t" // H3 - H4
+                "movq %%mm2, 16(%%"REG_c")              \n\t" // L3 - L4
+                "movq %%mm3, 24(%%"REG_c")              \n\t" // H3 - H4
+                "paddw %%mm4, %%mm4                     \n\t" // 2L2
+                "paddw %%mm5, %%mm5                     \n\t" // 2H2
+                "psubw %%mm2, %%mm4                     \n\t" // 2L2 - L3 + L4
+                "psubw %%mm3, %%mm5                     \n\t" // 2H2 - H3 + H4
+
+                "lea (%%"REG_a", %1), %0                \n\t"
+                "psllw $2, %%mm2                        \n\t" // 4L3 - 4L4
+                "psllw $2, %%mm3                        \n\t" // 4H3 - 4H4
+                "psubw %%mm2, %%mm4                     \n\t" // 2L2 - 5L3 + 5L4
+                "psubw %%mm3, %%mm5                     \n\t" // 2H2 - 5H3 + 5H4
 //50 opcodes so far
-		"movq (%0, %1, 2), %%mm2			\n\t"
-		"movq %%mm2, %%mm3				\n\t"
-		"punpcklbw %%mm7, %%mm2				\n\t" // L5
-		"punpckhbw %%mm7, %%mm3				\n\t" // H5
-		"psubw %%mm2, %%mm4				\n\t" // 2L2 - 5L3 + 5L4 - L5
-		"psubw %%mm3, %%mm5				\n\t" // 2H2 - 5H3 + 5H4 - H5
-		"psubw %%mm2, %%mm4				\n\t" // 2L2 - 5L3 + 5L4 - 2L5
-		"psubw %%mm3, %%mm5				\n\t" // 2H2 - 5H3 + 5H4 - 2H5
-
-		"movq (%%"REG_a", %1, 4), %%mm6			\n\t"
-		"punpcklbw %%mm7, %%mm6				\n\t" // L6
-		"psubw %%mm6, %%mm2				\n\t" // L5 - L6
-		"movq (%%"REG_a", %1, 4), %%mm6			\n\t"
-		"punpckhbw %%mm7, %%mm6				\n\t" // H6
-		"psubw %%mm6, %%mm3				\n\t" // H5 - H6
-
-		"paddw %%mm0, %%mm0				\n\t" // 2L4
-		"paddw %%mm1, %%mm1				\n\t" // 2H4
-		"psubw %%mm2, %%mm0				\n\t" // 2L4 - L5 + L6
-		"psubw %%mm3, %%mm1				\n\t" // 2H4 - H5 + H6
-
-		"psllw $2, %%mm2				\n\t" // 4L5 - 4L6
-		"psllw $2, %%mm3				\n\t" // 4H5 - 4H6
-		"psubw %%mm2, %%mm0				\n\t" // 2L4 - 5L5 + 5L6
-		"psubw %%mm3, %%mm1				\n\t" // 2H4 - 5H5 + 5H6
-
-		"movq (%0, %1, 4), %%mm2			\n\t"
-		"movq %%mm2, %%mm3				\n\t"
-		"punpcklbw %%mm7, %%mm2				\n\t" // L7
-		"punpckhbw %%mm7, %%mm3				\n\t" // H7
-
-		"paddw %%mm2, %%mm2				\n\t" // 2L7
-		"paddw %%mm3, %%mm3				\n\t" // 2H7
-		"psubw %%mm2, %%mm0				\n\t" // 2L4 - 5L5 + 5L6 - 2L7
-		"psubw %%mm3, %%mm1				\n\t" // 2H4 - 5H5 + 5H6 - 2H7
-
-		"movq (%%"REG_c"), %%mm2			\n\t" // 2L0 - 5L1 + 5L2 - 2L3
-		"movq 8(%%"REG_c"), %%mm3			\n\t" // 2H0 - 5H1 + 5H2 - 2H3
+                "movq (%0, %1, 2), %%mm2                \n\t"
+                "movq %%mm2, %%mm3                      \n\t"
+                "punpcklbw %%mm7, %%mm2                 \n\t" // L5
+                "punpckhbw %%mm7, %%mm3                 \n\t" // H5
+                "psubw %%mm2, %%mm4                     \n\t" // 2L2 - 5L3 + 5L4 - L5
+                "psubw %%mm3, %%mm5                     \n\t" // 2H2 - 5H3 + 5H4 - H5
+                "psubw %%mm2, %%mm4                     \n\t" // 2L2 - 5L3 + 5L4 - 2L5
+                "psubw %%mm3, %%mm5                     \n\t" // 2H2 - 5H3 + 5H4 - 2H5
+
+                "movq (%%"REG_a", %1, 4), %%mm6         \n\t"
+                "punpcklbw %%mm7, %%mm6                 \n\t" // L6
+                "psubw %%mm6, %%mm2                     \n\t" // L5 - L6
+                "movq (%%"REG_a", %1, 4), %%mm6         \n\t"
+                "punpckhbw %%mm7, %%mm6                 \n\t" // H6
+                "psubw %%mm6, %%mm3                     \n\t" // H5 - H6
+
+                "paddw %%mm0, %%mm0                     \n\t" // 2L4
+                "paddw %%mm1, %%mm1                     \n\t" // 2H4
+                "psubw %%mm2, %%mm0                     \n\t" // 2L4 - L5 + L6
+                "psubw %%mm3, %%mm1                     \n\t" // 2H4 - H5 + H6
+
+                "psllw $2, %%mm2                        \n\t" // 4L5 - 4L6
+                "psllw $2, %%mm3                        \n\t" // 4H5 - 4H6
+                "psubw %%mm2, %%mm0                     \n\t" // 2L4 - 5L5 + 5L6
+                "psubw %%mm3, %%mm1                     \n\t" // 2H4 - 5H5 + 5H6
+
+                "movq (%0, %1, 4), %%mm2                \n\t"
+                "movq %%mm2, %%mm3                      \n\t"
+                "punpcklbw %%mm7, %%mm2                 \n\t" // L7
+                "punpckhbw %%mm7, %%mm3                 \n\t" // H7
+
+                "paddw %%mm2, %%mm2                     \n\t" // 2L7
+                "paddw %%mm3, %%mm3                     \n\t" // 2H7
+                "psubw %%mm2, %%mm0                     \n\t" // 2L4 - 5L5 + 5L6 - 2L7
+                "psubw %%mm3, %%mm1                     \n\t" // 2H4 - 5H5 + 5H6 - 2H7
+
+                "movq (%%"REG_c"), %%mm2                \n\t" // 2L0 - 5L1 + 5L2 - 2L3
+                "movq 8(%%"REG_c"), %%mm3               \n\t" // 2H0 - 5H1 + 5H2 - 2H3
 
 #ifdef HAVE_MMX2
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"psubw %%mm0, %%mm6				\n\t"
-		"pmaxsw %%mm6, %%mm0				\n\t" // |2L4 - 5L5 + 5L6 - 2L7|
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"psubw %%mm1, %%mm6				\n\t"
-		"pmaxsw %%mm6, %%mm1				\n\t" // |2H4 - 5H5 + 5H6 - 2H7|
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"psubw %%mm2, %%mm6				\n\t"
-		"pmaxsw %%mm6, %%mm2				\n\t" // |2L0 - 5L1 + 5L2 - 2L3|
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"psubw %%mm3, %%mm6				\n\t"
-		"pmaxsw %%mm6, %%mm3				\n\t" // |2H0 - 5H1 + 5H2 - 2H3|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "psubw %%mm0, %%mm6                     \n\t"
+                "pmaxsw %%mm6, %%mm0                    \n\t" // |2L4 - 5L5 + 5L6 - 2L7|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "psubw %%mm1, %%mm6                     \n\t"
+                "pmaxsw %%mm6, %%mm1                    \n\t" // |2H4 - 5H5 + 5H6 - 2H7|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "psubw %%mm2, %%mm6                     \n\t"
+                "pmaxsw %%mm6, %%mm2                    \n\t" // |2L0 - 5L1 + 5L2 - 2L3|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "psubw %%mm3, %%mm6                     \n\t"
+                "pmaxsw %%mm6, %%mm3                    \n\t" // |2H0 - 5H1 + 5H2 - 2H3|
 #else
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"pcmpgtw %%mm0, %%mm6				\n\t"
-		"pxor %%mm6, %%mm0				\n\t"
-		"psubw %%mm6, %%mm0				\n\t" // |2L4 - 5L5 + 5L6 - 2L7|
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"pcmpgtw %%mm1, %%mm6				\n\t"
-		"pxor %%mm6, %%mm1				\n\t"
-		"psubw %%mm6, %%mm1				\n\t" // |2H4 - 5H5 + 5H6 - 2H7|
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"pcmpgtw %%mm2, %%mm6				\n\t"
-		"pxor %%mm6, %%mm2				\n\t"
-		"psubw %%mm6, %%mm2				\n\t" // |2L0 - 5L1 + 5L2 - 2L3|
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"pcmpgtw %%mm3, %%mm6				\n\t"
-		"pxor %%mm6, %%mm3				\n\t"
-		"psubw %%mm6, %%mm3				\n\t" // |2H0 - 5H1 + 5H2 - 2H3|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "pcmpgtw %%mm0, %%mm6                   \n\t"
+                "pxor %%mm6, %%mm0                      \n\t"
+                "psubw %%mm6, %%mm0                     \n\t" // |2L4 - 5L5 + 5L6 - 2L7|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "pcmpgtw %%mm1, %%mm6                   \n\t"
+                "pxor %%mm6, %%mm1                      \n\t"
+                "psubw %%mm6, %%mm1                     \n\t" // |2H4 - 5H5 + 5H6 - 2H7|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "pcmpgtw %%mm2, %%mm6                   \n\t"
+                "pxor %%mm6, %%mm2                      \n\t"
+                "psubw %%mm6, %%mm2                     \n\t" // |2L0 - 5L1 + 5L2 - 2L3|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "pcmpgtw %%mm3, %%mm6                   \n\t"
+                "pxor %%mm6, %%mm3                      \n\t"
+                "psubw %%mm6, %%mm3                     \n\t" // |2H0 - 5H1 + 5H2 - 2H3|
 #endif
 
 #ifdef HAVE_MMX2
-		"pminsw %%mm2, %%mm0				\n\t"
-		"pminsw %%mm3, %%mm1				\n\t"
+                "pminsw %%mm2, %%mm0                    \n\t"
+                "pminsw %%mm3, %%mm1                    \n\t"
 #else
-		"movq %%mm0, %%mm6				\n\t"
-		"psubusw %%mm2, %%mm6				\n\t"
-		"psubw %%mm6, %%mm0				\n\t"
-		"movq %%mm1, %%mm6				\n\t"
-		"psubusw %%mm3, %%mm6				\n\t"
-		"psubw %%mm6, %%mm1				\n\t"
+                "movq %%mm0, %%mm6                      \n\t"
+                "psubusw %%mm2, %%mm6                   \n\t"
+                "psubw %%mm6, %%mm0                     \n\t"
+                "movq %%mm1, %%mm6                      \n\t"
+                "psubusw %%mm3, %%mm6                   \n\t"
+                "psubw %%mm6, %%mm1                     \n\t"
 #endif
 
-		"movd %2, %%mm2					\n\t" // QP
-		"punpcklbw %%mm7, %%mm2				\n\t"
+                "movd %2, %%mm2                         \n\t" // QP
+                "punpcklbw %%mm7, %%mm2                 \n\t"
 
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"pcmpgtw %%mm4, %%mm6				\n\t" // sign(2L2 - 5L3 + 5L4 - 2L5)
-		"pxor %%mm6, %%mm4				\n\t"
-		"psubw %%mm6, %%mm4				\n\t" // |2L2 - 5L3 + 5L4 - 2L5|
-		"pcmpgtw %%mm5, %%mm7				\n\t" // sign(2H2 - 5H3 + 5H4 - 2H5)
-		"pxor %%mm7, %%mm5				\n\t"
-		"psubw %%mm7, %%mm5				\n\t" // |2H2 - 5H3 + 5H4 - 2H5|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "pcmpgtw %%mm4, %%mm6                   \n\t" // sign(2L2 - 5L3 + 5L4 - 2L5)
+                "pxor %%mm6, %%mm4                      \n\t"
+                "psubw %%mm6, %%mm4                     \n\t" // |2L2 - 5L3 + 5L4 - 2L5|
+                "pcmpgtw %%mm5, %%mm7                   \n\t" // sign(2H2 - 5H3 + 5H4 - 2H5)
+                "pxor %%mm7, %%mm5                      \n\t"
+                "psubw %%mm7, %%mm5                     \n\t" // |2H2 - 5H3 + 5H4 - 2H5|
 // 100 opcodes
-		"psllw $3, %%mm2				\n\t" // 8QP
-		"movq %%mm2, %%mm3				\n\t" // 8QP
-		"pcmpgtw %%mm4, %%mm2				\n\t"
-		"pcmpgtw %%mm5, %%mm3				\n\t"
-		"pand %%mm2, %%mm4				\n\t"
-		"pand %%mm3, %%mm5				\n\t"
-
-
-		"psubusw %%mm0, %%mm4				\n\t" // hd
-		"psubusw %%mm1, %%mm5				\n\t" // ld
-
-
-		"movq "MANGLE(w05)", %%mm2			\n\t" // 5
-		"pmullw %%mm2, %%mm4				\n\t"
-		"pmullw %%mm2, %%mm5				\n\t"
-		"movq "MANGLE(w20)", %%mm2			\n\t" // 32
-		"paddw %%mm2, %%mm4				\n\t"
-		"paddw %%mm2, %%mm5				\n\t"
-		"psrlw $6, %%mm4				\n\t"
-		"psrlw $6, %%mm5				\n\t"
-
-		"movq 16(%%"REG_c"), %%mm0			\n\t" // L3 - L4
-		"movq 24(%%"REG_c"), %%mm1			\n\t" // H3 - H4
-
-		"pxor %%mm2, %%mm2				\n\t"
-		"pxor %%mm3, %%mm3				\n\t"
-
-		"pcmpgtw %%mm0, %%mm2				\n\t" // sign (L3-L4)
-		"pcmpgtw %%mm1, %%mm3				\n\t" // sign (H3-H4)
-		"pxor %%mm2, %%mm0				\n\t"
-		"pxor %%mm3, %%mm1				\n\t"
-		"psubw %%mm2, %%mm0				\n\t" // |L3-L4|
-		"psubw %%mm3, %%mm1				\n\t" // |H3-H4|
-		"psrlw $1, %%mm0				\n\t" // |L3 - L4|/2
-		"psrlw $1, %%mm1				\n\t" // |H3 - H4|/2
-
-		"pxor %%mm6, %%mm2				\n\t"
-		"pxor %%mm7, %%mm3				\n\t"
-		"pand %%mm2, %%mm4				\n\t"
-		"pand %%mm3, %%mm5				\n\t"
+                "psllw $3, %%mm2                        \n\t" // 8QP
+                "movq %%mm2, %%mm3                      \n\t" // 8QP
+                "pcmpgtw %%mm4, %%mm2                   \n\t"
+                "pcmpgtw %%mm5, %%mm3                   \n\t"
+                "pand %%mm2, %%mm4                      \n\t"
+                "pand %%mm3, %%mm5                      \n\t"
+
+
+                "psubusw %%mm0, %%mm4                   \n\t" // hd
+                "psubusw %%mm1, %%mm5                   \n\t" // ld
+
+
+                "movq "MANGLE(w05)", %%mm2              \n\t" // 5
+                "pmullw %%mm2, %%mm4                    \n\t"
+                "pmullw %%mm2, %%mm5                    \n\t"
+                "movq "MANGLE(w20)", %%mm2              \n\t" // 32
+                "paddw %%mm2, %%mm4                     \n\t"
+                "paddw %%mm2, %%mm5                     \n\t"
+                "psrlw $6, %%mm4                        \n\t"
+                "psrlw $6, %%mm5                        \n\t"
+
+                "movq 16(%%"REG_c"), %%mm0              \n\t" // L3 - L4
+                "movq 24(%%"REG_c"), %%mm1              \n\t" // H3 - H4
+
+                "pxor %%mm2, %%mm2                      \n\t"
+                "pxor %%mm3, %%mm3                      \n\t"
+
+                "pcmpgtw %%mm0, %%mm2                   \n\t" // sign (L3-L4)
+                "pcmpgtw %%mm1, %%mm3                   \n\t" // sign (H3-H4)
+                "pxor %%mm2, %%mm0                      \n\t"
+                "pxor %%mm3, %%mm1                      \n\t"
+                "psubw %%mm2, %%mm0                     \n\t" // |L3-L4|
+                "psubw %%mm3, %%mm1                     \n\t" // |H3-H4|
+                "psrlw $1, %%mm0                        \n\t" // |L3 - L4|/2
+                "psrlw $1, %%mm1                        \n\t" // |H3 - H4|/2
+
+                "pxor %%mm6, %%mm2                      \n\t"
+                "pxor %%mm7, %%mm3                      \n\t"
+                "pand %%mm2, %%mm4                      \n\t"
+                "pand %%mm3, %%mm5                      \n\t"
 
 #ifdef HAVE_MMX2
-		"pminsw %%mm0, %%mm4				\n\t"
-		"pminsw %%mm1, %%mm5				\n\t"
+                "pminsw %%mm0, %%mm4                    \n\t"
+                "pminsw %%mm1, %%mm5                    \n\t"
 #else
-		"movq %%mm4, %%mm2				\n\t"
-		"psubusw %%mm0, %%mm2				\n\t"
-		"psubw %%mm2, %%mm4				\n\t"
-		"movq %%mm5, %%mm2				\n\t"
-		"psubusw %%mm1, %%mm2				\n\t"
-		"psubw %%mm2, %%mm5				\n\t"
+                "movq %%mm4, %%mm2                      \n\t"
+                "psubusw %%mm0, %%mm2                   \n\t"
+                "psubw %%mm2, %%mm4                     \n\t"
+                "movq %%mm5, %%mm2                      \n\t"
+                "psubusw %%mm1, %%mm2                   \n\t"
+                "psubw %%mm2, %%mm5                     \n\t"
 #endif
-		"pxor %%mm6, %%mm4				\n\t"
-		"pxor %%mm7, %%mm5				\n\t"
-		"psubw %%mm6, %%mm4				\n\t"
-		"psubw %%mm7, %%mm5				\n\t"
-		"packsswb %%mm5, %%mm4				\n\t"
-		"movq (%0), %%mm0				\n\t"
-		"paddb   %%mm4, %%mm0				\n\t"
-		"movq %%mm0, (%0)				\n\t"
-		"movq (%0, %1), %%mm0				\n\t"
-		"psubb %%mm4, %%mm0				\n\t"
-		"movq %%mm0, (%0, %1)				\n\t"
-
-		: "+r" (src)
-		: "r" ((long)stride), "m" (c->pQPb)
-		: "%"REG_a, "%"REG_c
-	);
+                "pxor %%mm6, %%mm4                      \n\t"
+                "pxor %%mm7, %%mm5                      \n\t"
+                "psubw %%mm6, %%mm4                     \n\t"
+                "psubw %%mm7, %%mm5                     \n\t"
+                "packsswb %%mm5, %%mm4                  \n\t"
+                "movq (%0), %%mm0                       \n\t"
+                "paddb   %%mm4, %%mm0                   \n\t"
+                "movq %%mm0, (%0)                       \n\t"
+                "movq (%0, %1), %%mm0                   \n\t"
+                "psubb %%mm4, %%mm0                     \n\t"
+                "movq %%mm0, (%0, %1)                   \n\t"
+
+                : "+r" (src)
+                : "r" ((long)stride), "m" (c->pQPb)
+                : "%"REG_a, "%"REG_c
+        );
 #else //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	const int l1= stride;
-	const int l2= stride + l1;
-	const int l3= stride + l2;
-	const int l4= stride + l3;
-	const int l5= stride + l4;
-	const int l6= stride + l5;
-	const int l7= stride + l6;
-	const int l8= stride + l7;
-//	const int l9= stride + l8;
-	int x;
-	src+= stride*3;
-	for(x=0; x<BLOCK_SIZE; x++)
-	{
-		const int middleEnergy= 5*(src[l5] - src[l4]) + 2*(src[l3] - src[l6]);
-		if(ABS(middleEnergy) < 8*c->QP)
-		{
-			const int q=(src[l4] - src[l5])/2;
-			const int leftEnergy=  5*(src[l3] - src[l2]) + 2*(src[l1] - src[l4]);
-			const int rightEnergy= 5*(src[l7] - src[l6]) + 2*(src[l5] - src[l8]);
-
-			int d= ABS(middleEnergy) - MIN( ABS(leftEnergy), ABS(rightEnergy) );
-			d= MAX(d, 0);
-
-			d= (5*d + 32) >> 6;
-			d*= SIGN(-middleEnergy);
-
-			if(q>0)
-			{
-				d= d<0 ? 0 : d;
-				d= d>q ? q : d;
-			}
-			else
-			{
-				d= d>0 ? 0 : d;
-				d= d<q ? q : d;
-			}
-
-        		src[l4]-= d;
-	        	src[l5]+= d;
-		}
-		src++;
-	}
+        const int l1= stride;
+        const int l2= stride + l1;
+        const int l3= stride + l2;
+        const int l4= stride + l3;
+        const int l5= stride + l4;
+        const int l6= stride + l5;
+        const int l7= stride + l6;
+        const int l8= stride + l7;
+//        const int l9= stride + l8;
+        int x;
+        src+= stride*3;
+        for(x=0; x<BLOCK_SIZE; x++)
+        {
+                const int middleEnergy= 5*(src[l5] - src[l4]) + 2*(src[l3] - src[l6]);
+                if(ABS(middleEnergy) < 8*c->QP)
+                {
+                        const int q=(src[l4] - src[l5])/2;
+                        const int leftEnergy=  5*(src[l3] - src[l2]) + 2*(src[l1] - src[l4]);
+                        const int rightEnergy= 5*(src[l7] - src[l6]) + 2*(src[l5] - src[l8]);
+
+                        int d= ABS(middleEnergy) - MIN( ABS(leftEnergy), ABS(rightEnergy) );
+                        d= MAX(d, 0);
+
+                        d= (5*d + 32) >> 6;
+                        d*= SIGN(-middleEnergy);
+
+                        if(q>0)
+                        {
+                                d= d<0 ? 0 : d;
+                                d= d>q ? q : d;
+                        }
+                        else
+                        {
+                                d= d>0 ? 0 : d;
+                                d= d<q ? q : d;
+                        }
+
+                        src[l4]-= d;
+                        src[l5]+= d;
+                }
+                src++;
+        }
 #endif //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 }
 #endif //HAVE_ALTIVEC
@@ -1183,36 +1183,36 @@ src-=8;
 static inline void RENAME(dering)(uint8_t src[], int stride, PPContext *c)
 {
 #if defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	asm volatile(
-		"pxor %%mm6, %%mm6				\n\t"
-		"pcmpeqb %%mm7, %%mm7				\n\t"
-		"movq %2, %%mm0					\n\t"
-		"punpcklbw %%mm6, %%mm0				\n\t"
-		"psrlw $1, %%mm0				\n\t"
-		"psubw %%mm7, %%mm0				\n\t"
-		"packuswb %%mm0, %%mm0				\n\t"
-		"movq %%mm0, %3					\n\t"
-
-		"lea (%0, %1), %%"REG_a"			\n\t"
-		"lea (%%"REG_a", %1, 4), %%"REG_d"		\n\t"
-
-//	0	1	2	3	4	5	6	7	8	9
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	edx	edx+%1	edx+2%1	%0+8%1	edx+4%1
+        asm volatile(
+                "pxor %%mm6, %%mm6                      \n\t"
+                "pcmpeqb %%mm7, %%mm7                   \n\t"
+                "movq %2, %%mm0                         \n\t"
+                "punpcklbw %%mm6, %%mm0                 \n\t"
+                "psrlw $1, %%mm0                        \n\t"
+                "psubw %%mm7, %%mm0                     \n\t"
+                "packuswb %%mm0, %%mm0                  \n\t"
+                "movq %%mm0, %3                         \n\t"
+
+                "lea (%0, %1), %%"REG_a"                \n\t"
+                "lea (%%"REG_a", %1, 4), %%"REG_d"      \n\t"
+
+//        0        1        2        3        4        5        6        7        8        9
+//        %0        eax        eax+%1        eax+2%1        %0+4%1        edx        edx+%1        edx+2%1        %0+8%1        edx+4%1
 
 #undef FIND_MIN_MAX
 #ifdef HAVE_MMX2
 #define REAL_FIND_MIN_MAX(addr)\
-		"movq " #addr ", %%mm0				\n\t"\
-		"pminub %%mm0, %%mm7				\n\t"\
-		"pmaxub %%mm0, %%mm6				\n\t"
+                "movq " #addr ", %%mm0                  \n\t"\
+                "pminub %%mm0, %%mm7                    \n\t"\
+                "pmaxub %%mm0, %%mm6                    \n\t"
 #else
 #define REAL_FIND_MIN_MAX(addr)\
-		"movq " #addr ", %%mm0				\n\t"\
-		"movq %%mm7, %%mm1				\n\t"\
-		"psubusb %%mm0, %%mm6				\n\t"\
-		"paddb %%mm0, %%mm6				\n\t"\
-		"psubusb %%mm0, %%mm1				\n\t"\
-		"psubb %%mm1, %%mm7				\n\t"
+                "movq " #addr ", %%mm0                  \n\t"\
+                "movq %%mm7, %%mm1                      \n\t"\
+                "psubusb %%mm0, %%mm6                   \n\t"\
+                "paddb %%mm0, %%mm6                     \n\t"\
+                "psubusb %%mm0, %%mm1                   \n\t"\
+                "psubb %%mm1, %%mm7                     \n\t"
 #endif
 #define FIND_MIN_MAX(addr)  REAL_FIND_MIN_MAX(addr)
 
@@ -1225,155 +1225,155 @@ FIND_MIN_MAX((%%REGd, %1))
 FIND_MIN_MAX((%%REGd, %1, 2))
 FIND_MIN_MAX((%0, %1, 8))
 
-		"movq %%mm7, %%mm4				\n\t"
-		"psrlq $8, %%mm7				\n\t"
+                "movq %%mm7, %%mm4                      \n\t"
+                "psrlq $8, %%mm7                        \n\t"
 #ifdef HAVE_MMX2
-		"pminub %%mm4, %%mm7				\n\t" // min of pixels
-		"pshufw $0xF9, %%mm7, %%mm4			\n\t"
-		"pminub %%mm4, %%mm7				\n\t" // min of pixels
-		"pshufw $0xFE, %%mm7, %%mm4			\n\t"
-		"pminub %%mm4, %%mm7				\n\t"
+                "pminub %%mm4, %%mm7                    \n\t" // min of pixels
+                "pshufw $0xF9, %%mm7, %%mm4             \n\t"
+                "pminub %%mm4, %%mm7                    \n\t" // min of pixels
+                "pshufw $0xFE, %%mm7, %%mm4             \n\t"
+                "pminub %%mm4, %%mm7                    \n\t"
 #else
-		"movq %%mm7, %%mm1				\n\t"
-		"psubusb %%mm4, %%mm1				\n\t"
-		"psubb %%mm1, %%mm7				\n\t"
-		"movq %%mm7, %%mm4				\n\t"
-		"psrlq $16, %%mm7				\n\t"
-		"movq %%mm7, %%mm1				\n\t"
-		"psubusb %%mm4, %%mm1				\n\t"
-		"psubb %%mm1, %%mm7				\n\t"
-		"movq %%mm7, %%mm4				\n\t"
-		"psrlq $32, %%mm7				\n\t"
-		"movq %%mm7, %%mm1				\n\t"
-		"psubusb %%mm4, %%mm1				\n\t"
-		"psubb %%mm1, %%mm7				\n\t"
+                "movq %%mm7, %%mm1                      \n\t"
+                "psubusb %%mm4, %%mm1                   \n\t"
+                "psubb %%mm1, %%mm7                     \n\t"
+                "movq %%mm7, %%mm4                      \n\t"
+                "psrlq $16, %%mm7                       \n\t"
+                "movq %%mm7, %%mm1                      \n\t"
+                "psubusb %%mm4, %%mm1                   \n\t"
+                "psubb %%mm1, %%mm7                     \n\t"
+                "movq %%mm7, %%mm4                      \n\t"
+                "psrlq $32, %%mm7                       \n\t"
+                "movq %%mm7, %%mm1                      \n\t"
+                "psubusb %%mm4, %%mm1                   \n\t"
+                "psubb %%mm1, %%mm7                     \n\t"
 #endif
 
 
-		"movq %%mm6, %%mm4				\n\t"
-		"psrlq $8, %%mm6				\n\t"
+                "movq %%mm6, %%mm4                      \n\t"
+                "psrlq $8, %%mm6                        \n\t"
 #ifdef HAVE_MMX2
-		"pmaxub %%mm4, %%mm6				\n\t" // max of pixels
-		"pshufw $0xF9, %%mm6, %%mm4			\n\t"
-		"pmaxub %%mm4, %%mm6				\n\t"
-		"pshufw $0xFE, %%mm6, %%mm4			\n\t"
-		"pmaxub %%mm4, %%mm6				\n\t"
+                "pmaxub %%mm4, %%mm6                    \n\t" // max of pixels
+                "pshufw $0xF9, %%mm6, %%mm4             \n\t"
+                "pmaxub %%mm4, %%mm6                    \n\t"
+                "pshufw $0xFE, %%mm6, %%mm4             \n\t"
+                "pmaxub %%mm4, %%mm6                    \n\t"
 #else
-		"psubusb %%mm4, %%mm6				\n\t"
-		"paddb %%mm4, %%mm6				\n\t"
-		"movq %%mm6, %%mm4				\n\t"
-		"psrlq $16, %%mm6				\n\t"
-		"psubusb %%mm4, %%mm6				\n\t"
-		"paddb %%mm4, %%mm6				\n\t"
-		"movq %%mm6, %%mm4				\n\t"
-		"psrlq $32, %%mm6				\n\t"
-		"psubusb %%mm4, %%mm6				\n\t"
-		"paddb %%mm4, %%mm6				\n\t"
+                "psubusb %%mm4, %%mm6                   \n\t"
+                "paddb %%mm4, %%mm6                     \n\t"
+                "movq %%mm6, %%mm4                      \n\t"
+                "psrlq $16, %%mm6                       \n\t"
+                "psubusb %%mm4, %%mm6                   \n\t"
+                "paddb %%mm4, %%mm6                     \n\t"
+                "movq %%mm6, %%mm4                      \n\t"
+                "psrlq $32, %%mm6                       \n\t"
+                "psubusb %%mm4, %%mm6                   \n\t"
+                "paddb %%mm4, %%mm6                     \n\t"
 #endif
-		"movq %%mm6, %%mm0				\n\t" // max
-		"psubb %%mm7, %%mm6				\n\t" // max - min
-		"movd %%mm6, %%ecx				\n\t"
-		"cmpb "MANGLE(deringThreshold)", %%cl		\n\t"
-		" jb 1f						\n\t"
-		"lea -24(%%"REG_SP"), %%"REG_c"			\n\t"
-		"and "ALIGN_MASK", %%"REG_c"			\n\t"
-		PAVGB(%%mm0, %%mm7)				      // a=(max + min)/2
-		"punpcklbw %%mm7, %%mm7				\n\t"
-		"punpcklbw %%mm7, %%mm7				\n\t"
-		"punpcklbw %%mm7, %%mm7				\n\t"
-		"movq %%mm7, (%%"REG_c")			\n\t"
-
-		"movq (%0), %%mm0				\n\t" // L10
-		"movq %%mm0, %%mm1				\n\t" // L10
-		"movq %%mm0, %%mm2				\n\t" // L10
-		"psllq $8, %%mm1				\n\t"
-		"psrlq $8, %%mm2				\n\t"
-		"movd -4(%0), %%mm3				\n\t"
-		"movd 8(%0), %%mm4				\n\t"
-		"psrlq $24, %%mm3				\n\t"
-		"psllq $56, %%mm4				\n\t"
-		"por %%mm3, %%mm1				\n\t" // L00
-		"por %%mm4, %%mm2				\n\t" // L20
-		"movq %%mm1, %%mm3				\n\t" // L00
-		PAVGB(%%mm2, %%mm1)				      // (L20 + L00)/2
-		PAVGB(%%mm0, %%mm1)				      // (L20 + L00 + 2L10)/4
-		"psubusb %%mm7, %%mm0				\n\t"
-		"psubusb %%mm7, %%mm2				\n\t"
-		"psubusb %%mm7, %%mm3				\n\t"
-		"pcmpeqb "MANGLE(b00)", %%mm0			\n\t" // L10 > a ? 0 : -1
-		"pcmpeqb "MANGLE(b00)", %%mm2			\n\t" // L20 > a ? 0 : -1
-		"pcmpeqb "MANGLE(b00)", %%mm3			\n\t" // L00 > a ? 0 : -1
-		"paddb %%mm2, %%mm0				\n\t"
-		"paddb %%mm3, %%mm0				\n\t"
-
-		"movq (%%"REG_a"), %%mm2			\n\t" // L11
-		"movq %%mm2, %%mm3				\n\t" // L11
-		"movq %%mm2, %%mm4				\n\t" // L11
-		"psllq $8, %%mm3				\n\t"
-		"psrlq $8, %%mm4				\n\t"
-		"movd -4(%%"REG_a"), %%mm5			\n\t"
-		"movd 8(%%"REG_a"), %%mm6			\n\t"
-		"psrlq $24, %%mm5				\n\t"
-		"psllq $56, %%mm6				\n\t"
-		"por %%mm5, %%mm3				\n\t" // L01
-		"por %%mm6, %%mm4				\n\t" // L21
-		"movq %%mm3, %%mm5				\n\t" // L01
-		PAVGB(%%mm4, %%mm3)				      // (L21 + L01)/2
-		PAVGB(%%mm2, %%mm3)				      // (L21 + L01 + 2L11)/4
-		"psubusb %%mm7, %%mm2				\n\t"
-		"psubusb %%mm7, %%mm4				\n\t"
-		"psubusb %%mm7, %%mm5				\n\t"
-		"pcmpeqb "MANGLE(b00)", %%mm2			\n\t" // L11 > a ? 0 : -1
-		"pcmpeqb "MANGLE(b00)", %%mm4			\n\t" // L21 > a ? 0 : -1
-		"pcmpeqb "MANGLE(b00)", %%mm5			\n\t" // L01 > a ? 0 : -1
-		"paddb %%mm4, %%mm2				\n\t"
-		"paddb %%mm5, %%mm2				\n\t"
+                "movq %%mm6, %%mm0                      \n\t" // max
+                "psubb %%mm7, %%mm6                     \n\t" // max - min
+                "movd %%mm6, %%ecx                      \n\t"
+                "cmpb "MANGLE(deringThreshold)", %%cl   \n\t"
+                " jb 1f                                 \n\t"
+                "lea -24(%%"REG_SP"), %%"REG_c"         \n\t"
+                "and "ALIGN_MASK", %%"REG_c"            \n\t"
+                PAVGB(%%mm0, %%mm7)                           // a=(max + min)/2
+                "punpcklbw %%mm7, %%mm7                 \n\t"
+                "punpcklbw %%mm7, %%mm7                 \n\t"
+                "punpcklbw %%mm7, %%mm7                 \n\t"
+                "movq %%mm7, (%%"REG_c")                \n\t"
+
+                "movq (%0), %%mm0                       \n\t" // L10
+                "movq %%mm0, %%mm1                      \n\t" // L10
+                "movq %%mm0, %%mm2                      \n\t" // L10
+                "psllq $8, %%mm1                        \n\t"
+                "psrlq $8, %%mm2                        \n\t"
+                "movd -4(%0), %%mm3                     \n\t"
+                "movd 8(%0), %%mm4                      \n\t"
+                "psrlq $24, %%mm3                       \n\t"
+                "psllq $56, %%mm4                       \n\t"
+                "por %%mm3, %%mm1                       \n\t" // L00
+                "por %%mm4, %%mm2                       \n\t" // L20
+                "movq %%mm1, %%mm3                      \n\t" // L00
+                PAVGB(%%mm2, %%mm1)                           // (L20 + L00)/2
+                PAVGB(%%mm0, %%mm1)                           // (L20 + L00 + 2L10)/4
+                "psubusb %%mm7, %%mm0                   \n\t"
+                "psubusb %%mm7, %%mm2                   \n\t"
+                "psubusb %%mm7, %%mm3                   \n\t"
+                "pcmpeqb "MANGLE(b00)", %%mm0           \n\t" // L10 > a ? 0 : -1
+                "pcmpeqb "MANGLE(b00)", %%mm2           \n\t" // L20 > a ? 0 : -1
+                "pcmpeqb "MANGLE(b00)", %%mm3           \n\t" // L00 > a ? 0 : -1
+                "paddb %%mm2, %%mm0                     \n\t"
+                "paddb %%mm3, %%mm0                     \n\t"
+
+                "movq (%%"REG_a"), %%mm2                \n\t" // L11
+                "movq %%mm2, %%mm3                      \n\t" // L11
+                "movq %%mm2, %%mm4                      \n\t" // L11
+                "psllq $8, %%mm3                        \n\t"
+                "psrlq $8, %%mm4                        \n\t"
+                "movd -4(%%"REG_a"), %%mm5              \n\t"
+                "movd 8(%%"REG_a"), %%mm6               \n\t"
+                "psrlq $24, %%mm5                       \n\t"
+                "psllq $56, %%mm6                       \n\t"
+                "por %%mm5, %%mm3                       \n\t" // L01
+                "por %%mm6, %%mm4                       \n\t" // L21
+                "movq %%mm3, %%mm5                      \n\t" // L01
+                PAVGB(%%mm4, %%mm3)                           // (L21 + L01)/2
+                PAVGB(%%mm2, %%mm3)                           // (L21 + L01 + 2L11)/4
+                "psubusb %%mm7, %%mm2                   \n\t"
+                "psubusb %%mm7, %%mm4                   \n\t"
+                "psubusb %%mm7, %%mm5                   \n\t"
+                "pcmpeqb "MANGLE(b00)", %%mm2           \n\t" // L11 > a ? 0 : -1
+                "pcmpeqb "MANGLE(b00)", %%mm4           \n\t" // L21 > a ? 0 : -1
+                "pcmpeqb "MANGLE(b00)", %%mm5           \n\t" // L01 > a ? 0 : -1
+                "paddb %%mm4, %%mm2                     \n\t"
+                "paddb %%mm5, %%mm2                     \n\t"
 // 0, 2, 3, 1
 #define REAL_DERING_CORE(dst,src,ppsx,psx,sx,pplx,plx,lx,t0,t1) \
-		"movq " #src ", " #sx "				\n\t" /* src[0] */\
-		"movq " #sx ", " #lx "				\n\t" /* src[0] */\
-		"movq " #sx ", " #t0 "				\n\t" /* src[0] */\
-		"psllq $8, " #lx "				\n\t"\
-		"psrlq $8, " #t0 "				\n\t"\
-		"movd -4" #src ", " #t1 "			\n\t"\
-		"psrlq $24, " #t1 "				\n\t"\
-		"por " #t1 ", " #lx "				\n\t" /* src[-1] */\
-		"movd 8" #src ", " #t1 "			\n\t"\
-		"psllq $56, " #t1 "				\n\t"\
-		"por " #t1 ", " #t0 "				\n\t" /* src[+1] */\
-		"movq " #lx ", " #t1 "				\n\t" /* src[-1] */\
-		PAVGB(t0, lx)				              /* (src[-1] + src[+1])/2 */\
-		PAVGB(sx, lx)				      /* (src[-1] + 2src[0] + src[+1])/4 */\
-		PAVGB(lx, pplx)					     \
-		"movq " #lx ", 8(%%"REG_c")			\n\t"\
-		"movq (%%"REG_c"), " #lx "			\n\t"\
-		"psubusb " #lx ", " #t1 "			\n\t"\
-		"psubusb " #lx ", " #t0 "			\n\t"\
-		"psubusb " #lx ", " #sx "			\n\t"\
-		"movq "MANGLE(b00)", " #lx "			\n\t"\
-		"pcmpeqb " #lx ", " #t1 "			\n\t" /* src[-1] > a ? 0 : -1*/\
-		"pcmpeqb " #lx ", " #t0 "			\n\t" /* src[+1] > a ? 0 : -1*/\
-		"pcmpeqb " #lx ", " #sx "			\n\t" /* src[0]  > a ? 0 : -1*/\
-		"paddb " #t1 ", " #t0 "				\n\t"\
-		"paddb " #t0 ", " #sx "				\n\t"\
+                "movq " #src ", " #sx "                 \n\t" /* src[0] */\
+                "movq " #sx ", " #lx "                  \n\t" /* src[0] */\
+                "movq " #sx ", " #t0 "                  \n\t" /* src[0] */\
+                "psllq $8, " #lx "                      \n\t"\
+                "psrlq $8, " #t0 "                      \n\t"\
+                "movd -4" #src ", " #t1 "               \n\t"\
+                "psrlq $24, " #t1 "                     \n\t"\
+                "por " #t1 ", " #lx "                   \n\t" /* src[-1] */\
+                "movd 8" #src ", " #t1 "                \n\t"\
+                "psllq $56, " #t1 "                     \n\t"\
+                "por " #t1 ", " #t0 "                   \n\t" /* src[+1] */\
+                "movq " #lx ", " #t1 "                  \n\t" /* src[-1] */\
+                PAVGB(t0, lx)                                 /* (src[-1] + src[+1])/2 */\
+                PAVGB(sx, lx)                                 /* (src[-1] + 2src[0] + src[+1])/4 */\
+                PAVGB(lx, pplx)                                     \
+                "movq " #lx ", 8(%%"REG_c")             \n\t"\
+                "movq (%%"REG_c"), " #lx "              \n\t"\
+                "psubusb " #lx ", " #t1 "               \n\t"\
+                "psubusb " #lx ", " #t0 "               \n\t"\
+                "psubusb " #lx ", " #sx "               \n\t"\
+                "movq "MANGLE(b00)", " #lx "            \n\t"\
+                "pcmpeqb " #lx ", " #t1 "               \n\t" /* src[-1] > a ? 0 : -1*/\
+                "pcmpeqb " #lx ", " #t0 "               \n\t" /* src[+1] > a ? 0 : -1*/\
+                "pcmpeqb " #lx ", " #sx "               \n\t" /* src[0]  > a ? 0 : -1*/\
+                "paddb " #t1 ", " #t0 "                 \n\t"\
+                "paddb " #t0 ", " #sx "                 \n\t"\
 \
-		PAVGB(plx, pplx)				      /* filtered */\
-		"movq " #dst ", " #t0 "				\n\t" /* dst */\
-		"movq " #t0 ", " #t1 "				\n\t" /* dst */\
-		"psubusb %3, " #t0 "				\n\t"\
-		"paddusb %3, " #t1 "				\n\t"\
-		PMAXUB(t0, pplx)\
-		PMINUB(t1, pplx, t0)\
-		"paddb " #sx ", " #ppsx "			\n\t"\
-		"paddb " #psx ", " #ppsx "			\n\t"\
-		"#paddb "MANGLE(b02)", " #ppsx "		\n\t"\
-		"pand "MANGLE(b08)", " #ppsx "			\n\t"\
-		"pcmpeqb " #lx ", " #ppsx "			\n\t"\
-		"pand " #ppsx ", " #pplx "			\n\t"\
-		"pandn " #dst ", " #ppsx "			\n\t"\
-		"por " #pplx ", " #ppsx "			\n\t"\
-		"movq " #ppsx ", " #dst "			\n\t"\
-		"movq 8(%%"REG_c"), " #lx "			\n\t"
+                PAVGB(plx, pplx)                              /* filtered */\
+                "movq " #dst ", " #t0 "                 \n\t" /* dst */\
+                "movq " #t0 ", " #t1 "                  \n\t" /* dst */\
+                "psubusb %3, " #t0 "                    \n\t"\
+                "paddusb %3, " #t1 "                    \n\t"\
+                PMAXUB(t0, pplx)\
+                PMINUB(t1, pplx, t0)\
+                "paddb " #sx ", " #ppsx "               \n\t"\
+                "paddb " #psx ", " #ppsx "              \n\t"\
+                "#paddb "MANGLE(b02)", " #ppsx "        \n\t"\
+                "pand "MANGLE(b08)", " #ppsx "          \n\t"\
+                "pcmpeqb " #lx ", " #ppsx "             \n\t"\
+                "pand " #ppsx ", " #pplx "              \n\t"\
+                "pandn " #dst ", " #ppsx "              \n\t"\
+                "por " #pplx ", " #ppsx "               \n\t"\
+                "movq " #ppsx ", " #dst "               \n\t"\
+                "movq 8(%%"REG_c"), " #lx "             \n\t"
 
 #define DERING_CORE(dst,src,ppsx,psx,sx,pplx,plx,lx,t0,t1) \
    REAL_DERING_CORE(dst,src,ppsx,psx,sx,pplx,plx,lx,t0,t1)
@@ -1392,149 +1392,149 @@ FIND_MIN_MAX((%0, %1, 8))
 1110111
 
 */
-//DERING_CORE(dst,src                  ,ppsx ,psx  ,sx   ,pplx ,plx  ,lx   ,t0   ,t1)
-DERING_CORE((%%REGa),(%%REGa, %1)        ,%%mm0,%%mm2,%%mm4,%%mm1,%%mm3,%%mm5,%%mm6,%%mm7)
-DERING_CORE((%%REGa, %1),(%%REGa, %1, 2) ,%%mm2,%%mm4,%%mm0,%%mm3,%%mm5,%%mm1,%%mm6,%%mm7)
-DERING_CORE((%%REGa, %1, 2),(%0, %1, 4) ,%%mm4,%%mm0,%%mm2,%%mm5,%%mm1,%%mm3,%%mm6,%%mm7)
-DERING_CORE((%0, %1, 4),(%%REGd)        ,%%mm0,%%mm2,%%mm4,%%mm1,%%mm3,%%mm5,%%mm6,%%mm7)
-DERING_CORE((%%REGd),(%%REGd, %1)        ,%%mm2,%%mm4,%%mm0,%%mm3,%%mm5,%%mm1,%%mm6,%%mm7)
-DERING_CORE((%%REGd, %1), (%%REGd, %1, 2),%%mm4,%%mm0,%%mm2,%%mm5,%%mm1,%%mm3,%%mm6,%%mm7)
-DERING_CORE((%%REGd, %1, 2),(%0, %1, 8) ,%%mm0,%%mm2,%%mm4,%%mm1,%%mm3,%%mm5,%%mm6,%%mm7)
-DERING_CORE((%0, %1, 8),(%%REGd, %1, 4) ,%%mm2,%%mm4,%%mm0,%%mm3,%%mm5,%%mm1,%%mm6,%%mm7)
-
-		"1:			\n\t"
-		: : "r" (src), "r" ((long)stride), "m" (c->pQPb), "m"(c->pQPb2)
-		: "%"REG_a, "%"REG_d, "%"REG_c
-	);
+//DERING_CORE(dst          ,src            ,ppsx ,psx  ,sx   ,pplx ,plx  ,lx   ,t0   ,t1)
+DERING_CORE((%%REGa)       ,(%%REGa, %1)   ,%%mm0,%%mm2,%%mm4,%%mm1,%%mm3,%%mm5,%%mm6,%%mm7)
+DERING_CORE((%%REGa, %1)   ,(%%REGa, %1, 2),%%mm2,%%mm4,%%mm0,%%mm3,%%mm5,%%mm1,%%mm6,%%mm7)
+DERING_CORE((%%REGa, %1, 2),(%0, %1, 4)    ,%%mm4,%%mm0,%%mm2,%%mm5,%%mm1,%%mm3,%%mm6,%%mm7)
+DERING_CORE((%0, %1, 4)    ,(%%REGd)       ,%%mm0,%%mm2,%%mm4,%%mm1,%%mm3,%%mm5,%%mm6,%%mm7)
+DERING_CORE((%%REGd)       ,(%%REGd, %1)   ,%%mm2,%%mm4,%%mm0,%%mm3,%%mm5,%%mm1,%%mm6,%%mm7)
+DERING_CORE((%%REGd, %1)   ,(%%REGd, %1, 2),%%mm4,%%mm0,%%mm2,%%mm5,%%mm1,%%mm3,%%mm6,%%mm7)
+DERING_CORE((%%REGd, %1, 2),(%0, %1, 8)    ,%%mm0,%%mm2,%%mm4,%%mm1,%%mm3,%%mm5,%%mm6,%%mm7)
+DERING_CORE((%0, %1, 8)    ,(%%REGd, %1, 4),%%mm2,%%mm4,%%mm0,%%mm3,%%mm5,%%mm1,%%mm6,%%mm7)
+
+                "1:                        \n\t"
+                : : "r" (src), "r" ((long)stride), "m" (c->pQPb), "m"(c->pQPb2)
+                : "%"REG_a, "%"REG_d, "%"REG_c
+        );
 #else //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	int y;
-	int min=255;
-	int max=0;
-	int avg;
-	uint8_t *p;
-	int s[10];
-	const int QP2= c->QP/2 + 1;
-
-	for(y=1; y<9; y++)
-	{
-		int x;
-		p= src + stride*y;
-		for(x=1; x<9; x++)
-		{
-			p++;
-			if(*p > max) max= *p;
-			if(*p < min) min= *p;
-		}
-	}
-	avg= (min + max + 1)>>1;
-
-	if(max - min <deringThreshold) return;
-
-	for(y=0; y<10; y++)
-	{
-		int t = 0;
-
-		if(src[stride*y + 0] > avg) t+= 1;
-		if(src[stride*y + 1] > avg) t+= 2;
-		if(src[stride*y + 2] > avg) t+= 4;
-		if(src[stride*y + 3] > avg) t+= 8;
-		if(src[stride*y + 4] > avg) t+= 16;
-		if(src[stride*y + 5] > avg) t+= 32;
-		if(src[stride*y + 6] > avg) t+= 64;
-		if(src[stride*y + 7] > avg) t+= 128;
-		if(src[stride*y + 8] > avg) t+= 256;
-		if(src[stride*y + 9] > avg) t+= 512;
-
-		t |= (~t)<<16;
-		t &= (t<<1) & (t>>1);
-		s[y] = t;
-	}
-
-	for(y=1; y<9; y++)
-	{
-		int t = s[y-1] & s[y] & s[y+1];
-		t|= t>>16;
-		s[y-1]= t;
-	}
-
-	for(y=1; y<9; y++)
-	{
-		int x;
-		int t = s[y-1];
-
-		p= src + stride*y;
-		for(x=1; x<9; x++)
-		{
-			p++;
-			if(t & (1<<x))
-			{
-				int f= (*(p-stride-1)) + 2*(*(p-stride)) + (*(p-stride+1))
-				      +2*(*(p     -1)) + 4*(*p         ) + 2*(*(p     +1))
-				      +(*(p+stride-1)) + 2*(*(p+stride)) + (*(p+stride+1));
-				f= (f + 8)>>4;
+        int y;
+        int min=255;
+        int max=0;
+        int avg;
+        uint8_t *p;
+        int s[10];
+        const int QP2= c->QP/2 + 1;
+
+        for(y=1; y<9; y++)
+        {
+                int x;
+                p= src + stride*y;
+                for(x=1; x<9; x++)
+                {
+                        p++;
+                        if(*p > max) max= *p;
+                        if(*p < min) min= *p;
+                }
+        }
+        avg= (min + max + 1)>>1;
+
+        if(max - min <deringThreshold) return;
+
+        for(y=0; y<10; y++)
+        {
+                int t = 0;
+
+                if(src[stride*y + 0] > avg) t+= 1;
+                if(src[stride*y + 1] > avg) t+= 2;
+                if(src[stride*y + 2] > avg) t+= 4;
+                if(src[stride*y + 3] > avg) t+= 8;
+                if(src[stride*y + 4] > avg) t+= 16;
+                if(src[stride*y + 5] > avg) t+= 32;
+                if(src[stride*y + 6] > avg) t+= 64;
+                if(src[stride*y + 7] > avg) t+= 128;
+                if(src[stride*y + 8] > avg) t+= 256;
+                if(src[stride*y + 9] > avg) t+= 512;
+
+                t |= (~t)<<16;
+                t &= (t<<1) & (t>>1);
+                s[y] = t;
+        }
+
+        for(y=1; y<9; y++)
+        {
+                int t = s[y-1] & s[y] & s[y+1];
+                t|= t>>16;
+                s[y-1]= t;
+        }
+
+        for(y=1; y<9; y++)
+        {
+                int x;
+                int t = s[y-1];
+
+                p= src + stride*y;
+                for(x=1; x<9; x++)
+                {
+                        p++;
+                        if(t & (1<<x))
+                        {
+                                int f= (*(p-stride-1)) + 2*(*(p-stride)) + (*(p-stride+1))
+                                      +2*(*(p     -1)) + 4*(*p         ) + 2*(*(p     +1))
+                                      +(*(p+stride-1)) + 2*(*(p+stride)) + (*(p+stride+1));
+                                f= (f + 8)>>4;
 
 #ifdef DEBUG_DERING_THRESHOLD
-				asm volatile("emms\n\t":);
-				{
-				static long long numPixels=0;
-				if(x!=1 && x!=8 && y!=1 && y!=8) numPixels++;
-//				if((max-min)<20 || (max-min)*QP<200)
-//				if((max-min)*QP < 500)
-//				if(max-min<QP/2)
-				if(max-min < 20)
-				{
-					static int numSkiped=0;
-					static int errorSum=0;
-					static int worstQP=0;
-					static int worstRange=0;
-					static int worstDiff=0;
-					int diff= (f - *p);
-					int absDiff= ABS(diff);
-					int error= diff*diff;
-
-					if(x==1 || x==8 || y==1 || y==8) continue;
-
-					numSkiped++;
-					if(absDiff > worstDiff)
-					{
-						worstDiff= absDiff;
-						worstQP= QP;
-						worstRange= max-min;
-					}
-					errorSum+= error;
-
-					if(1024LL*1024LL*1024LL % numSkiped == 0)
-					{
-						printf( "sum:%1.3f, skip:%d, wQP:%d, "
-							"wRange:%d, wDiff:%d, relSkip:%1.3f\n",
-							(float)errorSum/numSkiped, numSkiped, worstQP, worstRange,
-							worstDiff, (float)numSkiped/numPixels);
-					}
-				}
-				}
+                                asm volatile("emms\n\t":);
+                                {
+                                static long long numPixels=0;
+                                if(x!=1 && x!=8 && y!=1 && y!=8) numPixels++;
+//                                if((max-min)<20 || (max-min)*QP<200)
+//                                if((max-min)*QP < 500)
+//                                if(max-min<QP/2)
+                                if(max-min < 20)
+                                {
+                                        static int numSkiped=0;
+                                        static int errorSum=0;
+                                        static int worstQP=0;
+                                        static int worstRange=0;
+                                        static int worstDiff=0;
+                                        int diff= (f - *p);
+                                        int absDiff= ABS(diff);
+                                        int error= diff*diff;
+
+                                        if(x==1 || x==8 || y==1 || y==8) continue;
+
+                                        numSkiped++;
+                                        if(absDiff > worstDiff)
+                                        {
+                                                worstDiff= absDiff;
+                                                worstQP= QP;
+                                                worstRange= max-min;
+                                        }
+                                        errorSum+= error;
+
+                                        if(1024LL*1024LL*1024LL % numSkiped == 0)
+                                        {
+                                                printf( "sum:%1.3f, skip:%d, wQP:%d, "
+                                                        "wRange:%d, wDiff:%d, relSkip:%1.3f\n",
+                                                        (float)errorSum/numSkiped, numSkiped, worstQP, worstRange,
+                                                        worstDiff, (float)numSkiped/numPixels);
+                                        }
+                                }
+                                }
 #endif
-				if     (*p + QP2 < f) *p= *p + QP2;
-				else if(*p - QP2 > f) *p= *p - QP2;
-				else *p=f;
-			}
-		}
-	}
+                                if     (*p + QP2 < f) *p= *p + QP2;
+                                else if(*p - QP2 > f) *p= *p - QP2;
+                                else *p=f;
+                        }
+                }
+        }
 #ifdef DEBUG_DERING_THRESHOLD
-	if(max-min < 20)
-	{
-		for(y=1; y<9; y++)
-		{
-			int x;
-			int t = 0;
-			p= src + stride*y;
-			for(x=1; x<9; x++)
-			{
-				p++;
-				*p = MIN(*p + 20, 255);
-			}
-		}
-//		src[0] = src[7]=src[stride*7]=src[stride*7 + 7]=255;
-	}
+        if(max-min < 20)
+        {
+                for(y=1; y<9; y++)
+                {
+                        int x;
+                        int t = 0;
+                        p= src + stride*y;
+                        for(x=1; x<9; x++)
+                        {
+                                p++;
+                                *p = MIN(*p + 20, 255);
+                        }
+                }
+//                src[0] = src[7]=src[stride*7]=src[stride*7 + 7]=255;
+        }
 #endif
 #endif //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 }
@@ -1549,46 +1549,46 @@ DERING_CORE((%0, %1, 8),(%%REGd, %1, 4) ,%%mm2,%%mm4,%%mm0,%%mm3,%%mm5,%%mm1,%%m
 static inline void RENAME(deInterlaceInterpolateLinear)(uint8_t src[], int stride)
 {
 #if defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	src+= 4*stride;
-	asm volatile(
-		"lea (%0, %1), %%"REG_a"			\n\t"
-		"lea (%%"REG_a", %1, 4), %%"REG_c"		\n\t"
-//	0	1	2	3	4	5	6	7	8	9
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	ecx	ecx+%1	ecx+2%1	%0+8%1	ecx+4%1
-
-		"movq (%0), %%mm0				\n\t"
-		"movq (%%"REG_a", %1), %%mm1			\n\t"
-		PAVGB(%%mm1, %%mm0)
-		"movq %%mm0, (%%"REG_a")			\n\t"
-		"movq (%0, %1, 4), %%mm0			\n\t"
-		PAVGB(%%mm0, %%mm1)
-		"movq %%mm1, (%%"REG_a", %1, 2)			\n\t"
-		"movq (%%"REG_c", %1), %%mm1			\n\t"
-		PAVGB(%%mm1, %%mm0)
-		"movq %%mm0, (%%"REG_c")			\n\t"
-		"movq (%0, %1, 8), %%mm0			\n\t"
-		PAVGB(%%mm0, %%mm1)
-		"movq %%mm1, (%%"REG_c", %1, 2)			\n\t"
-
-		: : "r" (src), "r" ((long)stride)
-		: "%"REG_a, "%"REG_c
-	);
+        src+= 4*stride;
+        asm volatile(
+                "lea (%0, %1), %%"REG_a"                \n\t"
+                "lea (%%"REG_a", %1, 4), %%"REG_c"      \n\t"
+//      0       1       2       3       4       5       6       7       8       9
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  ecx     ecx+%1  ecx+2%1 %0+8%1  ecx+4%1
+
+                "movq (%0), %%mm0                       \n\t"
+                "movq (%%"REG_a", %1), %%mm1            \n\t"
+                PAVGB(%%mm1, %%mm0)
+                "movq %%mm0, (%%"REG_a")                \n\t"
+                "movq (%0, %1, 4), %%mm0                \n\t"
+                PAVGB(%%mm0, %%mm1)
+                "movq %%mm1, (%%"REG_a", %1, 2)         \n\t"
+                "movq (%%"REG_c", %1), %%mm1            \n\t"
+                PAVGB(%%mm1, %%mm0)
+                "movq %%mm0, (%%"REG_c")                \n\t"
+                "movq (%0, %1, 8), %%mm0                \n\t"
+                PAVGB(%%mm0, %%mm1)
+                "movq %%mm1, (%%"REG_c", %1, 2)         \n\t"
+
+                : : "r" (src), "r" ((long)stride)
+                : "%"REG_a, "%"REG_c
+        );
 #else
-	int a, b, x;
-	src+= 4*stride;
-
-	for(x=0; x<2; x++){
-		a= *(uint32_t*)&src[stride*0];
-		b= *(uint32_t*)&src[stride*2];
-		*(uint32_t*)&src[stride*1]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
-		a= *(uint32_t*)&src[stride*4];
-		*(uint32_t*)&src[stride*3]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
-		b= *(uint32_t*)&src[stride*6];
-		*(uint32_t*)&src[stride*5]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
-		a= *(uint32_t*)&src[stride*8];
-		*(uint32_t*)&src[stride*7]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
-		src += 4;
-	}
+        int a, b, x;
+        src+= 4*stride;
+
+        for(x=0; x<2; x++){
+                a= *(uint32_t*)&src[stride*0];
+                b= *(uint32_t*)&src[stride*2];
+                *(uint32_t*)&src[stride*1]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
+                a= *(uint32_t*)&src[stride*4];
+                *(uint32_t*)&src[stride*3]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
+                b= *(uint32_t*)&src[stride*6];
+                *(uint32_t*)&src[stride*5]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
+                a= *(uint32_t*)&src[stride*8];
+                *(uint32_t*)&src[stride*7]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
+                src += 4;
+        }
 #endif
 }
 
@@ -1602,58 +1602,58 @@ static inline void RENAME(deInterlaceInterpolateLinear)(uint8_t src[], int strid
 static inline void RENAME(deInterlaceInterpolateCubic)(uint8_t src[], int stride)
 {
 #if defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	src+= stride*3;
-	asm volatile(
-		"lea (%0, %1), %%"REG_a"			\n\t"
-		"lea (%%"REG_a", %1, 4), %%"REG_d"		\n\t"
-		"lea (%%"REG_d", %1, 4), %%"REG_c"		\n\t"
-		"add %1, %%"REG_c"				\n\t"
-		"pxor %%mm7, %%mm7				\n\t"
-//	0	1	2	3	4	5	6	7	8	9	10
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	edx	edx+%1	edx+2%1	%0+8%1	edx+4%1 ecx
+        src+= stride*3;
+        asm volatile(
+                "lea (%0, %1), %%"REG_a"                \n\t"
+                "lea (%%"REG_a", %1, 4), %%"REG_d"      \n\t"
+                "lea (%%"REG_d", %1, 4), %%"REG_c"      \n\t"
+                "add %1, %%"REG_c"                      \n\t"
+                "pxor %%mm7, %%mm7                      \n\t"
+//      0       1       2       3       4       5       6       7       8       9       10
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  edx     edx+%1  edx+2%1 %0+8%1  edx+4%1 ecx
 
 #define REAL_DEINT_CUBIC(a,b,c,d,e)\
-		"movq " #a ", %%mm0				\n\t"\
-		"movq " #b ", %%mm1				\n\t"\
-		"movq " #d ", %%mm2				\n\t"\
-		"movq " #e ", %%mm3				\n\t"\
-		PAVGB(%%mm2, %%mm1)					/* (b+d) /2 */\
-		PAVGB(%%mm3, %%mm0)					/* a(a+e) /2 */\
-		"movq %%mm0, %%mm2				\n\t"\
-		"punpcklbw %%mm7, %%mm0				\n\t"\
-		"punpckhbw %%mm7, %%mm2				\n\t"\
-		"movq %%mm1, %%mm3				\n\t"\
-		"punpcklbw %%mm7, %%mm1				\n\t"\
-		"punpckhbw %%mm7, %%mm3				\n\t"\
-		"psubw %%mm1, %%mm0				\n\t"	/* L(a+e - (b+d))/2 */\
-		"psubw %%mm3, %%mm2				\n\t"	/* H(a+e - (b+d))/2 */\
-		"psraw $3, %%mm0				\n\t"	/* L(a+e - (b+d))/16 */\
-		"psraw $3, %%mm2				\n\t"	/* H(a+e - (b+d))/16 */\
-		"psubw %%mm0, %%mm1				\n\t"	/* L(9b + 9d - a - e)/16 */\
-		"psubw %%mm2, %%mm3				\n\t"	/* H(9b + 9d - a - e)/16 */\
-		"packuswb %%mm3, %%mm1				\n\t"\
-		"movq %%mm1, " #c "				\n\t"
+                "movq " #a ", %%mm0                     \n\t"\
+                "movq " #b ", %%mm1                     \n\t"\
+                "movq " #d ", %%mm2                     \n\t"\
+                "movq " #e ", %%mm3                     \n\t"\
+                PAVGB(%%mm2, %%mm1)                             /* (b+d) /2 */\
+                PAVGB(%%mm3, %%mm0)                             /* a(a+e) /2 */\
+                "movq %%mm0, %%mm2                      \n\t"\
+                "punpcklbw %%mm7, %%mm0                 \n\t"\
+                "punpckhbw %%mm7, %%mm2                 \n\t"\
+                "movq %%mm1, %%mm3                      \n\t"\
+                "punpcklbw %%mm7, %%mm1                 \n\t"\
+                "punpckhbw %%mm7, %%mm3                 \n\t"\
+                "psubw %%mm1, %%mm0                     \n\t"   /* L(a+e - (b+d))/2 */\
+                "psubw %%mm3, %%mm2                     \n\t"   /* H(a+e - (b+d))/2 */\
+                "psraw $3, %%mm0                        \n\t"   /* L(a+e - (b+d))/16 */\
+                "psraw $3, %%mm2                        \n\t"   /* H(a+e - (b+d))/16 */\
+                "psubw %%mm0, %%mm1                     \n\t"   /* L(9b + 9d - a - e)/16 */\
+                "psubw %%mm2, %%mm3                     \n\t"   /* H(9b + 9d - a - e)/16 */\
+                "packuswb %%mm3, %%mm1                  \n\t"\
+                "movq %%mm1, " #c "                     \n\t"
 #define DEINT_CUBIC(a,b,c,d,e)  REAL_DEINT_CUBIC(a,b,c,d,e)
 
-DEINT_CUBIC((%0), (%%REGa, %1), (%%REGa, %1, 2), (%0, %1, 4), (%%REGd, %1))
-DEINT_CUBIC((%%REGa, %1), (%0, %1, 4), (%%REGd), (%%REGd, %1), (%0, %1, 8))
-DEINT_CUBIC((%0, %1, 4), (%%REGd, %1), (%%REGd, %1, 2), (%0, %1, 8), (%%REGc))
-DEINT_CUBIC((%%REGd, %1), (%0, %1, 8), (%%REGd, %1, 4), (%%REGc), (%%REGc, %1, 2))
+DEINT_CUBIC((%0)        , (%%REGa, %1), (%%REGa, %1, 2), (%0, %1, 4) , (%%REGd, %1))
+DEINT_CUBIC((%%REGa, %1), (%0, %1, 4) , (%%REGd)       , (%%REGd, %1), (%0, %1, 8))
+DEINT_CUBIC((%0, %1, 4) , (%%REGd, %1), (%%REGd, %1, 2), (%0, %1, 8) , (%%REGc))
+DEINT_CUBIC((%%REGd, %1), (%0, %1, 8) , (%%REGd, %1, 4), (%%REGc)    , (%%REGc, %1, 2))
 
-		: : "r" (src), "r" ((long)stride)
-		: "%"REG_a, "%"REG_d, "%"REG_c
-	);
+                : : "r" (src), "r" ((long)stride)
+                : "%"REG_a, "%"REG_d, "%"REG_c
+        );
 #else //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	int x;
-	src+= stride*3;
-	for(x=0; x<8; x++)
-	{
-		src[stride*3] = CLIP((-src[0]        + 9*src[stride*2] + 9*src[stride*4] - src[stride*6])>>4);
-		src[stride*5] = CLIP((-src[stride*2] + 9*src[stride*4] + 9*src[stride*6] - src[stride*8])>>4);
-		src[stride*7] = CLIP((-src[stride*4] + 9*src[stride*6] + 9*src[stride*8] - src[stride*10])>>4);
-		src[stride*9] = CLIP((-src[stride*6] + 9*src[stride*8] + 9*src[stride*10] - src[stride*12])>>4);
-		src++;
-	}
+        int x;
+        src+= stride*3;
+        for(x=0; x<8; x++)
+        {
+                src[stride*3] = CLIP((-src[0]        + 9*src[stride*2] + 9*src[stride*4] - src[stride*6])>>4);
+                src[stride*5] = CLIP((-src[stride*2] + 9*src[stride*4] + 9*src[stride*6] - src[stride*8])>>4);
+                src[stride*7] = CLIP((-src[stride*4] + 9*src[stride*6] + 9*src[stride*8] - src[stride*10])>>4);
+                src[stride*9] = CLIP((-src[stride*6] + 9*src[stride*8] + 9*src[stride*10] - src[stride*12])>>4);
+                src++;
+        }
 #endif //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 }
 
@@ -1667,73 +1667,73 @@ DEINT_CUBIC((%%REGd, %1), (%0, %1, 8), (%%REGd, %1, 4), (%%REGc), (%%REGc, %1, 2
 static inline void RENAME(deInterlaceFF)(uint8_t src[], int stride, uint8_t *tmp)
 {
 #if defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	src+= stride*4;
-	asm volatile(
-		"lea (%0, %1), %%"REG_a"			\n\t"
-		"lea (%%"REG_a", %1, 4), %%"REG_d"		\n\t"
-		"pxor %%mm7, %%mm7				\n\t"
-		"movq (%2), %%mm0				\n\t"
-//	0	1	2	3	4	5	6	7	8	9	10
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	edx	edx+%1	edx+2%1	%0+8%1	edx+4%1 ecx
+        src+= stride*4;
+        asm volatile(
+                "lea (%0, %1), %%"REG_a"                \n\t"
+                "lea (%%"REG_a", %1, 4), %%"REG_d"      \n\t"
+                "pxor %%mm7, %%mm7                      \n\t"
+                "movq (%2), %%mm0                       \n\t"
+//      0       1       2       3       4       5       6       7       8       9       10
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  edx     edx+%1  edx+2%1 %0+8%1  edx+4%1 ecx
 
 #define REAL_DEINT_FF(a,b,c,d)\
-		"movq " #a ", %%mm1				\n\t"\
-		"movq " #b ", %%mm2				\n\t"\
-		"movq " #c ", %%mm3				\n\t"\
-		"movq " #d ", %%mm4				\n\t"\
-		PAVGB(%%mm3, %%mm1)					\
-		PAVGB(%%mm4, %%mm0)					\
-		"movq %%mm0, %%mm3				\n\t"\
-		"punpcklbw %%mm7, %%mm0				\n\t"\
-		"punpckhbw %%mm7, %%mm3				\n\t"\
-		"movq %%mm1, %%mm4				\n\t"\
-		"punpcklbw %%mm7, %%mm1				\n\t"\
-		"punpckhbw %%mm7, %%mm4				\n\t"\
-		"psllw $2, %%mm1				\n\t"\
-		"psllw $2, %%mm4				\n\t"\
-		"psubw %%mm0, %%mm1				\n\t"\
-		"psubw %%mm3, %%mm4				\n\t"\
-		"movq %%mm2, %%mm5				\n\t"\
-		"movq %%mm2, %%mm0				\n\t"\
-		"punpcklbw %%mm7, %%mm2				\n\t"\
-		"punpckhbw %%mm7, %%mm5				\n\t"\
-		"paddw %%mm2, %%mm1				\n\t"\
-		"paddw %%mm5, %%mm4				\n\t"\
-		"psraw $2, %%mm1				\n\t"\
-		"psraw $2, %%mm4				\n\t"\
-		"packuswb %%mm4, %%mm1				\n\t"\
-		"movq %%mm1, " #b "				\n\t"\
+                "movq " #a ", %%mm1                     \n\t"\
+                "movq " #b ", %%mm2                     \n\t"\
+                "movq " #c ", %%mm3                     \n\t"\
+                "movq " #d ", %%mm4                     \n\t"\
+                PAVGB(%%mm3, %%mm1)                          \
+                PAVGB(%%mm4, %%mm0)                          \
+                "movq %%mm0, %%mm3                      \n\t"\
+                "punpcklbw %%mm7, %%mm0                 \n\t"\
+                "punpckhbw %%mm7, %%mm3                 \n\t"\
+                "movq %%mm1, %%mm4                      \n\t"\
+                "punpcklbw %%mm7, %%mm1                 \n\t"\
+                "punpckhbw %%mm7, %%mm4                 \n\t"\
+                "psllw $2, %%mm1                        \n\t"\
+                "psllw $2, %%mm4                        \n\t"\
+                "psubw %%mm0, %%mm1                     \n\t"\
+                "psubw %%mm3, %%mm4                     \n\t"\
+                "movq %%mm2, %%mm5                      \n\t"\
+                "movq %%mm2, %%mm0                      \n\t"\
+                "punpcklbw %%mm7, %%mm2                 \n\t"\
+                "punpckhbw %%mm7, %%mm5                 \n\t"\
+                "paddw %%mm2, %%mm1                     \n\t"\
+                "paddw %%mm5, %%mm4                     \n\t"\
+                "psraw $2, %%mm1                        \n\t"\
+                "psraw $2, %%mm4                        \n\t"\
+                "packuswb %%mm4, %%mm1                  \n\t"\
+                "movq %%mm1, " #b "                     \n\t"\
 
 #define DEINT_FF(a,b,c,d)  REAL_DEINT_FF(a,b,c,d)
 
-DEINT_FF((%0)       ,  (%%REGa)       , (%%REGa, %1), (%%REGa, %1, 2))
-DEINT_FF((%%REGa, %1), (%%REGa, %1, 2), (%0, %1, 4),  (%%REGd)       )
-DEINT_FF((%0, %1, 4),  (%%REGd)       , (%%REGd, %1), (%%REGd, %1, 2))
-DEINT_FF((%%REGd, %1), (%%REGd, %1, 2), (%0, %1, 8),  (%%REGd, %1, 4))
+DEINT_FF((%0)        , (%%REGa)       , (%%REGa, %1), (%%REGa, %1, 2))
+DEINT_FF((%%REGa, %1), (%%REGa, %1, 2), (%0, %1, 4) , (%%REGd)       )
+DEINT_FF((%0, %1, 4) , (%%REGd)       , (%%REGd, %1), (%%REGd, %1, 2))
+DEINT_FF((%%REGd, %1), (%%REGd, %1, 2), (%0, %1, 8) , (%%REGd, %1, 4))
 
-		"movq %%mm0, (%2)				\n\t"
-		: : "r" (src), "r" ((long)stride), "r"(tmp)
-		: "%"REG_a, "%"REG_d
-	);
+                "movq %%mm0, (%2)                       \n\t"
+                : : "r" (src), "r" ((long)stride), "r"(tmp)
+                : "%"REG_a, "%"REG_d
+        );
 #else //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	int x;
-	src+= stride*4;
-	for(x=0; x<8; x++)
-	{
-		int t1= tmp[x];
-		int t2= src[stride*1];
-
-		src[stride*1]= CLIP((-t1 + 4*src[stride*0] + 2*t2 + 4*src[stride*2] - src[stride*3] + 4)>>3);
-		t1= src[stride*4];
-		src[stride*3]= CLIP((-t2 + 4*src[stride*2] + 2*t1 + 4*src[stride*4] - src[stride*5] + 4)>>3);
-		t2= src[stride*6];
-		src[stride*5]= CLIP((-t1 + 4*src[stride*4] + 2*t2 + 4*src[stride*6] - src[stride*7] + 4)>>3);
-		t1= src[stride*8];
-		src[stride*7]= CLIP((-t2 + 4*src[stride*6] + 2*t1 + 4*src[stride*8] - src[stride*9] + 4)>>3);
-		tmp[x]= t1;
-
-		src++;
-	}
+        int x;
+        src+= stride*4;
+        for(x=0; x<8; x++)
+        {
+                int t1= tmp[x];
+                int t2= src[stride*1];
+
+                src[stride*1]= CLIP((-t1 + 4*src[stride*0] + 2*t2 + 4*src[stride*2] - src[stride*3] + 4)>>3);
+                t1= src[stride*4];
+                src[stride*3]= CLIP((-t2 + 4*src[stride*2] + 2*t1 + 4*src[stride*4] - src[stride*5] + 4)>>3);
+                t2= src[stride*6];
+                src[stride*5]= CLIP((-t1 + 4*src[stride*4] + 2*t2 + 4*src[stride*6] - src[stride*7] + 4)>>3);
+                t1= src[stride*8];
+                src[stride*7]= CLIP((-t2 + 4*src[stride*6] + 2*t1 + 4*src[stride*8] - src[stride*9] + 4)>>3);
+                tmp[x]= t1;
+
+                src++;
+        }
 #endif //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 }
 
@@ -1747,48 +1747,48 @@ DEINT_FF((%%REGd, %1), (%%REGd, %1, 2), (%0, %1, 8),  (%%REGd, %1, 4))
 static inline void RENAME(deInterlaceL5)(uint8_t src[], int stride, uint8_t *tmp, uint8_t *tmp2)
 {
 #if defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	src+= stride*4;
-	asm volatile(
-		"lea (%0, %1), %%"REG_a"			\n\t"
-		"lea (%%"REG_a", %1, 4), %%"REG_d"		\n\t"
-		"pxor %%mm7, %%mm7				\n\t"
-		"movq (%2), %%mm0				\n\t"
-		"movq (%3), %%mm1				\n\t"
-//	0	1	2	3	4	5	6	7	8	9	10
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	edx	edx+%1	edx+2%1	%0+8%1	edx+4%1 ecx
+        src+= stride*4;
+        asm volatile(
+                "lea (%0, %1), %%"REG_a"                \n\t"
+                "lea (%%"REG_a", %1, 4), %%"REG_d"      \n\t"
+                "pxor %%mm7, %%mm7                      \n\t"
+                "movq (%2), %%mm0                       \n\t"
+                "movq (%3), %%mm1                       \n\t"
+//      0       1       2       3       4       5       6       7       8       9       10
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  edx     edx+%1  edx+2%1 %0+8%1  edx+4%1 ecx
 
 #define REAL_DEINT_L5(t1,t2,a,b,c)\
-		"movq " #a ", %%mm2				\n\t"\
-		"movq " #b ", %%mm3				\n\t"\
-		"movq " #c ", %%mm4				\n\t"\
-		PAVGB(t2, %%mm3)					\
-		PAVGB(t1, %%mm4)					\
-		"movq %%mm2, %%mm5				\n\t"\
-		"movq %%mm2, " #t1 "				\n\t"\
-		"punpcklbw %%mm7, %%mm2				\n\t"\
-		"punpckhbw %%mm7, %%mm5				\n\t"\
-		"movq %%mm2, %%mm6				\n\t"\
-		"paddw %%mm2, %%mm2				\n\t"\
-		"paddw %%mm6, %%mm2				\n\t"\
-		"movq %%mm5, %%mm6				\n\t"\
-		"paddw %%mm5, %%mm5				\n\t"\
-		"paddw %%mm6, %%mm5				\n\t"\
-		"movq %%mm3, %%mm6				\n\t"\
-		"punpcklbw %%mm7, %%mm3				\n\t"\
-		"punpckhbw %%mm7, %%mm6				\n\t"\
-		"paddw %%mm3, %%mm3				\n\t"\
-		"paddw %%mm6, %%mm6				\n\t"\
-		"paddw %%mm3, %%mm2				\n\t"\
-		"paddw %%mm6, %%mm5				\n\t"\
-		"movq %%mm4, %%mm6				\n\t"\
-		"punpcklbw %%mm7, %%mm4				\n\t"\
-		"punpckhbw %%mm7, %%mm6				\n\t"\
-		"psubw %%mm4, %%mm2				\n\t"\
-		"psubw %%mm6, %%mm5				\n\t"\
-		"psraw $2, %%mm2				\n\t"\
-		"psraw $2, %%mm5				\n\t"\
-		"packuswb %%mm5, %%mm2				\n\t"\
-		"movq %%mm2, " #a "				\n\t"\
+                "movq " #a ", %%mm2                     \n\t"\
+                "movq " #b ", %%mm3                     \n\t"\
+                "movq " #c ", %%mm4                     \n\t"\
+                PAVGB(t2, %%mm3)                             \
+                PAVGB(t1, %%mm4)                             \
+                "movq %%mm2, %%mm5                      \n\t"\
+                "movq %%mm2, " #t1 "                    \n\t"\
+                "punpcklbw %%mm7, %%mm2                 \n\t"\
+                "punpckhbw %%mm7, %%mm5                 \n\t"\
+                "movq %%mm2, %%mm6                      \n\t"\
+                "paddw %%mm2, %%mm2                     \n\t"\
+                "paddw %%mm6, %%mm2                     \n\t"\
+                "movq %%mm5, %%mm6                      \n\t"\
+                "paddw %%mm5, %%mm5                     \n\t"\
+                "paddw %%mm6, %%mm5                     \n\t"\
+                "movq %%mm3, %%mm6                      \n\t"\
+                "punpcklbw %%mm7, %%mm3                 \n\t"\
+                "punpckhbw %%mm7, %%mm6                 \n\t"\
+                "paddw %%mm3, %%mm3                     \n\t"\
+                "paddw %%mm6, %%mm6                     \n\t"\
+                "paddw %%mm3, %%mm2                     \n\t"\
+                "paddw %%mm6, %%mm5                     \n\t"\
+                "movq %%mm4, %%mm6                      \n\t"\
+                "punpcklbw %%mm7, %%mm4                 \n\t"\
+                "punpckhbw %%mm7, %%mm6                 \n\t"\
+                "psubw %%mm4, %%mm2                     \n\t"\
+                "psubw %%mm6, %%mm5                     \n\t"\
+                "psraw $2, %%mm2                        \n\t"\
+                "psraw $2, %%mm5                        \n\t"\
+                "packuswb %%mm5, %%mm2                  \n\t"\
+                "movq %%mm2, " #a "                     \n\t"\
 
 #define DEINT_L5(t1,t2,a,b,c)  REAL_DEINT_L5(t1,t2,a,b,c)
 
@@ -1801,41 +1801,41 @@ DEINT_L5(%%mm1, %%mm0, (%%REGd)       , (%%REGd, %1)   , (%%REGd, %1, 2))
 DEINT_L5(%%mm0, %%mm1, (%%REGd, %1)   , (%%REGd, %1, 2), (%0, %1, 8)   )
 DEINT_L5(%%mm1, %%mm0, (%%REGd, %1, 2), (%0, %1, 8)    , (%%REGd, %1, 4))
 
-		"movq %%mm0, (%2)				\n\t"
-		"movq %%mm1, (%3)				\n\t"
-		: : "r" (src), "r" ((long)stride), "r"(tmp), "r"(tmp2)
-		: "%"REG_a, "%"REG_d
-	);
+                "movq %%mm0, (%2)                       \n\t"
+                "movq %%mm1, (%3)                       \n\t"
+                : : "r" (src), "r" ((long)stride), "r"(tmp), "r"(tmp2)
+                : "%"REG_a, "%"REG_d
+        );
 #else //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	int x;
-	src+= stride*4;
-	for(x=0; x<8; x++)
-	{
-		int t1= tmp[x];
-		int t2= tmp2[x];
-		int t3= src[0];
-
-		src[stride*0]= CLIP((-(t1 + src[stride*2]) + 2*(t2 + src[stride*1]) + 6*t3 + 4)>>3);
-		t1= src[stride*1];
-		src[stride*1]= CLIP((-(t2 + src[stride*3]) + 2*(t3 + src[stride*2]) + 6*t1 + 4)>>3);
-		t2= src[stride*2];
-		src[stride*2]= CLIP((-(t3 + src[stride*4]) + 2*(t1 + src[stride*3]) + 6*t2 + 4)>>3);
-		t3= src[stride*3];
-		src[stride*3]= CLIP((-(t1 + src[stride*5]) + 2*(t2 + src[stride*4]) + 6*t3 + 4)>>3);
-		t1= src[stride*4];
-		src[stride*4]= CLIP((-(t2 + src[stride*6]) + 2*(t3 + src[stride*5]) + 6*t1 + 4)>>3);
-		t2= src[stride*5];
-		src[stride*5]= CLIP((-(t3 + src[stride*7]) + 2*(t1 + src[stride*6]) + 6*t2 + 4)>>3);
-		t3= src[stride*6];
-		src[stride*6]= CLIP((-(t1 + src[stride*8]) + 2*(t2 + src[stride*7]) + 6*t3 + 4)>>3);
-		t1= src[stride*7];
-		src[stride*7]= CLIP((-(t2 + src[stride*9]) + 2*(t3 + src[stride*8]) + 6*t1 + 4)>>3);
-
-		tmp[x]= t3;
-		tmp2[x]= t1;
-
-		src++;
-	}
+        int x;
+        src+= stride*4;
+        for(x=0; x<8; x++)
+        {
+                int t1= tmp[x];
+                int t2= tmp2[x];
+                int t3= src[0];
+
+                src[stride*0]= CLIP((-(t1 + src[stride*2]) + 2*(t2 + src[stride*1]) + 6*t3 + 4)>>3);
+                t1= src[stride*1];
+                src[stride*1]= CLIP((-(t2 + src[stride*3]) + 2*(t3 + src[stride*2]) + 6*t1 + 4)>>3);
+                t2= src[stride*2];
+                src[stride*2]= CLIP((-(t3 + src[stride*4]) + 2*(t1 + src[stride*3]) + 6*t2 + 4)>>3);
+                t3= src[stride*3];
+                src[stride*3]= CLIP((-(t1 + src[stride*5]) + 2*(t2 + src[stride*4]) + 6*t3 + 4)>>3);
+                t1= src[stride*4];
+                src[stride*4]= CLIP((-(t2 + src[stride*6]) + 2*(t3 + src[stride*5]) + 6*t1 + 4)>>3);
+                t2= src[stride*5];
+                src[stride*5]= CLIP((-(t3 + src[stride*7]) + 2*(t1 + src[stride*6]) + 6*t2 + 4)>>3);
+                t3= src[stride*6];
+                src[stride*6]= CLIP((-(t1 + src[stride*8]) + 2*(t2 + src[stride*7]) + 6*t3 + 4)>>3);
+                t1= src[stride*7];
+                src[stride*7]= CLIP((-(t2 + src[stride*9]) + 2*(t3 + src[stride*8]) + 6*t1 + 4)>>3);
+
+                tmp[x]= t3;
+                tmp2[x]= t1;
+
+                src++;
+        }
 #endif //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 }
 
@@ -1849,95 +1849,95 @@ DEINT_L5(%%mm1, %%mm0, (%%REGd, %1, 2), (%0, %1, 8)    , (%%REGd, %1, 4))
 static inline void RENAME(deInterlaceBlendLinear)(uint8_t src[], int stride, uint8_t *tmp)
 {
 #if defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	src+= 4*stride;
-	asm volatile(
-		"lea (%0, %1), %%"REG_a"			\n\t"
-		"lea (%%"REG_a", %1, 4), %%"REG_d"		\n\t"
-//	0	1	2	3	4	5	6	7	8	9
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	edx	edx+%1	edx+2%1	%0+8%1	edx+4%1
-
-		"movq (%2), %%mm0				\n\t" // L0
-		"movq (%%"REG_a"), %%mm1			\n\t" // L2
-		PAVGB(%%mm1, %%mm0)				      // L0+L2
-		"movq (%0), %%mm2				\n\t" // L1
-		PAVGB(%%mm2, %%mm0)
-		"movq %%mm0, (%0)				\n\t"
-		"movq (%%"REG_a", %1), %%mm0			\n\t" // L3
-		PAVGB(%%mm0, %%mm2)				      // L1+L3
-		PAVGB(%%mm1, %%mm2)				      // 2L2 + L1 + L3
-		"movq %%mm2, (%%"REG_a")			\n\t"
-		"movq (%%"REG_a", %1, 2), %%mm2			\n\t" // L4
-		PAVGB(%%mm2, %%mm1)				      // L2+L4
-		PAVGB(%%mm0, %%mm1)				      // 2L3 + L2 + L4
-		"movq %%mm1, (%%"REG_a", %1)			\n\t"
-		"movq (%0, %1, 4), %%mm1			\n\t" // L5
-		PAVGB(%%mm1, %%mm0)				      // L3+L5
-		PAVGB(%%mm2, %%mm0)				      // 2L4 + L3 + L5
-		"movq %%mm0, (%%"REG_a", %1, 2)			\n\t"
-		"movq (%%"REG_d"), %%mm0			\n\t" // L6
-		PAVGB(%%mm0, %%mm2)				      // L4+L6
-		PAVGB(%%mm1, %%mm2)				      // 2L5 + L4 + L6
-		"movq %%mm2, (%0, %1, 4)			\n\t"
-		"movq (%%"REG_d", %1), %%mm2			\n\t" // L7
-		PAVGB(%%mm2, %%mm1)				      // L5+L7
-		PAVGB(%%mm0, %%mm1)				      // 2L6 + L5 + L7
-		"movq %%mm1, (%%"REG_d")			\n\t"
-		"movq (%%"REG_d", %1, 2), %%mm1			\n\t" // L8
-		PAVGB(%%mm1, %%mm0)				      // L6+L8
-		PAVGB(%%mm2, %%mm0)				      // 2L7 + L6 + L8
-		"movq %%mm0, (%%"REG_d", %1)			\n\t"
-		"movq (%0, %1, 8), %%mm0			\n\t" // L9
-		PAVGB(%%mm0, %%mm2)				      // L7+L9
-		PAVGB(%%mm1, %%mm2)				      // 2L8 + L7 + L9
-		"movq %%mm2, (%%"REG_d", %1, 2)			\n\t"
-		"movq %%mm1, (%2)				\n\t"
-
-		: : "r" (src), "r" ((long)stride), "r" (tmp)
-		: "%"REG_a, "%"REG_d
-	);
+        src+= 4*stride;
+        asm volatile(
+                "lea (%0, %1), %%"REG_a"                \n\t"
+                "lea (%%"REG_a", %1, 4), %%"REG_d"      \n\t"
+//      0       1       2       3       4       5       6       7       8       9
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  edx     edx+%1  edx+2%1 %0+8%1  edx+4%1
+
+                "movq (%2), %%mm0                       \n\t" // L0
+                "movq (%%"REG_a"), %%mm1                \n\t" // L2
+                PAVGB(%%mm1, %%mm0)                           // L0+L2
+                "movq (%0), %%mm2                       \n\t" // L1
+                PAVGB(%%mm2, %%mm0)
+                "movq %%mm0, (%0)                       \n\t"
+                "movq (%%"REG_a", %1), %%mm0            \n\t" // L3
+                PAVGB(%%mm0, %%mm2)                           // L1+L3
+                PAVGB(%%mm1, %%mm2)                           // 2L2 + L1 + L3
+                "movq %%mm2, (%%"REG_a")                \n\t"
+                "movq (%%"REG_a", %1, 2), %%mm2         \n\t" // L4
+                PAVGB(%%mm2, %%mm1)                           // L2+L4
+                PAVGB(%%mm0, %%mm1)                           // 2L3 + L2 + L4
+                "movq %%mm1, (%%"REG_a", %1)            \n\t"
+                "movq (%0, %1, 4), %%mm1                \n\t" // L5
+                PAVGB(%%mm1, %%mm0)                           // L3+L5
+                PAVGB(%%mm2, %%mm0)                           // 2L4 + L3 + L5
+                "movq %%mm0, (%%"REG_a", %1, 2)         \n\t"
+                "movq (%%"REG_d"), %%mm0                \n\t" // L6
+                PAVGB(%%mm0, %%mm2)                           // L4+L6
+                PAVGB(%%mm1, %%mm2)                           // 2L5 + L4 + L6
+                "movq %%mm2, (%0, %1, 4)                \n\t"
+                "movq (%%"REG_d", %1), %%mm2            \n\t" // L7
+                PAVGB(%%mm2, %%mm1)                           // L5+L7
+                PAVGB(%%mm0, %%mm1)                           // 2L6 + L5 + L7
+                "movq %%mm1, (%%"REG_d")                \n\t"
+                "movq (%%"REG_d", %1, 2), %%mm1         \n\t" // L8
+                PAVGB(%%mm1, %%mm0)                           // L6+L8
+                PAVGB(%%mm2, %%mm0)                           // 2L7 + L6 + L8
+                "movq %%mm0, (%%"REG_d", %1)            \n\t"
+                "movq (%0, %1, 8), %%mm0                \n\t" // L9
+                PAVGB(%%mm0, %%mm2)                           // L7+L9
+                PAVGB(%%mm1, %%mm2)                           // 2L8 + L7 + L9
+                "movq %%mm2, (%%"REG_d", %1, 2)         \n\t"
+                "movq %%mm1, (%2)                       \n\t"
+
+                : : "r" (src), "r" ((long)stride), "r" (tmp)
+                : "%"REG_a, "%"REG_d
+        );
 #else //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	int a, b, c, x;
-	src+= 4*stride;
-
-	for(x=0; x<2; x++){
-		a= *(uint32_t*)&tmp[stride*0];
-		b= *(uint32_t*)&src[stride*0];
-		c= *(uint32_t*)&src[stride*1];
-		a= (a&c) + (((a^c)&0xFEFEFEFEUL)>>1);
-		*(uint32_t*)&src[stride*0]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
-
-		a= *(uint32_t*)&src[stride*2];
-		b= (a&b) + (((a^b)&0xFEFEFEFEUL)>>1);
-		*(uint32_t*)&src[stride*1]= (c|b) - (((c^b)&0xFEFEFEFEUL)>>1);
-
-		b= *(uint32_t*)&src[stride*3];
-		c= (b&c) + (((b^c)&0xFEFEFEFEUL)>>1);
-		*(uint32_t*)&src[stride*2]= (c|a) - (((c^a)&0xFEFEFEFEUL)>>1);
-
-		c= *(uint32_t*)&src[stride*4];
-		a= (a&c) + (((a^c)&0xFEFEFEFEUL)>>1);
-		*(uint32_t*)&src[stride*3]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
-
-		a= *(uint32_t*)&src[stride*5];
-		b= (a&b) + (((a^b)&0xFEFEFEFEUL)>>1);
-		*(uint32_t*)&src[stride*4]= (c|b) - (((c^b)&0xFEFEFEFEUL)>>1);
-
-		b= *(uint32_t*)&src[stride*6];
-		c= (b&c) + (((b^c)&0xFEFEFEFEUL)>>1);
-		*(uint32_t*)&src[stride*5]= (c|a) - (((c^a)&0xFEFEFEFEUL)>>1);
-
-		c= *(uint32_t*)&src[stride*7];
-		a= (a&c) + (((a^c)&0xFEFEFEFEUL)>>1);
-		*(uint32_t*)&src[stride*6]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
-
-		a= *(uint32_t*)&src[stride*8];
-		b= (a&b) + (((a^b)&0xFEFEFEFEUL)>>1);
-		*(uint32_t*)&src[stride*7]= (c|b) - (((c^b)&0xFEFEFEFEUL)>>1);
-
-		*(uint32_t*)&tmp[stride*0]= c;
-		src += 4;
-		tmp += 4;
-	}
+        int a, b, c, x;
+        src+= 4*stride;
+
+        for(x=0; x<2; x++){
+                a= *(uint32_t*)&tmp[stride*0];
+                b= *(uint32_t*)&src[stride*0];
+                c= *(uint32_t*)&src[stride*1];
+                a= (a&c) + (((a^c)&0xFEFEFEFEUL)>>1);
+                *(uint32_t*)&src[stride*0]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
+
+                a= *(uint32_t*)&src[stride*2];
+                b= (a&b) + (((a^b)&0xFEFEFEFEUL)>>1);
+                *(uint32_t*)&src[stride*1]= (c|b) - (((c^b)&0xFEFEFEFEUL)>>1);
+
+                b= *(uint32_t*)&src[stride*3];
+                c= (b&c) + (((b^c)&0xFEFEFEFEUL)>>1);
+                *(uint32_t*)&src[stride*2]= (c|a) - (((c^a)&0xFEFEFEFEUL)>>1);
+
+                c= *(uint32_t*)&src[stride*4];
+                a= (a&c) + (((a^c)&0xFEFEFEFEUL)>>1);
+                *(uint32_t*)&src[stride*3]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
+
+                a= *(uint32_t*)&src[stride*5];
+                b= (a&b) + (((a^b)&0xFEFEFEFEUL)>>1);
+                *(uint32_t*)&src[stride*4]= (c|b) - (((c^b)&0xFEFEFEFEUL)>>1);
+
+                b= *(uint32_t*)&src[stride*6];
+                c= (b&c) + (((b^c)&0xFEFEFEFEUL)>>1);
+                *(uint32_t*)&src[stride*5]= (c|a) - (((c^a)&0xFEFEFEFEUL)>>1);
+
+                c= *(uint32_t*)&src[stride*7];
+                a= (a&c) + (((a^c)&0xFEFEFEFEUL)>>1);
+                *(uint32_t*)&src[stride*6]= (a|b) - (((a^b)&0xFEFEFEFEUL)>>1);
+
+                a= *(uint32_t*)&src[stride*8];
+                b= (a&b) + (((a^b)&0xFEFEFEFEUL)>>1);
+                *(uint32_t*)&src[stride*7]= (c|b) - (((c^b)&0xFEFEFEFEUL)>>1);
+
+                *(uint32_t*)&tmp[stride*0]= c;
+                src += 4;
+                tmp += 4;
+        }
 #endif //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 }
 
@@ -1950,119 +1950,119 @@ static inline void RENAME(deInterlaceBlendLinear)(uint8_t src[], int stride, uin
 static inline void RENAME(deInterlaceMedian)(uint8_t src[], int stride)
 {
 #ifdef HAVE_MMX
-	src+= 4*stride;
+        src+= 4*stride;
 #ifdef HAVE_MMX2
-	asm volatile(
-		"lea (%0, %1), %%"REG_a"			\n\t"
-		"lea (%%"REG_a", %1, 4), %%"REG_d"		\n\t"
-//	0	1	2	3	4	5	6	7	8	9
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	edx	edx+%1	edx+2%1	%0+8%1	edx+4%1
-
-		"movq (%0), %%mm0				\n\t" //
-		"movq (%%"REG_a", %1), %%mm2			\n\t" //
-		"movq (%%"REG_a"), %%mm1			\n\t" //
-		"movq %%mm0, %%mm3				\n\t"
-		"pmaxub %%mm1, %%mm0				\n\t" //
-		"pminub %%mm3, %%mm1				\n\t" //
-		"pmaxub %%mm2, %%mm1				\n\t" //
-		"pminub %%mm1, %%mm0				\n\t"
-		"movq %%mm0, (%%"REG_a")			\n\t"
-
-		"movq (%0, %1, 4), %%mm0			\n\t" //
-		"movq (%%"REG_a", %1, 2), %%mm1			\n\t" //
-		"movq %%mm2, %%mm3				\n\t"
-		"pmaxub %%mm1, %%mm2				\n\t" //
-		"pminub %%mm3, %%mm1				\n\t" //
-		"pmaxub %%mm0, %%mm1				\n\t" //
-		"pminub %%mm1, %%mm2				\n\t"
-		"movq %%mm2, (%%"REG_a", %1, 2)			\n\t"
-
-		"movq (%%"REG_d"), %%mm2			\n\t" //
-		"movq (%%"REG_d", %1), %%mm1			\n\t" //
-		"movq %%mm2, %%mm3				\n\t"
-		"pmaxub %%mm0, %%mm2				\n\t" //
-		"pminub %%mm3, %%mm0				\n\t" //
-		"pmaxub %%mm1, %%mm0				\n\t" //
-		"pminub %%mm0, %%mm2				\n\t"
-		"movq %%mm2, (%%"REG_d")			\n\t"
-
-		"movq (%%"REG_d", %1, 2), %%mm2			\n\t" //
-		"movq (%0, %1, 8), %%mm0			\n\t" //
-		"movq %%mm2, %%mm3				\n\t"
-		"pmaxub %%mm0, %%mm2				\n\t" //
-		"pminub %%mm3, %%mm0				\n\t" //
-		"pmaxub %%mm1, %%mm0				\n\t" //
-		"pminub %%mm0, %%mm2				\n\t"
-		"movq %%mm2, (%%"REG_d", %1, 2)			\n\t"
-
-
-		: : "r" (src), "r" ((long)stride)
-		: "%"REG_a, "%"REG_d
-	);
+        asm volatile(
+                "lea (%0, %1), %%"REG_a"                \n\t"
+                "lea (%%"REG_a", %1, 4), %%"REG_d"      \n\t"
+//      0       1       2       3       4       5       6       7       8       9
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  edx     edx+%1  edx+2%1 %0+8%1  edx+4%1
+
+                "movq (%0), %%mm0                       \n\t" //
+                "movq (%%"REG_a", %1), %%mm2            \n\t" //
+                "movq (%%"REG_a"), %%mm1                \n\t" //
+                "movq %%mm0, %%mm3                      \n\t"
+                "pmaxub %%mm1, %%mm0                    \n\t" //
+                "pminub %%mm3, %%mm1                    \n\t" //
+                "pmaxub %%mm2, %%mm1                    \n\t" //
+                "pminub %%mm1, %%mm0                    \n\t"
+                "movq %%mm0, (%%"REG_a")                \n\t"
+
+                "movq (%0, %1, 4), %%mm0                \n\t" //
+                "movq (%%"REG_a", %1, 2), %%mm1         \n\t" //
+                "movq %%mm2, %%mm3                      \n\t"
+                "pmaxub %%mm1, %%mm2                    \n\t" //
+                "pminub %%mm3, %%mm1                    \n\t" //
+                "pmaxub %%mm0, %%mm1                    \n\t" //
+                "pminub %%mm1, %%mm2                    \n\t"
+                "movq %%mm2, (%%"REG_a", %1, 2)         \n\t"
+
+                "movq (%%"REG_d"), %%mm2                \n\t" //
+                "movq (%%"REG_d", %1), %%mm1            \n\t" //
+                "movq %%mm2, %%mm3                      \n\t"
+                "pmaxub %%mm0, %%mm2                    \n\t" //
+                "pminub %%mm3, %%mm0                    \n\t" //
+                "pmaxub %%mm1, %%mm0                    \n\t" //
+                "pminub %%mm0, %%mm2                    \n\t"
+                "movq %%mm2, (%%"REG_d")                \n\t"
+
+                "movq (%%"REG_d", %1, 2), %%mm2         \n\t" //
+                "movq (%0, %1, 8), %%mm0                \n\t" //
+                "movq %%mm2, %%mm3                      \n\t"
+                "pmaxub %%mm0, %%mm2                    \n\t" //
+                "pminub %%mm3, %%mm0                    \n\t" //
+                "pmaxub %%mm1, %%mm0                    \n\t" //
+                "pminub %%mm0, %%mm2                    \n\t"
+                "movq %%mm2, (%%"REG_d", %1, 2)         \n\t"
+
+
+                : : "r" (src), "r" ((long)stride)
+                : "%"REG_a, "%"REG_d
+        );
 
 #else // MMX without MMX2
-	asm volatile(
-		"lea (%0, %1), %%"REG_a"			\n\t"
-		"lea (%%"REG_a", %1, 4), %%"REG_d"		\n\t"
-//	0	1	2	3	4	5	6	7	8	9
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	edx	edx+%1	edx+2%1	%0+8%1	edx+4%1
-		"pxor %%mm7, %%mm7				\n\t"
+        asm volatile(
+                "lea (%0, %1), %%"REG_a"                \n\t"
+                "lea (%%"REG_a", %1, 4), %%"REG_d"      \n\t"
+//      0       1       2       3       4       5       6       7       8       9
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  edx     edx+%1  edx+2%1 %0+8%1  edx+4%1
+                "pxor %%mm7, %%mm7                      \n\t"
 
 #define REAL_MEDIAN(a,b,c)\
-		"movq " #a ", %%mm0				\n\t"\
-		"movq " #b ", %%mm2				\n\t"\
-		"movq " #c ", %%mm1				\n\t"\
-		"movq %%mm0, %%mm3				\n\t"\
-		"movq %%mm1, %%mm4				\n\t"\
-		"movq %%mm2, %%mm5				\n\t"\
-		"psubusb %%mm1, %%mm3				\n\t"\
-		"psubusb %%mm2, %%mm4				\n\t"\
-		"psubusb %%mm0, %%mm5				\n\t"\
-		"pcmpeqb %%mm7, %%mm3				\n\t"\
-		"pcmpeqb %%mm7, %%mm4				\n\t"\
-		"pcmpeqb %%mm7, %%mm5				\n\t"\
-		"movq %%mm3, %%mm6				\n\t"\
-		"pxor %%mm4, %%mm3				\n\t"\
-		"pxor %%mm5, %%mm4				\n\t"\
-		"pxor %%mm6, %%mm5				\n\t"\
-		"por %%mm3, %%mm1				\n\t"\
-		"por %%mm4, %%mm2				\n\t"\
-		"por %%mm5, %%mm0				\n\t"\
-		"pand %%mm2, %%mm0				\n\t"\
-		"pand %%mm1, %%mm0				\n\t"\
-		"movq %%mm0, " #b "				\n\t"
+                "movq " #a ", %%mm0                     \n\t"\
+                "movq " #b ", %%mm2                     \n\t"\
+                "movq " #c ", %%mm1                     \n\t"\
+                "movq %%mm0, %%mm3                      \n\t"\
+                "movq %%mm1, %%mm4                      \n\t"\
+                "movq %%mm2, %%mm5                      \n\t"\
+                "psubusb %%mm1, %%mm3                   \n\t"\
+                "psubusb %%mm2, %%mm4                   \n\t"\
+                "psubusb %%mm0, %%mm5                   \n\t"\
+                "pcmpeqb %%mm7, %%mm3                   \n\t"\
+                "pcmpeqb %%mm7, %%mm4                   \n\t"\
+                "pcmpeqb %%mm7, %%mm5                   \n\t"\
+                "movq %%mm3, %%mm6                      \n\t"\
+                "pxor %%mm4, %%mm3                      \n\t"\
+                "pxor %%mm5, %%mm4                      \n\t"\
+                "pxor %%mm6, %%mm5                      \n\t"\
+                "por %%mm3, %%mm1                       \n\t"\
+                "por %%mm4, %%mm2                       \n\t"\
+                "por %%mm5, %%mm0                       \n\t"\
+                "pand %%mm2, %%mm0                      \n\t"\
+                "pand %%mm1, %%mm0                      \n\t"\
+                "movq %%mm0, " #b "                     \n\t"
 #define MEDIAN(a,b,c)  REAL_MEDIAN(a,b,c)
 
-MEDIAN((%0), (%%REGa), (%%REGa, %1))
+MEDIAN((%0)        , (%%REGa)       , (%%REGa, %1))
 MEDIAN((%%REGa, %1), (%%REGa, %1, 2), (%0, %1, 4))
-MEDIAN((%0, %1, 4), (%%REGd), (%%REGd, %1))
+MEDIAN((%0, %1, 4) , (%%REGd)       , (%%REGd, %1))
 MEDIAN((%%REGd, %1), (%%REGd, %1, 2), (%0, %1, 8))
 
-		: : "r" (src), "r" ((long)stride)
-		: "%"REG_a, "%"REG_d
-	);
+                : : "r" (src), "r" ((long)stride)
+                : "%"REG_a, "%"REG_d
+        );
 #endif //HAVE_MMX2
 #else //HAVE_MMX
-	int x, y;
-	src+= 4*stride;
-	// FIXME - there should be a way to do a few columns in parallel like w/mmx
-	for(x=0; x<8; x++)
-	{
-		uint8_t *colsrc = src;
-		for (y=0; y<4; y++)
-		{
-			int a, b, c, d, e, f;
-			a = colsrc[0       ];
-			b = colsrc[stride  ];
-			c = colsrc[stride*2];
-			d = (a-b)>>31;
-			e = (b-c)>>31;
-			f = (c-a)>>31;
-			colsrc[stride  ] = (a|(d^f)) & (b|(d^e)) & (c|(e^f));
-			colsrc += stride*2;
-		}
-		src++;
-	}
+        int x, y;
+        src+= 4*stride;
+        // FIXME - there should be a way to do a few columns in parallel like w/mmx
+        for(x=0; x<8; x++)
+        {
+                uint8_t *colsrc = src;
+                for (y=0; y<4; y++)
+                {
+                        int a, b, c, d, e, f;
+                        a = colsrc[0       ];
+                        b = colsrc[stride  ];
+                        c = colsrc[stride*2];
+                        d = (a-b)>>31;
+                        e = (b-c)>>31;
+                        f = (c-a)>>31;
+                        colsrc[stride  ] = (a|(d^f)) & (b|(d^e)) & (c|(e^f));
+                        colsrc += stride*2;
+                }
+                src++;
+        }
 #endif //HAVE_MMX
 }
 
@@ -2072,84 +2072,84 @@ MEDIAN((%%REGd, %1), (%%REGd, %1, 2), (%0, %1, 8))
  */
 static inline void RENAME(transpose1)(uint8_t *dst1, uint8_t *dst2, uint8_t *src, int srcStride)
 {
-	asm(
-		"lea (%0, %1), %%"REG_a"	\n\t"
-//	0	1	2	3	4	5	6	7	8	9
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	edx	edx+%1	edx+2%1	%0+8%1	edx+4%1
-		"movq (%0), %%mm0		\n\t" // 12345678
-		"movq (%%"REG_a"), %%mm1	\n\t" // abcdefgh
-		"movq %%mm0, %%mm2		\n\t" // 12345678
-		"punpcklbw %%mm1, %%mm0		\n\t" // 1a2b3c4d
-		"punpckhbw %%mm1, %%mm2		\n\t" // 5e6f7g8h
-
-		"movq (%%"REG_a", %1), %%mm1	\n\t"
-		"movq (%%"REG_a", %1, 2), %%mm3	\n\t"
-		"movq %%mm1, %%mm4		\n\t"
-		"punpcklbw %%mm3, %%mm1		\n\t"
-		"punpckhbw %%mm3, %%mm4		\n\t"
-
-		"movq %%mm0, %%mm3		\n\t"
-		"punpcklwd %%mm1, %%mm0		\n\t"
-		"punpckhwd %%mm1, %%mm3		\n\t"
-		"movq %%mm2, %%mm1		\n\t"
-		"punpcklwd %%mm4, %%mm2		\n\t"
-		"punpckhwd %%mm4, %%mm1		\n\t"
-
-		"movd %%mm0, 128(%2)		\n\t"
-		"psrlq $32, %%mm0		\n\t"
-		"movd %%mm0, 144(%2)		\n\t"
-		"movd %%mm3, 160(%2)		\n\t"
-		"psrlq $32, %%mm3		\n\t"
-		"movd %%mm3, 176(%2)		\n\t"
-		"movd %%mm3, 48(%3)		\n\t"
-		"movd %%mm2, 192(%2)		\n\t"
-		"movd %%mm2, 64(%3)		\n\t"
-		"psrlq $32, %%mm2		\n\t"
-		"movd %%mm2, 80(%3)		\n\t"
-		"movd %%mm1, 96(%3)		\n\t"
-		"psrlq $32, %%mm1		\n\t"
-		"movd %%mm1, 112(%3)		\n\t"
-
-		"lea (%%"REG_a", %1, 4), %%"REG_a"	\n\t"
-
-		"movq (%0, %1, 4), %%mm0	\n\t" // 12345678
-		"movq (%%"REG_a"), %%mm1	\n\t" // abcdefgh
-		"movq %%mm0, %%mm2		\n\t" // 12345678
-		"punpcklbw %%mm1, %%mm0		\n\t" // 1a2b3c4d
-		"punpckhbw %%mm1, %%mm2		\n\t" // 5e6f7g8h
-
-		"movq (%%"REG_a", %1), %%mm1	\n\t"
-		"movq (%%"REG_a", %1, 2), %%mm3	\n\t"
-		"movq %%mm1, %%mm4		\n\t"
-		"punpcklbw %%mm3, %%mm1		\n\t"
-		"punpckhbw %%mm3, %%mm4		\n\t"
-
-		"movq %%mm0, %%mm3		\n\t"
-		"punpcklwd %%mm1, %%mm0		\n\t"
-		"punpckhwd %%mm1, %%mm3		\n\t"
-		"movq %%mm2, %%mm1		\n\t"
-		"punpcklwd %%mm4, %%mm2		\n\t"
-		"punpckhwd %%mm4, %%mm1		\n\t"
-
-		"movd %%mm0, 132(%2)		\n\t"
-		"psrlq $32, %%mm0		\n\t"
-		"movd %%mm0, 148(%2)		\n\t"
-		"movd %%mm3, 164(%2)		\n\t"
-		"psrlq $32, %%mm3		\n\t"
-		"movd %%mm3, 180(%2)		\n\t"
-		"movd %%mm3, 52(%3)		\n\t"
-		"movd %%mm2, 196(%2)		\n\t"
-		"movd %%mm2, 68(%3)		\n\t"
-		"psrlq $32, %%mm2		\n\t"
-		"movd %%mm2, 84(%3)		\n\t"
-		"movd %%mm1, 100(%3)		\n\t"
-		"psrlq $32, %%mm1		\n\t"
-		"movd %%mm1, 116(%3)		\n\t"
-
-
-	:: "r" (src), "r" ((long)srcStride), "r" (dst1), "r" (dst2)
-	: "%"REG_a
-	);
+        asm(
+                "lea (%0, %1), %%"REG_a"                \n\t"
+//      0       1       2       3       4       5       6       7       8       9
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  edx     edx+%1  edx+2%1 %0+8%1  edx+4%1
+                "movq (%0), %%mm0                       \n\t" // 12345678
+                "movq (%%"REG_a"), %%mm1                \n\t" // abcdefgh
+                "movq %%mm0, %%mm2                      \n\t" // 12345678
+                "punpcklbw %%mm1, %%mm0                 \n\t" // 1a2b3c4d
+                "punpckhbw %%mm1, %%mm2                 \n\t" // 5e6f7g8h
+
+                "movq (%%"REG_a", %1), %%mm1            \n\t"
+                "movq (%%"REG_a", %1, 2), %%mm3         \n\t"
+                "movq %%mm1, %%mm4                      \n\t"
+                "punpcklbw %%mm3, %%mm1                 \n\t"
+                "punpckhbw %%mm3, %%mm4                 \n\t"
+
+                "movq %%mm0, %%mm3                      \n\t"
+                "punpcklwd %%mm1, %%mm0                 \n\t"
+                "punpckhwd %%mm1, %%mm3                 \n\t"
+                "movq %%mm2, %%mm1                      \n\t"
+                "punpcklwd %%mm4, %%mm2                 \n\t"
+                "punpckhwd %%mm4, %%mm1                 \n\t"
+
+                "movd %%mm0, 128(%2)                    \n\t"
+                "psrlq $32, %%mm0                       \n\t"
+                "movd %%mm0, 144(%2)                    \n\t"
+                "movd %%mm3, 160(%2)                    \n\t"
+                "psrlq $32, %%mm3                       \n\t"
+                "movd %%mm3, 176(%2)                    \n\t"
+                "movd %%mm3, 48(%3)                     \n\t"
+                "movd %%mm2, 192(%2)                    \n\t"
+                "movd %%mm2, 64(%3)                     \n\t"
+                "psrlq $32, %%mm2                       \n\t"
+                "movd %%mm2, 80(%3)                     \n\t"
+                "movd %%mm1, 96(%3)                     \n\t"
+                "psrlq $32, %%mm1                       \n\t"
+                "movd %%mm1, 112(%3)                    \n\t"
+
+                "lea (%%"REG_a", %1, 4), %%"REG_a"      \n\t"
+
+                "movq (%0, %1, 4), %%mm0                \n\t" // 12345678
+                "movq (%%"REG_a"), %%mm1                \n\t" // abcdefgh
+                "movq %%mm0, %%mm2                      \n\t" // 12345678
+                "punpcklbw %%mm1, %%mm0                 \n\t" // 1a2b3c4d
+                "punpckhbw %%mm1, %%mm2                 \n\t" // 5e6f7g8h
+
+                "movq (%%"REG_a", %1), %%mm1            \n\t"
+                "movq (%%"REG_a", %1, 2), %%mm3         \n\t"
+                "movq %%mm1, %%mm4                      \n\t"
+                "punpcklbw %%mm3, %%mm1                 \n\t"
+                "punpckhbw %%mm3, %%mm4                 \n\t"
+
+                "movq %%mm0, %%mm3                      \n\t"
+                "punpcklwd %%mm1, %%mm0                 \n\t"
+                "punpckhwd %%mm1, %%mm3                 \n\t"
+                "movq %%mm2, %%mm1                      \n\t"
+                "punpcklwd %%mm4, %%mm2                 \n\t"
+                "punpckhwd %%mm4, %%mm1                 \n\t"
+
+                "movd %%mm0, 132(%2)                    \n\t"
+                "psrlq $32, %%mm0                       \n\t"
+                "movd %%mm0, 148(%2)                    \n\t"
+                "movd %%mm3, 164(%2)                    \n\t"
+                "psrlq $32, %%mm3                       \n\t"
+                "movd %%mm3, 180(%2)                    \n\t"
+                "movd %%mm3, 52(%3)                     \n\t"
+                "movd %%mm2, 196(%2)                    \n\t"
+                "movd %%mm2, 68(%3)                     \n\t"
+                "psrlq $32, %%mm2                       \n\t"
+                "movd %%mm2, 84(%3)                     \n\t"
+                "movd %%mm1, 100(%3)                    \n\t"
+                "psrlq $32, %%mm1                       \n\t"
+                "movd %%mm1, 116(%3)                    \n\t"
+
+
+        :: "r" (src), "r" ((long)srcStride), "r" (dst1), "r" (dst2)
+        : "%"REG_a
+        );
 }
 
 /**
@@ -2157,414 +2157,414 @@ static inline void RENAME(transpose1)(uint8_t *dst1, uint8_t *dst2, uint8_t *src
  */
 static inline void RENAME(transpose2)(uint8_t *dst, int dstStride, uint8_t *src)
 {
-	asm(
-		"lea (%0, %1), %%"REG_a"	\n\t"
-		"lea (%%"REG_a",%1,4), %%"REG_d"\n\t"
-//	0	1	2	3	4	5	6	7	8	9
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	edx	edx+%1	edx+2%1	%0+8%1	edx+4%1
-		"movq (%2), %%mm0		\n\t" // 12345678
-		"movq 16(%2), %%mm1		\n\t" // abcdefgh
-		"movq %%mm0, %%mm2		\n\t" // 12345678
-		"punpcklbw %%mm1, %%mm0		\n\t" // 1a2b3c4d
-		"punpckhbw %%mm1, %%mm2		\n\t" // 5e6f7g8h
-
-		"movq 32(%2), %%mm1		\n\t"
-		"movq 48(%2), %%mm3		\n\t"
-		"movq %%mm1, %%mm4		\n\t"
-		"punpcklbw %%mm3, %%mm1		\n\t"
-		"punpckhbw %%mm3, %%mm4		\n\t"
-
-		"movq %%mm0, %%mm3		\n\t"
-		"punpcklwd %%mm1, %%mm0		\n\t"
-		"punpckhwd %%mm1, %%mm3		\n\t"
-		"movq %%mm2, %%mm1		\n\t"
-		"punpcklwd %%mm4, %%mm2		\n\t"
-		"punpckhwd %%mm4, %%mm1		\n\t"
-
-		"movd %%mm0, (%0)		\n\t"
-		"psrlq $32, %%mm0		\n\t"
-		"movd %%mm0, (%%"REG_a")	\n\t"
-		"movd %%mm3, (%%"REG_a", %1)	\n\t"
-		"psrlq $32, %%mm3		\n\t"
-		"movd %%mm3, (%%"REG_a", %1, 2)	\n\t"
-		"movd %%mm2, (%0, %1, 4)	\n\t"
-		"psrlq $32, %%mm2		\n\t"
-		"movd %%mm2, (%%"REG_d")	\n\t"
-		"movd %%mm1, (%%"REG_d", %1)	\n\t"
-		"psrlq $32, %%mm1		\n\t"
-		"movd %%mm1, (%%"REG_d", %1, 2)	\n\t"
-
-
-		"movq 64(%2), %%mm0		\n\t" // 12345678
-		"movq 80(%2), %%mm1		\n\t" // abcdefgh
-		"movq %%mm0, %%mm2		\n\t" // 12345678
-		"punpcklbw %%mm1, %%mm0		\n\t" // 1a2b3c4d
-		"punpckhbw %%mm1, %%mm2		\n\t" // 5e6f7g8h
-
-		"movq 96(%2), %%mm1		\n\t"
-		"movq 112(%2), %%mm3		\n\t"
-		"movq %%mm1, %%mm4		\n\t"
-		"punpcklbw %%mm3, %%mm1		\n\t"
-		"punpckhbw %%mm3, %%mm4		\n\t"
-
-		"movq %%mm0, %%mm3		\n\t"
-		"punpcklwd %%mm1, %%mm0		\n\t"
-		"punpckhwd %%mm1, %%mm3		\n\t"
-		"movq %%mm2, %%mm1		\n\t"
-		"punpcklwd %%mm4, %%mm2		\n\t"
-		"punpckhwd %%mm4, %%mm1		\n\t"
-
-		"movd %%mm0, 4(%0)		\n\t"
-		"psrlq $32, %%mm0		\n\t"
-		"movd %%mm0, 4(%%"REG_a")		\n\t"
-		"movd %%mm3, 4(%%"REG_a", %1)	\n\t"
-		"psrlq $32, %%mm3		\n\t"
-		"movd %%mm3, 4(%%"REG_a", %1, 2)	\n\t"
-		"movd %%mm2, 4(%0, %1, 4)	\n\t"
-		"psrlq $32, %%mm2		\n\t"
-		"movd %%mm2, 4(%%"REG_d")		\n\t"
-		"movd %%mm1, 4(%%"REG_d", %1)	\n\t"
-		"psrlq $32, %%mm1		\n\t"
-		"movd %%mm1, 4(%%"REG_d", %1, 2)	\n\t"
-
-	:: "r" (dst), "r" ((long)dstStride), "r" (src)
-	: "%"REG_a, "%"REG_d
-	);
+        asm(
+                "lea (%0, %1), %%"REG_a"                \n\t"
+                "lea (%%"REG_a",%1,4), %%"REG_d"        \n\t"
+//      0       1       2       3       4       5       6       7       8       9
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  edx     edx+%1  edx+2%1 %0+8%1  edx+4%1
+                "movq (%2), %%mm0                       \n\t" // 12345678
+                "movq 16(%2), %%mm1                     \n\t" // abcdefgh
+                "movq %%mm0, %%mm2                      \n\t" // 12345678
+                "punpcklbw %%mm1, %%mm0                 \n\t" // 1a2b3c4d
+                "punpckhbw %%mm1, %%mm2                 \n\t" // 5e6f7g8h
+
+                "movq 32(%2), %%mm1                     \n\t"
+                "movq 48(%2), %%mm3                     \n\t"
+                "movq %%mm1, %%mm4                      \n\t"
+                "punpcklbw %%mm3, %%mm1                 \n\t"
+                "punpckhbw %%mm3, %%mm4                 \n\t"
+
+                "movq %%mm0, %%mm3                      \n\t"
+                "punpcklwd %%mm1, %%mm0                 \n\t"
+                "punpckhwd %%mm1, %%mm3                 \n\t"
+                "movq %%mm2, %%mm1                      \n\t"
+                "punpcklwd %%mm4, %%mm2                 \n\t"
+                "punpckhwd %%mm4, %%mm1                 \n\t"
+
+                "movd %%mm0, (%0)                       \n\t"
+                "psrlq $32, %%mm0                       \n\t"
+                "movd %%mm0, (%%"REG_a")                \n\t"
+                "movd %%mm3, (%%"REG_a", %1)            \n\t"
+                "psrlq $32, %%mm3                       \n\t"
+                "movd %%mm3, (%%"REG_a", %1, 2)         \n\t"
+                "movd %%mm2, (%0, %1, 4)                \n\t"
+                "psrlq $32, %%mm2                       \n\t"
+                "movd %%mm2, (%%"REG_d")                \n\t"
+                "movd %%mm1, (%%"REG_d", %1)            \n\t"
+                "psrlq $32, %%mm1                       \n\t"
+                "movd %%mm1, (%%"REG_d", %1, 2)         \n\t"
+
+
+                "movq 64(%2), %%mm0                     \n\t" // 12345678
+                "movq 80(%2), %%mm1                     \n\t" // abcdefgh
+                "movq %%mm0, %%mm2                      \n\t" // 12345678
+                "punpcklbw %%mm1, %%mm0                 \n\t" // 1a2b3c4d
+                "punpckhbw %%mm1, %%mm2                 \n\t" // 5e6f7g8h
+
+                "movq 96(%2), %%mm1                     \n\t"
+                "movq 112(%2), %%mm3                    \n\t"
+                "movq %%mm1, %%mm4                      \n\t"
+                "punpcklbw %%mm3, %%mm1                 \n\t"
+                "punpckhbw %%mm3, %%mm4                 \n\t"
+
+                "movq %%mm0, %%mm3                      \n\t"
+                "punpcklwd %%mm1, %%mm0                 \n\t"
+                "punpckhwd %%mm1, %%mm3                 \n\t"
+                "movq %%mm2, %%mm1                      \n\t"
+                "punpcklwd %%mm4, %%mm2                 \n\t"
+                "punpckhwd %%mm4, %%mm1                 \n\t"
+
+                "movd %%mm0, 4(%0)                      \n\t"
+                "psrlq $32, %%mm0                       \n\t"
+                "movd %%mm0, 4(%%"REG_a")               \n\t"
+                "movd %%mm3, 4(%%"REG_a", %1)           \n\t"
+                "psrlq $32, %%mm3                       \n\t"
+                "movd %%mm3, 4(%%"REG_a", %1, 2)        \n\t"
+                "movd %%mm2, 4(%0, %1, 4)               \n\t"
+                "psrlq $32, %%mm2                       \n\t"
+                "movd %%mm2, 4(%%"REG_d")               \n\t"
+                "movd %%mm1, 4(%%"REG_d", %1)           \n\t"
+                "psrlq $32, %%mm1                       \n\t"
+                "movd %%mm1, 4(%%"REG_d", %1, 2)        \n\t"
+
+        :: "r" (dst), "r" ((long)dstStride), "r" (src)
+        : "%"REG_a, "%"REG_d
+        );
 }
 #endif //HAVE_MMX
 //static long test=0;
 
 #ifndef HAVE_ALTIVEC
 static inline void RENAME(tempNoiseReducer)(uint8_t *src, int stride,
-				    uint8_t *tempBlured, uint32_t *tempBluredPast, int *maxNoise)
+                                    uint8_t *tempBlured, uint32_t *tempBluredPast, int *maxNoise)
 {
-	// to save a register (FIXME do this outside of the loops)
-	tempBluredPast[127]= maxNoise[0];
-	tempBluredPast[128]= maxNoise[1];
-	tempBluredPast[129]= maxNoise[2];
+        // to save a register (FIXME do this outside of the loops)
+        tempBluredPast[127]= maxNoise[0];
+        tempBluredPast[128]= maxNoise[1];
+        tempBluredPast[129]= maxNoise[2];
 
 #define FAST_L2_DIFF
 //#define L1_DIFF //u should change the thresholds too if u try that one
 #if defined (HAVE_MMX2) || defined (HAVE_3DNOW)
-	asm volatile(
-		"lea (%2, %2, 2), %%"REG_a"			\n\t" // 3*stride
-		"lea (%2, %2, 4), %%"REG_d"			\n\t" // 5*stride
-		"lea (%%"REG_d", %2, 2), %%"REG_c"		\n\t" // 7*stride
-//	0	1	2	3	4	5	6	7	8	9
-//	%x	%x+%2	%x+2%2	%x+eax	%x+4%2	%x+edx	%x+2eax	%x+ecx	%x+8%2
+        asm volatile(
+                "lea (%2, %2, 2), %%"REG_a"             \n\t" // 3*stride
+                "lea (%2, %2, 4), %%"REG_d"             \n\t" // 5*stride
+                "lea (%%"REG_d", %2, 2), %%"REG_c"      \n\t" // 7*stride
+//      0       1       2       3       4       5       6       7       8       9
+//      %x      %x+%2   %x+2%2  %x+eax  %x+4%2  %x+edx  %x+2eax %x+ecx  %x+8%2
 //FIXME reorder?
 #ifdef L1_DIFF //needs mmx2
-		"movq (%0), %%mm0				\n\t" // L0
-		"psadbw (%1), %%mm0				\n\t" // |L0-R0|
-		"movq (%0, %2), %%mm1				\n\t" // L1
-		"psadbw (%1, %2), %%mm1				\n\t" // |L1-R1|
-		"movq (%0, %2, 2), %%mm2			\n\t" // L2
-		"psadbw (%1, %2, 2), %%mm2			\n\t" // |L2-R2|
-		"movq (%0, %%"REG_a"), %%mm3			\n\t" // L3
-		"psadbw (%1, %%"REG_a"), %%mm3			\n\t" // |L3-R3|
-
-		"movq (%0, %2, 4), %%mm4			\n\t" // L4
-		"paddw %%mm1, %%mm0				\n\t"
-		"psadbw (%1, %2, 4), %%mm4			\n\t" // |L4-R4|
-		"movq (%0, %%"REG_d"), %%mm5			\n\t" // L5
-		"paddw %%mm2, %%mm0				\n\t"
-		"psadbw (%1, %%"REG_d"), %%mm5			\n\t" // |L5-R5|
-		"movq (%0, %%"REG_a", 2), %%mm6			\n\t" // L6
-		"paddw %%mm3, %%mm0				\n\t"
-		"psadbw (%1, %%"REG_a", 2), %%mm6		\n\t" // |L6-R6|
-		"movq (%0, %%"REG_c"), %%mm7			\n\t" // L7
-		"paddw %%mm4, %%mm0				\n\t"
-		"psadbw (%1, %%"REG_c"), %%mm7			\n\t" // |L7-R7|
-		"paddw %%mm5, %%mm6				\n\t"
-		"paddw %%mm7, %%mm6				\n\t"
-		"paddw %%mm6, %%mm0				\n\t"
+                "movq (%0), %%mm0                       \n\t" // L0
+                "psadbw (%1), %%mm0                     \n\t" // |L0-R0|
+                "movq (%0, %2), %%mm1                   \n\t" // L1
+                "psadbw (%1, %2), %%mm1                 \n\t" // |L1-R1|
+                "movq (%0, %2, 2), %%mm2                \n\t" // L2
+                "psadbw (%1, %2, 2), %%mm2              \n\t" // |L2-R2|
+                "movq (%0, %%"REG_a"), %%mm3            \n\t" // L3
+                "psadbw (%1, %%"REG_a"), %%mm3          \n\t" // |L3-R3|
+
+                "movq (%0, %2, 4), %%mm4                \n\t" // L4
+                "paddw %%mm1, %%mm0                     \n\t"
+                "psadbw (%1, %2, 4), %%mm4              \n\t" // |L4-R4|
+                "movq (%0, %%"REG_d"), %%mm5            \n\t" // L5
+                "paddw %%mm2, %%mm0                     \n\t"
+                "psadbw (%1, %%"REG_d"), %%mm5          \n\t" // |L5-R5|
+                "movq (%0, %%"REG_a", 2), %%mm6         \n\t" // L6
+                "paddw %%mm3, %%mm0                     \n\t"
+                "psadbw (%1, %%"REG_a", 2), %%mm6       \n\t" // |L6-R6|
+                "movq (%0, %%"REG_c"), %%mm7            \n\t" // L7
+                "paddw %%mm4, %%mm0                     \n\t"
+                "psadbw (%1, %%"REG_c"), %%mm7          \n\t" // |L7-R7|
+                "paddw %%mm5, %%mm6                     \n\t"
+                "paddw %%mm7, %%mm6                     \n\t"
+                "paddw %%mm6, %%mm0                     \n\t"
 #else //L1_DIFF
 #if defined (FAST_L2_DIFF)
-		"pcmpeqb %%mm7, %%mm7				\n\t"
-		"movq "MANGLE(b80)", %%mm6			\n\t"
-		"pxor %%mm0, %%mm0				\n\t"
+                "pcmpeqb %%mm7, %%mm7                   \n\t"
+                "movq "MANGLE(b80)", %%mm6              \n\t"
+                "pxor %%mm0, %%mm0                      \n\t"
 #define REAL_L2_DIFF_CORE(a, b)\
-		"movq " #a ", %%mm5				\n\t"\
-		"movq " #b ", %%mm2				\n\t"\
-		"pxor %%mm7, %%mm2				\n\t"\
-		PAVGB(%%mm2, %%mm5)\
-		"paddb %%mm6, %%mm5				\n\t"\
-		"movq %%mm5, %%mm2				\n\t"\
-		"psllw $8, %%mm5				\n\t"\
-		"pmaddwd %%mm5, %%mm5				\n\t"\
-		"pmaddwd %%mm2, %%mm2				\n\t"\
-		"paddd %%mm2, %%mm5				\n\t"\
-		"psrld $14, %%mm5				\n\t"\
-		"paddd %%mm5, %%mm0				\n\t"
+                "movq " #a ", %%mm5                     \n\t"\
+                "movq " #b ", %%mm2                     \n\t"\
+                "pxor %%mm7, %%mm2                      \n\t"\
+                PAVGB(%%mm2, %%mm5)\
+                "paddb %%mm6, %%mm5                     \n\t"\
+                "movq %%mm5, %%mm2                      \n\t"\
+                "psllw $8, %%mm5                        \n\t"\
+                "pmaddwd %%mm5, %%mm5                   \n\t"\
+                "pmaddwd %%mm2, %%mm2                   \n\t"\
+                "paddd %%mm2, %%mm5                     \n\t"\
+                "psrld $14, %%mm5                       \n\t"\
+                "paddd %%mm5, %%mm0                     \n\t"
 
 #else //defined (FAST_L2_DIFF)
-		"pxor %%mm7, %%mm7				\n\t"
-		"pxor %%mm0, %%mm0				\n\t"
+                "pxor %%mm7, %%mm7                      \n\t"
+                "pxor %%mm0, %%mm0                      \n\t"
 #define REAL_L2_DIFF_CORE(a, b)\
-		"movq " #a ", %%mm5				\n\t"\
-		"movq " #b ", %%mm2				\n\t"\
-		"movq %%mm5, %%mm1				\n\t"\
-		"movq %%mm2, %%mm3				\n\t"\
-		"punpcklbw %%mm7, %%mm5				\n\t"\
-		"punpckhbw %%mm7, %%mm1				\n\t"\
-		"punpcklbw %%mm7, %%mm2				\n\t"\
-		"punpckhbw %%mm7, %%mm3				\n\t"\
-		"psubw %%mm2, %%mm5				\n\t"\
-		"psubw %%mm3, %%mm1				\n\t"\
-		"pmaddwd %%mm5, %%mm5				\n\t"\
-		"pmaddwd %%mm1, %%mm1				\n\t"\
-		"paddd %%mm1, %%mm5				\n\t"\
-		"paddd %%mm5, %%mm0				\n\t"
+                "movq " #a ", %%mm5                     \n\t"\
+                "movq " #b ", %%mm2                     \n\t"\
+                "movq %%mm5, %%mm1                      \n\t"\
+                "movq %%mm2, %%mm3                      \n\t"\
+                "punpcklbw %%mm7, %%mm5                 \n\t"\
+                "punpckhbw %%mm7, %%mm1                 \n\t"\
+                "punpcklbw %%mm7, %%mm2                 \n\t"\
+                "punpckhbw %%mm7, %%mm3                 \n\t"\
+                "psubw %%mm2, %%mm5                     \n\t"\
+                "psubw %%mm3, %%mm1                     \n\t"\
+                "pmaddwd %%mm5, %%mm5                   \n\t"\
+                "pmaddwd %%mm1, %%mm1                   \n\t"\
+                "paddd %%mm1, %%mm5                     \n\t"\
+                "paddd %%mm5, %%mm0                     \n\t"
 
 #endif //defined (FAST_L2_DIFF)
 
 #define L2_DIFF_CORE(a, b)  REAL_L2_DIFF_CORE(a, b)
 
-L2_DIFF_CORE((%0), (%1))
-L2_DIFF_CORE((%0, %2), (%1, %2))
-L2_DIFF_CORE((%0, %2, 2), (%1, %2, 2))
-L2_DIFF_CORE((%0, %%REGa), (%1, %%REGa))
-L2_DIFF_CORE((%0, %2, 4), (%1, %2, 4))
-L2_DIFF_CORE((%0, %%REGd), (%1, %%REGd))
+L2_DIFF_CORE((%0)          , (%1))
+L2_DIFF_CORE((%0, %2)      , (%1, %2))
+L2_DIFF_CORE((%0, %2, 2)   , (%1, %2, 2))
+L2_DIFF_CORE((%0, %%REGa)  , (%1, %%REGa))
+L2_DIFF_CORE((%0, %2, 4)   , (%1, %2, 4))
+L2_DIFF_CORE((%0, %%REGd)  , (%1, %%REGd))
 L2_DIFF_CORE((%0, %%REGa,2), (%1, %%REGa,2))
-L2_DIFF_CORE((%0, %%REGc), (%1, %%REGc))
+L2_DIFF_CORE((%0, %%REGc)  , (%1, %%REGc))
 
 #endif //L1_DIFF
 
-		"movq %%mm0, %%mm4				\n\t"
-		"psrlq $32, %%mm0				\n\t"
-		"paddd %%mm0, %%mm4				\n\t"
-		"movd %%mm4, %%ecx				\n\t"
-		"shll $2, %%ecx					\n\t"
-		"mov %3, %%"REG_d"				\n\t"
-		"addl -4(%%"REG_d"), %%ecx			\n\t"
-		"addl 4(%%"REG_d"), %%ecx			\n\t"
-		"addl -1024(%%"REG_d"), %%ecx			\n\t"
-		"addl $4, %%ecx					\n\t"
-		"addl 1024(%%"REG_d"), %%ecx			\n\t"
-		"shrl $3, %%ecx					\n\t"
-		"movl %%ecx, (%%"REG_d")			\n\t"
-
-//		"mov %3, %%"REG_c"				\n\t"
-//		"mov %%"REG_c", test				\n\t"
-//		"jmp 4f \n\t"
-		"cmpl 512(%%"REG_d"), %%ecx			\n\t"
-		" jb 2f						\n\t"
-		"cmpl 516(%%"REG_d"), %%ecx			\n\t"
-		" jb 1f						\n\t"
-
-		"lea (%%"REG_a", %2, 2), %%"REG_d"		\n\t" // 5*stride
-		"lea (%%"REG_d", %2, 2), %%"REG_c"		\n\t" // 7*stride
-		"movq (%0), %%mm0				\n\t" // L0
-		"movq (%0, %2), %%mm1				\n\t" // L1
-		"movq (%0, %2, 2), %%mm2			\n\t" // L2
-		"movq (%0, %%"REG_a"), %%mm3			\n\t" // L3
-		"movq (%0, %2, 4), %%mm4			\n\t" // L4
-		"movq (%0, %%"REG_d"), %%mm5			\n\t" // L5
-		"movq (%0, %%"REG_a", 2), %%mm6			\n\t" // L6
-		"movq (%0, %%"REG_c"), %%mm7			\n\t" // L7
-		"movq %%mm0, (%1)				\n\t" // L0
-		"movq %%mm1, (%1, %2)				\n\t" // L1
-		"movq %%mm2, (%1, %2, 2)			\n\t" // L2
-		"movq %%mm3, (%1, %%"REG_a")			\n\t" // L3
-		"movq %%mm4, (%1, %2, 4)			\n\t" // L4
-		"movq %%mm5, (%1, %%"REG_d")			\n\t" // L5
-		"movq %%mm6, (%1, %%"REG_a", 2)			\n\t" // L6
-		"movq %%mm7, (%1, %%"REG_c")			\n\t" // L7
-		"jmp 4f						\n\t"
-
-		"1:						\n\t"
-		"lea (%%"REG_a", %2, 2), %%"REG_d"		\n\t" // 5*stride
-		"lea (%%"REG_d", %2, 2), %%"REG_c"		\n\t" // 7*stride
-		"movq (%0), %%mm0				\n\t" // L0
-		PAVGB((%1), %%mm0)				      // L0
-		"movq (%0, %2), %%mm1				\n\t" // L1
-		PAVGB((%1, %2), %%mm1)				      // L1
-		"movq (%0, %2, 2), %%mm2			\n\t" // L2
-		PAVGB((%1, %2, 2), %%mm2)			      // L2
-		"movq (%0, %%"REG_a"), %%mm3			\n\t" // L3
-		PAVGB((%1, %%REGa), %%mm3)			      // L3
-		"movq (%0, %2, 4), %%mm4			\n\t" // L4
-		PAVGB((%1, %2, 4), %%mm4)			      // L4
-		"movq (%0, %%"REG_d"), %%mm5			\n\t" // L5
-		PAVGB((%1, %%REGd), %%mm5)			      // L5
-		"movq (%0, %%"REG_a", 2), %%mm6			\n\t" // L6
-		PAVGB((%1, %%REGa, 2), %%mm6)			      // L6
-		"movq (%0, %%"REG_c"), %%mm7			\n\t" // L7
-		PAVGB((%1, %%REGc), %%mm7)			      // L7
-		"movq %%mm0, (%1)				\n\t" // R0
-		"movq %%mm1, (%1, %2)				\n\t" // R1
-		"movq %%mm2, (%1, %2, 2)			\n\t" // R2
-		"movq %%mm3, (%1, %%"REG_a")			\n\t" // R3
-		"movq %%mm4, (%1, %2, 4)			\n\t" // R4
-		"movq %%mm5, (%1, %%"REG_d")			\n\t" // R5
-		"movq %%mm6, (%1, %%"REG_a", 2)			\n\t" // R6
-		"movq %%mm7, (%1, %%"REG_c")			\n\t" // R7
-		"movq %%mm0, (%0)				\n\t" // L0
-		"movq %%mm1, (%0, %2)				\n\t" // L1
-		"movq %%mm2, (%0, %2, 2)			\n\t" // L2
-		"movq %%mm3, (%0, %%"REG_a")			\n\t" // L3
-		"movq %%mm4, (%0, %2, 4)			\n\t" // L4
-		"movq %%mm5, (%0, %%"REG_d")			\n\t" // L5
-		"movq %%mm6, (%0, %%"REG_a", 2)			\n\t" // L6
-		"movq %%mm7, (%0, %%"REG_c")			\n\t" // L7
-		"jmp 4f						\n\t"
-
-		"2:						\n\t"
-		"cmpl 508(%%"REG_d"), %%ecx			\n\t"
-		" jb 3f						\n\t"
-
-		"lea (%%"REG_a", %2, 2), %%"REG_d"		\n\t" // 5*stride
-		"lea (%%"REG_d", %2, 2), %%"REG_c"		\n\t" // 7*stride
-		"movq (%0), %%mm0				\n\t" // L0
-		"movq (%0, %2), %%mm1				\n\t" // L1
-		"movq (%0, %2, 2), %%mm2			\n\t" // L2
-		"movq (%0, %%"REG_a"), %%mm3			\n\t" // L3
-		"movq (%1), %%mm4				\n\t" // R0
-		"movq (%1, %2), %%mm5				\n\t" // R1
-		"movq (%1, %2, 2), %%mm6			\n\t" // R2
-		"movq (%1, %%"REG_a"), %%mm7			\n\t" // R3
-		PAVGB(%%mm4, %%mm0)
-		PAVGB(%%mm5, %%mm1)
-		PAVGB(%%mm6, %%mm2)
-		PAVGB(%%mm7, %%mm3)
-		PAVGB(%%mm4, %%mm0)
-		PAVGB(%%mm5, %%mm1)
-		PAVGB(%%mm6, %%mm2)
-		PAVGB(%%mm7, %%mm3)
-		"movq %%mm0, (%1)				\n\t" // R0
-		"movq %%mm1, (%1, %2)				\n\t" // R1
-		"movq %%mm2, (%1, %2, 2)			\n\t" // R2
-		"movq %%mm3, (%1, %%"REG_a")			\n\t" // R3
-		"movq %%mm0, (%0)				\n\t" // L0
-		"movq %%mm1, (%0, %2)				\n\t" // L1
-		"movq %%mm2, (%0, %2, 2)			\n\t" // L2
-		"movq %%mm3, (%0, %%"REG_a")			\n\t" // L3
-
-		"movq (%0, %2, 4), %%mm0			\n\t" // L4
-		"movq (%0, %%"REG_d"), %%mm1			\n\t" // L5
-		"movq (%0, %%"REG_a", 2), %%mm2			\n\t" // L6
-		"movq (%0, %%"REG_c"), %%mm3			\n\t" // L7
-		"movq (%1, %2, 4), %%mm4			\n\t" // R4
-		"movq (%1, %%"REG_d"), %%mm5			\n\t" // R5
-		"movq (%1, %%"REG_a", 2), %%mm6			\n\t" // R6
-		"movq (%1, %%"REG_c"), %%mm7			\n\t" // R7
-		PAVGB(%%mm4, %%mm0)
-		PAVGB(%%mm5, %%mm1)
-		PAVGB(%%mm6, %%mm2)
-		PAVGB(%%mm7, %%mm3)
-		PAVGB(%%mm4, %%mm0)
-		PAVGB(%%mm5, %%mm1)
-		PAVGB(%%mm6, %%mm2)
-		PAVGB(%%mm7, %%mm3)
-		"movq %%mm0, (%1, %2, 4)			\n\t" // R4
-		"movq %%mm1, (%1, %%"REG_d")			\n\t" // R5
-		"movq %%mm2, (%1, %%"REG_a", 2)			\n\t" // R6
-		"movq %%mm3, (%1, %%"REG_c")			\n\t" // R7
-		"movq %%mm0, (%0, %2, 4)			\n\t" // L4
-		"movq %%mm1, (%0, %%"REG_d")			\n\t" // L5
-		"movq %%mm2, (%0, %%"REG_a", 2)			\n\t" // L6
-		"movq %%mm3, (%0, %%"REG_c")			\n\t" // L7
-		"jmp 4f						\n\t"
-
-		"3:						\n\t"
-		"lea (%%"REG_a", %2, 2), %%"REG_d"		\n\t" // 5*stride
-		"lea (%%"REG_d", %2, 2), %%"REG_c"		\n\t" // 7*stride
-		"movq (%0), %%mm0				\n\t" // L0
-		"movq (%0, %2), %%mm1				\n\t" // L1
-		"movq (%0, %2, 2), %%mm2			\n\t" // L2
-		"movq (%0, %%"REG_a"), %%mm3			\n\t" // L3
-		"movq (%1), %%mm4				\n\t" // R0
-		"movq (%1, %2), %%mm5				\n\t" // R1
-		"movq (%1, %2, 2), %%mm6			\n\t" // R2
-		"movq (%1, %%"REG_a"), %%mm7			\n\t" // R3
-		PAVGB(%%mm4, %%mm0)
-		PAVGB(%%mm5, %%mm1)
-		PAVGB(%%mm6, %%mm2)
-		PAVGB(%%mm7, %%mm3)
-		PAVGB(%%mm4, %%mm0)
-		PAVGB(%%mm5, %%mm1)
-		PAVGB(%%mm6, %%mm2)
-		PAVGB(%%mm7, %%mm3)
-		PAVGB(%%mm4, %%mm0)
-		PAVGB(%%mm5, %%mm1)
-		PAVGB(%%mm6, %%mm2)
-		PAVGB(%%mm7, %%mm3)
-		"movq %%mm0, (%1)				\n\t" // R0
-		"movq %%mm1, (%1, %2)				\n\t" // R1
-		"movq %%mm2, (%1, %2, 2)			\n\t" // R2
-		"movq %%mm3, (%1, %%"REG_a")			\n\t" // R3
-		"movq %%mm0, (%0)				\n\t" // L0
-		"movq %%mm1, (%0, %2)				\n\t" // L1
-		"movq %%mm2, (%0, %2, 2)			\n\t" // L2
-		"movq %%mm3, (%0, %%"REG_a")			\n\t" // L3
-
-		"movq (%0, %2, 4), %%mm0			\n\t" // L4
-		"movq (%0, %%"REG_d"), %%mm1			\n\t" // L5
-		"movq (%0, %%"REG_a", 2), %%mm2			\n\t" // L6
-		"movq (%0, %%"REG_c"), %%mm3			\n\t" // L7
-		"movq (%1, %2, 4), %%mm4			\n\t" // R4
-		"movq (%1, %%"REG_d"), %%mm5			\n\t" // R5
-		"movq (%1, %%"REG_a", 2), %%mm6			\n\t" // R6
-		"movq (%1, %%"REG_c"), %%mm7			\n\t" // R7
-		PAVGB(%%mm4, %%mm0)
-		PAVGB(%%mm5, %%mm1)
-		PAVGB(%%mm6, %%mm2)
-		PAVGB(%%mm7, %%mm3)
-		PAVGB(%%mm4, %%mm0)
-		PAVGB(%%mm5, %%mm1)
-		PAVGB(%%mm6, %%mm2)
-		PAVGB(%%mm7, %%mm3)
-		PAVGB(%%mm4, %%mm0)
-		PAVGB(%%mm5, %%mm1)
-		PAVGB(%%mm6, %%mm2)
-		PAVGB(%%mm7, %%mm3)
-		"movq %%mm0, (%1, %2, 4)			\n\t" // R4
-		"movq %%mm1, (%1, %%"REG_d")			\n\t" // R5
-		"movq %%mm2, (%1, %%"REG_a", 2)			\n\t" // R6
-		"movq %%mm3, (%1, %%"REG_c")			\n\t" // R7
-		"movq %%mm0, (%0, %2, 4)			\n\t" // L4
-		"movq %%mm1, (%0, %%"REG_d")			\n\t" // L5
-		"movq %%mm2, (%0, %%"REG_a", 2)			\n\t" // L6
-		"movq %%mm3, (%0, %%"REG_c")			\n\t" // L7
-
-		"4:						\n\t"
-
-		:: "r" (src), "r" (tempBlured), "r"((long)stride), "m" (tempBluredPast)
-		: "%"REG_a, "%"REG_d, "%"REG_c, "memory"
-		);
+                "movq %%mm0, %%mm4                      \n\t"
+                "psrlq $32, %%mm0                       \n\t"
+                "paddd %%mm0, %%mm4                     \n\t"
+                "movd %%mm4, %%ecx                      \n\t"
+                "shll $2, %%ecx                         \n\t"
+                "mov %3, %%"REG_d"                      \n\t"
+                "addl -4(%%"REG_d"), %%ecx              \n\t"
+                "addl 4(%%"REG_d"), %%ecx               \n\t"
+                "addl -1024(%%"REG_d"), %%ecx           \n\t"
+                "addl $4, %%ecx                         \n\t"
+                "addl 1024(%%"REG_d"), %%ecx            \n\t"
+                "shrl $3, %%ecx                         \n\t"
+                "movl %%ecx, (%%"REG_d")                \n\t"
+
+//                "mov %3, %%"REG_c"                      \n\t"
+//                "mov %%"REG_c", test                    \n\t"
+//                "jmp 4f                                 \n\t"
+                "cmpl 512(%%"REG_d"), %%ecx             \n\t"
+                " jb 2f                                 \n\t"
+                "cmpl 516(%%"REG_d"), %%ecx             \n\t"
+                " jb 1f                                 \n\t"
+
+                "lea (%%"REG_a", %2, 2), %%"REG_d"      \n\t" // 5*stride
+                "lea (%%"REG_d", %2, 2), %%"REG_c"      \n\t" // 7*stride
+                "movq (%0), %%mm0                       \n\t" // L0
+                "movq (%0, %2), %%mm1                   \n\t" // L1
+                "movq (%0, %2, 2), %%mm2                \n\t" // L2
+                "movq (%0, %%"REG_a"), %%mm3            \n\t" // L3
+                "movq (%0, %2, 4), %%mm4                \n\t" // L4
+                "movq (%0, %%"REG_d"), %%mm5            \n\t" // L5
+                "movq (%0, %%"REG_a", 2), %%mm6         \n\t" // L6
+                "movq (%0, %%"REG_c"), %%mm7            \n\t" // L7
+                "movq %%mm0, (%1)                       \n\t" // L0
+                "movq %%mm1, (%1, %2)                   \n\t" // L1
+                "movq %%mm2, (%1, %2, 2)                \n\t" // L2
+                "movq %%mm3, (%1, %%"REG_a")            \n\t" // L3
+                "movq %%mm4, (%1, %2, 4)                \n\t" // L4
+                "movq %%mm5, (%1, %%"REG_d")            \n\t" // L5
+                "movq %%mm6, (%1, %%"REG_a", 2)         \n\t" // L6
+                "movq %%mm7, (%1, %%"REG_c")            \n\t" // L7
+                "jmp 4f                                 \n\t"
+
+                "1:                                     \n\t"
+                "lea (%%"REG_a", %2, 2), %%"REG_d"      \n\t" // 5*stride
+                "lea (%%"REG_d", %2, 2), %%"REG_c"      \n\t" // 7*stride
+                "movq (%0), %%mm0                       \n\t" // L0
+                PAVGB((%1), %%mm0)                            // L0
+                "movq (%0, %2), %%mm1                   \n\t" // L1
+                PAVGB((%1, %2), %%mm1)                        // L1
+                "movq (%0, %2, 2), %%mm2                \n\t" // L2
+                PAVGB((%1, %2, 2), %%mm2)                     // L2
+                "movq (%0, %%"REG_a"), %%mm3            \n\t" // L3
+                PAVGB((%1, %%REGa), %%mm3)                    // L3
+                "movq (%0, %2, 4), %%mm4                \n\t" // L4
+                PAVGB((%1, %2, 4), %%mm4)                     // L4
+                "movq (%0, %%"REG_d"), %%mm5            \n\t" // L5
+                PAVGB((%1, %%REGd), %%mm5)                    // L5
+                "movq (%0, %%"REG_a", 2), %%mm6         \n\t" // L6
+                PAVGB((%1, %%REGa, 2), %%mm6)                 // L6
+                "movq (%0, %%"REG_c"), %%mm7            \n\t" // L7
+                PAVGB((%1, %%REGc), %%mm7)                    // L7
+                "movq %%mm0, (%1)                       \n\t" // R0
+                "movq %%mm1, (%1, %2)                   \n\t" // R1
+                "movq %%mm2, (%1, %2, 2)                \n\t" // R2
+                "movq %%mm3, (%1, %%"REG_a")            \n\t" // R3
+                "movq %%mm4, (%1, %2, 4)                \n\t" // R4
+                "movq %%mm5, (%1, %%"REG_d")            \n\t" // R5
+                "movq %%mm6, (%1, %%"REG_a", 2)         \n\t" // R6
+                "movq %%mm7, (%1, %%"REG_c")            \n\t" // R7
+                "movq %%mm0, (%0)                       \n\t" // L0
+                "movq %%mm1, (%0, %2)                   \n\t" // L1
+                "movq %%mm2, (%0, %2, 2)                \n\t" // L2
+                "movq %%mm3, (%0, %%"REG_a")            \n\t" // L3
+                "movq %%mm4, (%0, %2, 4)                \n\t" // L4
+                "movq %%mm5, (%0, %%"REG_d")            \n\t" // L5
+                "movq %%mm6, (%0, %%"REG_a", 2)         \n\t" // L6
+                "movq %%mm7, (%0, %%"REG_c")            \n\t" // L7
+                "jmp 4f                                 \n\t"
+
+                "2:                                     \n\t"
+                "cmpl 508(%%"REG_d"), %%ecx             \n\t"
+                " jb 3f                                 \n\t"
+
+                "lea (%%"REG_a", %2, 2), %%"REG_d"      \n\t" // 5*stride
+                "lea (%%"REG_d", %2, 2), %%"REG_c"      \n\t" // 7*stride
+                "movq (%0), %%mm0                       \n\t" // L0
+                "movq (%0, %2), %%mm1                   \n\t" // L1
+                "movq (%0, %2, 2), %%mm2                \n\t" // L2
+                "movq (%0, %%"REG_a"), %%mm3            \n\t" // L3
+                "movq (%1), %%mm4                       \n\t" // R0
+                "movq (%1, %2), %%mm5                   \n\t" // R1
+                "movq (%1, %2, 2), %%mm6                \n\t" // R2
+                "movq (%1, %%"REG_a"), %%mm7            \n\t" // R3
+                PAVGB(%%mm4, %%mm0)
+                PAVGB(%%mm5, %%mm1)
+                PAVGB(%%mm6, %%mm2)
+                PAVGB(%%mm7, %%mm3)
+                PAVGB(%%mm4, %%mm0)
+                PAVGB(%%mm5, %%mm1)
+                PAVGB(%%mm6, %%mm2)
+                PAVGB(%%mm7, %%mm3)
+                "movq %%mm0, (%1)                       \n\t" // R0
+                "movq %%mm1, (%1, %2)                   \n\t" // R1
+                "movq %%mm2, (%1, %2, 2)                \n\t" // R2
+                "movq %%mm3, (%1, %%"REG_a")            \n\t" // R3
+                "movq %%mm0, (%0)                       \n\t" // L0
+                "movq %%mm1, (%0, %2)                   \n\t" // L1
+                "movq %%mm2, (%0, %2, 2)                \n\t" // L2
+                "movq %%mm3, (%0, %%"REG_a")            \n\t" // L3
+
+                "movq (%0, %2, 4), %%mm0                \n\t" // L4
+                "movq (%0, %%"REG_d"), %%mm1            \n\t" // L5
+                "movq (%0, %%"REG_a", 2), %%mm2         \n\t" // L6
+                "movq (%0, %%"REG_c"), %%mm3            \n\t" // L7
+                "movq (%1, %2, 4), %%mm4                \n\t" // R4
+                "movq (%1, %%"REG_d"), %%mm5            \n\t" // R5
+                "movq (%1, %%"REG_a", 2), %%mm6         \n\t" // R6
+                "movq (%1, %%"REG_c"), %%mm7            \n\t" // R7
+                PAVGB(%%mm4, %%mm0)
+                PAVGB(%%mm5, %%mm1)
+                PAVGB(%%mm6, %%mm2)
+                PAVGB(%%mm7, %%mm3)
+                PAVGB(%%mm4, %%mm0)
+                PAVGB(%%mm5, %%mm1)
+                PAVGB(%%mm6, %%mm2)
+                PAVGB(%%mm7, %%mm3)
+                "movq %%mm0, (%1, %2, 4)                \n\t" // R4
+                "movq %%mm1, (%1, %%"REG_d")            \n\t" // R5
+                "movq %%mm2, (%1, %%"REG_a", 2)         \n\t" // R6
+                "movq %%mm3, (%1, %%"REG_c")            \n\t" // R7
+                "movq %%mm0, (%0, %2, 4)                \n\t" // L4
+                "movq %%mm1, (%0, %%"REG_d")            \n\t" // L5
+                "movq %%mm2, (%0, %%"REG_a", 2)         \n\t" // L6
+                "movq %%mm3, (%0, %%"REG_c")            \n\t" // L7
+                "jmp 4f                                 \n\t"
+
+                "3:                                     \n\t"
+                "lea (%%"REG_a", %2, 2), %%"REG_d"      \n\t" // 5*stride
+                "lea (%%"REG_d", %2, 2), %%"REG_c"      \n\t" // 7*stride
+                "movq (%0), %%mm0                       \n\t" // L0
+                "movq (%0, %2), %%mm1                   \n\t" // L1
+                "movq (%0, %2, 2), %%mm2                \n\t" // L2
+                "movq (%0, %%"REG_a"), %%mm3            \n\t" // L3
+                "movq (%1), %%mm4                       \n\t" // R0
+                "movq (%1, %2), %%mm5                   \n\t" // R1
+                "movq (%1, %2, 2), %%mm6                \n\t" // R2
+                "movq (%1, %%"REG_a"), %%mm7            \n\t" // R3
+                PAVGB(%%mm4, %%mm0)
+                PAVGB(%%mm5, %%mm1)
+                PAVGB(%%mm6, %%mm2)
+                PAVGB(%%mm7, %%mm3)
+                PAVGB(%%mm4, %%mm0)
+                PAVGB(%%mm5, %%mm1)
+                PAVGB(%%mm6, %%mm2)
+                PAVGB(%%mm7, %%mm3)
+                PAVGB(%%mm4, %%mm0)
+                PAVGB(%%mm5, %%mm1)
+                PAVGB(%%mm6, %%mm2)
+                PAVGB(%%mm7, %%mm3)
+                "movq %%mm0, (%1)                       \n\t" // R0
+                "movq %%mm1, (%1, %2)                   \n\t" // R1
+                "movq %%mm2, (%1, %2, 2)                \n\t" // R2
+                "movq %%mm3, (%1, %%"REG_a")            \n\t" // R3
+                "movq %%mm0, (%0)                       \n\t" // L0
+                "movq %%mm1, (%0, %2)                   \n\t" // L1
+                "movq %%mm2, (%0, %2, 2)                \n\t" // L2
+                "movq %%mm3, (%0, %%"REG_a")            \n\t" // L3
+
+                "movq (%0, %2, 4), %%mm0                \n\t" // L4
+                "movq (%0, %%"REG_d"), %%mm1            \n\t" // L5
+                "movq (%0, %%"REG_a", 2), %%mm2         \n\t" // L6
+                "movq (%0, %%"REG_c"), %%mm3            \n\t" // L7
+                "movq (%1, %2, 4), %%mm4                \n\t" // R4
+                "movq (%1, %%"REG_d"), %%mm5            \n\t" // R5
+                "movq (%1, %%"REG_a", 2), %%mm6         \n\t" // R6
+                "movq (%1, %%"REG_c"), %%mm7            \n\t" // R7
+                PAVGB(%%mm4, %%mm0)
+                PAVGB(%%mm5, %%mm1)
+                PAVGB(%%mm6, %%mm2)
+                PAVGB(%%mm7, %%mm3)
+                PAVGB(%%mm4, %%mm0)
+                PAVGB(%%mm5, %%mm1)
+                PAVGB(%%mm6, %%mm2)
+                PAVGB(%%mm7, %%mm3)
+                PAVGB(%%mm4, %%mm0)
+                PAVGB(%%mm5, %%mm1)
+                PAVGB(%%mm6, %%mm2)
+                PAVGB(%%mm7, %%mm3)
+                "movq %%mm0, (%1, %2, 4)                \n\t" // R4
+                "movq %%mm1, (%1, %%"REG_d")            \n\t" // R5
+                "movq %%mm2, (%1, %%"REG_a", 2)         \n\t" // R6
+                "movq %%mm3, (%1, %%"REG_c")            \n\t" // R7
+                "movq %%mm0, (%0, %2, 4)                \n\t" // L4
+                "movq %%mm1, (%0, %%"REG_d")            \n\t" // L5
+                "movq %%mm2, (%0, %%"REG_a", 2)         \n\t" // L6
+                "movq %%mm3, (%0, %%"REG_c")            \n\t" // L7
+
+                "4:                                     \n\t"
+
+                :: "r" (src), "r" (tempBlured), "r"((long)stride), "m" (tempBluredPast)
+                : "%"REG_a, "%"REG_d, "%"REG_c, "memory"
+                );
 //printf("%d\n", test);
 #else //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 {
-	int y;
-	int d=0;
-//	int sysd=0;
-	int i;
-
-	for(y=0; y<8; y++)
-	{
-		int x;
-		for(x=0; x<8; x++)
-		{
-			int ref= tempBlured[ x + y*stride ];
-			int cur= src[ x + y*stride ];
-			int d1=ref - cur;
-//			if(x==0 || x==7) d1+= d1>>1;
-//			if(y==0 || y==7) d1+= d1>>1;
-//			d+= ABS(d1);
-			d+= d1*d1;
-//			sysd+= d1;
-		}
-	}
-	i=d;
-	d= 	(
-		4*d
-		+(*(tempBluredPast-256))
-		+(*(tempBluredPast-1))+ (*(tempBluredPast+1))
-		+(*(tempBluredPast+256))
-		+4)>>3;
-	*tempBluredPast=i;
-//	((*tempBluredPast)*3 + d + 2)>>2;
+        int y;
+        int d=0;
+//        int sysd=0;
+        int i;
+
+        for(y=0; y<8; y++)
+        {
+                int x;
+                for(x=0; x<8; x++)
+                {
+                        int ref= tempBlured[ x + y*stride ];
+                        int cur= src[ x + y*stride ];
+                        int d1=ref - cur;
+//                        if(x==0 || x==7) d1+= d1>>1;
+//                        if(y==0 || y==7) d1+= d1>>1;
+//                        d+= ABS(d1);
+                        d+= d1*d1;
+//                        sysd+= d1;
+                }
+        }
+        i=d;
+        d=         (
+                4*d
+                +(*(tempBluredPast-256))
+                +(*(tempBluredPast-1))+ (*(tempBluredPast+1))
+                +(*(tempBluredPast+256))
+                +4)>>3;
+        *tempBluredPast=i;
+//        ((*tempBluredPast)*3 + d + 2)>>2;
 
 //printf("%d %d %d\n", maxNoise[0], maxNoise[1], maxNoise[2]);
 /*
@@ -2574,68 +2574,68 @@ Switch between
 64 48 36 27 20 15 11 (33) (approx)
 64 56 49 43 37 33 29 (200) (approx)
 */
-	if(d > maxNoise[1])
-	{
-		if(d < maxNoise[2])
-		{
-			for(y=0; y<8; y++)
-			{
-				int x;
-				for(x=0; x<8; x++)
-				{
-					int ref= tempBlured[ x + y*stride ];
-					int cur= src[ x + y*stride ];
-					tempBlured[ x + y*stride ]=
-					src[ x + y*stride ]=
-						(ref + cur + 1)>>1;
-				}
-			}
-		}
-		else
-		{
-			for(y=0; y<8; y++)
-			{
-				int x;
-				for(x=0; x<8; x++)
-				{
-					tempBlured[ x + y*stride ]= src[ x + y*stride ];
-				}
-			}
-		}
-	}
-	else
-	{
-		if(d < maxNoise[0])
-		{
-			for(y=0; y<8; y++)
-			{
-				int x;
-				for(x=0; x<8; x++)
-				{
-					int ref= tempBlured[ x + y*stride ];
-					int cur= src[ x + y*stride ];
-					tempBlured[ x + y*stride ]=
-					src[ x + y*stride ]=
-						(ref*7 + cur + 4)>>3;
-				}
-			}
-		}
-		else
-		{
-			for(y=0; y<8; y++)
-			{
-				int x;
-				for(x=0; x<8; x++)
-				{
-					int ref= tempBlured[ x + y*stride ];
-					int cur= src[ x + y*stride ];
-					tempBlured[ x + y*stride ]=
-					src[ x + y*stride ]=
-						(ref*3 + cur + 2)>>2;
-				}
-			}
-		}
-	}
+        if(d > maxNoise[1])
+        {
+                if(d < maxNoise[2])
+                {
+                        for(y=0; y<8; y++)
+                        {
+                                int x;
+                                for(x=0; x<8; x++)
+                                {
+                                        int ref= tempBlured[ x + y*stride ];
+                                        int cur= src[ x + y*stride ];
+                                        tempBlured[ x + y*stride ]=
+                                        src[ x + y*stride ]=
+                                                (ref + cur + 1)>>1;
+                                }
+                        }
+                }
+                else
+                {
+                        for(y=0; y<8; y++)
+                        {
+                                int x;
+                                for(x=0; x<8; x++)
+                                {
+                                        tempBlured[ x + y*stride ]= src[ x + y*stride ];
+                                }
+                        }
+                }
+        }
+        else
+        {
+                if(d < maxNoise[0])
+                {
+                        for(y=0; y<8; y++)
+                        {
+                                int x;
+                                for(x=0; x<8; x++)
+                                {
+                                        int ref= tempBlured[ x + y*stride ];
+                                        int cur= src[ x + y*stride ];
+                                        tempBlured[ x + y*stride ]=
+                                        src[ x + y*stride ]=
+                                                (ref*7 + cur + 4)>>3;
+                                }
+                        }
+                }
+                else
+                {
+                        for(y=0; y<8; y++)
+                        {
+                                int x;
+                                for(x=0; x<8; x++)
+                                {
+                                        int ref= tempBlured[ x + y*stride ];
+                                        int cur= src[ x + y*stride ];
+                                        tempBlured[ x + y*stride ]=
+                                        src[ x + y*stride ]=
+                                                (ref*3 + cur + 2)>>2;
+                                }
+                        }
+                }
+        }
 }
 #endif //defined (HAVE_MMX2) || defined (HAVE_3DNOW)
 }
@@ -2646,531 +2646,531 @@ Switch between
  * accurate deblock filter
  */
 static always_inline void RENAME(do_a_deblock)(uint8_t *src, int step, int stride, PPContext *c){
-	int64_t dc_mask, eq_mask, both_masks;
-	int64_t sums[10*8*2];
-	src+= step*3; // src points to begin of the 8x8 Block
+        int64_t dc_mask, eq_mask, both_masks;
+        int64_t sums[10*8*2];
+        src+= step*3; // src points to begin of the 8x8 Block
 //START_TIMER
 asm volatile(
-		"movq %0, %%mm7					\n\t"
-		"movq %1, %%mm6					\n\t"
+                "movq %0, %%mm7                         \n\t"
+                "movq %1, %%mm6                         \n\t"
                 : : "m" (c->mmxDcOffset[c->nonBQP]),  "m" (c->mmxDcThreshold[c->nonBQP])
                 );
 
 asm volatile(
-		"lea (%2, %3), %%"REG_a"			\n\t"
-//	0	1	2	3	4	5	6	7	8	9
-//	%1	eax	eax+%2	eax+2%2	%1+4%2	ecx	ecx+%2	ecx+2%2	%1+8%2	ecx+4%2
-
-		"movq (%2), %%mm0				\n\t"
-		"movq (%%"REG_a"), %%mm1			\n\t"
-                "movq %%mm1, %%mm3				\n\t"
-                "movq %%mm1, %%mm4				\n\t"
-		"psubb %%mm1, %%mm0				\n\t" // mm0 = differnece
-		"paddb %%mm7, %%mm0				\n\t"
-		"pcmpgtb %%mm6, %%mm0				\n\t"
-
-		"movq (%%"REG_a",%3), %%mm2			\n\t"
+                "lea (%2, %3), %%"REG_a"                \n\t"
+//      0       1       2       3       4       5       6       7       8       9
+//      %1      eax     eax+%2  eax+2%2 %1+4%2  ecx     ecx+%2  ecx+2%2 %1+8%2  ecx+4%2
+
+                "movq (%2), %%mm0                       \n\t"
+                "movq (%%"REG_a"), %%mm1                \n\t"
+                "movq %%mm1, %%mm3                      \n\t"
+                "movq %%mm1, %%mm4                      \n\t"
+                "psubb %%mm1, %%mm0                     \n\t" // mm0 = differnece
+                "paddb %%mm7, %%mm0                     \n\t"
+                "pcmpgtb %%mm6, %%mm0                   \n\t"
+
+                "movq (%%"REG_a",%3), %%mm2             \n\t"
                 PMAXUB(%%mm2, %%mm4)
                 PMINUB(%%mm2, %%mm3, %%mm5)
-		"psubb %%mm2, %%mm1				\n\t"
-		"paddb %%mm7, %%mm1				\n\t"
-		"pcmpgtb %%mm6, %%mm1				\n\t"
-		"paddb %%mm1, %%mm0				\n\t"
+                "psubb %%mm2, %%mm1                     \n\t"
+                "paddb %%mm7, %%mm1                     \n\t"
+                "pcmpgtb %%mm6, %%mm1                   \n\t"
+                "paddb %%mm1, %%mm0                     \n\t"
 
-		"movq (%%"REG_a", %3, 2), %%mm1			\n\t"
+                "movq (%%"REG_a", %3, 2), %%mm1         \n\t"
                 PMAXUB(%%mm1, %%mm4)
                 PMINUB(%%mm1, %%mm3, %%mm5)
-		"psubb %%mm1, %%mm2				\n\t"
-		"paddb %%mm7, %%mm2				\n\t"
-		"pcmpgtb %%mm6, %%mm2				\n\t"
-		"paddb %%mm2, %%mm0				\n\t"
+                "psubb %%mm1, %%mm2                     \n\t"
+                "paddb %%mm7, %%mm2                     \n\t"
+                "pcmpgtb %%mm6, %%mm2                   \n\t"
+                "paddb %%mm2, %%mm0                     \n\t"
 
-		"lea (%%"REG_a", %3, 4), %%"REG_a"		\n\t"
+                "lea (%%"REG_a", %3, 4), %%"REG_a"      \n\t"
 
-		"movq (%2, %3, 4), %%mm2			\n\t"
+                "movq (%2, %3, 4), %%mm2                \n\t"
                 PMAXUB(%%mm2, %%mm4)
                 PMINUB(%%mm2, %%mm3, %%mm5)
-		"psubb %%mm2, %%mm1				\n\t"
-		"paddb %%mm7, %%mm1				\n\t"
-		"pcmpgtb %%mm6, %%mm1				\n\t"
-		"paddb %%mm1, %%mm0				\n\t"
+                "psubb %%mm2, %%mm1                     \n\t"
+                "paddb %%mm7, %%mm1                     \n\t"
+                "pcmpgtb %%mm6, %%mm1                   \n\t"
+                "paddb %%mm1, %%mm0                     \n\t"
 
-		"movq (%%"REG_a"), %%mm1			\n\t"
+                "movq (%%"REG_a"), %%mm1                \n\t"
                 PMAXUB(%%mm1, %%mm4)
                 PMINUB(%%mm1, %%mm3, %%mm5)
-		"psubb %%mm1, %%mm2				\n\t"
-		"paddb %%mm7, %%mm2				\n\t"
-		"pcmpgtb %%mm6, %%mm2				\n\t"
-		"paddb %%mm2, %%mm0				\n\t"
+                "psubb %%mm1, %%mm2                     \n\t"
+                "paddb %%mm7, %%mm2                     \n\t"
+                "pcmpgtb %%mm6, %%mm2                   \n\t"
+                "paddb %%mm2, %%mm0                     \n\t"
 
-		"movq (%%"REG_a", %3), %%mm2			\n\t"
+                "movq (%%"REG_a", %3), %%mm2            \n\t"
                 PMAXUB(%%mm2, %%mm4)
                 PMINUB(%%mm2, %%mm3, %%mm5)
-		"psubb %%mm2, %%mm1				\n\t"
-		"paddb %%mm7, %%mm1				\n\t"
-		"pcmpgtb %%mm6, %%mm1				\n\t"
-		"paddb %%mm1, %%mm0				\n\t"
+                "psubb %%mm2, %%mm1                     \n\t"
+                "paddb %%mm7, %%mm1                     \n\t"
+                "pcmpgtb %%mm6, %%mm1                   \n\t"
+                "paddb %%mm1, %%mm0                     \n\t"
 
-		"movq (%%"REG_a", %3, 2), %%mm1			\n\t"
+                "movq (%%"REG_a", %3, 2), %%mm1         \n\t"
                 PMAXUB(%%mm1, %%mm4)
                 PMINUB(%%mm1, %%mm3, %%mm5)
-		"psubb %%mm1, %%mm2				\n\t"
-		"paddb %%mm7, %%mm2				\n\t"
-		"pcmpgtb %%mm6, %%mm2				\n\t"
-		"paddb %%mm2, %%mm0				\n\t"
+                "psubb %%mm1, %%mm2                     \n\t"
+                "paddb %%mm7, %%mm2                     \n\t"
+                "pcmpgtb %%mm6, %%mm2                   \n\t"
+                "paddb %%mm2, %%mm0                     \n\t"
 
-		"movq (%2, %3, 8), %%mm2			\n\t"
+                "movq (%2, %3, 8), %%mm2                \n\t"
                 PMAXUB(%%mm2, %%mm4)
                 PMINUB(%%mm2, %%mm3, %%mm5)
-		"psubb %%mm2, %%mm1				\n\t"
-		"paddb %%mm7, %%mm1				\n\t"
-		"pcmpgtb %%mm6, %%mm1				\n\t"
-		"paddb %%mm1, %%mm0				\n\t"
-
-		"movq (%%"REG_a", %3, 4), %%mm1			\n\t"
-		"psubb %%mm1, %%mm2				\n\t"
-		"paddb %%mm7, %%mm2				\n\t"
-		"pcmpgtb %%mm6, %%mm2				\n\t"
-		"paddb %%mm2, %%mm0				\n\t"
-		"psubusb %%mm3, %%mm4				\n\t"
-
-		"pxor %%mm6, %%mm6				\n\t"
-                "movq %4, %%mm7					\n\t" // QP,..., QP
-		"paddusb %%mm7, %%mm7				\n\t" // 2QP ... 2QP
-		"psubusb %%mm4, %%mm7				\n\t" // Diff >=2QP -> 0
-		"pcmpeqb %%mm6, %%mm7				\n\t" // Diff < 2QP -> 0
-		"pcmpeqb %%mm6, %%mm7				\n\t" // Diff < 2QP -> 0
-		"movq %%mm7, %1					\n\t"
-
-		"movq %5, %%mm7					\n\t"
-		"punpcklbw %%mm7, %%mm7				\n\t"
-		"punpcklbw %%mm7, %%mm7				\n\t"
-		"punpcklbw %%mm7, %%mm7				\n\t"
-		"psubb %%mm0, %%mm6				\n\t"
-		"pcmpgtb %%mm7, %%mm6				\n\t"
-		"movq %%mm6, %0					\n\t"
-
-		: "=m" (eq_mask), "=m" (dc_mask)
-		: "r" (src), "r" ((long)step), "m" (c->pQPb), "m"(c->ppMode.flatnessThreshold)
-		: "%"REG_a
-		);
-
-	both_masks = dc_mask & eq_mask;
-
-	if(both_masks){
-		long offset= -8*step;
-		int64_t *temp_sums= sums;
-
-		asm volatile(
-		"movq %2, %%mm0					\n\t"  // QP,..., QP
-		"pxor %%mm4, %%mm4				\n\t"
-
-		"movq (%0), %%mm6				\n\t"
-		"movq (%0, %1), %%mm5				\n\t"
-		"movq %%mm5, %%mm1				\n\t"
-		"movq %%mm6, %%mm2				\n\t"
-		"psubusb %%mm6, %%mm5				\n\t"
-		"psubusb %%mm1, %%mm2				\n\t"
-		"por %%mm5, %%mm2				\n\t" // ABS Diff of lines
-		"psubusb %%mm2, %%mm0				\n\t" // diff >= QP -> 0
-		"pcmpeqb %%mm4, %%mm0				\n\t" // diff >= QP -> FF
-
-		"pxor %%mm6, %%mm1				\n\t"
-		"pand %%mm0, %%mm1				\n\t"
-		"pxor %%mm1, %%mm6				\n\t"
-		// 0:QP  6:First
-
-		"movq (%0, %1, 8), %%mm5			\n\t"
-		"add %1, %0					\n\t" // %0 points to line 1 not 0
-		"movq (%0, %1, 8), %%mm7			\n\t"
-		"movq %%mm5, %%mm1				\n\t"
-		"movq %%mm7, %%mm2				\n\t"
-		"psubusb %%mm7, %%mm5				\n\t"
-		"psubusb %%mm1, %%mm2				\n\t"
-		"por %%mm5, %%mm2				\n\t" // ABS Diff of lines
-		"movq %2, %%mm0					\n\t"  // QP,..., QP
-		"psubusb %%mm2, %%mm0				\n\t" // diff >= QP -> 0
-		"pcmpeqb %%mm4, %%mm0				\n\t" // diff >= QP -> FF
-
-		"pxor %%mm7, %%mm1				\n\t"
-		"pand %%mm0, %%mm1				\n\t"
-		"pxor %%mm1, %%mm7				\n\t"
-
-		"movq %%mm6, %%mm5				\n\t"
-		"punpckhbw %%mm4, %%mm6				\n\t"
-		"punpcklbw %%mm4, %%mm5				\n\t"
-		// 4:0 5/6:First 7:Last
-
-		"movq %%mm5, %%mm0				\n\t"
-		"movq %%mm6, %%mm1				\n\t"
-		"psllw $2, %%mm0				\n\t"
-		"psllw $2, %%mm1				\n\t"
-		"paddw "MANGLE(w04)", %%mm0			\n\t"
-		"paddw "MANGLE(w04)", %%mm1			\n\t"
+                "psubb %%mm2, %%mm1                     \n\t"
+                "paddb %%mm7, %%mm1                     \n\t"
+                "pcmpgtb %%mm6, %%mm1                   \n\t"
+                "paddb %%mm1, %%mm0                     \n\t"
+
+                "movq (%%"REG_a", %3, 4), %%mm1         \n\t"
+                "psubb %%mm1, %%mm2                     \n\t"
+                "paddb %%mm7, %%mm2                     \n\t"
+                "pcmpgtb %%mm6, %%mm2                   \n\t"
+                "paddb %%mm2, %%mm0                     \n\t"
+                "psubusb %%mm3, %%mm4                   \n\t"
+
+                "pxor %%mm6, %%mm6                      \n\t"
+                "movq %4, %%mm7                         \n\t" // QP,..., QP
+                "paddusb %%mm7, %%mm7                   \n\t" // 2QP ... 2QP
+                "psubusb %%mm4, %%mm7                   \n\t" // Diff >=2QP -> 0
+                "pcmpeqb %%mm6, %%mm7                   \n\t" // Diff < 2QP -> 0
+                "pcmpeqb %%mm6, %%mm7                   \n\t" // Diff < 2QP -> 0
+                "movq %%mm7, %1                         \n\t"
+
+                "movq %5, %%mm7                         \n\t"
+                "punpcklbw %%mm7, %%mm7                 \n\t"
+                "punpcklbw %%mm7, %%mm7                 \n\t"
+                "punpcklbw %%mm7, %%mm7                 \n\t"
+                "psubb %%mm0, %%mm6                     \n\t"
+                "pcmpgtb %%mm7, %%mm6                   \n\t"
+                "movq %%mm6, %0                         \n\t"
+
+                : "=m" (eq_mask), "=m" (dc_mask)
+                : "r" (src), "r" ((long)step), "m" (c->pQPb), "m"(c->ppMode.flatnessThreshold)
+                : "%"REG_a
+                );
+
+        both_masks = dc_mask & eq_mask;
+
+        if(both_masks){
+                long offset= -8*step;
+                int64_t *temp_sums= sums;
+
+                asm volatile(
+                "movq %2, %%mm0                         \n\t"  // QP,..., QP
+                "pxor %%mm4, %%mm4                      \n\t"
+
+                "movq (%0), %%mm6                       \n\t"
+                "movq (%0, %1), %%mm5                   \n\t"
+                "movq %%mm5, %%mm1                      \n\t"
+                "movq %%mm6, %%mm2                      \n\t"
+                "psubusb %%mm6, %%mm5                   \n\t"
+                "psubusb %%mm1, %%mm2                   \n\t"
+                "por %%mm5, %%mm2                       \n\t" // ABS Diff of lines
+                "psubusb %%mm2, %%mm0                   \n\t" // diff >= QP -> 0
+                "pcmpeqb %%mm4, %%mm0                   \n\t" // diff >= QP -> FF
+
+                "pxor %%mm6, %%mm1                      \n\t"
+                "pand %%mm0, %%mm1                      \n\t"
+                "pxor %%mm1, %%mm6                      \n\t"
+                // 0:QP  6:First
+
+                "movq (%0, %1, 8), %%mm5                \n\t"
+                "add %1, %0                             \n\t" // %0 points to line 1 not 0
+                "movq (%0, %1, 8), %%mm7                \n\t"
+                "movq %%mm5, %%mm1                      \n\t"
+                "movq %%mm7, %%mm2                      \n\t"
+                "psubusb %%mm7, %%mm5                   \n\t"
+                "psubusb %%mm1, %%mm2                   \n\t"
+                "por %%mm5, %%mm2                       \n\t" // ABS Diff of lines
+                "movq %2, %%mm0                         \n\t"  // QP,..., QP
+                "psubusb %%mm2, %%mm0                   \n\t" // diff >= QP -> 0
+                "pcmpeqb %%mm4, %%mm0                   \n\t" // diff >= QP -> FF
+
+                "pxor %%mm7, %%mm1                      \n\t"
+                "pand %%mm0, %%mm1                      \n\t"
+                "pxor %%mm1, %%mm7                      \n\t"
+
+                "movq %%mm6, %%mm5                      \n\t"
+                "punpckhbw %%mm4, %%mm6                 \n\t"
+                "punpcklbw %%mm4, %%mm5                 \n\t"
+                // 4:0 5/6:First 7:Last
+
+                "movq %%mm5, %%mm0                      \n\t"
+                "movq %%mm6, %%mm1                      \n\t"
+                "psllw $2, %%mm0                        \n\t"
+                "psllw $2, %%mm1                        \n\t"
+                "paddw "MANGLE(w04)", %%mm0             \n\t"
+                "paddw "MANGLE(w04)", %%mm1             \n\t"
 
 #define NEXT\
-		"movq (%0), %%mm2				\n\t"\
-		"movq (%0), %%mm3				\n\t"\
-		"add %1, %0					\n\t"\
-		"punpcklbw %%mm4, %%mm2				\n\t"\
-		"punpckhbw %%mm4, %%mm3				\n\t"\
-		"paddw %%mm2, %%mm0				\n\t"\
-		"paddw %%mm3, %%mm1				\n\t"
+                "movq (%0), %%mm2                       \n\t"\
+                "movq (%0), %%mm3                       \n\t"\
+                "add %1, %0                             \n\t"\
+                "punpcklbw %%mm4, %%mm2                 \n\t"\
+                "punpckhbw %%mm4, %%mm3                 \n\t"\
+                "paddw %%mm2, %%mm0                     \n\t"\
+                "paddw %%mm3, %%mm1                     \n\t"
 
 #define PREV\
-		"movq (%0), %%mm2				\n\t"\
-		"movq (%0), %%mm3				\n\t"\
-		"add %1, %0					\n\t"\
-		"punpcklbw %%mm4, %%mm2				\n\t"\
-		"punpckhbw %%mm4, %%mm3				\n\t"\
-		"psubw %%mm2, %%mm0				\n\t"\
-		"psubw %%mm3, %%mm1				\n\t"
-
-
-		NEXT //0
-		NEXT //1
-		NEXT //2
-		"movq %%mm0, (%3)				\n\t"
-		"movq %%mm1, 8(%3)				\n\t"
-
-		NEXT //3
-		"psubw %%mm5, %%mm0				\n\t"
-		"psubw %%mm6, %%mm1				\n\t"
-		"movq %%mm0, 16(%3)				\n\t"
-		"movq %%mm1, 24(%3)				\n\t"
-
-		NEXT //4
-		"psubw %%mm5, %%mm0				\n\t"
-		"psubw %%mm6, %%mm1				\n\t"
-		"movq %%mm0, 32(%3)				\n\t"
-		"movq %%mm1, 40(%3)				\n\t"
-
-		NEXT //5
-		"psubw %%mm5, %%mm0				\n\t"
-		"psubw %%mm6, %%mm1				\n\t"
-		"movq %%mm0, 48(%3)				\n\t"
-		"movq %%mm1, 56(%3)				\n\t"
-
-		NEXT //6
-		"psubw %%mm5, %%mm0				\n\t"
-		"psubw %%mm6, %%mm1				\n\t"
-		"movq %%mm0, 64(%3)				\n\t"
-		"movq %%mm1, 72(%3)				\n\t"
-
-		"movq %%mm7, %%mm6				\n\t"
-		"punpckhbw %%mm4, %%mm7				\n\t"
-		"punpcklbw %%mm4, %%mm6				\n\t"
-
-		NEXT //7
-		"mov %4, %0					\n\t"
-		"add %1, %0					\n\t"
-		PREV //0
-		"movq %%mm0, 80(%3)				\n\t"
-		"movq %%mm1, 88(%3)				\n\t"
-
-		PREV //1
-		"paddw %%mm6, %%mm0				\n\t"
-		"paddw %%mm7, %%mm1				\n\t"
-		"movq %%mm0, 96(%3)				\n\t"
-		"movq %%mm1, 104(%3)				\n\t"
-
-		PREV //2
-		"paddw %%mm6, %%mm0				\n\t"
-		"paddw %%mm7, %%mm1				\n\t"
-		"movq %%mm0, 112(%3)				\n\t"
-		"movq %%mm1, 120(%3)				\n\t"
-
-		PREV //3
-		"paddw %%mm6, %%mm0				\n\t"
-		"paddw %%mm7, %%mm1				\n\t"
-		"movq %%mm0, 128(%3)				\n\t"
-		"movq %%mm1, 136(%3)				\n\t"
-
-		PREV //4
-		"paddw %%mm6, %%mm0				\n\t"
-		"paddw %%mm7, %%mm1				\n\t"
-		"movq %%mm0, 144(%3)				\n\t"
-		"movq %%mm1, 152(%3)				\n\t"
-
-		"mov %4, %0					\n\t" //FIXME
-
-		: "+&r"(src)
-		: "r" ((long)step), "m" (c->pQPb), "r"(sums), "g"(src)
-		);
-
-		src+= step; // src points to begin of the 8x8 Block
-
-		asm volatile(
-		"movq %4, %%mm6					\n\t"
-		"pcmpeqb %%mm5, %%mm5				\n\t"
-		"pxor %%mm6, %%mm5				\n\t"
-		"pxor %%mm7, %%mm7				\n\t"
-
-		"1:						\n\t"
-		"movq (%1), %%mm0				\n\t"
-		"movq 8(%1), %%mm1				\n\t"
-		"paddw 32(%1), %%mm0				\n\t"
-		"paddw 40(%1), %%mm1				\n\t"
-		"movq (%0, %3), %%mm2				\n\t"
-		"movq %%mm2, %%mm3				\n\t"
-		"movq %%mm2, %%mm4				\n\t"
-		"punpcklbw %%mm7, %%mm2				\n\t"
-		"punpckhbw %%mm7, %%mm3				\n\t"
-		"paddw %%mm2, %%mm0				\n\t"
-		"paddw %%mm3, %%mm1				\n\t"
-		"paddw %%mm2, %%mm0				\n\t"
-		"paddw %%mm3, %%mm1				\n\t"
-		"psrlw $4, %%mm0				\n\t"
-		"psrlw $4, %%mm1				\n\t"
-		"packuswb %%mm1, %%mm0				\n\t"
-		"pand %%mm6, %%mm0				\n\t"
-		"pand %%mm5, %%mm4				\n\t"
-		"por %%mm4, %%mm0				\n\t"
-		"movq %%mm0, (%0, %3)				\n\t"
-		"add $16, %1					\n\t"
-		"add %2, %0					\n\t"
-		" js 1b						\n\t"
-
-		: "+r"(offset), "+r"(temp_sums)
-		: "r" ((long)step), "r"(src - offset), "m"(both_masks)
-		);
-	}else
-		src+= step; // src points to begin of the 8x8 Block
-
-	if(eq_mask != -1LL){
-		uint8_t *temp_src= src;
-		asm volatile(
-		"pxor %%mm7, %%mm7				\n\t"
-		"lea -40(%%"REG_SP"), %%"REG_c"			\n\t" // make space for 4 8-byte vars
-		"and "ALIGN_MASK", %%"REG_c"			\n\t" // align
-//	0	1	2	3	4	5	6	7	8	9
-//	%0	eax	eax+%1	eax+2%1	%0+4%1	ecx	ecx+%1	ecx+2%1	%1+8%1	ecx+4%1
-
-		"movq (%0), %%mm0				\n\t"
-		"movq %%mm0, %%mm1				\n\t"
-		"punpcklbw %%mm7, %%mm0				\n\t" // low part of line 0
-		"punpckhbw %%mm7, %%mm1				\n\t" // high part of line 0
-
-		"movq (%0, %1), %%mm2				\n\t"
-		"lea (%0, %1, 2), %%"REG_a"			\n\t"
-		"movq %%mm2, %%mm3				\n\t"
-		"punpcklbw %%mm7, %%mm2				\n\t" // low part of line 1
-		"punpckhbw %%mm7, %%mm3				\n\t" // high part of line 1
-
-		"movq (%%"REG_a"), %%mm4			\n\t"
-		"movq %%mm4, %%mm5				\n\t"
-		"punpcklbw %%mm7, %%mm4				\n\t" // low part of line 2
-		"punpckhbw %%mm7, %%mm5				\n\t" // high part of line 2
-
-		"paddw %%mm0, %%mm0				\n\t" // 2L0
-		"paddw %%mm1, %%mm1				\n\t" // 2H0
-		"psubw %%mm4, %%mm2				\n\t" // L1 - L2
-		"psubw %%mm5, %%mm3				\n\t" // H1 - H2
-		"psubw %%mm2, %%mm0				\n\t" // 2L0 - L1 + L2
-		"psubw %%mm3, %%mm1				\n\t" // 2H0 - H1 + H2
-
-		"psllw $2, %%mm2				\n\t" // 4L1 - 4L2
-		"psllw $2, %%mm3				\n\t" // 4H1 - 4H2
-		"psubw %%mm2, %%mm0				\n\t" // 2L0 - 5L1 + 5L2
-		"psubw %%mm3, %%mm1				\n\t" // 2H0 - 5H1 + 5H2
-
-		"movq (%%"REG_a", %1), %%mm2			\n\t"
-		"movq %%mm2, %%mm3				\n\t"
-		"punpcklbw %%mm7, %%mm2				\n\t" // L3
-		"punpckhbw %%mm7, %%mm3				\n\t" // H3
-
-		"psubw %%mm2, %%mm0				\n\t" // 2L0 - 5L1 + 5L2 - L3
-		"psubw %%mm3, %%mm1				\n\t" // 2H0 - 5H1 + 5H2 - H3
-		"psubw %%mm2, %%mm0				\n\t" // 2L0 - 5L1 + 5L2 - 2L3
-		"psubw %%mm3, %%mm1				\n\t" // 2H0 - 5H1 + 5H2 - 2H3
-		"movq %%mm0, (%%"REG_c")			\n\t" // 2L0 - 5L1 + 5L2 - 2L3
-		"movq %%mm1, 8(%%"REG_c")			\n\t" // 2H0 - 5H1 + 5H2 - 2H3
-
-		"movq (%%"REG_a", %1, 2), %%mm0			\n\t"
-		"movq %%mm0, %%mm1				\n\t"
-		"punpcklbw %%mm7, %%mm0				\n\t" // L4
-		"punpckhbw %%mm7, %%mm1				\n\t" // H4
-
-		"psubw %%mm0, %%mm2				\n\t" // L3 - L4
-		"psubw %%mm1, %%mm3				\n\t" // H3 - H4
-		"movq %%mm2, 16(%%"REG_c")			\n\t" // L3 - L4
-		"movq %%mm3, 24(%%"REG_c")			\n\t" // H3 - H4
-		"paddw %%mm4, %%mm4				\n\t" // 2L2
-		"paddw %%mm5, %%mm5				\n\t" // 2H2
-		"psubw %%mm2, %%mm4				\n\t" // 2L2 - L3 + L4
-		"psubw %%mm3, %%mm5				\n\t" // 2H2 - H3 + H4
-
-		"lea (%%"REG_a", %1), %0			\n\t"
-		"psllw $2, %%mm2				\n\t" // 4L3 - 4L4
-		"psllw $2, %%mm3				\n\t" // 4H3 - 4H4
-		"psubw %%mm2, %%mm4				\n\t" // 2L2 - 5L3 + 5L4
-		"psubw %%mm3, %%mm5				\n\t" // 2H2 - 5H3 + 5H4
+                "movq (%0), %%mm2                       \n\t"\
+                "movq (%0), %%mm3                       \n\t"\
+                "add %1, %0                             \n\t"\
+                "punpcklbw %%mm4, %%mm2                 \n\t"\
+                "punpckhbw %%mm4, %%mm3                 \n\t"\
+                "psubw %%mm2, %%mm0                     \n\t"\
+                "psubw %%mm3, %%mm1                     \n\t"
+
+
+                NEXT //0
+                NEXT //1
+                NEXT //2
+                "movq %%mm0, (%3)                       \n\t"
+                "movq %%mm1, 8(%3)                      \n\t"
+
+                NEXT //3
+                "psubw %%mm5, %%mm0                     \n\t"
+                "psubw %%mm6, %%mm1                     \n\t"
+                "movq %%mm0, 16(%3)                     \n\t"
+                "movq %%mm1, 24(%3)                     \n\t"
+
+                NEXT //4
+                "psubw %%mm5, %%mm0                     \n\t"
+                "psubw %%mm6, %%mm1                     \n\t"
+                "movq %%mm0, 32(%3)                     \n\t"
+                "movq %%mm1, 40(%3)                     \n\t"
+
+                NEXT //5
+                "psubw %%mm5, %%mm0                     \n\t"
+                "psubw %%mm6, %%mm1                     \n\t"
+                "movq %%mm0, 48(%3)                     \n\t"
+                "movq %%mm1, 56(%3)                     \n\t"
+
+                NEXT //6
+                "psubw %%mm5, %%mm0                     \n\t"
+                "psubw %%mm6, %%mm1                     \n\t"
+                "movq %%mm0, 64(%3)                     \n\t"
+                "movq %%mm1, 72(%3)                     \n\t"
+
+                "movq %%mm7, %%mm6                      \n\t"
+                "punpckhbw %%mm4, %%mm7                 \n\t"
+                "punpcklbw %%mm4, %%mm6                 \n\t"
+
+                NEXT //7
+                "mov %4, %0                             \n\t"
+                "add %1, %0                             \n\t"
+                PREV //0
+                "movq %%mm0, 80(%3)                     \n\t"
+                "movq %%mm1, 88(%3)                     \n\t"
+
+                PREV //1
+                "paddw %%mm6, %%mm0                     \n\t"
+                "paddw %%mm7, %%mm1                     \n\t"
+                "movq %%mm0, 96(%3)                     \n\t"
+                "movq %%mm1, 104(%3)                    \n\t"
+
+                PREV //2
+                "paddw %%mm6, %%mm0                     \n\t"
+                "paddw %%mm7, %%mm1                     \n\t"
+                "movq %%mm0, 112(%3)                    \n\t"
+                "movq %%mm1, 120(%3)                    \n\t"
+
+                PREV //3
+                "paddw %%mm6, %%mm0                     \n\t"
+                "paddw %%mm7, %%mm1                     \n\t"
+                "movq %%mm0, 128(%3)                    \n\t"
+                "movq %%mm1, 136(%3)                    \n\t"
+
+                PREV //4
+                "paddw %%mm6, %%mm0                     \n\t"
+                "paddw %%mm7, %%mm1                     \n\t"
+                "movq %%mm0, 144(%3)                    \n\t"
+                "movq %%mm1, 152(%3)                    \n\t"
+
+                "mov %4, %0                             \n\t" //FIXME
+
+                : "+&r"(src)
+                : "r" ((long)step), "m" (c->pQPb), "r"(sums), "g"(src)
+                );
+
+                src+= step; // src points to begin of the 8x8 Block
+
+                asm volatile(
+                "movq %4, %%mm6                         \n\t"
+                "pcmpeqb %%mm5, %%mm5                   \n\t"
+                "pxor %%mm6, %%mm5                      \n\t"
+                "pxor %%mm7, %%mm7                      \n\t"
+
+                "1:                                     \n\t"
+                "movq (%1), %%mm0                       \n\t"
+                "movq 8(%1), %%mm1                      \n\t"
+                "paddw 32(%1), %%mm0                    \n\t"
+                "paddw 40(%1), %%mm1                    \n\t"
+                "movq (%0, %3), %%mm2                   \n\t"
+                "movq %%mm2, %%mm3                      \n\t"
+                "movq %%mm2, %%mm4                      \n\t"
+                "punpcklbw %%mm7, %%mm2                 \n\t"
+                "punpckhbw %%mm7, %%mm3                 \n\t"
+                "paddw %%mm2, %%mm0                     \n\t"
+                "paddw %%mm3, %%mm1                     \n\t"
+                "paddw %%mm2, %%mm0                     \n\t"
+                "paddw %%mm3, %%mm1                     \n\t"
+                "psrlw $4, %%mm0                        \n\t"
+                "psrlw $4, %%mm1                        \n\t"
+                "packuswb %%mm1, %%mm0                  \n\t"
+                "pand %%mm6, %%mm0                      \n\t"
+                "pand %%mm5, %%mm4                      \n\t"
+                "por %%mm4, %%mm0                       \n\t"
+                "movq %%mm0, (%0, %3)                   \n\t"
+                "add $16, %1                            \n\t"
+                "add %2, %0                             \n\t"
+                " js 1b                                 \n\t"
+
+                : "+r"(offset), "+r"(temp_sums)
+                : "r" ((long)step), "r"(src - offset), "m"(both_masks)
+                );
+        }else
+                src+= step; // src points to begin of the 8x8 Block
+
+        if(eq_mask != -1LL){
+                uint8_t *temp_src= src;
+                asm volatile(
+                "pxor %%mm7, %%mm7                      \n\t"
+                "lea -40(%%"REG_SP"), %%"REG_c"         \n\t" // make space for 4 8-byte vars
+                "and "ALIGN_MASK", %%"REG_c"            \n\t" // align
+//      0       1       2       3       4       5       6       7       8       9
+//      %0      eax     eax+%1  eax+2%1 %0+4%1  ecx     ecx+%1  ecx+2%1 %1+8%1  ecx+4%1
+
+                "movq (%0), %%mm0                       \n\t"
+                "movq %%mm0, %%mm1                      \n\t"
+                "punpcklbw %%mm7, %%mm0                 \n\t" // low part of line 0
+                "punpckhbw %%mm7, %%mm1                 \n\t" // high part of line 0
+
+                "movq (%0, %1), %%mm2                   \n\t"
+                "lea (%0, %1, 2), %%"REG_a"             \n\t"
+                "movq %%mm2, %%mm3                      \n\t"
+                "punpcklbw %%mm7, %%mm2                 \n\t" // low part of line 1
+                "punpckhbw %%mm7, %%mm3                 \n\t" // high part of line 1
+
+                "movq (%%"REG_a"), %%mm4                \n\t"
+                "movq %%mm4, %%mm5                      \n\t"
+                "punpcklbw %%mm7, %%mm4                 \n\t" // low part of line 2
+                "punpckhbw %%mm7, %%mm5                 \n\t" // high part of line 2
+
+                "paddw %%mm0, %%mm0                     \n\t" // 2L0
+                "paddw %%mm1, %%mm1                     \n\t" // 2H0
+                "psubw %%mm4, %%mm2                     \n\t" // L1 - L2
+                "psubw %%mm5, %%mm3                     \n\t" // H1 - H2
+                "psubw %%mm2, %%mm0                     \n\t" // 2L0 - L1 + L2
+                "psubw %%mm3, %%mm1                     \n\t" // 2H0 - H1 + H2
+
+                "psllw $2, %%mm2                        \n\t" // 4L1 - 4L2
+                "psllw $2, %%mm3                        \n\t" // 4H1 - 4H2
+                "psubw %%mm2, %%mm0                     \n\t" // 2L0 - 5L1 + 5L2
+                "psubw %%mm3, %%mm1                     \n\t" // 2H0 - 5H1 + 5H2
+
+                "movq (%%"REG_a", %1), %%mm2            \n\t"
+                "movq %%mm2, %%mm3                      \n\t"
+                "punpcklbw %%mm7, %%mm2                 \n\t" // L3
+                "punpckhbw %%mm7, %%mm3                 \n\t" // H3
+
+                "psubw %%mm2, %%mm0                     \n\t" // 2L0 - 5L1 + 5L2 - L3
+                "psubw %%mm3, %%mm1                     \n\t" // 2H0 - 5H1 + 5H2 - H3
+                "psubw %%mm2, %%mm0                     \n\t" // 2L0 - 5L1 + 5L2 - 2L3
+                "psubw %%mm3, %%mm1                     \n\t" // 2H0 - 5H1 + 5H2 - 2H3
+                "movq %%mm0, (%%"REG_c")                \n\t" // 2L0 - 5L1 + 5L2 - 2L3
+                "movq %%mm1, 8(%%"REG_c")               \n\t" // 2H0 - 5H1 + 5H2 - 2H3
+
+                "movq (%%"REG_a", %1, 2), %%mm0         \n\t"
+                "movq %%mm0, %%mm1                      \n\t"
+                "punpcklbw %%mm7, %%mm0                 \n\t" // L4
+                "punpckhbw %%mm7, %%mm1                 \n\t" // H4
+
+                "psubw %%mm0, %%mm2                     \n\t" // L3 - L4
+                "psubw %%mm1, %%mm3                     \n\t" // H3 - H4
+                "movq %%mm2, 16(%%"REG_c")              \n\t" // L3 - L4
+                "movq %%mm3, 24(%%"REG_c")              \n\t" // H3 - H4
+                "paddw %%mm4, %%mm4                     \n\t" // 2L2
+                "paddw %%mm5, %%mm5                     \n\t" // 2H2
+                "psubw %%mm2, %%mm4                     \n\t" // 2L2 - L3 + L4
+                "psubw %%mm3, %%mm5                     \n\t" // 2H2 - H3 + H4
+
+                "lea (%%"REG_a", %1), %0                \n\t"
+                "psllw $2, %%mm2                        \n\t" // 4L3 - 4L4
+                "psllw $2, %%mm3                        \n\t" // 4H3 - 4H4
+                "psubw %%mm2, %%mm4                     \n\t" // 2L2 - 5L3 + 5L4
+                "psubw %%mm3, %%mm5                     \n\t" // 2H2 - 5H3 + 5H4
 //50 opcodes so far
-		"movq (%0, %1, 2), %%mm2			\n\t"
-		"movq %%mm2, %%mm3				\n\t"
-		"punpcklbw %%mm7, %%mm2				\n\t" // L5
-		"punpckhbw %%mm7, %%mm3				\n\t" // H5
-		"psubw %%mm2, %%mm4				\n\t" // 2L2 - 5L3 + 5L4 - L5
-		"psubw %%mm3, %%mm5				\n\t" // 2H2 - 5H3 + 5H4 - H5
-		"psubw %%mm2, %%mm4				\n\t" // 2L2 - 5L3 + 5L4 - 2L5
-		"psubw %%mm3, %%mm5				\n\t" // 2H2 - 5H3 + 5H4 - 2H5
-
-		"movq (%%"REG_a", %1, 4), %%mm6			\n\t"
-		"punpcklbw %%mm7, %%mm6				\n\t" // L6
-		"psubw %%mm6, %%mm2				\n\t" // L5 - L6
-		"movq (%%"REG_a", %1, 4), %%mm6			\n\t"
-		"punpckhbw %%mm7, %%mm6				\n\t" // H6
-		"psubw %%mm6, %%mm3				\n\t" // H5 - H6
-
-		"paddw %%mm0, %%mm0				\n\t" // 2L4
-		"paddw %%mm1, %%mm1				\n\t" // 2H4
-		"psubw %%mm2, %%mm0				\n\t" // 2L4 - L5 + L6
-		"psubw %%mm3, %%mm1				\n\t" // 2H4 - H5 + H6
-
-		"psllw $2, %%mm2				\n\t" // 4L5 - 4L6
-		"psllw $2, %%mm3				\n\t" // 4H5 - 4H6
-		"psubw %%mm2, %%mm0				\n\t" // 2L4 - 5L5 + 5L6
-		"psubw %%mm3, %%mm1				\n\t" // 2H4 - 5H5 + 5H6
-
-		"movq (%0, %1, 4), %%mm2			\n\t"
-		"movq %%mm2, %%mm3				\n\t"
-		"punpcklbw %%mm7, %%mm2				\n\t" // L7
-		"punpckhbw %%mm7, %%mm3				\n\t" // H7
-
-		"paddw %%mm2, %%mm2				\n\t" // 2L7
-		"paddw %%mm3, %%mm3				\n\t" // 2H7
-		"psubw %%mm2, %%mm0				\n\t" // 2L4 - 5L5 + 5L6 - 2L7
-		"psubw %%mm3, %%mm1				\n\t" // 2H4 - 5H5 + 5H6 - 2H7
-
-		"movq (%%"REG_c"), %%mm2			\n\t" // 2L0 - 5L1 + 5L2 - 2L3
-		"movq 8(%%"REG_c"), %%mm3			\n\t" // 2H0 - 5H1 + 5H2 - 2H3
+                "movq (%0, %1, 2), %%mm2                \n\t"
+                "movq %%mm2, %%mm3                      \n\t"
+                "punpcklbw %%mm7, %%mm2                 \n\t" // L5
+                "punpckhbw %%mm7, %%mm3                 \n\t" // H5
+                "psubw %%mm2, %%mm4                     \n\t" // 2L2 - 5L3 + 5L4 - L5
+                "psubw %%mm3, %%mm5                     \n\t" // 2H2 - 5H3 + 5H4 - H5
+                "psubw %%mm2, %%mm4                     \n\t" // 2L2 - 5L3 + 5L4 - 2L5
+                "psubw %%mm3, %%mm5                     \n\t" // 2H2 - 5H3 + 5H4 - 2H5
+
+                "movq (%%"REG_a", %1, 4), %%mm6         \n\t"
+                "punpcklbw %%mm7, %%mm6                 \n\t" // L6
+                "psubw %%mm6, %%mm2                     \n\t" // L5 - L6
+                "movq (%%"REG_a", %1, 4), %%mm6         \n\t"
+                "punpckhbw %%mm7, %%mm6                 \n\t" // H6
+                "psubw %%mm6, %%mm3                     \n\t" // H5 - H6
+
+                "paddw %%mm0, %%mm0                     \n\t" // 2L4
+                "paddw %%mm1, %%mm1                     \n\t" // 2H4
+                "psubw %%mm2, %%mm0                     \n\t" // 2L4 - L5 + L6
+                "psubw %%mm3, %%mm1                     \n\t" // 2H4 - H5 + H6
+
+                "psllw $2, %%mm2                        \n\t" // 4L5 - 4L6
+                "psllw $2, %%mm3                        \n\t" // 4H5 - 4H6
+                "psubw %%mm2, %%mm0                     \n\t" // 2L4 - 5L5 + 5L6
+                "psubw %%mm3, %%mm1                     \n\t" // 2H4 - 5H5 + 5H6
+
+                "movq (%0, %1, 4), %%mm2                \n\t"
+                "movq %%mm2, %%mm3                      \n\t"
+                "punpcklbw %%mm7, %%mm2                 \n\t" // L7
+                "punpckhbw %%mm7, %%mm3                 \n\t" // H7
+
+                "paddw %%mm2, %%mm2                     \n\t" // 2L7
+                "paddw %%mm3, %%mm3                     \n\t" // 2H7
+                "psubw %%mm2, %%mm0                     \n\t" // 2L4 - 5L5 + 5L6 - 2L7
+                "psubw %%mm3, %%mm1                     \n\t" // 2H4 - 5H5 + 5H6 - 2H7
+
+                "movq (%%"REG_c"), %%mm2                \n\t" // 2L0 - 5L1 + 5L2 - 2L3
+                "movq 8(%%"REG_c"), %%mm3               \n\t" // 2H0 - 5H1 + 5H2 - 2H3
 
 #ifdef HAVE_MMX2
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"psubw %%mm0, %%mm6				\n\t"
-		"pmaxsw %%mm6, %%mm0				\n\t" // |2L4 - 5L5 + 5L6 - 2L7|
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"psubw %%mm1, %%mm6				\n\t"
-		"pmaxsw %%mm6, %%mm1				\n\t" // |2H4 - 5H5 + 5H6 - 2H7|
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"psubw %%mm2, %%mm6				\n\t"
-		"pmaxsw %%mm6, %%mm2				\n\t" // |2L0 - 5L1 + 5L2 - 2L3|
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"psubw %%mm3, %%mm6				\n\t"
-		"pmaxsw %%mm6, %%mm3				\n\t" // |2H0 - 5H1 + 5H2 - 2H3|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "psubw %%mm0, %%mm6                     \n\t"
+                "pmaxsw %%mm6, %%mm0                    \n\t" // |2L4 - 5L5 + 5L6 - 2L7|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "psubw %%mm1, %%mm6                     \n\t"
+                "pmaxsw %%mm6, %%mm1                    \n\t" // |2H4 - 5H5 + 5H6 - 2H7|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "psubw %%mm2, %%mm6                     \n\t"
+                "pmaxsw %%mm6, %%mm2                    \n\t" // |2L0 - 5L1 + 5L2 - 2L3|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "psubw %%mm3, %%mm6                     \n\t"
+                "pmaxsw %%mm6, %%mm3                    \n\t" // |2H0 - 5H1 + 5H2 - 2H3|
 #else
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"pcmpgtw %%mm0, %%mm6				\n\t"
-		"pxor %%mm6, %%mm0				\n\t"
-		"psubw %%mm6, %%mm0				\n\t" // |2L4 - 5L5 + 5L6 - 2L7|
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"pcmpgtw %%mm1, %%mm6				\n\t"
-		"pxor %%mm6, %%mm1				\n\t"
-		"psubw %%mm6, %%mm1				\n\t" // |2H4 - 5H5 + 5H6 - 2H7|
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"pcmpgtw %%mm2, %%mm6				\n\t"
-		"pxor %%mm6, %%mm2				\n\t"
-		"psubw %%mm6, %%mm2				\n\t" // |2L0 - 5L1 + 5L2 - 2L3|
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"pcmpgtw %%mm3, %%mm6				\n\t"
-		"pxor %%mm6, %%mm3				\n\t"
-		"psubw %%mm6, %%mm3				\n\t" // |2H0 - 5H1 + 5H2 - 2H3|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "pcmpgtw %%mm0, %%mm6                   \n\t"
+                "pxor %%mm6, %%mm0                      \n\t"
+                "psubw %%mm6, %%mm0                     \n\t" // |2L4 - 5L5 + 5L6 - 2L7|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "pcmpgtw %%mm1, %%mm6                   \n\t"
+                "pxor %%mm6, %%mm1                      \n\t"
+                "psubw %%mm6, %%mm1                     \n\t" // |2H4 - 5H5 + 5H6 - 2H7|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "pcmpgtw %%mm2, %%mm6                   \n\t"
+                "pxor %%mm6, %%mm2                      \n\t"
+                "psubw %%mm6, %%mm2                     \n\t" // |2L0 - 5L1 + 5L2 - 2L3|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "pcmpgtw %%mm3, %%mm6                   \n\t"
+                "pxor %%mm6, %%mm3                      \n\t"
+                "psubw %%mm6, %%mm3                     \n\t" // |2H0 - 5H1 + 5H2 - 2H3|
 #endif
 
 #ifdef HAVE_MMX2
-		"pminsw %%mm2, %%mm0				\n\t"
-		"pminsw %%mm3, %%mm1				\n\t"
+                "pminsw %%mm2, %%mm0                    \n\t"
+                "pminsw %%mm3, %%mm1                    \n\t"
 #else
-		"movq %%mm0, %%mm6				\n\t"
-		"psubusw %%mm2, %%mm6				\n\t"
-		"psubw %%mm6, %%mm0				\n\t"
-		"movq %%mm1, %%mm6				\n\t"
-		"psubusw %%mm3, %%mm6				\n\t"
-		"psubw %%mm6, %%mm1				\n\t"
+                "movq %%mm0, %%mm6                      \n\t"
+                "psubusw %%mm2, %%mm6                   \n\t"
+                "psubw %%mm6, %%mm0                     \n\t"
+                "movq %%mm1, %%mm6                      \n\t"
+                "psubusw %%mm3, %%mm6                   \n\t"
+                "psubw %%mm6, %%mm1                     \n\t"
 #endif
 
-		"movd %2, %%mm2					\n\t" // QP
-		"punpcklbw %%mm7, %%mm2				\n\t"
+                "movd %2, %%mm2                         \n\t" // QP
+                "punpcklbw %%mm7, %%mm2                 \n\t"
 
-		"movq %%mm7, %%mm6				\n\t" // 0
-		"pcmpgtw %%mm4, %%mm6				\n\t" // sign(2L2 - 5L3 + 5L4 - 2L5)
-		"pxor %%mm6, %%mm4				\n\t"
-		"psubw %%mm6, %%mm4				\n\t" // |2L2 - 5L3 + 5L4 - 2L5|
-		"pcmpgtw %%mm5, %%mm7				\n\t" // sign(2H2 - 5H3 + 5H4 - 2H5)
-		"pxor %%mm7, %%mm5				\n\t"
-		"psubw %%mm7, %%mm5				\n\t" // |2H2 - 5H3 + 5H4 - 2H5|
+                "movq %%mm7, %%mm6                      \n\t" // 0
+                "pcmpgtw %%mm4, %%mm6                   \n\t" // sign(2L2 - 5L3 + 5L4 - 2L5)
+                "pxor %%mm6, %%mm4                      \n\t"
+                "psubw %%mm6, %%mm4                     \n\t" // |2L2 - 5L3 + 5L4 - 2L5|
+                "pcmpgtw %%mm5, %%mm7                   \n\t" // sign(2H2 - 5H3 + 5H4 - 2H5)
+                "pxor %%mm7, %%mm5                      \n\t"
+                "psubw %%mm7, %%mm5                     \n\t" // |2H2 - 5H3 + 5H4 - 2H5|
 // 100 opcodes
-		"psllw $3, %%mm2				\n\t" // 8QP
-		"movq %%mm2, %%mm3				\n\t" // 8QP
-		"pcmpgtw %%mm4, %%mm2				\n\t"
-		"pcmpgtw %%mm5, %%mm3				\n\t"
-		"pand %%mm2, %%mm4				\n\t"
-		"pand %%mm3, %%mm5				\n\t"
-
-
-		"psubusw %%mm0, %%mm4				\n\t" // hd
-		"psubusw %%mm1, %%mm5				\n\t" // ld
-
-
-		"movq "MANGLE(w05)", %%mm2			\n\t" // 5
-		"pmullw %%mm2, %%mm4				\n\t"
-		"pmullw %%mm2, %%mm5				\n\t"
-		"movq "MANGLE(w20)", %%mm2			\n\t" // 32
-		"paddw %%mm2, %%mm4				\n\t"
-		"paddw %%mm2, %%mm5				\n\t"
-		"psrlw $6, %%mm4				\n\t"
-		"psrlw $6, %%mm5				\n\t"
-
-		"movq 16(%%"REG_c"), %%mm0			\n\t" // L3 - L4
-		"movq 24(%%"REG_c"), %%mm1			\n\t" // H3 - H4
-
-		"pxor %%mm2, %%mm2				\n\t"
-		"pxor %%mm3, %%mm3				\n\t"
-
-		"pcmpgtw %%mm0, %%mm2				\n\t" // sign (L3-L4)
-		"pcmpgtw %%mm1, %%mm3				\n\t" // sign (H3-H4)
-		"pxor %%mm2, %%mm0				\n\t"
-		"pxor %%mm3, %%mm1				\n\t"
-		"psubw %%mm2, %%mm0				\n\t" // |L3-L4|
-		"psubw %%mm3, %%mm1				\n\t" // |H3-H4|
-		"psrlw $1, %%mm0				\n\t" // |L3 - L4|/2
-		"psrlw $1, %%mm1				\n\t" // |H3 - H4|/2
-
-		"pxor %%mm6, %%mm2				\n\t"
-		"pxor %%mm7, %%mm3				\n\t"
-		"pand %%mm2, %%mm4				\n\t"
-		"pand %%mm3, %%mm5				\n\t"
+                "psllw $3, %%mm2                        \n\t" // 8QP
+                "movq %%mm2, %%mm3                      \n\t" // 8QP
+                "pcmpgtw %%mm4, %%mm2                   \n\t"
+                "pcmpgtw %%mm5, %%mm3                   \n\t"
+                "pand %%mm2, %%mm4                      \n\t"
+                "pand %%mm3, %%mm5                      \n\t"
+
+
+                "psubusw %%mm0, %%mm4                   \n\t" // hd
+                "psubusw %%mm1, %%mm5                   \n\t" // ld
+
+
+                "movq "MANGLE(w05)", %%mm2              \n\t" // 5
+                "pmullw %%mm2, %%mm4                    \n\t"
+                "pmullw %%mm2, %%mm5                    \n\t"
+                "movq "MANGLE(w20)", %%mm2              \n\t" // 32
+                "paddw %%mm2, %%mm4                     \n\t"
+                "paddw %%mm2, %%mm5                     \n\t"
+                "psrlw $6, %%mm4                        \n\t"
+                "psrlw $6, %%mm5                        \n\t"
+
+                "movq 16(%%"REG_c"), %%mm0              \n\t" // L3 - L4
+                "movq 24(%%"REG_c"), %%mm1              \n\t" // H3 - H4
+
+                "pxor %%mm2, %%mm2                      \n\t"
+                "pxor %%mm3, %%mm3                      \n\t"
+
+                "pcmpgtw %%mm0, %%mm2                   \n\t" // sign (L3-L4)
+                "pcmpgtw %%mm1, %%mm3                   \n\t" // sign (H3-H4)
+                "pxor %%mm2, %%mm0                      \n\t"
+                "pxor %%mm3, %%mm1                      \n\t"
+                "psubw %%mm2, %%mm0                     \n\t" // |L3-L4|
+                "psubw %%mm3, %%mm1                     \n\t" // |H3-H4|
+                "psrlw $1, %%mm0                        \n\t" // |L3 - L4|/2
+                "psrlw $1, %%mm1                        \n\t" // |H3 - H4|/2
+
+                "pxor %%mm6, %%mm2                      \n\t"
+                "pxor %%mm7, %%mm3                      \n\t"
+                "pand %%mm2, %%mm4                      \n\t"
+                "pand %%mm3, %%mm5                      \n\t"
 
 #ifdef HAVE_MMX2
-		"pminsw %%mm0, %%mm4				\n\t"
-		"pminsw %%mm1, %%mm5				\n\t"
+                "pminsw %%mm0, %%mm4                    \n\t"
+                "pminsw %%mm1, %%mm5                    \n\t"
 #else
-		"movq %%mm4, %%mm2				\n\t"
-		"psubusw %%mm0, %%mm2				\n\t"
-		"psubw %%mm2, %%mm4				\n\t"
-		"movq %%mm5, %%mm2				\n\t"
-		"psubusw %%mm1, %%mm2				\n\t"
-		"psubw %%mm2, %%mm5				\n\t"
+                "movq %%mm4, %%mm2                      \n\t"
+                "psubusw %%mm0, %%mm2                   \n\t"
+                "psubw %%mm2, %%mm4                     \n\t"
+                "movq %%mm5, %%mm2                      \n\t"
+                "psubusw %%mm1, %%mm2                   \n\t"
+                "psubw %%mm2, %%mm5                     \n\t"
 #endif
-		"pxor %%mm6, %%mm4				\n\t"
-		"pxor %%mm7, %%mm5				\n\t"
-		"psubw %%mm6, %%mm4				\n\t"
-		"psubw %%mm7, %%mm5				\n\t"
-		"packsswb %%mm5, %%mm4				\n\t"
-		"movq %3, %%mm1					\n\t"
-		"pandn %%mm4, %%mm1				\n\t"
-		"movq (%0), %%mm0				\n\t"
-		"paddb   %%mm1, %%mm0				\n\t"
-		"movq %%mm0, (%0)				\n\t"
-		"movq (%0, %1), %%mm0				\n\t"
-		"psubb %%mm1, %%mm0				\n\t"
-		"movq %%mm0, (%0, %1)				\n\t"
-
-		: "+r" (temp_src)
-		: "r" ((long)step), "m" (c->pQPb), "m"(eq_mask)
-		: "%"REG_a, "%"REG_c
-		);
-	}
+                "pxor %%mm6, %%mm4                      \n\t"
+                "pxor %%mm7, %%mm5                      \n\t"
+                "psubw %%mm6, %%mm4                     \n\t"
+                "psubw %%mm7, %%mm5                     \n\t"
+                "packsswb %%mm5, %%mm4                  \n\t"
+                "movq %3, %%mm1                         \n\t"
+                "pandn %%mm4, %%mm1                     \n\t"
+                "movq (%0), %%mm0                       \n\t"
+                "paddb   %%mm1, %%mm0                   \n\t"
+                "movq %%mm0, (%0)                       \n\t"
+                "movq (%0, %1), %%mm0                   \n\t"
+                "psubb %%mm1, %%mm0                     \n\t"
+                "movq %%mm0, (%0, %1)                   \n\t"
+
+                : "+r" (temp_src)
+                : "r" ((long)step), "m" (c->pQPb), "m"(eq_mask)
+                : "%"REG_a, "%"REG_c
+                );
+        }
 /*if(step==16){
     STOP_TIMER("step16")
 }else{
@@ -3180,7 +3180,7 @@ asm volatile(
 #endif //HAVE_MMX
 
 static void RENAME(postProcess)(uint8_t src[], int srcStride, uint8_t dst[], int dstStride, int width, int height,
-	QP_STORE_T QPs[], int QPStride, int isColor, PPContext *c);
+        QP_STORE_T QPs[], int QPStride, int isColor, PPContext *c);
 
 /**
  * Copies a block from src to dst and fixes the blacklevel
@@ -3189,69 +3189,69 @@ static void RENAME(postProcess)(uint8_t src[], int srcStride, uint8_t dst[], int
 #undef SCALED_CPY
 
 static inline void RENAME(blockCopy)(uint8_t dst[], int dstStride, uint8_t src[], int srcStride,
-	int levelFix, int64_t *packedOffsetAndScale)
+        int levelFix, int64_t *packedOffsetAndScale)
 {
 #ifndef HAVE_MMX
-	int i;
+        int i;
 #endif
-	if(levelFix)
-	{
+        if(levelFix)
+        {
 #ifdef HAVE_MMX
-					asm volatile(
-						"movq (%%"REG_a"), %%mm2	\n\t" // packedYOffset
-						"movq 8(%%"REG_a"), %%mm3	\n\t" // packedYScale
-						"lea (%2,%4), %%"REG_a"	\n\t"
-						"lea (%3,%5), %%"REG_d"	\n\t"
-						"pxor %%mm4, %%mm4	\n\t"
+                asm volatile(
+                        "movq (%%"REG_a"), %%mm2        \n\t" // packedYOffset
+                        "movq 8(%%"REG_a"), %%mm3       \n\t" // packedYScale
+                        "lea (%2,%4), %%"REG_a"         \n\t"
+                        "lea (%3,%5), %%"REG_d"         \n\t"
+                        "pxor %%mm4, %%mm4              \n\t"
 #ifdef HAVE_MMX2
-#define REAL_SCALED_CPY(src1, src2, dst1, dst2)					\
-						"movq " #src1 ", %%mm0	\n\t"\
-						"movq " #src1 ", %%mm5	\n\t"\
-						"movq " #src2 ", %%mm1	\n\t"\
-						"movq " #src2 ", %%mm6	\n\t"\
-						"punpcklbw %%mm0, %%mm0 \n\t"\
-						"punpckhbw %%mm5, %%mm5 \n\t"\
-						"punpcklbw %%mm1, %%mm1 \n\t"\
-						"punpckhbw %%mm6, %%mm6 \n\t"\
-						"pmulhuw %%mm3, %%mm0	\n\t"\
-						"pmulhuw %%mm3, %%mm5	\n\t"\
-						"pmulhuw %%mm3, %%mm1	\n\t"\
-						"pmulhuw %%mm3, %%mm6	\n\t"\
-						"psubw %%mm2, %%mm0	\n\t"\
-						"psubw %%mm2, %%mm5	\n\t"\
-						"psubw %%mm2, %%mm1	\n\t"\
-						"psubw %%mm2, %%mm6	\n\t"\
-						"packuswb %%mm5, %%mm0	\n\t"\
-						"packuswb %%mm6, %%mm1	\n\t"\
-						"movq %%mm0, " #dst1 "	\n\t"\
-						"movq %%mm1, " #dst2 "	\n\t"\
+#define REAL_SCALED_CPY(src1, src2, dst1, dst2)                                                \
+                        "movq " #src1 ", %%mm0          \n\t"\
+                        "movq " #src1 ", %%mm5          \n\t"\
+                        "movq " #src2 ", %%mm1          \n\t"\
+                        "movq " #src2 ", %%mm6          \n\t"\
+                        "punpcklbw %%mm0, %%mm0         \n\t"\
+                        "punpckhbw %%mm5, %%mm5         \n\t"\
+                        "punpcklbw %%mm1, %%mm1         \n\t"\
+                        "punpckhbw %%mm6, %%mm6         \n\t"\
+                        "pmulhuw %%mm3, %%mm0           \n\t"\
+                        "pmulhuw %%mm3, %%mm5           \n\t"\
+                        "pmulhuw %%mm3, %%mm1           \n\t"\
+                        "pmulhuw %%mm3, %%mm6           \n\t"\
+                        "psubw %%mm2, %%mm0             \n\t"\
+                        "psubw %%mm2, %%mm5             \n\t"\
+                        "psubw %%mm2, %%mm1             \n\t"\
+                        "psubw %%mm2, %%mm6             \n\t"\
+                        "packuswb %%mm5, %%mm0          \n\t"\
+                        "packuswb %%mm6, %%mm1          \n\t"\
+                        "movq %%mm0, " #dst1 "          \n\t"\
+                        "movq %%mm1, " #dst2 "          \n\t"\
 
 #else //HAVE_MMX2
-#define REAL_SCALED_CPY(src1, src2, dst1, dst2)					\
-						"movq " #src1 ", %%mm0	\n\t"\
-						"movq " #src1 ", %%mm5	\n\t"\
-						"punpcklbw %%mm4, %%mm0 \n\t"\
-						"punpckhbw %%mm4, %%mm5 \n\t"\
-						"psubw %%mm2, %%mm0	\n\t"\
-						"psubw %%mm2, %%mm5	\n\t"\
-						"movq " #src2 ", %%mm1	\n\t"\
-						"psllw $6, %%mm0	\n\t"\
-						"psllw $6, %%mm5	\n\t"\
-						"pmulhw %%mm3, %%mm0	\n\t"\
-						"movq " #src2 ", %%mm6	\n\t"\
-						"pmulhw %%mm3, %%mm5	\n\t"\
-						"punpcklbw %%mm4, %%mm1 \n\t"\
-						"punpckhbw %%mm4, %%mm6 \n\t"\
-						"psubw %%mm2, %%mm1	\n\t"\
-						"psubw %%mm2, %%mm6	\n\t"\
-						"psllw $6, %%mm1	\n\t"\
-						"psllw $6, %%mm6	\n\t"\
-						"pmulhw %%mm3, %%mm1	\n\t"\
-						"pmulhw %%mm3, %%mm6	\n\t"\
-						"packuswb %%mm5, %%mm0	\n\t"\
-						"packuswb %%mm6, %%mm1	\n\t"\
-						"movq %%mm0, " #dst1 "	\n\t"\
-						"movq %%mm1, " #dst2 "	\n\t"\
+#define REAL_SCALED_CPY(src1, src2, dst1, dst2)                                        \
+                        "movq " #src1 ", %%mm0          \n\t"\
+                        "movq " #src1 ", %%mm5          \n\t"\
+                        "punpcklbw %%mm4, %%mm0         \n\t"\
+                        "punpckhbw %%mm4, %%mm5         \n\t"\
+                        "psubw %%mm2, %%mm0             \n\t"\
+                        "psubw %%mm2, %%mm5             \n\t"\
+                        "movq " #src2 ", %%mm1          \n\t"\
+                        "psllw $6, %%mm0                \n\t"\
+                        "psllw $6, %%mm5                \n\t"\
+                        "pmulhw %%mm3, %%mm0            \n\t"\
+                        "movq " #src2 ", %%mm6          \n\t"\
+                        "pmulhw %%mm3, %%mm5            \n\t"\
+                        "punpcklbw %%mm4, %%mm1         \n\t"\
+                        "punpckhbw %%mm4, %%mm6         \n\t"\
+                        "psubw %%mm2, %%mm1             \n\t"\
+                        "psubw %%mm2, %%mm6             \n\t"\
+                        "psllw $6, %%mm1                \n\t"\
+                        "psllw $6, %%mm6                \n\t"\
+                        "pmulhw %%mm3, %%mm1            \n\t"\
+                        "pmulhw %%mm3, %%mm6            \n\t"\
+                        "packuswb %%mm5, %%mm0          \n\t"\
+                        "packuswb %%mm6, %%mm1          \n\t"\
+                        "movq %%mm0, " #dst1 "          \n\t"\
+                        "movq %%mm1, " #dst2 "          \n\t"\
 
 #endif //HAVE_MMX2
 #define SCALED_CPY(src1, src2, dst1, dst2)\
@@ -3260,60 +3260,60 @@ static inline void RENAME(blockCopy)(uint8_t dst[], int dstStride, uint8_t src[]
 SCALED_CPY((%2)       , (%2, %4)      , (%3)       , (%3, %5))
 SCALED_CPY((%2, %4, 2), (%%REGa, %4, 2), (%3, %5, 2), (%%REGd, %5, 2))
 SCALED_CPY((%2, %4, 4), (%%REGa, %4, 4), (%3, %5, 4), (%%REGd, %5, 4))
-						"lea (%%"REG_a",%4,4), %%"REG_a"	\n\t"
-						"lea (%%"REG_d",%5,4), %%"REG_d"	\n\t"
+                        "lea (%%"REG_a",%4,4), %%"REG_a"        \n\t"
+                        "lea (%%"REG_d",%5,4), %%"REG_d"        \n\t"
 SCALED_CPY((%%REGa, %4), (%%REGa, %4, 2), (%%REGd, %5), (%%REGd, %5, 2))
 
 
-						: "=&a" (packedOffsetAndScale)
-						: "0" (packedOffsetAndScale),
-						"r"(src),
-						"r"(dst),
-						"r" ((long)srcStride),
-						"r" ((long)dstStride)
-						: "%"REG_d
-					);
+                        : "=&a" (packedOffsetAndScale)
+                        : "0" (packedOffsetAndScale),
+                        "r"(src),
+                        "r"(dst),
+                        "r" ((long)srcStride),
+                        "r" ((long)dstStride)
+                        : "%"REG_d
+                                        );
 #else //HAVE_MMX
-				for(i=0; i<8; i++)
-					memcpy(	&(dst[dstStride*i]),
-						&(src[srcStride*i]), BLOCK_SIZE);
+        for(i=0; i<8; i++)
+                memcpy( &(dst[dstStride*i]),
+                        &(src[srcStride*i]), BLOCK_SIZE);
 #endif //HAVE_MMX
-	}
-	else
-	{
+        }
+        else
+        {
 #ifdef HAVE_MMX
-					asm volatile(
-						"lea (%0,%2), %%"REG_a"	\n\t"
-						"lea (%1,%3), %%"REG_d"	\n\t"
+        asm volatile(
+                "lea (%0,%2), %%"REG_a"                 \n\t"
+                "lea (%1,%3), %%"REG_d"                 \n\t"
 
-#define REAL_SIMPLE_CPY(src1, src2, dst1, dst2)				\
-						"movq " #src1 ", %%mm0	\n\t"\
-						"movq " #src2 ", %%mm1	\n\t"\
-						"movq %%mm0, " #dst1 "	\n\t"\
-						"movq %%mm1, " #dst2 "	\n\t"\
+#define REAL_SIMPLE_CPY(src1, src2, dst1, dst2)                              \
+                "movq " #src1 ", %%mm0          \n\t"\
+                "movq " #src2 ", %%mm1          \n\t"\
+                "movq %%mm0, " #dst1 "          \n\t"\
+                "movq %%mm1, " #dst2 "          \n\t"\
 
 #define SIMPLE_CPY(src1, src2, dst1, dst2)\
    REAL_SIMPLE_CPY(src1, src2, dst1, dst2)
 
-SIMPLE_CPY((%0)       , (%0, %2)      , (%1)       , (%1, %3))
+SIMPLE_CPY((%0)       , (%0, %2)       , (%1)       , (%1, %3))
 SIMPLE_CPY((%0, %2, 2), (%%REGa, %2, 2), (%1, %3, 2), (%%REGd, %3, 2))
 SIMPLE_CPY((%0, %2, 4), (%%REGa, %2, 4), (%1, %3, 4), (%%REGd, %3, 4))
-						"lea (%%"REG_a",%2,4), %%"REG_a"	\n\t"
-						"lea (%%"REG_d",%3,4), %%"REG_d"	\n\t"
+                "lea (%%"REG_a",%2,4), %%"REG_a"        \n\t"
+                "lea (%%"REG_d",%3,4), %%"REG_d"        \n\t"
 SIMPLE_CPY((%%REGa, %2), (%%REGa, %2, 2), (%%REGd, %3), (%%REGd, %3, 2))
 
-						: : "r" (src),
-						"r" (dst),
-						"r" ((long)srcStride),
-						"r" ((long)dstStride)
-						: "%"REG_a, "%"REG_d
-					);
+                : : "r" (src),
+                "r" (dst),
+                "r" ((long)srcStride),
+                "r" ((long)dstStride)
+                : "%"REG_a, "%"REG_d
+        );
 #else //HAVE_MMX
-				for(i=0; i<8; i++)
-					memcpy(	&(dst[dstStride*i]),
-						&(src[srcStride*i]), BLOCK_SIZE);
+        for(i=0; i<8; i++)
+                memcpy( &(dst[dstStride*i]),
+                        &(src[srcStride*i]), BLOCK_SIZE);
 #endif //HAVE_MMX
-	}
+        }
 }
 
 /**
@@ -3322,23 +3322,23 @@ SIMPLE_CPY((%%REGa, %2), (%%REGa, %2, 2), (%%REGd, %3), (%%REGd, %3, 2))
 static inline void RENAME(duplicate)(uint8_t src[], int stride)
 {
 #ifdef HAVE_MMX
-	asm volatile(
-		"movq (%0), %%mm0		\n\t"
-		"add %1, %0			\n\t"
-		"movq %%mm0, (%0)		\n\t"
-		"movq %%mm0, (%0, %1)		\n\t"
-		"movq %%mm0, (%0, %1, 2)	\n\t"
-		: "+r" (src)
-		: "r" ((long)-stride)
-	);
+        asm volatile(
+                "movq (%0), %%mm0               \n\t"
+                "add %1, %0                     \n\t"
+                "movq %%mm0, (%0)               \n\t"
+                "movq %%mm0, (%0, %1)           \n\t"
+                "movq %%mm0, (%0, %1, 2)        \n\t"
+                : "+r" (src)
+                : "r" ((long)-stride)
+        );
 #else
-	int i;
-	uint8_t *p=src;
-	for(i=0; i<3; i++)
-	{
-		p-= stride;
-		memcpy(p, src, 8);
-	}
+        int i;
+        uint8_t *p=src;
+        for(i=0; i<3; i++)
+        {
+                p-= stride;
+                memcpy(p, src, 8);
+        }
 #endif
 }
 
@@ -3346,502 +3346,502 @@ static inline void RENAME(duplicate)(uint8_t src[], int stride)
  * Filters array of bytes (Y or U or V values)
  */
 static void RENAME(postProcess)(uint8_t src[], int srcStride, uint8_t dst[], int dstStride, int width, int height,
-	QP_STORE_T QPs[], int QPStride, int isColor, PPContext *c2)
+        QP_STORE_T QPs[], int QPStride, int isColor, PPContext *c2)
 {
-	PPContext __attribute__((aligned(8))) c= *c2; //copy to stack for faster access
-	int x,y;
+        PPContext __attribute__((aligned(8))) c= *c2; //copy to stack for faster access
+        int x,y;
 #ifdef COMPILE_TIME_MODE
-	const int mode= COMPILE_TIME_MODE;
+        const int mode= COMPILE_TIME_MODE;
 #else
-	const int mode= isColor ? c.ppMode.chromMode : c.ppMode.lumMode;
+        const int mode= isColor ? c.ppMode.chromMode : c.ppMode.lumMode;
 #endif
-	int black=0, white=255; // blackest black and whitest white in the picture
-	int QPCorrecture= 256*256;
+        int black=0, white=255; // blackest black and whitest white in the picture
+        int QPCorrecture= 256*256;
 
-	int copyAhead;
+        int copyAhead;
 #ifdef HAVE_MMX
-	int i;
+        int i;
 #endif
 
-	const int qpHShift= isColor ? 4-c.hChromaSubSample : 4;
-	const int qpVShift= isColor ? 4-c.vChromaSubSample : 4;
+        const int qpHShift= isColor ? 4-c.hChromaSubSample : 4;
+        const int qpVShift= isColor ? 4-c.vChromaSubSample : 4;
 
-	//FIXME remove
-	uint64_t * const yHistogram= c.yHistogram;
-	uint8_t * const tempSrc= srcStride > 0 ? c.tempSrc : c.tempSrc - 23*srcStride;
-	uint8_t * const tempDst= dstStride > 0 ? c.tempDst : c.tempDst - 23*dstStride;
-	//const int mbWidth= isColor ? (width+7)>>3 : (width+15)>>4;
+        //FIXME remove
+        uint64_t * const yHistogram= c.yHistogram;
+        uint8_t * const tempSrc= srcStride > 0 ? c.tempSrc : c.tempSrc - 23*srcStride;
+        uint8_t * const tempDst= dstStride > 0 ? c.tempDst : c.tempDst - 23*dstStride;
+        //const int mbWidth= isColor ? (width+7)>>3 : (width+15)>>4;
 
 #ifdef HAVE_MMX
-	for(i=0; i<57; i++){
-		int offset= ((i*c.ppMode.baseDcDiff)>>8) + 1;
-		int threshold= offset*2 + 1;
-		c.mmxDcOffset[i]= 0x7F - offset;
-		c.mmxDcThreshold[i]= 0x7F - threshold;
-		c.mmxDcOffset[i]*= 0x0101010101010101LL;
-		c.mmxDcThreshold[i]*= 0x0101010101010101LL;
-	}
+        for(i=0; i<57; i++){
+                int offset= ((i*c.ppMode.baseDcDiff)>>8) + 1;
+                int threshold= offset*2 + 1;
+                c.mmxDcOffset[i]= 0x7F - offset;
+                c.mmxDcThreshold[i]= 0x7F - threshold;
+                c.mmxDcOffset[i]*= 0x0101010101010101LL;
+                c.mmxDcThreshold[i]*= 0x0101010101010101LL;
+        }
 #endif
 
-	if(mode & CUBIC_IPOL_DEINT_FILTER) copyAhead=16;
-	else if(   (mode & LINEAR_BLEND_DEINT_FILTER)
-		|| (mode & FFMPEG_DEINT_FILTER)
-		|| (mode & LOWPASS5_DEINT_FILTER)) copyAhead=14;
-	else if(   (mode & V_DEBLOCK)
-		|| (mode & LINEAR_IPOL_DEINT_FILTER)
-		|| (mode & MEDIAN_DEINT_FILTER)
-		|| (mode & V_A_DEBLOCK)) copyAhead=13;
-	else if(mode & V_X1_FILTER) copyAhead=11;
-//	else if(mode & V_RK1_FILTER) copyAhead=10;
-	else if(mode & DERING) copyAhead=9;
-	else copyAhead=8;
-
-	copyAhead-= 8;
-
-	if(!isColor)
-	{
-		uint64_t sum= 0;
-		int i;
-		uint64_t maxClipped;
-		uint64_t clipped;
-		double scale;
-
-		c.frameNum++;
-		// first frame is fscked so we ignore it
-		if(c.frameNum == 1) yHistogram[0]= width*height/64*15/256;
-
-		for(i=0; i<256; i++)
-		{
-			sum+= yHistogram[i];
-//			printf("%d ", yHistogram[i]);
-		}
-//		printf("\n\n");
-
-		/* we allways get a completly black picture first */
-		maxClipped= (uint64_t)(sum * c.ppMode.maxClippedThreshold);
-
-		clipped= sum;
-		for(black=255; black>0; black--)
-		{
-			if(clipped < maxClipped) break;
-			clipped-= yHistogram[black];
-		}
-
-		clipped= sum;
-		for(white=0; white<256; white++)
-		{
-			if(clipped < maxClipped) break;
-			clipped-= yHistogram[white];
-		}
-
-		scale= (double)(c.ppMode.maxAllowedY - c.ppMode.minAllowedY) / (double)(white-black);
+        if(mode & CUBIC_IPOL_DEINT_FILTER) copyAhead=16;
+        else if(   (mode & LINEAR_BLEND_DEINT_FILTER)
+                || (mode & FFMPEG_DEINT_FILTER)
+                || (mode & LOWPASS5_DEINT_FILTER)) copyAhead=14;
+        else if(   (mode & V_DEBLOCK)
+                || (mode & LINEAR_IPOL_DEINT_FILTER)
+                || (mode & MEDIAN_DEINT_FILTER)
+                || (mode & V_A_DEBLOCK)) copyAhead=13;
+        else if(mode & V_X1_FILTER) copyAhead=11;
+//        else if(mode & V_RK1_FILTER) copyAhead=10;
+        else if(mode & DERING) copyAhead=9;
+        else copyAhead=8;
+
+        copyAhead-= 8;
+
+        if(!isColor)
+        {
+                uint64_t sum= 0;
+                int i;
+                uint64_t maxClipped;
+                uint64_t clipped;
+                double scale;
+
+                c.frameNum++;
+                // first frame is fscked so we ignore it
+                if(c.frameNum == 1) yHistogram[0]= width*height/64*15/256;
+
+                for(i=0; i<256; i++)
+                {
+                        sum+= yHistogram[i];
+//                        printf("%d ", yHistogram[i]);
+                }
+//                printf("\n\n");
+
+                /* we allways get a completly black picture first */
+                maxClipped= (uint64_t)(sum * c.ppMode.maxClippedThreshold);
+
+                clipped= sum;
+                for(black=255; black>0; black--)
+                {
+                        if(clipped < maxClipped) break;
+                        clipped-= yHistogram[black];
+                }
+
+                clipped= sum;
+                for(white=0; white<256; white++)
+                {
+                        if(clipped < maxClipped) break;
+                        clipped-= yHistogram[white];
+                }
+
+                scale= (double)(c.ppMode.maxAllowedY - c.ppMode.minAllowedY) / (double)(white-black);
 
 #ifdef HAVE_MMX2
-		c.packedYScale= (uint16_t)(scale*256.0 + 0.5);
-		c.packedYOffset= (((black*c.packedYScale)>>8) - c.ppMode.minAllowedY) & 0xFFFF;
+                c.packedYScale= (uint16_t)(scale*256.0 + 0.5);
+                c.packedYOffset= (((black*c.packedYScale)>>8) - c.ppMode.minAllowedY) & 0xFFFF;
 #else
-		c.packedYScale= (uint16_t)(scale*1024.0 + 0.5);
-		c.packedYOffset= (black - c.ppMode.minAllowedY) & 0xFFFF;
+                c.packedYScale= (uint16_t)(scale*1024.0 + 0.5);
+                c.packedYOffset= (black - c.ppMode.minAllowedY) & 0xFFFF;
 #endif
 
-		c.packedYOffset|= c.packedYOffset<<32;
-		c.packedYOffset|= c.packedYOffset<<16;
-
-		c.packedYScale|= c.packedYScale<<32;
-		c.packedYScale|= c.packedYScale<<16;
-
-		if(mode & LEVEL_FIX)	QPCorrecture= (int)(scale*256*256 + 0.5);
-		else			QPCorrecture= 256*256;
-	}
-	else
-	{
-		c.packedYScale= 0x0100010001000100LL;
-		c.packedYOffset= 0;
-		QPCorrecture= 256*256;
-	}
-
-	/* copy & deinterlace first row of blocks */
-	y=-BLOCK_SIZE;
-	{
-		uint8_t *srcBlock= &(src[y*srcStride]);
-		uint8_t *dstBlock= tempDst + dstStride;
-
-		// From this point on it is guranteed that we can read and write 16 lines downward
-		// finish 1 block before the next otherwise we might have a problem
-		// with the L1 Cache of the P4 ... or only a few blocks at a time or soemthing
-		for(x=0; x<width; x+=BLOCK_SIZE)
-		{
+                c.packedYOffset|= c.packedYOffset<<32;
+                c.packedYOffset|= c.packedYOffset<<16;
+
+                c.packedYScale|= c.packedYScale<<32;
+                c.packedYScale|= c.packedYScale<<16;
+
+                if(mode & LEVEL_FIX)        QPCorrecture= (int)(scale*256*256 + 0.5);
+                else                        QPCorrecture= 256*256;
+        }
+        else
+        {
+                c.packedYScale= 0x0100010001000100LL;
+                c.packedYOffset= 0;
+                QPCorrecture= 256*256;
+        }
+
+        /* copy & deinterlace first row of blocks */
+        y=-BLOCK_SIZE;
+        {
+                uint8_t *srcBlock= &(src[y*srcStride]);
+                uint8_t *dstBlock= tempDst + dstStride;
+
+                // From this point on it is guranteed that we can read and write 16 lines downward
+                // finish 1 block before the next otherwise we might have a problem
+                // with the L1 Cache of the P4 ... or only a few blocks at a time or soemthing
+                for(x=0; x<width; x+=BLOCK_SIZE)
+                {
 
 #ifdef HAVE_MMX2
 /*
-			prefetchnta(srcBlock + (((x>>2)&6) + 5)*srcStride + 32);
-			prefetchnta(srcBlock + (((x>>2)&6) + 6)*srcStride + 32);
-			prefetcht0(dstBlock + (((x>>2)&6) + 5)*dstStride + 32);
-			prefetcht0(dstBlock + (((x>>2)&6) + 6)*dstStride + 32);
+                        prefetchnta(srcBlock + (((x>>2)&6) + 5)*srcStride + 32);
+                        prefetchnta(srcBlock + (((x>>2)&6) + 6)*srcStride + 32);
+                        prefetcht0(dstBlock + (((x>>2)&6) + 5)*dstStride + 32);
+                        prefetcht0(dstBlock + (((x>>2)&6) + 6)*dstStride + 32);
 */
 
-			asm(
-				"mov %4, %%"REG_a"		\n\t"
-				"shr $2, %%"REG_a"		\n\t"
-				"and $6, %%"REG_a"		\n\t"
-				"add %5, %%"REG_a"		\n\t"
-				"mov %%"REG_a", %%"REG_d"	\n\t"
-				"imul %1, %%"REG_a"		\n\t"
-				"imul %3, %%"REG_d"		\n\t"
-				"prefetchnta 32(%%"REG_a", %0)	\n\t"
-				"prefetcht0 32(%%"REG_d", %2)	\n\t"
-				"add %1, %%"REG_a"		\n\t"
-				"add %3, %%"REG_d"		\n\t"
-				"prefetchnta 32(%%"REG_a", %0)	\n\t"
-				"prefetcht0 32(%%"REG_d", %2)	\n\t"
-			:: "r" (srcBlock), "r" ((long)srcStride), "r" (dstBlock), "r" ((long)dstStride),
-			"g" ((long)x), "g" ((long)copyAhead)
-			: "%"REG_a, "%"REG_d
-			);
+                        asm(
+                                "mov %4, %%"REG_a"              \n\t"
+                                "shr $2, %%"REG_a"              \n\t"
+                                "and $6, %%"REG_a"              \n\t"
+                                "add %5, %%"REG_a"              \n\t"
+                                "mov %%"REG_a", %%"REG_d"       \n\t"
+                                "imul %1, %%"REG_a"             \n\t"
+                                "imul %3, %%"REG_d"             \n\t"
+                                "prefetchnta 32(%%"REG_a", %0)  \n\t"
+                                "prefetcht0 32(%%"REG_d", %2)   \n\t"
+                                "add %1, %%"REG_a"              \n\t"
+                                "add %3, %%"REG_d"              \n\t"
+                                "prefetchnta 32(%%"REG_a", %0)  \n\t"
+                                "prefetcht0 32(%%"REG_d", %2)   \n\t"
+                        :: "r" (srcBlock), "r" ((long)srcStride), "r" (dstBlock), "r" ((long)dstStride),
+                        "g" ((long)x), "g" ((long)copyAhead)
+                        : "%"REG_a, "%"REG_d
+                        );
 
 #elif defined(HAVE_3DNOW)
 //FIXME check if this is faster on an 3dnow chip or if its faster without the prefetch or ...
-/*			prefetch(srcBlock + (((x>>3)&3) + 5)*srcStride + 32);
-			prefetch(srcBlock + (((x>>3)&3) + 9)*srcStride + 32);
-			prefetchw(dstBlock + (((x>>3)&3) + 5)*dstStride + 32);
-			prefetchw(dstBlock + (((x>>3)&3) + 9)*dstStride + 32);
+/*                        prefetch(srcBlock + (((x>>3)&3) + 5)*srcStride + 32);
+                        prefetch(srcBlock + (((x>>3)&3) + 9)*srcStride + 32);
+                        prefetchw(dstBlock + (((x>>3)&3) + 5)*dstStride + 32);
+                        prefetchw(dstBlock + (((x>>3)&3) + 9)*dstStride + 32);
 */
 #endif
 
-			RENAME(blockCopy)(dstBlock + dstStride*8, dstStride,
-				srcBlock + srcStride*8, srcStride, mode & LEVEL_FIX, &c.packedYOffset);
-
-			RENAME(duplicate)(dstBlock + dstStride*8, dstStride);
-
-			if(mode & LINEAR_IPOL_DEINT_FILTER)
-				RENAME(deInterlaceInterpolateLinear)(dstBlock, dstStride);
-			else if(mode & LINEAR_BLEND_DEINT_FILTER)
-				RENAME(deInterlaceBlendLinear)(dstBlock, dstStride, c.deintTemp + x);
-			else if(mode & MEDIAN_DEINT_FILTER)
-				RENAME(deInterlaceMedian)(dstBlock, dstStride);
-			else if(mode & CUBIC_IPOL_DEINT_FILTER)
-				RENAME(deInterlaceInterpolateCubic)(dstBlock, dstStride);
-			else if(mode & FFMPEG_DEINT_FILTER)
-				RENAME(deInterlaceFF)(dstBlock, dstStride, c.deintTemp + x);
-			else if(mode & LOWPASS5_DEINT_FILTER)
-				RENAME(deInterlaceL5)(dstBlock, dstStride, c.deintTemp + x, c.deintTemp + width + x);
-/*			else if(mode & CUBIC_BLEND_DEINT_FILTER)
-				RENAME(deInterlaceBlendCubic)(dstBlock, dstStride);
+                        RENAME(blockCopy)(dstBlock + dstStride*8, dstStride,
+                                srcBlock + srcStride*8, srcStride, mode & LEVEL_FIX, &c.packedYOffset);
+
+                        RENAME(duplicate)(dstBlock + dstStride*8, dstStride);
+
+                        if(mode & LINEAR_IPOL_DEINT_FILTER)
+                                RENAME(deInterlaceInterpolateLinear)(dstBlock, dstStride);
+                        else if(mode & LINEAR_BLEND_DEINT_FILTER)
+                                RENAME(deInterlaceBlendLinear)(dstBlock, dstStride, c.deintTemp + x);
+                        else if(mode & MEDIAN_DEINT_FILTER)
+                                RENAME(deInterlaceMedian)(dstBlock, dstStride);
+                        else if(mode & CUBIC_IPOL_DEINT_FILTER)
+                                RENAME(deInterlaceInterpolateCubic)(dstBlock, dstStride);
+                        else if(mode & FFMPEG_DEINT_FILTER)
+                                RENAME(deInterlaceFF)(dstBlock, dstStride, c.deintTemp + x);
+                        else if(mode & LOWPASS5_DEINT_FILTER)
+                                RENAME(deInterlaceL5)(dstBlock, dstStride, c.deintTemp + x, c.deintTemp + width + x);
+/*                        else if(mode & CUBIC_BLEND_DEINT_FILTER)
+                                RENAME(deInterlaceBlendCubic)(dstBlock, dstStride);
 */
-			dstBlock+=8;
-			srcBlock+=8;
-		}
-		if(width==ABS(dstStride))
-			linecpy(dst, tempDst + 9*dstStride, copyAhead, dstStride);
-		else
-		{
-			int i;
-			for(i=0; i<copyAhead; i++)
-			{
-				memcpy(dst + i*dstStride, tempDst + (9+i)*dstStride, width);
-			}
-		}
-	}
+                        dstBlock+=8;
+                        srcBlock+=8;
+                }
+                if(width==ABS(dstStride))
+                        linecpy(dst, tempDst + 9*dstStride, copyAhead, dstStride);
+                else
+                {
+                        int i;
+                        for(i=0; i<copyAhead; i++)
+                        {
+                                memcpy(dst + i*dstStride, tempDst + (9+i)*dstStride, width);
+                        }
+                }
+        }
 
 //printf("\n");
-	for(y=0; y<height; y+=BLOCK_SIZE)
-	{
-		//1% speedup if these are here instead of the inner loop
-		uint8_t *srcBlock= &(src[y*srcStride]);
-		uint8_t *dstBlock= &(dst[y*dstStride]);
+        for(y=0; y<height; y+=BLOCK_SIZE)
+        {
+                //1% speedup if these are here instead of the inner loop
+                uint8_t *srcBlock= &(src[y*srcStride]);
+                uint8_t *dstBlock= &(dst[y*dstStride]);
 #ifdef HAVE_MMX
-		uint8_t *tempBlock1= c.tempBlocks;
-		uint8_t *tempBlock2= c.tempBlocks + 8;
+                uint8_t *tempBlock1= c.tempBlocks;
+                uint8_t *tempBlock2= c.tempBlocks + 8;
 #endif
-		int8_t *QPptr= &QPs[(y>>qpVShift)*QPStride];
-		int8_t *nonBQPptr= &c.nonBQPTable[(y>>qpVShift)*ABS(QPStride)];
-		int QP=0;
-		/* can we mess with a 8x16 block from srcBlock/dstBlock downwards and 1 line upwards
-		   if not than use a temporary buffer */
-		if(y+15 >= height)
-		{
-			int i;
-			/* copy from line (copyAhead) to (copyAhead+7) of src, these will be copied with
-			   blockcopy to dst later */
-			linecpy(tempSrc + srcStride*copyAhead, srcBlock + srcStride*copyAhead,
-				MAX(height-y-copyAhead, 0), srcStride);
-
-			/* duplicate last line of src to fill the void upto line (copyAhead+7) */
-			for(i=MAX(height-y, 8); i<copyAhead+8; i++)
-				memcpy(tempSrc + srcStride*i, src + srcStride*(height-1), ABS(srcStride));
-
-			/* copy up to (copyAhead+1) lines of dst (line -1 to (copyAhead-1))*/
-			linecpy(tempDst, dstBlock - dstStride, MIN(height-y+1, copyAhead+1), dstStride);
-
-			/* duplicate last line of dst to fill the void upto line (copyAhead) */
-			for(i=height-y+1; i<=copyAhead; i++)
-				memcpy(tempDst + dstStride*i, dst + dstStride*(height-1), ABS(dstStride));
-
-			dstBlock= tempDst + dstStride;
-			srcBlock= tempSrc;
-		}
+                int8_t *QPptr= &QPs[(y>>qpVShift)*QPStride];
+                int8_t *nonBQPptr= &c.nonBQPTable[(y>>qpVShift)*ABS(QPStride)];
+                int QP=0;
+                /* can we mess with a 8x16 block from srcBlock/dstBlock downwards and 1 line upwards
+                   if not than use a temporary buffer */
+                if(y+15 >= height)
+                {
+                        int i;
+                        /* copy from line (copyAhead) to (copyAhead+7) of src, these will be copied with
+                           blockcopy to dst later */
+                        linecpy(tempSrc + srcStride*copyAhead, srcBlock + srcStride*copyAhead,
+                                MAX(height-y-copyAhead, 0), srcStride);
+
+                        /* duplicate last line of src to fill the void upto line (copyAhead+7) */
+                        for(i=MAX(height-y, 8); i<copyAhead+8; i++)
+                                memcpy(tempSrc + srcStride*i, src + srcStride*(height-1), ABS(srcStride));
+
+                        /* copy up to (copyAhead+1) lines of dst (line -1 to (copyAhead-1))*/
+                        linecpy(tempDst, dstBlock - dstStride, MIN(height-y+1, copyAhead+1), dstStride);
+
+                        /* duplicate last line of dst to fill the void upto line (copyAhead) */
+                        for(i=height-y+1; i<=copyAhead; i++)
+                                memcpy(tempDst + dstStride*i, dst + dstStride*(height-1), ABS(dstStride));
+
+                        dstBlock= tempDst + dstStride;
+                        srcBlock= tempSrc;
+                }
 //printf("\n");
 
-		// From this point on it is guranteed that we can read and write 16 lines downward
-		// finish 1 block before the next otherwise we might have a problem
-		// with the L1 Cache of the P4 ... or only a few blocks at a time or soemthing
-		for(x=0; x<width; x+=BLOCK_SIZE)
-		{
-			const int stride= dstStride;
+                // From this point on it is guranteed that we can read and write 16 lines downward
+                // finish 1 block before the next otherwise we might have a problem
+                // with the L1 Cache of the P4 ... or only a few blocks at a time or soemthing
+                for(x=0; x<width; x+=BLOCK_SIZE)
+                {
+                        const int stride= dstStride;
 #ifdef HAVE_MMX
-			uint8_t *tmpXchg;
+                        uint8_t *tmpXchg;
 #endif
-			if(isColor)
-			{
-				QP= QPptr[x>>qpHShift];
-				c.nonBQP= nonBQPptr[x>>qpHShift];
-			}
-			else
-			{
-				QP= QPptr[x>>4];
-				QP= (QP* QPCorrecture + 256*128)>>16;
-				c.nonBQP= nonBQPptr[x>>4];
-				c.nonBQP= (c.nonBQP* QPCorrecture + 256*128)>>16;
-				yHistogram[ srcBlock[srcStride*12 + 4] ]++;
-			}
-			c.QP= QP;
+                        if(isColor)
+                        {
+                                QP= QPptr[x>>qpHShift];
+                                c.nonBQP= nonBQPptr[x>>qpHShift];
+                        }
+                        else
+                        {
+                                QP= QPptr[x>>4];
+                                QP= (QP* QPCorrecture + 256*128)>>16;
+                                c.nonBQP= nonBQPptr[x>>4];
+                                c.nonBQP= (c.nonBQP* QPCorrecture + 256*128)>>16;
+                                yHistogram[ srcBlock[srcStride*12 + 4] ]++;
+                        }
+                        c.QP= QP;
 #ifdef HAVE_MMX
-			asm volatile(
-				"movd %1, %%mm7					\n\t"
-				"packuswb %%mm7, %%mm7				\n\t" // 0, 0, 0, QP, 0, 0, 0, QP
-				"packuswb %%mm7, %%mm7				\n\t" // 0,QP, 0, QP, 0,QP, 0, QP
-				"packuswb %%mm7, %%mm7				\n\t" // QP,..., QP
-				"movq %%mm7, %0			\n\t"
-				: "=m" (c.pQPb)
-				: "r" (QP)
-			);
+                        asm volatile(
+                                "movd %1, %%mm7         \n\t"
+                                "packuswb %%mm7, %%mm7  \n\t" // 0, 0, 0, QP, 0, 0, 0, QP
+                                "packuswb %%mm7, %%mm7  \n\t" // 0,QP, 0, QP, 0,QP, 0, QP
+                                "packuswb %%mm7, %%mm7  \n\t" // QP,..., QP
+                                "movq %%mm7, %0         \n\t"
+                                : "=m" (c.pQPb)
+                                : "r" (QP)
+                        );
 #endif
 
 
 #ifdef HAVE_MMX2
 /*
-			prefetchnta(srcBlock + (((x>>2)&6) + 5)*srcStride + 32);
-			prefetchnta(srcBlock + (((x>>2)&6) + 6)*srcStride + 32);
-			prefetcht0(dstBlock + (((x>>2)&6) + 5)*dstStride + 32);
-			prefetcht0(dstBlock + (((x>>2)&6) + 6)*dstStride + 32);
+                        prefetchnta(srcBlock + (((x>>2)&6) + 5)*srcStride + 32);
+                        prefetchnta(srcBlock + (((x>>2)&6) + 6)*srcStride + 32);
+                        prefetcht0(dstBlock + (((x>>2)&6) + 5)*dstStride + 32);
+                        prefetcht0(dstBlock + (((x>>2)&6) + 6)*dstStride + 32);
 */
 
-			asm(
-				"mov %4, %%"REG_a"		\n\t"
-				"shr $2, %%"REG_a"		\n\t"
-				"and $6, %%"REG_a"		\n\t"
-				"add %5, %%"REG_a"		\n\t"
-				"mov %%"REG_a", %%"REG_d"	\n\t"
-				"imul %1, %%"REG_a"		\n\t"
-				"imul %3, %%"REG_d"		\n\t"
-				"prefetchnta 32(%%"REG_a", %0)	\n\t"
-				"prefetcht0 32(%%"REG_d", %2)	\n\t"
-				"add %1, %%"REG_a"		\n\t"
-				"add %3, %%"REG_d"		\n\t"
-				"prefetchnta 32(%%"REG_a", %0)	\n\t"
-				"prefetcht0 32(%%"REG_d", %2)	\n\t"
-			:: "r" (srcBlock), "r" ((long)srcStride), "r" (dstBlock), "r" ((long)dstStride),
-			 "g" ((long)x), "g" ((long)copyAhead)
-			: "%"REG_a, "%"REG_d
-			);
+                        asm(
+                                "mov %4, %%"REG_a"              \n\t"
+                                "shr $2, %%"REG_a"              \n\t"
+                                "and $6, %%"REG_a"              \n\t"
+                                "add %5, %%"REG_a"              \n\t"
+                                "mov %%"REG_a", %%"REG_d"       \n\t"
+                                "imul %1, %%"REG_a"             \n\t"
+                                "imul %3, %%"REG_d"             \n\t"
+                                "prefetchnta 32(%%"REG_a", %0)  \n\t"
+                                "prefetcht0 32(%%"REG_d", %2)   \n\t"
+                                "add %1, %%"REG_a"              \n\t"
+                                "add %3, %%"REG_d"              \n\t"
+                                "prefetchnta 32(%%"REG_a", %0)  \n\t"
+                                "prefetcht0 32(%%"REG_d", %2)   \n\t"
+                        :: "r" (srcBlock), "r" ((long)srcStride), "r" (dstBlock), "r" ((long)dstStride),
+                         "g" ((long)x), "g" ((long)copyAhead)
+                        : "%"REG_a, "%"REG_d
+                        );
 
 #elif defined(HAVE_3DNOW)
 //FIXME check if this is faster on an 3dnow chip or if its faster without the prefetch or ...
-/*			prefetch(srcBlock + (((x>>3)&3) + 5)*srcStride + 32);
-			prefetch(srcBlock + (((x>>3)&3) + 9)*srcStride + 32);
-			prefetchw(dstBlock + (((x>>3)&3) + 5)*dstStride + 32);
-			prefetchw(dstBlock + (((x>>3)&3) + 9)*dstStride + 32);
+/*                        prefetch(srcBlock + (((x>>3)&3) + 5)*srcStride + 32);
+                        prefetch(srcBlock + (((x>>3)&3) + 9)*srcStride + 32);
+                        prefetchw(dstBlock + (((x>>3)&3) + 5)*dstStride + 32);
+                        prefetchw(dstBlock + (((x>>3)&3) + 9)*dstStride + 32);
 */
 #endif
 
-			RENAME(blockCopy)(dstBlock + dstStride*copyAhead, dstStride,
-				srcBlock + srcStride*copyAhead, srcStride, mode & LEVEL_FIX, &c.packedYOffset);
-
-			if(mode & LINEAR_IPOL_DEINT_FILTER)
-				RENAME(deInterlaceInterpolateLinear)(dstBlock, dstStride);
-			else if(mode & LINEAR_BLEND_DEINT_FILTER)
-				RENAME(deInterlaceBlendLinear)(dstBlock, dstStride, c.deintTemp + x);
-			else if(mode & MEDIAN_DEINT_FILTER)
-				RENAME(deInterlaceMedian)(dstBlock, dstStride);
-			else if(mode & CUBIC_IPOL_DEINT_FILTER)
-				RENAME(deInterlaceInterpolateCubic)(dstBlock, dstStride);
-			else if(mode & FFMPEG_DEINT_FILTER)
-				RENAME(deInterlaceFF)(dstBlock, dstStride, c.deintTemp + x);
-			else if(mode & LOWPASS5_DEINT_FILTER)
-				RENAME(deInterlaceL5)(dstBlock, dstStride, c.deintTemp + x, c.deintTemp + width + x);
-/*			else if(mode & CUBIC_BLEND_DEINT_FILTER)
-				RENAME(deInterlaceBlendCubic)(dstBlock, dstStride);
+                        RENAME(blockCopy)(dstBlock + dstStride*copyAhead, dstStride,
+                                srcBlock + srcStride*copyAhead, srcStride, mode & LEVEL_FIX, &c.packedYOffset);
+
+                        if(mode & LINEAR_IPOL_DEINT_FILTER)
+                                RENAME(deInterlaceInterpolateLinear)(dstBlock, dstStride);
+                        else if(mode & LINEAR_BLEND_DEINT_FILTER)
+                                RENAME(deInterlaceBlendLinear)(dstBlock, dstStride, c.deintTemp + x);
+                        else if(mode & MEDIAN_DEINT_FILTER)
+                                RENAME(deInterlaceMedian)(dstBlock, dstStride);
+                        else if(mode & CUBIC_IPOL_DEINT_FILTER)
+                                RENAME(deInterlaceInterpolateCubic)(dstBlock, dstStride);
+                        else if(mode & FFMPEG_DEINT_FILTER)
+                                RENAME(deInterlaceFF)(dstBlock, dstStride, c.deintTemp + x);
+                        else if(mode & LOWPASS5_DEINT_FILTER)
+                                RENAME(deInterlaceL5)(dstBlock, dstStride, c.deintTemp + x, c.deintTemp + width + x);
+/*                        else if(mode & CUBIC_BLEND_DEINT_FILTER)
+                                RENAME(deInterlaceBlendCubic)(dstBlock, dstStride);
 */
 
-			/* only deblock if we have 2 blocks */
-			if(y + 8 < height)
-			{
-				if(mode & V_X1_FILTER)
-					RENAME(vertX1Filter)(dstBlock, stride, &c);
-				else if(mode & V_DEBLOCK)
-				{
-					const int t= RENAME(vertClassify)(dstBlock, stride, &c);
-
-					if(t==1)
-						RENAME(doVertLowPass)(dstBlock, stride, &c);
-					else if(t==2)
-						RENAME(doVertDefFilter)(dstBlock, stride, &c);
-				}else if(mode & V_A_DEBLOCK){
-					RENAME(do_a_deblock)(dstBlock, stride, 1, &c);
-				}
-			}
+                        /* only deblock if we have 2 blocks */
+                        if(y + 8 < height)
+                        {
+                                if(mode & V_X1_FILTER)
+                                        RENAME(vertX1Filter)(dstBlock, stride, &c);
+                                else if(mode & V_DEBLOCK)
+                                {
+                                        const int t= RENAME(vertClassify)(dstBlock, stride, &c);
+
+                                        if(t==1)
+                                                RENAME(doVertLowPass)(dstBlock, stride, &c);
+                                        else if(t==2)
+                                                RENAME(doVertDefFilter)(dstBlock, stride, &c);
+                                }else if(mode & V_A_DEBLOCK){
+                                        RENAME(do_a_deblock)(dstBlock, stride, 1, &c);
+                                }
+                        }
 
 #ifdef HAVE_MMX
-			RENAME(transpose1)(tempBlock1, tempBlock2, dstBlock, dstStride);
+                        RENAME(transpose1)(tempBlock1, tempBlock2, dstBlock, dstStride);
 #endif
-			/* check if we have a previous block to deblock it with dstBlock */
-			if(x - 8 >= 0)
-			{
+                        /* check if we have a previous block to deblock it with dstBlock */
+                        if(x - 8 >= 0)
+                        {
 #ifdef HAVE_MMX
-				if(mode & H_X1_FILTER)
-					RENAME(vertX1Filter)(tempBlock1, 16, &c);
-				else if(mode & H_DEBLOCK)
-				{
+                                if(mode & H_X1_FILTER)
+                                        RENAME(vertX1Filter)(tempBlock1, 16, &c);
+                                else if(mode & H_DEBLOCK)
+                                {
 //START_TIMER
-					const int t= RENAME(vertClassify)(tempBlock1, 16, &c);
+                                        const int t= RENAME(vertClassify)(tempBlock1, 16, &c);
 //STOP_TIMER("dc & minmax")
                                         if(t==1)
-						RENAME(doVertLowPass)(tempBlock1, 16, &c);
-					else if(t==2)
-						RENAME(doVertDefFilter)(tempBlock1, 16, &c);
-				}else if(mode & H_A_DEBLOCK){
-					RENAME(do_a_deblock)(tempBlock1, 16, 1, &c);
-				}
+                                                RENAME(doVertLowPass)(tempBlock1, 16, &c);
+                                        else if(t==2)
+                                                RENAME(doVertDefFilter)(tempBlock1, 16, &c);
+                                }else if(mode & H_A_DEBLOCK){
+                                        RENAME(do_a_deblock)(tempBlock1, 16, 1, &c);
+                                }
 
-				RENAME(transpose2)(dstBlock-4, dstStride, tempBlock1 + 4*16);
+                                RENAME(transpose2)(dstBlock-4, dstStride, tempBlock1 + 4*16);
 
 #else
-				if(mode & H_X1_FILTER)
-					horizX1Filter(dstBlock-4, stride, QP);
-				else if(mode & H_DEBLOCK)
-				{
+                                if(mode & H_X1_FILTER)
+                                        horizX1Filter(dstBlock-4, stride, QP);
+                                else if(mode & H_DEBLOCK)
+                                {
 #ifdef HAVE_ALTIVEC
-					unsigned char __attribute__ ((aligned(16))) tempBlock[272];
-					transpose_16x8_char_toPackedAlign_altivec(tempBlock, dstBlock - (4 + 1), stride);
+                                        unsigned char __attribute__ ((aligned(16))) tempBlock[272];
+                                        transpose_16x8_char_toPackedAlign_altivec(tempBlock, dstBlock - (4 + 1), stride);
 
-					const int t=vertClassify_altivec(tempBlock-48, 16, &c);
-					if(t==1) {
-						doVertLowPass_altivec(tempBlock-48, 16, &c);
+                                        const int t=vertClassify_altivec(tempBlock-48, 16, &c);
+                                        if(t==1) {
+                                                doVertLowPass_altivec(tempBlock-48, 16, &c);
                                                 transpose_8x16_char_fromPackedAlign_altivec(dstBlock - (4 + 1), tempBlock, stride);
                                         }
-					else if(t==2) {
-						doVertDefFilter_altivec(tempBlock-48, 16, &c);
+                                        else if(t==2) {
+                                                doVertDefFilter_altivec(tempBlock-48, 16, &c);
                                                 transpose_8x16_char_fromPackedAlign_altivec(dstBlock - (4 + 1), tempBlock, stride);
                                         }
 #else
-					const int t= RENAME(horizClassify)(dstBlock-4, stride, &c);
+                                        const int t= RENAME(horizClassify)(dstBlock-4, stride, &c);
 
-					if(t==1)
-						RENAME(doHorizLowPass)(dstBlock-4, stride, &c);
-					else if(t==2)
-						RENAME(doHorizDefFilter)(dstBlock-4, stride, &c);
+                                        if(t==1)
+                                                RENAME(doHorizLowPass)(dstBlock-4, stride, &c);
+                                        else if(t==2)
+                                                RENAME(doHorizDefFilter)(dstBlock-4, stride, &c);
 #endif
-				}else if(mode & H_A_DEBLOCK){
-					RENAME(do_a_deblock)(dstBlock-8, 1, stride, &c);
-				}
+                                }else if(mode & H_A_DEBLOCK){
+                                        RENAME(do_a_deblock)(dstBlock-8, 1, stride, &c);
+                                }
 #endif //HAVE_MMX
-				if(mode & DERING)
-				{
-				//FIXME filter first line
-					if(y>0) RENAME(dering)(dstBlock - stride - 8, stride, &c);
-				}
-
-				if(mode & TEMP_NOISE_FILTER)
-				{
-					RENAME(tempNoiseReducer)(dstBlock-8, stride,
-						c.tempBlured[isColor] + y*dstStride + x,
-						c.tempBluredPast[isColor] + (y>>3)*256 + (x>>3),
-						c.ppMode.maxTmpNoise);
-				}
-			}
-
-			dstBlock+=8;
-			srcBlock+=8;
+                                if(mode & DERING)
+                                {
+                                //FIXME filter first line
+                                        if(y>0) RENAME(dering)(dstBlock - stride - 8, stride, &c);
+                                }
+
+                                if(mode & TEMP_NOISE_FILTER)
+                                {
+                                        RENAME(tempNoiseReducer)(dstBlock-8, stride,
+                                                c.tempBlured[isColor] + y*dstStride + x,
+                                                c.tempBluredPast[isColor] + (y>>3)*256 + (x>>3),
+                                                c.ppMode.maxTmpNoise);
+                                }
+                        }
+
+                        dstBlock+=8;
+                        srcBlock+=8;
 
 #ifdef HAVE_MMX
-			tmpXchg= tempBlock1;
-			tempBlock1= tempBlock2;
-			tempBlock2 = tmpXchg;
+                        tmpXchg= tempBlock1;
+                        tempBlock1= tempBlock2;
+                        tempBlock2 = tmpXchg;
 #endif
-		}
-
-		if(mode & DERING)
-		{
-				if(y > 0) RENAME(dering)(dstBlock - dstStride - 8, dstStride, &c);
-		}
-
-		if((mode & TEMP_NOISE_FILTER))
-		{
-			RENAME(tempNoiseReducer)(dstBlock-8, dstStride,
-				c.tempBlured[isColor] + y*dstStride + x,
-				c.tempBluredPast[isColor] + (y>>3)*256 + (x>>3),
-				c.ppMode.maxTmpNoise);
-		}
-
-		/* did we use a tmp buffer for the last lines*/
-		if(y+15 >= height)
-		{
-			uint8_t *dstBlock= &(dst[y*dstStride]);
-			if(width==ABS(dstStride))
-				linecpy(dstBlock, tempDst + dstStride, height-y, dstStride);
-			else
-			{
-				int i;
-				for(i=0; i<height-y; i++)
-				{
-					memcpy(dstBlock + i*dstStride, tempDst + (i+1)*dstStride, width);
-				}
-			}
-		}
+                }
+
+                if(mode & DERING)
+                {
+                                if(y > 0) RENAME(dering)(dstBlock - dstStride - 8, dstStride, &c);
+                }
+
+                if((mode & TEMP_NOISE_FILTER))
+                {
+                        RENAME(tempNoiseReducer)(dstBlock-8, dstStride,
+                                c.tempBlured[isColor] + y*dstStride + x,
+                                c.tempBluredPast[isColor] + (y>>3)*256 + (x>>3),
+                                c.ppMode.maxTmpNoise);
+                }
+
+                /* did we use a tmp buffer for the last lines*/
+                if(y+15 >= height)
+                {
+                        uint8_t *dstBlock= &(dst[y*dstStride]);
+                        if(width==ABS(dstStride))
+                                linecpy(dstBlock, tempDst + dstStride, height-y, dstStride);
+                        else
+                        {
+                                int i;
+                                for(i=0; i<height-y; i++)
+                                {
+                                        memcpy(dstBlock + i*dstStride, tempDst + (i+1)*dstStride, width);
+                                }
+                        }
+                }
 /*
-		for(x=0; x<width; x+=32)
-		{
-			volatile int i;
-			i+=	+ dstBlock[x + 7*dstStride] + dstBlock[x + 8*dstStride]
-				+ dstBlock[x + 9*dstStride] + dstBlock[x +10*dstStride]
-				+ dstBlock[x +11*dstStride] + dstBlock[x +12*dstStride];
-//				+ dstBlock[x +13*dstStride]
-//				+ dstBlock[x +14*dstStride] + dstBlock[x +15*dstStride];
-		}*/
-	}
+                for(x=0; x<width; x+=32)
+                {
+                        volatile int i;
+                        i+=        + dstBlock[x + 7*dstStride] + dstBlock[x + 8*dstStride]
+                                + dstBlock[x + 9*dstStride] + dstBlock[x +10*dstStride]
+                                + dstBlock[x +11*dstStride] + dstBlock[x +12*dstStride];
+//                                + dstBlock[x +13*dstStride]
+//                                + dstBlock[x +14*dstStride] + dstBlock[x +15*dstStride];
+                }*/
+        }
 #ifdef HAVE_3DNOW
-	asm volatile("femms");
+        asm volatile("femms");
 #elif defined (HAVE_MMX)
-	asm volatile("emms");
+        asm volatile("emms");
 #endif
 
 #ifdef DEBUG_BRIGHTNESS
-	if(!isColor)
-	{
-		int max=1;
-		int i;
-		for(i=0; i<256; i++)
-			if(yHistogram[i] > max) max=yHistogram[i];
-
-		for(i=1; i<256; i++)
-		{
-			int x;
-			int start=yHistogram[i-1]/(max/256+1);
-			int end=yHistogram[i]/(max/256+1);
-			int inc= end > start ? 1 : -1;
-			for(x=start; x!=end+inc; x+=inc)
-				dst[ i*dstStride + x]+=128;
-		}
-
-		for(i=0; i<100; i+=2)
-		{
-			dst[ (white)*dstStride + i]+=128;
-			dst[ (black)*dstStride + i]+=128;
-		}
-
-	}
+        if(!isColor)
+        {
+                int max=1;
+                int i;
+                for(i=0; i<256; i++)
+                        if(yHistogram[i] > max) max=yHistogram[i];
+
+                for(i=1; i<256; i++)
+                {
+                        int x;
+                        int start=yHistogram[i-1]/(max/256+1);
+                        int end=yHistogram[i]/(max/256+1);
+                        int inc= end > start ? 1 : -1;
+                        for(x=start; x!=end+inc; x+=inc)
+                                dst[ i*dstStride + x]+=128;
+                }
+
+                for(i=0; i<100; i+=2)
+                {
+                        dst[ (white)*dstStride + i]+=128;
+                        dst[ (black)*dstStride + i]+=128;
+                }
+
+        }
 #endif
 
-	*c2= c; //copy local context back
+        *c2= c; //copy local context back
 
 }