yvu9 support (other planar yuv formats with other chroma subsamplings should be trivi...
authorMichael Niedermayer <michaelni@gmx.at>
Sun, 23 Jun 2002 18:51:43 +0000 (18:51 +0000)
committerMichael Niedermayer <michaelni@gmx.at>
Sun, 23 Jun 2002 18:51:43 +0000 (18:51 +0000)
disabled broken (=new) converters (they cause random sig11)
various bugfixes

Originally committed as revision 6533 to svn://svn.mplayerhq.hu/mplayer/trunk/postproc

postproc/swscale.c
postproc/swscale_template.c

index f21776182a1de162a105adfcc8ca011b45b15d24..35de538977a1ecc01d8b36058b8db9b708120257 100644 (file)
@@ -17,8 +17,8 @@
 */
 
 /*
-  supported Input formats: YV12, I420, IYUV, YUY2, BGR32, BGR24, BGR16, BGR15, RGB32, RGB24, Y8, Y800
-  supported output formats: YV12, I420, IYUV, BGR15, BGR16, BGR24, BGR32, Y8, Y800
+  supported Input formats: YV12, I420, IYUV, YUY2, BGR32, BGR24, BGR16, BGR15, RGB32, RGB24, Y8, Y800, YVU9
+  supported output formats: YV12, I420, IYUV, BGR15, BGR16, BGR24, BGR32, Y8, Y800, YVU9
   BGR15/16 support dithering
   
   unscaled special converters
@@ -106,10 +106,10 @@ untested special converters
 #define isSupportedIn(x)  ((x)==IMGFMT_YV12 || (x)==IMGFMT_I420 || (x)==IMGFMT_YUY2 \
                        || (x)==IMGFMT_BGR32|| (x)==IMGFMT_BGR24|| (x)==IMGFMT_BGR16|| (x)==IMGFMT_BGR15\
                        || (x)==IMGFMT_RGB32|| (x)==IMGFMT_RGB24\
-                       || (x)==IMGFMT_Y800)
+                       || (x)==IMGFMT_Y800 || (x)==IMGFMT_YVU9)
 #define isSupportedOut(x) ((x)==IMGFMT_YV12 || (x)==IMGFMT_I420 \
                        || (x)==IMGFMT_BGR32|| (x)==IMGFMT_BGR24|| (x)==IMGFMT_BGR16|| (x)==IMGFMT_BGR15\
-                       || (x)==IMGFMT_Y800)
+                       || (x)==IMGFMT_Y800 || (x)==IMGFMT_YVU9)
 #define isRGB(x)       (((x)&IMGFMT_RGB_MASK)==IMGFMT_RGB)
 #define isBGR(x)       (((x)&IMGFMT_BGR_MASK)==IMGFMT_BGR)
 #define isPacked(x)    ((x)==IMGFMT_YUY2 || isRGB(x) || isBGR(x))
@@ -265,6 +265,7 @@ void in_asm_used_var_warning_killer()
 #endif
 
 static int testFormat[]={
+IMGFMT_YVU9,
 IMGFMT_YV12,
 //IMGFMT_IYUV,
 IMGFMT_I420,
@@ -335,7 +336,7 @@ static void doTest(uint8_t *ref[3], int refStride[3], int w, int h, int srcForma
        ssdU/= w*h/4;
        ssdV/= w*h/4;
        
-       if(ssdY>10 || ssdU>10 || ssdV>10){
+       if(ssdY>100 || ssdU>50 || ssdV>50){
                printf(" %s %dx%d -> %s %4dx%4d flags=%2d SSD=%5lld,%5lld,%5lld\n", 
                        vo_format_name(srcFormat), srcW, srcH, 
                        vo_format_name(dstFormat), dstW, dstH,
@@ -382,13 +383,13 @@ static void selfTest(uint8_t *src[3], int stride[3], int w, int h){
        }
 }
 
-static inline void yuv2yuvXinC(int16_t *lumFilter, int16_t **lumSrc, int lumFilterSize,
+static inline void yuv2yuvXinC(SwsContext *c, int16_t *lumFilter, int16_t **lumSrc, int lumFilterSize,
                                    int16_t *chrFilter, int16_t **chrSrc, int chrFilterSize,
-                                   uint8_t *dest, uint8_t *uDest, uint8_t *vDest, int dstW)
+                                   uint8_t *dest, uint8_t *uDest, uint8_t *vDest)
 {
        //FIXME Optimize (just quickly writen not opti..)
        int i;
-       for(i=0; i<dstW; i++)
+       for(i=0; i<c->dstW; i++)
        {
                int val=0;
                int j;
@@ -399,7 +400,7 @@ static inline void yuv2yuvXinC(int16_t *lumFilter, int16_t **lumSrc, int lumFilt
        }
 
        if(uDest != NULL)
-               for(i=0; i<(dstW>>1); i++)
+               for(i=0; i<c->chrDstW; i++)
                {
                        int u=0;
                        int v=0;
@@ -1657,7 +1658,7 @@ static void bgr24toyv12Wrapper(SwsContext *c, uint8_t* src[], int srcStride[], i
  * bring pointers in YUV order instead of YVU
  */
 static inline void orderYUV(int format, uint8_t * sortedP[], int sortedStride[], uint8_t * p[], int stride[]){
-       if(format == IMGFMT_YV12){
+       if(format == IMGFMT_YV12 || format == IMGFMT_YVU9){
                sortedP[0]= p[0];
                sortedP[1]= p[1];
                sortedP[2]= p[2];
@@ -1726,14 +1727,14 @@ static void simpleCopy(SwsContext *c, uint8_t* srcParam[], int srcStrideParam[],
                int plane;
                for(plane=0; plane<3; plane++)
                {
-                       int length= plane==0 ? c->srcW  : ((c->srcW+1)>>1);
-                       int y=      plane==0 ? srcSliceY: ((srcSliceY+1)>>1);
-                       int height= plane==0 ? srcSliceH: ((srcSliceH+1)>>1);
+                       int length= plane==0 ? c->srcW  : -((-c->srcW  )>>c->chrDstHSubSample);
+                       int y=      plane==0 ? srcSliceY: -((-srcSliceY)>>c->chrDstVSubSample);
+                       int height= plane==0 ? srcSliceH: -((-srcSliceH)>>c->chrDstVSubSample);
 
                        if((isGray(c->srcFormat) || isGray(c->dstFormat)) && plane>0)
                        {
                                if(!isGray(c->dstFormat))
-                                       memset(dst[plane], 0, dstStride[plane]*height);
+                                       memset(dst[plane], 128, dstStride[plane]*height);
                        }
                        else
                        {
@@ -1774,6 +1775,7 @@ static void getSubSampleFactors(int *h, int *v, int format){
                break;
        case IMGFMT_YV12:
        case IMGFMT_I420:
+       case IMGFMT_Y800: //FIXME remove after different subsamplings are fully implemented
                *h=1;
                *v=1;
                break;
@@ -1801,7 +1803,8 @@ SwsContext *getSwsContext(int srcW, int srcH, int srcFormat, int dstW, int dstH,
                asm volatile("emms\n\t"::: "memory");
 #endif
        if(swScale==NULL) globalInit();
-
+//srcFormat= IMGFMT_Y800;
+//srcFormat= IMGFMT_YVU9;
        /* avoid dupplicate Formats, so we dont need to check to much */
        srcFormat = remove_dup_fourcc(srcFormat);
        dstFormat = remove_dup_fourcc(dstFormat);
@@ -1852,6 +1855,38 @@ SwsContext *getSwsContext(int srcW, int srcH, int srcFormat, int dstW, int dstH,
        if(srcFilter->lumH!=NULL && srcFilter->lumH->length>1) usesFilter=1;
        if(srcFilter->chrV!=NULL && srcFilter->chrV->length>1) usesFilter=1;
        if(srcFilter->chrH!=NULL && srcFilter->chrH->length>1) usesFilter=1;
+
+       getSubSampleFactors(&c->chrSrcHSubSample, &c->chrSrcVSubSample, srcFormat);
+       getSubSampleFactors(&c->chrDstHSubSample, &c->chrDstVSubSample, dstFormat);
+
+       // reuse chroma for 2 pixles rgb/bgr unless user wants full chroma interpolation
+       if((isBGR(dstFormat) || isRGB(dstFormat)) && !(flags&SWS_FULL_CHR_H_INT)) c->chrDstHSubSample=1;
+
+       // drop eery 2. pixel for chroma calculation unless user wants full chroma
+       if((isBGR(srcFormat) || isRGB(srcFormat) || srcFormat==IMGFMT_YUY2) && !(flags&SWS_FULL_CHR_V)) 
+               c->chrSrcVSubSample=1;
+
+       // drop eery 2. pixel for chroma calculation unless user wants full chroma
+       if((isBGR(srcFormat) || isRGB(srcFormat)) && !(flags&SWS_FULL_CHR_H_INP)) 
+               c->chrSrcHSubSample=1;
+
+       c->chrIntHSubSample= c->chrDstHSubSample;
+       c->chrIntVSubSample= c->chrSrcVSubSample;
+       
+       // note the -((-x)>>y) is so that we allways round toward +inf
+       c->chrSrcW= -((-srcW) >> c->chrSrcHSubSample);
+       c->chrSrcH= -((-srcH) >> c->chrSrcVSubSample);
+       c->chrDstW= -((-dstW) >> c->chrDstHSubSample);
+       c->chrDstH= -((-dstH) >> c->chrDstVSubSample);
+/*     printf("%d %d %d %d / %d %d %d %d //\n", 
+       c->chrSrcW,
+c->chrSrcH,
+c->chrDstW,
+c->chrDstH,
+srcW,
+srcH,
+dstW,
+dstH);*/
        
        /* unscaled special Cases */
        if(unscaled && !usesFilter)
@@ -1877,7 +1912,7 @@ SwsContext *getSwsContext(int srcW, int srcH, int srcFormat, int dstW, int dstH,
                        return c;
                }
                /* yuv2bgr */
-               if(isPlanarYUV(srcFormat) && isBGR(dstFormat))
+               if((srcFormat==IMGFMT_YV12 || srcFormat==IMGFMT_I420) && isBGR(dstFormat))
                {
                        // FIXME multiple yuv2rgb converters wont work that way cuz that thing is full of globals&statics
 #ifdef WORDS_BIGENDIAN
@@ -1895,10 +1930,14 @@ SwsContext *getSwsContext(int srcW, int srcH, int srcFormat, int dstW, int dstH,
                                        vo_format_name(srcFormat), vo_format_name(dstFormat));
                        return c;
                }
-
+#if 1
                /* simple copy */
-               if(srcFormat == dstFormat 
-                  || ((isPlanarYUV(srcFormat)||isGray(srcFormat)) && (isPlanarYUV(dstFormat)||isGray(dstFormat))))
+               if(   srcFormat == dstFormat
+                  || (srcFormat==IMGFMT_YV12 && dstFormat==IMGFMT_I420)
+                  || (srcFormat==IMGFMT_I420 && dstFormat==IMGFMT_YV12)
+                  || (isPlanarYUV(srcFormat) && isGray(dstFormat))
+                  || (isPlanarYUV(dstFormat) && isGray(srcFormat))
+                 )
                {
                        c->swScale= simpleCopy;
 
@@ -1907,7 +1946,7 @@ SwsContext *getSwsContext(int srcW, int srcH, int srcFormat, int dstW, int dstH,
                                        vo_format_name(srcFormat), vo_format_name(dstFormat));
                        return c;
                }
-
+#endif
                /* bgr32to24 & rgb32to24*/
                if((srcFormat==IMGFMT_BGR32 && dstFormat==IMGFMT_BGR24)
                 ||(srcFormat==IMGFMT_RGB32 && dstFormat==IMGFMT_RGB24))
@@ -1943,7 +1982,7 @@ SwsContext *getSwsContext(int srcW, int srcH, int srcFormat, int dstW, int dstH,
                                        vo_format_name(srcFormat), vo_format_name(dstFormat));
                        return c;
                }
-               
+
                /* bgr24to32 & rgb24to32*/
                if((srcFormat==IMGFMT_BGR24 && dstFormat==IMGFMT_BGR32)
                 ||(srcFormat==IMGFMT_RGB24 && dstFormat==IMGFMT_RGB32))
@@ -2003,6 +2042,7 @@ SwsContext *getSwsContext(int srcW, int srcH, int srcFormat, int dstW, int dstH,
                        return c;
                }
 
+#if 0 //segfaults
                /* bgr15to32 */
                if((srcFormat==IMGFMT_BGR15 && dstFormat==IMGFMT_BGR32)
                 ||(srcFormat==IMGFMT_RGB15 && dstFormat==IMGFMT_RGB32))
@@ -2014,7 +2054,7 @@ SwsContext *getSwsContext(int srcW, int srcH, int srcFormat, int dstW, int dstH,
                                        vo_format_name(srcFormat), vo_format_name(dstFormat));
                        return c;
                }
-
+#endif
                /* bgr16to24 */
                if((srcFormat==IMGFMT_BGR16 && dstFormat==IMGFMT_BGR24)
                 ||(srcFormat==IMGFMT_RGB16 && dstFormat==IMGFMT_RGB24))
@@ -2027,6 +2067,7 @@ SwsContext *getSwsContext(int srcW, int srcH, int srcFormat, int dstW, int dstH,
                        return c;
                }
 
+#if 0 //segfaults
                /* bgr16to32 */
                if((srcFormat==IMGFMT_BGR16 && dstFormat==IMGFMT_BGR32)
                 ||(srcFormat==IMGFMT_RGB16 && dstFormat==IMGFMT_RGB32))
@@ -2038,7 +2079,7 @@ SwsContext *getSwsContext(int srcW, int srcH, int srcFormat, int dstW, int dstH,
                                        vo_format_name(srcFormat), vo_format_name(dstFormat));
                        return c;
                }
-
+#endif
                /* bgr24toYV12 */
                if(srcFormat==IMGFMT_BGR24 && dstFormat==IMGFMT_YV12)
                {
@@ -2063,37 +2104,6 @@ SwsContext *getSwsContext(int srcW, int srcH, int srcFormat, int dstW, int dstH,
        else
                c->canMMX2BeUsed=0;
 
-       getSubSampleFactors(&c->chrSrcHSubSample, &c->chrSrcVSubSample, srcFormat);
-       getSubSampleFactors(&c->chrDstHSubSample, &c->chrDstVSubSample, dstFormat);
-
-       // reuse chroma for 2 pixles rgb/bgr unless user wants full chroma interpolation
-       if((isBGR(dstFormat) || isRGB(dstFormat)) && !(flags&SWS_FULL_CHR_H_INT)) c->chrDstHSubSample=1;
-
-       // drop eery 2. pixel for chroma calculation unless user wants full chroma
-       if((isBGR(srcFormat) || isRGB(srcFormat) || srcFormat==IMGFMT_YUY2) && !(flags&SWS_FULL_CHR_V)) 
-               c->chrSrcVSubSample=1;
-
-       // drop eery 2. pixel for chroma calculation unless user wants full chroma
-       if((isBGR(srcFormat) || isRGB(srcFormat)) && !(flags&SWS_FULL_CHR_H_INP)) 
-               c->chrSrcHSubSample=1;
-
-       c->chrIntHSubSample= c->chrDstHSubSample;
-       c->chrIntVSubSample= c->chrSrcVSubSample;
-       
-       // note the -((-x)>>y) is so that we allways round toward +inf
-       c->chrSrcW= -((-srcW) >> c->chrSrcHSubSample);
-       c->chrSrcH= -((-srcH) >> c->chrSrcVSubSample);
-       c->chrDstW= -((-dstW) >> c->chrDstHSubSample);
-       c->chrDstH= -((-dstH) >> c->chrDstVSubSample);
-/*     printf("%d %d %d %d / %d %d %d %d //\n", 
-       c->chrSrcW,
-c->chrSrcH,
-c->chrDstW,
-c->chrDstH,
-srcW,
-srcH,
-dstW,
-dstH);*/
        c->chrXInc= ((c->chrSrcW<<16) + (c->chrDstW>>1))/c->chrDstW;
        c->chrYInc= ((c->chrSrcH<<16) + (c->chrDstH>>1))/c->chrDstH;
 
@@ -2126,7 +2136,7 @@ dstH);*/
                                 srcW      ,       dstW, filterAlign, 1<<14, flags,
                                 srcFilter->lumH, dstFilter->lumH);
                initFilter(&c->hChrFilter, &c->hChrFilterPos, &c->hChrFilterSize, c->chrXInc,
-                               (srcW+1)>>1, c->chrDstW, filterAlign, 1<<14, flags,
+                                c->chrSrcW, c->chrDstW, filterAlign, 1<<14, flags,
                                 srcFilter->chrH, dstFilter->chrH);
 
 #ifdef ARCH_X86
@@ -2151,7 +2161,7 @@ dstH);*/
                        srcH      ,        dstH, 1, (1<<12)-4, flags,
                        srcFilter->lumV, dstFilter->lumV);
        initFilter(&c->vChrFilter, &c->vChrFilterPos, &c->vChrFilterSize, c->chrYInc,
-                       (srcH+1)>>1, c->chrDstH, 1, (1<<12)-4, flags,
+                       c->chrSrcH, c->chrDstH, 1, (1<<12)-4, flags,
                         srcFilter->chrV, dstFilter->chrV);
 
        // Calculate Buffer Sizes so that they wont run out while handling these damn slices
@@ -2161,12 +2171,12 @@ dstH);*/
        {
                int chrI= i*c->chrDstH / dstH;
                int nextSlice= MAX(c->vLumFilterPos[i   ] + c->vLumFilterSize - 1,
-                                ((c->vChrFilterPos[chrI] + c->vChrFilterSize - 1)<<1));
-               nextSlice&= ~1; // Slices start at even boundaries
+                                ((c->vChrFilterPos[chrI] + c->vChrFilterSize - 1)<<c->chrSrcVSubSample));
+               nextSlice&= ~3; // Slices start at boundaries which are divisable through 4
                if(c->vLumFilterPos[i   ] + c->vLumBufSize < nextSlice)
                        c->vLumBufSize= nextSlice - c->vLumFilterPos[i   ];
-               if(c->vChrFilterPos[chrI] + c->vChrBufSize < (nextSlice>>1))
-                       c->vChrBufSize= (nextSlice>>1) - c->vChrFilterPos[chrI];
+               if(c->vChrFilterPos[chrI] + c->vChrBufSize < (nextSlice>>c->chrSrcVSubSample))
+                       c->vChrBufSize= (nextSlice>>c->chrSrcVSubSample) - c->vChrFilterPos[chrI];
        }
 
        // allocate pixbufs (we use dynamic allocation because otherwise we would need to
index 89ddb4b0e55fcdcf7763532ca7cc3dfb83cf2cf1..b50a10f71b7eee0b5d9484295185411c9895684e 100644 (file)
 
 static inline void RENAME(yuv2yuvX)(int16_t *lumFilter, int16_t **lumSrc, int lumFilterSize,
                                    int16_t *chrFilter, int16_t **chrSrc, int chrFilterSize,
-                                   uint8_t *dest, uint8_t *uDest, uint8_t *vDest, int dstW,
+                                   uint8_t *dest, uint8_t *uDest, uint8_t *vDest, int dstW, int chrDstW,
                                    int16_t * lumMmxFilter, int16_t * chrMmxFilter)
 {
 #ifdef HAVE_MMX
@@ -664,14 +664,14 @@ static inline void RENAME(yuv2yuvX)(int16_t *lumFilter, int16_t **lumSrc, int lu
                asm volatile(
                                YSCALEYUV2YV12X(0)
                                :: "m" (-chrFilterSize), "r" (chrSrc+chrFilterSize),
-                               "r" (chrMmxFilter+chrFilterSize*4), "r" (uDest), "m" (dstW>>1)
+                               "r" (chrMmxFilter+chrFilterSize*4), "r" (uDest), "m" (chrDstW)
                                : "%eax", "%edx", "%esi"
                        );
 
                asm volatile(
                                YSCALEYUV2YV12X(4096)
                                :: "m" (-chrFilterSize), "r" (chrSrc+chrFilterSize),
-                               "r" (chrMmxFilter+chrFilterSize*4), "r" (vDest), "m" (dstW>>1)
+                               "r" (chrMmxFilter+chrFilterSize*4), "r" (vDest), "m" (chrDstW)
                                : "%eax", "%edx", "%esi"
                        );
        }
@@ -683,29 +683,29 @@ static inline void RENAME(yuv2yuvX)(int16_t *lumFilter, int16_t **lumSrc, int lu
                        : "%eax", "%edx", "%esi"
                );
 #else
-yuv2yuvXinC(lumFilter, lumSrc, lumFilterSize,
+yuv2yuvXinC(c, lumFilter, lumSrc, lumFilterSize,
            chrFilter, chrSrc, chrFilterSize,
-           dest, uDest, vDest, dstW);
+           dest, uDest, vDest);
 #endif
 }
 
 static inline void RENAME(yuv2yuv1)(int16_t *lumSrc, int16_t *chrSrc,
-                                   uint8_t *dest, uint8_t *uDest, uint8_t *vDest, int dstW)
+                                   uint8_t *dest, uint8_t *uDest, uint8_t *vDest, int dstW, int chrDstW)
 {
 #ifdef HAVE_MMX
        if(uDest != NULL)
        {
                asm volatile(
                                YSCALEYUV2YV121
-                               :: "r" (chrSrc + (dstW>>1)), "r" (uDest + (dstW>>1)),
-                               "g" (-(dstW>>1))
+                               :: "r" (chrSrc + chrDstW), "r" (uDest + chrDstW),
+                               "g" (-chrDstW)
                                : "%eax"
                        );
 
                asm volatile(
                                YSCALEYUV2YV121
-                               :: "r" (chrSrc + 2048 + (dstW>>1)), "r" (vDest + (dstW>>1)),
-                               "g" (-(dstW>>1))
+                               :: "r" (chrSrc + 2048 + chrDstW), "r" (vDest + chrDstW),
+                               "g" (-chrDstW)
                                : "%eax"
                        );
        }
@@ -731,7 +731,7 @@ static inline void RENAME(yuv2yuv1)(int16_t *lumSrc, int16_t *chrSrc,
        }
 
        if(uDest != NULL)
-               for(i=0; i<(dstW>>1); i++)
+               for(i=0; i<chrDstW; i++)
                {
                        int u=chrSrc[i]>>7;
                        int v=chrSrc[i + 2048]>>7;
@@ -2582,6 +2582,7 @@ static void RENAME(swScale)(SwsContext *c, uint8_t* srcParam[], int srcStridePar
        const int dstW= c->dstW;
        const int dstH= c->dstH;
        const int chrDstW= c->chrDstW;
+       const int chrSrcW= c->chrSrcW;
        const int lumXInc= c->lumXInc;
        const int chrXInc= c->chrXInc;
        const int dstFormat= c->dstFormat;
@@ -2609,6 +2610,8 @@ static void RENAME(swScale)(SwsContext *c, uint8_t* srcParam[], int srcStridePar
        uint8_t *funnyYCode= c->funnyYCode;
        uint8_t *funnyUVCode= c->funnyUVCode;
        uint8_t *formatConvBuffer= c->formatConvBuffer;
+       const int chrSrcSliceY= srcSliceY >> c->chrSrcVSubSample;
+       const int chrSrcSliceH= -((-srcSliceH) >> c->chrSrcVSubSample);
 
        /* vars whch will change and which we need to storw back in the context */
        int dstY= c->dstY;
@@ -2629,7 +2632,7 @@ static void RENAME(swScale)(SwsContext *c, uint8_t* srcParam[], int srcStridePar
                srcStride[1]= srcStrideParam[2];
                srcStride[2]= srcStrideParam[1];
        }
-       else if(c->srcFormat==IMGFMT_YV12){
+       else if(c->srcFormat==IMGFMT_YV12 || c->srcFormat==IMGFMT_YVU9){
                src[0]= srcParam[0];
                src[1]= srcParam[1];
                src[2]= srcParam[2];
@@ -2726,7 +2729,7 @@ i--;
                ASSERT(firstChrSrcY >= lastInChrBuf - vChrBufSize + 1)
 
                // Do we have enough lines in this slice to output the dstY line
-               if(lastLumSrcY < srcSliceY + srcSliceH && lastChrSrcY < ((srcSliceY + srcSliceH + 1)>>1))
+               if(lastLumSrcY < srcSliceY + srcSliceH && lastChrSrcY < -((-srcSliceY - srcSliceH)>>c->chrSrcVSubSample))
                {
                        //Do horizontal scaling
                        while(lastInLumBuf < lastLumSrcY)
@@ -2746,16 +2749,16 @@ i--;
                        }
                        while(lastInChrBuf < lastChrSrcY)
                        {
-                               uint8_t *src1= src[1]+(lastInChrBuf + 1 - (srcSliceY>>1))*srcStride[1];
-                               uint8_t *src2= src[2]+(lastInChrBuf + 1 - (srcSliceY>>1))*srcStride[2];
+                               uint8_t *src1= src[1]+(lastInChrBuf + 1 - chrSrcSliceY)*srcStride[1];
+                               uint8_t *src2= src[2]+(lastInChrBuf + 1 - chrSrcSliceY)*srcStride[2];
                                chrBufIndex++;
                                ASSERT(chrBufIndex < 2*vChrBufSize)
-                               ASSERT(lastInChrBuf + 1 - (srcSliceY>>1) < ((srcSliceH+1)>>1))
-                               ASSERT(lastInChrBuf + 1 - (srcSliceY>>1) >= 0)
+                               ASSERT(lastInChrBuf + 1 - chrSrcSliceY < (chrSrcSliceH))
+                               ASSERT(lastInChrBuf + 1 - chrSrcSliceY >= 0)
                                //FIXME replace parameters through context struct (some at least)
 
                                if(!(isGray(srcFormat) || isGray(dstFormat)))
-                                       RENAME(hcscale)(chrPixBuf[ chrBufIndex ], chrDstW, src1, src2, (srcW+1)>>1, chrXInc,
+                                       RENAME(hcscale)(chrPixBuf[ chrBufIndex ], chrDstW, src1, src2, chrSrcW, chrXInc,
                                                flags, canMMX2BeUsed, hChrFilter, hChrFilterPos, hChrFilterSize,
                                                funnyUVCode, c->srcFormat, formatConvBuffer, 
                                                c->chrMmx2Filter, c->chrMmx2FilterPos);
@@ -2770,8 +2773,8 @@ i--;
 /*             printf("%d %d Last:%d %d LastInBuf:%d %d Index:%d %d Y:%d FSize: %d %d BSize: %d %d\n",
                        firstChrSrcY,firstLumSrcY,lastChrSrcY,lastLumSrcY,
                        lastInChrBuf,lastInLumBuf,chrBufIndex,lumBufIndex,dstY,vChrFilterSize,vLumFilterSize,
-                       vChrBufSize, vLumBufSize);
-*/
+                       vChrBufSize, vLumBufSize);*/
+
                        //Do horizontal scaling
                        while(lastInLumBuf+1 < srcSliceY + srcSliceH)
                        {
@@ -2786,17 +2789,17 @@ i--;
                                                c->lumMmx2Filter, c->lumMmx2FilterPos);
                                lastInLumBuf++;
                        }
-                       while(lastInChrBuf+1 < ((srcSliceY + srcSliceH)>>1))
+                       while(lastInChrBuf+1 < (chrSrcSliceY + chrSrcSliceH))
                        {
-                               uint8_t *src1= src[1]+(lastInChrBuf + 1 - (srcSliceY>>1))*srcStride[1];
-                               uint8_t *src2= src[2]+(lastInChrBuf + 1 - (srcSliceY>>1))*srcStride[2];
+                               uint8_t *src1= src[1]+(lastInChrBuf + 1 - chrSrcSliceY)*srcStride[1];
+                               uint8_t *src2= src[2]+(lastInChrBuf + 1 - chrSrcSliceY)*srcStride[2];
                                chrBufIndex++;
                                ASSERT(chrBufIndex < 2*vChrBufSize)
-                               ASSERT(lastInChrBuf + 1 - (srcSliceY>>1) < ((srcSliceH+1)>>1))
-                               ASSERT(lastInChrBuf + 1 - (srcSliceY>>1) >= 0)
+                               ASSERT(lastInChrBuf + 1 - chrSrcSliceY < chrSrcSliceH)
+                               ASSERT(lastInChrBuf + 1 - chrSrcSliceY >= 0)
 
                                if(!(isGray(srcFormat) || isGray(dstFormat)))
-                                       RENAME(hcscale)(chrPixBuf[ chrBufIndex ], chrDstW, src1, src2, (srcW+1)>>1, chrXInc,
+                                       RENAME(hcscale)(chrPixBuf[ chrBufIndex ], chrDstW, src1, src2, chrSrcW, chrXInc,
                                                flags, canMMX2BeUsed, hChrFilter, hChrFilterPos, hChrFilterSize,
                                                funnyUVCode, c->srcFormat, formatConvBuffer, 
                                                c->chrMmx2Filter, c->chrMmx2FilterPos);
@@ -2823,17 +2826,17 @@ i--;
                        {
                                int16_t *lumBuf = lumPixBuf[0];
                                int16_t *chrBuf= chrPixBuf[0];
-                               RENAME(yuv2yuv1)(lumBuf, chrBuf, dest, uDest, vDest, dstW);
+                               RENAME(yuv2yuv1)(lumBuf, chrBuf, dest, uDest, vDest, dstW, chrDstW);
                        }
                        else //General YV12
                        {
                                int16_t **lumSrcPtr= lumPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize;
                                int16_t **chrSrcPtr= chrPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize;
                                RENAME(yuv2yuvX)(
-                                       vLumFilter+dstY*vLumFilterSize     , lumSrcPtr, vLumFilterSize,
-                                       vChrFilter+(dstY>>1)*vChrFilterSize, chrSrcPtr, vChrFilterSize,
-                                       dest, uDest, vDest, dstW,
-                                       lumMmxFilter+dstY*vLumFilterSize*4, chrMmxFilter+(dstY>>1)*vChrFilterSize*4);
+                                       vLumFilter+dstY*vLumFilterSize   , lumSrcPtr, vLumFilterSize,
+                                       vChrFilter+chrDstY*vChrFilterSize, chrSrcPtr, vChrFilterSize,
+                                       dest, uDest, vDest, dstW, chrDstW,
+                                       lumMmxFilter+dstY*vLumFilterSize*4, chrMmxFilter+chrDstY*vChrFilterSize*4);
                        }
                }
                else
@@ -2874,11 +2877,11 @@ i--;
                int16_t **chrSrcPtr= chrPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize;
                if(isPlanarYUV(dstFormat)) //YV12
                {
-                       if(dstY&1) uDest=vDest= NULL; //FIXME split functions in lumi / chromi
-                       yuv2yuvXinC(
-                               vLumFilter+dstY*vLumFilterSize     , lumSrcPtr, vLumFilterSize,
-                               vChrFilter+(dstY>>1)*vChrFilterSize, chrSrcPtr, vChrFilterSize,
-                               dest, uDest, vDest, dstW);
+                       if(dstY&1) uDest=vDest= NULL;
+                       yuv2yuvXinC(c, 
+                               vLumFilter+dstY*vLumFilterSize   , lumSrcPtr, vLumFilterSize,
+                               vChrFilter+chrDstY*vChrFilterSize, chrSrcPtr, vChrFilterSize,
+                               dest, uDest, vDest);
                }
                else
                {