[SCM] libav/experimental: Unscaled converters for YUYV->YUV420P YUYV->YUV422P UYVY->YUV420P UYVY->YUV422P
siretart at users.alioth.debian.org
siretart at users.alioth.debian.org
Sun Jun 30 16:41:41 UTC 2013
The following commit has been merged in the experimental branch:
commit 0411072ee3c1b1b727cd6f76eda48c4c3b84103b
Author: Michael Niedermayer <michaelni at gmx.at>
Date: Thu Mar 19 03:45:29 2009 +0000
Unscaled converters for
YUYV->YUV420P
YUYV->YUV422P
UYVY->YUV420P
UYVY->YUV422P
Originally committed as revision 28997 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale
diff --git a/libswscale/rgb2rgb.c b/libswscale/rgb2rgb.c
index ad69265..59b8e30 100644
--- a/libswscale/rgb2rgb.c
+++ b/libswscale/rgb2rgb.c
@@ -87,6 +87,19 @@ void (*yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2, const uint8_t *sr
long width, long height,
long srcStride1, long srcStride2,
long srcStride3, long dstStride);
+void (*uyvytoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src,
+ long width, long height,
+ long lumStride, long chromStride, long srcStride);
+void (*uyvytoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src,
+ long width, long height,
+ long lumStride, long chromStride, long srcStride);
+void (*yuyvtoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src,
+ long width, long height,
+ long lumStride, long chromStride, long srcStride);
+void (*yuyvtoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src,
+ long width, long height,
+ long lumStride, long chromStride, long srcStride);
+
#if ARCH_X86 && CONFIG_GPL
DECLARE_ASM_CONST(8, uint64_t, mmx_null) = 0x0000000000000000ULL;
diff --git a/libswscale/rgb2rgb.h b/libswscale/rgb2rgb.h
index df912c8..3850ef2 100644
--- a/libswscale/rgb2rgb.h
+++ b/libswscale/rgb2rgb.h
@@ -142,6 +142,20 @@ extern void (*yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2, const uint
long srcStride1, long srcStride2,
long srcStride3, long dstStride);
+
+extern void (*uyvytoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src,
+ long width, long height,
+ long lumStride, long chromStride, long srcStride);
+extern void (*uyvytoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src,
+ long width, long height,
+ long lumStride, long chromStride, long srcStride);
+extern void (*yuyvtoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src,
+ long width, long height,
+ long lumStride, long chromStride, long srcStride);
+extern void (*yuyvtoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src,
+ long width, long height,
+ long lumStride, long chromStride, long srcStride);
+
void sws_rgb2rgb_init(int flags);
#endif /* SWSCALE_RGB2RGB_H */
diff --git a/libswscale/rgb2rgb_template.c b/libswscale/rgb2rgb_template.c
index fb4ac23..fa6de09 100644
--- a/libswscale/rgb2rgb_template.c
+++ b/libswscale/rgb2rgb_template.c
@@ -2701,6 +2701,245 @@ static inline void RENAME(yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2
#endif
}
+static void RENAME(extract_even)(const uint8_t *src, uint8_t *dst, x86_reg count)
+{
+ dst += count;
+ src += 2*count;
+ count= - count;
+
+#if HAVE_MMX
+ if(count <= -16){
+ count += 15;
+ __asm__ volatile(
+ "pcmpeqw %%mm7, %%mm7 \n\t"
+ "psrlw $8, %%mm7 \n\t"
+ "1: \n\t"
+ "movq -30(%1, %0, 2), %%mm0 \n\t"
+ "movq -22(%1, %0, 2), %%mm1 \n\t"
+ "movq -14(%1, %0, 2), %%mm2 \n\t"
+ "movq -6(%1, %0, 2), %%mm3 \n\t"
+ "pand %%mm7, %%mm0 \n\t"
+ "pand %%mm7, %%mm1 \n\t"
+ "pand %%mm7, %%mm2 \n\t"
+ "pand %%mm7, %%mm3 \n\t"
+ "packuswb %%mm1, %%mm0 \n\t"
+ "packuswb %%mm3, %%mm2 \n\t"
+ MOVNTQ" %%mm0,-15(%2, %0) \n\t"
+ MOVNTQ" %%mm2,- 7(%2, %0) \n\t"
+ "add $16, %0 \n\t"
+ " js 1b \n\t"
+ : "+r"(count)
+ : "r"(src), "r"(dst)
+ );
+ count -= 15;
+ }
+#endif
+ while(count<0){
+ dst[count]= src[2*count];
+ count++;
+ }
+}
+
+static void RENAME(extract_even2)(const uint8_t *src, uint8_t *dst0, uint8_t *dst1, x86_reg count)
+{
+ dst0+= count;
+ dst1+= count;
+ src += 4*count;
+ count= - count;
+#if HAVE_MMX
+ if(count <= -8){
+ count += 7;
+ __asm__ volatile(
+ "pcmpeqw %%mm7, %%mm7 \n\t"
+ "psrlw $8, %%mm7 \n\t"
+ "1: \n\t"
+ "movq -28(%1, %0, 4), %%mm0 \n\t"
+ "movq -20(%1, %0, 4), %%mm1 \n\t"
+ "movq -12(%1, %0, 4), %%mm2 \n\t"
+ "movq -4(%1, %0, 4), %%mm3 \n\t"
+ "pand %%mm7, %%mm0 \n\t"
+ "pand %%mm7, %%mm1 \n\t"
+ "pand %%mm7, %%mm2 \n\t"
+ "pand %%mm7, %%mm3 \n\t"
+ "packuswb %%mm1, %%mm0 \n\t"
+ "packuswb %%mm3, %%mm2 \n\t"
+ "movq %%mm0, %%mm1 \n\t"
+ "movq %%mm2, %%mm3 \n\t"
+ "psrlw $8, %%mm0 \n\t"
+ "psrlw $8, %%mm2 \n\t"
+ "pand %%mm7, %%mm1 \n\t"
+ "pand %%mm7, %%mm3 \n\t"
+ "packuswb %%mm2, %%mm0 \n\t"
+ "packuswb %%mm3, %%mm1 \n\t"
+ MOVNTQ" %%mm0,- 7(%3, %0) \n\t"
+ MOVNTQ" %%mm1,- 7(%2, %0) \n\t"
+ "add $8, %0 \n\t"
+ " js 1b \n\t"
+ : "+r"(count)
+ : "r"(src), "r"(dst0), "r"(dst1)
+ );
+ count -= 7;
+ }
+#endif
+ while(count<0){
+ dst0[count]= src[4*count+0];
+ dst1[count]= src[4*count+2];
+ count++;
+ }
+}
+
+static void RENAME(extract_odd2)(const uint8_t *src, uint8_t *dst0, uint8_t *dst1, x86_reg count)
+{
+ dst0+= count;
+ dst1+= count;
+ src += 4*count;
+ count= - count;
+#if HAVE_MMX
+ if(count <= -8){
+ count += 7;
+ __asm__ volatile(
+ "pcmpeqw %%mm7, %%mm7 \n\t"
+ "psrlw $8, %%mm7 \n\t"
+ "1: \n\t"
+ "movq -28(%1, %0, 4), %%mm0 \n\t"
+ "movq -20(%1, %0, 4), %%mm1 \n\t"
+ "movq -12(%1, %0, 4), %%mm2 \n\t"
+ "movq -4(%1, %0, 4), %%mm3 \n\t"
+ "psrlw $8, %%mm0 \n\t"
+ "psrlw $8, %%mm1 \n\t"
+ "psrlw $8, %%mm2 \n\t"
+ "psrlw $8, %%mm3 \n\t"
+ "packuswb %%mm1, %%mm0 \n\t"
+ "packuswb %%mm3, %%mm2 \n\t"
+ "movq %%mm0, %%mm1 \n\t"
+ "movq %%mm2, %%mm3 \n\t"
+ "psrlw $8, %%mm0 \n\t"
+ "psrlw $8, %%mm2 \n\t"
+ "pand %%mm7, %%mm1 \n\t"
+ "pand %%mm7, %%mm3 \n\t"
+ "packuswb %%mm2, %%mm0 \n\t"
+ "packuswb %%mm3, %%mm1 \n\t"
+ MOVNTQ" %%mm0,- 7(%3, %0) \n\t"
+ MOVNTQ" %%mm1,- 7(%2, %0) \n\t"
+ "add $8, %0 \n\t"
+ " js 1b \n\t"
+ : "+r"(count)
+ : "r"(src), "r"(dst0), "r"(dst1)
+ );
+ count -= 7;
+ }
+#endif
+ while(count<0){
+ dst0[count]= src[4*count+0];
+ dst1[count]= src[4*count+2];
+ count++;
+ }
+}
+
+static void RENAME(yuyvtoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src,
+ long width, long height,
+ long lumStride, long chromStride, long srcStride)
+{
+ long y;
+ const long chromWidth= -((-width)>>1);
+
+ for (y=0; y<height; y++){
+ RENAME(extract_even)(src, ydst, width);
+ if(!(y&1)){
+ RENAME(extract_odd2)(src, udst, vdst, chromWidth);
+ udst+= chromStride;
+ vdst+= chromStride;
+ }
+
+ src += srcStride;
+ ydst+= lumStride;
+ }
+#if HAVE_MMX
+ __asm__(
+ EMMS" \n\t"
+ SFENCE" \n\t"
+ ::: "memory"
+ );
+#endif
+}
+
+static void RENAME(yuyvtoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src,
+ long width, long height,
+ long lumStride, long chromStride, long srcStride)
+{
+ long y;
+ const long chromWidth= -((-width)>>1);
+
+ for (y=0; y<height; y++){
+ RENAME(extract_even)(src, ydst, width);
+ RENAME(extract_odd2)(src, udst, vdst, chromWidth);
+
+ src += srcStride;
+ ydst+= lumStride;
+ udst+= chromStride;
+ vdst+= chromStride;
+ }
+#if HAVE_MMX
+ __asm__(
+ EMMS" \n\t"
+ SFENCE" \n\t"
+ ::: "memory"
+ );
+#endif
+}
+
+static void RENAME(uyvytoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src,
+ long width, long height,
+ long lumStride, long chromStride, long srcStride)
+{
+ long y;
+ const long chromWidth= -((-width)>>1);
+
+ for (y=0; y<height; y++){
+ RENAME(extract_even)(src+1, ydst, width);
+ if(!(y&1)){
+ RENAME(extract_even2)(src, udst, vdst, chromWidth);
+ udst+= chromStride;
+ vdst+= chromStride;
+ }
+
+ src += srcStride;
+ ydst+= lumStride;
+ }
+#if HAVE_MMX
+ __asm__(
+ EMMS" \n\t"
+ SFENCE" \n\t"
+ ::: "memory"
+ );
+#endif
+}
+
+static void RENAME(uyvytoyuv422)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src,
+ long width, long height,
+ long lumStride, long chromStride, long srcStride)
+{
+ long y;
+ const long chromWidth= -((-width)>>1);
+
+ for (y=0; y<height; y++){
+ RENAME(extract_even)(src+1, ydst, width);
+ RENAME(extract_even2)(src, udst, vdst, chromWidth);
+
+ src += srcStride;
+ ydst+= lumStride;
+ udst+= chromStride;
+ vdst+= chromStride;
+ }
+#if HAVE_MMX
+ __asm__(
+ EMMS" \n\t"
+ SFENCE" \n\t"
+ ::: "memory"
+ );
+#endif
+}
+
static inline void RENAME(rgb2rgb_init)(void){
rgb15to16 = RENAME(rgb15to16);
rgb15tobgr24 = RENAME(rgb15tobgr24);
@@ -2725,11 +2964,15 @@ static inline void RENAME(rgb2rgb_init)(void){
yuv422ptoyuy2 = RENAME(yuv422ptoyuy2);
yuv422ptouyvy = RENAME(yuv422ptouyvy);
yuy2toyv12 = RENAME(yuy2toyv12);
-// uyvytoyv12 = RENAME(uyvytoyv12);
// yvu9toyv12 = RENAME(yvu9toyv12);
planar2x = RENAME(planar2x);
rgb24toyv12 = RENAME(rgb24toyv12);
interleaveBytes = RENAME(interleaveBytes);
vu9_to_vu12 = RENAME(vu9_to_vu12);
yvu9_to_yuy2 = RENAME(yvu9_to_yuy2);
+
+ uyvytoyuv420 = RENAME(uyvytoyuv420);
+ uyvytoyuv422 = RENAME(uyvytoyuv422);
+ yuyvtoyuv420 = RENAME(yuyvtoyuv420);
+ yuyvtoyuv422 = RENAME(yuyvtoyuv422);
}
diff --git a/libswscale/swscale.c b/libswscale/swscale.c
index 5974ea5..e1101b4 100644
--- a/libswscale/swscale.c
+++ b/libswscale/swscale.c
@@ -1745,6 +1745,50 @@ static int YUV422PToUyvyWrapper(SwsContext *c, uint8_t* src[], int srcStride[],
return srcSliceH;
}
+static int YUYV2YUV420Wrapper(SwsContext *c, uint8_t* src[], int srcStride[], int srcSliceY,
+ int srcSliceH, uint8_t* dstParam[], int dstStride[]){
+ uint8_t *ydst=dstParam[0] + dstStride[0]*srcSliceY;
+ uint8_t *udst=dstParam[1] + dstStride[1]*srcSliceY/2;
+ uint8_t *vdst=dstParam[2] + dstStride[2]*srcSliceY/2;
+
+ yuyvtoyuv420(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0], dstStride[1], srcStride[0]);
+
+ return srcSliceH;
+}
+
+static int YUYV2YUV422Wrapper(SwsContext *c, uint8_t* src[], int srcStride[], int srcSliceY,
+ int srcSliceH, uint8_t* dstParam[], int dstStride[]){
+ uint8_t *ydst=dstParam[0] + dstStride[0]*srcSliceY;
+ uint8_t *udst=dstParam[1] + dstStride[1]*srcSliceY/2;
+ uint8_t *vdst=dstParam[2] + dstStride[2]*srcSliceY/2;
+
+ yuyvtoyuv422(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0], dstStride[1], srcStride[0]);
+
+ return srcSliceH;
+}
+
+static int UYVY2YUV420Wrapper(SwsContext *c, uint8_t* src[], int srcStride[], int srcSliceY,
+ int srcSliceH, uint8_t* dstParam[], int dstStride[]){
+ uint8_t *ydst=dstParam[0] + dstStride[0]*srcSliceY;
+ uint8_t *udst=dstParam[1] + dstStride[1]*srcSliceY/2;
+ uint8_t *vdst=dstParam[2] + dstStride[2]*srcSliceY/2;
+
+ uyvytoyuv420(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0], dstStride[1], srcStride[0]);
+
+ return srcSliceH;
+}
+
+static int UYVY2YUV422Wrapper(SwsContext *c, uint8_t* src[], int srcStride[], int srcSliceY,
+ int srcSliceH, uint8_t* dstParam[], int dstStride[]){
+ uint8_t *ydst=dstParam[0] + dstStride[0]*srcSliceY;
+ uint8_t *udst=dstParam[1] + dstStride[1]*srcSliceY/2;
+ uint8_t *vdst=dstParam[2] + dstStride[2]*srcSliceY/2;
+
+ uyvytoyuv422(ydst, udst, vdst, src[0], c->srcW, srcSliceH, dstStride[0], dstStride[1], srcStride[0]);
+
+ return srcSliceH;
+}
+
static int pal2rgbWrapper(SwsContext *c, uint8_t* src[], int srcStride[], int srcSliceY,
int srcSliceH, uint8_t* dst[], int dstStride[]){
const enum PixelFormat srcFormat= c->srcFormat;
@@ -2399,7 +2443,16 @@ SwsContext *sws_getContext(int srcW, int srcH, enum PixelFormat srcFormat, int d
else if (dstFormat == PIX_FMT_UYVY422)
c->swScale= PlanarToUyvyWrapper;
}
+
+ if(srcFormat == PIX_FMT_YUYV422 && dstFormat == PIX_FMT_YUV420P)
+ c->swScale= YUYV2YUV420Wrapper;
+ if(srcFormat == PIX_FMT_UYVY422 && dstFormat == PIX_FMT_YUV420P)
+ c->swScale= UYVY2YUV420Wrapper;
}
+ if(srcFormat == PIX_FMT_YUYV422 && dstFormat == PIX_FMT_YUV422P)
+ c->swScale= YUYV2YUV422Wrapper;
+ if(srcFormat == PIX_FMT_UYVY422 && dstFormat == PIX_FMT_YUV422P)
+ c->swScale= UYVY2YUV422Wrapper;
#ifdef COMPILE_ALTIVEC
if ((c->flags & SWS_CPU_CAPS_ALTIVEC) &&
--
Libav/FFmpeg packaging
More information about the pkg-multimedia-commits
mailing list