summaryrefslogtreecommitdiffstats
path: root/libswscale
diff options
context:
space:
mode:
Diffstat (limited to 'libswscale')
-rw-r--r--libswscale/rgb2rgb.c13
-rw-r--r--libswscale/rgb2rgb_template.c83
-rw-r--r--libswscale/swscale.c4
-rw-r--r--libswscale/swscale_template.c31
-rw-r--r--libswscale/yuv2rgb.c5
-rw-r--r--libswscale/yuv2rgb_altivec.c10
-rw-r--r--libswscale/yuv2rgb_template.c2
-rw-r--r--libswscale/yuv2rgb_vis.c6
8 files changed, 72 insertions, 82 deletions
diff --git a/libswscale/rgb2rgb.c b/libswscale/rgb2rgb.c
index 99be2a4f27..ad69265c37 100644
--- a/libswscale/rgb2rgb.c
+++ b/libswscale/rgb2rgb.c
@@ -94,6 +94,7 @@ DECLARE_ASM_CONST(8, uint64_t, mmx_one) = 0xFFFFFFFFFFFFFFFFULL;
DECLARE_ASM_CONST(8, uint64_t, mask32b) = 0x000000FF000000FFULL;
DECLARE_ASM_CONST(8, uint64_t, mask32g) = 0x0000FF000000FF00ULL;
DECLARE_ASM_CONST(8, uint64_t, mask32r) = 0x00FF000000FF0000ULL;
+DECLARE_ASM_CONST(8, uint64_t, mask32a) = 0xFF000000FF000000ULL;
DECLARE_ASM_CONST(8, uint64_t, mask32) = 0x00FFFFFF00FFFFFFULL;
DECLARE_ASM_CONST(8, uint64_t, mask3216br) = 0x00F800F800F800F8ULL;
DECLARE_ASM_CONST(8, uint64_t, mask3216g) = 0x0000FC000000FC00ULL;
@@ -281,7 +282,7 @@ void rgb24to32(const uint8_t *src, uint8_t *dst, long src_size)
{
#ifdef WORDS_BIGENDIAN
/* RGB24 (= R,G,B) -> BGR32 (= A,R,G,B) */
- dst[4*i + 0] = 0;
+ dst[4*i + 0] = 255;
dst[4*i + 1] = src[3*i + 0];
dst[4*i + 2] = src[3*i + 1];
dst[4*i + 3] = src[3*i + 2];
@@ -289,7 +290,7 @@ void rgb24to32(const uint8_t *src, uint8_t *dst, long src_size)
dst[4*i + 0] = src[3*i + 2];
dst[4*i + 1] = src[3*i + 1];
dst[4*i + 2] = src[3*i + 0];
- dst[4*i + 3] = 0;
+ dst[4*i + 3] = 255;
#endif
}
}
@@ -305,7 +306,7 @@ void rgb16tobgr32(const uint8_t *src, uint8_t *dst, long src_size)
register uint16_t bgr;
bgr = *s++;
#ifdef WORDS_BIGENDIAN
- *d++ = 0;
+ *d++ = 255;
*d++ = (bgr&0x1F)<<3;
*d++ = (bgr&0x7E0)>>3;
*d++ = (bgr&0xF800)>>8;
@@ -313,7 +314,7 @@ void rgb16tobgr32(const uint8_t *src, uint8_t *dst, long src_size)
*d++ = (bgr&0xF800)>>8;
*d++ = (bgr&0x7E0)>>3;
*d++ = (bgr&0x1F)<<3;
- *d++ = 0;
+ *d++ = 255;
#endif
}
}
@@ -369,7 +370,7 @@ void rgb15tobgr32(const uint8_t *src, uint8_t *dst, long src_size)
register uint16_t bgr;
bgr = *s++;
#ifdef WORDS_BIGENDIAN
- *d++ = 0;
+ *d++ = 255;
*d++ = (bgr&0x1F)<<3;
*d++ = (bgr&0x3E0)>>2;
*d++ = (bgr&0x7C00)>>7;
@@ -377,7 +378,7 @@ void rgb15tobgr32(const uint8_t *src, uint8_t *dst, long src_size)
*d++ = (bgr&0x7C00)>>7;
*d++ = (bgr&0x3E0)>>2;
*d++ = (bgr&0x1F)<<3;
- *d++ = 0;
+ *d++ = 255;
#endif
}
}
diff --git a/libswscale/rgb2rgb_template.c b/libswscale/rgb2rgb_template.c
index f8915e8df2..b03f6424f7 100644
--- a/libswscale/rgb2rgb_template.c
+++ b/libswscale/rgb2rgb_template.c
@@ -83,7 +83,7 @@ static inline void RENAME(rgb24tobgr32)(const uint8_t *src, uint8_t *dst, long s
#if HAVE_MMX
__asm__ volatile(PREFETCH" %0"::"m"(*s):"memory");
mm_end = end - 23;
- __asm__ volatile("movq %0, %%mm7"::"m"(mask32):"memory");
+ __asm__ volatile("movq %0, %%mm7"::"m"(mask32a):"memory");
while (s < mm_end)
{
__asm__ volatile(
@@ -96,10 +96,10 @@ static inline void RENAME(rgb24tobgr32)(const uint8_t *src, uint8_t *dst, long s
"punpckldq 15%1, %%mm2 \n\t"
"movd 18%1, %%mm3 \n\t"
"punpckldq 21%1, %%mm3 \n\t"
- "pand %%mm7, %%mm0 \n\t"
- "pand %%mm7, %%mm1 \n\t"
- "pand %%mm7, %%mm2 \n\t"
- "pand %%mm7, %%mm3 \n\t"
+ "por %%mm7, %%mm0 \n\t"
+ "por %%mm7, %%mm1 \n\t"
+ "por %%mm7, %%mm2 \n\t"
+ "por %%mm7, %%mm3 \n\t"
MOVNTQ" %%mm0, %0 \n\t"
MOVNTQ" %%mm1, 8%0 \n\t"
MOVNTQ" %%mm2, 16%0 \n\t"
@@ -117,7 +117,7 @@ static inline void RENAME(rgb24tobgr32)(const uint8_t *src, uint8_t *dst, long s
{
#ifdef WORDS_BIGENDIAN
/* RGB24 (= R,G,B) -> RGB32 (= A,B,G,R) */
- *dest++ = 0;
+ *dest++ = 255;
*dest++ = s[2];
*dest++ = s[1];
*dest++ = s[0];
@@ -126,7 +126,7 @@ static inline void RENAME(rgb24tobgr32)(const uint8_t *src, uint8_t *dst, long s
*dest++ = *s++;
*dest++ = *s++;
*dest++ = *s++;
- *dest++ = 0;
+ *dest++ = 255;
#endif
}
}
@@ -1202,6 +1202,25 @@ static inline void RENAME(rgb16tobgr24)(const uint8_t *src, uint8_t *dst, long s
}
}
+/*
+ * mm0 = 00 B3 00 B2 00 B1 00 B0
+ * mm1 = 00 G3 00 G2 00 G1 00 G0
+ * mm2 = 00 R3 00 R2 00 R1 00 R0
+ * mm6 = FF FF FF FF FF FF FF FF
+ * mm7 = 00 00 00 00 00 00 00 00
+ */
+#define PACK_RGB32 \
+ "packuswb %%mm7, %%mm0 \n\t" /* 00 00 00 00 B3 B2 B1 B0 */ \
+ "packuswb %%mm7, %%mm1 \n\t" /* 00 00 00 00 G3 G2 G1 G0 */ \
+ "packuswb %%mm7, %%mm2 \n\t" /* 00 00 00 00 R3 R2 R1 R0 */ \
+ "punpcklbw %%mm1, %%mm0 \n\t" /* G3 B3 G2 B2 G1 B1 G0 B0 */ \
+ "punpcklbw %%mm6, %%mm2 \n\t" /* FF R3 FF R2 FF R1 FF R0 */ \
+ "movq %%mm0, %%mm3 \n\t" \
+ "punpcklwd %%mm2, %%mm0 \n\t" /* FF R1 G1 B1 FF R0 G0 B0 */ \
+ "punpckhwd %%mm2, %%mm3 \n\t" /* FF R3 G3 B3 FF R2 G2 B2 */ \
+ MOVNTQ" %%mm0, %0 \n\t" \
+ MOVNTQ" %%mm3, 8%0 \n\t" \
+
static inline void RENAME(rgb15to32)(const uint8_t *src, uint8_t *dst, long src_size)
{
const uint16_t *end;
@@ -1214,6 +1233,7 @@ static inline void RENAME(rgb15to32)(const uint8_t *src, uint8_t *dst, long src_
#if HAVE_MMX
__asm__ volatile(PREFETCH" %0"::"m"(*s):"memory");
__asm__ volatile("pxor %%mm7,%%mm7 \n\t":::"memory");
+ __asm__ volatile("pcmpeqd %%mm6,%%mm6 \n\t":::"memory");
mm_end = end - 3;
while (s < mm_end)
{
@@ -1228,25 +1248,7 @@ static inline void RENAME(rgb15to32)(const uint8_t *src, uint8_t *dst, long src_
"psllq $3, %%mm0 \n\t"
"psrlq $2, %%mm1 \n\t"
"psrlq $7, %%mm2 \n\t"
- "movq %%mm0, %%mm3 \n\t"
- "movq %%mm1, %%mm4 \n\t"
- "movq %%mm2, %%mm5 \n\t"
- "punpcklwd %%mm7, %%mm0 \n\t"
- "punpcklwd %%mm7, %%mm1 \n\t"
- "punpcklwd %%mm7, %%mm2 \n\t"
- "punpckhwd %%mm7, %%mm3 \n\t"
- "punpckhwd %%mm7, %%mm4 \n\t"
- "punpckhwd %%mm7, %%mm5 \n\t"
- "psllq $8, %%mm1 \n\t"
- "psllq $16, %%mm2 \n\t"
- "por %%mm1, %%mm0 \n\t"
- "por %%mm2, %%mm0 \n\t"
- "psllq $8, %%mm4 \n\t"
- "psllq $16, %%mm5 \n\t"
- "por %%mm4, %%mm3 \n\t"
- "por %%mm5, %%mm3 \n\t"
- MOVNTQ" %%mm0, %0 \n\t"
- MOVNTQ" %%mm3, 8%0 \n\t"
+ PACK_RGB32
:"=m"(*d)
:"m"(*s),"m"(mask15b),"m"(mask15g),"m"(mask15r)
:"memory");
@@ -1265,7 +1267,7 @@ static inline void RENAME(rgb15to32)(const uint8_t *src, uint8_t *dst, long src_
register uint16_t bgr;
bgr = *s++;
#ifdef WORDS_BIGENDIAN
- *d++ = 0;
+ *d++ = 255;
*d++ = (bgr&0x7C00)>>7;
*d++ = (bgr&0x3E0)>>2;
*d++ = (bgr&0x1F)<<3;
@@ -1273,7 +1275,7 @@ static inline void RENAME(rgb15to32)(const uint8_t *src, uint8_t *dst, long src_
*d++ = (bgr&0x1F)<<3;
*d++ = (bgr&0x3E0)>>2;
*d++ = (bgr&0x7C00)>>7;
- *d++ = 0;
+ *d++ = 255;
#endif
#endif
@@ -1292,6 +1294,7 @@ static inline void RENAME(rgb16to32)(const uint8_t *src, uint8_t *dst, long src_
#if HAVE_MMX
__asm__ volatile(PREFETCH" %0"::"m"(*s):"memory");
__asm__ volatile("pxor %%mm7,%%mm7 \n\t":::"memory");
+ __asm__ volatile("pcmpeqd %%mm6,%%mm6 \n\t":::"memory");
mm_end = end - 3;
while (s < mm_end)
{
@@ -1306,25 +1309,7 @@ static inline void RENAME(rgb16to32)(const uint8_t *src, uint8_t *dst, long src_
"psllq $3, %%mm0 \n\t"
"psrlq $3, %%mm1 \n\t"
"psrlq $8, %%mm2 \n\t"
- "movq %%mm0, %%mm3 \n\t"
- "movq %%mm1, %%mm4 \n\t"
- "movq %%mm2, %%mm5 \n\t"
- "punpcklwd %%mm7, %%mm0 \n\t"
- "punpcklwd %%mm7, %%mm1 \n\t"
- "punpcklwd %%mm7, %%mm2 \n\t"
- "punpckhwd %%mm7, %%mm3 \n\t"
- "punpckhwd %%mm7, %%mm4 \n\t"
- "punpckhwd %%mm7, %%mm5 \n\t"
- "psllq $8, %%mm1 \n\t"
- "psllq $16, %%mm2 \n\t"
- "por %%mm1, %%mm0 \n\t"
- "por %%mm2, %%mm0 \n\t"
- "psllq $8, %%mm4 \n\t"
- "psllq $16, %%mm5 \n\t"
- "por %%mm4, %%mm3 \n\t"
- "por %%mm5, %%mm3 \n\t"
- MOVNTQ" %%mm0, %0 \n\t"
- MOVNTQ" %%mm3, 8%0 \n\t"
+ PACK_RGB32
:"=m"(*d)
:"m"(*s),"m"(mask16b),"m"(mask16g),"m"(mask16r)
:"memory");
@@ -1339,7 +1324,7 @@ static inline void RENAME(rgb16to32)(const uint8_t *src, uint8_t *dst, long src_
register uint16_t bgr;
bgr = *s++;
#ifdef WORDS_BIGENDIAN
- *d++ = 0;
+ *d++ = 255;
*d++ = (bgr&0xF800)>>8;
*d++ = (bgr&0x7E0)>>3;
*d++ = (bgr&0x1F)<<3;
@@ -1347,7 +1332,7 @@ static inline void RENAME(rgb16to32)(const uint8_t *src, uint8_t *dst, long src_
*d++ = (bgr&0x1F)<<3;
*d++ = (bgr&0x7E0)>>3;
*d++ = (bgr&0xF800)>>8;
- *d++ = 0;
+ *d++ = 255;
#endif
}
}
diff --git a/libswscale/swscale.c b/libswscale/swscale.c
index 1fe421e0d6..c0f4b97684 100644
--- a/libswscale/swscale.c
+++ b/libswscale/swscale.c
@@ -927,7 +927,7 @@ static inline void yuv2rgbXinC_full(SwsContext *c, int16_t *lumFilter, int16_t *
aidx--;
case PIX_FMT_RGBA:
YSCALE_YUV_2_RGBX_FULL_C(1<<21)
- dest[aidx]= 0;
+ dest[aidx]= 255;
dest[0]= R>>22;
dest[1]= G>>22;
dest[2]= B>>22;
@@ -941,7 +941,7 @@ static inline void yuv2rgbXinC_full(SwsContext *c, int16_t *lumFilter, int16_t *
aidx--;
case PIX_FMT_BGRA:
YSCALE_YUV_2_RGBX_FULL_C(1<<21)
- dest[aidx]= 0;
+ dest[aidx]= 255;
dest[0]= B>>22;
dest[1]= G>>22;
dest[2]= R>>22;
diff --git a/libswscale/swscale_template.c b/libswscale/swscale_template.c
index ba2b6f99c6..6f2e243052 100644
--- a/libswscale/swscale_template.c
+++ b/libswscale/swscale_template.c
@@ -1031,7 +1031,7 @@ static inline void RENAME(yuv2packedX)(SwsContext *c, int16_t *lumFilter, int16_
case PIX_FMT_RGB32:
YSCALEYUV2PACKEDX_ACCURATE
YSCALEYUV2RGBX
- "pxor %%mm7, %%mm7 \n\t"
+ "pcmpeqd %%mm7, %%mm7 \n\t"
WRITEBGR32(%4, %5, %%REGa, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
YSCALEYUV2PACKEDX_END
@@ -1097,7 +1097,7 @@ static inline void RENAME(yuv2packedX)(SwsContext *c, int16_t *lumFilter, int16_
case PIX_FMT_RGB32:
YSCALEYUV2PACKEDX
YSCALEYUV2RGBX
- "pxor %%mm7, %%mm7 \n\t"
+ "pcmpeqd %%mm7, %%mm7 \n\t"
WRITEBGR32(%4, %5, %%REGa, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
YSCALEYUV2PACKEDX_END
return;
@@ -1196,7 +1196,7 @@ static inline void RENAME(yuv2packed2)(SwsContext *c, uint16_t *buf0, uint16_t *
"mov %4, %%"REG_b" \n\t"
"push %%"REG_BP" \n\t"
YSCALEYUV2RGB(%%REGBP, %5)
- "pxor %%mm7, %%mm7 \n\t"
+ "pcmpeqd %%mm7, %%mm7 \n\t"
WRITEBGR32(%%REGb, 8280(%5), %%REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
"pop %%"REG_BP" \n\t"
"mov "ESP_OFFSET"(%5), %%"REG_b" \n\t"
@@ -1312,7 +1312,7 @@ static inline void RENAME(yuv2packed1)(SwsContext *c, uint16_t *buf0, uint16_t *
"mov %4, %%"REG_b" \n\t"
"push %%"REG_BP" \n\t"
YSCALEYUV2RGB1(%%REGBP, %5)
- "pxor %%mm7, %%mm7 \n\t"
+ "pcmpeqd %%mm7, %%mm7 \n\t"
WRITEBGR32(%%REGb, 8280(%5), %%REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
"pop %%"REG_BP" \n\t"
"mov "ESP_OFFSET"(%5), %%"REG_b" \n\t"
@@ -1405,7 +1405,7 @@ static inline void RENAME(yuv2packed1)(SwsContext *c, uint16_t *buf0, uint16_t *
"mov %4, %%"REG_b" \n\t"
"push %%"REG_BP" \n\t"
YSCALEYUV2RGB1b(%%REGBP, %5)
- "pxor %%mm7, %%mm7 \n\t"
+ "pcmpeqd %%mm7, %%mm7 \n\t"
WRITEBGR32(%%REGb, 8280(%5), %%REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
"pop %%"REG_BP" \n\t"
"mov "ESP_OFFSET"(%5), %%"REG_b" \n\t"
@@ -1642,7 +1642,7 @@ BGR2Y(uint16_t, bgr15ToY, 0, 0, 0, 0x001F, 0x03E0, 0x7C00, RY<<10, GY<<5, BY
BGR2Y(uint16_t, rgb16ToY, 0, 0, 0, 0xF800, 0x07E0, 0x001F, RY , GY<<5, BY<<11, RGB2YUV_SHIFT+8)
BGR2Y(uint16_t, rgb15ToY, 0, 0, 0, 0x7C00, 0x03E0, 0x001F, RY , GY<<5, BY<<10, RGB2YUV_SHIFT+7)
-#define BGR2UV(type, name, shr, shg, shb, maskr, maskg, maskb, RU, GU, BU, RV, GV, BV, S)\
+#define BGR2UV(type, name, shr, shg, shb, maska, maskr, maskg, maskb, RU, GU, BU, RV, GV, BV, S)\
static inline void RENAME(name)(uint8_t *dstU, uint8_t *dstV, uint8_t *src, uint8_t *dummy, long width, uint32_t *unused)\
{\
int i;\
@@ -1663,9 +1663,10 @@ static inline void RENAME(name ## _half)(uint8_t *dstU, uint8_t *dstV, uint8_t *
{\
int pix0= ((type*)src)[2*i+0];\
int pix1= ((type*)src)[2*i+1];\
- int g= (pix0&maskg)+(pix1&maskg);\
+ int g= (pix0&(maskg|maska))+(pix1&(maskg|maska));\
int b= ((pix0+pix1-g)&(maskb|(2*maskb)))>>shb;\
int r= ((pix0+pix1-g)&(maskr|(2*maskr)))>>shr;\
+ g&= maskg|(2*maskg);\
\
g>>=shg;\
\
@@ -1674,12 +1675,12 @@ static inline void RENAME(name ## _half)(uint8_t *dstU, uint8_t *dstV, uint8_t *
}\
}
-BGR2UV(uint32_t, bgr32ToUV,16, 0, 0, 0xFF0000, 0xFF00, 0x00FF, RU<< 8, GU , BU<< 8, RV<< 8, GV , BV<< 8, RGB2YUV_SHIFT+8)
-BGR2UV(uint32_t, rgb32ToUV, 0, 0,16, 0x00FF, 0xFF00, 0xFF0000, RU<< 8, GU , BU<< 8, RV<< 8, GV , BV<< 8, RGB2YUV_SHIFT+8)
-BGR2UV(uint16_t, bgr16ToUV, 0, 0, 0, 0x001F, 0x07E0, 0xF800, RU<<11, GU<<5, BU , RV<<11, GV<<5, BV , RGB2YUV_SHIFT+8)
-BGR2UV(uint16_t, bgr15ToUV, 0, 0, 0, 0x001F, 0x03E0, 0x7C00, RU<<10, GU<<5, BU , RV<<10, GV<<5, BV , RGB2YUV_SHIFT+7)
-BGR2UV(uint16_t, rgb16ToUV, 0, 0, 0, 0xF800, 0x07E0, 0x001F, RU , GU<<5, BU<<11, RV , GV<<5, BV<<11, RGB2YUV_SHIFT+8)
-BGR2UV(uint16_t, rgb15ToUV, 0, 0, 0, 0x7C00, 0x03E0, 0x001F, RU , GU<<5, BU<<10, RV , GV<<5, BV<<10, RGB2YUV_SHIFT+7)
+BGR2UV(uint32_t, bgr32ToUV,16, 0, 0, 0xFF000000, 0xFF0000, 0xFF00, 0x00FF, RU<< 8, GU , BU<< 8, RV<< 8, GV , BV<< 8, RGB2YUV_SHIFT+8)
+BGR2UV(uint32_t, rgb32ToUV, 0, 0,16, 0xFF000000, 0x00FF, 0xFF00, 0xFF0000, RU<< 8, GU , BU<< 8, RV<< 8, GV , BV<< 8, RGB2YUV_SHIFT+8)
+BGR2UV(uint16_t, bgr16ToUV, 0, 0, 0, 0, 0x001F, 0x07E0, 0xF800, RU<<11, GU<<5, BU , RV<<11, GV<<5, BV , RGB2YUV_SHIFT+8)
+BGR2UV(uint16_t, bgr15ToUV, 0, 0, 0, 0, 0x001F, 0x03E0, 0x7C00, RU<<10, GU<<5, BU , RV<<10, GV<<5, BV , RGB2YUV_SHIFT+7)
+BGR2UV(uint16_t, rgb16ToUV, 0, 0, 0, 0, 0xF800, 0x07E0, 0x001F, RU , GU<<5, BU<<11, RV , GV<<5, BV<<11, RGB2YUV_SHIFT+8)
+BGR2UV(uint16_t, rgb15ToUV, 0, 0, 0, 0, 0x7C00, 0x03E0, 0x001F, RU , GU<<5, BU<<10, RV , GV<<5, BV<<10, RGB2YUV_SHIFT+7)
#if HAVE_MMX
static inline void RENAME(bgr24ToY_mmx)(uint8_t *dst, uint8_t *src, long width, int srcFormat)
@@ -2218,7 +2219,7 @@ static inline void RENAME(hyscale)(SwsContext *c, uint16_t *dst, long dstWidth,
}
else // fast bilinear upscale / crap downscale
{
-#if ARCH_X86
+#if ARCH_X86 && CONFIG_GPL
#if HAVE_MMX2
int i;
#if defined(PIC)
@@ -2491,7 +2492,7 @@ inline static void RENAME(hcscale)(SwsContext *c, uint16_t *dst, long dstWidth,
}
else // fast bilinear upscale / crap downscale
{
-#if ARCH_X86
+#if ARCH_X86 && CONFIG_GPL
#if HAVE_MMX2
int i;
#if defined(PIC)
diff --git a/libswscale/yuv2rgb.c b/libswscale/yuv2rgb.c
index 53a0d4e7b8..65af412c2c 100644
--- a/libswscale/yuv2rgb.c
+++ b/libswscale/yuv2rgb.c
@@ -533,7 +533,7 @@ av_cold int sws_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4], int
uint8_t *y_table;
uint16_t *y_table16;
uint32_t *y_table32;
- int i, base, rbase, gbase, bbase;
+ int i, base, rbase, gbase, bbase, abase;
const int yoffs = fullRange ? 384 : 326;
int64_t crv = inv_table[0];
@@ -659,12 +659,13 @@ av_cold int sws_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4], int
rbase = base + (isRgb ? 16 : 0);
gbase = base + 8;
bbase = base + (isRgb ? 0 : 16);
+ abase = (base + 24) & 31;
c->yuvTable = av_malloc(1024*3*4);
y_table32 = c->yuvTable;
yb = -(384<<16) - oy;
for (i = 0; i < 1024; i++) {
uint8_t yval = av_clip_uint8((yb + 0x8000) >> 16);
- y_table32[i ] = yval << rbase;
+ y_table32[i ] = (yval << rbase) + (255 << abase);
y_table32[i+1024] = yval << gbase;
y_table32[i+2048] = yval << bbase;
yb += cy;
diff --git a/libswscale/yuv2rgb_altivec.c b/libswscale/yuv2rgb_altivec.c
index 784c197d70..b3a87a0360 100644
--- a/libswscale/yuv2rgb_altivec.c
+++ b/libswscale/yuv2rgb_altivec.c
@@ -237,7 +237,7 @@ do { \
((vector unsigned short)vec_max (x,((vector signed short) {0})), \
(vector unsigned short)vec_max (y,((vector signed short) {0})))
-//#define out_pixels(a,b,c,ptr) vec_mstrgb32(__typeof__(a),((__typeof__ (a)){0}),a,a,a,ptr)
+//#define out_pixels(a,b,c,ptr) vec_mstrgb32(__typeof__(a),((__typeof__ (a)){255}),a,a,a,ptr)
static inline void cvtyuvtoRGB (SwsContext *c,
@@ -438,10 +438,10 @@ static int altivec_##name (SwsContext *c, \
}
-#define out_abgr(a,b,c,ptr) vec_mstrgb32(__typeof__(a),((__typeof__ (a)){0}),c,b,a,ptr)
-#define out_bgra(a,b,c,ptr) vec_mstrgb32(__typeof__(a),c,b,a,((__typeof__ (a)){0}),ptr)
-#define out_rgba(a,b,c,ptr) vec_mstrgb32(__typeof__(a),a,b,c,((__typeof__ (a)){0}),ptr)
-#define out_argb(a,b,c,ptr) vec_mstrgb32(__typeof__(a),((__typeof__ (a)){0}),a,b,c,ptr)
+#define out_abgr(a,b,c,ptr) vec_mstrgb32(__typeof__(a),((__typeof__ (a)){255}),c,b,a,ptr)
+#define out_bgra(a,b,c,ptr) vec_mstrgb32(__typeof__(a),c,b,a,((__typeof__ (a)){255}),ptr)
+#define out_rgba(a,b,c,ptr) vec_mstrgb32(__typeof__(a),a,b,c,((__typeof__ (a)){255}),ptr)
+#define out_argb(a,b,c,ptr) vec_mstrgb32(__typeof__(a),((__typeof__ (a)){255}),a,b,c,ptr)
#define out_rgb24(a,b,c,ptr) vec_mstrgb24(a,b,c,ptr)
#define out_bgr24(a,b,c,ptr) vec_mstbgr24(a,b,c,ptr)
diff --git a/libswscale/yuv2rgb_template.c b/libswscale/yuv2rgb_template.c
index 0b5e6f9ca1..f55568b0ab 100644
--- a/libswscale/yuv2rgb_template.c
+++ b/libswscale/yuv2rgb_template.c
@@ -446,7 +446,7 @@ static inline int RENAME(yuv420_rgb32)(SwsContext *c, uint8_t* src[], int srcStr
YUV2RGB_INIT
YUV2RGB
- "pxor %%mm3, %%mm3;" /* zero mm3 */
+ "pcmpeqd %%mm3, %%mm3;" /* fill mm3 */
RGB_PLANAR2PACKED32
YUV2RGB_ENDLOOP(4)
diff --git a/libswscale/yuv2rgb_vis.c b/libswscale/yuv2rgb_vis.c
index 5c9698510d..2e2737aa9f 100644
--- a/libswscale/yuv2rgb_vis.c
+++ b/libswscale/yuv2rgb_vis.c
@@ -80,6 +80,7 @@
+// FIXME: must be changed to set alpha to 255 instead of 0
static int vis_420P_ARGB32(SwsContext *c, uint8_t* src[], int srcStride[], int srcSliceY,
int srcSliceH, uint8_t* dst[], int dstStride[]){
int y, out1, out2, out3, out4, out5, out6;
@@ -131,6 +132,7 @@ static int vis_420P_ARGB32(SwsContext *c, uint8_t* src[], int srcStride[], int s
return srcSliceH;
}
+// FIXME: must be changed to set alpha to 255 instead of 0
static int vis_422P_ARGB32(SwsContext *c, uint8_t* src[], int srcStride[], int srcSliceY,
int srcSliceH, uint8_t* dst[], int dstStride[]){
int y, out1, out2, out3, out4, out5, out6;
@@ -196,11 +198,11 @@ SwsFunc sws_yuv2rgb_init_vis(SwsContext *c) {
c->sparc_coeffs[4]=(((int16_t)c->vOffset*(int16_t)c->vrCoeff>>11) & 0xffff) * 0x0001000100010001ULL;
if (c->dstFormat == PIX_FMT_RGB32 && c->srcFormat == PIX_FMT_YUV422P && (c->dstW & 7)==0) {
- av_log(c, AV_LOG_INFO, "SPARC VIS accelerated YUV422P -> RGB32\n");
+ av_log(c, AV_LOG_INFO, "SPARC VIS accelerated YUV422P -> RGB32 (WARNING: alpha value is wrong)\n");
return vis_422P_ARGB32;
}
else if (c->dstFormat == PIX_FMT_RGB32 && c->srcFormat == PIX_FMT_YUV420P && (c->dstW & 7)==0) {
- av_log(c, AV_LOG_INFO, "SPARC VIS accelerated YUV420P -> RGB32\n");
+ av_log(c, AV_LOG_INFO, "SPARC VIS accelerated YUV420P -> RGB32 (WARNING: alpha value is wrong)\n");
return vis_420P_ARGB32;
}
return NULL;
generated by cgit v1.2.3 (git 2.25.1) at 2024-05-25 03:43:41 +0000