mirror of
https://github.com/xenia-project/FFmpeg.git
synced 2025-02-28 16:16:41 +00:00
Merge commit '945673f792eb4141860ed66664156df685f7b7f0'
* commit '945673f792eb4141860ed66664156df685f7b7f0': dsputil_template: K&R formatting cosmetics Conflicts: libavcodec/dsputil_template.c Merged-by: Michael Niedermayer <michaelni@gmx.at>
This commit is contained in:
commit
fdf4eb7911
@ -31,18 +31,19 @@
|
|||||||
|
|
||||||
#if BIT_DEPTH == 8
|
#if BIT_DEPTH == 8
|
||||||
/* draw the edges of width 'w' of an image of size width, height */
|
/* draw the edges of width 'w' of an image of size width, height */
|
||||||
//FIXME check that this is ok for mpeg4 interlaced
|
// FIXME: Check that this is OK for MPEG-4 interlaced.
|
||||||
static void FUNCC(draw_edges)(uint8_t *_buf, int _wrap, int width, int height, int w, int h, int sides)
|
static void FUNCC(draw_edges)(uint8_t *_buf, int _wrap, int width, int height,
|
||||||
|
int w, int h, int sides)
|
||||||
{
|
{
|
||||||
pixel *buf = (pixel*)_buf;
|
pixel *buf = (pixel *) _buf;
|
||||||
int wrap = _wrap / sizeof(pixel);
|
int wrap = _wrap / sizeof(pixel);
|
||||||
pixel *ptr = buf, *last_line;
|
pixel *ptr = buf, *last_line;
|
||||||
int i;
|
int i;
|
||||||
|
|
||||||
/* left and right */
|
/* left and right */
|
||||||
for(i=0;i<height;i++) {
|
for (i = 0; i < height; i++) {
|
||||||
memset(ptr - w, ptr[0], w);
|
memset(ptr - w, ptr[0], w);
|
||||||
memset(ptr + width, ptr[width-1], w);
|
memset(ptr + width, ptr[width - 1], w);
|
||||||
ptr += wrap;
|
ptr += wrap;
|
||||||
}
|
}
|
||||||
|
|
||||||
@ -50,23 +51,25 @@ static void FUNCC(draw_edges)(uint8_t *_buf, int _wrap, int width, int height, i
|
|||||||
buf -= w;
|
buf -= w;
|
||||||
last_line = buf + (height - 1) * wrap;
|
last_line = buf + (height - 1) * wrap;
|
||||||
if (sides & EDGE_TOP)
|
if (sides & EDGE_TOP)
|
||||||
for(i = 0; i < h; i++)
|
for (i = 0; i < h; i++)
|
||||||
memcpy(buf - (i + 1) * wrap, buf, (width + w + w) * sizeof(pixel)); // top
|
// top
|
||||||
|
memcpy(buf - (i + 1) * wrap, buf, (width + w + w) * sizeof(pixel));
|
||||||
if (sides & EDGE_BOTTOM)
|
if (sides & EDGE_BOTTOM)
|
||||||
for (i = 0; i < h; i++)
|
for (i = 0; i < h; i++)
|
||||||
memcpy(last_line + (i + 1) * wrap, last_line, (width + w + w) * sizeof(pixel)); // bottom
|
// bottom
|
||||||
|
memcpy(last_line + (i + 1) * wrap, last_line,
|
||||||
|
(width + w + w) * sizeof(pixel));
|
||||||
}
|
}
|
||||||
#endif
|
#endif
|
||||||
|
|
||||||
static void FUNCC(get_pixels)(int16_t *av_restrict block,
|
static void FUNCC(get_pixels)(int16_t *av_restrict block, const uint8_t *_pixels,
|
||||||
const uint8_t *_pixels,
|
|
||||||
int line_size)
|
int line_size)
|
||||||
{
|
{
|
||||||
const pixel *pixels = (const pixel *) _pixels;
|
const pixel *pixels = (const pixel *) _pixels;
|
||||||
int i;
|
int i;
|
||||||
|
|
||||||
/* read the pixels */
|
/* read the pixels */
|
||||||
for(i=0;i<8;i++) {
|
for (i = 0; i < 8; i++) {
|
||||||
block[0] = pixels[0];
|
block[0] = pixels[0];
|
||||||
block[1] = pixels[1];
|
block[1] = pixels[1];
|
||||||
block[2] = pixels[2];
|
block[2] = pixels[2];
|
||||||
@ -75,20 +78,20 @@ static void FUNCC(get_pixels)(int16_t *av_restrict block,
|
|||||||
block[5] = pixels[5];
|
block[5] = pixels[5];
|
||||||
block[6] = pixels[6];
|
block[6] = pixels[6];
|
||||||
block[7] = pixels[7];
|
block[7] = pixels[7];
|
||||||
pixels += line_size / sizeof(pixel);
|
pixels += line_size / sizeof(pixel);
|
||||||
block += 8;
|
block += 8;
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
#if BIT_DEPTH == 8
|
#if BIT_DEPTH == 8
|
||||||
static void FUNCC(clear_block)(int16_t *block)
|
static void FUNCC(clear_block)(int16_t *block)
|
||||||
{
|
{
|
||||||
memset(block, 0, sizeof(int16_t)*64);
|
memset(block, 0, sizeof(int16_t) * 64);
|
||||||
}
|
}
|
||||||
|
|
||||||
static void FUNCC(clear_blocks)(int16_t *blocks)
|
static void FUNCC(clear_blocks)(int16_t *blocks)
|
||||||
{
|
{
|
||||||
memset(blocks, 0, sizeof(int16_t)*6*64);
|
memset(blocks, 0, sizeof(int16_t) * 6 * 64);
|
||||||
}
|
}
|
||||||
#endif
|
#endif
|
||||||
|
|
||||||
@ -96,154 +99,244 @@ static void FUNCC(clear_blocks)(int16_t *blocks)
|
|||||||
#include "hpel_template.c"
|
#include "hpel_template.c"
|
||||||
#endif
|
#endif
|
||||||
|
|
||||||
#define PIXOP2(OPNAME, OP) \
|
#define PIXOP2(OPNAME, OP) \
|
||||||
static inline void FUNC(OPNAME ## _no_rnd_pixels8_l2)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
|
static inline void FUNC(OPNAME ## _no_rnd_pixels8_l2)(uint8_t *dst, \
|
||||||
int src_stride1, int src_stride2, int h){\
|
const uint8_t *src1, \
|
||||||
int i;\
|
const uint8_t *src2, \
|
||||||
for(i=0; i<h; i++){\
|
int dst_stride, \
|
||||||
pixel4 a,b;\
|
int src_stride1, \
|
||||||
a= AV_RN4P(&src1[i*src_stride1 ]);\
|
int src_stride2, \
|
||||||
b= AV_RN4P(&src2[i*src_stride2 ]);\
|
int h) \
|
||||||
OP(*((pixel4*)&dst[i*dst_stride ]), no_rnd_avg_pixel4(a, b));\
|
{ \
|
||||||
a= AV_RN4P(&src1[i*src_stride1+4*sizeof(pixel)]);\
|
int i; \
|
||||||
b= AV_RN4P(&src2[i*src_stride2+4*sizeof(pixel)]);\
|
\
|
||||||
OP(*((pixel4*)&dst[i*dst_stride+4*sizeof(pixel)]), no_rnd_avg_pixel4(a, b));\
|
for (i = 0; i < h; i++) { \
|
||||||
}\
|
pixel4 a, b; \
|
||||||
}\
|
a = AV_RN4P(&src1[i * src_stride1]); \
|
||||||
\
|
b = AV_RN4P(&src2[i * src_stride2]); \
|
||||||
static inline void FUNC(OPNAME ## _no_rnd_pixels16_l2)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
|
OP(*((pixel4 *) &dst[i * dst_stride]), \
|
||||||
int src_stride1, int src_stride2, int h){\
|
no_rnd_avg_pixel4(a, b)); \
|
||||||
FUNC(OPNAME ## _no_rnd_pixels8_l2)(dst , src1 , src2 , dst_stride, src_stride1, src_stride2, h);\
|
a = AV_RN4P(&src1[i * src_stride1 + 4 * sizeof(pixel)]); \
|
||||||
FUNC(OPNAME ## _no_rnd_pixels8_l2)(dst+8*sizeof(pixel), src1+8*sizeof(pixel), src2+8*sizeof(pixel), dst_stride, src_stride1, src_stride2, h);\
|
b = AV_RN4P(&src2[i * src_stride2 + 4 * sizeof(pixel)]); \
|
||||||
}\
|
OP(*((pixel4 *) &dst[i * dst_stride + 4 * sizeof(pixel)]), \
|
||||||
\
|
no_rnd_avg_pixel4(a, b)); \
|
||||||
static inline void FUNC(OPNAME ## _pixels8_l4)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4,\
|
} \
|
||||||
int dst_stride, int src_stride1, int src_stride2,int src_stride3,int src_stride4, int h){\
|
} \
|
||||||
/* FIXME HIGH BIT DEPTH */\
|
\
|
||||||
int i;\
|
static inline void FUNC(OPNAME ## _no_rnd_pixels16_l2)(uint8_t *dst, \
|
||||||
for(i=0; i<h; i++){\
|
const uint8_t *src1, \
|
||||||
uint32_t a, b, c, d, l0, l1, h0, h1;\
|
const uint8_t *src2, \
|
||||||
a= AV_RN32(&src1[i*src_stride1]);\
|
int dst_stride, \
|
||||||
b= AV_RN32(&src2[i*src_stride2]);\
|
int src_stride1, \
|
||||||
c= AV_RN32(&src3[i*src_stride3]);\
|
int src_stride2, \
|
||||||
d= AV_RN32(&src4[i*src_stride4]);\
|
int h) \
|
||||||
l0= (a&0x03030303UL)\
|
{ \
|
||||||
+ (b&0x03030303UL)\
|
FUNC(OPNAME ## _no_rnd_pixels8_l2)(dst, src1, src2, dst_stride, \
|
||||||
+ 0x02020202UL;\
|
src_stride1, src_stride2, h); \
|
||||||
h0= ((a&0xFCFCFCFCUL)>>2)\
|
FUNC(OPNAME ## _no_rnd_pixels8_l2)(dst + 8 * sizeof(pixel), \
|
||||||
+ ((b&0xFCFCFCFCUL)>>2);\
|
src1 + 8 * sizeof(pixel), \
|
||||||
l1= (c&0x03030303UL)\
|
src2 + 8 * sizeof(pixel), \
|
||||||
+ (d&0x03030303UL);\
|
dst_stride, src_stride1, \
|
||||||
h1= ((c&0xFCFCFCFCUL)>>2)\
|
src_stride2, h); \
|
||||||
+ ((d&0xFCFCFCFCUL)>>2);\
|
} \
|
||||||
OP(*((uint32_t*)&dst[i*dst_stride]), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
|
\
|
||||||
a= AV_RN32(&src1[i*src_stride1+4]);\
|
static inline void FUNC(OPNAME ## _pixels8_l4)(uint8_t *dst, \
|
||||||
b= AV_RN32(&src2[i*src_stride2+4]);\
|
const uint8_t *src1, \
|
||||||
c= AV_RN32(&src3[i*src_stride3+4]);\
|
const uint8_t *src2, \
|
||||||
d= AV_RN32(&src4[i*src_stride4+4]);\
|
const uint8_t *src3, \
|
||||||
l0= (a&0x03030303UL)\
|
const uint8_t *src4, \
|
||||||
+ (b&0x03030303UL)\
|
int dst_stride, \
|
||||||
+ 0x02020202UL;\
|
int src_stride1, \
|
||||||
h0= ((a&0xFCFCFCFCUL)>>2)\
|
int src_stride2, \
|
||||||
+ ((b&0xFCFCFCFCUL)>>2);\
|
int src_stride3, \
|
||||||
l1= (c&0x03030303UL)\
|
int src_stride4, \
|
||||||
+ (d&0x03030303UL);\
|
int h) \
|
||||||
h1= ((c&0xFCFCFCFCUL)>>2)\
|
{ \
|
||||||
+ ((d&0xFCFCFCFCUL)>>2);\
|
/* FIXME HIGH BIT DEPTH */ \
|
||||||
OP(*((uint32_t*)&dst[i*dst_stride+4]), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
|
int i; \
|
||||||
}\
|
\
|
||||||
}\
|
for (i = 0; i < h; i++) { \
|
||||||
\
|
uint32_t a, b, c, d, l0, l1, h0, h1; \
|
||||||
static inline void FUNC(OPNAME ## _no_rnd_pixels8_l4)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4,\
|
a = AV_RN32(&src1[i * src_stride1]); \
|
||||||
int dst_stride, int src_stride1, int src_stride2,int src_stride3,int src_stride4, int h){\
|
b = AV_RN32(&src2[i * src_stride2]); \
|
||||||
/* FIXME HIGH BIT DEPTH*/\
|
c = AV_RN32(&src3[i * src_stride3]); \
|
||||||
int i;\
|
d = AV_RN32(&src4[i * src_stride4]); \
|
||||||
for(i=0; i<h; i++){\
|
l0 = (a & 0x03030303UL) + \
|
||||||
uint32_t a, b, c, d, l0, l1, h0, h1;\
|
(b & 0x03030303UL) + \
|
||||||
a= AV_RN32(&src1[i*src_stride1]);\
|
0x02020202UL; \
|
||||||
b= AV_RN32(&src2[i*src_stride2]);\
|
h0 = ((a & 0xFCFCFCFCUL) >> 2) + \
|
||||||
c= AV_RN32(&src3[i*src_stride3]);\
|
((b & 0xFCFCFCFCUL) >> 2); \
|
||||||
d= AV_RN32(&src4[i*src_stride4]);\
|
l1 = (c & 0x03030303UL) + \
|
||||||
l0= (a&0x03030303UL)\
|
(d & 0x03030303UL); \
|
||||||
+ (b&0x03030303UL)\
|
h1 = ((c & 0xFCFCFCFCUL) >> 2) + \
|
||||||
+ 0x01010101UL;\
|
((d & 0xFCFCFCFCUL) >> 2); \
|
||||||
h0= ((a&0xFCFCFCFCUL)>>2)\
|
OP(*((uint32_t *) &dst[i * dst_stride]), \
|
||||||
+ ((b&0xFCFCFCFCUL)>>2);\
|
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL)); \
|
||||||
l1= (c&0x03030303UL)\
|
a = AV_RN32(&src1[i * src_stride1 + 4]); \
|
||||||
+ (d&0x03030303UL);\
|
b = AV_RN32(&src2[i * src_stride2 + 4]); \
|
||||||
h1= ((c&0xFCFCFCFCUL)>>2)\
|
c = AV_RN32(&src3[i * src_stride3 + 4]); \
|
||||||
+ ((d&0xFCFCFCFCUL)>>2);\
|
d = AV_RN32(&src4[i * src_stride4 + 4]); \
|
||||||
OP(*((uint32_t*)&dst[i*dst_stride]), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
|
l0 = (a & 0x03030303UL) + \
|
||||||
a= AV_RN32(&src1[i*src_stride1+4]);\
|
(b & 0x03030303UL) + \
|
||||||
b= AV_RN32(&src2[i*src_stride2+4]);\
|
0x02020202UL; \
|
||||||
c= AV_RN32(&src3[i*src_stride3+4]);\
|
h0 = ((a & 0xFCFCFCFCUL) >> 2) + \
|
||||||
d= AV_RN32(&src4[i*src_stride4+4]);\
|
((b & 0xFCFCFCFCUL) >> 2); \
|
||||||
l0= (a&0x03030303UL)\
|
l1 = (c & 0x03030303UL) + \
|
||||||
+ (b&0x03030303UL)\
|
(d & 0x03030303UL); \
|
||||||
+ 0x01010101UL;\
|
h1 = ((c & 0xFCFCFCFCUL) >> 2) + \
|
||||||
h0= ((a&0xFCFCFCFCUL)>>2)\
|
((d & 0xFCFCFCFCUL) >> 2); \
|
||||||
+ ((b&0xFCFCFCFCUL)>>2);\
|
OP(*((uint32_t *) &dst[i * dst_stride + 4]), \
|
||||||
l1= (c&0x03030303UL)\
|
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL)); \
|
||||||
+ (d&0x03030303UL);\
|
} \
|
||||||
h1= ((c&0xFCFCFCFCUL)>>2)\
|
} \
|
||||||
+ ((d&0xFCFCFCFCUL)>>2);\
|
\
|
||||||
OP(*((uint32_t*)&dst[i*dst_stride+4]), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
|
static inline void FUNC(OPNAME ## _no_rnd_pixels8_l4)(uint8_t *dst, \
|
||||||
}\
|
const uint8_t *src1, \
|
||||||
}\
|
const uint8_t *src2, \
|
||||||
static inline void FUNC(OPNAME ## _pixels16_l4)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4,\
|
const uint8_t *src3, \
|
||||||
int dst_stride, int src_stride1, int src_stride2,int src_stride3,int src_stride4, int h){\
|
const uint8_t *src4, \
|
||||||
FUNC(OPNAME ## _pixels8_l4)(dst , src1 , src2 , src3 , src4 , dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
|
int dst_stride, \
|
||||||
FUNC(OPNAME ## _pixels8_l4)(dst+8*sizeof(pixel), src1+8*sizeof(pixel), src2+8*sizeof(pixel), src3+8*sizeof(pixel), src4+8*sizeof(pixel), dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
|
int src_stride1, \
|
||||||
}\
|
int src_stride2, \
|
||||||
static inline void FUNC(OPNAME ## _no_rnd_pixels16_l4)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4,\
|
int src_stride3, \
|
||||||
int dst_stride, int src_stride1, int src_stride2,int src_stride3,int src_stride4, int h){\
|
int src_stride4, \
|
||||||
FUNC(OPNAME ## _no_rnd_pixels8_l4)(dst , src1 , src2 , src3 , src4 , dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
|
int h) \
|
||||||
FUNC(OPNAME ## _no_rnd_pixels8_l4)(dst+8*sizeof(pixel), src1+8*sizeof(pixel), src2+8*sizeof(pixel), src3+8*sizeof(pixel), src4+8*sizeof(pixel), dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
|
{ \
|
||||||
}\
|
/* FIXME HIGH BIT DEPTH */ \
|
||||||
\
|
int i; \
|
||||||
static inline void FUNCC(OPNAME ## _pixels8_xy2)(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)\
|
\
|
||||||
{\
|
for (i = 0; i < h; i++) { \
|
||||||
/* FIXME HIGH BIT DEPTH */\
|
uint32_t a, b, c, d, l0, l1, h0, h1; \
|
||||||
int j;\
|
a = AV_RN32(&src1[i * src_stride1]); \
|
||||||
for(j=0; j<2; j++){\
|
b = AV_RN32(&src2[i * src_stride2]); \
|
||||||
int i;\
|
c = AV_RN32(&src3[i * src_stride3]); \
|
||||||
const uint32_t a= AV_RN32(pixels );\
|
d = AV_RN32(&src4[i * src_stride4]); \
|
||||||
const uint32_t b= AV_RN32(pixels+1);\
|
l0 = (a & 0x03030303UL) + \
|
||||||
uint32_t l0= (a&0x03030303UL)\
|
(b & 0x03030303UL) + \
|
||||||
+ (b&0x03030303UL)\
|
0x01010101UL; \
|
||||||
+ 0x02020202UL;\
|
h0 = ((a & 0xFCFCFCFCUL) >> 2) + \
|
||||||
uint32_t h0= ((a&0xFCFCFCFCUL)>>2)\
|
((b & 0xFCFCFCFCUL) >> 2); \
|
||||||
+ ((b&0xFCFCFCFCUL)>>2);\
|
l1 = (c & 0x03030303UL) + \
|
||||||
uint32_t l1,h1;\
|
(d & 0x03030303UL); \
|
||||||
\
|
h1 = ((c & 0xFCFCFCFCUL) >> 2) + \
|
||||||
pixels+=line_size;\
|
((d & 0xFCFCFCFCUL) >> 2); \
|
||||||
for(i=0; i<h; i+=2){\
|
OP(*((uint32_t *) &dst[i * dst_stride]), \
|
||||||
uint32_t a= AV_RN32(pixels );\
|
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL)); \
|
||||||
uint32_t b= AV_RN32(pixels+1);\
|
a = AV_RN32(&src1[i * src_stride1 + 4]); \
|
||||||
l1= (a&0x03030303UL)\
|
b = AV_RN32(&src2[i * src_stride2 + 4]); \
|
||||||
+ (b&0x03030303UL);\
|
c = AV_RN32(&src3[i * src_stride3 + 4]); \
|
||||||
h1= ((a&0xFCFCFCFCUL)>>2)\
|
d = AV_RN32(&src4[i * src_stride4 + 4]); \
|
||||||
+ ((b&0xFCFCFCFCUL)>>2);\
|
l0 = (a & 0x03030303UL) + \
|
||||||
OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
|
(b & 0x03030303UL) + \
|
||||||
pixels+=line_size;\
|
0x01010101UL; \
|
||||||
block +=line_size;\
|
h0 = ((a & 0xFCFCFCFCUL) >> 2) + \
|
||||||
a= AV_RN32(pixels );\
|
((b & 0xFCFCFCFCUL) >> 2); \
|
||||||
b= AV_RN32(pixels+1);\
|
l1 = (c & 0x03030303UL) + \
|
||||||
l0= (a&0x03030303UL)\
|
(d & 0x03030303UL); \
|
||||||
+ (b&0x03030303UL)\
|
h1 = ((c & 0xFCFCFCFCUL) >> 2) + \
|
||||||
+ 0x02020202UL;\
|
((d & 0xFCFCFCFCUL) >> 2); \
|
||||||
h0= ((a&0xFCFCFCFCUL)>>2)\
|
OP(*((uint32_t *) &dst[i * dst_stride + 4]), \
|
||||||
+ ((b&0xFCFCFCFCUL)>>2);\
|
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL)); \
|
||||||
OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
|
} \
|
||||||
pixels+=line_size;\
|
} \
|
||||||
block +=line_size;\
|
static inline void FUNC(OPNAME ## _pixels16_l4)(uint8_t *dst, \
|
||||||
}\
|
const uint8_t *src1, \
|
||||||
pixels+=4-line_size*(h+1);\
|
const uint8_t *src2, \
|
||||||
block +=4-line_size*h;\
|
const uint8_t *src3, \
|
||||||
}\
|
const uint8_t *src4, \
|
||||||
}\
|
int dst_stride, \
|
||||||
\
|
int src_stride1, \
|
||||||
CALL_2X_PIXELS(FUNCC(OPNAME ## _pixels16_xy2), FUNCC(OPNAME ## _pixels8_xy2), 8*sizeof(pixel))\
|
int src_stride2, \
|
||||||
|
int src_stride3, \
|
||||||
|
int src_stride4, \
|
||||||
|
int h) \
|
||||||
|
{ \
|
||||||
|
FUNC(OPNAME ## _pixels8_l4)(dst, src1, src2, src3, src4, dst_stride, \
|
||||||
|
src_stride1, src_stride2, src_stride3, \
|
||||||
|
src_stride4, h); \
|
||||||
|
FUNC(OPNAME ## _pixels8_l4)(dst + 8 * sizeof(pixel), \
|
||||||
|
src1 + 8 * sizeof(pixel), \
|
||||||
|
src2 + 8 * sizeof(pixel), \
|
||||||
|
src3 + 8 * sizeof(pixel), \
|
||||||
|
src4 + 8 * sizeof(pixel), \
|
||||||
|
dst_stride, src_stride1, src_stride2, \
|
||||||
|
src_stride3, src_stride4, h); \
|
||||||
|
} \
|
||||||
|
static inline void FUNC(OPNAME ## _no_rnd_pixels16_l4)(uint8_t *dst, \
|
||||||
|
const uint8_t *src1, \
|
||||||
|
const uint8_t *src2, \
|
||||||
|
const uint8_t *src3, \
|
||||||
|
const uint8_t *src4, \
|
||||||
|
int dst_stride, \
|
||||||
|
int src_stride1, \
|
||||||
|
int src_stride2, \
|
||||||
|
int src_stride3, \
|
||||||
|
int src_stride4, \
|
||||||
|
int h) \
|
||||||
|
{ \
|
||||||
|
FUNC(OPNAME ## _no_rnd_pixels8_l4)(dst, src1, src2, src3, src4, \
|
||||||
|
dst_stride, src_stride1, src_stride2, \
|
||||||
|
src_stride3, src_stride4, h); \
|
||||||
|
FUNC(OPNAME ## _no_rnd_pixels8_l4)(dst + 8 * sizeof(pixel), \
|
||||||
|
src1 + 8 * sizeof(pixel), \
|
||||||
|
src2 + 8 * sizeof(pixel), \
|
||||||
|
src3 + 8 * sizeof(pixel), \
|
||||||
|
src4 + 8 * sizeof(pixel), \
|
||||||
|
dst_stride, src_stride1, src_stride2, \
|
||||||
|
src_stride3, src_stride4, h); \
|
||||||
|
} \
|
||||||
|
\
|
||||||
|
static inline void FUNCC(OPNAME ## _pixels8_xy2)(uint8_t *block, \
|
||||||
|
const uint8_t *pixels, \
|
||||||
|
ptrdiff_t line_size, \
|
||||||
|
int h) \
|
||||||
|
{ \
|
||||||
|
/* FIXME HIGH BIT DEPTH */ \
|
||||||
|
int j; \
|
||||||
|
\
|
||||||
|
for (j = 0; j < 2; j++) { \
|
||||||
|
int i; \
|
||||||
|
const uint32_t a = AV_RN32(pixels); \
|
||||||
|
const uint32_t b = AV_RN32(pixels + 1); \
|
||||||
|
uint32_t l0 = (a & 0x03030303UL) + \
|
||||||
|
(b & 0x03030303UL) + \
|
||||||
|
0x02020202UL; \
|
||||||
|
uint32_t h0 = ((a & 0xFCFCFCFCUL) >> 2) + \
|
||||||
|
((b & 0xFCFCFCFCUL) >> 2); \
|
||||||
|
uint32_t l1, h1; \
|
||||||
|
\
|
||||||
|
pixels += line_size; \
|
||||||
|
for (i = 0; i < h; i += 2) { \
|
||||||
|
uint32_t a = AV_RN32(pixels); \
|
||||||
|
uint32_t b = AV_RN32(pixels + 1); \
|
||||||
|
l1 = (a & 0x03030303UL) + \
|
||||||
|
(b & 0x03030303UL); \
|
||||||
|
h1 = ((a & 0xFCFCFCFCUL) >> 2) + \
|
||||||
|
((b & 0xFCFCFCFCUL) >> 2); \
|
||||||
|
OP(*((uint32_t *) block), \
|
||||||
|
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL)); \
|
||||||
|
pixels += line_size; \
|
||||||
|
block += line_size; \
|
||||||
|
a = AV_RN32(pixels); \
|
||||||
|
b = AV_RN32(pixels + 1); \
|
||||||
|
l0 = (a & 0x03030303UL) + \
|
||||||
|
(b & 0x03030303UL) + \
|
||||||
|
0x02020202UL; \
|
||||||
|
h0 = ((a & 0xFCFCFCFCUL) >> 2) + \
|
||||||
|
((b & 0xFCFCFCFCUL) >> 2); \
|
||||||
|
OP(*((uint32_t *) block), \
|
||||||
|
h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL)); \
|
||||||
|
pixels += line_size; \
|
||||||
|
block += line_size; \
|
||||||
|
} \
|
||||||
|
pixels += 4 - line_size * (h + 1); \
|
||||||
|
block += 4 - line_size * h; \
|
||||||
|
} \
|
||||||
|
} \
|
||||||
|
\
|
||||||
|
CALL_2X_PIXELS(FUNCC(OPNAME ## _pixels16_xy2), \
|
||||||
|
FUNCC(OPNAME ## _pixels8_xy2), \
|
||||||
|
8 * sizeof(pixel)) \
|
||||||
|
|
||||||
#define op_avg(a, b) a = rnd_avg_pixel4(a, b)
|
#define op_avg(a, b) a = rnd_avg_pixel4(a, b)
|
||||||
#define op_put(a, b) a = b
|
#define op_put(a, b) a = b
|
||||||
|
Loading…
x
Reference in New Issue
Block a user