* DSP utils
*/
+#include "libavutil/attributes.h"
#include "libavutil/imgutils.h"
+#include "libavutil/internal.h"
#include "avcodec.h"
+#include "copy_block.h"
+#include "dct.h"
#include "dsputil.h"
#include "simple_idct.h"
#include "faandct.h"
#include "faanidct.h"
+#include "imgconvert.h"
#include "mathops.h"
#include "mpegvideo.h"
#include "config.h"
-uint8_t ff_cropTbl[256 + 2 * MAX_NEG_CROP] = {0, };
uint32_t ff_squareTbl[512] = {0, };
-#define BIT_DEPTH 9
-#include "dsputil_template.c"
-#undef BIT_DEPTH
-
-#define BIT_DEPTH 10
+#define BIT_DEPTH 16
#include "dsputil_template.c"
#undef BIT_DEPTH
#define pb_7f (~0UL/255 * 0x7f)
#define pb_80 (~0UL/255 * 0x80)
-const uint8_t ff_zigzag_direct[64] = {
- 0, 1, 8, 16, 9, 2, 3, 10,
- 17, 24, 32, 25, 18, 11, 4, 5,
- 12, 19, 26, 33, 40, 48, 41, 34,
- 27, 20, 13, 6, 7, 14, 21, 28,
- 35, 42, 49, 56, 57, 50, 43, 36,
- 29, 22, 15, 23, 30, 37, 44, 51,
- 58, 59, 52, 45, 38, 31, 39, 46,
- 53, 60, 61, 54, 47, 55, 62, 63
-};
-
/* Specific zigzag scan for 248 idct. NOTE that unlike the
specification, we interleave the fields */
const uint8_t ff_zigzag248_direct[64] = {
53, 61, 54, 62, 39, 47, 55, 63,
};
-/* not permutated inverse zigzag_direct + 1 for MMX quantizer */
-DECLARE_ALIGNED(16, uint16_t, ff_inv_zigzag_direct16)[64];
-
const uint8_t ff_alternate_horizontal_scan[64] = {
0, 1, 2, 3, 8, 9, 16, 17,
10, 11, 4, 5, 6, 7, 15, 14,
static const uint8_t idct_sse2_row_perm[8] = {0, 4, 1, 5, 2, 6, 3, 7};
-void ff_init_scantable(uint8_t *permutation, ScanTable *st, const uint8_t *src_scantable){
+av_cold void ff_init_scantable(uint8_t *permutation, ScanTable *st,
+ const uint8_t *src_scantable)
+{
int i;
int end;
}
}
-void ff_init_scantable_permutation(uint8_t *idct_permutation,
- int idct_permutation_type)
+av_cold void ff_init_scantable_permutation(uint8_t *idct_permutation,
+ int idct_permutation_type)
{
int i;
}
}
+static void add_pixels8_c(uint8_t *restrict pixels,
+ int16_t *block,
+ int line_size)
+{
+ int i;
+
+ for(i=0;i<8;i++) {
+ pixels[0] += block[0];
+ pixels[1] += block[1];
+ pixels[2] += block[2];
+ pixels[3] += block[3];
+ pixels[4] += block[4];
+ pixels[5] += block[5];
+ pixels[6] += block[6];
+ pixels[7] += block[7];
+ pixels += line_size;
+ block += 8;
+ }
+}
+
static void add_pixels_clamped_c(const int16_t *block, uint8_t *restrict pixels,
int line_size)
{
#define QPEL_MC(r, OPNAME, RND, OP) \
static void OPNAME ## mpeg4_qpel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){\
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
+ const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
int i;\
for(i=0; i<h; i++)\
{\
\
static void OPNAME ## mpeg4_qpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
const int w=8;\
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
+ const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
int i;\
for(i=0; i<w; i++)\
{\
}\
\
static void OPNAME ## mpeg4_qpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){\
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
+ const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
int i;\
\
for(i=0; i<h; i++)\
}\
\
static void OPNAME ## mpeg4_qpel16_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
+ const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
int i;\
const int w=16;\
for(i=0; i<w; i++)\
}\
}\
\
-static void OPNAME ## qpel8_mc10_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc10_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t half[64];\
put ## RND ## mpeg4_qpel8_h_lowpass(half, src, 8, stride, 8);\
OPNAME ## pixels8_l2_8(dst, src, half, stride, stride, 8, 8);\
}\
\
-static void OPNAME ## qpel8_mc20_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc20_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
OPNAME ## mpeg4_qpel8_h_lowpass(dst, src, stride, stride, 8);\
}\
\
-static void OPNAME ## qpel8_mc30_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc30_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t half[64];\
put ## RND ## mpeg4_qpel8_h_lowpass(half, src, 8, stride, 8);\
OPNAME ## pixels8_l2_8(dst, src+1, half, stride, stride, 8, 8);\
}\
\
-static void OPNAME ## qpel8_mc01_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc01_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t half[64];\
copy_block9(full, src, 16, stride, 9);\
OPNAME ## pixels8_l2_8(dst, full, half, stride, 16, 8, 8);\
}\
\
-static void OPNAME ## qpel8_mc02_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc02_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
copy_block9(full, src, 16, stride, 9);\
OPNAME ## mpeg4_qpel8_v_lowpass(dst, full, stride, 16);\
}\
\
-static void OPNAME ## qpel8_mc03_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc03_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t half[64];\
copy_block9(full, src, 16, stride, 9);\
put ## RND ## mpeg4_qpel8_v_lowpass(half, full, 8, 16);\
OPNAME ## pixels8_l2_8(dst, full+16, half, stride, 16, 8, 8);\
}\
-void ff_ ## OPNAME ## qpel8_mc11_old_c(uint8_t *dst, uint8_t *src, int stride){\
+void ff_ ## OPNAME ## qpel8_mc11_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfV[64];\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l4_8(dst, full, halfH, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
}\
-static void OPNAME ## qpel8_mc11_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc11_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfHV[64];\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfH, halfHV, stride, 8, 8, 8);\
}\
-void ff_ ## OPNAME ## qpel8_mc31_old_c(uint8_t *dst, uint8_t *src, int stride){\
+void ff_ ## OPNAME ## qpel8_mc31_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfV[64];\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l4_8(dst, full+1, halfH, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
}\
-static void OPNAME ## qpel8_mc31_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc31_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfHV[64];\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfH, halfHV, stride, 8, 8, 8);\
}\
-void ff_ ## OPNAME ## qpel8_mc13_old_c(uint8_t *dst, uint8_t *src, int stride){\
+void ff_ ## OPNAME ## qpel8_mc13_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfV[64];\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l4_8(dst, full+16, halfH+8, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
}\
-static void OPNAME ## qpel8_mc13_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc13_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfHV[64];\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfH+8, halfHV, stride, 8, 8, 8);\
}\
-void ff_ ## OPNAME ## qpel8_mc33_old_c(uint8_t *dst, uint8_t *src, int stride){\
+void ff_ ## OPNAME ## qpel8_mc33_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfV[64];\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l4_8(dst, full+17, halfH+8, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
}\
-static void OPNAME ## qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfHV[64];\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfH+8, halfHV, stride, 8, 8, 8);\
}\
-static void OPNAME ## qpel8_mc21_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc21_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t halfH[72];\
uint8_t halfHV[64];\
put ## RND ## mpeg4_qpel8_h_lowpass(halfH, src, 8, stride, 9);\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfH, halfHV, stride, 8, 8, 8);\
}\
-static void OPNAME ## qpel8_mc23_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc23_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t halfH[72];\
uint8_t halfHV[64];\
put ## RND ## mpeg4_qpel8_h_lowpass(halfH, src, 8, stride, 9);\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfH+8, halfHV, stride, 8, 8, 8);\
}\
-void ff_ ## OPNAME ## qpel8_mc12_old_c(uint8_t *dst, uint8_t *src, int stride){\
+void ff_ ## OPNAME ## qpel8_mc12_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfV[64];\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfV, halfHV, stride, 8, 8, 8);\
}\
-static void OPNAME ## qpel8_mc12_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc12_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t halfH[72];\
copy_block9(full, src, 16, stride, 9);\
put ## RND ## pixels8_l2_8(halfH, halfH, full, 8, 8, 16, 9);\
OPNAME ## mpeg4_qpel8_v_lowpass(dst, halfH, stride, 8);\
}\
-void ff_ ## OPNAME ## qpel8_mc32_old_c(uint8_t *dst, uint8_t *src, int stride){\
+void ff_ ## OPNAME ## qpel8_mc32_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfV[64];\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfV, halfHV, stride, 8, 8, 8);\
}\
-static void OPNAME ## qpel8_mc32_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc32_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[16*9];\
uint8_t halfH[72];\
copy_block9(full, src, 16, stride, 9);\
put ## RND ## pixels8_l2_8(halfH, halfH, full+1, 8, 8, 16, 9);\
OPNAME ## mpeg4_qpel8_v_lowpass(dst, halfH, stride, 8);\
}\
-static void OPNAME ## qpel8_mc22_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel8_mc22_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t halfH[72];\
put ## RND ## mpeg4_qpel8_h_lowpass(halfH, src, 8, stride, 9);\
OPNAME ## mpeg4_qpel8_v_lowpass(dst, halfH, stride, 8);\
}\
\
-static void OPNAME ## qpel16_mc10_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc10_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t half[256];\
put ## RND ## mpeg4_qpel16_h_lowpass(half, src, 16, stride, 16);\
OPNAME ## pixels16_l2_8(dst, src, half, stride, stride, 16, 16);\
}\
\
-static void OPNAME ## qpel16_mc20_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc20_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
OPNAME ## mpeg4_qpel16_h_lowpass(dst, src, stride, stride, 16);\
}\
\
-static void OPNAME ## qpel16_mc30_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc30_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t half[256];\
put ## RND ## mpeg4_qpel16_h_lowpass(half, src, 16, stride, 16);\
OPNAME ## pixels16_l2_8(dst, src+1, half, stride, stride, 16, 16);\
}\
\
-static void OPNAME ## qpel16_mc01_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc01_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t half[256];\
copy_block17(full, src, 24, stride, 17);\
OPNAME ## pixels16_l2_8(dst, full, half, stride, 24, 16, 16);\
}\
\
-static void OPNAME ## qpel16_mc02_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc02_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
copy_block17(full, src, 24, stride, 17);\
OPNAME ## mpeg4_qpel16_v_lowpass(dst, full, stride, 24);\
}\
\
-static void OPNAME ## qpel16_mc03_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc03_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t half[256];\
copy_block17(full, src, 24, stride, 17);\
put ## RND ## mpeg4_qpel16_v_lowpass(half, full, 16, 24);\
OPNAME ## pixels16_l2_8(dst, full+24, half, stride, 24, 16, 16);\
}\
-void ff_ ## OPNAME ## qpel16_mc11_old_c(uint8_t *dst, uint8_t *src, int stride){\
+void ff_ ## OPNAME ## qpel16_mc11_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfV[256];\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l4_8(dst, full, halfH, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
}\
-static void OPNAME ## qpel16_mc11_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc11_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfHV[256];\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfH, halfHV, stride, 16, 16, 16);\
}\
-void ff_ ## OPNAME ## qpel16_mc31_old_c(uint8_t *dst, uint8_t *src, int stride){\
+void ff_ ## OPNAME ## qpel16_mc31_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfV[256];\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l4_8(dst, full+1, halfH, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
}\
-static void OPNAME ## qpel16_mc31_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc31_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfHV[256];\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfH, halfHV, stride, 16, 16, 16);\
}\
-void ff_ ## OPNAME ## qpel16_mc13_old_c(uint8_t *dst, uint8_t *src, int stride){\
+void ff_ ## OPNAME ## qpel16_mc13_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfV[256];\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l4_8(dst, full+24, halfH+16, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
}\
-static void OPNAME ## qpel16_mc13_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc13_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfHV[256];\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfH+16, halfHV, stride, 16, 16, 16);\
}\
-void ff_ ## OPNAME ## qpel16_mc33_old_c(uint8_t *dst, uint8_t *src, int stride){\
+void ff_ ## OPNAME ## qpel16_mc33_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfV[256];\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l4_8(dst, full+25, halfH+16, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
}\
-static void OPNAME ## qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfHV[256];\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfH+16, halfHV, stride, 16, 16, 16);\
}\
-static void OPNAME ## qpel16_mc21_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc21_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t halfH[272];\
uint8_t halfHV[256];\
put ## RND ## mpeg4_qpel16_h_lowpass(halfH, src, 16, stride, 17);\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfH, halfHV, stride, 16, 16, 16);\
}\
-static void OPNAME ## qpel16_mc23_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc23_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t halfH[272];\
uint8_t halfHV[256];\
put ## RND ## mpeg4_qpel16_h_lowpass(halfH, src, 16, stride, 17);\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfH+16, halfHV, stride, 16, 16, 16);\
}\
-void ff_ ## OPNAME ## qpel16_mc12_old_c(uint8_t *dst, uint8_t *src, int stride){\
+void ff_ ## OPNAME ## qpel16_mc12_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfV[256];\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfV, halfHV, stride, 16, 16, 16);\
}\
-static void OPNAME ## qpel16_mc12_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc12_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t halfH[272];\
copy_block17(full, src, 24, stride, 17);\
put ## RND ## pixels16_l2_8(halfH, halfH, full, 16, 16, 24, 17);\
OPNAME ## mpeg4_qpel16_v_lowpass(dst, halfH, stride, 16);\
}\
-void ff_ ## OPNAME ## qpel16_mc32_old_c(uint8_t *dst, uint8_t *src, int stride){\
+void ff_ ## OPNAME ## qpel16_mc32_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfV[256];\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfV, halfHV, stride, 16, 16, 16);\
}\
-static void OPNAME ## qpel16_mc32_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc32_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t full[24*17];\
uint8_t halfH[272];\
copy_block17(full, src, 24, stride, 17);\
put ## RND ## pixels16_l2_8(halfH, halfH, full+1, 16, 16, 24, 17);\
OPNAME ## mpeg4_qpel16_v_lowpass(dst, halfH, stride, 16);\
}\
-static void OPNAME ## qpel16_mc22_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## qpel16_mc22_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
+{\
uint8_t halfH[272];\
put ## RND ## mpeg4_qpel16_h_lowpass(halfH, src, 16, stride, 17);\
OPNAME ## mpeg4_qpel16_v_lowpass(dst, halfH, stride, 16);\
#undef op_put
#undef op_put_no_rnd
+void ff_put_pixels8x8_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
+ put_pixels8_8_c(dst, src, stride, 8);
+}
+void ff_avg_pixels8x8_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
+ avg_pixels8_8_c(dst, src, stride, 8);
+}
+void ff_put_pixels16x16_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
+ put_pixels16_8_c(dst, src, stride, 16);
+}
+void ff_avg_pixels16x16_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
+ avg_pixels16_8_c(dst, src, stride, 16);
+}
+
#define put_qpel8_mc00_c ff_put_pixels8x8_c
#define avg_qpel8_mc00_c ff_avg_pixels8x8_c
#define put_qpel16_mc00_c ff_put_pixels16x16_c
#define avg_qpel16_mc00_c ff_avg_pixels16x16_c
#define put_no_rnd_qpel8_mc00_c ff_put_pixels8x8_c
-#define put_no_rnd_qpel16_mc00_c ff_put_pixels16x16_8_c
+#define put_no_rnd_qpel16_mc00_c ff_put_pixels16x16_c
static void wmv2_mspel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
+ const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
int i;
for(i=0; i<h; i++){
}
#if CONFIG_RV40_DECODER
-void ff_put_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){
+void ff_put_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
put_pixels16_xy2_8_c(dst, src, stride, 16);
}
-void ff_avg_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){
+void ff_avg_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
avg_pixels16_xy2_8_c(dst, src, stride, 16);
}
-void ff_put_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){
+void ff_put_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
put_pixels8_xy2_8_c(dst, src, stride, 8);
}
-void ff_avg_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){
+void ff_avg_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
avg_pixels8_xy2_8_c(dst, src, stride, 8);
}
#endif /* CONFIG_RV40_DECODER */
static void wmv2_mspel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int w){
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
+ const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
int i;
for(i=0; i<w; i++){
}
}
-static void put_mspel8_mc10_c(uint8_t *dst, uint8_t *src, int stride){
+static void put_mspel8_mc10_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
uint8_t half[64];
wmv2_mspel8_h_lowpass(half, src, 8, stride, 8);
put_pixels8_l2_8(dst, src, half, stride, stride, 8, 8);
}
-static void put_mspel8_mc20_c(uint8_t *dst, uint8_t *src, int stride){
+static void put_mspel8_mc20_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
wmv2_mspel8_h_lowpass(dst, src, stride, stride, 8);
}
-static void put_mspel8_mc30_c(uint8_t *dst, uint8_t *src, int stride){
+static void put_mspel8_mc30_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
uint8_t half[64];
wmv2_mspel8_h_lowpass(half, src, 8, stride, 8);
put_pixels8_l2_8(dst, src+1, half, stride, stride, 8, 8);
}
-static void put_mspel8_mc02_c(uint8_t *dst, uint8_t *src, int stride){
+static void put_mspel8_mc02_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
wmv2_mspel8_v_lowpass(dst, src, stride, stride, 8);
}
-static void put_mspel8_mc12_c(uint8_t *dst, uint8_t *src, int stride){
+static void put_mspel8_mc12_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
uint8_t halfH[88];
uint8_t halfV[64];
uint8_t halfHV[64];
wmv2_mspel8_v_lowpass(halfHV, halfH+8, 8, 8, 8);
put_pixels8_l2_8(dst, halfV, halfHV, stride, 8, 8, 8);
}
-static void put_mspel8_mc32_c(uint8_t *dst, uint8_t *src, int stride){
+static void put_mspel8_mc32_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
uint8_t halfH[88];
uint8_t halfV[64];
uint8_t halfHV[64];
wmv2_mspel8_v_lowpass(halfHV, halfH+8, 8, 8, 8);
put_pixels8_l2_8(dst, halfV, halfHV, stride, 8, 8, 8);
}
-static void put_mspel8_mc22_c(uint8_t *dst, uint8_t *src, int stride){
+static void put_mspel8_mc22_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
+{
uint8_t halfH[88];
wmv2_mspel8_h_lowpass(halfH, src-stride, 8, stride, 11);
wmv2_mspel8_v_lowpass(dst, halfH+8, stride, 8, 8);
}
-static void h263_v_loop_filter_c(uint8_t *src, int stride, int qscale){
- if(CONFIG_H263_DECODER || CONFIG_H263_ENCODER) {
- int x;
- const int strength= ff_h263_loop_filter_strength[qscale];
-
- for(x=0; x<8; x++){
- int d1, d2, ad1;
- int p0= src[x-2*stride];
- int p1= src[x-1*stride];
- int p2= src[x+0*stride];
- int p3= src[x+1*stride];
- int d = (p0 - p3 + 4*(p2 - p1)) / 8;
-
- if (d<-2*strength) d1= 0;
- else if(d<- strength) d1=-2*strength - d;
- else if(d< strength) d1= d;
- else if(d< 2*strength) d1= 2*strength - d;
- else d1= 0;
-
- p1 += d1;
- p2 -= d1;
- if(p1&256) p1= ~(p1>>31);
- if(p2&256) p2= ~(p2>>31);
-
- src[x-1*stride] = p1;
- src[x+0*stride] = p2;
-
- ad1= FFABS(d1)>>1;
-
- d2= av_clip((p0-p3)/4, -ad1, ad1);
-
- src[x-2*stride] = p0 - d2;
- src[x+ stride] = p3 + d2;
- }
- }
-}
-
-static void h263_h_loop_filter_c(uint8_t *src, int stride, int qscale){
- if(CONFIG_H263_DECODER || CONFIG_H263_ENCODER) {
- int y;
- const int strength= ff_h263_loop_filter_strength[qscale];
-
- for(y=0; y<8; y++){
- int d1, d2, ad1;
- int p0= src[y*stride-2];
- int p1= src[y*stride-1];
- int p2= src[y*stride+0];
- int p3= src[y*stride+1];
- int d = (p0 - p3 + 4*(p2 - p1)) / 8;
-
- if (d<-2*strength) d1= 0;
- else if(d<- strength) d1=-2*strength - d;
- else if(d< strength) d1= d;
- else if(d< 2*strength) d1= 2*strength - d;
- else d1= 0;
-
- p1 += d1;
- p2 -= d1;
- if(p1&256) p1= ~(p1>>31);
- if(p2&256) p2= ~(p2>>31);
-
- src[y*stride-1] = p1;
- src[y*stride+0] = p2;
-
- ad1= FFABS(d1)>>1;
-
- d2= av_clip((p0-p3)/4, -ad1, ad1);
-
- src[y*stride-2] = p0 - d2;
- src[y*stride+1] = p3 + d2;
- }
- }
-}
-
-static void h261_loop_filter_c(uint8_t *src, int stride){
- int x,y,xy,yz;
- int temp[64];
-
- for(x=0; x<8; x++){
- temp[x ] = 4*src[x ];
- temp[x + 7*8] = 4*src[x + 7*stride];
- }
- for(y=1; y<7; y++){
- for(x=0; x<8; x++){
- xy = y * stride + x;
- yz = y * 8 + x;
- temp[yz] = src[xy - stride] + 2*src[xy] + src[xy + stride];
- }
- }
-
- for(y=0; y<8; y++){
- src[ y*stride] = (temp[ y*8] + 2)>>2;
- src[7+y*stride] = (temp[7+y*8] + 2)>>2;
- for(x=1; x<7; x++){
- xy = y * stride + x;
- yz = y * 8 + x;
- src[xy] = (temp[yz-1] + 2*temp[yz] + temp[yz+1] + 8)>>4;
- }
- }
-}
-
static inline int pix_abs16_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
{
int s, i;
}
}
-/**
- * Permute an 8x8 block.
- * @param block the block which will be permuted according to the given permutation vector
- * @param permutation the permutation vector
- * @param last the last non zero coefficient in scantable order, used to speed the permutation up
- * @param scantable the used scantable, this is only used to speed the permutation up, the block is not
- * (inverse) permutated to scantable order!
- */
-void ff_block_permute(int16_t *block, uint8_t *permutation, const uint8_t *scantable, int last)
-{
- int i;
- int16_t temp[64];
-
- if(last<=0) return;
- //if(permutation[1]==1) return; //FIXME it is ok but not clean and might fail for some permutations
-
- for(i=0; i<=last; i++){
- const int j= scantable[i];
- temp[j]= block[j];
- block[j]=0;
- }
-
- for(i=0; i<=last; i++){
- const int j= scantable[i];
- const int perm_j= permutation[j];
- block[perm_j]= temp[j];
- }
-}
-
static int zero_cmp(void *s, uint8_t *a, uint8_t *b, int stride, int h){
return 0;
}
static void add_bytes_c(uint8_t *dst, uint8_t *src, int w){
long i;
- for(i=0; i<=w-sizeof(long); i+=sizeof(long)){
+ for (i = 0; i <= w - (int) sizeof(long); i += sizeof(long)) {
long a = *(long*)(src+i);
long b = *(long*)(dst+i);
*(long*)(dst+i) = ((a&pb_7f) + (b&pb_7f)) ^ ((a^b)&pb_80);
}
}else
#endif
- for(i=0; i<=w-sizeof(long); i+=sizeof(long)){
+ for (i = 0; i <= w - (int) sizeof(long); i += sizeof(long)) {
long a = *(long*)(src1+i);
long b = *(long*)(src2+i);
*(long*)(dst+i) = ((a|pb_80) - (b&pb_7f)) ^ ((a^b^pb_80)&pb_80);
return score;
}
+#define WRAPPER8_16_SQ(name8, name16)\
+static int name16(void /*MpegEncContext*/ *s, uint8_t *dst, uint8_t *src, int stride, int h){\
+ int score=0;\
+ score +=name8(s, dst , src , stride, 8);\
+ score +=name8(s, dst+8 , src+8 , stride, 8);\
+ if(h==16){\
+ dst += 8*stride;\
+ src += 8*stride;\
+ score +=name8(s, dst , src , stride, 8);\
+ score +=name8(s, dst+8 , src+8 , stride, 8);\
+ }\
+ return score;\
+}
+
WRAPPER8_16_SQ(hadamard8_diff8x8_c, hadamard8_diff16_c)
WRAPPER8_16_SQ(hadamard8_intra8x8_c, hadamard8_intra16_c)
WRAPPER8_16_SQ(dct_sad8x8_c, dct_sad16_c)
return res;
}
-static void apply_window_int16_c(int16_t *output, const int16_t *input,
- const int16_t *window, unsigned int len)
-{
- int i;
- int len2 = len >> 1;
-
- for (i = 0; i < len2; i++) {
- int16_t w = window[i];
- output[i] = (MUL16(input[i], w) + (1 << 14)) >> 15;
- output[len-i-1] = (MUL16(input[len-i-1], w) + (1 << 14)) >> 15;
- }
-}
-
static void vector_clip_int32_c(int32_t *dst, const int32_t *src, int32_t min,
int32_t max, unsigned int len)
{
} while (len > 0);
}
-static void ff_jref_idct_put(uint8_t *dest, int line_size, int16_t *block)
+static void jref_idct_put(uint8_t *dest, int line_size, int16_t *block)
{
ff_j_rev_dct (block);
put_pixels_clamped_c(block, dest, line_size);
}
-static void ff_jref_idct_add(uint8_t *dest, int line_size, int16_t *block)
+static void jref_idct_add(uint8_t *dest, int line_size, int16_t *block)
{
ff_j_rev_dct (block);
add_pixels_clamped_c(block, dest, line_size);
{
int i;
- for(i=0;i<256;i++) ff_cropTbl[i + MAX_NEG_CROP] = i;
- for(i=0;i<MAX_NEG_CROP;i++) {
- ff_cropTbl[i] = 0;
- ff_cropTbl[i + MAX_NEG_CROP + 256] = 255;
- }
-
for(i=0;i<512;i++) {
ff_squareTbl[i] = (i - 256) * (i - 256);
}
-
- for(i=0; i<64; i++) ff_inv_zigzag_direct16[ff_zigzag_direct[i]]= i+1;
}
int ff_check_alignment(void){
c->idct_permutation_type = FF_NO_IDCT_PERM;
} else {
if(avctx->idct_algo==FF_IDCT_INT){
- c->idct_put= ff_jref_idct_put;
- c->idct_add= ff_jref_idct_add;
+ c->idct_put= jref_idct_put;
+ c->idct_add= jref_idct_add;
c->idct = ff_j_rev_dct;
c->idct_permutation_type= FF_LIBMPEG2_IDCT_PERM;
}else if(avctx->idct_algo==FF_IDCT_FAAN){
c->bswap_buf= bswap_buf;
c->bswap16_buf = bswap16_buf;
- if (CONFIG_H263_DECODER || CONFIG_H263_ENCODER) {
- c->h263_h_loop_filter= h263_h_loop_filter_c;
- c->h263_v_loop_filter= h263_v_loop_filter_c;
- }
-
- c->h261_loop_filter= h261_loop_filter_c;
-
c->try_8x8basis= try_8x8basis_c;
c->add_8x8basis= add_8x8basis_c;
c->vector_clipf = vector_clipf_c;
c->scalarproduct_int16 = scalarproduct_int16_c;
c->scalarproduct_and_madd_int16 = scalarproduct_and_madd_int16_c;
- c->apply_window_int16 = apply_window_int16_c;
c->vector_clip_int32 = vector_clip_int32_c;
c->shrink[0]= av_image_copy_plane;
c->shrink[2]= ff_shrink44;
c->shrink[3]= ff_shrink88;
-#define hpel_funcs(prefix, idx, num) \
- c->prefix ## _pixels_tab idx [0] = prefix ## _pixels ## num ## _8_c; \
- c->prefix ## _pixels_tab idx [1] = prefix ## _pixels ## num ## _x2_8_c; \
- c->prefix ## _pixels_tab idx [2] = prefix ## _pixels ## num ## _y2_8_c; \
- c->prefix ## _pixels_tab idx [3] = prefix ## _pixels ## num ## _xy2_8_c
-
- hpel_funcs(put, [0], 16);
- hpel_funcs(put, [1], 8);
- hpel_funcs(put, [2], 4);
- hpel_funcs(put, [3], 2);
- hpel_funcs(put_no_rnd, [0], 16);
- hpel_funcs(put_no_rnd, [1], 8);
- hpel_funcs(avg, [0], 16);
- hpel_funcs(avg, [1], 8);
- hpel_funcs(avg, [2], 4);
- hpel_funcs(avg, [3], 2);
- hpel_funcs(avg_no_rnd,, 16);
+ c->add_pixels8 = add_pixels8_c;
#undef FUNC
#undef FUNCC
#define FUNC(f, depth) f ## _ ## depth
#define FUNCC(f, depth) f ## _ ## depth ## _c
-#define BIT_DEPTH_FUNCS(depth, dct)\
- c->get_pixels = FUNCC(get_pixels ## dct , depth);\
- c->draw_edges = FUNCC(draw_edges , depth);\
- c->clear_block = FUNCC(clear_block ## dct , depth);\
- c->clear_blocks = FUNCC(clear_blocks ## dct , depth);\
- c->add_pixels8 = FUNCC(add_pixels8 ## dct , depth);\
- c->add_pixels4 = FUNCC(add_pixels4 ## dct , depth);\
+ c->draw_edges = FUNCC(draw_edges, 8);
+ c->clear_block = FUNCC(clear_block, 8);
+ c->clear_blocks = FUNCC(clear_blocks, 8);
+
+#define BIT_DEPTH_FUNCS(depth) \
+ c->get_pixels = FUNCC(get_pixels, depth);
switch (avctx->bits_per_raw_sample) {
case 9:
- if (c->dct_bits == 32) {
- BIT_DEPTH_FUNCS(9, _32);
- } else {
- BIT_DEPTH_FUNCS(9, _16);
- }
- break;
case 10:
- if (c->dct_bits == 32) {
- BIT_DEPTH_FUNCS(10, _32);
- } else {
- BIT_DEPTH_FUNCS(10, _16);
- }
+ BIT_DEPTH_FUNCS(16);
break;
default:
- BIT_DEPTH_FUNCS(8, _16);
+ BIT_DEPTH_FUNCS(8);
break;
}
- if (HAVE_MMX) ff_dsputil_init_mmx (c, avctx);
- if (ARCH_ARM) ff_dsputil_init_arm (c, avctx);
- if (HAVE_VIS) ff_dsputil_init_vis (c, avctx);
- if (ARCH_ALPHA) ff_dsputil_init_alpha (c, avctx);
- if (ARCH_PPC) ff_dsputil_init_ppc (c, avctx);
- if (ARCH_SH4) ff_dsputil_init_sh4 (c, avctx);
- if (ARCH_BFIN) ff_dsputil_init_bfin (c, avctx);
+ if (ARCH_ARM)
+ ff_dsputil_init_arm(c, avctx);
+ if (ARCH_BFIN)
+ ff_dsputil_init_bfin(c, avctx);
+ if (ARCH_PPC)
+ ff_dsputil_init_ppc(c, avctx);
+ if (ARCH_SH4)
+ ff_dsputil_init_sh4(c, avctx);
+ if (HAVE_VIS)
+ ff_dsputil_init_vis(c, avctx);
+ if (ARCH_X86)
+ ff_dsputil_init_x86(c, avctx);
ff_init_scantable_permutation(c->idct_permutation,
c->idct_permutation_type);