| de6d9b64 |
/* |
| 162cffca |
* DSP utils
* Copyright (c) 2000, 2001 Fabrice Bellard
* Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
* |
| b78e7197 |
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or |
| ff4ec49e |
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either |
| b78e7197 |
* version 2.1 of the License, or (at your option) any later version. |
| de6d9b64 |
* |
| b78e7197 |
* FFmpeg is distributed in the hope that it will be useful, |
| de6d9b64 |
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
| ff4ec49e |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details. |
| de6d9b64 |
* |
| ff4ec49e |
* You should have received a copy of the GNU Lesser General Public |
| b78e7197 |
* License along with FFmpeg; if not, write to the Free Software |
| 5509bffa |
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
| de6d9b64 |
*/ |
| 115329f1 |
|
| 6fee1b90 |
#include "libavutil/attributes.h" |
| 218aefce |
#include "libavutil/internal.h" |
| de6d9b64 |
#include "avcodec.h" |
| b9ba5253 |
#include "copy_block.h" |
| b0368839 |
#include "simple_idct.h" |
| 2d604443 |
#include "me_cmp.h" |
| af818f7a |
#include "mpegvideo.h"
#include "config.h" |
| 5596c60c |
|
| 05563cca |
uint32_t ff_square_tab[512] = { 0, }; |
| de6d9b64 |
|
| 82bb3048 |
static int sse4_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| 26efc54e |
{ |
| 635ec127 |
int s = 0, i; |
| 05563cca |
uint32_t *sq = ff_square_tab + 256; |
| 26efc54e |
for (i = 0; i < h; i++) { |
| bb12c2e9 |
s += sq[pix1[0] - pix2[0]];
s += sq[pix1[1] - pix2[1]];
s += sq[pix1[2] - pix2[2]];
s += sq[pix1[3] - pix2[3]]; |
| 9c12c6ff |
pix1 += stride;
pix2 += stride; |
| 26efc54e |
}
return s;
}
|
| 82bb3048 |
static int sse8_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| 1457ab52 |
{ |
| 635ec127 |
int s = 0, i; |
| 05563cca |
uint32_t *sq = ff_square_tab + 256; |
| 1457ab52 |
|
| bb198e19 |
for (i = 0; i < h; i++) { |
| bb12c2e9 |
s += sq[pix1[0] - pix2[0]];
s += sq[pix1[1] - pix2[1]];
s += sq[pix1[2] - pix2[2]];
s += sq[pix1[3] - pix2[3]];
s += sq[pix1[4] - pix2[4]];
s += sq[pix1[5] - pix2[5]];
s += sq[pix1[6] - pix2[6]];
s += sq[pix1[7] - pix2[7]]; |
| 9c12c6ff |
pix1 += stride;
pix2 += stride; |
| 1457ab52 |
}
return s;
}
|
| 82bb3048 |
static int sse16_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| 9c76bd48 |
{ |
| 635ec127 |
int s = 0, i; |
| 05563cca |
uint32_t *sq = ff_square_tab + 256; |
| 9c76bd48 |
|
| bb198e19 |
for (i = 0; i < h; i++) { |
| bb12c2e9 |
s += sq[pix1[0] - pix2[0]];
s += sq[pix1[1] - pix2[1]];
s += sq[pix1[2] - pix2[2]];
s += sq[pix1[3] - pix2[3]];
s += sq[pix1[4] - pix2[4]];
s += sq[pix1[5] - pix2[5]];
s += sq[pix1[6] - pix2[6]];
s += sq[pix1[7] - pix2[7]];
s += sq[pix1[8] - pix2[8]];
s += sq[pix1[9] - pix2[9]]; |
| 6b026927 |
s += sq[pix1[10] - pix2[10]];
s += sq[pix1[11] - pix2[11]];
s += sq[pix1[12] - pix2[12]];
s += sq[pix1[13] - pix2[13]];
s += sq[pix1[14] - pix2[14]];
s += sq[pix1[15] - pix2[15]]; |
| 2a006cd3 |
|
| 9c12c6ff |
pix1 += stride;
pix2 += stride; |
| 9c76bd48 |
}
return s;
}
|
| 88bd7fdc |
static int sum_abs_dctelem_c(int16_t *block) |
| 1edbfe19 |
{ |
| bb12c2e9 |
int sum = 0, i;
for (i = 0; i < 64; i++)
sum += FFABS(block[i]); |
| 1edbfe19 |
return sum;
}
|
| 23578226 |
#define avg2(a, b) (((a) + (b) + 1) >> 1)
#define avg4(a, b, c, d) (((a) + (b) + (c) + (d) + 2) >> 2) |
| de6d9b64 |
|
| 82bb3048 |
static inline int pix_abs16_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| de6d9b64 |
{ |
| 635ec127 |
int s = 0, i; |
| de6d9b64 |
|
| bb12c2e9 |
for (i = 0; i < h; i++) {
s += abs(pix1[0] - pix2[0]);
s += abs(pix1[1] - pix2[1]);
s += abs(pix1[2] - pix2[2]);
s += abs(pix1[3] - pix2[3]);
s += abs(pix1[4] - pix2[4]);
s += abs(pix1[5] - pix2[5]);
s += abs(pix1[6] - pix2[6]);
s += abs(pix1[7] - pix2[7]);
s += abs(pix1[8] - pix2[8]);
s += abs(pix1[9] - pix2[9]);
s += abs(pix1[10] - pix2[10]);
s += abs(pix1[11] - pix2[11]);
s += abs(pix1[12] - pix2[12]);
s += abs(pix1[13] - pix2[13]);
s += abs(pix1[14] - pix2[14]);
s += abs(pix1[15] - pix2[15]); |
| 9c12c6ff |
pix1 += stride;
pix2 += stride; |
| de6d9b64 |
}
return s;
}
|
| 4edd74bd |
static inline int pix_median_abs16_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
ptrdiff_t stride, int h)
{
int s = 0, i, j;
#define V(x) (pix1[x] - pix2[x])
s += abs(V(0));
s += abs(V(1) - V(0));
s += abs(V(2) - V(1));
s += abs(V(3) - V(2));
s += abs(V(4) - V(3));
s += abs(V(5) - V(4));
s += abs(V(6) - V(5));
s += abs(V(7) - V(6));
s += abs(V(8) - V(7));
s += abs(V(9) - V(8));
s += abs(V(10) - V(9));
s += abs(V(11) - V(10));
s += abs(V(12) - V(11));
s += abs(V(13) - V(12));
s += abs(V(14) - V(13));
s += abs(V(15) - V(14));
pix1 += stride;
pix2 += stride;
for (i = 1; i < h; i++) {
s += abs(V(0) - V(-stride));
for (j = 1; j < 16; j++)
s += abs(V(j) - mid_pred(V(j-stride), V(j-1), V(j-stride) + V(j-1) - V(j-stride-1)));
pix1 += stride;
pix2 += stride;
}
#undef V
return s;
}
|
| 82bb3048 |
static int pix_abs16_x2_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| de6d9b64 |
{ |
| 635ec127 |
int s = 0, i; |
| de6d9b64 |
|
| bb12c2e9 |
for (i = 0; i < h; i++) {
s += abs(pix1[0] - avg2(pix2[0], pix2[1]));
s += abs(pix1[1] - avg2(pix2[1], pix2[2]));
s += abs(pix1[2] - avg2(pix2[2], pix2[3]));
s += abs(pix1[3] - avg2(pix2[3], pix2[4]));
s += abs(pix1[4] - avg2(pix2[4], pix2[5]));
s += abs(pix1[5] - avg2(pix2[5], pix2[6]));
s += abs(pix1[6] - avg2(pix2[6], pix2[7]));
s += abs(pix1[7] - avg2(pix2[7], pix2[8]));
s += abs(pix1[8] - avg2(pix2[8], pix2[9]));
s += abs(pix1[9] - avg2(pix2[9], pix2[10]));
s += abs(pix1[10] - avg2(pix2[10], pix2[11]));
s += abs(pix1[11] - avg2(pix2[11], pix2[12]));
s += abs(pix1[12] - avg2(pix2[12], pix2[13]));
s += abs(pix1[13] - avg2(pix2[13], pix2[14]));
s += abs(pix1[14] - avg2(pix2[14], pix2[15]));
s += abs(pix1[15] - avg2(pix2[15], pix2[16])); |
| 9c12c6ff |
pix1 += stride;
pix2 += stride; |
| de6d9b64 |
}
return s;
}
|
| 82bb3048 |
static int pix_abs16_y2_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| de6d9b64 |
{ |
| 635ec127 |
int s = 0, i; |
| 9c12c6ff |
uint8_t *pix3 = pix2 + stride; |
| de6d9b64 |
|
| bb12c2e9 |
for (i = 0; i < h; i++) {
s += abs(pix1[0] - avg2(pix2[0], pix3[0]));
s += abs(pix1[1] - avg2(pix2[1], pix3[1]));
s += abs(pix1[2] - avg2(pix2[2], pix3[2]));
s += abs(pix1[3] - avg2(pix2[3], pix3[3]));
s += abs(pix1[4] - avg2(pix2[4], pix3[4]));
s += abs(pix1[5] - avg2(pix2[5], pix3[5]));
s += abs(pix1[6] - avg2(pix2[6], pix3[6]));
s += abs(pix1[7] - avg2(pix2[7], pix3[7]));
s += abs(pix1[8] - avg2(pix2[8], pix3[8]));
s += abs(pix1[9] - avg2(pix2[9], pix3[9]));
s += abs(pix1[10] - avg2(pix2[10], pix3[10]));
s += abs(pix1[11] - avg2(pix2[11], pix3[11]));
s += abs(pix1[12] - avg2(pix2[12], pix3[12]));
s += abs(pix1[13] - avg2(pix2[13], pix3[13]));
s += abs(pix1[14] - avg2(pix2[14], pix3[14]));
s += abs(pix1[15] - avg2(pix2[15], pix3[15])); |
| 9c12c6ff |
pix1 += stride;
pix2 += stride;
pix3 += stride; |
| de6d9b64 |
}
return s;
}
|
| 82bb3048 |
static int pix_abs16_xy2_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| de6d9b64 |
{ |
| 635ec127 |
int s = 0, i; |
| 9c12c6ff |
uint8_t *pix3 = pix2 + stride; |
| de6d9b64 |
|
| bb12c2e9 |
for (i = 0; i < h; i++) {
s += abs(pix1[0] - avg4(pix2[0], pix2[1], pix3[0], pix3[1]));
s += abs(pix1[1] - avg4(pix2[1], pix2[2], pix3[1], pix3[2]));
s += abs(pix1[2] - avg4(pix2[2], pix2[3], pix3[2], pix3[3]));
s += abs(pix1[3] - avg4(pix2[3], pix2[4], pix3[3], pix3[4]));
s += abs(pix1[4] - avg4(pix2[4], pix2[5], pix3[4], pix3[5]));
s += abs(pix1[5] - avg4(pix2[5], pix2[6], pix3[5], pix3[6]));
s += abs(pix1[6] - avg4(pix2[6], pix2[7], pix3[6], pix3[7]));
s += abs(pix1[7] - avg4(pix2[7], pix2[8], pix3[7], pix3[8]));
s += abs(pix1[8] - avg4(pix2[8], pix2[9], pix3[8], pix3[9]));
s += abs(pix1[9] - avg4(pix2[9], pix2[10], pix3[9], pix3[10]));
s += abs(pix1[10] - avg4(pix2[10], pix2[11], pix3[10], pix3[11]));
s += abs(pix1[11] - avg4(pix2[11], pix2[12], pix3[11], pix3[12]));
s += abs(pix1[12] - avg4(pix2[12], pix2[13], pix3[12], pix3[13]));
s += abs(pix1[13] - avg4(pix2[13], pix2[14], pix3[13], pix3[14]));
s += abs(pix1[14] - avg4(pix2[14], pix2[15], pix3[14], pix3[15]));
s += abs(pix1[15] - avg4(pix2[15], pix2[16], pix3[15], pix3[16])); |
| 9c12c6ff |
pix1 += stride;
pix2 += stride;
pix3 += stride; |
| de6d9b64 |
}
return s;
}
|
| 82bb3048 |
static inline int pix_abs8_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| ba6802de |
{ |
| 635ec127 |
int s = 0, i; |
| ba6802de |
|
| bb12c2e9 |
for (i = 0; i < h; i++) {
s += abs(pix1[0] - pix2[0]);
s += abs(pix1[1] - pix2[1]);
s += abs(pix1[2] - pix2[2]);
s += abs(pix1[3] - pix2[3]);
s += abs(pix1[4] - pix2[4]);
s += abs(pix1[5] - pix2[5]);
s += abs(pix1[6] - pix2[6]);
s += abs(pix1[7] - pix2[7]); |
| 9c12c6ff |
pix1 += stride;
pix2 += stride; |
| ba6802de |
}
return s;
}
|
| 4edd74bd |
static inline int pix_median_abs8_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2,
ptrdiff_t stride, int h)
{
int s = 0, i, j;
#define V(x) (pix1[x] - pix2[x])
s += abs(V(0));
s += abs(V(1) - V(0));
s += abs(V(2) - V(1));
s += abs(V(3) - V(2));
s += abs(V(4) - V(3));
s += abs(V(5) - V(4));
s += abs(V(6) - V(5));
s += abs(V(7) - V(6));
pix1 += stride;
pix2 += stride;
for (i = 1; i < h; i++) {
s += abs(V(0) - V(-stride));
for (j = 1; j < 8; j++)
s += abs(V(j) - mid_pred(V(j-stride), V(j-1), V(j-stride) + V(j-1) - V(j-stride-1)));
pix1 += stride;
pix2 += stride;
}
#undef V
return s;
}
|
| 82bb3048 |
static int pix_abs8_x2_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| ba6802de |
{ |
| 635ec127 |
int s = 0, i; |
| ba6802de |
|
| bb12c2e9 |
for (i = 0; i < h; i++) {
s += abs(pix1[0] - avg2(pix2[0], pix2[1]));
s += abs(pix1[1] - avg2(pix2[1], pix2[2]));
s += abs(pix1[2] - avg2(pix2[2], pix2[3]));
s += abs(pix1[3] - avg2(pix2[3], pix2[4]));
s += abs(pix1[4] - avg2(pix2[4], pix2[5]));
s += abs(pix1[5] - avg2(pix2[5], pix2[6]));
s += abs(pix1[6] - avg2(pix2[6], pix2[7]));
s += abs(pix1[7] - avg2(pix2[7], pix2[8])); |
| 9c12c6ff |
pix1 += stride;
pix2 += stride; |
| ba6802de |
}
return s;
}
|
| 82bb3048 |
static int pix_abs8_y2_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| ba6802de |
{ |
| 635ec127 |
int s = 0, i; |
| 9c12c6ff |
uint8_t *pix3 = pix2 + stride; |
| ba6802de |
|
| bb12c2e9 |
for (i = 0; i < h; i++) {
s += abs(pix1[0] - avg2(pix2[0], pix3[0]));
s += abs(pix1[1] - avg2(pix2[1], pix3[1]));
s += abs(pix1[2] - avg2(pix2[2], pix3[2]));
s += abs(pix1[3] - avg2(pix2[3], pix3[3]));
s += abs(pix1[4] - avg2(pix2[4], pix3[4]));
s += abs(pix1[5] - avg2(pix2[5], pix3[5]));
s += abs(pix1[6] - avg2(pix2[6], pix3[6]));
s += abs(pix1[7] - avg2(pix2[7], pix3[7])); |
| 9c12c6ff |
pix1 += stride;
pix2 += stride;
pix3 += stride; |
| ba6802de |
}
return s;
}
|
| 82bb3048 |
static int pix_abs8_xy2_c(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| ba6802de |
{ |
| 635ec127 |
int s = 0, i; |
| 9c12c6ff |
uint8_t *pix3 = pix2 + stride; |
| ba6802de |
|
| bb12c2e9 |
for (i = 0; i < h; i++) {
s += abs(pix1[0] - avg4(pix2[0], pix2[1], pix3[0], pix3[1]));
s += abs(pix1[1] - avg4(pix2[1], pix2[2], pix3[1], pix3[2]));
s += abs(pix1[2] - avg4(pix2[2], pix2[3], pix3[2], pix3[3]));
s += abs(pix1[3] - avg4(pix2[3], pix2[4], pix3[3], pix3[4]));
s += abs(pix1[4] - avg4(pix2[4], pix2[5], pix3[4], pix3[5]));
s += abs(pix1[5] - avg4(pix2[5], pix2[6], pix3[5], pix3[6]));
s += abs(pix1[6] - avg4(pix2[6], pix2[7], pix3[6], pix3[7]));
s += abs(pix1[7] - avg4(pix2[7], pix2[8], pix3[7], pix3[8])); |
| 9c12c6ff |
pix1 += stride;
pix2 += stride;
pix3 += stride; |
| ba6802de |
}
return s;
}
|
| 9c12c6ff |
static int nsse16_c(MpegEncContext *c, uint8_t *s1, uint8_t *s2,
ptrdiff_t stride, int h) |
| bb12c2e9 |
{ |
| 635ec127 |
int score1 = 0, score2 = 0, x, y; |
| bb12c2e9 |
for (y = 0; y < h; y++) {
for (x = 0; x < 16; x++)
score1 += (s1[x] - s2[x]) * (s1[x] - s2[x]);
if (y + 1 < h) {
for (x = 0; x < 15; x++)
score2 += FFABS(s1[x] - s1[x + stride] -
s1[x + 1] + s1[x + stride + 1]) -
FFABS(s2[x] - s2[x + stride] -
s2[x + 1] + s2[x + stride + 1]); |
| e6a2ac34 |
} |
| bb12c2e9 |
s1 += stride;
s2 += stride; |
| e6a2ac34 |
} |
| d4c5d2ad |
|
| bb12c2e9 |
if (c)
return score1 + FFABS(score2) * c->avctx->nsse_weight;
else
return score1 + FFABS(score2) * 8; |
| e6a2ac34 |
}
|
| 9c12c6ff |
static int nsse8_c(MpegEncContext *c, uint8_t *s1, uint8_t *s2,
ptrdiff_t stride, int h) |
| bb12c2e9 |
{ |
| 635ec127 |
int score1 = 0, score2 = 0, x, y; |
| bb12c2e9 |
for (y = 0; y < h; y++) {
for (x = 0; x < 8; x++)
score1 += (s1[x] - s2[x]) * (s1[x] - s2[x]);
if (y + 1 < h) {
for (x = 0; x < 7; x++)
score2 += FFABS(s1[x] - s1[x + stride] -
s1[x + 1] + s1[x + stride + 1]) -
FFABS(s2[x] - s2[x + stride] -
s2[x + 1] + s2[x + stride + 1]); |
| e6a2ac34 |
} |
| bb12c2e9 |
s1 += stride;
s2 += stride; |
| e6a2ac34 |
} |
| 115329f1 |
|
| bb12c2e9 |
if (c)
return score1 + FFABS(score2) * c->avctx->nsse_weight;
else
return score1 + FFABS(score2) * 8; |
| e6a2ac34 |
}
|
| 82bb3048 |
static int zero_cmp(MpegEncContext *s, uint8_t *a, uint8_t *b, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| bb12c2e9 |
{ |
| 622348f9 |
return 0;
}
|
| 2d604443 |
void ff_set_cmp(MECmpContext *c, me_cmp_func *cmp, int type) |
| bb12c2e9 |
{ |
| 622348f9 |
int i; |
| 115329f1 |
|
| bb12c2e9 |
memset(cmp, 0, sizeof(void *) * 6); |
| 115329f1 |
|
| bb12c2e9 |
for (i = 0; i < 6; i++) {
switch (type & 0xFF) { |
| 622348f9 |
case FF_CMP_SAD: |
| bb12c2e9 |
cmp[i] = c->sad[i]; |
| 622348f9 |
break; |
| 4edd74bd |
case FF_CMP_MEDIAN_SAD:
cmp[i] = c->median_sad[i];
break; |
| 622348f9 |
case FF_CMP_SATD: |
| bb12c2e9 |
cmp[i] = c->hadamard8_diff[i]; |
| 622348f9 |
break;
case FF_CMP_SSE: |
| bb12c2e9 |
cmp[i] = c->sse[i]; |
| 622348f9 |
break;
case FF_CMP_DCT: |
| bb12c2e9 |
cmp[i] = c->dct_sad[i]; |
| 622348f9 |
break; |
| 27c61ac5 |
case FF_CMP_DCT264: |
| bb12c2e9 |
cmp[i] = c->dct264_sad[i]; |
| 27c61ac5 |
break; |
| 0fd6aea1 |
case FF_CMP_DCTMAX: |
| bb12c2e9 |
cmp[i] = c->dct_max[i]; |
| 0fd6aea1 |
break; |
| 622348f9 |
case FF_CMP_PSNR: |
| bb12c2e9 |
cmp[i] = c->quant_psnr[i]; |
| 622348f9 |
break;
case FF_CMP_BIT: |
| bb12c2e9 |
cmp[i] = c->bit[i]; |
| 622348f9 |
break;
case FF_CMP_RD: |
| bb12c2e9 |
cmp[i] = c->rd[i]; |
| 622348f9 |
break;
case FF_CMP_VSAD: |
| bb12c2e9 |
cmp[i] = c->vsad[i]; |
| 622348f9 |
break;
case FF_CMP_VSSE: |
| bb12c2e9 |
cmp[i] = c->vsse[i]; |
| 622348f9 |
break;
case FF_CMP_ZERO: |
| bb12c2e9 |
cmp[i] = zero_cmp; |
| 622348f9 |
break; |
| e6a2ac34 |
case FF_CMP_NSSE: |
| bb12c2e9 |
cmp[i] = c->nsse[i]; |
| e6a2ac34 |
break; |
| 05aec7bb |
#if CONFIG_DWT |
| 26efc54e |
case FF_CMP_W53:
cmp[i]= c->w53[i];
break;
case FF_CMP_W97:
cmp[i]= c->w97[i];
break; |
| 3a6fc8fa |
#endif |
| 622348f9 |
default: |
| bb12c2e9 |
av_log(NULL, AV_LOG_ERROR,
"internal error in cmp function selection\n"); |
| 622348f9 |
}
}
}
|
| bb12c2e9 |
#define BUTTERFLY2(o1, o2, i1, i2) \
o1 = (i1) + (i2); \
o2 = (i1) - (i2); |
| 1457ab52 |
|
| bb12c2e9 |
#define BUTTERFLY1(x, y) \
{ \
int a, b; \
a = x; \
b = y; \
x = a + b; \
y = a - b; \
} |
| 1457ab52 |
|
| bb12c2e9 |
#define BUTTERFLYA(x, y) (FFABS((x) + (y)) + FFABS((x) - (y))) |
| 1457ab52 |
|
| 82bb3048 |
static int hadamard8_diff8x8_c(MpegEncContext *s, uint8_t *dst, |
| 9c12c6ff |
uint8_t *src, ptrdiff_t stride, int h) |
| bb12c2e9 |
{ |
| 635ec127 |
int i, temp[64], sum = 0; |
| 115329f1 |
|
| dd7c2e9b |
av_assert2(h == 8); |
| 115329f1 |
|
| bb12c2e9 |
for (i = 0; i < 8; i++) {
// FIXME: try pointer walks
BUTTERFLY2(temp[8 * i + 0], temp[8 * i + 1],
src[stride * i + 0] - dst[stride * i + 0],
src[stride * i + 1] - dst[stride * i + 1]);
BUTTERFLY2(temp[8 * i + 2], temp[8 * i + 3],
src[stride * i + 2] - dst[stride * i + 2],
src[stride * i + 3] - dst[stride * i + 3]);
BUTTERFLY2(temp[8 * i + 4], temp[8 * i + 5],
src[stride * i + 4] - dst[stride * i + 4],
src[stride * i + 5] - dst[stride * i + 5]);
BUTTERFLY2(temp[8 * i + 6], temp[8 * i + 7],
src[stride * i + 6] - dst[stride * i + 6],
src[stride * i + 7] - dst[stride * i + 7]);
BUTTERFLY1(temp[8 * i + 0], temp[8 * i + 2]);
BUTTERFLY1(temp[8 * i + 1], temp[8 * i + 3]);
BUTTERFLY1(temp[8 * i + 4], temp[8 * i + 6]);
BUTTERFLY1(temp[8 * i + 5], temp[8 * i + 7]);
BUTTERFLY1(temp[8 * i + 0], temp[8 * i + 4]);
BUTTERFLY1(temp[8 * i + 1], temp[8 * i + 5]);
BUTTERFLY1(temp[8 * i + 2], temp[8 * i + 6]);
BUTTERFLY1(temp[8 * i + 3], temp[8 * i + 7]); |
| 1457ab52 |
}
|
| bb12c2e9 |
for (i = 0; i < 8; i++) {
BUTTERFLY1(temp[8 * 0 + i], temp[8 * 1 + i]);
BUTTERFLY1(temp[8 * 2 + i], temp[8 * 3 + i]);
BUTTERFLY1(temp[8 * 4 + i], temp[8 * 5 + i]);
BUTTERFLY1(temp[8 * 6 + i], temp[8 * 7 + i]); |
| 115329f1 |
|
| bb12c2e9 |
BUTTERFLY1(temp[8 * 0 + i], temp[8 * 2 + i]);
BUTTERFLY1(temp[8 * 1 + i], temp[8 * 3 + i]);
BUTTERFLY1(temp[8 * 4 + i], temp[8 * 6 + i]);
BUTTERFLY1(temp[8 * 5 + i], temp[8 * 7 + i]); |
| 1457ab52 |
|
| bb12c2e9 |
sum += BUTTERFLYA(temp[8 * 0 + i], temp[8 * 4 + i]) +
BUTTERFLYA(temp[8 * 1 + i], temp[8 * 5 + i]) +
BUTTERFLYA(temp[8 * 2 + i], temp[8 * 6 + i]) +
BUTTERFLYA(temp[8 * 3 + i], temp[8 * 7 + i]); |
| 1457ab52 |
}
return sum;
}
|
| 82bb3048 |
static int hadamard8_intra8x8_c(MpegEncContext *s, uint8_t *src, |
| 9c12c6ff |
uint8_t *dummy, ptrdiff_t stride, int h) |
| bb12c2e9 |
{ |
| 635ec127 |
int i, temp[64], sum = 0; |
| 115329f1 |
|
| dd7c2e9b |
av_assert2(h == 8); |
| 115329f1 |
|
| bb12c2e9 |
for (i = 0; i < 8; i++) {
// FIXME: try pointer walks
BUTTERFLY2(temp[8 * i + 0], temp[8 * i + 1],
src[stride * i + 0], src[stride * i + 1]);
BUTTERFLY2(temp[8 * i + 2], temp[8 * i + 3],
src[stride * i + 2], src[stride * i + 3]);
BUTTERFLY2(temp[8 * i + 4], temp[8 * i + 5],
src[stride * i + 4], src[stride * i + 5]);
BUTTERFLY2(temp[8 * i + 6], temp[8 * i + 7],
src[stride * i + 6], src[stride * i + 7]);
BUTTERFLY1(temp[8 * i + 0], temp[8 * i + 2]);
BUTTERFLY1(temp[8 * i + 1], temp[8 * i + 3]);
BUTTERFLY1(temp[8 * i + 4], temp[8 * i + 6]);
BUTTERFLY1(temp[8 * i + 5], temp[8 * i + 7]);
BUTTERFLY1(temp[8 * i + 0], temp[8 * i + 4]);
BUTTERFLY1(temp[8 * i + 1], temp[8 * i + 5]);
BUTTERFLY1(temp[8 * i + 2], temp[8 * i + 6]);
BUTTERFLY1(temp[8 * i + 3], temp[8 * i + 7]); |
| 1457ab52 |
}
|
| bb12c2e9 |
for (i = 0; i < 8; i++) {
BUTTERFLY1(temp[8 * 0 + i], temp[8 * 1 + i]);
BUTTERFLY1(temp[8 * 2 + i], temp[8 * 3 + i]);
BUTTERFLY1(temp[8 * 4 + i], temp[8 * 5 + i]);
BUTTERFLY1(temp[8 * 6 + i], temp[8 * 7 + i]); |
| 115329f1 |
|
| bb12c2e9 |
BUTTERFLY1(temp[8 * 0 + i], temp[8 * 2 + i]);
BUTTERFLY1(temp[8 * 1 + i], temp[8 * 3 + i]);
BUTTERFLY1(temp[8 * 4 + i], temp[8 * 6 + i]);
BUTTERFLY1(temp[8 * 5 + i], temp[8 * 7 + i]); |
| 115329f1 |
sum += |
| bb12c2e9 |
BUTTERFLYA(temp[8 * 0 + i], temp[8 * 4 + i])
+ BUTTERFLYA(temp[8 * 1 + i], temp[8 * 5 + i])
+ BUTTERFLYA(temp[8 * 2 + i], temp[8 * 6 + i])
+ BUTTERFLYA(temp[8 * 3 + i], temp[8 * 7 + i]); |
| 1457ab52 |
} |
| 115329f1 |
|
| bb12c2e9 |
sum -= FFABS(temp[8 * 0] + temp[8 * 4]); // -mean |
| 115329f1 |
|
| 1457ab52 |
return sum;
}
|
| 82bb3048 |
static int dct_sad8x8_c(MpegEncContext *s, uint8_t *src1, |
| 9c12c6ff |
uint8_t *src2, ptrdiff_t stride, int h) |
| bb12c2e9 |
{ |
| 88bd7fdc |
LOCAL_ALIGNED_16(int16_t, temp, [64]); |
| 115329f1 |
|
| dd7c2e9b |
av_assert2(h == 8); |
| 1457ab52 |
|
| bc488ec2 |
s->pdsp.diff_pixels_unaligned(temp, src1, src2, stride); |
| a9aee08d |
s->fdsp.fdct(temp); |
| 2d604443 |
return s->mecc.sum_abs_dctelem(temp); |
| 1457ab52 |
}
|
| b250f9c6 |
#if CONFIG_GPL |
| bb12c2e9 |
#define DCT8_1D \
{ \
const int s07 = SRC(0) + SRC(7); \
const int s16 = SRC(1) + SRC(6); \
const int s25 = SRC(2) + SRC(5); \
const int s34 = SRC(3) + SRC(4); \
const int a0 = s07 + s34; \
const int a1 = s16 + s25; \
const int a2 = s07 - s34; \
const int a3 = s16 - s25; \
const int d07 = SRC(0) - SRC(7); \
const int d16 = SRC(1) - SRC(6); \
const int d25 = SRC(2) - SRC(5); \
const int d34 = SRC(3) - SRC(4); \
const int a4 = d16 + d25 + (d07 + (d07 >> 1)); \
const int a5 = d07 - d34 - (d25 + (d25 >> 1)); \
const int a6 = d07 + d34 - (d16 + (d16 >> 1)); \
const int a7 = d16 - d25 + (d34 + (d34 >> 1)); \
DST(0, a0 + a1); \
DST(1, a4 + (a7 >> 2)); \
DST(2, a2 + (a3 >> 1)); \
DST(3, a5 + (a6 >> 2)); \
DST(4, a0 - a1); \
DST(5, a6 - (a5 >> 2)); \
DST(6, (a2 >> 1) - a3); \
DST(7, (a4 >> 2) - a7); \
}
|
| 82bb3048 |
static int dct264_sad8x8_c(MpegEncContext *s, uint8_t *src1, |
| 9c12c6ff |
uint8_t *src2, ptrdiff_t stride, int h) |
| bb12c2e9 |
{ |
| 88bd7fdc |
int16_t dct[8][8]; |
| 635ec127 |
int i, sum = 0; |
| 27c61ac5 |
|
| bc488ec2 |
s->pdsp.diff_pixels_unaligned(dct[0], src1, src2, stride); |
| 27c61ac5 |
#define SRC(x) dct[i][x] |
| bb12c2e9 |
#define DST(x, v) dct[i][x] = v
for (i = 0; i < 8; i++) |
| 27c61ac5 |
DCT8_1D
#undef SRC
#undef DST
#define SRC(x) dct[x][i] |
| bb12c2e9 |
#define DST(x, v) sum += FFABS(v)
for (i = 0; i < 8; i++)
DCT8_1D |
| 27c61ac5 |
#undef SRC
#undef DST |
| bb12c2e9 |
return sum; |
| 27c61ac5 |
}
#endif
|
| 82bb3048 |
static int dct_max8x8_c(MpegEncContext *s, uint8_t *src1, |
| 9c12c6ff |
uint8_t *src2, ptrdiff_t stride, int h) |
| bb12c2e9 |
{ |
| 88bd7fdc |
LOCAL_ALIGNED_16(int16_t, temp, [64]); |
| bb12c2e9 |
int sum = 0, i; |
| 115329f1 |
|
| dd7c2e9b |
av_assert2(h == 8); |
| 0fd6aea1 |
|
| bc488ec2 |
s->pdsp.diff_pixels_unaligned(temp, src1, src2, stride); |
| a9aee08d |
s->fdsp.fdct(temp); |
| 0fd6aea1 |
|
| bb12c2e9 |
for (i = 0; i < 64; i++)
sum = FFMAX(sum, FFABS(temp[i])); |
| 115329f1 |
|
| 0fd6aea1 |
return sum;
}
|
| 82bb3048 |
static int quant_psnr8x8_c(MpegEncContext *s, uint8_t *src1, |
| 9c12c6ff |
uint8_t *src2, ptrdiff_t stride, int h) |
| bb12c2e9 |
{
LOCAL_ALIGNED_16(int16_t, temp, [64 * 2]);
int16_t *const bak = temp + 64;
int sum = 0, i; |
| 1457ab52 |
|
| dd7c2e9b |
av_assert2(h == 8); |
| bb12c2e9 |
s->mb_intra = 0; |
| 115329f1 |
|
| bc488ec2 |
s->pdsp.diff_pixels_unaligned(temp, src1, src2, stride); |
| 115329f1 |
|
| bb12c2e9 |
memcpy(bak, temp, 64 * sizeof(int16_t)); |
| 115329f1 |
|
| bb12c2e9 |
s->block_last_index[0 /* FIXME */] =
s->fast_dct_quantize(s, temp, 0 /* FIXME */, s->qscale, &i); |
| d50635cd |
s->dct_unquantize_inter(s, temp, 0, s->qscale); |
| bb12c2e9 |
ff_simple_idct_8(temp); // FIXME |
| 115329f1 |
|
| bb12c2e9 |
for (i = 0; i < 64; i++)
sum += (temp[i] - bak[i]) * (temp[i] - bak[i]); |
| 115329f1 |
|
| 1457ab52 |
return sum;
}
|
| 82bb3048 |
static int rd8x8_c(MpegEncContext *s, uint8_t *src1, uint8_t *src2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| bb12c2e9 |
{
const uint8_t *scantable = s->intra_scantable.permutated; |
| 88bd7fdc |
LOCAL_ALIGNED_16(int16_t, temp, [64]); |
| 40d11227 |
LOCAL_ALIGNED_16(uint8_t, lsrc1, [64]);
LOCAL_ALIGNED_16(uint8_t, lsrc2, [64]); |
| e6dba5df |
int i, last, run, bits, level, distortion, start_i; |
| bb12c2e9 |
const int esc_length = s->ac_esc_length; |
| 635ec127 |
uint8_t *length, *last_length; |
| 115329f1 |
|
| dd7c2e9b |
av_assert2(h == 8); |
| bb198e19 |
|
| 90d43b52 |
copy_block8(lsrc1, src1, 8, stride, 8);
copy_block8(lsrc2, src2, 8, stride, 8); |
| 3a87ac94 |
|
| f46bb608 |
s->pdsp.diff_pixels(temp, lsrc1, lsrc2, 8); |
| 67725183 |
|
| bb12c2e9 |
s->block_last_index[0 /* FIXME */] =
last =
s->fast_dct_quantize(s, temp, 0 /* FIXME */, s->qscale, &i); |
| 67725183 |
|
| bb12c2e9 |
bits = 0; |
| 115329f1 |
|
| 3a87ac94 |
if (s->mb_intra) { |
| bb12c2e9 |
start_i = 1;
length = s->intra_ac_vlc_length;
last_length = s->intra_ac_vlc_last_length;
bits += s->luma_dc_vlc_length[temp[0] + 256]; // FIXME: chroma |
| 3a87ac94 |
} else { |
| bb12c2e9 |
start_i = 0;
length = s->inter_ac_vlc_length;
last_length = s->inter_ac_vlc_last_length;
}
if (last >= start_i) {
run = 0;
for (i = start_i; i < last; i++) {
int j = scantable[i];
level = temp[j];
if (level) {
level += 64;
if ((level & (~127)) == 0)
bits += length[UNI_AC_ENC_INDEX(run, level)];
else
bits += esc_length;
run = 0;
} else |
| 3a87ac94 |
run++;
} |
| bb12c2e9 |
i = scantable[last]; |
| 115329f1 |
|
| bb12c2e9 |
level = temp[i] + 64; |
| 1d0eab1d |
|
| 72c2d8a3 |
av_assert2(level - 64); |
| 115329f1 |
|
| bb12c2e9 |
if ((level & (~127)) == 0) {
bits += last_length[UNI_AC_ENC_INDEX(run, level)];
} else
bits += esc_length; |
| 67725183 |
}
|
| bb12c2e9 |
if (last >= 0) {
if (s->mb_intra) |
| d50635cd |
s->dct_unquantize_intra(s, temp, 0, s->qscale);
else
s->dct_unquantize_inter(s, temp, 0, s->qscale); |
| 3a87ac94 |
} |
| 115329f1 |
|
| e3fcb143 |
s->idsp.idct_add(lsrc2, 8, temp); |
| 115329f1 |
|
| 2d604443 |
distortion = s->mecc.sse[1](NULL, lsrc2, lsrc1, 8, 8); |
| 3a87ac94 |
|
| bb12c2e9 |
return distortion + ((bits * s->qscale * s->qscale * 109 + 64) >> 7); |
| 3a87ac94 |
}
|
| 82bb3048 |
static int bit8x8_c(MpegEncContext *s, uint8_t *src1, uint8_t *src2, |
| 9c12c6ff |
ptrdiff_t stride, int h) |
| bb12c2e9 |
{
const uint8_t *scantable = s->intra_scantable.permutated; |
| 88bd7fdc |
LOCAL_ALIGNED_16(int16_t, temp, [64]); |
| 3a87ac94 |
int i, last, run, bits, level, start_i; |
| bb12c2e9 |
const int esc_length = s->ac_esc_length; |
| 635ec127 |
uint8_t *length, *last_length; |
| bb198e19 |
|
| dd7c2e9b |
av_assert2(h == 8); |
| 115329f1 |
|
| bc488ec2 |
s->pdsp.diff_pixels_unaligned(temp, src1, src2, stride); |
| 3a87ac94 |
|
| bb12c2e9 |
s->block_last_index[0 /* FIXME */] =
last =
s->fast_dct_quantize(s, temp, 0 /* FIXME */, s->qscale, &i); |
| 67725183 |
|
| bb12c2e9 |
bits = 0; |
| 115329f1 |
|
| 3a87ac94 |
if (s->mb_intra) { |
| bb12c2e9 |
start_i = 1;
length = s->intra_ac_vlc_length;
last_length = s->intra_ac_vlc_last_length;
bits += s->luma_dc_vlc_length[temp[0] + 256]; // FIXME: chroma |
| 3a87ac94 |
} else { |
| bb12c2e9 |
start_i = 0;
length = s->inter_ac_vlc_length;
last_length = s->inter_ac_vlc_last_length;
}
if (last >= start_i) {
run = 0;
for (i = start_i; i < last; i++) {
int j = scantable[i];
level = temp[j];
if (level) {
level += 64;
if ((level & (~127)) == 0)
bits += length[UNI_AC_ENC_INDEX(run, level)];
else
bits += esc_length;
run = 0;
} else |
| 3a87ac94 |
run++;
} |
| bb12c2e9 |
i = scantable[last]; |
| 115329f1 |
|
| bb12c2e9 |
level = temp[i] + 64; |
| 115329f1 |
|
| 72c2d8a3 |
av_assert2(level - 64); |
| 115329f1 |
|
| bb12c2e9 |
if ((level & (~127)) == 0)
bits += last_length[UNI_AC_ENC_INDEX(run, level)];
else
bits += esc_length; |
| 3a87ac94 |
}
return bits;
}
|
| bb12c2e9 |
#define VSAD_INTRA(size) \ |
| 82bb3048 |
static int vsad_intra ## size ## _c(MpegEncContext *c, \ |
| bb12c2e9 |
uint8_t *s, uint8_t *dummy, \ |
| 9c12c6ff |
ptrdiff_t stride, int h) \ |
| bb12c2e9 |
{ \ |
| 635ec127 |
int score = 0, x, y; \ |
| bb12c2e9 |
\
for (y = 1; y < h; y++) { \
for (x = 0; x < size; x += 4) { \
score += FFABS(s[x] - s[x + stride]) + \
FFABS(s[x + 1] - s[x + stride + 1]) + \
FFABS(s[x + 2] - s[x + 2 + stride]) + \
FFABS(s[x + 3] - s[x + 3 + stride]); \
} \
s += stride; \
} \
\
return score; \ |
| 7fb7f636 |
}
VSAD_INTRA(8)
VSAD_INTRA(16) |
| 622348f9 |
|
| ac1077ab |
#define VSAD(size) \
static int vsad ## size ## _c(MpegEncContext *c, \
uint8_t *s1, uint8_t *s2, \ |
| ea41e6d6 |
ptrdiff_t stride, int h) \ |
| ac1077ab |
{ \
int score = 0, x, y; \
\
for (y = 1; y < h; y++) { \
for (x = 0; x < size; x++) \
score += FFABS(s1[x] - s2[x] - s1[x + stride] + s2[x + stride]); \
s1 += stride; \
s2 += stride; \
} \
\
return score; \
}
VSAD(8)
VSAD(16) |
| 622348f9 |
|
| bb12c2e9 |
#define SQ(a) ((a) * (a))
#define VSSE_INTRA(size) \ |
| 82bb3048 |
static int vsse_intra ## size ## _c(MpegEncContext *c, \ |
| bb12c2e9 |
uint8_t *s, uint8_t *dummy, \ |
| 9c12c6ff |
ptrdiff_t stride, int h) \ |
| bb12c2e9 |
{ \ |
| 635ec127 |
int score = 0, x, y; \ |
| bb12c2e9 |
\
for (y = 1; y < h; y++) { \
for (x = 0; x < size; x += 4) { \
score += SQ(s[x] - s[x + stride]) + \
SQ(s[x + 1] - s[x + stride + 1]) + \
SQ(s[x + 2] - s[x + stride + 2]) + \
SQ(s[x + 3] - s[x + stride + 3]); \
} \
s += stride; \
} \
\
return score; \ |
| 7fb7f636 |
}
VSSE_INTRA(8)
VSSE_INTRA(16) |
| 622348f9 |
|
| ac1077ab |
#define VSSE(size) \
static int vsse ## size ## _c(MpegEncContext *c, uint8_t *s1, uint8_t *s2, \ |
| ea41e6d6 |
ptrdiff_t stride, int h) \ |
| ac1077ab |
{ \
int score = 0, x, y; \
\
for (y = 1; y < h; y++) { \
for (x = 0; x < size; x++) \
score += SQ(s1[x] - s2[x] - s1[x + stride] + s2[x + stride]); \
s1 += stride; \
s2 += stride; \
} \
\
return score; \
}
VSSE(8)
VSSE(16) |
| 622348f9 |
|
| bb12c2e9 |
#define WRAPPER8_16_SQ(name8, name16) \ |
| 82bb3048 |
static int name16(MpegEncContext *s, uint8_t *dst, uint8_t *src, \ |
| 9c12c6ff |
ptrdiff_t stride, int h) \ |
| bb12c2e9 |
{ \
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; \ |
| 059866eb |
}
|
| 9fbd14ac |
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) |
| b250f9c6 |
#if CONFIG_GPL |
| 9fbd14ac |
WRAPPER8_16_SQ(dct264_sad8x8_c, dct264_sad16_c) |
| 60900991 |
#endif |
| 9fbd14ac |
WRAPPER8_16_SQ(dct_max8x8_c, dct_max16_c)
WRAPPER8_16_SQ(quant_psnr8x8_c, quant_psnr16_c)
WRAPPER8_16_SQ(rd8x8_c, rd16_c)
WRAPPER8_16_SQ(bit8x8_c, bit16_c) |
| 1457ab52 |
|
| 2d604443 |
av_cold void ff_me_cmp_init_static(void) |
| e0eac44e |
{ |
| d2975f8d |
int i; |
| e0eac44e |
|
| bb12c2e9 |
for (i = 0; i < 512; i++) |
| 05563cca |
ff_square_tab[i] = (i - 256) * (i - 256); |
| 59cf08ce |
} |
| 92ddb692 |
|
| dd7c2e9b |
int ff_check_alignment(void)
{
static int did_fail = 0; |
| 29b42c66 |
LOCAL_ALIGNED_16(int, aligned, [4]); |
| 6dc7d5da |
|
| dd7c2e9b |
if ((intptr_t)aligned & 15) {
if (!did_fail) { |
| b250f9c6 |
#if HAVE_MMX || HAVE_ALTIVEC |
| 6dc7d5da |
av_log(NULL, AV_LOG_ERROR, |
| c1173617 |
"Compiler did not align stack variables. Libavcodec has been miscompiled\n"
"and may be very slow or crash. This is not a bug in libavcodec,\n" |
| 5e4c7ca2 |
"but in the compiler. You may try recompiling using gcc >= 4.2.\n" |
| 2d2b5a14 |
"Do not report crashes to FFmpeg developers.\n"); |
| 6dc7d5da |
#endif
did_fail=1;
}
return -1;
}
return 0;
} |
| 92ddb692 |
|
| 2d604443 |
av_cold void ff_me_cmp_init(MECmpContext *c, AVCodecContext *avctx) |
| 59cf08ce |
{ |
| 6dc7d5da |
ff_check_alignment();
|
| 1edbfe19 |
c->sum_abs_dctelem = sum_abs_dctelem_c; |
| bb12c2e9 |
|
| 45553457 |
/* TODO [0] 16 [1] 8 */ |
| bb198e19 |
c->pix_abs[0][0] = pix_abs16_c;
c->pix_abs[0][1] = pix_abs16_x2_c;
c->pix_abs[0][2] = pix_abs16_y2_c;
c->pix_abs[0][3] = pix_abs16_xy2_c;
c->pix_abs[1][0] = pix_abs8_c;
c->pix_abs[1][1] = pix_abs8_x2_c;
c->pix_abs[1][2] = pix_abs8_y2_c;
c->pix_abs[1][3] = pix_abs8_xy2_c; |
| eb4b3dd3 |
|
| bb12c2e9 |
#define SET_CMP_FUNC(name) \
c->name[0] = name ## 16_c; \
c->name[1] = name ## 8x8_c; |
| 115329f1 |
|
| bb198e19 |
SET_CMP_FUNC(hadamard8_diff) |
| bb12c2e9 |
c->hadamard8_diff[4] = hadamard8_intra16_c;
c->hadamard8_diff[5] = hadamard8_intra8x8_c; |
| bb198e19 |
SET_CMP_FUNC(dct_sad) |
| 0fd6aea1 |
SET_CMP_FUNC(dct_max) |
| b250f9c6 |
#if CONFIG_GPL |
| 27c61ac5 |
SET_CMP_FUNC(dct264_sad) |
| 60900991 |
#endif |
| bb12c2e9 |
c->sad[0] = pix_abs16_c;
c->sad[1] = pix_abs8_c;
c->sse[0] = sse16_c;
c->sse[1] = sse8_c;
c->sse[2] = sse4_c; |
| bb198e19 |
SET_CMP_FUNC(quant_psnr)
SET_CMP_FUNC(rd)
SET_CMP_FUNC(bit) |
| bb12c2e9 |
c->vsad[0] = vsad16_c; |
| ac1077ab |
c->vsad[1] = vsad8_c; |
| bb12c2e9 |
c->vsad[4] = vsad_intra16_c;
c->vsad[5] = vsad_intra8_c;
c->vsse[0] = vsse16_c; |
| ac1077ab |
c->vsse[1] = vsse8_c; |
| bb12c2e9 |
c->vsse[4] = vsse_intra16_c;
c->vsse[5] = vsse_intra8_c;
c->nsse[0] = nsse16_c;
c->nsse[1] = nsse8_c; |
| aeef76b3 |
#if CONFIG_SNOW_DECODER || CONFIG_SNOW_ENCODER |
| 05aec7bb |
ff_dsputil_init_dwt(c); |
| 3a6fc8fa |
#endif |
| 26efc54e |
|
| 67e6a9f5 |
if (ARCH_ALPHA) |
| 3a2d1465 |
ff_me_cmp_init_alpha(c, avctx); |
| 67e6a9f5 |
if (ARCH_ARM) |
| 2d604443 |
ff_me_cmp_init_arm(c, avctx); |
| 67e6a9f5 |
if (ARCH_PPC) |
| 2d604443 |
ff_me_cmp_init_ppc(c, avctx); |
| 67e6a9f5 |
if (ARCH_X86) |
| 2d604443 |
ff_me_cmp_init_x86(c, avctx); |
| 709bb45c |
if (ARCH_MIPS)
ff_me_cmp_init_mips(c, avctx); |
| 4edd74bd |
c->median_sad[0] = pix_median_abs16_c;
c->median_sad[1] = pix_median_abs8_c; |
| 1e3104cd |
} |