2004-03-08 04:13:45 +02:00
|
|
|
/*
|
|
|
|
* Copyright (C) 2004 the ffmpeg project
|
|
|
|
*
|
2006-10-07 18:30:46 +03:00
|
|
|
* This file is part of FFmpeg.
|
|
|
|
*
|
|
|
|
* FFmpeg is free software; you can redistribute it and/or
|
2004-03-08 04:13:45 +02:00
|
|
|
* modify it under the terms of the GNU Lesser General Public
|
|
|
|
* License as published by the Free Software Foundation; either
|
2006-10-07 18:30:46 +03:00
|
|
|
* version 2.1 of the License, or (at your option) any later version.
|
2004-03-08 04:13:45 +02:00
|
|
|
*
|
2006-10-07 18:30:46 +03:00
|
|
|
* FFmpeg is distributed in the hope that it will be useful,
|
2004-03-08 04:13:45 +02:00
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
* Lesser General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU Lesser General Public
|
2006-10-07 18:30:46 +03:00
|
|
|
* License along with FFmpeg; if not, write to the Free Software
|
2006-01-13 00:43:26 +02:00
|
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
2004-03-08 04:13:45 +02:00
|
|
|
*/
|
|
|
|
|
|
|
|
/**
|
2010-04-20 17:45:34 +03:00
|
|
|
* @file
|
2005-12-17 20:14:38 +02:00
|
|
|
* Standard C DSP-oriented functions cribbed from the original VP3
|
2004-03-08 04:13:45 +02:00
|
|
|
* source code.
|
|
|
|
*/
|
|
|
|
|
2012-07-17 18:47:43 +03:00
|
|
|
#include "libavutil/attributes.h"
|
2012-08-06 16:49:32 +03:00
|
|
|
#include "libavutil/common.h"
|
2014-02-10 20:58:28 +03:00
|
|
|
#include "libavutil/intreadwrite.h"
|
2014-03-31 19:12:58 +03:00
|
|
|
|
2004-03-08 04:13:45 +02:00
|
|
|
#include "avcodec.h"
|
2013-02-08 00:37:00 +03:00
|
|
|
#include "rnd_avg.h"
|
2012-07-17 18:47:43 +03:00
|
|
|
#include "vp3dsp.h"
|
2004-03-08 04:13:45 +02:00
|
|
|
|
|
|
|
#define IdctAdjustBeforeShift 8
|
|
|
|
#define xC1S7 64277
|
|
|
|
#define xC2S6 60547
|
|
|
|
#define xC3S5 54491
|
|
|
|
#define xC4S4 46341
|
|
|
|
#define xC5S3 36410
|
|
|
|
#define xC6S2 25080
|
|
|
|
#define xC7S1 12785
|
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
#define M(a, b) (((a) * (b)) >> 16)
|
2006-07-20 02:17:18 +03:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
static av_always_inline void idct(uint8_t *dst, int stride,
|
|
|
|
int16_t *input, int type)
|
2004-03-08 04:13:45 +02:00
|
|
|
{
|
2005-05-17 14:12:04 +03:00
|
|
|
int16_t *ip = input;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2006-07-20 02:21:20 +03:00
|
|
|
int A, B, C, D, Ad, Bd, Cd, Dd, E, F, G, H;
|
|
|
|
int Ed, Gd, Add, Bdd, Fd, Hd;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2005-08-14 18:42:40 +03:00
|
|
|
int i;
|
2005-12-17 20:14:38 +02:00
|
|
|
|
2004-03-08 04:13:45 +02:00
|
|
|
/* Inverse DCT on the rows now */
|
|
|
|
for (i = 0; i < 8; i++) {
|
|
|
|
/* Check for non-zero values */
|
2014-03-31 19:12:58 +03:00
|
|
|
if (ip[0 * 8] | ip[1 * 8] | ip[2 * 8] | ip[3 * 8] |
|
|
|
|
ip[4 * 8] | ip[5 * 8] | ip[6 * 8] | ip[7 * 8]) {
|
2013-03-12 17:28:12 +03:00
|
|
|
A = M(xC1S7, ip[1 * 8]) + M(xC7S1, ip[7 * 8]);
|
|
|
|
B = M(xC7S1, ip[1 * 8]) - M(xC1S7, ip[7 * 8]);
|
|
|
|
C = M(xC3S5, ip[3 * 8]) + M(xC5S3, ip[5 * 8]);
|
|
|
|
D = M(xC3S5, ip[5 * 8]) - M(xC5S3, ip[3 * 8]);
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2006-07-20 02:21:20 +03:00
|
|
|
Ad = M(xC4S4, (A - C));
|
|
|
|
Bd = M(xC4S4, (B - D));
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2006-07-20 02:21:20 +03:00
|
|
|
Cd = A + C;
|
|
|
|
Dd = B + D;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2013-03-12 17:28:12 +03:00
|
|
|
E = M(xC4S4, (ip[0 * 8] + ip[4 * 8]));
|
|
|
|
F = M(xC4S4, (ip[0 * 8] - ip[4 * 8]));
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2013-03-12 17:28:12 +03:00
|
|
|
G = M(xC2S6, ip[2 * 8]) + M(xC6S2, ip[6 * 8]);
|
|
|
|
H = M(xC6S2, ip[2 * 8]) - M(xC2S6, ip[6 * 8]);
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2006-07-20 02:21:20 +03:00
|
|
|
Ed = E - G;
|
|
|
|
Gd = E + G;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2006-07-20 02:21:20 +03:00
|
|
|
Add = F + Ad;
|
|
|
|
Bdd = Bd - H;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2006-07-20 02:21:20 +03:00
|
|
|
Fd = F - Ad;
|
|
|
|
Hd = Bd + H;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
|
|
|
/* Final sequence of operations over-write original inputs. */
|
2014-03-31 19:12:58 +03:00
|
|
|
ip[0 * 8] = Gd + Cd;
|
|
|
|
ip[7 * 8] = Gd - Cd;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2013-03-12 17:28:12 +03:00
|
|
|
ip[1 * 8] = Add + Hd;
|
|
|
|
ip[2 * 8] = Add - Hd;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
ip[3 * 8] = Ed + Dd;
|
|
|
|
ip[4 * 8] = Ed - Dd;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2013-03-12 17:28:12 +03:00
|
|
|
ip[5 * 8] = Fd + Bdd;
|
|
|
|
ip[6 * 8] = Fd - Bdd;
|
2004-03-08 04:13:45 +02:00
|
|
|
}
|
|
|
|
|
2013-03-12 17:28:12 +03:00
|
|
|
ip += 1; /* next row */
|
2004-03-08 04:13:45 +02:00
|
|
|
}
|
2005-12-17 20:14:38 +02:00
|
|
|
|
2005-05-17 14:12:04 +03:00
|
|
|
ip = input;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
for (i = 0; i < 8; i++) {
|
2004-03-08 04:13:45 +02:00
|
|
|
/* Check for non-zero values (bitwise or faster than ||) */
|
2014-03-31 19:12:58 +03:00
|
|
|
if (ip[1] | ip[2] | ip[3] |
|
|
|
|
ip[4] | ip[5] | ip[6] | ip[7]) {
|
2013-03-12 17:28:12 +03:00
|
|
|
A = M(xC1S7, ip[1]) + M(xC7S1, ip[7]);
|
|
|
|
B = M(xC7S1, ip[1]) - M(xC1S7, ip[7]);
|
|
|
|
C = M(xC3S5, ip[3]) + M(xC5S3, ip[5]);
|
|
|
|
D = M(xC3S5, ip[5]) - M(xC5S3, ip[3]);
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2006-07-20 02:21:20 +03:00
|
|
|
Ad = M(xC4S4, (A - C));
|
|
|
|
Bd = M(xC4S4, (B - D));
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2006-07-20 02:21:20 +03:00
|
|
|
Cd = A + C;
|
|
|
|
Dd = B + D;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2013-03-12 17:28:12 +03:00
|
|
|
E = M(xC4S4, (ip[0] + ip[4])) + 8;
|
|
|
|
F = M(xC4S4, (ip[0] - ip[4])) + 8;
|
2006-07-20 02:51:11 +03:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
if (type == 1) { // HACK
|
|
|
|
E += 16 * 128;
|
|
|
|
F += 16 * 128;
|
2006-07-20 02:51:11 +03:00
|
|
|
}
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2013-03-12 17:28:12 +03:00
|
|
|
G = M(xC2S6, ip[2]) + M(xC6S2, ip[6]);
|
|
|
|
H = M(xC6S2, ip[2]) - M(xC2S6, ip[6]);
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2006-07-20 02:21:20 +03:00
|
|
|
Ed = E - G;
|
|
|
|
Gd = E + G;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2006-07-20 02:21:20 +03:00
|
|
|
Add = F + Ad;
|
|
|
|
Bdd = Bd - H;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
2006-07-20 02:21:20 +03:00
|
|
|
Fd = F - Ad;
|
|
|
|
Hd = Bd + H;
|
2004-03-08 04:13:45 +02:00
|
|
|
|
|
|
|
/* Final sequence of operations over-write original inputs. */
|
2013-03-12 17:28:12 +03:00
|
|
|
if (type == 1) {
|
2014-03-31 19:12:58 +03:00
|
|
|
dst[0 * stride] = av_clip_uint8((Gd + Cd) >> 4);
|
|
|
|
dst[7 * stride] = av_clip_uint8((Gd - Cd) >> 4);
|
2005-12-17 20:14:38 +02:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
dst[1 * stride] = av_clip_uint8((Add + Hd) >> 4);
|
|
|
|
dst[2 * stride] = av_clip_uint8((Add - Hd) >> 4);
|
2005-12-17 20:14:38 +02:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
dst[3 * stride] = av_clip_uint8((Ed + Dd) >> 4);
|
|
|
|
dst[4 * stride] = av_clip_uint8((Ed - Dd) >> 4);
|
2005-12-17 20:14:38 +02:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
dst[5 * stride] = av_clip_uint8((Fd + Bdd) >> 4);
|
|
|
|
dst[6 * stride] = av_clip_uint8((Fd - Bdd) >> 4);
|
|
|
|
} else {
|
|
|
|
dst[0 * stride] = av_clip_uint8(dst[0 * stride] + ((Gd + Cd) >> 4));
|
|
|
|
dst[7 * stride] = av_clip_uint8(dst[7 * stride] + ((Gd - Cd) >> 4));
|
2005-12-17 20:14:38 +02:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
dst[1 * stride] = av_clip_uint8(dst[1 * stride] + ((Add + Hd) >> 4));
|
|
|
|
dst[2 * stride] = av_clip_uint8(dst[2 * stride] + ((Add - Hd) >> 4));
|
2005-12-17 20:14:38 +02:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
dst[3 * stride] = av_clip_uint8(dst[3 * stride] + ((Ed + Dd) >> 4));
|
|
|
|
dst[4 * stride] = av_clip_uint8(dst[4 * stride] + ((Ed - Dd) >> 4));
|
2005-12-17 20:14:38 +02:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
dst[5 * stride] = av_clip_uint8(dst[5 * stride] + ((Fd + Bdd) >> 4));
|
|
|
|
dst[6 * stride] = av_clip_uint8(dst[6 * stride] + ((Fd - Bdd) >> 4));
|
2005-05-17 14:12:04 +03:00
|
|
|
}
|
2004-03-08 04:13:45 +02:00
|
|
|
} else {
|
2013-03-12 17:28:12 +03:00
|
|
|
if (type == 1) {
|
2014-03-31 19:12:58 +03:00
|
|
|
dst[0*stride] =
|
|
|
|
dst[1*stride] =
|
|
|
|
dst[2*stride] =
|
|
|
|
dst[3*stride] =
|
|
|
|
dst[4*stride] =
|
|
|
|
dst[5*stride] =
|
|
|
|
dst[6*stride] =
|
|
|
|
dst[7*stride] = av_clip_uint8(128 + ((xC4S4 * ip[0] + (IdctAdjustBeforeShift << 16)) >> 20));
|
|
|
|
} else {
|
|
|
|
if (ip[0]) {
|
|
|
|
int v = (xC4S4 * ip[0] + (IdctAdjustBeforeShift << 16)) >> 20;
|
|
|
|
dst[0 * stride] = av_clip_uint8(dst[0 * stride] + v);
|
|
|
|
dst[1 * stride] = av_clip_uint8(dst[1 * stride] + v);
|
|
|
|
dst[2 * stride] = av_clip_uint8(dst[2 * stride] + v);
|
|
|
|
dst[3 * stride] = av_clip_uint8(dst[3 * stride] + v);
|
|
|
|
dst[4 * stride] = av_clip_uint8(dst[4 * stride] + v);
|
|
|
|
dst[5 * stride] = av_clip_uint8(dst[5 * stride] + v);
|
|
|
|
dst[6 * stride] = av_clip_uint8(dst[6 * stride] + v);
|
|
|
|
dst[7 * stride] = av_clip_uint8(dst[7 * stride] + v);
|
2005-05-17 14:12:04 +03:00
|
|
|
}
|
|
|
|
}
|
2004-03-08 04:13:45 +02:00
|
|
|
}
|
|
|
|
|
2013-03-12 17:28:12 +03:00
|
|
|
ip += 8; /* next column */
|
2005-05-17 14:12:04 +03:00
|
|
|
dst++;
|
2004-03-08 04:13:45 +02:00
|
|
|
}
|
|
|
|
}
|
2005-05-17 14:12:04 +03:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
static void vp3_idct_put_c(uint8_t *dest /* align 8 */, int line_size,
|
|
|
|
int16_t *block /* align 16 */)
|
2013-01-20 03:02:29 +03:00
|
|
|
{
|
2005-05-17 14:12:04 +03:00
|
|
|
idct(dest, line_size, block, 1);
|
2013-01-18 18:43:04 +03:00
|
|
|
memset(block, 0, sizeof(*block) * 64);
|
2005-05-17 14:12:04 +03:00
|
|
|
}
|
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
static void vp3_idct_add_c(uint8_t *dest /* align 8 */, int line_size,
|
|
|
|
int16_t *block /* align 16 */)
|
2013-01-20 03:02:29 +03:00
|
|
|
{
|
2005-05-17 14:12:04 +03:00
|
|
|
idct(dest, line_size, block, 2);
|
2013-01-18 18:43:04 +03:00
|
|
|
memset(block, 0, sizeof(*block) * 64);
|
2005-05-17 14:12:04 +03:00
|
|
|
}
|
2008-10-04 13:26:17 +03:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
static void vp3_idct_dc_add_c(uint8_t *dest /* align 8 */, int line_size,
|
|
|
|
int16_t *block /* align 16 */)
|
2013-01-20 03:02:29 +03:00
|
|
|
{
|
2010-05-28 10:22:04 +03:00
|
|
|
int i, dc = (block[0] + 15) >> 5;
|
2010-04-17 05:04:30 +03:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
for (i = 0; i < 8; i++) {
|
2012-03-06 03:01:19 +03:00
|
|
|
dest[0] = av_clip_uint8(dest[0] + dc);
|
|
|
|
dest[1] = av_clip_uint8(dest[1] + dc);
|
|
|
|
dest[2] = av_clip_uint8(dest[2] + dc);
|
|
|
|
dest[3] = av_clip_uint8(dest[3] + dc);
|
|
|
|
dest[4] = av_clip_uint8(dest[4] + dc);
|
|
|
|
dest[5] = av_clip_uint8(dest[5] + dc);
|
|
|
|
dest[6] = av_clip_uint8(dest[6] + dc);
|
|
|
|
dest[7] = av_clip_uint8(dest[7] + dc);
|
2014-03-31 19:12:58 +03:00
|
|
|
dest += line_size;
|
2010-04-17 05:04:30 +03:00
|
|
|
}
|
2013-01-18 18:43:04 +03:00
|
|
|
block[0] = 0;
|
2010-04-17 05:04:30 +03:00
|
|
|
}
|
|
|
|
|
2012-07-17 18:47:43 +03:00
|
|
|
static void vp3_v_loop_filter_c(uint8_t *first_pixel, int stride,
|
|
|
|
int *bounding_values)
|
2008-10-04 13:26:17 +03:00
|
|
|
{
|
|
|
|
unsigned char *end;
|
|
|
|
int filter_value;
|
2014-03-31 19:12:58 +03:00
|
|
|
const int nstride = -stride;
|
2008-10-04 13:26:17 +03:00
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
for (end = first_pixel + 8; first_pixel < end; first_pixel++) {
|
|
|
|
filter_value = (first_pixel[2 * nstride] - first_pixel[stride]) +
|
|
|
|
(first_pixel[0] - first_pixel[nstride]) * 3;
|
2008-10-04 13:26:17 +03:00
|
|
|
filter_value = bounding_values[(filter_value + 4) >> 3];
|
2014-03-31 19:12:58 +03:00
|
|
|
|
2008-10-04 13:26:17 +03:00
|
|
|
first_pixel[nstride] = av_clip_uint8(first_pixel[nstride] + filter_value);
|
2014-03-31 19:12:58 +03:00
|
|
|
first_pixel[0] = av_clip_uint8(first_pixel[0] - filter_value);
|
2008-10-04 13:26:17 +03:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-07-17 18:47:43 +03:00
|
|
|
static void vp3_h_loop_filter_c(uint8_t *first_pixel, int stride,
|
|
|
|
int *bounding_values)
|
2008-10-04 13:26:17 +03:00
|
|
|
{
|
|
|
|
unsigned char *end;
|
|
|
|
int filter_value;
|
|
|
|
|
2014-03-31 19:12:58 +03:00
|
|
|
for (end = first_pixel + 8 * stride; first_pixel != end; first_pixel += stride) {
|
|
|
|
filter_value = (first_pixel[-2] - first_pixel[1]) +
|
|
|
|
(first_pixel[ 0] - first_pixel[-1]) * 3;
|
2008-10-04 13:26:17 +03:00
|
|
|
filter_value = bounding_values[(filter_value + 4) >> 3];
|
2014-03-31 19:12:58 +03:00
|
|
|
|
2008-10-04 13:26:17 +03:00
|
|
|
first_pixel[-1] = av_clip_uint8(first_pixel[-1] + filter_value);
|
|
|
|
first_pixel[ 0] = av_clip_uint8(first_pixel[ 0] - filter_value);
|
|
|
|
}
|
|
|
|
}
|
2012-07-17 18:47:43 +03:00
|
|
|
|
2013-01-21 22:02:33 +03:00
|
|
|
static void put_no_rnd_pixels_l2(uint8_t *dst, const uint8_t *src1,
|
|
|
|
const uint8_t *src2, ptrdiff_t stride, int h)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < h; i++) {
|
|
|
|
uint32_t a, b;
|
|
|
|
|
2013-01-22 23:45:35 +03:00
|
|
|
a = AV_RN32(&src1[i * stride]);
|
|
|
|
b = AV_RN32(&src2[i * stride]);
|
2013-01-21 22:02:33 +03:00
|
|
|
AV_WN32A(&dst[i * stride], no_rnd_avg32(a, b));
|
2013-01-22 23:45:35 +03:00
|
|
|
a = AV_RN32(&src1[i * stride + 4]);
|
|
|
|
b = AV_RN32(&src2[i * stride + 4]);
|
2013-01-21 22:02:33 +03:00
|
|
|
AV_WN32A(&dst[i * stride + 4], no_rnd_avg32(a, b));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-07-17 18:47:43 +03:00
|
|
|
av_cold void ff_vp3dsp_init(VP3DSPContext *c, int flags)
|
|
|
|
{
|
2013-01-21 22:02:33 +03:00
|
|
|
c->put_no_rnd_pixels_l2 = put_no_rnd_pixels_l2;
|
|
|
|
|
2012-07-17 18:47:43 +03:00
|
|
|
c->idct_put = vp3_idct_put_c;
|
|
|
|
c->idct_add = vp3_idct_add_c;
|
|
|
|
c->idct_dc_add = vp3_idct_dc_add_c;
|
|
|
|
c->v_loop_filter = vp3_v_loop_filter_c;
|
|
|
|
c->h_loop_filter = vp3_h_loop_filter_c;
|
|
|
|
|
|
|
|
if (ARCH_ARM)
|
|
|
|
ff_vp3dsp_init_arm(c, flags);
|
|
|
|
if (ARCH_PPC)
|
|
|
|
ff_vp3dsp_init_ppc(c, flags);
|
|
|
|
if (ARCH_X86)
|
|
|
|
ff_vp3dsp_init_x86(c, flags);
|
|
|
|
}
|