2
* Copyright (C) 2004 the ffmpeg project
4
* This file is part of FFmpeg.
6
* FFmpeg is free software; you can redistribute it and/or
7
* modify it under the terms of the GNU Lesser General Public
8
* License as published by the Free Software Foundation; either
9
* version 2.1 of the License, or (at your option) any later version.
11
* FFmpeg is distributed in the hope that it will be useful,
12
* but WITHOUT ANY WARRANTY; without even the implied warranty of
13
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14
* Lesser General Public License for more details.
16
* You should have received a copy of the GNU Lesser General Public
17
* License along with FFmpeg; if not, write to the Free Software
18
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
23
* Standard C DSP-oriented functions cribbed from the original VP3
31
#define IdctAdjustBeforeShift 8
40
#define M(a,b) (((a) * (b))>>16)
42
static av_always_inline void idct(uint8_t *dst, int stride, int16_t *input, int type)
45
uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
47
int A, B, C, D, Ad, Bd, Cd, Dd, E, F, G, H;
48
int Ed, Gd, Add, Bdd, Fd, Hd;
52
/* Inverse DCT on the rows now */
53
for (i = 0; i < 8; i++) {
54
/* Check for non-zero values */
55
if ( ip[0] | ip[1] | ip[2] | ip[3] | ip[4] | ip[5] | ip[6] | ip[7] ) {
56
A = M(xC1S7, ip[1]) + M(xC7S1, ip[7]);
57
B = M(xC7S1, ip[1]) - M(xC1S7, ip[7]);
58
C = M(xC3S5, ip[3]) + M(xC5S3, ip[5]);
59
D = M(xC3S5, ip[5]) - M(xC5S3, ip[3]);
61
Ad = M(xC4S4, (A - C));
62
Bd = M(xC4S4, (B - D));
67
E = M(xC4S4, (ip[0] + ip[4]));
68
F = M(xC4S4, (ip[0] - ip[4]));
70
G = M(xC2S6, ip[2]) + M(xC6S2, ip[6]);
71
H = M(xC6S2, ip[2]) - M(xC2S6, ip[6]);
82
/* Final sequence of operations over-write original inputs. */
96
ip += 8; /* next row */
101
for ( i = 0; i < 8; i++) {
102
/* Check for non-zero values (bitwise or faster than ||) */
103
if ( ip[1 * 8] | ip[2 * 8] | ip[3 * 8] |
104
ip[4 * 8] | ip[5 * 8] | ip[6 * 8] | ip[7 * 8] ) {
106
A = M(xC1S7, ip[1*8]) + M(xC7S1, ip[7*8]);
107
B = M(xC7S1, ip[1*8]) - M(xC1S7, ip[7*8]);
108
C = M(xC3S5, ip[3*8]) + M(xC5S3, ip[5*8]);
109
D = M(xC3S5, ip[5*8]) - M(xC5S3, ip[3*8]);
111
Ad = M(xC4S4, (A - C));
112
Bd = M(xC4S4, (B - D));
117
E = M(xC4S4, (ip[0*8] + ip[4*8])) + 8;
118
F = M(xC4S4, (ip[0*8] - ip[4*8])) + 8;
125
G = M(xC2S6, ip[2*8]) + M(xC6S2, ip[6*8]);
126
H = M(xC6S2, ip[2*8]) - M(xC2S6, ip[6*8]);
137
/* Final sequence of operations over-write original inputs. */
139
ip[0*8] = (Gd + Cd ) >> 4;
140
ip[7*8] = (Gd - Cd ) >> 4;
142
ip[1*8] = (Add + Hd ) >> 4;
143
ip[2*8] = (Add - Hd ) >> 4;
145
ip[3*8] = (Ed + Dd ) >> 4;
146
ip[4*8] = (Ed - Dd ) >> 4;
148
ip[5*8] = (Fd + Bdd ) >> 4;
149
ip[6*8] = (Fd - Bdd ) >> 4;
151
dst[0*stride] = cm[(Gd + Cd ) >> 4];
152
dst[7*stride] = cm[(Gd - Cd ) >> 4];
154
dst[1*stride] = cm[(Add + Hd ) >> 4];
155
dst[2*stride] = cm[(Add - Hd ) >> 4];
157
dst[3*stride] = cm[(Ed + Dd ) >> 4];
158
dst[4*stride] = cm[(Ed - Dd ) >> 4];
160
dst[5*stride] = cm[(Fd + Bdd ) >> 4];
161
dst[6*stride] = cm[(Fd - Bdd ) >> 4];
163
dst[0*stride] = cm[dst[0*stride] + ((Gd + Cd ) >> 4)];
164
dst[7*stride] = cm[dst[7*stride] + ((Gd - Cd ) >> 4)];
166
dst[1*stride] = cm[dst[1*stride] + ((Add + Hd ) >> 4)];
167
dst[2*stride] = cm[dst[2*stride] + ((Add - Hd ) >> 4)];
169
dst[3*stride] = cm[dst[3*stride] + ((Ed + Dd ) >> 4)];
170
dst[4*stride] = cm[dst[4*stride] + ((Ed - Dd ) >> 4)];
172
dst[5*stride] = cm[dst[5*stride] + ((Fd + Bdd ) >> 4)];
173
dst[6*stride] = cm[dst[6*stride] + ((Fd - Bdd ) >> 4)];
185
ip[7*8] = ((xC4S4 * ip[0*8] + (IdctAdjustBeforeShift<<16))>>20);
194
dst[7*stride]= 128 + ((xC4S4 * ip[0*8] + (IdctAdjustBeforeShift<<16))>>20);
197
int v= ((xC4S4 * ip[0*8] + (IdctAdjustBeforeShift<<16))>>20);
198
dst[0*stride] = cm[dst[0*stride] + v];
199
dst[1*stride] = cm[dst[1*stride] + v];
200
dst[2*stride] = cm[dst[2*stride] + v];
201
dst[3*stride] = cm[dst[3*stride] + v];
202
dst[4*stride] = cm[dst[4*stride] + v];
203
dst[5*stride] = cm[dst[5*stride] + v];
204
dst[6*stride] = cm[dst[6*stride] + v];
205
dst[7*stride] = cm[dst[7*stride] + v];
210
ip++; /* next column */
215
void ff_vp3_idct_c(DCTELEM *block/* align 16*/){
216
idct(NULL, 0, block, 0);
219
void ff_vp3_idct_put_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/){
220
idct(dest, line_size, block, 1);
223
void ff_vp3_idct_add_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/){
224
idct(dest, line_size, block, 2);