Statistics
| Branch: | Revision:

ffmpeg / libavcodec / h264idct.c @ 5509bffa

History | View | Annotate | Download (5.27 KB)

1 0fa8158d Michael Niedermayer
/*
2
 * H.264 IDCT
3
 * Copyright (c) 2004 Michael Niedermayer <michaelni@gmx.at>
4
 *
5
 * This library is free software; you can redistribute it and/or
6
 * modify it under the terms of the GNU Lesser General Public
7
 * License as published by the Free Software Foundation; either
8
 * version 2 of the License, or (at your option) any later version.
9
 *
10
 * This library is distributed in the hope that it will be useful,
11
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
13
 * Lesser General Public License for more details.
14
 *
15
 * You should have received a copy of the GNU Lesser General Public
16
 * License along with this library; if not, write to the Free Software
17 5509bffa Diego Biurrun
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
18 0fa8158d Michael Niedermayer
 *
19
 */
20 115329f1 Diego Biurrun
21 0fa8158d Michael Niedermayer
/**
22
 * @file h264-idct.c
23
 * H.264 IDCT.
24
 * @author Michael Niedermayer <michaelni@gmx.at>
25
 */
26 115329f1 Diego Biurrun
27 0fa8158d Michael Niedermayer
#include "dsputil.h"
28
29
static always_inline void idct_internal(uint8_t *dst, DCTELEM *block, int stride, int block_stride, int shift, int add){
30
    int i;
31
    uint8_t *cm = cropTbl + MAX_NEG_CROP;
32
33
    block[0] += 1<<(shift-1);
34
35
    for(i=0; i<4; i++){
36
        const int z0=  block[0 + block_stride*i]     +  block[2 + block_stride*i];
37
        const int z1=  block[0 + block_stride*i]     -  block[2 + block_stride*i];
38
        const int z2= (block[1 + block_stride*i]>>1) -  block[3 + block_stride*i];
39
        const int z3=  block[1 + block_stride*i]     + (block[3 + block_stride*i]>>1);
40
41
        block[0 + block_stride*i]= z0 + z3;
42
        block[1 + block_stride*i]= z1 + z2;
43
        block[2 + block_stride*i]= z1 - z2;
44
        block[3 + block_stride*i]= z0 - z3;
45
    }
46
47
    for(i=0; i<4; i++){
48
        const int z0=  block[i + block_stride*0]     +  block[i + block_stride*2];
49
        const int z1=  block[i + block_stride*0]     -  block[i + block_stride*2];
50
        const int z2= (block[i + block_stride*1]>>1) -  block[i + block_stride*3];
51
        const int z3=  block[i + block_stride*1]     + (block[i + block_stride*3]>>1);
52
53
        dst[i + 0*stride]= cm[ add*dst[i + 0*stride] + ((z0 + z3) >> shift) ];
54
        dst[i + 1*stride]= cm[ add*dst[i + 1*stride] + ((z1 + z2) >> shift) ];
55
        dst[i + 2*stride]= cm[ add*dst[i + 2*stride] + ((z1 - z2) >> shift) ];
56
        dst[i + 3*stride]= cm[ add*dst[i + 3*stride] + ((z0 - z3) >> shift) ];
57
    }
58
}
59
60
void ff_h264_idct_add_c(uint8_t *dst, DCTELEM *block, int stride){
61
    idct_internal(dst, block, stride, 4, 6, 1);
62
}
63
64
void ff_h264_lowres_idct_add_c(uint8_t *dst, int stride, DCTELEM *block){
65
    idct_internal(dst, block, stride, 8, 3, 1);
66
}
67
68
void ff_h264_lowres_idct_put_c(uint8_t *dst, int stride, DCTELEM *block){
69
    idct_internal(dst, block, stride, 8, 3, 0);
70
}
71 43efd19a Loren Merritt
72
void ff_h264_idct8_add_c(uint8_t *dst, DCTELEM *block, int stride){
73
    int i;
74
    DCTELEM (*src)[8] = (DCTELEM(*)[8])block;
75
    uint8_t *cm = cropTbl + MAX_NEG_CROP;
76
77
    block[0] += 32;
78
79
    for( i = 0; i < 8; i++ )
80
    {
81
        const int a0 =  src[i][0] + src[i][4];
82
        const int a2 =  src[i][0] - src[i][4];
83
        const int a4 = (src[i][2]>>1) - src[i][6];
84
        const int a6 = (src[i][6]>>1) + src[i][2];
85
86
        const int b0 = a0 + a6;
87
        const int b2 = a2 + a4;
88
        const int b4 = a2 - a4;
89
        const int b6 = a0 - a6;
90
91
        const int a1 = -src[i][3] + src[i][5] - src[i][7] - (src[i][7]>>1);
92
        const int a3 =  src[i][1] + src[i][7] - src[i][3] - (src[i][3]>>1);
93
        const int a5 = -src[i][1] + src[i][7] + src[i][5] + (src[i][5]>>1);
94
        const int a7 =  src[i][3] + src[i][5] + src[i][1] + (src[i][1]>>1);
95
96
        const int b1 = (a7>>2) + a1;
97
        const int b3 =  a3 + (a5>>2);
98
        const int b5 = (a3>>2) - a5;
99
        const int b7 =  a7 - (a1>>2);
100
101
        src[i][0] = b0 + b7;
102
        src[i][7] = b0 - b7;
103
        src[i][1] = b2 + b5;
104
        src[i][6] = b2 - b5;
105
        src[i][2] = b4 + b3;
106
        src[i][5] = b4 - b3;
107
        src[i][3] = b6 + b1;
108
        src[i][4] = b6 - b1;
109
    }
110
    for( i = 0; i < 8; i++ )
111
    {
112
        const int a0 =  src[0][i] + src[4][i];
113
        const int a2 =  src[0][i] - src[4][i];
114
        const int a4 = (src[2][i]>>1) - src[6][i];
115
        const int a6 = (src[6][i]>>1) + src[2][i];
116
117
        const int b0 = a0 + a6;
118
        const int b2 = a2 + a4;
119
        const int b4 = a2 - a4;
120
        const int b6 = a0 - a6;
121
122
        const int a1 = -src[3][i] + src[5][i] - src[7][i] - (src[7][i]>>1);
123
        const int a3 =  src[1][i] + src[7][i] - src[3][i] - (src[3][i]>>1);
124
        const int a5 = -src[1][i] + src[7][i] + src[5][i] + (src[5][i]>>1);
125
        const int a7 =  src[3][i] + src[5][i] + src[1][i] + (src[1][i]>>1);
126
127
        const int b1 = (a7>>2) + a1;
128
        const int b3 =  a3 + (a5>>2);
129
        const int b5 = (a3>>2) - a5;
130
        const int b7 =  a7 - (a1>>2);
131
132
        dst[i + 0*stride] = cm[ dst[i + 0*stride] + ((b0 + b7) >> 6) ];
133
        dst[i + 1*stride] = cm[ dst[i + 1*stride] + ((b2 + b5) >> 6) ];
134
        dst[i + 2*stride] = cm[ dst[i + 2*stride] + ((b4 + b3) >> 6) ];
135
        dst[i + 3*stride] = cm[ dst[i + 3*stride] + ((b6 + b1) >> 6) ];
136
        dst[i + 4*stride] = cm[ dst[i + 4*stride] + ((b6 - b1) >> 6) ];
137
        dst[i + 5*stride] = cm[ dst[i + 5*stride] + ((b4 - b3) >> 6) ];
138
        dst[i + 6*stride] = cm[ dst[i + 6*stride] + ((b2 - b5) >> 6) ];
139
        dst[i + 7*stride] = cm[ dst[i + 7*stride] + ((b0 - b7) >> 6) ];
140
    }
141
}