GCC Code Coverage Report
Directory: ../../../ffmpeg/ Exec Total Coverage
File: src/libavcodec/h264idct_template.c Lines: 191 191 100.0 %
Date: 2021-04-18 21:26:34 Branches: 70 70 100.0 %

Line Branch Exec Source
1
/*
2
 * H.264 IDCT
3
 * Copyright (c) 2004-2011 Michael Niedermayer <michaelni@gmx.at>
4
 *
5
 * This file is part of FFmpeg.
6
 *
7
 * FFmpeg is free software; you can redistribute it and/or
8
 * modify it under the terms of the GNU Lesser General Public
9
 * License as published by the Free Software Foundation; either
10
 * version 2.1 of the License, or (at your option) any later version.
11
 *
12
 * FFmpeg is distributed in the hope that it will be useful,
13
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
15
 * Lesser General Public License for more details.
16
 *
17
 * You should have received a copy of the GNU Lesser General Public
18
 * License along with FFmpeg; if not, write to the Free Software
19
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20
 */
21
22
/**
23
 * @file
24
 * H.264 IDCT.
25
 * @author Michael Niedermayer <michaelni@gmx.at>
26
 */
27
28
#include "bit_depth_template.c"
29
#include "libavutil/common.h"
30
#include "h264dec.h"
31
#include "h264idct.h"
32
33
104103628
void FUNCC(ff_h264_idct_add)(uint8_t *_dst, int16_t *_block, int stride)
34
{
35
    int i;
36
104103628
    pixel *dst = (pixel*)_dst;
37
104103628
    dctcoef *block = (dctcoef*)_block;
38
104103628
    stride >>= sizeof(pixel)-1;
39
40
104103628
    block[0] += 1 << 5;
41
42
520518140
    for(i=0; i<4; i++){
43
416414512
        const SUINT z0=  block[i + 4*0]     +  (unsigned)block[i + 4*2];
44
416414512
        const SUINT z1=  block[i + 4*0]     -  (unsigned)block[i + 4*2];
45
416414512
        const SUINT z2= (block[i + 4*1]>>1) -  (unsigned)block[i + 4*3];
46
416414512
        const SUINT z3=  block[i + 4*1]     + (unsigned)(block[i + 4*3]>>1);
47
48
416414512
        block[i + 4*0]= z0 + z3;
49
416414512
        block[i + 4*1]= z1 + z2;
50
416414512
        block[i + 4*2]= z1 - z2;
51
416414512
        block[i + 4*3]= z0 - z3;
52
    }
53
54
520518140
    for(i=0; i<4; i++){
55
416414512
        const SUINT z0=  block[0 + 4*i]     +  (SUINT)block[2 + 4*i];
56
416414512
        const SUINT z1=  block[0 + 4*i]     -  (SUINT)block[2 + 4*i];
57
416414512
        const SUINT z2= (block[1 + 4*i]>>1) -  (SUINT)block[3 + 4*i];
58
416414512
        const SUINT z3=  block[1 + 4*i]     + (SUINT)(block[3 + 4*i]>>1);
59
60
416414512
        dst[i + 0*stride]= av_clip_pixel(dst[i + 0*stride] + ((int)(z0 + z3) >> 6));
61
416414512
        dst[i + 1*stride]= av_clip_pixel(dst[i + 1*stride] + ((int)(z1 + z2) >> 6));
62
416414512
        dst[i + 2*stride]= av_clip_pixel(dst[i + 2*stride] + ((int)(z1 - z2) >> 6));
63
416414512
        dst[i + 3*stride]= av_clip_pixel(dst[i + 3*stride] + ((int)(z0 - z3) >> 6));
64
    }
65
66
104103628
    memset(block, 0, 16 * sizeof(dctcoef));
67
}
68
69
13272096
void FUNCC(ff_h264_idct8_add)(uint8_t *_dst, int16_t *_block, int stride){
70
    int i;
71
13272096
    pixel *dst = (pixel*)_dst;
72
13272096
    dctcoef *block = (dctcoef*)_block;
73
13272096
    stride >>= sizeof(pixel)-1;
74
75
13272096
    block[0] += 32;
76
77
119448864
    for( i = 0; i < 8; i++ )
78
    {
79
106176768
        const unsigned int a0 =  block[i+0*8] + (unsigned)block[i+4*8];
80
106176768
        const unsigned int a2 =  block[i+0*8] - (unsigned)block[i+4*8];
81
106176768
        const unsigned int a4 = (block[i+2*8]>>1) - (unsigned)block[i+6*8];
82
106176768
        const unsigned int a6 = (block[i+6*8]>>1) + (unsigned)block[i+2*8];
83
84
106176768
        const unsigned int b0 = a0 + a6;
85
106176768
        const unsigned int b2 = a2 + a4;
86
106176768
        const unsigned int b4 = a2 - a4;
87
106176768
        const unsigned int b6 = a0 - a6;
88
89
106176768
        const int a1 = -block[i+3*8] + (unsigned)block[i+5*8] - block[i+7*8] - (block[i+7*8]>>1);
90
106176768
        const int a3 =  block[i+1*8] + (unsigned)block[i+7*8] - block[i+3*8] - (block[i+3*8]>>1);
91
106176768
        const int a5 = -block[i+1*8] + (unsigned)block[i+7*8] + block[i+5*8] + (block[i+5*8]>>1);
92
106176768
        const int a7 =  block[i+3*8] + (unsigned)block[i+5*8] + block[i+1*8] + (block[i+1*8]>>1);
93
94
106176768
        const int b1 = (a7>>2) + (unsigned)a1;
95
106176768
        const int b3 =  (unsigned)a3 + (a5>>2);
96
106176768
        const int b5 = (a3>>2) - (unsigned)a5;
97
106176768
        const int b7 =  (unsigned)a7 - (a1>>2);
98
99
106176768
        block[i+0*8] = b0 + b7;
100
106176768
        block[i+7*8] = b0 - b7;
101
106176768
        block[i+1*8] = b2 + b5;
102
106176768
        block[i+6*8] = b2 - b5;
103
106176768
        block[i+2*8] = b4 + b3;
104
106176768
        block[i+5*8] = b4 - b3;
105
106176768
        block[i+3*8] = b6 + b1;
106
106176768
        block[i+4*8] = b6 - b1;
107
    }
108
119448864
    for( i = 0; i < 8; i++ )
109
    {
110
106176768
        const unsigned a0 =  block[0+i*8] + (unsigned)block[4+i*8];
111
106176768
        const unsigned a2 =  block[0+i*8] - (unsigned)block[4+i*8];
112
106176768
        const unsigned a4 = (block[2+i*8]>>1) - (unsigned)block[6+i*8];
113
106176768
        const unsigned a6 = (block[6+i*8]>>1) + (unsigned)block[2+i*8];
114
115
106176768
        const unsigned b0 = a0 + a6;
116
106176768
        const unsigned b2 = a2 + a4;
117
106176768
        const unsigned b4 = a2 - a4;
118
106176768
        const unsigned b6 = a0 - a6;
119
120
106176768
        const int a1 = -(unsigned)block[3+i*8] + block[5+i*8] - block[7+i*8] - (block[7+i*8]>>1);
121
106176768
        const int a3 =  (unsigned)block[1+i*8] + block[7+i*8] - block[3+i*8] - (block[3+i*8]>>1);
122
106176768
        const int a5 = -(unsigned)block[1+i*8] + block[7+i*8] + block[5+i*8] + (block[5+i*8]>>1);
123
106176768
        const int a7 =  (unsigned)block[3+i*8] + block[5+i*8] + block[1+i*8] + (block[1+i*8]>>1);
124
125
106176768
        const unsigned b1 = (a7>>2) + (unsigned)a1;
126
106176768
        const unsigned b3 =  (unsigned)a3 + (a5>>2);
127
106176768
        const unsigned b5 = (a3>>2) - (unsigned)a5;
128
106176768
        const unsigned b7 =  (unsigned)a7 - (a1>>2);
129
130
106176768
        dst[i + 0*stride] = av_clip_pixel( dst[i + 0*stride] + ((int)(b0 + b7) >> 6) );
131
106176768
        dst[i + 1*stride] = av_clip_pixel( dst[i + 1*stride] + ((int)(b2 + b5) >> 6) );
132
106176768
        dst[i + 2*stride] = av_clip_pixel( dst[i + 2*stride] + ((int)(b4 + b3) >> 6) );
133
106176768
        dst[i + 3*stride] = av_clip_pixel( dst[i + 3*stride] + ((int)(b6 + b1) >> 6) );
134
106176768
        dst[i + 4*stride] = av_clip_pixel( dst[i + 4*stride] + ((int)(b6 - b1) >> 6) );
135
106176768
        dst[i + 5*stride] = av_clip_pixel( dst[i + 5*stride] + ((int)(b4 - b3) >> 6) );
136
106176768
        dst[i + 6*stride] = av_clip_pixel( dst[i + 6*stride] + ((int)(b2 - b5) >> 6) );
137
106176768
        dst[i + 7*stride] = av_clip_pixel( dst[i + 7*stride] + ((int)(b0 - b7) >> 6) );
138
    }
139
140
13272096
    memset(block, 0, 64 * sizeof(dctcoef));
141
}
142
143
// assumes all AC coefs are 0
144
56308154
void FUNCC(ff_h264_idct_dc_add)(uint8_t *_dst, int16_t *_block, int stride){
145
    int i, j;
146
56308154
    pixel *dst = (pixel*)_dst;
147
56308154
    dctcoef *block = (dctcoef*)_block;
148
56308154
    int dc = (block[0] + 32) >> 6;
149
56308154
    stride /= sizeof(pixel);
150
56308154
    block[0] = 0;
151
281540770
    for( j = 0; j < 4; j++ )
152
    {
153
1126163080
        for( i = 0; i < 4; i++ )
154
900930464
            dst[i] = av_clip_pixel( dst[i] + dc );
155
225232616
        dst += stride;
156
    }
157
}
158
159
1190710
void FUNCC(ff_h264_idct8_dc_add)(uint8_t *_dst, int16_t *_block, int stride){
160
    int i, j;
161
1190710
    pixel *dst = (pixel*)_dst;
162
1190710
    dctcoef *block = (dctcoef*)_block;
163
1190710
    int dc = (block[0] + 32) >> 6;
164
1190710
    block[0] = 0;
165
1190710
    stride /= sizeof(pixel);
166
10716390
    for( j = 0; j < 8; j++ )
167
    {
168
85731120
        for( i = 0; i < 8; i++ )
169
76205440
            dst[i] = av_clip_pixel( dst[i] + dc );
170
9525680
        dst += stride;
171
    }
172
}
173
174
8988780
void FUNCC(ff_h264_idct_add16)(uint8_t *dst, const int *block_offset, int16_t *block, int stride, const uint8_t nnzc[15*8]){
175
    int i;
176
152809260
    for(i=0; i<16; i++){
177
143820480
        int nnz = nnzc[ scan8[i] ];
178
143820480
        if(nnz){
179

54571774
            if(nnz==1 && ((dctcoef*)block)[i*16]) FUNCC(ff_h264_idct_dc_add)(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
180
45666086
            else                                  FUNCC(ff_h264_idct_add   )(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
181
        }
182
    }
183
}
184
185
1689148
void FUNCC(ff_h264_idct_add16intra)(uint8_t *dst, const int *block_offset, int16_t *block, int stride, const uint8_t nnzc[15*8]){
186
    int i;
187
28715516
    for(i=0; i<16; i++){
188
27026368
        if(nnzc[ scan8[i] ])             FUNCC(ff_h264_idct_add   )(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
189
19693814
        else if(((dctcoef*)block)[i*16]) FUNCC(ff_h264_idct_dc_add)(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
190
    }
191
}
192
193
1696090
void FUNCC(ff_h264_idct8_add4)(uint8_t *dst, const int *block_offset, int16_t *block, int stride, const uint8_t nnzc[15*8]){
194
    int i;
195
8480450
    for(i=0; i<16; i+=4){
196
6784360
        int nnz = nnzc[ scan8[i] ];
197
6784360
        if(nnz){
198

4366042
            if(nnz==1 && ((dctcoef*)block)[i*16]) FUNCC(ff_h264_idct8_dc_add)(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
199
3659796
            else                                  FUNCC(ff_h264_idct8_add   )(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
200
        }
201
    }
202
}
203
204
8519042
void FUNCC(ff_h264_idct_add8)(uint8_t **dest, const int *block_offset, int16_t *block, int stride, const uint8_t nnzc[15*8]){
205
    int i, j;
206
25557126
    for(j=1; j<3; j++){
207
85190420
        for(i=j*16; i<j*16+4; i++){
208
68152336
            if(nnzc[ scan8[i] ])
209
15978338
                FUNCC(ff_h264_idct_add   )(dest[j-1] + block_offset[i], block + i*16*sizeof(pixel), stride);
210
52173998
            else if(((dctcoef*)block)[i*16])
211
24445688
                FUNCC(ff_h264_idct_dc_add)(dest[j-1] + block_offset[i], block + i*16*sizeof(pixel), stride);
212
        }
213
    }
214
}
215
216
1350636
void FUNCC(ff_h264_idct_add8_422)(uint8_t **dest, const int *block_offset, int16_t *block, int stride, const uint8_t nnzc[15*8]){
217
    int i, j;
218
219
4051908
    for(j=1; j<3; j++){
220
13506360
        for(i=j*16; i<j*16+4; i++){
221
10805088
            if(nnzc[ scan8[i] ])
222
3641714
                FUNCC(ff_h264_idct_add   )(dest[j-1] + block_offset[i], block + i*16*sizeof(pixel), stride);
223
7163374
            else if(((dctcoef*)block)[i*16])
224
4581790
                FUNCC(ff_h264_idct_dc_add)(dest[j-1] + block_offset[i], block + i*16*sizeof(pixel), stride);
225
        }
226
    }
227
228
4051908
    for(j=1; j<3; j++){
229
13506360
        for(i=j*16+4; i<j*16+8; i++){
230
10805088
            if(nnzc[ scan8[i+4] ])
231
3691780
                FUNCC(ff_h264_idct_add   )(dest[j-1] + block_offset[i+4], block + i*16*sizeof(pixel), stride);
232
7113308
            else if(((dctcoef*)block)[i*16])
233
4646866
                FUNCC(ff_h264_idct_dc_add)(dest[j-1] + block_offset[i+4], block + i*16*sizeof(pixel), stride);
234
        }
235
    }
236
}
237
238
/**
239
 * IDCT transforms the 16 dc values and dequantizes them.
240
 * @param qmul quantization parameter
241
 */
242
1145660
void FUNCC(ff_h264_luma_dc_dequant_idct)(int16_t *_output, int16_t *_input, int qmul){
243
#define stride 16
244
    int i;
245
    int temp[16];
246
    static const uint8_t x_offset[4]={0, 2*stride, 8*stride, 10*stride};
247
1145660
    dctcoef *input = (dctcoef*)_input;
248
1145660
    dctcoef *output = (dctcoef*)_output;
249
250
5728300
    for(i=0; i<4; i++){
251
4582640
        const int z0= input[4*i+0] + input[4*i+1];
252
4582640
        const int z1= input[4*i+0] - input[4*i+1];
253
4582640
        const int z2= input[4*i+2] - input[4*i+3];
254
4582640
        const int z3= input[4*i+2] + input[4*i+3];
255
256
4582640
        temp[4*i+0]= z0+z3;
257
4582640
        temp[4*i+1]= z0-z3;
258
4582640
        temp[4*i+2]= z1-z2;
259
4582640
        temp[4*i+3]= z1+z2;
260
    }
261
262
5728300
    for(i=0; i<4; i++){
263
4582640
        const int offset= x_offset[i];
264
4582640
        const SUINT z0= temp[4*0+i] + temp[4*2+i];
265
4582640
        const SUINT z1= temp[4*0+i] - temp[4*2+i];
266
4582640
        const SUINT z2= temp[4*1+i] - temp[4*3+i];
267
4582640
        const SUINT z3= temp[4*1+i] + temp[4*3+i];
268
269
4582640
        output[stride* 0+offset]= (int)((z0 + z3)*qmul + 128 ) >> 8;
270
4582640
        output[stride* 1+offset]= (int)((z1 + z2)*qmul + 128 ) >> 8;
271
4582640
        output[stride* 4+offset]= (int)((z1 - z2)*qmul + 128 ) >> 8;
272
4582640
        output[stride* 5+offset]= (int)((z0 - z3)*qmul + 128 ) >> 8;
273
    }
274
#undef stride
275
}
276
277
2252062
void FUNCC(ff_h264_chroma422_dc_dequant_idct)(int16_t *_block, int qmul){
278
2252062
    const int stride= 16*2;
279
2252062
    const int xStride= 16;
280
    int i;
281
    unsigned temp[8];
282
    static const uint8_t x_offset[2]={0, 16};
283
2252062
    dctcoef *block = (dctcoef*)_block;
284
285
11260310
    for(i=0; i<4; i++){
286
9008248
        temp[2*i+0] = block[stride*i + xStride*0] + (unsigned)block[stride*i + xStride*1];
287
9008248
        temp[2*i+1] = block[stride*i + xStride*0] - (unsigned)block[stride*i + xStride*1];
288
    }
289
290
6756186
    for(i=0; i<2; i++){
291
4504124
        const int offset= x_offset[i];
292
4504124
        const SUINT z0= temp[2*0+i] + temp[2*2+i];
293
4504124
        const SUINT z1= temp[2*0+i] - temp[2*2+i];
294
4504124
        const SUINT z2= temp[2*1+i] - temp[2*3+i];
295
4504124
        const SUINT z3= temp[2*1+i] + temp[2*3+i];
296
297
4504124
        block[stride*0+offset]= (int)((z0 + z3)*qmul + 128) >> 8;
298
4504124
        block[stride*1+offset]= (int)((z1 + z2)*qmul + 128) >> 8;
299
4504124
        block[stride*2+offset]= (int)((z1 - z2)*qmul + 128) >> 8;
300
4504124
        block[stride*3+offset]= (int)((z0 - z3)*qmul + 128) >> 8;
301
    }
302
}
303
304
12086092
void FUNCC(ff_h264_chroma_dc_dequant_idct)(int16_t *_block, int qmul){
305
12086092
    const int stride= 16*2;
306
12086092
    const int xStride= 16;
307
    SUINT a,b,c,d,e;
308
12086092
    dctcoef *block = (dctcoef*)_block;
309
310
12086092
    a= block[stride*0 + xStride*0];
311
12086092
    b= block[stride*0 + xStride*1];
312
12086092
    c= block[stride*1 + xStride*0];
313
12086092
    d= block[stride*1 + xStride*1];
314
315
12086092
    e= a-b;
316
12086092
    a= a+b;
317
12086092
    b= c-d;
318
12086092
    c= c+d;
319
320
12086092
    block[stride*0 + xStride*0]= (int)((a+c)*qmul) >> 7;
321
12086092
    block[stride*0 + xStride*1]= (int)((e+b)*qmul) >> 7;
322
12086092
    block[stride*1 + xStride*0]= (int)((a-c)*qmul) >> 7;
323
12086092
    block[stride*1 + xStride*1]= (int)((e-b)*qmul) >> 7;
324
}