FFmpeg coverage


Directory: ../../../ffmpeg/
File: src/libavcodec/h264idct_template.c
Date: 2021-09-26 18:22:30
Exec Total Coverage
Lines: 191 191 100.0%
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 104103412 void FUNCC(ff_h264_idct_add)(uint8_t *_dst, int16_t *_block, int stride)
34 {
35 int i;
36 104103412 pixel *dst = (pixel*)_dst;
37 104103412 dctcoef *block = (dctcoef*)_block;
38 104103412 stride >>= sizeof(pixel)-1;
39
40 104103412 block[0] += 1 << 5;
41
42
2/2
✓ Branch 0 taken 208206824 times.
✓ Branch 1 taken 52051706 times.
520517060 for(i=0; i<4; i++){
43 416413648 const SUINT z0= block[i + 4*0] + (unsigned)block[i + 4*2];
44 416413648 const SUINT z1= block[i + 4*0] - (unsigned)block[i + 4*2];
45 416413648 const SUINT z2= (block[i + 4*1]>>1) - (unsigned)block[i + 4*3];
46 416413648 const SUINT z3= block[i + 4*1] + (unsigned)(block[i + 4*3]>>1);
47
48 416413648 block[i + 4*0]= z0 + z3;
49 416413648 block[i + 4*1]= z1 + z2;
50 416413648 block[i + 4*2]= z1 - z2;
51 416413648 block[i + 4*3]= z0 - z3;
52 }
53
54
2/2
✓ Branch 0 taken 208206824 times.
✓ Branch 1 taken 52051706 times.
520517060 for(i=0; i<4; i++){
55 416413648 const SUINT z0= block[0 + 4*i] + (SUINT)block[2 + 4*i];
56 416413648 const SUINT z1= block[0 + 4*i] - (SUINT)block[2 + 4*i];
57 416413648 const SUINT z2= (block[1 + 4*i]>>1) - (SUINT)block[3 + 4*i];
58 416413648 const SUINT z3= block[1 + 4*i] + (SUINT)(block[3 + 4*i]>>1);
59
60 416413648 dst[i + 0*stride]= av_clip_pixel(dst[i + 0*stride] + ((int)(z0 + z3) >> 6));
61 416413648 dst[i + 1*stride]= av_clip_pixel(dst[i + 1*stride] + ((int)(z1 + z2) >> 6));
62 416413648 dst[i + 2*stride]= av_clip_pixel(dst[i + 2*stride] + ((int)(z1 - z2) >> 6));
63 416413648 dst[i + 3*stride]= av_clip_pixel(dst[i + 3*stride] + ((int)(z0 - z3) >> 6));
64 }
65
66 104103412 memset(block, 0, 16 * sizeof(dctcoef));
67 }
68
69 13272110 void FUNCC(ff_h264_idct8_add)(uint8_t *_dst, int16_t *_block, int stride){
70 int i;
71 13272110 pixel *dst = (pixel*)_dst;
72 13272110 dctcoef *block = (dctcoef*)_block;
73 13272110 stride >>= sizeof(pixel)-1;
74
75 13272110 block[0] += 32;
76
77
2/2
✓ Branch 0 taken 53088440 times.
✓ Branch 1 taken 6636055 times.
119448990 for( i = 0; i < 8; i++ )
78 {
79 106176880 const unsigned int a0 = block[i+0*8] + (unsigned)block[i+4*8];
80 106176880 const unsigned int a2 = block[i+0*8] - (unsigned)block[i+4*8];
81 106176880 const unsigned int a4 = (block[i+2*8]>>1) - (unsigned)block[i+6*8];
82 106176880 const unsigned int a6 = (block[i+6*8]>>1) + (unsigned)block[i+2*8];
83
84 106176880 const unsigned int b0 = a0 + a6;
85 106176880 const unsigned int b2 = a2 + a4;
86 106176880 const unsigned int b4 = a2 - a4;
87 106176880 const unsigned int b6 = a0 - a6;
88
89 106176880 const int a1 = -block[i+3*8] + (unsigned)block[i+5*8] - block[i+7*8] - (block[i+7*8]>>1);
90 106176880 const int a3 = block[i+1*8] + (unsigned)block[i+7*8] - block[i+3*8] - (block[i+3*8]>>1);
91 106176880 const int a5 = -block[i+1*8] + (unsigned)block[i+7*8] + block[i+5*8] + (block[i+5*8]>>1);
92 106176880 const int a7 = block[i+3*8] + (unsigned)block[i+5*8] + block[i+1*8] + (block[i+1*8]>>1);
93
94 106176880 const int b1 = (a7>>2) + (unsigned)a1;
95 106176880 const int b3 = (unsigned)a3 + (a5>>2);
96 106176880 const int b5 = (a3>>2) - (unsigned)a5;
97 106176880 const int b7 = (unsigned)a7 - (a1>>2);
98
99 106176880 block[i+0*8] = b0 + b7;
100 106176880 block[i+7*8] = b0 - b7;
101 106176880 block[i+1*8] = b2 + b5;
102 106176880 block[i+6*8] = b2 - b5;
103 106176880 block[i+2*8] = b4 + b3;
104 106176880 block[i+5*8] = b4 - b3;
105 106176880 block[i+3*8] = b6 + b1;
106 106176880 block[i+4*8] = b6 - b1;
107 }
108
2/2
✓ Branch 0 taken 53088440 times.
✓ Branch 1 taken 6636055 times.
119448990 for( i = 0; i < 8; i++ )
109 {
110 106176880 const unsigned a0 = block[0+i*8] + (unsigned)block[4+i*8];
111 106176880 const unsigned a2 = block[0+i*8] - (unsigned)block[4+i*8];
112 106176880 const unsigned a4 = (block[2+i*8]>>1) - (unsigned)block[6+i*8];
113 106176880 const unsigned a6 = (block[6+i*8]>>1) + (unsigned)block[2+i*8];
114
115 106176880 const unsigned b0 = a0 + a6;
116 106176880 const unsigned b2 = a2 + a4;
117 106176880 const unsigned b4 = a2 - a4;
118 106176880 const unsigned b6 = a0 - a6;
119
120 106176880 const int a1 = -(unsigned)block[3+i*8] + block[5+i*8] - block[7+i*8] - (block[7+i*8]>>1);
121 106176880 const int a3 = (unsigned)block[1+i*8] + block[7+i*8] - block[3+i*8] - (block[3+i*8]>>1);
122 106176880 const int a5 = -(unsigned)block[1+i*8] + block[7+i*8] + block[5+i*8] + (block[5+i*8]>>1);
123 106176880 const int a7 = (unsigned)block[3+i*8] + block[5+i*8] + block[1+i*8] + (block[1+i*8]>>1);
124
125 106176880 const unsigned b1 = (a7>>2) + (unsigned)a1;
126 106176880 const unsigned b3 = (unsigned)a3 + (a5>>2);
127 106176880 const unsigned b5 = (a3>>2) - (unsigned)a5;
128 106176880 const unsigned b7 = (unsigned)a7 - (a1>>2);
129
130 106176880 dst[i + 0*stride] = av_clip_pixel( dst[i + 0*stride] + ((int)(b0 + b7) >> 6) );
131 106176880 dst[i + 1*stride] = av_clip_pixel( dst[i + 1*stride] + ((int)(b2 + b5) >> 6) );
132 106176880 dst[i + 2*stride] = av_clip_pixel( dst[i + 2*stride] + ((int)(b4 + b3) >> 6) );
133 106176880 dst[i + 3*stride] = av_clip_pixel( dst[i + 3*stride] + ((int)(b6 + b1) >> 6) );
134 106176880 dst[i + 4*stride] = av_clip_pixel( dst[i + 4*stride] + ((int)(b6 - b1) >> 6) );
135 106176880 dst[i + 5*stride] = av_clip_pixel( dst[i + 5*stride] + ((int)(b4 - b3) >> 6) );
136 106176880 dst[i + 6*stride] = av_clip_pixel( dst[i + 6*stride] + ((int)(b2 - b5) >> 6) );
137 106176880 dst[i + 7*stride] = av_clip_pixel( dst[i + 7*stride] + ((int)(b0 - b7) >> 6) );
138 }
139
140 13272110 memset(block, 0, 64 * sizeof(dctcoef));
141 }
142
143 // assumes all AC coefs are 0
144 56294700 void FUNCC(ff_h264_idct_dc_add)(uint8_t *_dst, int16_t *_block, int stride){
145 int i, j;
146 56294700 pixel *dst = (pixel*)_dst;
147 56294700 dctcoef *block = (dctcoef*)_block;
148 56294700 int dc = (block[0] + 32) >> 6;
149 56294700 stride /= sizeof(pixel);
150 56294700 block[0] = 0;
151
2/2
✓ Branch 0 taken 112589400 times.
✓ Branch 1 taken 28147350 times.
281473500 for( j = 0; j < 4; j++ )
152 {
153
2/2
✓ Branch 0 taken 450357600 times.
✓ Branch 1 taken 112589400 times.
1125894000 for( i = 0; i < 4; i++ )
154 900715200 dst[i] = av_clip_pixel( dst[i] + dc );
155 225178800 dst += stride;
156 }
157 }
158
159 1190698 void FUNCC(ff_h264_idct8_dc_add)(uint8_t *_dst, int16_t *_block, int stride){
160 int i, j;
161 1190698 pixel *dst = (pixel*)_dst;
162 1190698 dctcoef *block = (dctcoef*)_block;
163 1190698 int dc = (block[0] + 32) >> 6;
164 1190698 block[0] = 0;
165 1190698 stride /= sizeof(pixel);
166
2/2
✓ Branch 0 taken 4762792 times.
✓ Branch 1 taken 595349 times.
10716282 for( j = 0; j < 8; j++ )
167 {
168
2/2
✓ Branch 0 taken 38102336 times.
✓ Branch 1 taken 4762792 times.
85730256 for( i = 0; i < 8; i++ )
169 76204672 dst[i] = av_clip_pixel( dst[i] + dc );
170 9525584 dst += stride;
171 }
172 }
173
174 8988718 void FUNCC(ff_h264_idct_add16)(uint8_t *dst, const int *block_offset,
175 int16_t *block, int stride,
176 const uint8_t nnzc[5 * 8])
177 {
178 int i;
179
2/2
✓ Branch 0 taken 71909744 times.
✓ Branch 1 taken 4494359 times.
152808206 for(i=0; i<16; i++){
180 143819488 int nnz = nnzc[ scan8[i] ];
181
2/2
✓ Branch 0 taken 27285875 times.
✓ Branch 1 taken 44623869 times.
143819488 if(nnz){
182
4/4
✓ Branch 0 taken 12373387 times.
✓ Branch 1 taken 14912488 times.
✓ Branch 2 taken 4452830 times.
✓ Branch 3 taken 7920557 times.
54571750 if(nnz==1 && ((dctcoef*)block)[i*16]) FUNCC(ff_h264_idct_dc_add)(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
183 45666090 else FUNCC(ff_h264_idct_add )(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
184 }
185 }
186 }
187
188 1688256 void FUNCC(ff_h264_idct_add16intra)(uint8_t *dst, const int *block_offset,
189 int16_t *block, int stride,
190 const uint8_t nnzc[5 * 8])
191 {
192 int i;
193
2/2
✓ Branch 0 taken 13506048 times.
✓ Branch 1 taken 844128 times.
28700352 for(i=0; i<16; i++){
194
2/2
✓ Branch 0 taken 3666264 times.
✓ Branch 1 taken 9839784 times.
27012096 if(nnzc[ scan8[i] ]) FUNCC(ff_h264_idct_add )(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
195
2/2
✓ Branch 0 taken 4830289 times.
✓ Branch 1 taken 5009495 times.
19679568 else if(((dctcoef*)block)[i*16]) FUNCC(ff_h264_idct_dc_add)(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
196 }
197 }
198
199 1696090 void FUNCC(ff_h264_idct8_add4)(uint8_t *dst, const int *block_offset,
200 int16_t *block, int stride,
201 const uint8_t nnzc[5 * 8])
202 {
203 int i;
204
2/2
✓ Branch 0 taken 3392180 times.
✓ Branch 1 taken 848045 times.
8480450 for(i=0; i<16; i+=4){
205 6784360 int nnz = nnzc[ scan8[i] ];
206
2/2
✓ Branch 0 taken 2183022 times.
✓ Branch 1 taken 1209158 times.
6784360 if(nnz){
207
4/4
✓ Branch 0 taken 660003 times.
✓ Branch 1 taken 1523019 times.
✓ Branch 2 taken 353117 times.
✓ Branch 3 taken 306886 times.
4366044 if(nnz==1 && ((dctcoef*)block)[i*16]) FUNCC(ff_h264_idct8_dc_add)(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
208 3659810 else FUNCC(ff_h264_idct8_add )(dst + block_offset[i], block + i*16*sizeof(pixel), stride);
209 }
210 }
211 }
212
213 8518926 void FUNCC(ff_h264_idct_add8)(uint8_t **dest, const int *block_offset, int16_t *block, int stride, const uint8_t nnzc[15*8]){
214 int i, j;
215
2/2
✓ Branch 0 taken 8518926 times.
✓ Branch 1 taken 4259463 times.
25556778 for(j=1; j<3; j++){
216
2/2
✓ Branch 0 taken 34075704 times.
✓ Branch 1 taken 8518926 times.
85189260 for(i=j*16; i<j*16+4; i++){
217
2/2
✓ Branch 0 taken 7989098 times.
✓ Branch 1 taken 26086606 times.
68151408 if(nnzc[ scan8[i] ])
218 15978196 FUNCC(ff_h264_idct_add )(dest[j-1] + block_offset[i], block + i*16*sizeof(pixel), stride);
219
2/2
✓ Branch 0 taken 12222788 times.
✓ Branch 1 taken 13863818 times.
52173212 else if(((dctcoef*)block)[i*16])
220 24445576 FUNCC(ff_h264_idct_dc_add)(dest[j-1] + block_offset[i], block + i*16*sizeof(pixel), stride);
221 }
222 }
223 }
224
225 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]){
226 int i, j;
227
228
2/2
✓ Branch 0 taken 1350636 times.
✓ Branch 1 taken 675318 times.
4051908 for(j=1; j<3; j++){
229
2/2
✓ Branch 0 taken 5402544 times.
✓ Branch 1 taken 1350636 times.
13506360 for(i=j*16; i<j*16+4; i++){
230
2/2
✓ Branch 0 taken 1820857 times.
✓ Branch 1 taken 3581687 times.
10805088 if(nnzc[ scan8[i] ])
231 3641714 FUNCC(ff_h264_idct_add )(dest[j-1] + block_offset[i], block + i*16*sizeof(pixel), stride);
232
2/2
✓ Branch 0 taken 2290895 times.
✓ Branch 1 taken 1290792 times.
7163374 else if(((dctcoef*)block)[i*16])
233 4581790 FUNCC(ff_h264_idct_dc_add)(dest[j-1] + block_offset[i], block + i*16*sizeof(pixel), stride);
234 }
235 }
236
237
2/2
✓ Branch 0 taken 1350636 times.
✓ Branch 1 taken 675318 times.
4051908 for(j=1; j<3; j++){
238
2/2
✓ Branch 0 taken 5402544 times.
✓ Branch 1 taken 1350636 times.
13506360 for(i=j*16+4; i<j*16+8; i++){
239
2/2
✓ Branch 0 taken 1845890 times.
✓ Branch 1 taken 3556654 times.
10805088 if(nnzc[ scan8[i+4] ])
240 3691780 FUNCC(ff_h264_idct_add )(dest[j-1] + block_offset[i+4], block + i*16*sizeof(pixel), stride);
241
2/2
✓ Branch 0 taken 2323433 times.
✓ Branch 1 taken 1233221 times.
7113308 else if(((dctcoef*)block)[i*16])
242 4646866 FUNCC(ff_h264_idct_dc_add)(dest[j-1] + block_offset[i+4], block + i*16*sizeof(pixel), stride);
243 }
244 }
245 }
246
247 /**
248 * IDCT transforms the 16 dc values and dequantizes them.
249 * @param qmul quantization parameter
250 */
251 1144828 void FUNCC(ff_h264_luma_dc_dequant_idct)(int16_t *_output, int16_t *_input, int qmul){
252 #define stride 16
253 int i;
254 int temp[16];
255 static const uint8_t x_offset[4]={0, 2*stride, 8*stride, 10*stride};
256 1144828 dctcoef *input = (dctcoef*)_input;
257 1144828 dctcoef *output = (dctcoef*)_output;
258
259
2/2
✓ Branch 0 taken 2289656 times.
✓ Branch 1 taken 572414 times.
5724140 for(i=0; i<4; i++){
260 4579312 const int z0= input[4*i+0] + input[4*i+1];
261 4579312 const int z1= input[4*i+0] - input[4*i+1];
262 4579312 const int z2= input[4*i+2] - input[4*i+3];
263 4579312 const int z3= input[4*i+2] + input[4*i+3];
264
265 4579312 temp[4*i+0]= z0+z3;
266 4579312 temp[4*i+1]= z0-z3;
267 4579312 temp[4*i+2]= z1-z2;
268 4579312 temp[4*i+3]= z1+z2;
269 }
270
271
2/2
✓ Branch 0 taken 2289656 times.
✓ Branch 1 taken 572414 times.
5724140 for(i=0; i<4; i++){
272 4579312 const int offset= x_offset[i];
273 4579312 const SUINT z0= temp[4*0+i] + temp[4*2+i];
274 4579312 const SUINT z1= temp[4*0+i] - temp[4*2+i];
275 4579312 const SUINT z2= temp[4*1+i] - temp[4*3+i];
276 4579312 const SUINT z3= temp[4*1+i] + temp[4*3+i];
277
278 4579312 output[stride* 0+offset]= (int)((z0 + z3)*qmul + 128 ) >> 8;
279 4579312 output[stride* 1+offset]= (int)((z1 + z2)*qmul + 128 ) >> 8;
280 4579312 output[stride* 4+offset]= (int)((z1 - z2)*qmul + 128 ) >> 8;
281 4579312 output[stride* 5+offset]= (int)((z0 - z3)*qmul + 128 ) >> 8;
282 }
283 #undef stride
284 }
285
286 2252062 void FUNCC(ff_h264_chroma422_dc_dequant_idct)(int16_t *_block, int qmul){
287 2252062 const int stride= 16*2;
288 2252062 const int xStride= 16;
289 int i;
290 unsigned temp[8];
291 static const uint8_t x_offset[2]={0, 16};
292 2252062 dctcoef *block = (dctcoef*)_block;
293
294
2/2
✓ Branch 0 taken 4504124 times.
✓ Branch 1 taken 1126031 times.
11260310 for(i=0; i<4; i++){
295 9008248 temp[2*i+0] = block[stride*i + xStride*0] + (unsigned)block[stride*i + xStride*1];
296 9008248 temp[2*i+1] = block[stride*i + xStride*0] - (unsigned)block[stride*i + xStride*1];
297 }
298
299
2/2
✓ Branch 0 taken 2252062 times.
✓ Branch 1 taken 1126031 times.
6756186 for(i=0; i<2; i++){
300 4504124 const int offset= x_offset[i];
301 4504124 const SUINT z0= temp[2*0+i] + temp[2*2+i];
302 4504124 const SUINT z1= temp[2*0+i] - temp[2*2+i];
303 4504124 const SUINT z2= temp[2*1+i] - temp[2*3+i];
304 4504124 const SUINT z3= temp[2*1+i] + temp[2*3+i];
305
306 4504124 block[stride*0+offset]= (int)((z0 + z3)*qmul + 128) >> 8;
307 4504124 block[stride*1+offset]= (int)((z1 + z2)*qmul + 128) >> 8;
308 4504124 block[stride*2+offset]= (int)((z1 - z2)*qmul + 128) >> 8;
309 4504124 block[stride*3+offset]= (int)((z0 - z3)*qmul + 128) >> 8;
310 }
311 }
312
313 12086092 void FUNCC(ff_h264_chroma_dc_dequant_idct)(int16_t *_block, int qmul){
314 12086092 const int stride= 16*2;
315 12086092 const int xStride= 16;
316 SUINT a,b,c,d,e;
317 12086092 dctcoef *block = (dctcoef*)_block;
318
319 12086092 a= block[stride*0 + xStride*0];
320 12086092 b= block[stride*0 + xStride*1];
321 12086092 c= block[stride*1 + xStride*0];
322 12086092 d= block[stride*1 + xStride*1];
323
324 12086092 e= a-b;
325 12086092 a= a+b;
326 12086092 b= c-d;
327 12086092 c= c+d;
328
329 12086092 block[stride*0 + xStride*0]= (int)((a+c)*qmul) >> 7;
330 12086092 block[stride*0 + xStride*1]= (int)((e+b)*qmul) >> 7;
331 12086092 block[stride*1 + xStride*0]= (int)((a-c)*qmul) >> 7;
332 12086092 block[stride*1 + xStride*1]= (int)((e-b)*qmul) >> 7;
333 }
334