FFmpeg coverage


Directory: ../../../ffmpeg/
File: src/libavcodec/h264dsp_template.c
Date: 2024-10-04 17:46:48
Exec Total Coverage
Lines: 162 162 100.0%
Functions: 74 140 52.9%
Branches: 70 70 100.0%

Line Branch Exec Source
1 /*
2 * H.26L/H.264/AVC/JVT/14496-10/... encoder/decoder
3 * Copyright (c) 2003-2011 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 / AVC / MPEG-4 part10 DSP functions.
25 * @author Michael Niedermayer <michaelni@gmx.at>
26 */
27
28 #include "bit_depth_template.c"
29
30 #define op_scale1(x) block[x] = av_clip_pixel( (block[x]*weight + offset) >> log2_denom )
31 #define op_scale2(x) dst[x] = av_clip_pixel( (src[x]*weights + dst[x]*weightd + offset) >> (log2_denom+1))
32 #define H264_WEIGHT(W) \
33 static void FUNCC(weight_h264_pixels ## W)(uint8_t *_block, ptrdiff_t stride, int height, \
34 int log2_denom, int weight, int offset) \
35 { \
36 int y; \
37 pixel *block = (pixel*)_block; \
38 stride >>= sizeof(pixel)-1; \
39 offset = (unsigned)offset << (log2_denom + (BIT_DEPTH-8)); \
40 if(log2_denom) offset += 1<<(log2_denom-1); \
41 for (y = 0; y < height; y++, block += stride) { \
42 op_scale1(0); \
43 op_scale1(1); \
44 if(W==2) continue; \
45 op_scale1(2); \
46 op_scale1(3); \
47 if(W==4) continue; \
48 op_scale1(4); \
49 op_scale1(5); \
50 op_scale1(6); \
51 op_scale1(7); \
52 if(W==8) continue; \
53 op_scale1(8); \
54 op_scale1(9); \
55 op_scale1(10); \
56 op_scale1(11); \
57 op_scale1(12); \
58 op_scale1(13); \
59 op_scale1(14); \
60 op_scale1(15); \
61 } \
62 } \
63 static void FUNCC(biweight_h264_pixels ## W)(uint8_t *_dst, uint8_t *_src, ptrdiff_t stride, int height, \
64 int log2_denom, int weightd, int weights, int offset) \
65 { \
66 int y; \
67 pixel *dst = (pixel*)_dst; \
68 pixel *src = (pixel*)_src; \
69 stride >>= sizeof(pixel)-1; \
70 offset = (unsigned)offset << (BIT_DEPTH-8); \
71 offset = (unsigned)((offset + 1) | 1) << log2_denom; \
72 for (y = 0; y < height; y++, dst += stride, src += stride) { \
73 op_scale2(0); \
74 op_scale2(1); \
75 if(W==2) continue; \
76 op_scale2(2); \
77 op_scale2(3); \
78 if(W==4) continue; \
79 op_scale2(4); \
80 op_scale2(5); \
81 op_scale2(6); \
82 op_scale2(7); \
83 if(W==8) continue; \
84 op_scale2(8); \
85 op_scale2(9); \
86 op_scale2(10); \
87 op_scale2(11); \
88 op_scale2(12); \
89 op_scale2(13); \
90 op_scale2(14); \
91 op_scale2(15); \
92 } \
93 }
94
95
4/4
✓ Branch 0 taken 3595955 times.
✓ Branch 1 taken 672437 times.
✓ Branch 2 taken 9614472 times.
✓ Branch 3 taken 635068 times.
27765728 H264_WEIGHT(16)
96
4/4
✓ Branch 0 taken 4241088 times.
✓ Branch 1 taken 531832 times.
✓ Branch 2 taken 3562912 times.
✓ Branch 3 taken 475486 times.
16671664 H264_WEIGHT(8)
97
4/4
✓ Branch 0 taken 1273166 times.
✓ Branch 1 taken 151830 times.
✓ Branch 2 taken 3111408 times.
✓ Branch 3 taken 697530 times.
9072808 H264_WEIGHT(4)
98
4/4
✓ Branch 0 taken 642584 times.
✓ Branch 1 taken 98228 times.
✓ Branch 2 taken 1261776 times.
✓ Branch 3 taken 447120 times.
4005176 H264_WEIGHT(2)
99
100 #undef op_scale1
101 #undef op_scale2
102 #undef H264_WEIGHT
103
104 75228902 static av_always_inline av_flatten void FUNCC(h264_loop_filter_luma)(uint8_t *p_pix, ptrdiff_t xstride, ptrdiff_t ystride, int inner_iters, int alpha, int beta, int8_t *tc0)
105 {
106 75228902 pixel *pix = (pixel*)p_pix;
107 int i, d;
108 75228902 xstride >>= sizeof(pixel)-1;
109 75228902 ystride >>= sizeof(pixel)-1;
110 75228902 alpha <<= BIT_DEPTH - 8;
111 75228902 beta <<= BIT_DEPTH - 8;
112
2/2
✓ Branch 0 taken 150457804 times.
✓ Branch 1 taken 37614451 times.
376144510 for( i = 0; i < 4; i++ ) {
113 300915608 const int tc_orig = tc0[i] * (1 << (BIT_DEPTH - 8));
114
2/2
✓ Branch 0 taken 25250498 times.
✓ Branch 1 taken 125207306 times.
300915608 if( tc_orig < 0 ) {
115 50500996 pix += inner_iters*ystride;
116 50500996 continue;
117 }
118
2/2
✓ Branch 0 taken 498424576 times.
✓ Branch 1 taken 125207306 times.
1247263764 for( d = 0; d < inner_iters; d++ ) {
119 996849152 const int p0 = pix[-1*xstride];
120 996849152 const int p1 = pix[-2*xstride];
121 996849152 const int p2 = pix[-3*xstride];
122 996849152 const int q0 = pix[0];
123 996849152 const int q1 = pix[1*xstride];
124 996849152 const int q2 = pix[2*xstride];
125
126
2/2
✓ Branch 0 taken 427196195 times.
✓ Branch 1 taken 71228381 times.
996849152 if( FFABS( p0 - q0 ) < alpha &&
127
2/2
✓ Branch 0 taken 321120798 times.
✓ Branch 1 taken 106075397 times.
854392390 FFABS( p1 - p0 ) < beta &&
128
2/2
✓ Branch 0 taken 268818597 times.
✓ Branch 1 taken 52302201 times.
642241596 FFABS( q1 - q0 ) < beta ) {
129
130 537637194 int tc = tc_orig;
131 int i_delta;
132
133
2/2
✓ Branch 0 taken 227849078 times.
✓ Branch 1 taken 40969519 times.
537637194 if( FFABS( p2 - p0 ) < beta ) {
134
2/2
✓ Branch 0 taken 213053887 times.
✓ Branch 1 taken 14795191 times.
455698156 if(tc_orig)
135 426107774 pix[-2*xstride] = p1 + av_clip( (( p2 + ( ( p0 + q0 + 1 ) >> 1 ) ) >> 1) - p1, -tc_orig, tc_orig );
136 455698156 tc++;
137 }
138
2/2
✓ Branch 0 taken 226686453 times.
✓ Branch 1 taken 42132144 times.
537637194 if( FFABS( q2 - q0 ) < beta ) {
139
2/2
✓ Branch 0 taken 211918378 times.
✓ Branch 1 taken 14768075 times.
453372906 if(tc_orig)
140 423836756 pix[ xstride] = q1 + av_clip( (( q2 + ( ( p0 + q0 + 1 ) >> 1 ) ) >> 1) - q1, -tc_orig, tc_orig );
141 453372906 tc++;
142 }
143
144 537637194 i_delta = av_clip( (((q0 - p0 ) * 4) + (p1 - q1) + 4) >> 3, -tc, tc );
145 537637194 pix[-xstride] = av_clip_pixel( p0 + i_delta ); /* p0' */
146 537637194 pix[0] = av_clip_pixel( q0 - i_delta ); /* q0' */
147 }
148 996849152 pix += ystride;
149 }
150 }
151 75228902 }
152 38175120 static void FUNCC(h264_v_loop_filter_luma)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
153 {
154 38175120 FUNCC(h264_loop_filter_luma)(pix, stride, sizeof(pixel), 4, alpha, beta, tc0);
155 38175120 }
156 36452620 static void FUNCC(h264_h_loop_filter_luma)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
157 {
158 36452620 FUNCC(h264_loop_filter_luma)(pix, sizeof(pixel), stride, 4, alpha, beta, tc0);
159 36452620 }
160 601162 static void FUNCC(h264_h_loop_filter_luma_mbaff)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
161 {
162 601162 FUNCC(h264_loop_filter_luma)(pix, sizeof(pixel), stride, 2, alpha, beta, tc0);
163 601162 }
164
165 11271522 static av_always_inline av_flatten void FUNCC(h264_loop_filter_luma_intra)(uint8_t *p_pix, ptrdiff_t xstride, ptrdiff_t ystride, int inner_iters, int alpha, int beta)
166 {
167 11271522 pixel *pix = (pixel*)p_pix;
168 int d;
169 11271522 xstride >>= sizeof(pixel)-1;
170 11271522 ystride >>= sizeof(pixel)-1;
171 11271522 alpha <<= BIT_DEPTH - 8;
172 11271522 beta <<= BIT_DEPTH - 8;
173
2/2
✓ Branch 0 taken 87165368 times.
✓ Branch 1 taken 5635761 times.
185602258 for( d = 0; d < 4 * inner_iters; d++ ) {
174 174330736 const int p2 = pix[-3*xstride];
175 174330736 const int p1 = pix[-2*xstride];
176 174330736 const int p0 = pix[-1*xstride];
177
178 174330736 const int q0 = pix[ 0*xstride];
179 174330736 const int q1 = pix[ 1*xstride];
180 174330736 const int q2 = pix[ 2*xstride];
181
182
2/2
✓ Branch 0 taken 76663663 times.
✓ Branch 1 taken 10501705 times.
174330736 if( FFABS( p0 - q0 ) < alpha &&
183
2/2
✓ Branch 0 taken 61458206 times.
✓ Branch 1 taken 15205457 times.
153327326 FFABS( p1 - p0 ) < beta &&
184
2/2
✓ Branch 0 taken 52928700 times.
✓ Branch 1 taken 8529506 times.
122916412 FFABS( q1 - q0 ) < beta ) {
185
186
2/2
✓ Branch 0 taken 43979562 times.
✓ Branch 1 taken 8949138 times.
105857400 if(FFABS( p0 - q0 ) < (( alpha >> 2 ) + 2 )){
187
2/2
✓ Branch 0 taken 39597378 times.
✓ Branch 1 taken 4382184 times.
87959124 if( FFABS( p2 - p0 ) < beta)
188 {
189 79194756 const int p3 = pix[-4*xstride];
190 /* p0', p1', p2' */
191 79194756 pix[-1*xstride] = ( p2 + 2*p1 + 2*p0 + 2*q0 + q1 + 4 ) >> 3;
192 79194756 pix[-2*xstride] = ( p2 + p1 + p0 + q0 + 2 ) >> 2;
193 79194756 pix[-3*xstride] = ( 2*p3 + 3*p2 + p1 + p0 + q0 + 4 ) >> 3;
194 } else {
195 /* p0' */
196 8764368 pix[-1*xstride] = ( 2*p1 + p0 + q1 + 2 ) >> 2;
197 }
198
2/2
✓ Branch 0 taken 39368188 times.
✓ Branch 1 taken 4611374 times.
87959124 if( FFABS( q2 - q0 ) < beta)
199 {
200 78736376 const int q3 = pix[3*xstride];
201 /* q0', q1', q2' */
202 78736376 pix[0*xstride] = ( p1 + 2*p0 + 2*q0 + 2*q1 + q2 + 4 ) >> 3;
203 78736376 pix[1*xstride] = ( p0 + q0 + q1 + q2 + 2 ) >> 2;
204 78736376 pix[2*xstride] = ( 2*q3 + 3*q2 + q1 + q0 + p0 + 4 ) >> 3;
205 } else {
206 /* q0' */
207 9222748 pix[0*xstride] = ( 2*q1 + q0 + p1 + 2 ) >> 2;
208 }
209 }else{
210 /* p0', q0' */
211 17898276 pix[-1*xstride] = ( 2*p1 + p0 + q1 + 2 ) >> 2;
212 17898276 pix[ 0*xstride] = ( 2*q1 + q0 + p1 + 2 ) >> 2;
213 }
214 }
215 174330736 pix += ystride;
216 }
217 11271522 }
218 4657142 static void FUNCC(h264_v_loop_filter_luma_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
219 {
220 4657142 FUNCC(h264_loop_filter_luma_intra)(pix, stride, sizeof(pixel), 4, alpha, beta);
221 4657142 }
222 5862678 static void FUNCC(h264_h_loop_filter_luma_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
223 {
224 5862678 FUNCC(h264_loop_filter_luma_intra)(pix, sizeof(pixel), stride, 4, alpha, beta);
225 5862678 }
226 751702 static void FUNCC(h264_h_loop_filter_luma_mbaff_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
227 {
228 751702 FUNCC(h264_loop_filter_luma_intra)(pix, sizeof(pixel), stride, 2, alpha, beta);
229 751702 }
230
231 89419874 static av_always_inline av_flatten void FUNCC(h264_loop_filter_chroma)(uint8_t *p_pix, ptrdiff_t xstride, ptrdiff_t ystride, int inner_iters, int alpha, int beta, int8_t *tc0)
232 {
233 89419874 pixel *pix = (pixel*)p_pix;
234 int i, d;
235 89419874 alpha <<= BIT_DEPTH - 8;
236 89419874 beta <<= BIT_DEPTH - 8;
237 89419874 xstride >>= sizeof(pixel)-1;
238 89419874 ystride >>= sizeof(pixel)-1;
239
2/2
✓ Branch 0 taken 178839748 times.
✓ Branch 1 taken 44709937 times.
447099370 for( i = 0; i < 4; i++ ) {
240 357679496 const int tc = ((tc0[i] - 1U) << (BIT_DEPTH - 8)) + 1;
241
2/2
✓ Branch 0 taken 23872242 times.
✓ Branch 1 taken 154967506 times.
357679496 if( tc <= 0 ) {
242 47744484 pix += inner_iters*ystride;
243 47744484 continue;
244 }
245
2/2
✓ Branch 0 taken 318401470 times.
✓ Branch 1 taken 154967506 times.
946737952 for( d = 0; d < inner_iters; d++ ) {
246 636802940 const int p0 = pix[-1*xstride];
247 636802940 const int p1 = pix[-2*xstride];
248 636802940 const int q0 = pix[0];
249 636802940 const int q1 = pix[1*xstride];
250
251
2/2
✓ Branch 0 taken 311833973 times.
✓ Branch 1 taken 6567497 times.
636802940 if( FFABS( p0 - q0 ) < alpha &&
252
2/2
✓ Branch 0 taken 293218170 times.
✓ Branch 1 taken 18615803 times.
623667946 FFABS( p1 - p0 ) < beta &&
253
2/2
✓ Branch 0 taken 280765942 times.
✓ Branch 1 taken 12452228 times.
586436340 FFABS( q1 - q0 ) < beta ) {
254
255 561531884 int delta = av_clip( ((q0 - p0) * 4 + (p1 - q1) + 4) >> 3, -tc, tc );
256
257 561531884 pix[-xstride] = av_clip_pixel( p0 + delta ); /* p0' */
258 561531884 pix[0] = av_clip_pixel( q0 - delta ); /* q0' */
259 }
260 636802940 pix += ystride;
261 }
262 }
263 89419874 }
264 48538392 static void FUNCC(h264_v_loop_filter_chroma)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
265 {
266 48538392 FUNCC(h264_loop_filter_chroma)(pix, stride, sizeof(pixel), 2, alpha, beta, tc0);
267 48538392 }
268 36806954 static void FUNCC(h264_h_loop_filter_chroma)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
269 {
270 36806954 FUNCC(h264_loop_filter_chroma)(pix, sizeof(pixel), stride, 2, alpha, beta, tc0);
271 36806954 }
272 1300824 static void FUNCC(h264_h_loop_filter_chroma_mbaff)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
273 {
274 1300824 FUNCC(h264_loop_filter_chroma)(pix, sizeof(pixel), stride, 1, alpha, beta, tc0);
275 1300824 }
276 2773272 static void FUNCC(h264_h_loop_filter_chroma422)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
277 {
278 2773272 FUNCC(h264_loop_filter_chroma)(pix, sizeof(pixel), stride, 4, alpha, beta, tc0);
279 2773272 }
280 432 static void FUNCC(h264_h_loop_filter_chroma422_mbaff)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
281 {
282 432 FUNCC(h264_loop_filter_chroma)(pix, sizeof(pixel), stride, 2, alpha, beta, tc0);
283 432 }
284
285 21984202 static av_always_inline av_flatten void FUNCC(h264_loop_filter_chroma_intra)(uint8_t *p_pix, ptrdiff_t xstride, ptrdiff_t ystride, int inner_iters, int alpha, int beta)
286 {
287 21984202 pixel *pix = (pixel*)p_pix;
288 int d;
289 21984202 xstride >>= sizeof(pixel)-1;
290 21984202 ystride >>= sizeof(pixel)-1;
291 21984202 alpha <<= BIT_DEPTH - 8;
292 21984202 beta <<= BIT_DEPTH - 8;
293
2/2
✓ Branch 0 taken 95806392 times.
✓ Branch 1 taken 10992101 times.
213596986 for( d = 0; d < 4 * inner_iters; d++ ) {
294 191612784 const int p0 = pix[-1*xstride];
295 191612784 const int p1 = pix[-2*xstride];
296 191612784 const int q0 = pix[0];
297 191612784 const int q1 = pix[1*xstride];
298
299
2/2
✓ Branch 0 taken 92926717 times.
✓ Branch 1 taken 2879675 times.
191612784 if( FFABS( p0 - q0 ) < alpha &&
300
2/2
✓ Branch 0 taken 87889693 times.
✓ Branch 1 taken 5037024 times.
185853434 FFABS( p1 - p0 ) < beta &&
301
2/2
✓ Branch 0 taken 84343855 times.
✓ Branch 1 taken 3545838 times.
175779386 FFABS( q1 - q0 ) < beta ) {
302
303 168687710 pix[-xstride] = ( 2*p1 + p0 + q1 + 2 ) >> 2; /* p0' */
304 168687710 pix[0] = ( 2*q1 + q0 + p1 + 2 ) >> 2; /* q0' */
305 }
306 191612784 pix += ystride;
307 }
308 21984202 }
309 9011836 static void FUNCC(h264_v_loop_filter_chroma_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
310 {
311 9011836 FUNCC(h264_loop_filter_chroma_intra)(pix, stride, sizeof(pixel), 2, alpha, beta);
312 9011836 }
313 8970066 static void FUNCC(h264_h_loop_filter_chroma_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
314 {
315 8970066 FUNCC(h264_loop_filter_chroma_intra)(pix, sizeof(pixel), stride, 2, alpha, beta);
316 8970066 }
317 1063536 static void FUNCC(h264_h_loop_filter_chroma_mbaff_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
318 {
319 1063536 FUNCC(h264_loop_filter_chroma_intra)(pix, sizeof(pixel), stride, 1, alpha, beta);
320 1063536 }
321 2499164 static void FUNCC(h264_h_loop_filter_chroma422_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
322 {
323 2499164 FUNCC(h264_loop_filter_chroma_intra)(pix, sizeof(pixel), stride, 4, alpha, beta);
324 2499164 }
325 439600 static void FUNCC(h264_h_loop_filter_chroma422_mbaff_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
326 {
327 439600 FUNCC(h264_loop_filter_chroma_intra)(pix, sizeof(pixel), stride, 2, alpha, beta);
328 439600 }
329