FFmpeg coverage

Directory: ../../../ffmpeg/
File: src/libavcodec/simple_idct_template.c
Date: 2025-03-26 11:21:49
Exec Total Coverage
Lines: 121 121 100.0%
Functions: 22 26 84.6%
Branches: 57 58 98.3%
Line Branch Exec Source
1 /*
2 * Simple IDCT
3 *
4 * Copyright (c) 2001 Michael Niedermayer <michaelni@gmx.at>
5 *
6 * This file is part of FFmpeg.
7 *
8 * FFmpeg is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
12 *
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
17 *
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with FFmpeg; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 */
22
23 /**
24 * @file
25 * simpleidct in C.
26 */
27
28 /* Based upon some commented-out C code from mpeg2dec (idct_mmx.c
29 * written by Aaron Holtzman <aholtzma@ess.engr.uvic.ca>). */
30
31 #include "bit_depth_template.c"
32
33 #undef W1
34 #undef W2
35 #undef W3
36 #undef W4
37 #undef W5
38 #undef W6
39 #undef W7
40 #undef ROW_SHIFT
41 #undef COL_SHIFT
42 #undef DC_SHIFT
43 #undef MUL
44 #undef MAC
45
46 #if BIT_DEPTH == 8
47
48 #define W1 22725 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
49 #define W2 21407 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
50 #define W3 19266 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
51 #define W4 16383 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
52 #define W5 12873 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
53 #define W6 8867 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
54 #define W7 4520 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
55
56 #define ROW_SHIFT 11
57 #define COL_SHIFT 20
58 #define DC_SHIFT 3
59
60 #define MUL(a, b) MUL16(a, b)
61 #define MAC(a, b, c) MAC16(a, b, c)
62
63 #elif BIT_DEPTH == 10 || BIT_DEPTH == 12
64
65 # if BIT_DEPTH == 10
66 #define W1 22725 // 90901
67 #define W2 21407 // 85627
68 #define W3 19265 // 77062
69 #define W4 16384 // 65535
70 #define W5 12873 // 51491
71 #define W6 8867 // 35468
72 #define W7 4520 // 18081
73
74 # ifdef EXTRA_SHIFT
75 #define ROW_SHIFT 13
76 #define COL_SHIFT 18
77 #define DC_SHIFT 1
78 # elif IN_IDCT_DEPTH == 32
79 #define ROW_SHIFT 13
80 #define COL_SHIFT 21
81 #define DC_SHIFT 2
82 # else
83 #define ROW_SHIFT 12
84 #define COL_SHIFT 19
85 #define DC_SHIFT 2
86 # endif
87
88 # else
89 #define W1 45451
90 #define W2 42813
91 #define W3 38531
92 #define W4 32767
93 #define W5 25746
94 #define W6 17734
95 #define W7 9041
96
97 #define ROW_SHIFT 16
98 #define COL_SHIFT 17
99 #define DC_SHIFT -1
100 # endif
101
102 #define MUL(a, b) ((int)((SUINT)(a) * (b)))
103 #define MAC(a, b, c) ((a) += (SUINT)(b) * (c))
104
105 #else
106
107 #error "Unsupported bitdepth"
108
109 #endif
110
111 #ifdef EXTRA_SHIFT
112 21924096 static inline void FUNC(idctRowCondDC_extrashift)(int16_t *row, int extra_shift)
113 #else
114 1477603664 static inline void FUNC6(idctRowCondDC)(idctin *row, int extra_shift)
115 #endif
116 {
117 SUINT a0, a1, a2, a3, b0, b1, b2, b3;
118
119 // TODO: Add DC-only support for int32_t input
120 #if IN_IDCT_DEPTH == 16
121 #if HAVE_FAST_64BIT
122 #define ROW0_MASK (0xffffULL << 48 * HAVE_BIGENDIAN)
123
4/4
✓ Branch 0 taken 547950713 times.
✓ Branch 1 taken 200593087 times.
✓ Branch 2 taken 13064919 times.
✓ Branch 3 taken 8859177 times.
 1499442384 if (((AV_RN64A(row) & ~ROW0_MASK) | AV_RN64A(row+4)) == 0) {
124 uint64_t temp;
125
3/4
✓ Branch 0 taken 536224548 times.
✓ Branch 1 taken 11726165 times.
✗ Branch 2 not taken.
✓ Branch 3 taken 13064919 times.
 1098527950 if (DC_SHIFT - extra_shift >= 0) {
126 1072449096 temp = (row[0] * (1 << (DC_SHIFT - extra_shift))) & 0xffff;
127 } else {
128 26078854 temp = ((row[0] + (1<<(extra_shift - DC_SHIFT-1))) >> (extra_shift - DC_SHIFT)) & 0xffff;
129 }
130 1098527950 temp += temp * (1 << 16);
131 1098527950 temp += temp * ((uint64_t) 1 << 32);
132 1098527950 AV_WN64A(row, temp);
133 1098527950 AV_WN64A(row + 4, temp);
134 1098527950 return;
135 }
136 #else
137 if (!(AV_RN32A(row+2) |
138 AV_RN32A(row+4) |
139 AV_RN32A(row+6) |
140 row[1])) {
141 uint32_t temp;
142 if (DC_SHIFT - extra_shift >= 0) {
143 temp = (row[0] * (1 << (DC_SHIFT - extra_shift))) & 0xffff;
144 } else {
145 temp = ((row[0] + (1<<(extra_shift - DC_SHIFT-1))) >> (extra_shift - DC_SHIFT)) & 0xffff;
146 }
147 temp += temp * (1 << 16);
148 AV_WN32A(row, temp);
149 AV_WN32A(row+2, temp);
150 AV_WN32A(row+4, temp);
151 AV_WN32A(row+6, temp);
152 return;
153 }
154 #endif
155 #endif
156
157 400999810 a0 = ((SUINT)W4 * row[0]) + (1 << (ROW_SHIFT + extra_shift - 1));
158 400999810 a1 = a0;
159 400999810 a2 = a0;
160 400999810 a3 = a0;
161
162 400999810 a0 += (SUINT)W2 * row[2];
163 400999810 a1 += (SUINT)W6 * row[2];
164 400999810 a2 -= (SUINT)W6 * row[2];
165 400999810 a3 -= (SUINT)W2 * row[2];
166
167 400999810 b0 = MUL(W1, row[1]);
168 400999810 MAC(b0, W3, row[3]);
169 400999810 b1 = MUL(W3, row[1]);
170 400999810 MAC(b1, -W7, row[3]);
171 400999810 b2 = MUL(W5, row[1]);
172 400999810 MAC(b2, -W1, row[3]);
173 400999810 b3 = MUL(W7, row[1]);
174 400999810 MAC(b3, -W5, row[3]);
175
176 #if IN_IDCT_DEPTH == 32
177
2/2
✓ Branch 0 taken 4223 times.
✓ Branch 1 taken 38465 times.
 85376 if (AV_RN64A(row + 4) | AV_RN64A(row + 6)) {
178 #else
179
4/4
✓ Branch 0 taken 82280698 times.
✓ Branch 1 taken 118312389 times.
✓ Branch 2 taken 6710882 times.
✓ Branch 3 taken 2148295 times.
 400914434 if (AV_RN64A(row + 4)) {
180 #endif
181 165760028 a0 += (SUINT) W4*row[4] + (SUINT)W6*row[6];
182 165760028 a1 += (SUINT)- W4*row[4] - (SUINT)W2*row[6];
183 165760028 a2 += (SUINT)- W4*row[4] + (SUINT)W2*row[6];
184 165760028 a3 += (SUINT) W4*row[4] - (SUINT)W6*row[6];
185
186 165760028 MAC(b0, W5, row[5]);
187 165760028 MAC(b0, W7, row[7]);
188
189 165760028 MAC(b1, -W1, row[5]);
190 165760028 MAC(b1, -W5, row[7]);
191
192 165760028 MAC(b2, W7, row[5]);
193 165760028 MAC(b2, W3, row[7]);
194
195 165760028 MAC(b3, W3, row[5]);
196 165760028 MAC(b3, -W1, row[7]);
197 }
198
199 400999810 row[0] = (int)(a0 + b0) >> (ROW_SHIFT + extra_shift);
200 400999810 row[7] = (int)(a0 - b0) >> (ROW_SHIFT + extra_shift);
201 400999810 row[1] = (int)(a1 + b1) >> (ROW_SHIFT + extra_shift);
202 400999810 row[6] = (int)(a1 - b1) >> (ROW_SHIFT + extra_shift);
203 400999810 row[2] = (int)(a2 + b2) >> (ROW_SHIFT + extra_shift);
204 400999810 row[5] = (int)(a2 - b2) >> (ROW_SHIFT + extra_shift);
205 400999810 row[3] = (int)(a3 + b3) >> (ROW_SHIFT + extra_shift);
206 400999810 row[4] = (int)(a3 - b3) >> (ROW_SHIFT + extra_shift);
207 85376 }
208
209 #define IDCT_COLS do { \
210 a0 = (SUINT)W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4)); \
211 a1 = a0; \
212 a2 = a0; \
213 a3 = a0; \
214 \
215 a0 += (SUINT) W2*col[8*2]; \
216 a1 += (SUINT) W6*col[8*2]; \
217 a2 += (SUINT)-W6*col[8*2]; \
218 a3 += (SUINT)-W2*col[8*2]; \
219 \
220 b0 = MUL(W1, col[8*1]); \
221 b1 = MUL(W3, col[8*1]); \
222 b2 = MUL(W5, col[8*1]); \
223 b3 = MUL(W7, col[8*1]); \
224 \
225 MAC(b0, W3, col[8*3]); \
226 MAC(b1, -W7, col[8*3]); \
227 MAC(b2, -W1, col[8*3]); \
228 MAC(b3, -W5, col[8*3]); \
229 \
230 if (col[8*4]) { \
231 a0 += (SUINT) W4*col[8*4]; \
232 a1 += (SUINT)-W4*col[8*4]; \
233 a2 += (SUINT)-W4*col[8*4]; \
234 a3 += (SUINT) W4*col[8*4]; \
235 } \
236 \
237 if (col[8*5]) { \
238 MAC(b0, W5, col[8*5]); \
239 MAC(b1, -W1, col[8*5]); \
240 MAC(b2, W7, col[8*5]); \
241 MAC(b3, W3, col[8*5]); \
242 } \
243 \
244 if (col[8*6]) { \
245 a0 += (SUINT) W6*col[8*6]; \
246 a1 += (SUINT)-W2*col[8*6]; \
247 a2 += (SUINT) W2*col[8*6]; \
248 a3 += (SUINT)-W6*col[8*6]; \
249 } \
250 \
251 if (col[8*7]) { \
252 MAC(b0, W7, col[8*7]); \
253 MAC(b1, -W5, col[8*7]); \
254 MAC(b2, W3, col[8*7]); \
255 MAC(b3, -W1, col[8*7]); \
256 } \
257 } while (0)
258
259 #ifdef EXTRA_SHIFT
260 21924096 static inline void FUNC(idctSparseCol_extrashift)(int16_t *col)
261 #else
262 84597728 static inline void FUNC6(idctSparseCol)(idctin *col)
263 #endif
264 {
265 unsigned a0, a1, a2, a3, b0, b1, b2, b3;
266
267
16/16
✓ Branch 0 taken 17129675 times.
✓ Branch 1 taken 34953845 times.
✓ Branch 2 taken 15340586 times.
✓ Branch 3 taken 36742934 times.
✓ Branch 4 taken 12208678 times.
✓ Branch 5 taken 39874842 times.
✓ Branch 6 taken 10910348 times.
✓ Branch 7 taken 41173172 times.
✓ Branch 8 taken 8566104 times.
✓ Branch 9 taken 13357992 times.
✓ Branch 10 taken 8394116 times.
✓ Branch 11 taken 13529980 times.
✓ Branch 12 taken 6694259 times.
✓ Branch 13 taken 15229837 times.
✓ Branch 14 taken 5669425 times.
✓ Branch 15 taken 16254671 times.
 106521824 IDCT_COLS;
268
269 106521824 col[0 ] = ((int)(a0 + b0) >> COL_SHIFT);
270 106521824 col[8 ] = ((int)(a1 + b1) >> COL_SHIFT);
271 106521824 col[16] = ((int)(a2 + b2) >> COL_SHIFT);
272 106521824 col[24] = ((int)(a3 + b3) >> COL_SHIFT);
273 106521824 col[32] = ((int)(a3 - b3) >> COL_SHIFT);
274 106521824 col[40] = ((int)(a2 - b2) >> COL_SHIFT);
275 106521824 col[48] = ((int)(a1 - b1) >> COL_SHIFT);
276 106521824 col[56] = ((int)(a0 - b0) >> COL_SHIFT);
277 106521824 }
278
279 #ifndef PRORES_ONLY
280 #ifndef EXTRA_SHIFT
281 1117527008 static inline void FUNC6(idctSparseColPut)(pixel *dest, ptrdiff_t line_size,
282 idctin *col)
283 {
284 SUINT a0, a1, a2, a3, b0, b1, b2, b3;
285
286
8/8
✓ Branch 0 taken 119008792 times.
✓ Branch 1 taken 439754712 times.
✓ Branch 2 taken 112751000 times.
✓ Branch 3 taken 446012504 times.
✓ Branch 4 taken 64421727 times.
✓ Branch 5 taken 494341777 times.
✓ Branch 6 taken 74543527 times.
✓ Branch 7 taken 484219977 times.
 1117527008 IDCT_COLS;
287
288 1117527008 dest[0] = av_clip_pixel((int)(a0 + b0) >> COL_SHIFT);
289 1117527008 dest += line_size;
290 1117527008 dest[0] = av_clip_pixel((int)(a1 + b1) >> COL_SHIFT);
291 1117527008 dest += line_size;
292 1117527008 dest[0] = av_clip_pixel((int)(a2 + b2) >> COL_SHIFT);
293 1117527008 dest += line_size;
294 1117527008 dest[0] = av_clip_pixel((int)(a3 + b3) >> COL_SHIFT);
295 1117527008 dest += line_size;
296 1117527008 dest[0] = av_clip_pixel((int)(a3 - b3) >> COL_SHIFT);
297 1117527008 dest += line_size;
298 1117527008 dest[0] = av_clip_pixel((int)(a2 - b2) >> COL_SHIFT);
299 1117527008 dest += line_size;
300 1117527008 dest[0] = av_clip_pixel((int)(a1 - b1) >> COL_SHIFT);
301 1117527008 dest += line_size;
302 1117527008 dest[0] = av_clip_pixel((int)(a0 - b0) >> COL_SHIFT);
303 1117527008 }
304
305 275013856 static inline void FUNC6(idctSparseColAdd)(pixel *dest, ptrdiff_t line_size,
306 idctin *col)
307 {
308 unsigned a0, a1, a2, a3, b0, b1, b2, b3;
309
310
8/8
✓ Branch 0 taken 40263037 times.
✓ Branch 1 taken 97243891 times.
✓ Branch 2 taken 33919159 times.
✓ Branch 3 taken 103587769 times.
✓ Branch 4 taken 28101566 times.
✓ Branch 5 taken 109405362 times.
✓ Branch 6 taken 62707427 times.
✓ Branch 7 taken 74799501 times.
 275013856 IDCT_COLS;
311
312 275013856 dest[0] = av_clip_pixel(dest[0] + ((int)(a0 + b0) >> COL_SHIFT));
313 275013856 dest += line_size;
314 275013856 dest[0] = av_clip_pixel(dest[0] + ((int)(a1 + b1) >> COL_SHIFT));
315 275013856 dest += line_size;
316 275013856 dest[0] = av_clip_pixel(dest[0] + ((int)(a2 + b2) >> COL_SHIFT));
317 275013856 dest += line_size;
318 275013856 dest[0] = av_clip_pixel(dest[0] + ((int)(a3 + b3) >> COL_SHIFT));
319 275013856 dest += line_size;
320 275013856 dest[0] = av_clip_pixel(dest[0] + ((int)(a3 - b3) >> COL_SHIFT));
321 275013856 dest += line_size;
322 275013856 dest[0] = av_clip_pixel(dest[0] + ((int)(a2 - b2) >> COL_SHIFT));
323 275013856 dest += line_size;
324 275013856 dest[0] = av_clip_pixel(dest[0] + ((int)(a1 - b1) >> COL_SHIFT));
325 275013856 dest += line_size;
326 275013856 dest[0] = av_clip_pixel(dest[0] + ((int)(a0 - b0) >> COL_SHIFT));
327 275013856 }
328
329 139690876 void FUNC6(ff_simple_idct_put)(uint8_t *dest_, ptrdiff_t line_size, int16_t *block_)
330 {
331 139690876 idctin *block = (idctin *)block_;
332 139690876 pixel *dest = (pixel *)dest_;
333 int i;
334
335 139690876 line_size /= sizeof(pixel);
336
337
2/2
✓ Branch 0 taken 558763504 times.
✓ Branch 1 taken 69845438 times.
 1257217884 for (i = 0; i < 8; i++)
338 1117527008 FUNC6(idctRowCondDC)(block + i*8, 0);
339
340
2/2
✓ Branch 0 taken 558763504 times.
✓ Branch 1 taken 69845438 times.
 1257217884 for (i = 0; i < 8; i++)
341 1117527008 FUNC6(idctSparseColPut)(dest + i, line_size, block + i);
342 139690876 }
343
344 #if IN_IDCT_DEPTH == 16
345 34346776 void FUNC6(ff_simple_idct_add)(uint8_t *dest_, ptrdiff_t line_size, int16_t *block)
346 {
347 34346776 pixel *dest = (pixel *)dest_;
348 int i;
349
350 34346776 line_size /= sizeof(pixel);
351
352
2/2
✓ Branch 0 taken 137387104 times.
✓ Branch 1 taken 17173388 times.
 309120984 for (i = 0; i < 8; i++)
353 274774208 FUNC6(idctRowCondDC)(block + i*8, 0);
354
355
2/2
✓ Branch 0 taken 137387104 times.
✓ Branch 1 taken 17173388 times.
 309120984 for (i = 0; i < 8; i++)
356 274774208 FUNC6(idctSparseColAdd)(dest + i, line_size, block + i);
357 34346776 }
358
359 8128552 void FUNC6(ff_simple_idct)(int16_t *block)
360 {
361 int i;
362
363
2/2
✓ Branch 0 taken 32514208 times.
✓ Branch 1 taken 4064276 times.
 73156968 for (i = 0; i < 8; i++)
364 65028416 FUNC6(idctRowCondDC)(block + i*8, 0);
365
366
2/2
✓ Branch 0 taken 32514208 times.
✓ Branch 1 taken 4064276 times.
 73156968 for (i = 0; i < 8; i++)
367 65028416 FUNC6(idctSparseCol)(block + i);
368 8128552 }
369 #endif
370 #endif
371 #endif /* PRORES_ONLY */
372