Details | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 6148 | serge | 1 | /* |
| 2 | * (c) 2002 Fabrice Bellard |
||
| 3 | * |
||
| 4 | * This file is part of FFmpeg. |
||
| 5 | * |
||
| 6 | * FFmpeg is free software; you can redistribute it and/or |
||
| 7 | * modify it under the terms of the GNU Lesser General Public |
||
| 8 | * License as published by the Free Software Foundation; either |
||
| 9 | * version 2.1 of the License, or (at your option) any later version. |
||
| 10 | * |
||
| 11 | * FFmpeg is distributed in the hope that it will be useful, |
||
| 12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
||
| 13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
||
| 14 | * Lesser General Public License for more details. |
||
| 15 | * |
||
| 16 | * You should have received a copy of the GNU Lesser General Public |
||
| 17 | * License along with FFmpeg; if not, write to the Free Software |
||
| 18 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
||
| 19 | */ |
||
| 20 | |||
| 21 | /** |
||
| 22 | * @file |
||
| 23 | * FFT and MDCT tests. |
||
| 24 | */ |
||
| 25 | |||
| 26 | #include "libavutil/cpu.h" |
||
| 27 | #include "libavutil/mathematics.h" |
||
| 28 | #include "libavutil/lfg.h" |
||
| 29 | #include "libavutil/log.h" |
||
| 30 | #include "libavutil/time.h" |
||
| 31 | #include "fft.h" |
||
| 32 | #if CONFIG_FFT_FLOAT |
||
| 33 | #include "dct.h" |
||
| 34 | #include "rdft.h" |
||
| 35 | #endif |
||
| 36 | #include |
||
| 37 | #if HAVE_UNISTD_H |
||
| 38 | #include |
||
| 39 | #endif |
||
| 40 | #include |
||
| 41 | #include |
||
| 42 | #include |
||
| 43 | |||
| 44 | /* reference fft */ |
||
| 45 | |||
| 46 | #define MUL16(a,b) ((a) * (b)) |
||
| 47 | |||
| 48 | #define CMAC(pre, pim, are, aim, bre, bim) \ |
||
| 49 | {\ |
||
| 50 | pre += (MUL16(are, bre) - MUL16(aim, bim));\ |
||
| 51 | pim += (MUL16(are, bim) + MUL16(bre, aim));\ |
||
| 52 | } |
||
| 53 | |||
| 54 | #if CONFIG_FFT_FLOAT |
||
| 55 | # define RANGE 1.0 |
||
| 56 | # define REF_SCALE(x, bits) (x) |
||
| 57 | # define FMT "%10.6f" |
||
| 58 | #elif CONFIG_FFT_FIXED_32 |
||
| 59 | # define RANGE 8388608 |
||
| 60 | # define REF_SCALE(x, bits) (x) |
||
| 61 | # define FMT "%6d" |
||
| 62 | #else |
||
| 63 | # define RANGE 16384 |
||
| 64 | # define REF_SCALE(x, bits) ((x) / (1<<(bits))) |
||
| 65 | # define FMT "%6d" |
||
| 66 | #endif |
||
| 67 | |||
| 68 | struct { |
||
| 69 | float re, im; |
||
| 70 | } *exptab; |
||
| 71 | |||
| 72 | static void fft_ref_init(int nbits, int inverse) |
||
| 73 | { |
||
| 74 | int n, i; |
||
| 75 | double c1, s1, alpha; |
||
| 76 | |||
| 77 | n = 1 << nbits; |
||
| 78 | exptab = av_malloc((n / 2) * sizeof(*exptab)); |
||
| 79 | |||
| 80 | for (i = 0; i < (n/2); i++) { |
||
| 81 | alpha = 2 * M_PI * (float)i / (float)n; |
||
| 82 | c1 = cos(alpha); |
||
| 83 | s1 = sin(alpha); |
||
| 84 | if (!inverse) |
||
| 85 | s1 = -s1; |
||
| 86 | exptab[i].re = c1; |
||
| 87 | exptab[i].im = s1; |
||
| 88 | } |
||
| 89 | } |
||
| 90 | |||
| 91 | static void fft_ref(FFTComplex *tabr, FFTComplex *tab, int nbits) |
||
| 92 | { |
||
| 93 | int n, i, j, k, n2; |
||
| 94 | double tmp_re, tmp_im, s, c; |
||
| 95 | FFTComplex *q; |
||
| 96 | |||
| 97 | n = 1 << nbits; |
||
| 98 | n2 = n >> 1; |
||
| 99 | for (i = 0; i < n; i++) { |
||
| 100 | tmp_re = 0; |
||
| 101 | tmp_im = 0; |
||
| 102 | q = tab; |
||
| 103 | for (j = 0; j < n; j++) { |
||
| 104 | k = (i * j) & (n - 1); |
||
| 105 | if (k >= n2) { |
||
| 106 | c = -exptab[k - n2].re; |
||
| 107 | s = -exptab[k - n2].im; |
||
| 108 | } else { |
||
| 109 | c = exptab[k].re; |
||
| 110 | s = exptab[k].im; |
||
| 111 | } |
||
| 112 | CMAC(tmp_re, tmp_im, c, s, q->re, q->im); |
||
| 113 | q++; |
||
| 114 | } |
||
| 115 | tabr[i].re = REF_SCALE(tmp_re, nbits); |
||
| 116 | tabr[i].im = REF_SCALE(tmp_im, nbits); |
||
| 117 | } |
||
| 118 | } |
||
| 119 | |||
| 120 | static void imdct_ref(FFTSample *out, FFTSample *in, int nbits) |
||
| 121 | { |
||
| 122 | int n = 1< |
||
| 123 | int k, i, a; |
||
| 124 | double sum, f; |
||
| 125 | |||
| 126 | for (i = 0; i < n; i++) { |
||
| 127 | sum = 0; |
||
| 128 | for (k = 0; k < n/2; k++) { |
||
| 129 | a = (2 * i + 1 + (n / 2)) * (2 * k + 1); |
||
| 130 | f = cos(M_PI * a / (double)(2 * n)); |
||
| 131 | sum += f * in[k]; |
||
| 132 | } |
||
| 133 | out[i] = REF_SCALE(-sum, nbits - 2); |
||
| 134 | } |
||
| 135 | } |
||
| 136 | |||
| 137 | /* NOTE: no normalisation by 1 / N is done */ |
||
| 138 | static void mdct_ref(FFTSample *output, FFTSample *input, int nbits) |
||
| 139 | { |
||
| 140 | int n = 1< |
||
| 141 | int k, i; |
||
| 142 | double a, s; |
||
| 143 | |||
| 144 | /* do it by hand */ |
||
| 145 | for (k = 0; k < n/2; k++) { |
||
| 146 | s = 0; |
||
| 147 | for (i = 0; i < n; i++) { |
||
| 148 | a = (2*M_PI*(2*i+1+n/2)*(2*k+1) / (4 * n)); |
||
| 149 | s += input[i] * cos(a); |
||
| 150 | } |
||
| 151 | output[k] = REF_SCALE(s, nbits - 1); |
||
| 152 | } |
||
| 153 | } |
||
| 154 | |||
| 155 | #if CONFIG_FFT_FLOAT |
||
| 156 | static void idct_ref(FFTSample *output, FFTSample *input, int nbits) |
||
| 157 | { |
||
| 158 | int n = 1< |
||
| 159 | int k, i; |
||
| 160 | double a, s; |
||
| 161 | |||
| 162 | /* do it by hand */ |
||
| 163 | for (i = 0; i < n; i++) { |
||
| 164 | s = 0.5 * input[0]; |
||
| 165 | for (k = 1; k < n; k++) { |
||
| 166 | a = M_PI*k*(i+0.5) / n; |
||
| 167 | s += input[k] * cos(a); |
||
| 168 | } |
||
| 169 | output[i] = 2 * s / n; |
||
| 170 | } |
||
| 171 | } |
||
| 172 | static void dct_ref(FFTSample *output, FFTSample *input, int nbits) |
||
| 173 | { |
||
| 174 | int n = 1< |
||
| 175 | int k, i; |
||
| 176 | double a, s; |
||
| 177 | |||
| 178 | /* do it by hand */ |
||
| 179 | for (k = 0; k < n; k++) { |
||
| 180 | s = 0; |
||
| 181 | for (i = 0; i < n; i++) { |
||
| 182 | a = M_PI*k*(i+0.5) / n; |
||
| 183 | s += input[i] * cos(a); |
||
| 184 | } |
||
| 185 | output[k] = s; |
||
| 186 | } |
||
| 187 | } |
||
| 188 | #endif |
||
| 189 | |||
| 190 | |||
| 191 | static FFTSample frandom(AVLFG *prng) |
||
| 192 | { |
||
| 193 | return (int16_t)av_lfg_get(prng) / 32768.0 * RANGE; |
||
| 194 | } |
||
| 195 | |||
| 196 | static int check_diff(FFTSample *tab1, FFTSample *tab2, int n, double scale) |
||
| 197 | { |
||
| 198 | int i; |
||
| 199 | double max= 0; |
||
| 200 | double error= 0; |
||
| 201 | int err = 0; |
||
| 202 | |||
| 203 | for (i = 0; i < n; i++) { |
||
| 204 | double e = fabsf(tab1[i] - (tab2[i] / scale)) / RANGE; |
||
| 205 | if (e >= 1e-3) { |
||
| 206 | av_log(NULL, AV_LOG_ERROR, "ERROR %5d: "FMT" "FMT"\n", |
||
| 207 | i, tab1[i], tab2[i]); |
||
| 208 | err = 1; |
||
| 209 | } |
||
| 210 | error+= e*e; |
||
| 211 | if(e>max) max= e; |
||
| 212 | } |
||
| 213 | av_log(NULL, AV_LOG_INFO, "max:%f e:%g\n", max, sqrt(error/n)); |
||
| 214 | return err; |
||
| 215 | } |
||
| 216 | |||
| 217 | |||
| 218 | static void help(void) |
||
| 219 | { |
||
| 220 | av_log(NULL, AV_LOG_INFO,"usage: fft-test [-h] [-s] [-i] [-n b]\n" |
||
| 221 | "-h print this help\n" |
||
| 222 | "-s speed test\n" |
||
| 223 | "-m (I)MDCT test\n" |
||
| 224 | "-d (I)DCT test\n" |
||
| 225 | "-r (I)RDFT test\n" |
||
| 226 | "-i inverse transform test\n" |
||
| 227 | "-n b set the transform size to 2^b\n" |
||
| 228 | "-f x set scale factor for output data of (I)MDCT to x\n" |
||
| 229 | ); |
||
| 230 | } |
||
| 231 | |||
| 232 | enum tf_transform { |
||
| 233 | TRANSFORM_FFT, |
||
| 234 | TRANSFORM_MDCT, |
||
| 235 | TRANSFORM_RDFT, |
||
| 236 | TRANSFORM_DCT, |
||
| 237 | }; |
||
| 238 | |||
| 239 | #if !HAVE_GETOPT |
||
| 240 | #include "compat/getopt.c" |
||
| 241 | #endif |
||
| 242 | |||
| 243 | int main(int argc, char **argv) |
||
| 244 | { |
||
| 245 | FFTComplex *tab, *tab1, *tab_ref; |
||
| 246 | FFTSample *tab2; |
||
| 247 | int it, i, c; |
||
| 248 | int cpuflags; |
||
| 249 | int do_speed = 0; |
||
| 250 | int err = 1; |
||
| 251 | enum tf_transform transform = TRANSFORM_FFT; |
||
| 252 | int do_inverse = 0; |
||
| 253 | FFTContext s1, *s = &s1; |
||
| 254 | FFTContext m1, *m = &m1; |
||
| 255 | #if CONFIG_FFT_FLOAT |
||
| 256 | RDFTContext r1, *r = &r1; |
||
| 257 | DCTContext d1, *d = &d1; |
||
| 258 | int fft_size_2; |
||
| 259 | #endif |
||
| 260 | int fft_nbits, fft_size; |
||
| 261 | double scale = 1.0; |
||
| 262 | AVLFG prng; |
||
| 263 | av_lfg_init(&prng, 1); |
||
| 264 | |||
| 265 | fft_nbits = 9; |
||
| 266 | for(;;) { |
||
| 267 | c = getopt(argc, argv, "hsimrdn:f:c:"); |
||
| 268 | if (c == -1) |
||
| 269 | break; |
||
| 270 | switch(c) { |
||
| 271 | case 'h': |
||
| 272 | help(); |
||
| 273 | return 1; |
||
| 274 | case 's': |
||
| 275 | do_speed = 1; |
||
| 276 | break; |
||
| 277 | case 'i': |
||
| 278 | do_inverse = 1; |
||
| 279 | break; |
||
| 280 | case 'm': |
||
| 281 | transform = TRANSFORM_MDCT; |
||
| 282 | break; |
||
| 283 | case 'r': |
||
| 284 | transform = TRANSFORM_RDFT; |
||
| 285 | break; |
||
| 286 | case 'd': |
||
| 287 | transform = TRANSFORM_DCT; |
||
| 288 | break; |
||
| 289 | case 'n': |
||
| 290 | fft_nbits = atoi(optarg); |
||
| 291 | break; |
||
| 292 | case 'f': |
||
| 293 | scale = atof(optarg); |
||
| 294 | break; |
||
| 295 | case 'c': |
||
| 296 | cpuflags = av_get_cpu_flags(); |
||
| 297 | |||
| 298 | if (av_parse_cpu_caps(&cpuflags, optarg) < 0) |
||
| 299 | return 1; |
||
| 300 | |||
| 301 | av_force_cpu_flags(cpuflags); |
||
| 302 | break; |
||
| 303 | } |
||
| 304 | } |
||
| 305 | |||
| 306 | fft_size = 1 << fft_nbits; |
||
| 307 | tab = av_malloc(fft_size * sizeof(FFTComplex)); |
||
| 308 | tab1 = av_malloc(fft_size * sizeof(FFTComplex)); |
||
| 309 | tab_ref = av_malloc(fft_size * sizeof(FFTComplex)); |
||
| 310 | tab2 = av_malloc(fft_size * sizeof(FFTSample)); |
||
| 311 | |||
| 312 | switch (transform) { |
||
| 313 | case TRANSFORM_MDCT: |
||
| 314 | av_log(NULL, AV_LOG_INFO,"Scale factor is set to %f\n", scale); |
||
| 315 | if (do_inverse) |
||
| 316 | av_log(NULL, AV_LOG_INFO,"IMDCT"); |
||
| 317 | else |
||
| 318 | av_log(NULL, AV_LOG_INFO,"MDCT"); |
||
| 319 | ff_mdct_init(m, fft_nbits, do_inverse, scale); |
||
| 320 | break; |
||
| 321 | case TRANSFORM_FFT: |
||
| 322 | if (do_inverse) |
||
| 323 | av_log(NULL, AV_LOG_INFO,"IFFT"); |
||
| 324 | else |
||
| 325 | av_log(NULL, AV_LOG_INFO,"FFT"); |
||
| 326 | ff_fft_init(s, fft_nbits, do_inverse); |
||
| 327 | fft_ref_init(fft_nbits, do_inverse); |
||
| 328 | break; |
||
| 329 | #if CONFIG_FFT_FLOAT |
||
| 330 | case TRANSFORM_RDFT: |
||
| 331 | if (do_inverse) |
||
| 332 | av_log(NULL, AV_LOG_INFO,"IDFT_C2R"); |
||
| 333 | else |
||
| 334 | av_log(NULL, AV_LOG_INFO,"DFT_R2C"); |
||
| 335 | ff_rdft_init(r, fft_nbits, do_inverse ? IDFT_C2R : DFT_R2C); |
||
| 336 | fft_ref_init(fft_nbits, do_inverse); |
||
| 337 | break; |
||
| 338 | # if CONFIG_DCT |
||
| 339 | case TRANSFORM_DCT: |
||
| 340 | if (do_inverse) |
||
| 341 | av_log(NULL, AV_LOG_INFO,"DCT_III"); |
||
| 342 | else |
||
| 343 | av_log(NULL, AV_LOG_INFO,"DCT_II"); |
||
| 344 | ff_dct_init(d, fft_nbits, do_inverse ? DCT_III : DCT_II); |
||
| 345 | break; |
||
| 346 | # endif |
||
| 347 | #endif |
||
| 348 | default: |
||
| 349 | av_log(NULL, AV_LOG_ERROR, "Requested transform not supported\n"); |
||
| 350 | return 1; |
||
| 351 | } |
||
| 352 | av_log(NULL, AV_LOG_INFO," %d test\n", fft_size); |
||
| 353 | |||
| 354 | /* generate random data */ |
||
| 355 | |||
| 356 | for (i = 0; i < fft_size; i++) { |
||
| 357 | tab1[i].re = frandom(&prng); |
||
| 358 | tab1[i].im = frandom(&prng); |
||
| 359 | } |
||
| 360 | |||
| 361 | /* checking result */ |
||
| 362 | av_log(NULL, AV_LOG_INFO,"Checking...\n"); |
||
| 363 | |||
| 364 | switch (transform) { |
||
| 365 | case TRANSFORM_MDCT: |
||
| 366 | if (do_inverse) { |
||
| 367 | imdct_ref((FFTSample *)tab_ref, (FFTSample *)tab1, fft_nbits); |
||
| 368 | m->imdct_calc(m, tab2, (FFTSample *)tab1); |
||
| 369 | err = check_diff((FFTSample *)tab_ref, tab2, fft_size, scale); |
||
| 370 | } else { |
||
| 371 | mdct_ref((FFTSample *)tab_ref, (FFTSample *)tab1, fft_nbits); |
||
| 372 | |||
| 373 | m->mdct_calc(m, tab2, (FFTSample *)tab1); |
||
| 374 | |||
| 375 | err = check_diff((FFTSample *)tab_ref, tab2, fft_size / 2, scale); |
||
| 376 | } |
||
| 377 | break; |
||
| 378 | case TRANSFORM_FFT: |
||
| 379 | memcpy(tab, tab1, fft_size * sizeof(FFTComplex)); |
||
| 380 | s->fft_permute(s, tab); |
||
| 381 | s->fft_calc(s, tab); |
||
| 382 | |||
| 383 | fft_ref(tab_ref, tab1, fft_nbits); |
||
| 384 | err = check_diff((FFTSample *)tab_ref, (FFTSample *)tab, fft_size * 2, 1.0); |
||
| 385 | break; |
||
| 386 | #if CONFIG_FFT_FLOAT |
||
| 387 | case TRANSFORM_RDFT: |
||
| 388 | fft_size_2 = fft_size >> 1; |
||
| 389 | if (do_inverse) { |
||
| 390 | tab1[ 0].im = 0; |
||
| 391 | tab1[fft_size_2].im = 0; |
||
| 392 | for (i = 1; i < fft_size_2; i++) { |
||
| 393 | tab1[fft_size_2+i].re = tab1[fft_size_2-i].re; |
||
| 394 | tab1[fft_size_2+i].im = -tab1[fft_size_2-i].im; |
||
| 395 | } |
||
| 396 | |||
| 397 | memcpy(tab2, tab1, fft_size * sizeof(FFTSample)); |
||
| 398 | tab2[1] = tab1[fft_size_2].re; |
||
| 399 | |||
| 400 | r->rdft_calc(r, tab2); |
||
| 401 | fft_ref(tab_ref, tab1, fft_nbits); |
||
| 402 | for (i = 0; i < fft_size; i++) { |
||
| 403 | tab[i].re = tab2[i]; |
||
| 404 | tab[i].im = 0; |
||
| 405 | } |
||
| 406 | err = check_diff((float *)tab_ref, (float *)tab, fft_size * 2, 0.5); |
||
| 407 | } else { |
||
| 408 | for (i = 0; i < fft_size; i++) { |
||
| 409 | tab2[i] = tab1[i].re; |
||
| 410 | tab1[i].im = 0; |
||
| 411 | } |
||
| 412 | r->rdft_calc(r, tab2); |
||
| 413 | fft_ref(tab_ref, tab1, fft_nbits); |
||
| 414 | tab_ref[0].im = tab_ref[fft_size_2].re; |
||
| 415 | err = check_diff((float *)tab_ref, (float *)tab2, fft_size, 1.0); |
||
| 416 | } |
||
| 417 | break; |
||
| 418 | case TRANSFORM_DCT: |
||
| 419 | memcpy(tab, tab1, fft_size * sizeof(FFTComplex)); |
||
| 420 | d->dct_calc(d, (FFTSample *)tab); |
||
| 421 | if (do_inverse) { |
||
| 422 | idct_ref((FFTSample*)tab_ref, (FFTSample *)tab1, fft_nbits); |
||
| 423 | } else { |
||
| 424 | dct_ref((FFTSample*)tab_ref, (FFTSample *)tab1, fft_nbits); |
||
| 425 | } |
||
| 426 | err = check_diff((float *)tab_ref, (float *)tab, fft_size, 1.0); |
||
| 427 | break; |
||
| 428 | #endif |
||
| 429 | } |
||
| 430 | |||
| 431 | /* do a speed test */ |
||
| 432 | |||
| 433 | if (do_speed) { |
||
| 434 | int64_t time_start, duration; |
||
| 435 | int nb_its; |
||
| 436 | |||
| 437 | av_log(NULL, AV_LOG_INFO,"Speed test...\n"); |
||
| 438 | /* we measure during about 1 seconds */ |
||
| 439 | nb_its = 1; |
||
| 440 | for(;;) { |
||
| 441 | time_start = av_gettime(); |
||
| 442 | for (it = 0; it < nb_its; it++) { |
||
| 443 | switch (transform) { |
||
| 444 | case TRANSFORM_MDCT: |
||
| 445 | if (do_inverse) { |
||
| 446 | m->imdct_calc(m, (FFTSample *)tab, (FFTSample *)tab1); |
||
| 447 | } else { |
||
| 448 | m->mdct_calc(m, (FFTSample *)tab, (FFTSample *)tab1); |
||
| 449 | } |
||
| 450 | break; |
||
| 451 | case TRANSFORM_FFT: |
||
| 452 | memcpy(tab, tab1, fft_size * sizeof(FFTComplex)); |
||
| 453 | s->fft_calc(s, tab); |
||
| 454 | break; |
||
| 455 | #if CONFIG_FFT_FLOAT |
||
| 456 | case TRANSFORM_RDFT: |
||
| 457 | memcpy(tab2, tab1, fft_size * sizeof(FFTSample)); |
||
| 458 | r->rdft_calc(r, tab2); |
||
| 459 | break; |
||
| 460 | case TRANSFORM_DCT: |
||
| 461 | memcpy(tab2, tab1, fft_size * sizeof(FFTSample)); |
||
| 462 | d->dct_calc(d, tab2); |
||
| 463 | break; |
||
| 464 | #endif |
||
| 465 | } |
||
| 466 | } |
||
| 467 | duration = av_gettime() - time_start; |
||
| 468 | if (duration >= 1000000) |
||
| 469 | break; |
||
| 470 | nb_its *= 2; |
||
| 471 | } |
||
| 472 | av_log(NULL, AV_LOG_INFO,"time: %0.1f us/transform [total time=%0.2f s its=%d]\n", |
||
| 473 | (double)duration / nb_its, |
||
| 474 | (double)duration / 1000000.0, |
||
| 475 | nb_its); |
||
| 476 | } |
||
| 477 | |||
| 478 | switch (transform) { |
||
| 479 | case TRANSFORM_MDCT: |
||
| 480 | ff_mdct_end(m); |
||
| 481 | break; |
||
| 482 | case TRANSFORM_FFT: |
||
| 483 | ff_fft_end(s); |
||
| 484 | break; |
||
| 485 | #if CONFIG_FFT_FLOAT |
||
| 486 | case TRANSFORM_RDFT: |
||
| 487 | ff_rdft_end(r); |
||
| 488 | break; |
||
| 489 | # if CONFIG_DCT |
||
| 490 | case TRANSFORM_DCT: |
||
| 491 | ff_dct_end(d); |
||
| 492 | break; |
||
| 493 | # endif |
||
| 494 | #endif |
||
| 495 | } |
||
| 496 | |||
| 497 | av_free(tab); |
||
| 498 | av_free(tab1); |
||
| 499 | av_free(tab2); |
||
| 500 | av_free(tab_ref); |
||
| 501 | av_free(exptab); |
||
| 502 | |||
| 503 | if (err) |
||
| 504 | printf("Error: %d.\n", err); |
||
| 505 | |||
| 506 | return !!err; |
||
| 507 | } |