1 /* 2 * Copyright (C) 2011-2013 Michael Niedermayer (michaelni@gmx.at) 3 * 4 * This file is part of libswresample 5 * 6 * libswresample 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 * libswresample 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 libswresample; if not, write to the Free Software 18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA 19 */ 20 21 module ffmpeg.libswresample.swresample; 22 23 extern (C): 24 import ffmpeg; @nogc nothrow: 25 26 /** 27 * @file 28 * @ingroup lswr 29 * libswresample public header 30 */ 31 32 /** 33 * @defgroup lswr libswresample 34 * @{ 35 * 36 * Audio resampling, sample format conversion and mixing library. 37 * 38 * Interaction with lswr is done through SwrContext, which is 39 * allocated with swr_alloc() or swr_alloc_set_opts(). It is opaque, so all parameters 40 * must be set with the @ref avoptions API. 41 * 42 * The first thing you will need to do in order to use lswr is to allocate 43 * SwrContext. This can be done with swr_alloc() or swr_alloc_set_opts(). If you 44 * are using the former, you must set options through the @ref avoptions API. 45 * The latter function provides the same feature, but it allows you to set some 46 * common options in the same statement. 47 * 48 * For example the following code will setup conversion from planar float sample 49 * format to interleaved signed 16-bit integer, downsampling from 48kHz to 50 * 44.1kHz and downmixing from 5.1 channels to stereo (using the default mixing 51 * matrix). This is using the swr_alloc() function. 52 * @code 53 * SwrContext *swr = swr_alloc(); 54 * av_opt_set_channel_layout(swr, "in_channel_layout", AV_CH_LAYOUT_5POINT1, 0); 55 * av_opt_set_channel_layout(swr, "out_channel_layout", AV_CH_LAYOUT_STEREO, 0); 56 * av_opt_set_int(swr, "in_sample_rate", 48000, 0); 57 * av_opt_set_int(swr, "out_sample_rate", 44100, 0); 58 * av_opt_set_sample_fmt(swr, "in_sample_fmt", AV_SAMPLE_FMT_FLTP, 0); 59 * av_opt_set_sample_fmt(swr, "out_sample_fmt", AV_SAMPLE_FMT_S16, 0); 60 * @endcode 61 * 62 * The same job can be done using swr_alloc_set_opts() as well: 63 * @code 64 * SwrContext *swr = swr_alloc_set_opts(NULL, // we're allocating a new context 65 * AV_CH_LAYOUT_STEREO, // out_ch_layout 66 * AV_SAMPLE_FMT_S16, // out_sample_fmt 67 * 44100, // out_sample_rate 68 * AV_CH_LAYOUT_5POINT1, // in_ch_layout 69 * AV_SAMPLE_FMT_FLTP, // in_sample_fmt 70 * 48000, // in_sample_rate 71 * 0, // log_offset 72 * NULL); // log_ctx 73 * @endcode 74 * 75 * Once all values have been set, it must be initialized with swr_init(). If 76 * you need to change the conversion parameters, you can change the parameters 77 * using @ref AVOptions, as described above in the first example; or by using 78 * swr_alloc_set_opts(), but with the first argument the allocated context. 79 * You must then call swr_init() again. 80 * 81 * The conversion itself is done by repeatedly calling swr_convert(). 82 * Note that the samples may get buffered in swr if you provide insufficient 83 * output space or if sample rate conversion is done, which requires "future" 84 * samples. Samples that do not require future input can be retrieved at any 85 * time by using swr_convert() (in_count can be set to 0). 86 * At the end of conversion the resampling buffer can be flushed by calling 87 * swr_convert() with NULL in and 0 in_count. 88 * 89 * The samples used in the conversion process can be managed with the libavutil 90 * @ref lavu_sampmanip "samples manipulation" API, including av_samples_alloc() 91 * function used in the following example. 92 * 93 * The delay between input and output, can at any time be found by using 94 * swr_get_delay(). 95 * 96 * The following code demonstrates the conversion loop assuming the parameters 97 * from above and caller-defined functions get_input() and handle_output(): 98 * @code 99 * uint8_t **input; 100 * int in_samples; 101 * 102 * while (get_input(&input, &in_samples)) { 103 * uint8_t *output; 104 * int out_samples = av_rescale_rnd(swr_get_delay(swr, 48000) + 105 * in_samples, 44100, 48000, AV_ROUND_UP); 106 * av_samples_alloc(&output, NULL, 2, out_samples, 107 * AV_SAMPLE_FMT_S16, 0); 108 * out_samples = swr_convert(swr, &output, out_samples, 109 * input, in_samples); 110 * handle_output(output, out_samples); 111 * av_freep(&output); 112 * } 113 * @endcode 114 * 115 * When the conversion is finished, the conversion 116 * context and everything associated with it must be freed with swr_free(). 117 * A swr_close() function is also available, but it exists mainly for 118 * compatibility with libavresample, and is not required to be called. 119 * 120 * There will be no memory leak if the data is not completely flushed before 121 * swr_free(). 122 */ 123 124 /** 125 * @name Option constants 126 * These constants are used for the @ref avoptions interface for lswr. 127 * @{ 128 * 129 */ 130 131 enum SWR_FLAG_RESAMPLE = 1; ///< Force resampling even if equal sample rate 132 //TODO use int resample ? 133 //long term TODO can we enable this dynamically? 134 135 /** Dithering algorithms */ 136 enum SwrDitherType 137 { 138 SWR_DITHER_NONE = 0, 139 SWR_DITHER_RECTANGULAR = 1, 140 SWR_DITHER_TRIANGULAR = 2, 141 SWR_DITHER_TRIANGULAR_HIGHPASS = 3, 142 143 SWR_DITHER_NS = 64, ///< not part of API/ABI 144 SWR_DITHER_NS_LIPSHITZ = 65, 145 SWR_DITHER_NS_F_WEIGHTED = 66, 146 SWR_DITHER_NS_MODIFIED_E_WEIGHTED = 67, 147 SWR_DITHER_NS_IMPROVED_E_WEIGHTED = 68, 148 SWR_DITHER_NS_SHIBATA = 69, 149 SWR_DITHER_NS_LOW_SHIBATA = 70, 150 SWR_DITHER_NS_HIGH_SHIBATA = 71, 151 SWR_DITHER_NB = 72 ///< not part of API/ABI 152 } 153 154 /** Resampling Engines */ 155 enum SwrEngine 156 { 157 SWR_ENGINE_SWR = 0, /**< SW Resampler */ 158 SWR_ENGINE_SOXR = 1, /**< SoX Resampler */ 159 SWR_ENGINE_NB = 2 ///< not part of API/ABI 160 } 161 162 /** Resampling Filter Types */ 163 enum SwrFilterType 164 { 165 SWR_FILTER_TYPE_CUBIC = 0, /**< Cubic */ 166 SWR_FILTER_TYPE_BLACKMAN_NUTTALL = 1, /**< Blackman Nuttall windowed sinc */ 167 SWR_FILTER_TYPE_KAISER = 2 /**< Kaiser windowed sinc */ 168 } 169 170 /** 171 * @} 172 */ 173 174 /** 175 * The libswresample context. Unlike libavcodec and libavformat, this structure 176 * is opaque. This means that if you would like to set options, you must use 177 * the @ref avoptions API and cannot directly set values to members of the 178 * structure. 179 */ 180 struct SwrContext; 181 182 /** 183 * Get the AVClass for SwrContext. It can be used in combination with 184 * AV_OPT_SEARCH_FAKE_OBJ for examining options. 185 * 186 * @see av_opt_find(). 187 * @return the AVClass of SwrContext 188 */ 189 const(AVClass)* swr_get_class (); 190 191 /** 192 * @name SwrContext constructor functions 193 * @{ 194 */ 195 196 /** 197 * Allocate SwrContext. 198 * 199 * If you use this function you will need to set the parameters (manually or 200 * with swr_alloc_set_opts()) before calling swr_init(). 201 * 202 * @see swr_alloc_set_opts(), swr_init(), swr_free() 203 * @return NULL on error, allocated context otherwise 204 */ 205 SwrContext* swr_alloc (); 206 207 /** 208 * Initialize context after user parameters have been set. 209 * @note The context must be configured using the AVOption API. 210 * 211 * @see av_opt_set_int() 212 * @see av_opt_set_dict() 213 * 214 * @param[in,out] s Swr context to initialize 215 * @return AVERROR error code in case of failure. 216 */ 217 int swr_init (SwrContext* s); 218 219 /** 220 * Check whether an swr context has been initialized or not. 221 * 222 * @param[in] s Swr context to check 223 * @see swr_init() 224 * @return positive if it has been initialized, 0 if not initialized 225 */ 226 int swr_is_initialized (SwrContext* s); 227 228 /** 229 * Allocate SwrContext if needed and set/reset common parameters. 230 * 231 * This function does not require s to be allocated with swr_alloc(). On the 232 * other hand, swr_alloc() can use swr_alloc_set_opts() to set the parameters 233 * on the allocated context. 234 * 235 * @param s existing Swr context if available, or NULL if not 236 * @param out_ch_layout output channel layout (AV_CH_LAYOUT_*) 237 * @param out_sample_fmt output sample format (AV_SAMPLE_FMT_*). 238 * @param out_sample_rate output sample rate (frequency in Hz) 239 * @param in_ch_layout input channel layout (AV_CH_LAYOUT_*) 240 * @param in_sample_fmt input sample format (AV_SAMPLE_FMT_*). 241 * @param in_sample_rate input sample rate (frequency in Hz) 242 * @param log_offset logging level offset 243 * @param log_ctx parent logging context, can be NULL 244 * 245 * @see swr_init(), swr_free() 246 * @return NULL on error, allocated context otherwise 247 */ 248 SwrContext* swr_alloc_set_opts ( 249 SwrContext* s, 250 long out_ch_layout, 251 AVSampleFormat out_sample_fmt, 252 int out_sample_rate, 253 long in_ch_layout, 254 AVSampleFormat in_sample_fmt, 255 int in_sample_rate, 256 int log_offset, 257 void* log_ctx); 258 259 /** 260 * @} 261 * 262 * @name SwrContext destructor functions 263 * @{ 264 */ 265 266 /** 267 * Free the given SwrContext and set the pointer to NULL. 268 * 269 * @param[in] s a pointer to a pointer to Swr context 270 */ 271 void swr_free (SwrContext** s); 272 273 /** 274 * Closes the context so that swr_is_initialized() returns 0. 275 * 276 * The context can be brought back to life by running swr_init(), 277 * swr_init() can also be used without swr_close(). 278 * This function is mainly provided for simplifying the usecase 279 * where one tries to support libavresample and libswresample. 280 * 281 * @param[in,out] s Swr context to be closed 282 */ 283 void swr_close (SwrContext* s); 284 285 /** 286 * @} 287 * 288 * @name Core conversion functions 289 * @{ 290 */ 291 292 /** Convert audio. 293 * 294 * in and in_count can be set to 0 to flush the last few samples out at the 295 * end. 296 * 297 * If more input is provided than output space, then the input will be buffered. 298 * You can avoid this buffering by using swr_get_out_samples() to retrieve an 299 * upper bound on the required number of output samples for the given number of 300 * input samples. Conversion will run directly without copying whenever possible. 301 * 302 * @param s allocated Swr context, with parameters set 303 * @param out output buffers, only the first one need be set in case of packed audio 304 * @param out_count amount of space available for output in samples per channel 305 * @param in input buffers, only the first one need to be set in case of packed audio 306 * @param in_count number of input samples available in one channel 307 * 308 * @return number of samples output per channel, negative value on error 309 */ 310 int swr_convert ( 311 SwrContext* s, 312 ubyte** out_, 313 int out_count, 314 const(ubyte*)* in_, 315 int in_count); 316 317 /** 318 * Convert the next timestamp from input to output 319 * timestamps are in 1/(in_sample_rate * out_sample_rate) units. 320 * 321 * @note There are 2 slightly differently behaving modes. 322 * @li When automatic timestamp compensation is not used, (min_compensation >= FLT_MAX) 323 * in this case timestamps will be passed through with delays compensated 324 * @li When automatic timestamp compensation is used, (min_compensation < FLT_MAX) 325 * in this case the output timestamps will match output sample numbers. 326 * See ffmpeg-resampler(1) for the two modes of compensation. 327 * 328 * @param s[in] initialized Swr context 329 * @param pts[in] timestamp for the next input sample, INT64_MIN if unknown 330 * @see swr_set_compensation(), swr_drop_output(), and swr_inject_silence() are 331 * function used internally for timestamp compensation. 332 * @return the output timestamp for the next output sample 333 */ 334 long swr_next_pts (SwrContext* s, long pts); 335 336 /** 337 * @} 338 * 339 * @name Low-level option setting functions 340 * These functons provide a means to set low-level options that is not possible 341 * with the AVOption API. 342 * @{ 343 */ 344 345 /** 346 * Activate resampling compensation ("soft" compensation). This function is 347 * internally called when needed in swr_next_pts(). 348 * 349 * @param[in,out] s allocated Swr context. If it is not initialized, 350 * or SWR_FLAG_RESAMPLE is not set, swr_init() is 351 * called with the flag set. 352 * @param[in] sample_delta delta in PTS per sample 353 * @param[in] compensation_distance number of samples to compensate for 354 * @return >= 0 on success, AVERROR error codes if: 355 * @li @c s is NULL, 356 * @li @c compensation_distance is less than 0, 357 * @li @c compensation_distance is 0 but sample_delta is not, 358 * @li compensation unsupported by resampler, or 359 * @li swr_init() fails when called. 360 */ 361 int swr_set_compensation (SwrContext* s, int sample_delta, int compensation_distance); 362 363 /** 364 * Set a customized input channel mapping. 365 * 366 * @param[in,out] s allocated Swr context, not yet initialized 367 * @param[in] channel_map customized input channel mapping (array of channel 368 * indexes, -1 for a muted channel) 369 * @return >= 0 on success, or AVERROR error code in case of failure. 370 */ 371 int swr_set_channel_mapping (SwrContext* s, const(int)* channel_map); 372 373 /** 374 * Generate a channel mixing matrix. 375 * 376 * This function is the one used internally by libswresample for building the 377 * default mixing matrix. It is made public just as a utility function for 378 * building custom matrices. 379 * 380 * @param in_layout input channel layout 381 * @param out_layout output channel layout 382 * @param center_mix_level mix level for the center channel 383 * @param surround_mix_level mix level for the surround channel(s) 384 * @param lfe_mix_level mix level for the low-frequency effects channel 385 * @param rematrix_maxval if 1.0, coefficients will be normalized to prevent 386 * overflow. if INT_MAX, coefficients will not be 387 * normalized. 388 * @param[out] matrix mixing coefficients; matrix[i + stride * o] is 389 * the weight of input channel i in output channel o. 390 * @param stride distance between adjacent input channels in the 391 * matrix array 392 * @param matrix_encoding matrixed stereo downmix mode (e.g. dplii) 393 * @param log_ctx parent logging context, can be NULL 394 * @return 0 on success, negative AVERROR code on failure 395 */ 396 int swr_build_matrix ( 397 ulong in_layout, 398 ulong out_layout, 399 double center_mix_level, 400 double surround_mix_level, 401 double lfe_mix_level, 402 double rematrix_maxval, 403 double rematrix_volume, 404 double* matrix, 405 int stride, 406 AVMatrixEncoding matrix_encoding, 407 void* log_ctx); 408 409 /** 410 * Set a customized remix matrix. 411 * 412 * @param s allocated Swr context, not yet initialized 413 * @param matrix remix coefficients; matrix[i + stride * o] is 414 * the weight of input channel i in output channel o 415 * @param stride offset between lines of the matrix 416 * @return >= 0 on success, or AVERROR error code in case of failure. 417 */ 418 int swr_set_matrix (SwrContext* s, const(double)* matrix, int stride); 419 420 /** 421 * @} 422 * 423 * @name Sample handling functions 424 * @{ 425 */ 426 427 /** 428 * Drops the specified number of output samples. 429 * 430 * This function, along with swr_inject_silence(), is called by swr_next_pts() 431 * if needed for "hard" compensation. 432 * 433 * @param s allocated Swr context 434 * @param count number of samples to be dropped 435 * 436 * @return >= 0 on success, or a negative AVERROR code on failure 437 */ 438 int swr_drop_output (SwrContext* s, int count); 439 440 /** 441 * Injects the specified number of silence samples. 442 * 443 * This function, along with swr_drop_output(), is called by swr_next_pts() 444 * if needed for "hard" compensation. 445 * 446 * @param s allocated Swr context 447 * @param count number of samples to be dropped 448 * 449 * @return >= 0 on success, or a negative AVERROR code on failure 450 */ 451 int swr_inject_silence (SwrContext* s, int count); 452 453 /** 454 * Gets the delay the next input sample will experience relative to the next output sample. 455 * 456 * Swresample can buffer data if more input has been provided than available 457 * output space, also converting between sample rates needs a delay. 458 * This function returns the sum of all such delays. 459 * The exact delay is not necessarily an integer value in either input or 460 * output sample rate. Especially when downsampling by a large value, the 461 * output sample rate may be a poor choice to represent the delay, similarly 462 * for upsampling and the input sample rate. 463 * 464 * @param s swr context 465 * @param base timebase in which the returned delay will be: 466 * @li if it's set to 1 the returned delay is in seconds 467 * @li if it's set to 1000 the returned delay is in milliseconds 468 * @li if it's set to the input sample rate then the returned 469 * delay is in input samples 470 * @li if it's set to the output sample rate then the returned 471 * delay is in output samples 472 * @li if it's the least common multiple of in_sample_rate and 473 * out_sample_rate then an exact rounding-free delay will be 474 * returned 475 * @returns the delay in 1 / @c base units. 476 */ 477 long swr_get_delay (SwrContext* s, long base); 478 479 /** 480 * Find an upper bound on the number of samples that the next swr_convert 481 * call will output, if called with in_samples of input samples. This 482 * depends on the internal state, and anything changing the internal state 483 * (like further swr_convert() calls) will may change the number of samples 484 * swr_get_out_samples() returns for the same number of input samples. 485 * 486 * @param in_samples number of input samples. 487 * @note any call to swr_inject_silence(), swr_convert(), swr_next_pts() 488 * or swr_set_compensation() invalidates this limit 489 * @note it is recommended to pass the correct available buffer size 490 * to all functions like swr_convert() even if swr_get_out_samples() 491 * indicates that less would be used. 492 * @returns an upper bound on the number of samples that the next swr_convert 493 * will output or a negative value to indicate an error 494 */ 495 int swr_get_out_samples (SwrContext* s, int in_samples); 496 497 /** 498 * @} 499 * 500 * @name Configuration accessors 501 * @{ 502 */ 503 504 /** 505 * Return the @ref LIBSWRESAMPLE_VERSION_INT constant. 506 * 507 * This is useful to check if the build-time libswresample has the same version 508 * as the run-time one. 509 * 510 * @returns the unsigned int-typed version 511 */ 512 uint swresample_version (); 513 514 /** 515 * Return the swr build-time configuration. 516 * 517 * @returns the build-time @c ./configure flags 518 */ 519 const(char)* swresample_configuration (); 520 521 /** 522 * Return the swr license. 523 * 524 * @returns the license of libswresample, determined at build-time 525 */ 526 const(char)* swresample_license (); 527 528 /** 529 * @} 530 * 531 * @name AVFrame based API 532 * @{ 533 */ 534 535 /** 536 * Convert the samples in the input AVFrame and write them to the output AVFrame. 537 * 538 * Input and output AVFrames must have channel_layout, sample_rate and format set. 539 * 540 * If the output AVFrame does not have the data pointers allocated the nb_samples 541 * field will be set using av_frame_get_buffer() 542 * is called to allocate the frame. 543 * 544 * The output AVFrame can be NULL or have fewer allocated samples than required. 545 * In this case, any remaining samples not written to the output will be added 546 * to an internal FIFO buffer, to be returned at the next call to this function 547 * or to swr_convert(). 548 * 549 * If converting sample rate, there may be data remaining in the internal 550 * resampling delay buffer. swr_get_delay() tells the number of 551 * remaining samples. To get this data as output, call this function or 552 * swr_convert() with NULL input. 553 * 554 * If the SwrContext configuration does not match the output and 555 * input AVFrame settings the conversion does not take place and depending on 556 * which AVFrame is not matching AVERROR_OUTPUT_CHANGED, AVERROR_INPUT_CHANGED 557 * or the result of a bitwise-OR of them is returned. 558 * 559 * @see swr_delay() 560 * @see swr_convert() 561 * @see swr_get_delay() 562 * 563 * @param swr audio resample context 564 * @param output output AVFrame 565 * @param input input AVFrame 566 * @return 0 on success, AVERROR on failure or nonmatching 567 * configuration. 568 */ 569 int swr_convert_frame (SwrContext* swr, AVFrame* output, const(AVFrame)* input); 570 571 /** 572 * Configure or reconfigure the SwrContext using the information 573 * provided by the AVFrames. 574 * 575 * The original resampling context is reset even on failure. 576 * The function calls swr_close() internally if the context is open. 577 * 578 * @see swr_close(); 579 * 580 * @param swr audio resample context 581 * @param output output AVFrame 582 * @param input input AVFrame 583 * @return 0 on success, AVERROR on failure. 584 */ 585 int swr_config_frame (SwrContext* swr, const(AVFrame)* out_, const(AVFrame)* in_); 586 587 /** 588 * @} 589 * @} 590 */ 591 592 /* SWRESAMPLE_SWRESAMPLE_H */