1 /* 2 * This file is part of FFmpeg. 3 * 4 * FFmpeg is free software; you can redistribute it and/or 5 * modify it under the terms of the GNU Lesser General Public 6 * License as published by the Free Software Foundation; either 7 * version 2.1 of the License, or (at your option) any later version. 8 * 9 * FFmpeg is distributed in the hope that it will be useful, 10 * but WITHOUT ANY WARRANTY; without even the implied warranty of 11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 12 * Lesser General Public License for more details. 13 * 14 * You should have received a copy of the GNU Lesser General Public 15 * License along with FFmpeg; if not, write to the Free Software 16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA 17 */ 18 19 /** 20 * @file 21 * @ingroup lavu_frame 22 * reference-counted frame API 23 */ 24 25 #ifndef AVUTIL_FRAME_H 26 #define AVUTIL_FRAME_H 27 28 #include <stddef.h> 29 #include <stdint.h> 30 31 #include "avutil.h" 32 #include "buffer.h" 33 #include "channel_layout.h" 34 #include "dict.h" 35 #include "rational.h" 36 #include "samplefmt.h" 37 #include "pixfmt.h" 38 #include "version.h" 39 40 41 /** 42 * @defgroup lavu_frame AVFrame 43 * @ingroup lavu_data 44 * 45 * @{ 46 * AVFrame is an abstraction for reference-counted raw multimedia data. 47 */ 48 49 enum AVFrameSideDataType { 50 /** 51 * The data is the AVPanScan struct defined in libavcodec. 52 */ 53 AV_FRAME_DATA_PANSCAN, 54 /** 55 * ATSC A53 Part 4 Closed Captions. 56 * A53 CC bitstream is stored as uint8_t in AVFrameSideData.data. 57 * The number of bytes of CC data is AVFrameSideData.size. 58 */ 59 AV_FRAME_DATA_A53_CC, 60 /** 61 * Stereoscopic 3d metadata. 62 * The data is the AVStereo3D struct defined in libavutil/stereo3d.h. 63 */ 64 AV_FRAME_DATA_STEREO3D, 65 /** 66 * The data is the AVMatrixEncoding enum defined in libavutil/channel_layout.h. 67 */ 68 AV_FRAME_DATA_MATRIXENCODING, 69 /** 70 * Metadata relevant to a downmix procedure. 71 * The data is the AVDownmixInfo struct defined in libavutil/downmix_info.h. 72 */ 73 AV_FRAME_DATA_DOWNMIX_INFO, 74 /** 75 * ReplayGain information in the form of the AVReplayGain struct. 76 */ 77 AV_FRAME_DATA_REPLAYGAIN, 78 /** 79 * This side data contains a 3x3 transformation matrix describing an affine 80 * transformation that needs to be applied to the frame for correct 81 * presentation. 82 * 83 * See libavutil/display.h for a detailed description of the data. 84 */ 85 AV_FRAME_DATA_DISPLAYMATRIX, 86 /** 87 * Active Format Description data consisting of a single byte as specified 88 * in ETSI TS 101 154 using AVActiveFormatDescription enum. 89 */ 90 AV_FRAME_DATA_AFD, 91 /** 92 * Motion vectors exported by some codecs (on demand through the export_mvs 93 * flag set in the libavcodec AVCodecContext flags2 option). 94 * The data is the AVMotionVector struct defined in 95 * libavutil/motion_vector.h. 96 */ 97 AV_FRAME_DATA_MOTION_VECTORS, 98 /** 99 * Recommmends skipping the specified number of samples. This is exported 100 * only if the "skip_manual" AVOption is set in libavcodec. 101 * This has the same format as AV_PKT_DATA_SKIP_SAMPLES. 102 * @code 103 * u32le number of samples to skip from start of this packet 104 * u32le number of samples to skip from end of this packet 105 * u8 reason for start skip 106 * u8 reason for end skip (0=padding silence, 1=convergence) 107 * @endcode 108 */ 109 AV_FRAME_DATA_SKIP_SAMPLES, 110 /** 111 * This side data must be associated with an audio frame and corresponds to 112 * enum AVAudioServiceType defined in avcodec.h. 113 */ 114 AV_FRAME_DATA_AUDIO_SERVICE_TYPE, 115 /** 116 * Mastering display metadata associated with a video frame. The payload is 117 * an AVMasteringDisplayMetadata type and contains information about the 118 * mastering display color volume. 119 */ 120 AV_FRAME_DATA_MASTERING_DISPLAY_METADATA, 121 /** 122 * The GOP timecode in 25 bit timecode format. Data format is 64-bit integer. 123 * This is set on the first frame of a GOP that has a temporal reference of 0. 124 */ 125 AV_FRAME_DATA_GOP_TIMECODE, 126 127 /** 128 * The data represents the AVSphericalMapping structure defined in 129 * libavutil/spherical.h. 130 */ 131 AV_FRAME_DATA_SPHERICAL, 132 133 /** 134 * Content light level (based on CTA-861.3). This payload contains data in 135 * the form of the AVContentLightMetadata struct. 136 */ 137 AV_FRAME_DATA_CONTENT_LIGHT_LEVEL, 138 139 /** 140 * The data contains an ICC profile as an opaque octet buffer following the 141 * format described by ISO 15076-1 with an optional name defined in the 142 * metadata key entry "name". 143 */ 144 AV_FRAME_DATA_ICC_PROFILE, 145 146 /** 147 * Timecode which conforms to SMPTE ST 12-1. The data is an array of 4 uint32_t 148 * where the first uint32_t describes how many (1-3) of the other timecodes are used. 149 * The timecode format is described in the documentation of av_timecode_get_smpte_from_framenum() 150 * function in libavutil/timecode.h. 151 */ 152 AV_FRAME_DATA_S12M_TIMECODE, 153 154 /** 155 * HDR dynamic metadata associated with a video frame. The payload is 156 * an AVDynamicHDRPlus type and contains information for color 157 * volume transform - application 4 of SMPTE 2094-40:2016 standard. 158 */ 159 AV_FRAME_DATA_DYNAMIC_HDR_PLUS, 160 161 /** 162 * Regions Of Interest, the data is an array of AVRegionOfInterest type, the number of 163 * array element is implied by AVFrameSideData.size / AVRegionOfInterest.self_size. 164 */ 165 AV_FRAME_DATA_REGIONS_OF_INTEREST, 166 167 /** 168 * Encoding parameters for a video frame, as described by AVVideoEncParams. 169 */ 170 AV_FRAME_DATA_VIDEO_ENC_PARAMS, 171 172 /** 173 * User data unregistered metadata associated with a video frame. 174 * This is the H.26[45] UDU SEI message, and shouldn't be used for any other purpose 175 * The data is stored as uint8_t in AVFrameSideData.data which is 16 bytes of 176 * uuid_iso_iec_11578 followed by AVFrameSideData.size - 16 bytes of user_data_payload_byte. 177 */ 178 AV_FRAME_DATA_SEI_UNREGISTERED, 179 180 /** 181 * Film grain parameters for a frame, described by AVFilmGrainParams. 182 * Must be present for every frame which should have film grain applied. 183 */ 184 AV_FRAME_DATA_FILM_GRAIN_PARAMS, 185 186 /** 187 * Bounding boxes for object detection and classification, 188 * as described by AVDetectionBBoxHeader. 189 */ 190 AV_FRAME_DATA_DETECTION_BBOXES, 191 192 /** 193 * Dolby Vision RPU raw data, suitable for passing to x265 194 * or other libraries. Array of uint8_t, with NAL emulation 195 * bytes intact. 196 */ 197 AV_FRAME_DATA_DOVI_RPU_BUFFER, 198 199 /** 200 * Parsed Dolby Vision metadata, suitable for passing to a software 201 * implementation. The payload is the AVDOVIMetadata struct defined in 202 * libavutil/dovi_meta.h. 203 */ 204 AV_FRAME_DATA_DOVI_METADATA, 205 206 /** 207 * HDR Vivid dynamic metadata associated with a video frame. The payload is 208 * an AVDynamicHDRVivid type and contains information for color 209 * volume transform - CUVA 005.1-2021. 210 */ 211 AV_FRAME_DATA_DYNAMIC_HDR_VIVID, 212 }; 213 214 enum AVActiveFormatDescription { 215 AV_AFD_SAME = 8, 216 AV_AFD_4_3 = 9, 217 AV_AFD_16_9 = 10, 218 AV_AFD_14_9 = 11, 219 AV_AFD_4_3_SP_14_9 = 13, 220 AV_AFD_16_9_SP_14_9 = 14, 221 AV_AFD_SP_4_3 = 15, 222 }; 223 224 225 /** 226 * Structure to hold side data for an AVFrame. 227 * 228 * sizeof(AVFrameSideData) is not a part of the public ABI, so new fields may be added 229 * to the end with a minor bump. 230 */ 231 typedef struct AVFrameSideData { 232 enum AVFrameSideDataType type; 233 uint8_t *data; 234 size_t size; 235 AVDictionary *metadata; 236 AVBufferRef *buf; 237 } AVFrameSideData; 238 239 /** 240 * Structure describing a single Region Of Interest. 241 * 242 * When multiple regions are defined in a single side-data block, they 243 * should be ordered from most to least important - some encoders are only 244 * capable of supporting a limited number of distinct regions, so will have 245 * to truncate the list. 246 * 247 * When overlapping regions are defined, the first region containing a given 248 * area of the frame applies. 249 */ 250 typedef struct AVRegionOfInterest { 251 /** 252 * Must be set to the size of this data structure (that is, 253 * sizeof(AVRegionOfInterest)). 254 */ 255 uint32_t self_size; 256 /** 257 * Distance in pixels from the top edge of the frame to the top and 258 * bottom edges and from the left edge of the frame to the left and 259 * right edges of the rectangle defining this region of interest. 260 * 261 * The constraints on a region are encoder dependent, so the region 262 * actually affected may be slightly larger for alignment or other 263 * reasons. 264 */ 265 int top; 266 int bottom; 267 int left; 268 int right; 269 /** 270 * Quantisation offset. 271 * 272 * Must be in the range -1 to +1. A value of zero indicates no quality 273 * change. A negative value asks for better quality (less quantisation), 274 * while a positive value asks for worse quality (greater quantisation). 275 * 276 * The range is calibrated so that the extreme values indicate the 277 * largest possible offset - if the rest of the frame is encoded with the 278 * worst possible quality, an offset of -1 indicates that this region 279 * should be encoded with the best possible quality anyway. Intermediate 280 * values are then interpolated in some codec-dependent way. 281 * 282 * For example, in 10-bit H.264 the quantisation parameter varies between 283 * -12 and 51. A typical qoffset value of -1/10 therefore indicates that 284 * this region should be encoded with a QP around one-tenth of the full 285 * range better than the rest of the frame. So, if most of the frame 286 * were to be encoded with a QP of around 30, this region would get a QP 287 * of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3). 288 * An extreme value of -1 would indicate that this region should be 289 * encoded with the best possible quality regardless of the treatment of 290 * the rest of the frame - that is, should be encoded at a QP of -12. 291 */ 292 AVRational qoffset; 293 } AVRegionOfInterest; 294 295 /** 296 * This structure describes decoded (raw) audio or video data. 297 * 298 * AVFrame must be allocated using av_frame_alloc(). Note that this only 299 * allocates the AVFrame itself, the buffers for the data must be managed 300 * through other means (see below). 301 * AVFrame must be freed with av_frame_free(). 302 * 303 * AVFrame is typically allocated once and then reused multiple times to hold 304 * different data (e.g. a single AVFrame to hold frames received from a 305 * decoder). In such a case, av_frame_unref() will free any references held by 306 * the frame and reset it to its original clean state before it 307 * is reused again. 308 * 309 * The data described by an AVFrame is usually reference counted through the 310 * AVBuffer API. The underlying buffer references are stored in AVFrame.buf / 311 * AVFrame.extended_buf. An AVFrame is considered to be reference counted if at 312 * least one reference is set, i.e. if AVFrame.buf[0] != NULL. In such a case, 313 * every single data plane must be contained in one of the buffers in 314 * AVFrame.buf or AVFrame.extended_buf. 315 * There may be a single buffer for all the data, or one separate buffer for 316 * each plane, or anything in between. 317 * 318 * sizeof(AVFrame) is not a part of the public ABI, so new fields may be added 319 * to the end with a minor bump. 320 * 321 * Fields can be accessed through AVOptions, the name string used, matches the 322 * C structure field name for fields accessible through AVOptions. The AVClass 323 * for AVFrame can be obtained from avcodec_get_frame_class() 324 */ 325 typedef struct AVFrame { 326 #define AV_NUM_DATA_POINTERS 8 327 /** 328 * pointer to the picture/channel planes. 329 * This might be different from the first allocated byte. For video, 330 * it could even point to the end of the image data. 331 * 332 * All pointers in data and extended_data must point into one of the 333 * AVBufferRef in buf or extended_buf. 334 * 335 * Some decoders access areas outside 0,0 - width,height, please 336 * see avcodec_align_dimensions2(). Some filters and swscale can read 337 * up to 16 bytes beyond the planes, if these filters are to be used, 338 * then 16 extra bytes must be allocated. 339 * 340 * NOTE: Pointers not needed by the format MUST be set to NULL. 341 * 342 * @attention In case of video, the data[] pointers can point to the 343 * end of image data in order to reverse line order, when used in 344 * combination with negative values in the linesize[] array. 345 */ 346 uint8_t *data[AV_NUM_DATA_POINTERS]; 347 348 /** 349 * For video, a positive or negative value, which is typically indicating 350 * the size in bytes of each picture line, but it can also be: 351 * - the negative byte size of lines for vertical flipping 352 * (with data[n] pointing to the end of the data 353 * - a positive or negative multiple of the byte size as for accessing 354 * even and odd fields of a frame (possibly flipped) 355 * 356 * For audio, only linesize[0] may be set. For planar audio, each channel 357 * plane must be the same size. 358 * 359 * For video the linesizes should be multiples of the CPUs alignment 360 * preference, this is 16 or 32 for modern desktop CPUs. 361 * Some code requires such alignment other code can be slower without 362 * correct alignment, for yet other it makes no difference. 363 * 364 * @note The linesize may be larger than the size of usable data -- there 365 * may be extra padding present for performance reasons. 366 * 367 * @attention In case of video, line size values can be negative to achieve 368 * a vertically inverted iteration over image lines. 369 */ 370 int linesize[AV_NUM_DATA_POINTERS]; 371 372 /** 373 * pointers to the data planes/channels. 374 * 375 * For video, this should simply point to data[]. 376 * 377 * For planar audio, each channel has a separate data pointer, and 378 * linesize[0] contains the size of each channel buffer. 379 * For packed audio, there is just one data pointer, and linesize[0] 380 * contains the total size of the buffer for all channels. 381 * 382 * Note: Both data and extended_data should always be set in a valid frame, 383 * but for planar audio with more channels that can fit in data, 384 * extended_data must be used in order to access all channels. 385 */ 386 uint8_t **extended_data; 387 388 /** 389 * @name Video dimensions 390 * Video frames only. The coded dimensions (in pixels) of the video frame, 391 * i.e. the size of the rectangle that contains some well-defined values. 392 * 393 * @note The part of the frame intended for display/presentation is further 394 * restricted by the @ref cropping "Cropping rectangle". 395 * @{ 396 */ 397 int width, height; 398 /** 399 * @} 400 */ 401 402 /** 403 * number of audio samples (per channel) described by this frame 404 */ 405 int nb_samples; 406 407 /** 408 * format of the frame, -1 if unknown or unset 409 * Values correspond to enum AVPixelFormat for video frames, 410 * enum AVSampleFormat for audio) 411 */ 412 int format; 413 414 /** 415 * 1 -> keyframe, 0-> not 416 */ 417 int key_frame; 418 419 /** 420 * Picture type of the frame. 421 */ 422 enum AVPictureType pict_type; 423 424 /** 425 * Sample aspect ratio for the video frame, 0/1 if unknown/unspecified. 426 */ 427 AVRational sample_aspect_ratio; 428 429 /** 430 * Presentation timestamp in time_base units (time when frame should be shown to user). 431 */ 432 int64_t pts; 433 434 /** 435 * DTS copied from the AVPacket that triggered returning this frame. (if frame threading isn't used) 436 * This is also the Presentation time of this AVFrame calculated from 437 * only AVPacket.dts values without pts values. 438 */ 439 int64_t pkt_dts; 440 441 /** 442 * Time base for the timestamps in this frame. 443 * In the future, this field may be set on frames output by decoders or 444 * filters, but its value will be by default ignored on input to encoders 445 * or filters. 446 */ 447 AVRational time_base; 448 449 /** 450 * picture number in bitstream order 451 */ 452 int coded_picture_number; 453 /** 454 * picture number in display order 455 */ 456 int display_picture_number; 457 458 /** 459 * quality (between 1 (good) and FF_LAMBDA_MAX (bad)) 460 */ 461 int quality; 462 463 /** 464 * for some private data of the user 465 */ 466 void *opaque; 467 468 /** 469 * When decoding, this signals how much the picture must be delayed. 470 * extra_delay = repeat_pict / (2*fps) 471 */ 472 int repeat_pict; 473 474 /** 475 * The content of the picture is interlaced. 476 */ 477 int interlaced_frame; 478 479 /** 480 * If the content is interlaced, is top field displayed first. 481 */ 482 int top_field_first; 483 484 /** 485 * Tell user application that palette has changed from previous frame. 486 */ 487 int palette_has_changed; 488 489 /** 490 * reordered opaque 64 bits (generally an integer or a double precision float 491 * PTS but can be anything). 492 * The user sets AVCodecContext.reordered_opaque to represent the input at 493 * that time, 494 * the decoder reorders values as needed and sets AVFrame.reordered_opaque 495 * to exactly one of the values provided by the user through AVCodecContext.reordered_opaque 496 */ 497 int64_t reordered_opaque; 498 499 /** 500 * Sample rate of the audio data. 501 */ 502 int sample_rate; 503 504 #if FF_API_OLD_CHANNEL_LAYOUT 505 /** 506 * Channel layout of the audio data. 507 * @deprecated use ch_layout instead 508 */ 509 attribute_deprecated 510 uint64_t channel_layout; 511 #endif 512 513 /** 514 * AVBuffer references backing the data for this frame. All the pointers in 515 * data and extended_data must point inside one of the buffers in buf or 516 * extended_buf. This array must be filled contiguously -- if buf[i] is 517 * non-NULL then buf[j] must also be non-NULL for all j < i. 518 * 519 * There may be at most one AVBuffer per data plane, so for video this array 520 * always contains all the references. For planar audio with more than 521 * AV_NUM_DATA_POINTERS channels, there may be more buffers than can fit in 522 * this array. Then the extra AVBufferRef pointers are stored in the 523 * extended_buf array. 524 */ 525 AVBufferRef *buf[AV_NUM_DATA_POINTERS]; 526 527 /** 528 * For planar audio which requires more than AV_NUM_DATA_POINTERS 529 * AVBufferRef pointers, this array will hold all the references which 530 * cannot fit into AVFrame.buf. 531 * 532 * Note that this is different from AVFrame.extended_data, which always 533 * contains all the pointers. This array only contains the extra pointers, 534 * which cannot fit into AVFrame.buf. 535 * 536 * This array is always allocated using av_malloc() by whoever constructs 537 * the frame. It is freed in av_frame_unref(). 538 */ 539 AVBufferRef **extended_buf; 540 /** 541 * Number of elements in extended_buf. 542 */ 543 int nb_extended_buf; 544 545 AVFrameSideData **side_data; 546 int nb_side_data; 547 548 /** 549 * @defgroup lavu_frame_flags AV_FRAME_FLAGS 550 * @ingroup lavu_frame 551 * Flags describing additional frame properties. 552 * 553 * @{ 554 */ 555 556 /** 557 * The frame data may be corrupted, e.g. due to decoding errors. 558 */ 559 #define AV_FRAME_FLAG_CORRUPT (1 << 0) 560 /** 561 * A flag to mark the frames which need to be decoded, but shouldn't be output. 562 */ 563 #define AV_FRAME_FLAG_DISCARD (1 << 2) 564 /** 565 * @} 566 */ 567 568 /** 569 * Frame flags, a combination of @ref lavu_frame_flags 570 */ 571 int flags; 572 573 /** 574 * MPEG vs JPEG YUV range. 575 * - encoding: Set by user 576 * - decoding: Set by libavcodec 577 */ 578 enum AVColorRange color_range; 579 580 enum AVColorPrimaries color_primaries; 581 582 enum AVColorTransferCharacteristic color_trc; 583 584 /** 585 * YUV colorspace type. 586 * - encoding: Set by user 587 * - decoding: Set by libavcodec 588 */ 589 enum AVColorSpace colorspace; 590 591 enum AVChromaLocation chroma_location; 592 593 /** 594 * frame timestamp estimated using various heuristics, in stream time base 595 * - encoding: unused 596 * - decoding: set by libavcodec, read by user. 597 */ 598 int64_t best_effort_timestamp; 599 600 /** 601 * reordered pos from the last AVPacket that has been input into the decoder 602 * - encoding: unused 603 * - decoding: Read by user. 604 */ 605 int64_t pkt_pos; 606 607 /** 608 * duration of the corresponding packet, expressed in 609 * AVStream->time_base units, 0 if unknown. 610 * - encoding: unused 611 * - decoding: Read by user. 612 */ 613 int64_t pkt_duration; 614 615 /** 616 * metadata. 617 * - encoding: Set by user. 618 * - decoding: Set by libavcodec. 619 */ 620 AVDictionary *metadata; 621 622 /** 623 * decode error flags of the frame, set to a combination of 624 * FF_DECODE_ERROR_xxx flags if the decoder produced a frame, but there 625 * were errors during the decoding. 626 * - encoding: unused 627 * - decoding: set by libavcodec, read by user. 628 */ 629 int decode_error_flags; 630 #define FF_DECODE_ERROR_INVALID_BITSTREAM 1 631 #define FF_DECODE_ERROR_MISSING_REFERENCE 2 632 #define FF_DECODE_ERROR_CONCEALMENT_ACTIVE 4 633 #define FF_DECODE_ERROR_DECODE_SLICES 8 634 635 #if FF_API_OLD_CHANNEL_LAYOUT 636 /** 637 * number of audio channels, only used for audio. 638 * - encoding: unused 639 * - decoding: Read by user. 640 * @deprecated use ch_layout instead 641 */ 642 attribute_deprecated 643 int channels; 644 #endif 645 646 /** 647 * size of the corresponding packet containing the compressed 648 * frame. 649 * It is set to a negative value if unknown. 650 * - encoding: unused 651 * - decoding: set by libavcodec, read by user. 652 */ 653 int pkt_size; 654 655 /** 656 * For hwaccel-format frames, this should be a reference to the 657 * AVHWFramesContext describing the frame. 658 */ 659 AVBufferRef *hw_frames_ctx; 660 661 /** 662 * AVBufferRef for free use by the API user. FFmpeg will never check the 663 * contents of the buffer ref. FFmpeg calls av_buffer_unref() on it when 664 * the frame is unreferenced. av_frame_copy_props() calls create a new 665 * reference with av_buffer_ref() for the target frame's opaque_ref field. 666 * 667 * This is unrelated to the opaque field, although it serves a similar 668 * purpose. 669 */ 670 AVBufferRef *opaque_ref; 671 672 /** 673 * @anchor cropping 674 * @name Cropping 675 * Video frames only. The number of pixels to discard from the the 676 * top/bottom/left/right border of the frame to obtain the sub-rectangle of 677 * the frame intended for presentation. 678 * @{ 679 */ 680 size_t crop_top; 681 size_t crop_bottom; 682 size_t crop_left; 683 size_t crop_right; 684 /** 685 * @} 686 */ 687 688 /** 689 * AVBufferRef for internal use by a single libav* library. 690 * Must not be used to transfer data between libraries. 691 * Has to be NULL when ownership of the frame leaves the respective library. 692 * 693 * Code outside the FFmpeg libs should never check or change the contents of the buffer ref. 694 * 695 * FFmpeg calls av_buffer_unref() on it when the frame is unreferenced. 696 * av_frame_copy_props() calls create a new reference with av_buffer_ref() 697 * for the target frame's private_ref field. 698 */ 699 AVBufferRef *private_ref; 700 701 /** 702 * Channel layout of the audio data. 703 */ 704 AVChannelLayout ch_layout; 705 } AVFrame; 706 707 708 #if FF_API_COLORSPACE_NAME 709 /** 710 * Get the name of a colorspace. 711 * @return a static string identifying the colorspace; can be NULL. 712 * @deprecated use av_color_space_name() 713 */ 714 attribute_deprecated 715 const char *av_get_colorspace_name(enum AVColorSpace val); 716 #endif 717 /** 718 * Allocate an AVFrame and set its fields to default values. The resulting 719 * struct must be freed using av_frame_free(). 720 * 721 * @return An AVFrame filled with default values or NULL on failure. 722 * 723 * @note this only allocates the AVFrame itself, not the data buffers. Those 724 * must be allocated through other means, e.g. with av_frame_get_buffer() or 725 * manually. 726 */ 727 AVFrame *av_frame_alloc(void); 728 729 /** 730 * Free the frame and any dynamically allocated objects in it, 731 * e.g. extended_data. If the frame is reference counted, it will be 732 * unreferenced first. 733 * 734 * @param frame frame to be freed. The pointer will be set to NULL. 735 */ 736 void av_frame_free(AVFrame **frame); 737 738 /** 739 * Set up a new reference to the data described by the source frame. 740 * 741 * Copy frame properties from src to dst and create a new reference for each 742 * AVBufferRef from src. 743 * 744 * If src is not reference counted, new buffers are allocated and the data is 745 * copied. 746 * 747 * @warning: dst MUST have been either unreferenced with av_frame_unref(dst), 748 * or newly allocated with av_frame_alloc() before calling this 749 * function, or undefined behavior will occur. 750 * 751 * @return 0 on success, a negative AVERROR on error 752 */ 753 int av_frame_ref(AVFrame *dst, const AVFrame *src); 754 755 /** 756 * Create a new frame that references the same data as src. 757 * 758 * This is a shortcut for av_frame_alloc()+av_frame_ref(). 759 * 760 * @return newly created AVFrame on success, NULL on error. 761 */ 762 AVFrame *av_frame_clone(const AVFrame *src); 763 764 /** 765 * Unreference all the buffers referenced by frame and reset the frame fields. 766 */ 767 void av_frame_unref(AVFrame *frame); 768 769 /** 770 * Move everything contained in src to dst and reset src. 771 * 772 * @warning: dst is not unreferenced, but directly overwritten without reading 773 * or deallocating its contents. Call av_frame_unref(dst) manually 774 * before calling this function to ensure that no memory is leaked. 775 */ 776 void av_frame_move_ref(AVFrame *dst, AVFrame *src); 777 778 /** 779 * Allocate new buffer(s) for audio or video data. 780 * 781 * The following fields must be set on frame before calling this function: 782 * - format (pixel format for video, sample format for audio) 783 * - width and height for video 784 * - nb_samples and ch_layout for audio 785 * 786 * This function will fill AVFrame.data and AVFrame.buf arrays and, if 787 * necessary, allocate and fill AVFrame.extended_data and AVFrame.extended_buf. 788 * For planar formats, one buffer will be allocated for each plane. 789 * 790 * @warning: if frame already has been allocated, calling this function will 791 * leak memory. In addition, undefined behavior can occur in certain 792 * cases. 793 * 794 * @param frame frame in which to store the new buffers. 795 * @param align Required buffer size alignment. If equal to 0, alignment will be 796 * chosen automatically for the current CPU. It is highly 797 * recommended to pass 0 here unless you know what you are doing. 798 * 799 * @return 0 on success, a negative AVERROR on error. 800 */ 801 int av_frame_get_buffer(AVFrame *frame, int align); 802 803 /** 804 * Check if the frame data is writable. 805 * 806 * @return A positive value if the frame data is writable (which is true if and 807 * only if each of the underlying buffers has only one reference, namely the one 808 * stored in this frame). Return 0 otherwise. 809 * 810 * If 1 is returned the answer is valid until av_buffer_ref() is called on any 811 * of the underlying AVBufferRefs (e.g. through av_frame_ref() or directly). 812 * 813 * @see av_frame_make_writable(), av_buffer_is_writable() 814 */ 815 int av_frame_is_writable(AVFrame *frame); 816 817 /** 818 * Ensure that the frame data is writable, avoiding data copy if possible. 819 * 820 * Do nothing if the frame is writable, allocate new buffers and copy the data 821 * if it is not. 822 * 823 * @return 0 on success, a negative AVERROR on error. 824 * 825 * @see av_frame_is_writable(), av_buffer_is_writable(), 826 * av_buffer_make_writable() 827 */ 828 int av_frame_make_writable(AVFrame *frame); 829 830 /** 831 * Copy the frame data from src to dst. 832 * 833 * This function does not allocate anything, dst must be already initialized and 834 * allocated with the same parameters as src. 835 * 836 * This function only copies the frame data (i.e. the contents of the data / 837 * extended data arrays), not any other properties. 838 * 839 * @return >= 0 on success, a negative AVERROR on error. 840 */ 841 int av_frame_copy(AVFrame *dst, const AVFrame *src); 842 843 /** 844 * Copy only "metadata" fields from src to dst. 845 * 846 * Metadata for the purpose of this function are those fields that do not affect 847 * the data layout in the buffers. E.g. pts, sample rate (for audio) or sample 848 * aspect ratio (for video), but not width/height or channel layout. 849 * Side data is also copied. 850 */ 851 int av_frame_copy_props(AVFrame *dst, const AVFrame *src); 852 853 /** 854 * Get the buffer reference a given data plane is stored in. 855 * 856 * @param plane index of the data plane of interest in frame->extended_data. 857 * 858 * @return the buffer reference that contains the plane or NULL if the input 859 * frame is not valid. 860 */ 861 AVBufferRef *av_frame_get_plane_buffer(AVFrame *frame, int plane); 862 863 /** 864 * Add a new side data to a frame. 865 * 866 * @param frame a frame to which the side data should be added 867 * @param type type of the added side data 868 * @param size size of the side data 869 * 870 * @return newly added side data on success, NULL on error 871 */ 872 AVFrameSideData *av_frame_new_side_data(AVFrame *frame, 873 enum AVFrameSideDataType type, 874 size_t size); 875 876 /** 877 * Add a new side data to a frame from an existing AVBufferRef 878 * 879 * @param frame a frame to which the side data should be added 880 * @param type the type of the added side data 881 * @param buf an AVBufferRef to add as side data. The ownership of 882 * the reference is transferred to the frame. 883 * 884 * @return newly added side data on success, NULL on error. On failure 885 * the frame is unchanged and the AVBufferRef remains owned by 886 * the caller. 887 */ 888 AVFrameSideData *av_frame_new_side_data_from_buf(AVFrame *frame, 889 enum AVFrameSideDataType type, 890 AVBufferRef *buf); 891 892 /** 893 * @return a pointer to the side data of a given type on success, NULL if there 894 * is no side data with such type in this frame. 895 */ 896 AVFrameSideData *av_frame_get_side_data(const AVFrame *frame, 897 enum AVFrameSideDataType type); 898 899 /** 900 * Remove and free all side data instances of the given type. 901 */ 902 void av_frame_remove_side_data(AVFrame *frame, enum AVFrameSideDataType type); 903 904 905 /** 906 * Flags for frame cropping. 907 */ 908 enum { 909 /** 910 * Apply the maximum possible cropping, even if it requires setting the 911 * AVFrame.data[] entries to unaligned pointers. Passing unaligned data 912 * to FFmpeg API is generally not allowed, and causes undefined behavior 913 * (such as crashes). You can pass unaligned data only to FFmpeg APIs that 914 * are explicitly documented to accept it. Use this flag only if you 915 * absolutely know what you are doing. 916 */ 917 AV_FRAME_CROP_UNALIGNED = 1 << 0, 918 }; 919 920 /** 921 * Crop the given video AVFrame according to its crop_left/crop_top/crop_right/ 922 * crop_bottom fields. If cropping is successful, the function will adjust the 923 * data pointers and the width/height fields, and set the crop fields to 0. 924 * 925 * In all cases, the cropping boundaries will be rounded to the inherent 926 * alignment of the pixel format. In some cases, such as for opaque hwaccel 927 * formats, the left/top cropping is ignored. The crop fields are set to 0 even 928 * if the cropping was rounded or ignored. 929 * 930 * @param frame the frame which should be cropped 931 * @param flags Some combination of AV_FRAME_CROP_* flags, or 0. 932 * 933 * @return >= 0 on success, a negative AVERROR on error. If the cropping fields 934 * were invalid, AVERROR(ERANGE) is returned, and nothing is changed. 935 */ 936 int av_frame_apply_cropping(AVFrame *frame, int flags); 937 938 /** 939 * @return a string identifying the side data type 940 */ 941 const char *av_frame_side_data_name(enum AVFrameSideDataType type); 942 943 /** 944 * @} 945 */ 946 947 #endif /* AVUTIL_FRAME_H */ 948