FFmpeg 6.0.1
frame.h
Go to the documentation of this file.
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
50 /**
51 * The data is the AVPanScan struct defined in libavcodec.
52 */
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 */
60 /**
61 * Stereoscopic 3d metadata.
62 * The data is the AVStereo3D struct defined in libavutil/stereo3d.h.
63 */
65 /**
66 * The data is the AVMatrixEncoding enum defined in libavutil/channel_layout.h.
67 */
69 /**
70 * Metadata relevant to a downmix procedure.
71 * The data is the AVDownmixInfo struct defined in libavutil/downmix_info.h.
72 */
74 /**
75 * ReplayGain information in the form of the AVReplayGain struct.
76 */
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 */
86 /**
87 * Active Format Description data consisting of a single byte as specified
88 * in ETSI TS 101 154 using AVActiveFormatDescription enum.
89 */
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 */
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 */
110 /**
111 * This side data must be associated with an audio frame and corresponds to
112 * enum AVAudioServiceType defined in avcodec.h.
113 */
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 */
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 */
126
127 /**
128 * The data represents the AVSphericalMapping structure defined in
129 * libavutil/spherical.h.
130 */
132
133 /**
134 * Content light level (based on CTA-861.3). This payload contains data in
135 * the form of the AVContentLightMetadata struct.
136 */
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 */
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 */
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 */
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 */
166
167 /**
168 * Encoding parameters for a video frame, as described by AVVideoEncParams.
169 */
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 */
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 */
185
186 /**
187 * Bounding boxes for object detection and classification,
188 * as described by AVDetectionBBoxHeader.
189 */
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 */
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 */
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 */
212
213 /**
214 * Ambient viewing environment metadata, as defined by H.274.
215 */
217};
218
227};
228
229
230/**
231 * Structure to hold side data for an AVFrame.
232 *
233 * sizeof(AVFrameSideData) is not a part of the public ABI, so new fields may be added
234 * to the end with a minor bump.
235 */
236typedef struct AVFrameSideData {
238 uint8_t *data;
239 size_t size;
243
244/**
245 * Structure describing a single Region Of Interest.
246 *
247 * When multiple regions are defined in a single side-data block, they
248 * should be ordered from most to least important - some encoders are only
249 * capable of supporting a limited number of distinct regions, so will have
250 * to truncate the list.
251 *
252 * When overlapping regions are defined, the first region containing a given
253 * area of the frame applies.
254 */
255typedef struct AVRegionOfInterest {
256 /**
257 * Must be set to the size of this data structure (that is,
258 * sizeof(AVRegionOfInterest)).
259 */
260 uint32_t self_size;
261 /**
262 * Distance in pixels from the top edge of the frame to the top and
263 * bottom edges and from the left edge of the frame to the left and
264 * right edges of the rectangle defining this region of interest.
265 *
266 * The constraints on a region are encoder dependent, so the region
267 * actually affected may be slightly larger for alignment or other
268 * reasons.
269 */
270 int top;
272 int left;
273 int right;
274 /**
275 * Quantisation offset.
276 *
277 * Must be in the range -1 to +1. A value of zero indicates no quality
278 * change. A negative value asks for better quality (less quantisation),
279 * while a positive value asks for worse quality (greater quantisation).
280 *
281 * The range is calibrated so that the extreme values indicate the
282 * largest possible offset - if the rest of the frame is encoded with the
283 * worst possible quality, an offset of -1 indicates that this region
284 * should be encoded with the best possible quality anyway. Intermediate
285 * values are then interpolated in some codec-dependent way.
286 *
287 * For example, in 10-bit H.264 the quantisation parameter varies between
288 * -12 and 51. A typical qoffset value of -1/10 therefore indicates that
289 * this region should be encoded with a QP around one-tenth of the full
290 * range better than the rest of the frame. So, if most of the frame
291 * were to be encoded with a QP of around 30, this region would get a QP
292 * of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
293 * An extreme value of -1 would indicate that this region should be
294 * encoded with the best possible quality regardless of the treatment of
295 * the rest of the frame - that is, should be encoded at a QP of -12.
296 */
299
300/**
301 * This structure describes decoded (raw) audio or video data.
302 *
303 * AVFrame must be allocated using av_frame_alloc(). Note that this only
304 * allocates the AVFrame itself, the buffers for the data must be managed
305 * through other means (see below).
306 * AVFrame must be freed with av_frame_free().
307 *
308 * AVFrame is typically allocated once and then reused multiple times to hold
309 * different data (e.g. a single AVFrame to hold frames received from a
310 * decoder). In such a case, av_frame_unref() will free any references held by
311 * the frame and reset it to its original clean state before it
312 * is reused again.
313 *
314 * The data described by an AVFrame is usually reference counted through the
315 * AVBuffer API. The underlying buffer references are stored in AVFrame.buf /
316 * AVFrame.extended_buf. An AVFrame is considered to be reference counted if at
317 * least one reference is set, i.e. if AVFrame.buf[0] != NULL. In such a case,
318 * every single data plane must be contained in one of the buffers in
319 * AVFrame.buf or AVFrame.extended_buf.
320 * There may be a single buffer for all the data, or one separate buffer for
321 * each plane, or anything in between.
322 *
323 * sizeof(AVFrame) is not a part of the public ABI, so new fields may be added
324 * to the end with a minor bump.
325 *
326 * Fields can be accessed through AVOptions, the name string used, matches the
327 * C structure field name for fields accessible through AVOptions. The AVClass
328 * for AVFrame can be obtained from avcodec_get_frame_class()
329 */
330typedef struct AVFrame {
331#define AV_NUM_DATA_POINTERS 8
332 /**
333 * pointer to the picture/channel planes.
334 * This might be different from the first allocated byte. For video,
335 * it could even point to the end of the image data.
336 *
337 * All pointers in data and extended_data must point into one of the
338 * AVBufferRef in buf or extended_buf.
339 *
340 * Some decoders access areas outside 0,0 - width,height, please
341 * see avcodec_align_dimensions2(). Some filters and swscale can read
342 * up to 16 bytes beyond the planes, if these filters are to be used,
343 * then 16 extra bytes must be allocated.
344 *
345 * NOTE: Pointers not needed by the format MUST be set to NULL.
346 *
347 * @attention In case of video, the data[] pointers can point to the
348 * end of image data in order to reverse line order, when used in
349 * combination with negative values in the linesize[] array.
350 */
352
353 /**
354 * For video, a positive or negative value, which is typically indicating
355 * the size in bytes of each picture line, but it can also be:
356 * - the negative byte size of lines for vertical flipping
357 * (with data[n] pointing to the end of the data
358 * - a positive or negative multiple of the byte size as for accessing
359 * even and odd fields of a frame (possibly flipped)
360 *
361 * For audio, only linesize[0] may be set. For planar audio, each channel
362 * plane must be the same size.
363 *
364 * For video the linesizes should be multiples of the CPUs alignment
365 * preference, this is 16 or 32 for modern desktop CPUs.
366 * Some code requires such alignment other code can be slower without
367 * correct alignment, for yet other it makes no difference.
368 *
369 * @note The linesize may be larger than the size of usable data -- there
370 * may be extra padding present for performance reasons.
371 *
372 * @attention In case of video, line size values can be negative to achieve
373 * a vertically inverted iteration over image lines.
374 */
376
377 /**
378 * pointers to the data planes/channels.
379 *
380 * For video, this should simply point to data[].
381 *
382 * For planar audio, each channel has a separate data pointer, and
383 * linesize[0] contains the size of each channel buffer.
384 * For packed audio, there is just one data pointer, and linesize[0]
385 * contains the total size of the buffer for all channels.
386 *
387 * Note: Both data and extended_data should always be set in a valid frame,
388 * but for planar audio with more channels that can fit in data,
389 * extended_data must be used in order to access all channels.
390 */
391 uint8_t **extended_data;
392
393 /**
394 * @name Video dimensions
395 * Video frames only. The coded dimensions (in pixels) of the video frame,
396 * i.e. the size of the rectangle that contains some well-defined values.
397 *
398 * @note The part of the frame intended for display/presentation is further
399 * restricted by the @ref cropping "Cropping rectangle".
400 * @{
401 */
403 /**
404 * @}
405 */
406
407 /**
408 * number of audio samples (per channel) described by this frame
409 */
411
412 /**
413 * format of the frame, -1 if unknown or unset
414 * Values correspond to enum AVPixelFormat for video frames,
415 * enum AVSampleFormat for audio)
416 */
418
419 /**
420 * 1 -> keyframe, 0-> not
421 */
423
424 /**
425 * Picture type of the frame.
426 */
428
429 /**
430 * Sample aspect ratio for the video frame, 0/1 if unknown/unspecified.
431 */
433
434 /**
435 * Presentation timestamp in time_base units (time when frame should be shown to user).
436 */
437 int64_t pts;
438
439 /**
440 * DTS copied from the AVPacket that triggered returning this frame. (if frame threading isn't used)
441 * This is also the Presentation time of this AVFrame calculated from
442 * only AVPacket.dts values without pts values.
443 */
444 int64_t pkt_dts;
445
446 /**
447 * Time base for the timestamps in this frame.
448 * In the future, this field may be set on frames output by decoders or
449 * filters, but its value will be by default ignored on input to encoders
450 * or filters.
451 */
453
454#if FF_API_FRAME_PICTURE_NUMBER
455 /**
456 * picture number in bitstream order
457 */
460 /**
461 * picture number in display order
462 */
465#endif
466
467 /**
468 * quality (between 1 (good) and FF_LAMBDA_MAX (bad))
469 */
471
472 /**
473 * for some private data of the user
474 */
475 void *opaque;
476
477 /**
478 * When decoding, this signals how much the picture must be delayed.
479 * extra_delay = repeat_pict / (2*fps)
480 */
482
483 /**
484 * The content of the picture is interlaced.
485 */
487
488 /**
489 * If the content is interlaced, is top field displayed first.
490 */
492
493 /**
494 * Tell user application that palette has changed from previous frame.
495 */
497
498#if FF_API_REORDERED_OPAQUE
499 /**
500 * reordered opaque 64 bits (generally an integer or a double precision float
501 * PTS but can be anything).
502 * The user sets AVCodecContext.reordered_opaque to represent the input at
503 * that time,
504 * the decoder reorders values as needed and sets AVFrame.reordered_opaque
505 * to exactly one of the values provided by the user through AVCodecContext.reordered_opaque
506 *
507 * @deprecated Use AV_CODEC_FLAG_COPY_OPAQUE instead
508 */
511#endif
512
513 /**
514 * Sample rate of the audio data.
515 */
517
518#if FF_API_OLD_CHANNEL_LAYOUT
519 /**
520 * Channel layout of the audio data.
521 * @deprecated use ch_layout instead
522 */
525#endif
526
527 /**
528 * AVBuffer references backing the data for this frame. All the pointers in
529 * data and extended_data must point inside one of the buffers in buf or
530 * extended_buf. This array must be filled contiguously -- if buf[i] is
531 * non-NULL then buf[j] must also be non-NULL for all j < i.
532 *
533 * There may be at most one AVBuffer per data plane, so for video this array
534 * always contains all the references. For planar audio with more than
535 * AV_NUM_DATA_POINTERS channels, there may be more buffers than can fit in
536 * this array. Then the extra AVBufferRef pointers are stored in the
537 * extended_buf array.
538 */
540
541 /**
542 * For planar audio which requires more than AV_NUM_DATA_POINTERS
543 * AVBufferRef pointers, this array will hold all the references which
544 * cannot fit into AVFrame.buf.
545 *
546 * Note that this is different from AVFrame.extended_data, which always
547 * contains all the pointers. This array only contains the extra pointers,
548 * which cannot fit into AVFrame.buf.
549 *
550 * This array is always allocated using av_malloc() by whoever constructs
551 * the frame. It is freed in av_frame_unref().
552 */
554 /**
555 * Number of elements in extended_buf.
556 */
558
561
562/**
563 * @defgroup lavu_frame_flags AV_FRAME_FLAGS
564 * @ingroup lavu_frame
565 * Flags describing additional frame properties.
566 *
567 * @{
568 */
569
570/**
571 * The frame data may be corrupted, e.g. due to decoding errors.
572 */
573#define AV_FRAME_FLAG_CORRUPT (1 << 0)
574/**
575 * A flag to mark the frames which need to be decoded, but shouldn't be output.
576 */
577#define AV_FRAME_FLAG_DISCARD (1 << 2)
578/**
579 * @}
580 */
581
582 /**
583 * Frame flags, a combination of @ref lavu_frame_flags
584 */
585 int flags;
586
587 /**
588 * MPEG vs JPEG YUV range.
589 * - encoding: Set by user
590 * - decoding: Set by libavcodec
591 */
593
595
597
598 /**
599 * YUV colorspace type.
600 * - encoding: Set by user
601 * - decoding: Set by libavcodec
602 */
604
606
607 /**
608 * frame timestamp estimated using various heuristics, in stream time base
609 * - encoding: unused
610 * - decoding: set by libavcodec, read by user.
611 */
613
614 /**
615 * reordered pos from the last AVPacket that has been input into the decoder
616 * - encoding: unused
617 * - decoding: Read by user.
618 */
619 int64_t pkt_pos;
620
621#if FF_API_PKT_DURATION
622 /**
623 * duration of the corresponding packet, expressed in
624 * AVStream->time_base units, 0 if unknown.
625 * - encoding: unused
626 * - decoding: Read by user.
627 *
628 * @deprecated use duration instead
629 */
632#endif
633
634 /**
635 * metadata.
636 * - encoding: Set by user.
637 * - decoding: Set by libavcodec.
638 */
640
641 /**
642 * decode error flags of the frame, set to a combination of
643 * FF_DECODE_ERROR_xxx flags if the decoder produced a frame, but there
644 * were errors during the decoding.
645 * - encoding: unused
646 * - decoding: set by libavcodec, read by user.
647 */
649#define FF_DECODE_ERROR_INVALID_BITSTREAM 1
650#define FF_DECODE_ERROR_MISSING_REFERENCE 2
651#define FF_DECODE_ERROR_CONCEALMENT_ACTIVE 4
652#define FF_DECODE_ERROR_DECODE_SLICES 8
653
654#if FF_API_OLD_CHANNEL_LAYOUT
655 /**
656 * number of audio channels, only used for audio.
657 * - encoding: unused
658 * - decoding: Read by user.
659 * @deprecated use ch_layout instead
660 */
663#endif
664
665 /**
666 * size of the corresponding packet containing the compressed
667 * frame.
668 * It is set to a negative value if unknown.
669 * - encoding: unused
670 * - decoding: set by libavcodec, read by user.
671 */
673
674 /**
675 * For hwaccel-format frames, this should be a reference to the
676 * AVHWFramesContext describing the frame.
677 */
679
680 /**
681 * AVBufferRef for free use by the API user. FFmpeg will never check the
682 * contents of the buffer ref. FFmpeg calls av_buffer_unref() on it when
683 * the frame is unreferenced. av_frame_copy_props() calls create a new
684 * reference with av_buffer_ref() for the target frame's opaque_ref field.
685 *
686 * This is unrelated to the opaque field, although it serves a similar
687 * purpose.
688 */
690
691 /**
692 * @anchor cropping
693 * @name Cropping
694 * Video frames only. The number of pixels to discard from the the
695 * top/bottom/left/right border of the frame to obtain the sub-rectangle of
696 * the frame intended for presentation.
697 * @{
698 */
699 size_t crop_top;
701 size_t crop_left;
703 /**
704 * @}
705 */
706
707 /**
708 * AVBufferRef for internal use by a single libav* library.
709 * Must not be used to transfer data between libraries.
710 * Has to be NULL when ownership of the frame leaves the respective library.
711 *
712 * Code outside the FFmpeg libs should never check or change the contents of the buffer ref.
713 *
714 * FFmpeg calls av_buffer_unref() on it when the frame is unreferenced.
715 * av_frame_copy_props() calls create a new reference with av_buffer_ref()
716 * for the target frame's private_ref field.
717 */
719
720 /**
721 * Channel layout of the audio data.
722 */
724
725 /**
726 * Duration of the frame, in the same units as pts. 0 if unknown.
727 */
728 int64_t duration;
729} AVFrame;
730
731
732/**
733 * Allocate an AVFrame and set its fields to default values. The resulting
734 * struct must be freed using av_frame_free().
735 *
736 * @return An AVFrame filled with default values or NULL on failure.
737 *
738 * @note this only allocates the AVFrame itself, not the data buffers. Those
739 * must be allocated through other means, e.g. with av_frame_get_buffer() or
740 * manually.
741 */
743
744/**
745 * Free the frame and any dynamically allocated objects in it,
746 * e.g. extended_data. If the frame is reference counted, it will be
747 * unreferenced first.
748 *
749 * @param frame frame to be freed. The pointer will be set to NULL.
750 */
752
753/**
754 * Set up a new reference to the data described by the source frame.
755 *
756 * Copy frame properties from src to dst and create a new reference for each
757 * AVBufferRef from src.
758 *
759 * If src is not reference counted, new buffers are allocated and the data is
760 * copied.
761 *
762 * @warning: dst MUST have been either unreferenced with av_frame_unref(dst),
763 * or newly allocated with av_frame_alloc() before calling this
764 * function, or undefined behavior will occur.
765 *
766 * @return 0 on success, a negative AVERROR on error
767 */
768int av_frame_ref(AVFrame *dst, const AVFrame *src);
769
770/**
771 * Create a new frame that references the same data as src.
772 *
773 * This is a shortcut for av_frame_alloc()+av_frame_ref().
774 *
775 * @return newly created AVFrame on success, NULL on error.
776 */
778
779/**
780 * Unreference all the buffers referenced by frame and reset the frame fields.
781 */
783
784/**
785 * Move everything contained in src to dst and reset src.
786 *
787 * @warning: dst is not unreferenced, but directly overwritten without reading
788 * or deallocating its contents. Call av_frame_unref(dst) manually
789 * before calling this function to ensure that no memory is leaked.
790 */
792
793/**
794 * Allocate new buffer(s) for audio or video data.
795 *
796 * The following fields must be set on frame before calling this function:
797 * - format (pixel format for video, sample format for audio)
798 * - width and height for video
799 * - nb_samples and ch_layout for audio
800 *
801 * This function will fill AVFrame.data and AVFrame.buf arrays and, if
802 * necessary, allocate and fill AVFrame.extended_data and AVFrame.extended_buf.
803 * For planar formats, one buffer will be allocated for each plane.
804 *
805 * @warning: if frame already has been allocated, calling this function will
806 * leak memory. In addition, undefined behavior can occur in certain
807 * cases.
808 *
809 * @param frame frame in which to store the new buffers.
810 * @param align Required buffer size alignment. If equal to 0, alignment will be
811 * chosen automatically for the current CPU. It is highly
812 * recommended to pass 0 here unless you know what you are doing.
813 *
814 * @return 0 on success, a negative AVERROR on error.
815 */
817
818/**
819 * Check if the frame data is writable.
820 *
821 * @return A positive value if the frame data is writable (which is true if and
822 * only if each of the underlying buffers has only one reference, namely the one
823 * stored in this frame). Return 0 otherwise.
824 *
825 * If 1 is returned the answer is valid until av_buffer_ref() is called on any
826 * of the underlying AVBufferRefs (e.g. through av_frame_ref() or directly).
827 *
828 * @see av_frame_make_writable(), av_buffer_is_writable()
829 */
831
832/**
833 * Ensure that the frame data is writable, avoiding data copy if possible.
834 *
835 * Do nothing if the frame is writable, allocate new buffers and copy the data
836 * if it is not. Non-refcounted frames behave as non-writable, i.e. a copy
837 * is always made.
838 *
839 * @return 0 on success, a negative AVERROR on error.
840 *
841 * @see av_frame_is_writable(), av_buffer_is_writable(),
842 * av_buffer_make_writable()
843 */
845
846/**
847 * Copy the frame data from src to dst.
848 *
849 * This function does not allocate anything, dst must be already initialized and
850 * allocated with the same parameters as src.
851 *
852 * This function only copies the frame data (i.e. the contents of the data /
853 * extended data arrays), not any other properties.
854 *
855 * @return >= 0 on success, a negative AVERROR on error.
856 */
857int av_frame_copy(AVFrame *dst, const AVFrame *src);
858
859/**
860 * Copy only "metadata" fields from src to dst.
861 *
862 * Metadata for the purpose of this function are those fields that do not affect
863 * the data layout in the buffers. E.g. pts, sample rate (for audio) or sample
864 * aspect ratio (for video), but not width/height or channel layout.
865 * Side data is also copied.
866 */
867int av_frame_copy_props(AVFrame *dst, const AVFrame *src);
868
869/**
870 * Get the buffer reference a given data plane is stored in.
871 *
872 * @param frame the frame to get the plane's buffer from
873 * @param plane index of the data plane of interest in frame->extended_data.
874 *
875 * @return the buffer reference that contains the plane or NULL if the input
876 * frame is not valid.
877 */
879
880/**
881 * Add a new side data to a frame.
882 *
883 * @param frame a frame to which the side data should be added
884 * @param type type of the added side data
885 * @param size size of the side data
886 *
887 * @return newly added side data on success, NULL on error
888 */
890 enum AVFrameSideDataType type,
891 size_t size);
892
893/**
894 * Add a new side data to a frame from an existing AVBufferRef
895 *
896 * @param frame a frame to which the side data should be added
897 * @param type the type of the added side data
898 * @param buf an AVBufferRef to add as side data. The ownership of
899 * the reference is transferred to the frame.
900 *
901 * @return newly added side data on success, NULL on error. On failure
902 * the frame is unchanged and the AVBufferRef remains owned by
903 * the caller.
904 */
906 enum AVFrameSideDataType type,
907 AVBufferRef *buf);
908
909/**
910 * @return a pointer to the side data of a given type on success, NULL if there
911 * is no side data with such type in this frame.
912 */
914 enum AVFrameSideDataType type);
915
916/**
917 * Remove and free all side data instances of the given type.
918 */
920
921
922/**
923 * Flags for frame cropping.
924 */
925enum {
926 /**
927 * Apply the maximum possible cropping, even if it requires setting the
928 * AVFrame.data[] entries to unaligned pointers. Passing unaligned data
929 * to FFmpeg API is generally not allowed, and causes undefined behavior
930 * (such as crashes). You can pass unaligned data only to FFmpeg APIs that
931 * are explicitly documented to accept it. Use this flag only if you
932 * absolutely know what you are doing.
933 */
935};
936
937/**
938 * Crop the given video AVFrame according to its crop_left/crop_top/crop_right/
939 * crop_bottom fields. If cropping is successful, the function will adjust the
940 * data pointers and the width/height fields, and set the crop fields to 0.
941 *
942 * In all cases, the cropping boundaries will be rounded to the inherent
943 * alignment of the pixel format. In some cases, such as for opaque hwaccel
944 * formats, the left/top cropping is ignored. The crop fields are set to 0 even
945 * if the cropping was rounded or ignored.
946 *
947 * @param frame the frame which should be cropped
948 * @param flags Some combination of AV_FRAME_CROP_* flags, or 0.
949 *
950 * @return >= 0 on success, a negative AVERROR on error. If the cropping fields
951 * were invalid, AVERROR(ERANGE) is returned, and nothing is changed.
952 */
954
955/**
956 * @return a string identifying the side data type
957 */
959
960/**
961 * @}
962 */
963
964#endif /* AVUTIL_FRAME_H */
#define attribute_deprecated
Definition: attributes.h:100
Convenience header that includes libavutil's core.
refcounted data buffer API
Public libavutil channel layout APIs header.
static AVFrame * frame
Definition: demux_decode.c:54
Public dictionary API.
#define AV_NUM_DATA_POINTERS
Definition: frame.h:331
struct AVDictionary AVDictionary
Definition: dict.h:94
void av_frame_unref(AVFrame *frame)
Unreference all the buffers referenced by frame and reset the frame fields.
AVFrameSideData * av_frame_get_side_data(const AVFrame *frame, enum AVFrameSideDataType type)
void av_frame_remove_side_data(AVFrame *frame, enum AVFrameSideDataType type)
Remove and free all side data instances of the given type.
int av_frame_is_writable(AVFrame *frame)
Check if the frame data is writable.
AVActiveFormatDescription
Definition: frame.h:219
int av_frame_get_buffer(AVFrame *frame, int align)
Allocate new buffer(s) for audio or video data.
void av_frame_move_ref(AVFrame *dst, AVFrame *src)
Move everything contained in src to dst and reset src.
int av_frame_ref(AVFrame *dst, const AVFrame *src)
Set up a new reference to the data described by the source frame.
AVBufferRef * av_frame_get_plane_buffer(AVFrame *frame, int plane)
Get the buffer reference a given data plane is stored in.
void av_frame_free(AVFrame **frame)
Free the frame and any dynamically allocated objects in it, e.g.
AVFrameSideData * av_frame_new_side_data(AVFrame *frame, enum AVFrameSideDataType type, size_t size)
Add a new side data to a frame.
AVFrameSideData * av_frame_new_side_data_from_buf(AVFrame *frame, enum AVFrameSideDataType type, AVBufferRef *buf)
Add a new side data to a frame from an existing AVBufferRef.
int av_frame_copy_props(AVFrame *dst, const AVFrame *src)
Copy only "metadata" fields from src to dst.
AVFrame * av_frame_alloc(void)
Allocate an AVFrame and set its fields to default values.
AVFrame * av_frame_clone(const AVFrame *src)
Create a new frame that references the same data as src.
int av_frame_make_writable(AVFrame *frame)
Ensure that the frame data is writable, avoiding data copy if possible.
AVFrameSideDataType
Definition: frame.h:49
int av_frame_apply_cropping(AVFrame *frame, int flags)
Crop the given video AVFrame according to its crop_left/crop_top/crop_right/ crop_bottom fields.
int av_frame_copy(AVFrame *dst, const AVFrame *src)
Copy the frame data from src to dst.
const char * av_frame_side_data_name(enum AVFrameSideDataType type)
@ AV_AFD_14_9
Definition: frame.h:223
@ AV_AFD_SAME
Definition: frame.h:220
@ AV_AFD_16_9_SP_14_9
Definition: frame.h:225
@ AV_AFD_4_3_SP_14_9
Definition: frame.h:224
@ AV_AFD_SP_4_3
Definition: frame.h:226
@ AV_AFD_4_3
Definition: frame.h:221
@ AV_AFD_16_9
Definition: frame.h:222
@ AV_FRAME_CROP_UNALIGNED
Apply the maximum possible cropping, even if it requires setting the AVFrame.data[] entries to unalig...
Definition: frame.h:934
@ AV_FRAME_DATA_GOP_TIMECODE
The GOP timecode in 25 bit timecode format.
Definition: frame.h:125
@ AV_FRAME_DATA_DOVI_METADATA
Parsed Dolby Vision metadata, suitable for passing to a software implementation.
Definition: frame.h:204
@ AV_FRAME_DATA_SPHERICAL
The data represents the AVSphericalMapping structure defined in libavutil/spherical....
Definition: frame.h:131
@ AV_FRAME_DATA_DYNAMIC_HDR_VIVID
HDR Vivid dynamic metadata associated with a video frame.
Definition: frame.h:211
@ AV_FRAME_DATA_CONTENT_LIGHT_LEVEL
Content light level (based on CTA-861.3).
Definition: frame.h:137
@ AV_FRAME_DATA_AMBIENT_VIEWING_ENVIRONMENT
Ambient viewing environment metadata, as defined by H.274.
Definition: frame.h:216
@ AV_FRAME_DATA_PANSCAN
The data is the AVPanScan struct defined in libavcodec.
Definition: frame.h:53
@ AV_FRAME_DATA_DISPLAYMATRIX
This side data contains a 3x3 transformation matrix describing an affine transformation that needs to...
Definition: frame.h:85
@ AV_FRAME_DATA_A53_CC
ATSC A53 Part 4 Closed Captions.
Definition: frame.h:59
@ AV_FRAME_DATA_SEI_UNREGISTERED
User data unregistered metadata associated with a video frame.
Definition: frame.h:178
@ AV_FRAME_DATA_AUDIO_SERVICE_TYPE
This side data must be associated with an audio frame and corresponds to enum AVAudioServiceType defi...
Definition: frame.h:114
@ AV_FRAME_DATA_DYNAMIC_HDR_PLUS
HDR dynamic metadata associated with a video frame.
Definition: frame.h:159
@ AV_FRAME_DATA_REPLAYGAIN
ReplayGain information in the form of the AVReplayGain struct.
Definition: frame.h:77
@ AV_FRAME_DATA_SKIP_SAMPLES
Recommmends skipping the specified number of samples.
Definition: frame.h:109
@ AV_FRAME_DATA_MASTERING_DISPLAY_METADATA
Mastering display metadata associated with a video frame.
Definition: frame.h:120
@ AV_FRAME_DATA_DOWNMIX_INFO
Metadata relevant to a downmix procedure.
Definition: frame.h:73
@ AV_FRAME_DATA_FILM_GRAIN_PARAMS
Film grain parameters for a frame, described by AVFilmGrainParams.
Definition: frame.h:184
@ AV_FRAME_DATA_VIDEO_ENC_PARAMS
Encoding parameters for a video frame, as described by AVVideoEncParams.
Definition: frame.h:170
@ AV_FRAME_DATA_DETECTION_BBOXES
Bounding boxes for object detection and classification, as described by AVDetectionBBoxHeader.
Definition: frame.h:190
@ AV_FRAME_DATA_ICC_PROFILE
The data contains an ICC profile as an opaque octet buffer following the format described by ISO 1507...
Definition: frame.h:144
@ AV_FRAME_DATA_AFD
Active Format Description data consisting of a single byte as specified in ETSI TS 101 154 using AVAc...
Definition: frame.h:90
@ AV_FRAME_DATA_S12M_TIMECODE
Timecode which conforms to SMPTE ST 12-1.
Definition: frame.h:152
@ AV_FRAME_DATA_MATRIXENCODING
The data is the AVMatrixEncoding enum defined in libavutil/channel_layout.h.
Definition: frame.h:68
@ AV_FRAME_DATA_DOVI_RPU_BUFFER
Dolby Vision RPU raw data, suitable for passing to x265 or other libraries.
Definition: frame.h:197
@ AV_FRAME_DATA_STEREO3D
Stereoscopic 3d metadata.
Definition: frame.h:64
@ AV_FRAME_DATA_MOTION_VECTORS
Motion vectors exported by some codecs (on demand through the export_mvs flag set in the libavcodec A...
Definition: frame.h:97
@ AV_FRAME_DATA_REGIONS_OF_INTEREST
Regions Of Interest, the data is an array of AVRegionOfInterest type, the number of array element is ...
Definition: frame.h:165
AVPictureType
Definition: avutil.h:272
swscale version macros
pixel format definitions
AVChromaLocation
Location of chroma samples.
Definition: pixfmt.h:680
AVColorRange
Visual content value range.
Definition: pixfmt.h:626
AVColorPrimaries
Chromaticity coordinates of the source primaries.
Definition: pixfmt.h:533
AVColorTransferCharacteristic
Color Transfer Characteristic.
Definition: pixfmt.h:558
AVColorSpace
YUV colorspace type.
Definition: pixfmt.h:587
Utilties for rational number calculation.
A reference to a data buffer.
Definition: buffer.h:82
An AVChannelLayout holds information about the channel layout of audio data.
Structure to hold side data for an AVFrame.
Definition: frame.h:236
enum AVFrameSideDataType type
Definition: frame.h:237
AVDictionary * metadata
Definition: frame.h:240
size_t size
Definition: frame.h:239
uint8_t * data
Definition: frame.h:238
AVBufferRef * buf
Definition: frame.h:241
This structure describes decoded (raw) audio or video data.
Definition: frame.h:330
int nb_samples
number of audio samples (per channel) described by this frame
Definition: frame.h:410
int64_t pts
Presentation timestamp in time_base units (time when frame should be shown to user).
Definition: frame.h:437
int64_t best_effort_timestamp
frame timestamp estimated using various heuristics, in stream time base
Definition: frame.h:612
attribute_deprecated int64_t reordered_opaque
reordered opaque 64 bits (generally an integer or a double precision float PTS but can be anything).
Definition: frame.h:510
size_t crop_right
Definition: frame.h:702
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
Definition: frame.h:351
enum AVChromaLocation chroma_location
Definition: frame.h:605
int width
Definition: frame.h:402
AVBufferRef ** extended_buf
For planar audio which requires more than AV_NUM_DATA_POINTERS AVBufferRef pointers,...
Definition: frame.h:553
AVBufferRef * hw_frames_ctx
For hwaccel-format frames, this should be a reference to the AVHWFramesContext describing the frame.
Definition: frame.h:678
int key_frame
1 -> keyframe, 0-> not
Definition: frame.h:422
AVRational time_base
Time base for the timestamps in this frame.
Definition: frame.h:452
void * opaque
for some private data of the user
Definition: frame.h:475
int64_t pkt_pos
reordered pos from the last AVPacket that has been input into the decoder
Definition: frame.h:619
int pkt_size
size of the corresponding packet containing the compressed frame.
Definition: frame.h:672
int decode_error_flags
decode error flags of the frame, set to a combination of FF_DECODE_ERROR_xxx flags if the decoder pro...
Definition: frame.h:648
int height
Definition: frame.h:402
AVFrameSideData ** side_data
Definition: frame.h:559
int flags
Frame flags, a combination of AV_FRAME_FLAGS.
Definition: frame.h:585
AVBufferRef * buf[AV_NUM_DATA_POINTERS]
AVBuffer references backing the data for this frame.
Definition: frame.h:539
attribute_deprecated int channels
number of audio channels, only used for audio.
Definition: frame.h:662
enum AVColorPrimaries color_primaries
Definition: frame.h:594
AVDictionary * metadata
metadata.
Definition: frame.h:639
int palette_has_changed
Tell user application that palette has changed from previous frame.
Definition: frame.h:496
AVRational sample_aspect_ratio
Sample aspect ratio for the video frame, 0/1 if unknown/unspecified.
Definition: frame.h:432
AVBufferRef * private_ref
AVBufferRef for internal use by a single libav* library.
Definition: frame.h:718
int top_field_first
If the content is interlaced, is top field displayed first.
Definition: frame.h:491
attribute_deprecated uint64_t channel_layout
Channel layout of the audio data.
Definition: frame.h:524
int quality
quality (between 1 (good) and FF_LAMBDA_MAX (bad))
Definition: frame.h:470
int interlaced_frame
The content of the picture is interlaced.
Definition: frame.h:486
AVBufferRef * opaque_ref
AVBufferRef for free use by the API user.
Definition: frame.h:689
size_t crop_top
Definition: frame.h:699
enum AVColorRange color_range
MPEG vs JPEG YUV range.
Definition: frame.h:592
int nb_side_data
Definition: frame.h:560
attribute_deprecated int64_t pkt_duration
duration of the corresponding packet, expressed in AVStream->time_base units, 0 if unknown.
Definition: frame.h:631
enum AVColorSpace colorspace
YUV colorspace type.
Definition: frame.h:603
int nb_extended_buf
Number of elements in extended_buf.
Definition: frame.h:557
attribute_deprecated int coded_picture_number
picture number in bitstream order
Definition: frame.h:459
int64_t pkt_dts
DTS copied from the AVPacket that triggered returning this frame.
Definition: frame.h:444
int linesize[AV_NUM_DATA_POINTERS]
For video, a positive or negative value, which is typically indicating the size in bytes of each pict...
Definition: frame.h:375
size_t crop_left
Definition: frame.h:701
enum AVColorTransferCharacteristic color_trc
Definition: frame.h:596
int64_t duration
Duration of the frame, in the same units as pts.
Definition: frame.h:728
int sample_rate
Sample rate of the audio data.
Definition: frame.h:516
AVChannelLayout ch_layout
Channel layout of the audio data.
Definition: frame.h:723
size_t crop_bottom
Definition: frame.h:700
int format
format of the frame, -1 if unknown or unset Values correspond to enum AVPixelFormat for video frames,...
Definition: frame.h:417
int repeat_pict
When decoding, this signals how much the picture must be delayed.
Definition: frame.h:481
attribute_deprecated int display_picture_number
picture number in display order
Definition: frame.h:464
enum AVPictureType pict_type
Picture type of the frame.
Definition: frame.h:427
uint8_t ** extended_data
pointers to the data planes/channels.
Definition: frame.h:391
Rational number (pair of numerator and denominator).
Definition: rational.h:58
Structure describing a single Region Of Interest.
Definition: frame.h:255
uint32_t self_size
Must be set to the size of this data structure (that is, sizeof(AVRegionOfInterest)).
Definition: frame.h:260
AVRational qoffset
Quantisation offset.
Definition: frame.h:297
int top
Distance in pixels from the top edge of the frame to the top and bottom edges and from the left edge ...
Definition: frame.h:270