FFmpeg
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
frame.h
Go to the documentation of this file.
1 /*
2  *
3  * This file is part of FFmpeg.
4  *
5  * FFmpeg is free software; you can redistribute it and/or
6  * modify it under the terms of the GNU Lesser General Public
7  * License as published by the Free Software Foundation; either
8  * version 2.1 of the License, or (at your option) any later version.
9  *
10  * FFmpeg is distributed in the hope that it will be useful,
11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13  * Lesser General Public License for more details.
14  *
15  * You should have received a copy of the GNU Lesser General Public
16  * License along with FFmpeg; if not, write to the Free Software
17  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
18  */
19 
20 /**
21  * @file
22  * @ingroup lavu_frame
23  * reference-counted frame API
24  */
25 
26 #ifndef AVUTIL_FRAME_H
27 #define AVUTIL_FRAME_H
28 
29 #include <stdint.h>
30 
31 #include "avutil.h"
32 #include "buffer.h"
33 #include "dict.h"
34 #include "rational.h"
35 #include "samplefmt.h"
36 #include "version.h"
37 
38 
41  AVCOL_SPC_BT709 = 1, ///< also ITU-R BT1361 / IEC 61966-2-4 xvYCC709 / SMPTE RP177 Annex B
44  AVCOL_SPC_BT470BG = 5, ///< also ITU-R BT601-6 625 / ITU-R BT1358 625 / ITU-R BT1700 625 PAL & SECAM / IEC 61966-2-4 xvYCC601
45  AVCOL_SPC_SMPTE170M = 6, ///< also ITU-R BT601-6 525 / ITU-R BT1358 525 / ITU-R BT1700 NTSC / functionally identical to above
47  AVCOL_SPC_YCOCG = 8, ///< Used by Dirac / VC-2 and H.264 FRext, see ITU-T SG16
48  AVCOL_SPC_BT2020_NCL = 9, ///< ITU-R BT2020 non-constant luminance system
49  AVCOL_SPC_BT2020_CL = 10, ///< ITU-R BT2020 constant luminance system
50  AVCOL_SPC_NB , ///< Not part of ABI
51 };
52 #define AVCOL_SPC_YCGCO AVCOL_SPC_YCOCG
53 
56  AVCOL_RANGE_MPEG = 1, ///< the normal 219*2^(n-8) "MPEG" YUV ranges
57  AVCOL_RANGE_JPEG = 2, ///< the normal 2^n-1 "JPEG" YUV ranges
58  AVCOL_RANGE_NB , ///< Not part of ABI
59 };
60 
61 
62 /**
63  * @defgroup lavu_frame AVFrame
64  * @ingroup lavu_data
65  *
66  * @{
67  * AVFrame is an abstraction for reference-counted raw multimedia data.
68  */
69 
71  /**
72  * The data is the AVPanScan struct defined in libavcodec.
73  */
75  /**
76  * ATSC A53 Part 4 Closed Captions.
77  * A53 CC bitstream is stored as uint8_t in AVFrameSideData.data.
78  * The number of bytes of CC data is AVFrameSideData.size.
79  */
81  /**
82  * Stereoscopic 3d metadata.
83  * The data is the AVStereo3D struct defined in libavutil/stereo3d.h.
84  */
86  /**
87  * The data is the AVMatrixEncoding enum defined in libavutil/channel_layout.h.
88  */
90  /**
91  * Metadata relevant to a downmix procedure.
92  * The data is the AVDownmixInfo struct defined in libavutil/downmix_info.h.
93  */
95 };
96 
97 typedef struct AVFrameSideData {
100  int size;
103 
104 /**
105  * This structure describes decoded (raw) audio or video data.
106  *
107  * AVFrame must be allocated using av_frame_alloc(). Note that this only
108  * allocates the AVFrame itself, the buffers for the data must be managed
109  * through other means (see below).
110  * AVFrame must be freed with av_frame_free().
111  *
112  * AVFrame is typically allocated once and then reused multiple times to hold
113  * different data (e.g. a single AVFrame to hold frames received from a
114  * decoder). In such a case, av_frame_unref() will free any references held by
115  * the frame and reset it to its original clean state before it
116  * is reused again.
117  *
118  * The data described by an AVFrame is usually reference counted through the
119  * AVBuffer API. The underlying buffer references are stored in AVFrame.buf /
120  * AVFrame.extended_buf. An AVFrame is considered to be reference counted if at
121  * least one reference is set, i.e. if AVFrame.buf[0] != NULL. In such a case,
122  * every single data plane must be contained in one of the buffers in
123  * AVFrame.buf or AVFrame.extended_buf.
124  * There may be a single buffer for all the data, or one separate buffer for
125  * each plane, or anything in between.
126  *
127  * sizeof(AVFrame) is not a part of the public ABI, so new fields may be added
128  * to the end with a minor bump.
129  * Similarly fields that are marked as to be only accessed by
130  * av_opt_ptr() can be reordered. This allows 2 forks to add fields
131  * without breaking compatibility with each other.
132  */
133 typedef struct AVFrame {
134 #define AV_NUM_DATA_POINTERS 8
135  /**
136  * pointer to the picture/channel planes.
137  * This might be different from the first allocated byte
138  *
139  * Some decoders access areas outside 0,0 - width,height, please
140  * see avcodec_align_dimensions2(). Some filters and swscale can read
141  * up to 16 bytes beyond the planes, if these filters are to be used,
142  * then 16 extra bytes must be allocated.
143  */
145 
146  /**
147  * For video, size in bytes of each picture line.
148  * For audio, size in bytes of each plane.
149  *
150  * For audio, only linesize[0] may be set. For planar audio, each channel
151  * plane must be the same size.
152  *
153  * For video the linesizes should be multiplies of the CPUs alignment
154  * preference, this is 16 or 32 for modern desktop CPUs.
155  * Some code requires such alignment other code can be slower without
156  * correct alignment, for yet other it makes no difference.
157  *
158  * @note The linesize may be larger than the size of usable data -- there
159  * may be extra padding present for performance reasons.
160  */
162 
163  /**
164  * pointers to the data planes/channels.
165  *
166  * For video, this should simply point to data[].
167  *
168  * For planar audio, each channel has a separate data pointer, and
169  * linesize[0] contains the size of each channel buffer.
170  * For packed audio, there is just one data pointer, and linesize[0]
171  * contains the total size of the buffer for all channels.
172  *
173  * Note: Both data and extended_data should always be set in a valid frame,
174  * but for planar audio with more channels that can fit in data,
175  * extended_data must be used in order to access all channels.
176  */
178 
179  /**
180  * width and height of the video frame
181  */
182  int width, height;
183 
184  /**
185  * number of audio samples (per channel) described by this frame
186  */
188 
189  /**
190  * format of the frame, -1 if unknown or unset
191  * Values correspond to enum AVPixelFormat for video frames,
192  * enum AVSampleFormat for audio)
193  */
194  int format;
195 
196  /**
197  * 1 -> keyframe, 0-> not
198  */
200 
201  /**
202  * Picture type of the frame.
203  */
205 
206 #if FF_API_AVFRAME_LAVC
209 #endif
210 
211  /**
212  * Sample aspect ratio for the video frame, 0/1 if unknown/unspecified.
213  */
215 
216  /**
217  * Presentation timestamp in time_base units (time when frame should be shown to user).
218  */
219  int64_t pts;
220 
221  /**
222  * PTS copied from the AVPacket that was decoded to produce this frame.
223  */
224  int64_t pkt_pts;
225 
226  /**
227  * DTS copied from the AVPacket that triggered returning this frame. (if frame threading isnt used)
228  * This is also the Presentation time of this AVFrame calculated from
229  * only AVPacket.dts values without pts values.
230  */
231  int64_t pkt_dts;
232 
233  /**
234  * picture number in bitstream order
235  */
237  /**
238  * picture number in display order
239  */
241 
242  /**
243  * quality (between 1 (good) and FF_LAMBDA_MAX (bad))
244  */
245  int quality;
246 
247 #if FF_API_AVFRAME_LAVC
250 
251  /**
252  * QP table
253  */
255  int8_t *qscale_table;
256  /**
257  * QP store stride
258  */
260  int qstride;
261 
264 
265  /**
266  * mbskip_table[mb]>=1 if MB didn't change
267  * stride= mb_width = (width+15)>>4
268  */
271 
272  /**
273  * motion vector table
274  * @code
275  * example:
276  * int mv_sample_log2= 4 - motion_subsample_log2;
277  * int mb_width= (width+15)>>4;
278  * int mv_stride= (mb_width << mv_sample_log2) + 1;
279  * motion_val[direction][x + y*mv_stride][0->mv_x, 1->mv_y];
280  * @endcode
281  */
283  int16_t (*motion_val[2])[2];
284 
285  /**
286  * macroblock type table
287  * mb_type_base + mb_width + 2
288  */
290  uint32_t *mb_type;
291 
292  /**
293  * DCT coefficients
294  */
296  short *dct_coeff;
297 
298  /**
299  * motion reference frame index
300  * the order in which these are stored can depend on the codec.
301  */
303  int8_t *ref_index[2];
304 #endif
305 
306  /**
307  * for some private data of the user
308  */
309  void *opaque;
310 
311  /**
312  * error
313  */
315 
316 #if FF_API_AVFRAME_LAVC
318  int type;
319 #endif
320 
321  /**
322  * When decoding, this signals how much the picture must be delayed.
323  * extra_delay = repeat_pict / (2*fps)
324  */
326 
327  /**
328  * The content of the picture is interlaced.
329  */
331 
332  /**
333  * If the content is interlaced, is top field displayed first.
334  */
336 
337  /**
338  * Tell user application that palette has changed from previous frame.
339  */
341 
342 #if FF_API_AVFRAME_LAVC
345 
346  /**
347  * Pan scan.
348  */
351 #endif
352 
353  /**
354  * reordered opaque 64bit (generally an integer or a double precision float
355  * PTS but can be anything).
356  * The user sets AVCodecContext.reordered_opaque to represent the input at
357  * that time,
358  * the decoder reorders values as needed and sets AVFrame.reordered_opaque
359  * to exactly one of the values provided by the user through AVCodecContext.reordered_opaque
360  * @deprecated in favor of pkt_pts
361  */
363 
364 #if FF_API_AVFRAME_LAVC
365  /**
366  * @deprecated this field is unused
367  */
369 
374 
375  /**
376  * log2 of the size of the block which a single vector in motion_val represents:
377  * (4->16x16, 3->8x8, 2-> 4x4, 1-> 2x2)
378  */
381 #endif
382 
383  /**
384  * Sample rate of the audio data.
385  */
387 
388  /**
389  * Channel layout of the audio data.
390  */
391  uint64_t channel_layout;
392 
393  /**
394  * AVBuffer references backing the data for this frame. If all elements of
395  * this array are NULL, then this frame is not reference counted.
396  *
397  * There may be at most one AVBuffer per data plane, so for video this array
398  * always contains all the references. For planar audio with more than
399  * AV_NUM_DATA_POINTERS channels, there may be more buffers than can fit in
400  * this array. Then the extra AVBufferRef pointers are stored in the
401  * extended_buf array.
402  */
404 
405  /**
406  * For planar audio which requires more than AV_NUM_DATA_POINTERS
407  * AVBufferRef pointers, this array will hold all the references which
408  * cannot fit into AVFrame.buf.
409  *
410  * Note that this is different from AVFrame.extended_data, which always
411  * contains all the pointers. This array only contains the extra pointers,
412  * which cannot fit into AVFrame.buf.
413  *
414  * This array is always allocated using av_malloc() by whoever constructs
415  * the frame. It is freed in av_frame_unref().
416  */
418  /**
419  * Number of elements in extended_buf.
420  */
422 
425 
426 /**
427  * @defgroup lavu_frame_flags AV_FRAME_FLAGS
428  * Flags describing additional frame properties.
429  *
430  * @{
431  */
432 
433 /**
434  * The frame data may be corrupted, e.g. due to decoding errors.
435  */
436 #define AV_FRAME_FLAG_CORRUPT (1 << 0)
437 /**
438  * @}
439  */
440 
441  /**
442  * Frame flags, a combination of @ref lavu_frame_flags
443  */
444  int flags;
445 
446  /**
447  * frame timestamp estimated using various heuristics, in stream time base
448  * Code outside libavcodec should access this field using:
449  * av_frame_get_best_effort_timestamp(frame)
450  * - encoding: unused
451  * - decoding: set by libavcodec, read by user.
452  */
454 
455  /**
456  * reordered pos from the last AVPacket that has been input into the decoder
457  * Code outside libavcodec should access this field using:
458  * av_frame_get_pkt_pos(frame)
459  * - encoding: unused
460  * - decoding: Read by user.
461  */
462  int64_t pkt_pos;
463 
464  /**
465  * duration of the corresponding packet, expressed in
466  * AVStream->time_base units, 0 if unknown.
467  * Code outside libavcodec should access this field using:
468  * av_frame_get_pkt_duration(frame)
469  * - encoding: unused
470  * - decoding: Read by user.
471  */
472  int64_t pkt_duration;
473 
474  /**
475  * metadata.
476  * Code outside libavcodec should access this field using:
477  * av_frame_get_metadata(frame)
478  * - encoding: Set by user.
479  * - decoding: Set by libavcodec.
480  */
482 
483  /**
484  * decode error flags of the frame, set to a combination of
485  * FF_DECODE_ERROR_xxx flags if the decoder produced a frame, but there
486  * were errors during the decoding.
487  * Code outside libavcodec should access this field using:
488  * av_frame_get_decode_error_flags(frame)
489  * - encoding: unused
490  * - decoding: set by libavcodec, read by user.
491  */
493 #define FF_DECODE_ERROR_INVALID_BITSTREAM 1
494 #define FF_DECODE_ERROR_MISSING_REFERENCE 2
495 
496  /**
497  * number of audio channels, only used for audio.
498  * Code outside libavcodec should access this field using:
499  * av_frame_get_channels(frame)
500  * - encoding: unused
501  * - decoding: Read by user.
502  */
503  int channels;
504 
505  /**
506  * size of the corresponding packet containing the compressed
507  * frame. It must be accessed using av_frame_get_pkt_size() and
508  * av_frame_set_pkt_size().
509  * It is set to a negative value if unknown.
510  * - encoding: unused
511  * - decoding: set by libavcodec, read by user.
512  */
513  int pkt_size;
514 
515  /**
516  * YUV colorspace type.
517  * It must be accessed using av_frame_get_colorspace() and
518  * av_frame_set_colorspace().
519  * - encoding: Set by user
520  * - decoding: Set by libavcodec
521  */
523 
524  /**
525  * MPEG vs JPEG YUV range.
526  * It must be accessed using av_frame_get_color_range() and
527  * av_frame_set_color_range().
528  * - encoding: Set by user
529  * - decoding: Set by libavcodec
530  */
532 
533 
534  /**
535  * Not to be accessed directly from outside libavutil
536  */
538 } AVFrame;
539 
540 /**
541  * Accessors for some AVFrame fields.
542  * The position of these field in the structure is not part of the ABI,
543  * they should not be accessed directly outside libavcodec.
544  */
547 int64_t av_frame_get_pkt_duration (const AVFrame *frame);
548 void av_frame_set_pkt_duration (AVFrame *frame, int64_t val);
549 int64_t av_frame_get_pkt_pos (const AVFrame *frame);
550 void av_frame_set_pkt_pos (AVFrame *frame, int64_t val);
551 int64_t av_frame_get_channel_layout (const AVFrame *frame);
553 int av_frame_get_channels (const AVFrame *frame);
564 int8_t *av_frame_get_qp_table(AVFrame *f, int *stride, int *type);
570 
571 /**
572  * Get the name of a colorspace.
573  * @return a static string identifying the colorspace; can be NULL.
574  */
575 const char *av_get_colorspace_name(enum AVColorSpace val);
576 
577 /**
578  * Allocate an AVFrame and set its fields to default values. The resulting
579  * struct must be freed using av_frame_free().
580  *
581  * @return An AVFrame filled with default values or NULL on failure.
582  *
583  * @note this only allocates the AVFrame itself, not the data buffers. Those
584  * must be allocated through other means, e.g. with av_frame_get_buffer() or
585  * manually.
586  */
587 AVFrame *av_frame_alloc(void);
588 
589 /**
590  * Free the frame and any dynamically allocated objects in it,
591  * e.g. extended_data. If the frame is reference counted, it will be
592  * unreferenced first.
593  *
594  * @param frame frame to be freed. The pointer will be set to NULL.
595  */
596 void av_frame_free(AVFrame **frame);
597 
598 /**
599  * Set up a new reference to the data described by the source frame.
600  *
601  * Copy frame properties from src to dst and create a new reference for each
602  * AVBufferRef from src.
603  *
604  * If src is not reference counted, new buffers are allocated and the data is
605  * copied.
606  *
607  * @return 0 on success, a negative AVERROR on error
608  */
609 int av_frame_ref(AVFrame *dst, const AVFrame *src);
610 
611 /**
612  * Create a new frame that references the same data as src.
613  *
614  * This is a shortcut for av_frame_alloc()+av_frame_ref().
615  *
616  * @return newly created AVFrame on success, NULL on error.
617  */
619 
620 /**
621  * Unreference all the buffers referenced by frame and reset the frame fields.
622  */
624 
625 /**
626  * Move everythnig contained in src to dst and reset src.
627  */
628 void av_frame_move_ref(AVFrame *dst, AVFrame *src);
629 
630 /**
631  * Allocate new buffer(s) for audio or video data.
632  *
633  * The following fields must be set on frame before calling this function:
634  * - format (pixel format for video, sample format for audio)
635  * - width and height for video
636  * - nb_samples and channel_layout for audio
637  *
638  * This function will fill AVFrame.data and AVFrame.buf arrays and, if
639  * necessary, allocate and fill AVFrame.extended_data and AVFrame.extended_buf.
640  * For planar formats, one buffer will be allocated for each plane.
641  *
642  * @param frame frame in which to store the new buffers.
643  * @param align required buffer size alignment
644  *
645  * @return 0 on success, a negative AVERROR on error.
646  */
648 
649 /**
650  * Check if the frame data is writable.
651  *
652  * @return A positive value if the frame data is writable (which is true if and
653  * only if each of the underlying buffers has only one reference, namely the one
654  * stored in this frame). Return 0 otherwise.
655  *
656  * If 1 is returned the answer is valid until av_buffer_ref() is called on any
657  * of the underlying AVBufferRefs (e.g. through av_frame_ref() or directly).
658  *
659  * @see av_frame_make_writable(), av_buffer_is_writable()
660  */
662 
663 /**
664  * Ensure that the frame data is writable, avoiding data copy if possible.
665  *
666  * Do nothing if the frame is writable, allocate new buffers and copy the data
667  * if it is not.
668  *
669  * @return 0 on success, a negative AVERROR on error.
670  *
671  * @see av_frame_is_writable(), av_buffer_is_writable(),
672  * av_buffer_make_writable()
673  */
675 
676 /**
677  * Copy the frame data from src to dst.
678  *
679  * This function does not allocate anything, dst must be already initialized and
680  * allocated with the same parameters as src.
681  *
682  * This function only copies the frame data (i.e. the contents of the data /
683  * extended data arrays), not any other properties.
684  *
685  * @return >= 0 on success, a negative AVERROR on error.
686  */
687 int av_frame_copy(AVFrame *dst, const AVFrame *src);
688 
689 /**
690  * Copy only "metadata" fields from src to dst.
691  *
692  * Metadata for the purpose of this function are those fields that do not affect
693  * the data layout in the buffers. E.g. pts, sample rate (for audio) or sample
694  * aspect ratio (for video), but not width/height or channel layout.
695  * Side data is also copied.
696  */
697 int av_frame_copy_props(AVFrame *dst, const AVFrame *src);
698 
699 /**
700  * Get the buffer reference a given data plane is stored in.
701  *
702  * @param plane index of the data plane of interest in frame->extended_data.
703  *
704  * @return the buffer reference that contains the plane or NULL if the input
705  * frame is not valid.
706  */
708 
709 /**
710  * Add a new side data to a frame.
711  *
712  * @param frame a frame to which the side data should be added
713  * @param type type of the added side data
714  * @param size size of the side data
715  *
716  * @return newly added side data on success, NULL on error
717  */
720  int size);
721 
722 /**
723  * @return a pointer to the side data of a given type on success, NULL if there
724  * is no side data with such type in this frame.
725  */
728 
729 /**
730  * @}
731  */
732 
733 #endif /* AVUTIL_FRAME_H */