frame: allow align=0 (meaning automatic) for av_frame_get_buffer()
[libav.git] / libavutil / frame.h
1 /*
2 * This file is part of Libav.
3 *
4 * Libav 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 * Libav 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 Libav; 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 "dict.h"
34 #include "rational.h"
35 #include "samplefmt.h"
36 #include "pixfmt.h"
37 #include "version.h"
38
39
40 /**
41 * @defgroup lavu_frame AVFrame
42 * @ingroup lavu_data
43 *
44 * @{
45 * AVFrame is an abstraction for reference-counted raw multimedia data.
46 */
47
48 enum AVFrameSideDataType {
49 /**
50 * The data is the AVPanScan struct defined in libavcodec.
51 */
52 AV_FRAME_DATA_PANSCAN,
53 /**
54 * ATSC A53 Part 4 Closed Captions.
55 * A53 CC bitstream is stored as uint8_t in AVFrameSideData.data.
56 * The number of bytes of CC data is AVFrameSideData.size.
57 */
58 AV_FRAME_DATA_A53_CC,
59 /**
60 * Stereoscopic 3d metadata.
61 * The data is the AVStereo3D struct defined in libavutil/stereo3d.h.
62 */
63 AV_FRAME_DATA_STEREO3D,
64 /**
65 * The data is the AVMatrixEncoding enum defined in libavutil/channel_layout.h.
66 */
67 AV_FRAME_DATA_MATRIXENCODING,
68 /**
69 * Metadata relevant to a downmix procedure.
70 * The data is the AVDownmixInfo struct defined in libavutil/downmix_info.h.
71 */
72 AV_FRAME_DATA_DOWNMIX_INFO,
73 /**
74 * ReplayGain information in the form of the AVReplayGain struct.
75 */
76 AV_FRAME_DATA_REPLAYGAIN,
77 /**
78 * This side data contains a 3x3 transformation matrix describing an affine
79 * transformation that needs to be applied to the frame for correct
80 * presentation.
81 *
82 * See libavutil/display.h for a detailed description of the data.
83 */
84 AV_FRAME_DATA_DISPLAYMATRIX,
85 /**
86 * Active Format Description data consisting of a single byte as specified
87 * in ETSI TS 101 154 using enum AVActiveFormatDescription.
88 */
89 AV_FRAME_DATA_AFD,
90
91 /**
92 * This side data must be associated with an audio frame and corresponds to
93 * enum AVAudioServiceType defined in avcodec.h.
94 */
95 AV_FRAME_DATA_AUDIO_SERVICE_TYPE,
96
97 /**
98 * The data represents the AVSphericalMapping structure defined in
99 * libavutil/spherical.h.
100 */
101 AV_FRAME_DATA_SPHERICAL,
102 };
103
104 enum AVActiveFormatDescription {
105 AV_AFD_SAME = 8,
106 AV_AFD_4_3 = 9,
107 AV_AFD_16_9 = 10,
108 AV_AFD_14_9 = 11,
109 AV_AFD_4_3_SP_14_9 = 13,
110 AV_AFD_16_9_SP_14_9 = 14,
111 AV_AFD_SP_4_3 = 15,
112 };
113
114 typedef struct AVFrameSideData {
115 enum AVFrameSideDataType type;
116 uint8_t *data;
117 int size;
118 AVDictionary *metadata;
119 } AVFrameSideData;
120
121 /**
122 * This structure describes decoded (raw) audio or video data.
123 *
124 * AVFrame must be allocated using av_frame_alloc(). Note that this only
125 * allocates the AVFrame itself, the buffers for the data must be managed
126 * through other means (see below).
127 * AVFrame must be freed with av_frame_free().
128 *
129 * AVFrame is typically allocated once and then reused multiple times to hold
130 * different data (e.g. a single AVFrame to hold frames received from a
131 * decoder). In such a case, av_frame_unref() will free any references held by
132 * the frame and reset it to its original clean state before it
133 * is reused again.
134 *
135 * The data described by an AVFrame is usually reference counted through the
136 * AVBuffer API. The underlying buffer references are stored in AVFrame.buf /
137 * AVFrame.extended_buf. An AVFrame is considered to be reference counted if at
138 * least one reference is set, i.e. if AVFrame.buf[0] != NULL. In such a case,
139 * every single data plane must be contained in one of the buffers in
140 * AVFrame.buf or AVFrame.extended_buf.
141 * There may be a single buffer for all the data, or one separate buffer for
142 * each plane, or anything in between.
143 *
144 * sizeof(AVFrame) is not a part of the public ABI, so new fields may be added
145 * to the end with a minor bump.
146 */
147 typedef struct AVFrame {
148 #define AV_NUM_DATA_POINTERS 8
149 /**
150 * pointer to the picture/channel planes.
151 * This might be different from the first allocated byte
152 */
153 uint8_t *data[AV_NUM_DATA_POINTERS];
154
155 /**
156 * For video, size in bytes of each picture line.
157 * For audio, size in bytes of each plane.
158 *
159 * For audio, only linesize[0] may be set. For planar audio, each channel
160 * plane must be the same size.
161 *
162 * @note The linesize may be larger than the size of usable data -- there
163 * may be extra padding present for performance reasons.
164 */
165 int linesize[AV_NUM_DATA_POINTERS];
166
167 /**
168 * pointers to the data planes/channels.
169 *
170 * For video, this should simply point to data[].
171 *
172 * For planar audio, each channel has a separate data pointer, and
173 * linesize[0] contains the size of each channel buffer.
174 * For packed audio, there is just one data pointer, and linesize[0]
175 * contains the total size of the buffer for all channels.
176 *
177 * Note: Both data and extended_data should always be set in a valid frame,
178 * but for planar audio with more channels that can fit in data,
179 * extended_data must be used in order to access all channels.
180 */
181 uint8_t **extended_data;
182
183 /**
184 * @name Video dimensions
185 * Video frames only. The coded dimensions (in pixels) of the video frame,
186 * i.e. the size of the rectangle that contains some well-defined values.
187 *
188 * @note The part of the frame intended for display/presentation is further
189 * restricted by the @ref cropping "Cropping rectangle".
190 * @{
191 */
192 int width, height;
193 /**
194 * @}
195 */
196
197 /**
198 * number of audio samples (per channel) described by this frame
199 */
200 int nb_samples;
201
202 /**
203 * format of the frame, -1 if unknown or unset
204 * Values correspond to enum AVPixelFormat for video frames,
205 * enum AVSampleFormat for audio)
206 */
207 int format;
208
209 /**
210 * 1 -> keyframe, 0-> not
211 */
212 int key_frame;
213
214 /**
215 * Picture type of the frame.
216 */
217 enum AVPictureType pict_type;
218
219 /**
220 * Sample aspect ratio for the video frame, 0/1 if unknown/unspecified.
221 */
222 AVRational sample_aspect_ratio;
223
224 /**
225 * Presentation timestamp in time_base units (time when frame should be shown to user).
226 */
227 int64_t pts;
228
229 #if FF_API_PKT_PTS
230 /**
231 * PTS copied from the AVPacket that was decoded to produce this frame.
232 * @deprecated use the pts field instead
233 */
234 attribute_deprecated
235 int64_t pkt_pts;
236 #endif
237
238 /**
239 * DTS copied from the AVPacket that triggered returning this frame.
240 */
241 int64_t pkt_dts;
242
243 /**
244 * picture number in bitstream order
245 */
246 int coded_picture_number;
247 /**
248 * picture number in display order
249 */
250 int display_picture_number;
251
252 /**
253 * quality (between 1 (good) and FF_LAMBDA_MAX (bad))
254 */
255 int quality;
256
257 /**
258 * for some private data of the user
259 */
260 void *opaque;
261
262 #if FF_API_ERROR_FRAME
263 /**
264 * @deprecated unused
265 */
266 attribute_deprecated
267 uint64_t error[AV_NUM_DATA_POINTERS];
268 #endif
269
270 /**
271 * When decoding, this signals how much the picture must be delayed.
272 * extra_delay = repeat_pict / (2*fps)
273 */
274 int repeat_pict;
275
276 /**
277 * The content of the picture is interlaced.
278 */
279 int interlaced_frame;
280
281 /**
282 * If the content is interlaced, is top field displayed first.
283 */
284 int top_field_first;
285
286 /**
287 * Tell user application that palette has changed from previous frame.
288 */
289 int palette_has_changed;
290
291 /**
292 * reordered opaque 64 bits (generally an integer or a double precision float
293 * PTS but can be anything).
294 * The user sets AVCodecContext.reordered_opaque to represent the input at
295 * that time,
296 * the decoder reorders values as needed and sets AVFrame.reordered_opaque
297 * to exactly one of the values provided by the user through AVCodecContext.reordered_opaque
298 * @deprecated in favor of pkt_pts
299 */
300 int64_t reordered_opaque;
301
302 /**
303 * Sample rate of the audio data.
304 */
305 int sample_rate;
306
307 /**
308 * Channel layout of the audio data.
309 */
310 uint64_t channel_layout;
311
312 /**
313 * AVBuffer references backing the data for this frame. If all elements of
314 * this array are NULL, then this frame is not reference counted. This array
315 * must be filled contiguously -- if buf[i] is non-NULL then buf[j] must
316 * also be non-NULL for all j < i.
317 *
318 * There may be at most one AVBuffer per data plane, so for video this array
319 * always contains all the references. For planar audio with more than
320 * AV_NUM_DATA_POINTERS channels, there may be more buffers than can fit in
321 * this array. Then the extra AVBufferRef pointers are stored in the
322 * extended_buf array.
323 */
324 AVBufferRef *buf[AV_NUM_DATA_POINTERS];
325
326 /**
327 * For planar audio which requires more than AV_NUM_DATA_POINTERS
328 * AVBufferRef pointers, this array will hold all the references which
329 * cannot fit into AVFrame.buf.
330 *
331 * Note that this is different from AVFrame.extended_data, which always
332 * contains all the pointers. This array only contains the extra pointers,
333 * which cannot fit into AVFrame.buf.
334 *
335 * This array is always allocated using av_malloc() by whoever constructs
336 * the frame. It is freed in av_frame_unref().
337 */
338 AVBufferRef **extended_buf;
339 /**
340 * Number of elements in extended_buf.
341 */
342 int nb_extended_buf;
343
344 AVFrameSideData **side_data;
345 int nb_side_data;
346
347 /**
348 * @defgroup lavu_frame_flags AV_FRAME_FLAGS
349 * Flags describing additional frame properties.
350 *
351 * @{
352 */
353
354 /**
355 * The frame data may be corrupted, e.g. due to decoding errors.
356 */
357 #define AV_FRAME_FLAG_CORRUPT (1 << 0)
358 /**
359 * @}
360 */
361
362 /**
363 * Frame flags, a combination of @ref lavu_frame_flags
364 */
365 int flags;
366
367 enum AVColorRange color_range;
368
369 enum AVColorPrimaries color_primaries;
370
371 enum AVColorTransferCharacteristic color_trc;
372
373 enum AVColorSpace colorspace;
374
375 enum AVChromaLocation chroma_location;
376
377 /**
378 * For hwaccel-format frames, this should be a reference to the
379 * AVHWFramesContext describing the frame.
380 */
381 AVBufferRef *hw_frames_ctx;
382
383 /**
384 * @anchor cropping
385 * @name Cropping
386 * Video frames only. The number of pixels to discard from the the
387 * top/bottom/left/right border of the frame to obtain the sub-rectangle of
388 * the frame intended for presentation.
389 * @{
390 */
391 size_t crop_top;
392 size_t crop_bottom;
393 size_t crop_left;
394 size_t crop_right;
395 /**
396 * @}
397 */
398 } AVFrame;
399
400 /**
401 * Allocate an AVFrame and set its fields to default values. The resulting
402 * struct must be freed using av_frame_free().
403 *
404 * @return An AVFrame filled with default values or NULL on failure.
405 *
406 * @note this only allocates the AVFrame itself, not the data buffers. Those
407 * must be allocated through other means, e.g. with av_frame_get_buffer() or
408 * manually.
409 */
410 AVFrame *av_frame_alloc(void);
411
412 /**
413 * Free the frame and any dynamically allocated objects in it,
414 * e.g. extended_data. If the frame is reference counted, it will be
415 * unreferenced first.
416 *
417 * @param frame frame to be freed. The pointer will be set to NULL.
418 */
419 void av_frame_free(AVFrame **frame);
420
421 /**
422 * Set up a new reference to the data described by the source frame.
423 *
424 * Copy frame properties from src to dst and create a new reference for each
425 * AVBufferRef from src.
426 *
427 * If src is not reference counted, new buffers are allocated and the data is
428 * copied.
429 *
430 * @warning: dst MUST have been either unreferenced with av_frame_unref(dst),
431 * or newly allocated with av_frame_alloc() before calling this
432 * function, or undefined behavior will occur.
433 *
434 * @return 0 on success, a negative AVERROR on error
435 */
436 int av_frame_ref(AVFrame *dst, const AVFrame *src);
437
438 /**
439 * Create a new frame that references the same data as src.
440 *
441 * This is a shortcut for av_frame_alloc()+av_frame_ref().
442 *
443 * @return newly created AVFrame on success, NULL on error.
444 */
445 AVFrame *av_frame_clone(const AVFrame *src);
446
447 /**
448 * Unreference all the buffers referenced by frame and reset the frame fields.
449 */
450 void av_frame_unref(AVFrame *frame);
451
452 /**
453 * Move everything contained in src to dst and reset src.
454 *
455 * @warning: dst is not unreferenced, but directly overwritten without reading
456 * or deallocating its contents. Call av_frame_unref(dst) manually
457 * before calling this function to ensure that no memory is leaked.
458 */
459 void av_frame_move_ref(AVFrame *dst, AVFrame *src);
460
461 /**
462 * Allocate new buffer(s) for audio or video data.
463 *
464 * The following fields must be set on frame before calling this function:
465 * - format (pixel format for video, sample format for audio)
466 * - width and height for video
467 * - nb_samples and channel_layout for audio
468 *
469 * This function will fill AVFrame.data and AVFrame.buf arrays and, if
470 * necessary, allocate and fill AVFrame.extended_data and AVFrame.extended_buf.
471 * For planar formats, one buffer will be allocated for each plane.
472 *
473 * @warning: if frame already has been allocated, calling this function will
474 * leak memory. In addition, undefined behavior can occur in certain
475 * cases.
476 *
477 * @param frame frame in which to store the new buffers.
478 * @param align Required buffer size alignment. If equal to 0, alignment will be
479 * chosen automatically for the current CPU. It is highly
480 * recommended to pass 0 here unless you know what you are doing.
481 *
482 * @return 0 on success, a negative AVERROR on error.
483 */
484 int av_frame_get_buffer(AVFrame *frame, int align);
485
486 /**
487 * Check if the frame data is writable.
488 *
489 * @return A positive value if the frame data is writable (which is true if and
490 * only if each of the underlying buffers has only one reference, namely the one
491 * stored in this frame). Return 0 otherwise.
492 *
493 * If 1 is returned the answer is valid until av_buffer_ref() is called on any
494 * of the underlying AVBufferRefs (e.g. through av_frame_ref() or directly).
495 *
496 * @see av_frame_make_writable(), av_buffer_is_writable()
497 */
498 int av_frame_is_writable(AVFrame *frame);
499
500 /**
501 * Ensure that the frame data is writable, avoiding data copy if possible.
502 *
503 * Do nothing if the frame is writable, allocate new buffers and copy the data
504 * if it is not.
505 *
506 * @return 0 on success, a negative AVERROR on error.
507 *
508 * @see av_frame_is_writable(), av_buffer_is_writable(),
509 * av_buffer_make_writable()
510 */
511 int av_frame_make_writable(AVFrame *frame);
512
513 /**
514 * Copy the frame data from src to dst.
515 *
516 * This function does not allocate anything, dst must be already initialized and
517 * allocated with the same parameters as src.
518 *
519 * This function only copies the frame data (i.e. the contents of the data /
520 * extended data arrays), not any other properties.
521 *
522 * @return >= 0 on success, a negative AVERROR on error.
523 */
524 int av_frame_copy(AVFrame *dst, const AVFrame *src);
525
526 /**
527 * Copy only "metadata" fields from src to dst.
528 *
529 * Metadata for the purpose of this function are those fields that do not affect
530 * the data layout in the buffers. E.g. pts, sample rate (for audio) or sample
531 * aspect ratio (for video), but not width/height or channel layout.
532 * Side data is also copied.
533 */
534 int av_frame_copy_props(AVFrame *dst, const AVFrame *src);
535
536 /**
537 * Get the buffer reference a given data plane is stored in.
538 *
539 * @param plane index of the data plane of interest in frame->extended_data.
540 *
541 * @return the buffer reference that contains the plane or NULL if the input
542 * frame is not valid.
543 */
544 AVBufferRef *av_frame_get_plane_buffer(AVFrame *frame, int plane);
545
546 /**
547 * Add a new side data to a frame.
548 *
549 * @param frame a frame to which the side data should be added
550 * @param type type of the added side data
551 * @param size size of the side data
552 *
553 * @return newly added side data on success, NULL on error
554 */
555 AVFrameSideData *av_frame_new_side_data(AVFrame *frame,
556 enum AVFrameSideDataType type,
557 int size);
558
559 /**
560 * @return a pointer to the side data of a given type on success, NULL if there
561 * is no side data with such type in this frame.
562 */
563 AVFrameSideData *av_frame_get_side_data(const AVFrame *frame,
564 enum AVFrameSideDataType type);
565
566 /**
567 * If side data of the supplied type exists in the frame, free it and remove it
568 * from the frame.
569 */
570 void av_frame_remove_side_data(AVFrame *frame, enum AVFrameSideDataType type);
571
572 /**
573 * @}
574 */
575
576 #endif /* AVUTIL_FRAME_H */