3 * This file is part of FFmpeg.
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.
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.
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
23 * reference-counted frame API
26 #ifndef AVUTIL_FRAME_H
27 #define AVUTIL_FRAME_H
35 #include "samplefmt.h"
41 AVCOL_SPC_BT709 = 1, ///< also ITU-R BT1361 / IEC 61966-2-4 xvYCC709 / SMPTE RP177 Annex B
42 AVCOL_SPC_UNSPECIFIED = 2,
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
46 AVCOL_SPC_SMPTE240M = 7,
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
52 #define AVCOL_SPC_YCGCO AVCOL_SPC_YCOCG
55 AVCOL_RANGE_UNSPECIFIED = 0,
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
63 * @defgroup lavu_frame AVFrame
67 * AVFrame is an abstraction for reference-counted raw multimedia data.
70 enum AVFrameSideDataType {
72 * The data is the AVPanScan struct defined in libavcodec.
74 AV_FRAME_DATA_PANSCAN,
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.
82 * Stereoscopic 3d metadata.
83 * The data is the AVStereo3D struct defined in libavutil/stereo3d.h.
85 AV_FRAME_DATA_STEREO3D,
87 * The data is the AVMatrixEncoding enum defined in libavutil/channel_layout.h.
89 AV_FRAME_DATA_MATRIXENCODING,
91 * Metadata relevant to a downmix procedure.
92 * The data is the AVDownmixInfo struct defined in libavutil/downmix_info.h.
94 AV_FRAME_DATA_DOWNMIX_INFO,
96 * ReplayGain information in the form of the AVReplayGain struct.
98 AV_FRAME_DATA_REPLAYGAIN,
100 * This side data contains a 3x3 transformation matrix describing an affine
101 * transformation that needs to be applied to the frame for correct
104 * See libavutil/display.h for a detailed description of the data.
106 AV_FRAME_DATA_DISPLAYMATRIX,
109 typedef struct AVFrameSideData {
110 enum AVFrameSideDataType type;
113 AVDictionary *metadata;
117 * This structure describes decoded (raw) audio or video data.
119 * AVFrame must be allocated using av_frame_alloc(). Note that this only
120 * allocates the AVFrame itself, the buffers for the data must be managed
121 * through other means (see below).
122 * AVFrame must be freed with av_frame_free().
124 * AVFrame is typically allocated once and then reused multiple times to hold
125 * different data (e.g. a single AVFrame to hold frames received from a
126 * decoder). In such a case, av_frame_unref() will free any references held by
127 * the frame and reset it to its original clean state before it
130 * The data described by an AVFrame is usually reference counted through the
131 * AVBuffer API. The underlying buffer references are stored in AVFrame.buf /
132 * AVFrame.extended_buf. An AVFrame is considered to be reference counted if at
133 * least one reference is set, i.e. if AVFrame.buf[0] != NULL. In such a case,
134 * every single data plane must be contained in one of the buffers in
135 * AVFrame.buf or AVFrame.extended_buf.
136 * There may be a single buffer for all the data, or one separate buffer for
137 * each plane, or anything in between.
139 * sizeof(AVFrame) is not a part of the public ABI, so new fields may be added
140 * to the end with a minor bump.
141 * Similarly fields that are marked as to be only accessed by
142 * av_opt_ptr() can be reordered. This allows 2 forks to add fields
143 * without breaking compatibility with each other.
145 typedef struct AVFrame {
146 #define AV_NUM_DATA_POINTERS 8
148 * pointer to the picture/channel planes.
149 * This might be different from the first allocated byte
151 * Some decoders access areas outside 0,0 - width,height, please
152 * see avcodec_align_dimensions2(). Some filters and swscale can read
153 * up to 16 bytes beyond the planes, if these filters are to be used,
154 * then 16 extra bytes must be allocated.
156 uint8_t *data[AV_NUM_DATA_POINTERS];
159 * For video, size in bytes of each picture line.
160 * For audio, size in bytes of each plane.
162 * For audio, only linesize[0] may be set. For planar audio, each channel
163 * plane must be the same size.
165 * For video the linesizes should be multiplies of the CPUs alignment
166 * preference, this is 16 or 32 for modern desktop CPUs.
167 * Some code requires such alignment other code can be slower without
168 * correct alignment, for yet other it makes no difference.
170 * @note The linesize may be larger than the size of usable data -- there
171 * may be extra padding present for performance reasons.
173 int linesize[AV_NUM_DATA_POINTERS];
176 * pointers to the data planes/channels.
178 * For video, this should simply point to data[].
180 * For planar audio, each channel has a separate data pointer, and
181 * linesize[0] contains the size of each channel buffer.
182 * For packed audio, there is just one data pointer, and linesize[0]
183 * contains the total size of the buffer for all channels.
185 * Note: Both data and extended_data should always be set in a valid frame,
186 * but for planar audio with more channels that can fit in data,
187 * extended_data must be used in order to access all channels.
189 uint8_t **extended_data;
192 * width and height of the video frame
197 * number of audio samples (per channel) described by this frame
202 * format of the frame, -1 if unknown or unset
203 * Values correspond to enum AVPixelFormat for video frames,
204 * enum AVSampleFormat for audio)
209 * 1 -> keyframe, 0-> not
214 * Picture type of the frame.
216 enum AVPictureType pict_type;
218 #if FF_API_AVFRAME_LAVC
220 uint8_t *base[AV_NUM_DATA_POINTERS];
224 * Sample aspect ratio for the video frame, 0/1 if unknown/unspecified.
226 AVRational sample_aspect_ratio;
229 * Presentation timestamp in time_base units (time when frame should be shown to user).
234 * PTS copied from the AVPacket that was decoded to produce this frame.
239 * DTS copied from the AVPacket that triggered returning this frame. (if frame threading isnt used)
240 * This is also the Presentation time of this AVFrame calculated from
241 * only AVPacket.dts values without pts values.
246 * picture number in bitstream order
248 int coded_picture_number;
250 * picture number in display order
252 int display_picture_number;
255 * quality (between 1 (good) and FF_LAMBDA_MAX (bad))
259 #if FF_API_AVFRAME_LAVC
267 int8_t *qscale_table;
278 * mbskip_table[mb]>=1 if MB didn't change
279 * stride= mb_width = (width+15)>>4
282 uint8_t *mbskip_table;
285 * motion vector table
288 * int mv_sample_log2= 4 - motion_subsample_log2;
289 * int mb_width= (width+15)>>4;
290 * int mv_stride= (mb_width << mv_sample_log2) + 1;
291 * motion_val[direction][x + y*mv_stride][0->mv_x, 1->mv_y];
294 int16_t (*motion_val[2])[2];
297 * macroblock type table
298 * mb_type_base + mb_width + 2
310 * motion reference frame index
311 * the order in which these are stored can depend on the codec.
314 int8_t *ref_index[2];
318 * for some private data of the user
325 uint64_t error[AV_NUM_DATA_POINTERS];
327 #if FF_API_AVFRAME_LAVC
333 * When decoding, this signals how much the picture must be delayed.
334 * extra_delay = repeat_pict / (2*fps)
339 * The content of the picture is interlaced.
341 int interlaced_frame;
344 * If the content is interlaced, is top field displayed first.
349 * Tell user application that palette has changed from previous frame.
351 int palette_has_changed;
353 #if FF_API_AVFRAME_LAVC
361 struct AVPanScan *pan_scan;
365 * reordered opaque 64bit (generally an integer or a double precision float
366 * PTS but can be anything).
367 * The user sets AVCodecContext.reordered_opaque to represent the input at
369 * the decoder reorders values as needed and sets AVFrame.reordered_opaque
370 * to exactly one of the values provided by the user through AVCodecContext.reordered_opaque
371 * @deprecated in favor of pkt_pts
373 int64_t reordered_opaque;
375 #if FF_API_AVFRAME_LAVC
377 * @deprecated this field is unused
379 attribute_deprecated void *hwaccel_picture_private;
382 struct AVCodecContext *owner;
387 * log2 of the size of the block which a single vector in motion_val represents:
388 * (4->16x16, 3->8x8, 2-> 4x4, 1-> 2x2)
390 uint8_t motion_subsample_log2;
394 * Sample rate of the audio data.
399 * Channel layout of the audio data.
401 uint64_t channel_layout;
404 * AVBuffer references backing the data for this frame. If all elements of
405 * this array are NULL, then this frame is not reference counted.
407 * There may be at most one AVBuffer per data plane, so for video this array
408 * always contains all the references. For planar audio with more than
409 * AV_NUM_DATA_POINTERS channels, there may be more buffers than can fit in
410 * this array. Then the extra AVBufferRef pointers are stored in the
411 * extended_buf array.
413 AVBufferRef *buf[AV_NUM_DATA_POINTERS];
416 * For planar audio which requires more than AV_NUM_DATA_POINTERS
417 * AVBufferRef pointers, this array will hold all the references which
418 * cannot fit into AVFrame.buf.
420 * Note that this is different from AVFrame.extended_data, which always
421 * contains all the pointers. This array only contains the extra pointers,
422 * which cannot fit into AVFrame.buf.
424 * This array is always allocated using av_malloc() by whoever constructs
425 * the frame. It is freed in av_frame_unref().
427 AVBufferRef **extended_buf;
429 * Number of elements in extended_buf.
433 AVFrameSideData **side_data;
437 * @defgroup lavu_frame_flags AV_FRAME_FLAGS
438 * Flags describing additional frame properties.
444 * The frame data may be corrupted, e.g. due to decoding errors.
446 #define AV_FRAME_FLAG_CORRUPT (1 << 0)
452 * Frame flags, a combination of @ref lavu_frame_flags
457 * frame timestamp estimated using various heuristics, in stream time base
458 * Code outside libavcodec should access this field using:
459 * av_frame_get_best_effort_timestamp(frame)
461 * - decoding: set by libavcodec, read by user.
463 int64_t best_effort_timestamp;
466 * reordered pos from the last AVPacket that has been input into the decoder
467 * Code outside libavcodec should access this field using:
468 * av_frame_get_pkt_pos(frame)
470 * - decoding: Read by user.
475 * duration of the corresponding packet, expressed in
476 * AVStream->time_base units, 0 if unknown.
477 * Code outside libavcodec should access this field using:
478 * av_frame_get_pkt_duration(frame)
480 * - decoding: Read by user.
482 int64_t pkt_duration;
486 * Code outside libavcodec should access this field using:
487 * av_frame_get_metadata(frame)
488 * - encoding: Set by user.
489 * - decoding: Set by libavcodec.
491 AVDictionary *metadata;
494 * decode error flags of the frame, set to a combination of
495 * FF_DECODE_ERROR_xxx flags if the decoder produced a frame, but there
496 * were errors during the decoding.
497 * Code outside libavcodec should access this field using:
498 * av_frame_get_decode_error_flags(frame)
500 * - decoding: set by libavcodec, read by user.
502 int decode_error_flags;
503 #define FF_DECODE_ERROR_INVALID_BITSTREAM 1
504 #define FF_DECODE_ERROR_MISSING_REFERENCE 2
507 * number of audio channels, only used for audio.
508 * Code outside libavcodec should access this field using:
509 * av_frame_get_channels(frame)
511 * - decoding: Read by user.
516 * size of the corresponding packet containing the compressed
517 * frame. It must be accessed using av_frame_get_pkt_size() and
518 * av_frame_set_pkt_size().
519 * It is set to a negative value if unknown.
521 * - decoding: set by libavcodec, read by user.
526 * YUV colorspace type.
527 * It must be accessed using av_frame_get_colorspace() and
528 * av_frame_set_colorspace().
529 * - encoding: Set by user
530 * - decoding: Set by libavcodec
532 enum AVColorSpace colorspace;
535 * MPEG vs JPEG YUV range.
536 * It must be accessed using av_frame_get_color_range() and
537 * av_frame_set_color_range().
538 * - encoding: Set by user
539 * - decoding: Set by libavcodec
541 enum AVColorRange color_range;
545 * Not to be accessed directly from outside libavutil
547 AVBufferRef *qp_table_buf;
551 * Accessors for some AVFrame fields.
552 * The position of these field in the structure is not part of the ABI,
553 * they should not be accessed directly outside libavcodec.
555 int64_t av_frame_get_best_effort_timestamp(const AVFrame *frame);
556 void av_frame_set_best_effort_timestamp(AVFrame *frame, int64_t val);
557 int64_t av_frame_get_pkt_duration (const AVFrame *frame);
558 void av_frame_set_pkt_duration (AVFrame *frame, int64_t val);
559 int64_t av_frame_get_pkt_pos (const AVFrame *frame);
560 void av_frame_set_pkt_pos (AVFrame *frame, int64_t val);
561 int64_t av_frame_get_channel_layout (const AVFrame *frame);
562 void av_frame_set_channel_layout (AVFrame *frame, int64_t val);
563 int av_frame_get_channels (const AVFrame *frame);
564 void av_frame_set_channels (AVFrame *frame, int val);
565 int av_frame_get_sample_rate (const AVFrame *frame);
566 void av_frame_set_sample_rate (AVFrame *frame, int val);
567 AVDictionary *av_frame_get_metadata (const AVFrame *frame);
568 void av_frame_set_metadata (AVFrame *frame, AVDictionary *val);
569 int av_frame_get_decode_error_flags (const AVFrame *frame);
570 void av_frame_set_decode_error_flags (AVFrame *frame, int val);
571 int av_frame_get_pkt_size(const AVFrame *frame);
572 void av_frame_set_pkt_size(AVFrame *frame, int val);
573 AVDictionary **avpriv_frame_get_metadatap(AVFrame *frame);
574 int8_t *av_frame_get_qp_table(AVFrame *f, int *stride, int *type);
575 int av_frame_set_qp_table(AVFrame *f, AVBufferRef *buf, int stride, int type);
576 enum AVColorSpace av_frame_get_colorspace(const AVFrame *frame);
577 void av_frame_set_colorspace(AVFrame *frame, enum AVColorSpace val);
578 enum AVColorRange av_frame_get_color_range(const AVFrame *frame);
579 void av_frame_set_color_range(AVFrame *frame, enum AVColorRange val);
582 * Get the name of a colorspace.
583 * @return a static string identifying the colorspace; can be NULL.
585 const char *av_get_colorspace_name(enum AVColorSpace val);
588 * Allocate an AVFrame and set its fields to default values. The resulting
589 * struct must be freed using av_frame_free().
591 * @return An AVFrame filled with default values or NULL on failure.
593 * @note this only allocates the AVFrame itself, not the data buffers. Those
594 * must be allocated through other means, e.g. with av_frame_get_buffer() or
597 AVFrame *av_frame_alloc(void);
600 * Free the frame and any dynamically allocated objects in it,
601 * e.g. extended_data. If the frame is reference counted, it will be
602 * unreferenced first.
604 * @param frame frame to be freed. The pointer will be set to NULL.
606 void av_frame_free(AVFrame **frame);
609 * Set up a new reference to the data described by the source frame.
611 * Copy frame properties from src to dst and create a new reference for each
612 * AVBufferRef from src.
614 * If src is not reference counted, new buffers are allocated and the data is
617 * @return 0 on success, a negative AVERROR on error
619 int av_frame_ref(AVFrame *dst, const AVFrame *src);
622 * Create a new frame that references the same data as src.
624 * This is a shortcut for av_frame_alloc()+av_frame_ref().
626 * @return newly created AVFrame on success, NULL on error.
628 AVFrame *av_frame_clone(const AVFrame *src);
631 * Unreference all the buffers referenced by frame and reset the frame fields.
633 void av_frame_unref(AVFrame *frame);
636 * Move everythnig contained in src to dst and reset src.
638 void av_frame_move_ref(AVFrame *dst, AVFrame *src);
641 * Allocate new buffer(s) for audio or video data.
643 * The following fields must be set on frame before calling this function:
644 * - format (pixel format for video, sample format for audio)
645 * - width and height for video
646 * - nb_samples and channel_layout for audio
648 * This function will fill AVFrame.data and AVFrame.buf arrays and, if
649 * necessary, allocate and fill AVFrame.extended_data and AVFrame.extended_buf.
650 * For planar formats, one buffer will be allocated for each plane.
652 * @param frame frame in which to store the new buffers.
653 * @param align required buffer size alignment
655 * @return 0 on success, a negative AVERROR on error.
657 int av_frame_get_buffer(AVFrame *frame, int align);
660 * Check if the frame data is writable.
662 * @return A positive value if the frame data is writable (which is true if and
663 * only if each of the underlying buffers has only one reference, namely the one
664 * stored in this frame). Return 0 otherwise.
666 * If 1 is returned the answer is valid until av_buffer_ref() is called on any
667 * of the underlying AVBufferRefs (e.g. through av_frame_ref() or directly).
669 * @see av_frame_make_writable(), av_buffer_is_writable()
671 int av_frame_is_writable(AVFrame *frame);
674 * Ensure that the frame data is writable, avoiding data copy if possible.
676 * Do nothing if the frame is writable, allocate new buffers and copy the data
679 * @return 0 on success, a negative AVERROR on error.
681 * @see av_frame_is_writable(), av_buffer_is_writable(),
682 * av_buffer_make_writable()
684 int av_frame_make_writable(AVFrame *frame);
687 * Copy the frame data from src to dst.
689 * This function does not allocate anything, dst must be already initialized and
690 * allocated with the same parameters as src.
692 * This function only copies the frame data (i.e. the contents of the data /
693 * extended data arrays), not any other properties.
695 * @return >= 0 on success, a negative AVERROR on error.
697 int av_frame_copy(AVFrame *dst, const AVFrame *src);
700 * Copy only "metadata" fields from src to dst.
702 * Metadata for the purpose of this function are those fields that do not affect
703 * the data layout in the buffers. E.g. pts, sample rate (for audio) or sample
704 * aspect ratio (for video), but not width/height or channel layout.
705 * Side data is also copied.
707 int av_frame_copy_props(AVFrame *dst, const AVFrame *src);
710 * Get the buffer reference a given data plane is stored in.
712 * @param plane index of the data plane of interest in frame->extended_data.
714 * @return the buffer reference that contains the plane or NULL if the input
715 * frame is not valid.
717 AVBufferRef *av_frame_get_plane_buffer(AVFrame *frame, int plane);
720 * Add a new side data to a frame.
722 * @param frame a frame to which the side data should be added
723 * @param type type of the added side data
724 * @param size size of the side data
726 * @return newly added side data on success, NULL on error
728 AVFrameSideData *av_frame_new_side_data(AVFrame *frame,
729 enum AVFrameSideDataType type,
733 * @return a pointer to the side data of a given type on success, NULL if there
734 * is no side data with such type in this frame.
736 AVFrameSideData *av_frame_get_side_data(const AVFrame *frame,
737 enum AVFrameSideDataType type);
740 * If side data of the supplied type exists in the frame, free it and remove it
743 void av_frame_remove_side_data(AVFrame *frame, enum AVFrameSideDataType type);
749 #endif /* AVUTIL_FRAME_H */