a71c12caca9f2f8f799c8b2be0ce3dd55aa55004
[libav.git] / libavcodec / avcodec.h
1 #ifndef AVCODEC_H
2 #define AVCODEC_H
3
4 /**
5 * @file avcodec.h
6 * external api header.
7 */
8
9
10 #ifdef __cplusplus
11 extern "C" {
12 #endif
13
14 #include "common.h"
15
16 #define LIBAVCODEC_VERSION_INT 0x000406
17 #define LIBAVCODEC_VERSION "0.4.6"
18 #define LIBAVCODEC_BUILD 4663
19 #define LIBAVCODEC_BUILD_STR "4663"
20
21 #define LIBAVCODEC_IDENT "FFmpeg" LIBAVCODEC_VERSION "b" LIBAVCODEC_BUILD_STR
22
23 enum CodecID {
24 CODEC_ID_NONE,
25 CODEC_ID_MPEG1VIDEO,
26 CODEC_ID_H263,
27 CODEC_ID_RV10,
28 CODEC_ID_MP2,
29 CODEC_ID_MP3LAME,
30 CODEC_ID_VORBIS,
31 CODEC_ID_AC3,
32 CODEC_ID_MJPEG,
33 CODEC_ID_MJPEGB,
34 CODEC_ID_MPEG4,
35 CODEC_ID_RAWVIDEO,
36 CODEC_ID_MSMPEG4V1,
37 CODEC_ID_MSMPEG4V2,
38 CODEC_ID_MSMPEG4V3,
39 CODEC_ID_WMV1,
40 CODEC_ID_WMV2,
41 CODEC_ID_H263P,
42 CODEC_ID_H263I,
43 CODEC_ID_SVQ1,
44 CODEC_ID_DVVIDEO,
45 CODEC_ID_DVAUDIO,
46 CODEC_ID_WMAV1,
47 CODEC_ID_WMAV2,
48 CODEC_ID_MACE3,
49 CODEC_ID_MACE6,
50 CODEC_ID_HUFFYUV,
51 CODEC_ID_CYUV,
52 CODEC_ID_H264,
53
54 /* various pcm "codecs" */
55 CODEC_ID_PCM_S16LE,
56 CODEC_ID_PCM_S16BE,
57 CODEC_ID_PCM_U16LE,
58 CODEC_ID_PCM_U16BE,
59 CODEC_ID_PCM_S8,
60 CODEC_ID_PCM_U8,
61 CODEC_ID_PCM_MULAW,
62 CODEC_ID_PCM_ALAW,
63
64 /* various adpcm codecs */
65 CODEC_ID_ADPCM_IMA_QT,
66 CODEC_ID_ADPCM_IMA_WAV,
67 CODEC_ID_ADPCM_MS,
68 };
69
70 enum CodecType {
71 CODEC_TYPE_UNKNOWN = -1,
72 CODEC_TYPE_VIDEO,
73 CODEC_TYPE_AUDIO,
74 };
75
76 /**
77 * Pixel format.
78 */
79 enum PixelFormat {
80 PIX_FMT_YUV420P,
81 PIX_FMT_YUV422,
82 PIX_FMT_RGB24, ///< 3 bytes, R is first
83 PIX_FMT_BGR24, ///< 3 bytes, B is first
84 PIX_FMT_YUV422P,
85 PIX_FMT_YUV444P,
86 PIX_FMT_RGBA32, ///< always stored in cpu endianness
87 PIX_FMT_YUV410P,
88 PIX_FMT_YUV411P,
89 PIX_FMT_RGB565, ///< always stored in cpu endianness
90 PIX_FMT_RGB555, ///< always stored in cpu endianness, most significant bit to 1
91 PIX_FMT_GRAY8,
92 PIX_FMT_MONOWHITE, ///< 0 is white
93 PIX_FMT_MONOBLACK, ///< 0 is black
94 PIX_FMT_PAL8, ///< 8 bit with RGBA palette
95 PIX_FMT_NB,
96 };
97
98 /* currently unused, may be used if 24/32 bits samples ever supported */
99 enum SampleFormat {
100 SAMPLE_FMT_S16 = 0, ///< signed 16 bits
101 };
102
103 /* in bytes */
104 #define AVCODEC_MAX_AUDIO_FRAME_SIZE 131072
105
106 /**
107 * Required number of zero bytes at the end of the input bitstream for decoding.
108 * to avoid overreading (and possibly segfaulting)
109 */
110 #define FF_INPUT_BUFFER_PADDING_SIZE 8
111
112 /* motion estimation type, EPZS by default */
113 enum Motion_Est_ID {
114 ME_ZERO = 1,
115 ME_FULL,
116 ME_LOG,
117 ME_PHODS,
118 ME_EPZS,
119 ME_X1
120 };
121
122 typedef struct RcOverride{
123 int start_frame;
124 int end_frame;
125 int qscale; // if this is 0 then quality_factor will be used instead
126 float quality_factor;
127 } RcOverride;
128
129 /* only for ME compatiblity with old apps */
130 extern int motion_estimation_method;
131
132 /* ME algos sorted by quality */
133 static const int Motion_Est_QTab[] = { ME_ZERO, ME_PHODS, ME_LOG,
134 ME_X1, ME_EPZS, ME_FULL };
135
136
137 #define FF_MAX_B_FRAMES 8
138
139 /* encoding support
140 these flags can be passed in AVCodecContext.flags before initing
141 Note: note not everything is supported yet
142 */
143
144 #define CODEC_FLAG_HQ 0x0001 ///< brute force MB-type decission mode (slow)
145 #define CODEC_FLAG_QSCALE 0x0002 ///< use fixed qscale
146 #define CODEC_FLAG_4MV 0x0004 ///< 4 MV per MB allowed
147 #define CODEC_FLAG_QPEL 0x0010 ///< use qpel MC
148 #define CODEC_FLAG_GMC 0x0020 ///< use GMC
149 #define CODEC_FLAG_PART 0x0080 ///< use data partitioning
150 /* parent program gurantees that the input for b-frame containing streams is not written to
151 for at least s->max_b_frames+1 frames, if this is not set than the input will be copied */
152 #define CODEC_FLAG_INPUT_PRESERVED 0x0100
153 #define CODEC_FLAG_PASS1 0x0200 ///< use internal 2pass ratecontrol in first pass mode
154 #define CODEC_FLAG_PASS2 0x0400 ///< use internal 2pass ratecontrol in second pass mode
155 #define CODEC_FLAG_EXTERN_HUFF 0x1000 ///< use external huffman table (for mjpeg)
156 #define CODEC_FLAG_GRAY 0x2000 ///< only decode/encode grayscale
157 #define CODEC_FLAG_EMU_EDGE 0x4000///< dont draw edges
158 #define CODEC_FLAG_PSNR 0x8000 ///< error[?] variables will be set during encoding
159 #define CODEC_FLAG_TRUNCATED 0x00010000 /** input bitstream might be truncated at a random location instead
160 of only at frame boundaries */
161 #define CODEC_FLAG_NORMALIZE_AQP 0x00020000 ///< normalize adaptive quantization
162 #define CODEC_FLAG_INTERLACED_DCT 0x00040000 ///< use interlaced dct
163 #define CODEC_FLAG_LOW_DELAY 0x00080000 ///< force low delay / will fail on b frames
164 #define CODEC_FLAG_ALT_SCAN 0x00100000 ///< use alternate scan
165 #define CODEC_FLAG_TRELLIS_QUANT 0x00200000 ///< use trellis quantization
166 #define CODEC_FLAG_GLOBAL_HEADER 0x00400000 ///< place global headers in extradata instead of every keyframe
167 #define CODEC_FLAG_BITEXACT 0x00800000 ///< use only bitexact stuff (except (i)dct)
168 /* Fx : Flag for h263+ extra options */
169 #define CODEC_FLAG_H263P_AIC 0x01000000 ///< Advanced intra coding
170 #define CODEC_FLAG_H263P_UMV 0x02000000 ///< Unlimited motion vector
171 /* For advanced prediction mode, we reuse the 4MV flag */
172 /* Unsupported options :
173 * Syntax Arithmetic coding (SAC)
174 * Deblocking filter internal loop
175 * Slice structured
176 * Reference Picture Selection
177 * Independant Segment Decoding
178 * Alternative Inter * VLC
179 * Modified Quantization */
180 /* /Fx */
181 /* codec capabilities */
182
183 #define CODEC_CAP_DRAW_HORIZ_BAND 0x0001 ///< decoder can use draw_horiz_band callback
184 #define CODEC_CAP_DR1 0x0002 ///< direct rendering method 1
185 /* if 'parse_only' field is true, then avcodec_parse_frame() can be
186 used */
187 #define CODEC_CAP_PARSE_ONLY 0x0004
188 #define CODEC_CAP_TRUNCATED 0x0008
189
190 #define FF_COMMON_FRAME \
191 /**\
192 * pointer to the picture planes.\
193 * this might be different from the first allocated byte\
194 * - encoding: \
195 * - decoding: \
196 */\
197 uint8_t *data[4];\
198 int linesize[4];\
199 /**\
200 * pointer to the first allocated byte of the picture. can be used in get_buffer/release_buffer\
201 * this isnt used by lavc unless the default get/release_buffer() is used\
202 * - encoding: \
203 * - decoding: \
204 */\
205 uint8_t *base[4];\
206 /**\
207 * 1 -> keyframe, 0-> not\
208 * - encoding: set by lavc\
209 * - decoding: set by lavc\
210 */\
211 int key_frame;\
212 \
213 /**\
214 * picture type of the frame, see ?_TYPE below.\
215 * - encoding: set by lavc for coded_picture (and set by user for input)\
216 * - decoding: set by lavc\
217 */\
218 int pict_type;\
219 \
220 /**\
221 * presentation timestamp in micro seconds (time when frame should be shown to user)\
222 * if 0 then the frame_rate will be used as reference\
223 * - encoding: MUST be set by user\
224 * - decoding: set by lavc\
225 */\
226 int64_t pts;\
227 \
228 /**\
229 * picture number in bitstream order.\
230 * - encoding: set by\
231 * - decoding: set by lavc\
232 */\
233 int coded_picture_number;\
234 /**\
235 * picture number in display order.\
236 * - encoding: set by\
237 * - decoding: set by lavc\
238 */\
239 int display_picture_number;\
240 \
241 /**\
242 * quality (between 1 (good) and 31 (bad)) \
243 * - encoding: set by lavc for coded_picture (and set by user for input)\
244 * - decoding: set by lavc\
245 */\
246 float quality; \
247 \
248 /**\
249 * buffer age (1->was last buffer and dint change, 2->..., ...).\
250 * set to something large if the buffer has not been used yet \
251 * - encoding: unused\
252 * - decoding: MUST be set by get_buffer()\
253 */\
254 int age;\
255 \
256 /**\
257 * is this picture used as reference\
258 * - encoding: unused\
259 * - decoding: set by lavc (before get_buffer() call))\
260 */\
261 int reference;\
262 \
263 /**\
264 * QP table\
265 * - encoding: unused\
266 * - decoding: set by lavc\
267 */\
268 int8_t *qscale_table;\
269 /**\
270 * QP store stride\
271 * - encoding: unused\
272 * - decoding: set by lavc\
273 */\
274 int qstride;\
275 \
276 /**\
277 * mbskip_table[mb]>=1 if MB didnt change\
278 * stride= mb_width = (width+15)>>4\
279 * - encoding: unused\
280 * - decoding: set by lavc\
281 */\
282 uint8_t *mbskip_table;\
283 \
284 /**\
285 * for some private data of the user\
286 * - encoding: unused\
287 * - decoding: set by user\
288 */\
289 void *opaque;\
290 \
291 /**\
292 * error\
293 * - encoding: set by lavc if flags&CODEC_FLAG_PSNR\
294 * - decoding: unused\
295 */\
296 uint64_t error[4];\
297 \
298 /**\
299 * type of the buffer (to keep track of who has to dealloc data[*])\
300 * - encoding: set by the one who allocs it\
301 * - decoding: set by the one who allocs it\
302 * Note: user allocated (direct rendering) & internal buffers can not coexist currently\
303 */\
304 int type;\
305 \
306 /**\
307 * when decoding, this signal how much the picture must be delayed.\
308 * extra_delay = repeat_pict / (2*fps)\
309 * - encoding: unused\
310 * - decoding: set by lavc\
311 */\
312 int repeat_pict;
313
314
315 #define FF_BUFFER_TYPE_INTERNAL 1
316 #define FF_BUFFER_TYPE_USER 2 ///< Direct rendering buffers
317 #define FF_BUFFER_TYPE_SHARED 4 ///< buffer from somewher else, dont dealloc
318
319
320 #define FF_I_TYPE 1 // Intra
321 #define FF_P_TYPE 2 // Predicted
322 #define FF_B_TYPE 3 // Bi-dir predicted
323 #define FF_S_TYPE 4 // S(GMC)-VOP MPEG4
324 #define FF_SI_TYPE 5
325 #define FF_SP_TYPE 6
326
327 /**
328 * Audio Video Frame.
329 */
330 typedef struct AVFrame {
331 FF_COMMON_FRAME
332 } AVFrame;
333
334 #define DEFAULT_FRAME_RATE_BASE 1001000
335
336 /**
337 * main external api structure.
338 */
339 typedef struct AVCodecContext {
340 /**
341 * the average bitrate.
342 * - encoding: set by user. unused for constant quantizer encoding
343 * - decoding: set by lavc. 0 or some bitrate if this info is available in the stream
344 */
345 int bit_rate;
346
347 /**
348 * number of bits the bitstream is allowed to diverge from the reference.
349 * the reference can be CBR (for CBR pass1) or VBR (for pass2)
350 * - encoding: set by user. unused for constant quantizer encoding
351 * - decoding: unused
352 */
353 int bit_rate_tolerance;
354
355 /**
356 * CODEC_FLAG_*.
357 * - encoding: set by user.
358 * - decoding: set by user.
359 */
360 int flags;
361
362 /**
363 * some codecs needs additionnal format info. It is stored here
364 * - encoding: set by user.
365 * - decoding: set by lavc. (FIXME is this ok?)
366 */
367 int sub_id;
368
369 /**
370 * motion estimation algorithm used for video coding.
371 * - encoding: MUST be set by user.
372 * - decoding: unused
373 */
374 int me_method;
375
376 /**
377 * some codecs need / can use extra-data like huffman tables.
378 * mjpeg: huffman tables
379 * rv10: additional flags
380 * mpeg4: global headers (they can be in the bitstream or here)
381 * - encoding: set/allocated/freed by lavc.
382 * - decoding: set/allocated/freed by user.
383 */
384 void *extradata;
385 int extradata_size;
386
387 /* video only */
388 /**
389 * frames per sec multiplied by frame_rate_base.
390 * for variable fps this is the precission, so if the timestamps
391 * can be specified in msec precssion then this is 1000*frame_rate_base
392 * - encoding: MUST be set by user
393 * - decoding: set by lavc. 0 or the frame_rate if available
394 */
395 int frame_rate;
396
397 /**
398 * width / height.
399 * - encoding: MUST be set by user.
400 * - decoding: set by user, some codecs might override / change it during playback
401 */
402 int width, height;
403
404 #define FF_ASPECT_SQUARE 1
405 #define FF_ASPECT_4_3_625 2
406 #define FF_ASPECT_4_3_525 3
407 #define FF_ASPECT_16_9_625 4
408 #define FF_ASPECT_16_9_525 5
409 #define FF_ASPECT_EXTENDED 15
410
411 /**
412 * the number of pictures in a group of pitures, or 0 for intra_only.
413 * - encoding: set by user.
414 * - decoding: unused
415 */
416 int gop_size;
417
418 /**
419 * pixel format, see PIX_FMT_xxx.
420 * - encoding: unused
421 * - decoding: set by lavc.
422 */
423 enum PixelFormat pix_fmt;
424
425 /**
426 * Frame rate emulation. If not zero lower layer (i.e. format handler)
427 * has to read frames at native frame rate.
428 * - encoding: set by user.
429 * - decoding: unused.
430 */
431 int rate_emu;
432
433 /**
434 * if non NULL, 'draw_horiz_band' is called by the libavcodec
435 * decoder to draw an horizontal band. It improve cache usage. Not
436 * all codecs can do that. You must check the codec capabilities
437 * before
438 * - encoding: unused
439 * - decoding: set by user.
440 */
441 void (*draw_horiz_band)(struct AVCodecContext *s,
442 uint8_t **src_ptr, int linesize,
443 int y, int width, int height);
444
445 /* audio only */
446 int sample_rate; ///< samples per sec
447 int channels;
448 int sample_fmt; ///< sample format, currenly unused
449
450 /* the following data should not be initialized */
451 int frame_size; ///< in samples, initialized when calling 'init'
452 int frame_number; ///< audio or video frame number
453 int real_pict_num; ///< returns the real picture number of previous encoded frame
454
455 /**
456 * number of frames the decoded output will be delayed relative to
457 * the encoded input.
458 * - encoding: set by lavc.
459 * - decoding: unused
460 */
461 int delay;
462
463 /* - encoding parameters */
464 float qcompress; ///< amount of qscale change between easy & hard scenes (0.0-1.0)
465 float qblur; ///< amount of qscale smoothing over time (0.0-1.0)
466
467 /**
468 * minimum quantizer.
469 * - encoding: set by user.
470 * - decoding: unused
471 */
472 int qmin;
473
474 /**
475 * maximum quantizer.
476 * - encoding: set by user.
477 * - decoding: unused
478 */
479 int qmax;
480
481 /**
482 * maximum quantizer difference etween frames.
483 * - encoding: set by user.
484 * - decoding: unused
485 */
486 int max_qdiff;
487
488 /**
489 * maximum number of b frames between non b frames.
490 * note: the output will be delayed by max_b_frames+1 relative to the input
491 * - encoding: set by user.
492 * - decoding: unused
493 */
494 int max_b_frames;
495
496 /**
497 * qscale factor between ip and b frames.
498 * - encoding: set by user.
499 * - decoding: unused
500 */
501 float b_quant_factor;
502
503 /** obsolete FIXME remove */
504 int rc_strategy;
505 int b_frame_strategy;
506
507 /**
508 * hurry up amount.
509 * - encoding: unused
510 * - decoding: set by user. 1-> skip b frames, 2-> skip idct/dequant too, 5-> skip everything except header
511 */
512 int hurry_up;
513
514 struct AVCodec *codec;
515
516 void *priv_data;
517
518 /* The following data is for RTP friendly coding */
519 /* By now only H.263/H.263+/MPEG4 coder honours this */
520 int rtp_mode; /* 1 for activate RTP friendly-mode */
521 /* highers numbers represent more error-prone */
522 /* enviroments, by now just "1" exist */
523
524 int rtp_payload_size; /* The size of the RTP payload, the coder will */
525 /* do it's best to deliver a chunk with size */
526 /* below rtp_payload_size, the chunk will start */
527 /* with a start code on some codecs like H.263 */
528 /* This doesn't take account of any particular */
529 /* headers inside the transmited RTP payload */
530
531
532 /* The RTP callcack: This function is called */
533 /* every time the encoder as a packet to send */
534 /* Depends on the encoder if the data starts */
535 /* with a Start Code (it should) H.263 does */
536 void (*rtp_callback)(void *data, int size, int packet_number);
537
538 /* statistics, used for 2-pass encoding */
539 int mv_bits;
540 int header_bits;
541 int i_tex_bits;
542 int p_tex_bits;
543 int i_count;
544 int p_count;
545 int skip_count;
546 int misc_bits;
547
548 /**
549 * number of bits used for the previously encoded frame.
550 * - encoding: set by lavc
551 * - decoding: unused
552 */
553 int frame_bits;
554
555 /**
556 * private data of the user, can be used to carry app specific stuff.
557 * - encoding: set by user
558 * - decoding: set by user
559 */
560 void *opaque;
561
562 char codec_name[32];
563 enum CodecType codec_type; /* see CODEC_TYPE_xxx */
564 enum CodecID codec_id; /* see CODEC_ID_xxx */
565
566 /**
567 * fourcc (LSB first, so "ABCD" -> ('D'<<24) + ('C'<<16) + ('B'<<8) + 'A').
568 * this is used to workaround some encoder bugs
569 * - encoding: unused
570 * - decoding: set by user, will be converted to upper case by lavc during init
571 */
572 unsigned int codec_tag;
573
574 /**
575 * workaround bugs in encoders which sometimes cannot be detected automatically.
576 * - encoding: unused
577 * - decoding: set by user
578 */
579 int workaround_bugs;
580 #define FF_BUG_AUTODETECT 1 ///< autodetection
581 #define FF_BUG_OLD_MSMPEG4 2
582 #define FF_BUG_XVID_ILACE 4
583 #define FF_BUG_UMP4 8
584 #define FF_BUG_NO_PADDING 16
585 #define FF_BUG_AC_VLC 32
586 #define FF_BUG_QPEL_CHROMA 64
587 #define FF_BUG_STD_QPEL 128
588 #define FF_BUG_QPEL_CHROMA2 256
589 #define FF_BUG_DIRECT_BLOCKSIZE 512
590 #define FF_BUG_EDGE 1024
591 //#define FF_BUG_FAKE_SCALABILITY 16 //autodetection should work 100%
592
593 /**
594 * luma single coeff elimination threshold.
595 * - encoding: set by user
596 * - decoding: unused
597 */
598 int luma_elim_threshold;
599
600 /**
601 * chroma single coeff elimination threshold.
602 * - encoding: set by user
603 * - decoding: unused
604 */
605 int chroma_elim_threshold;
606
607 /**
608 * strictly follow the std (MPEG4, ...).
609 * - encoding: set by user
610 * - decoding: unused
611 */
612 int strict_std_compliance;
613
614 /**
615 * qscale offset between ip and b frames.
616 * if > 0 then the last p frame quantizer will be used (q= lastp_q*factor+offset)
617 * if < 0 then normal ratecontrol will be done (q= -normal_q*factor+offset)
618 * - encoding: set by user.
619 * - decoding: unused
620 */
621 float b_quant_offset;
622
623 /**
624 * error resilience higher values will detect more errors but may missdetect
625 * some more or less valid parts as errors.
626 * - encoding: unused
627 * - decoding: set by user
628 */
629 int error_resilience;
630 #define FF_ER_CAREFULL 1
631 #define FF_ER_COMPLIANT 2
632 #define FF_ER_AGGRESSIVE 3
633 #define FF_ER_VERY_AGGRESSIVE 4
634
635 /**
636 * called at the beginning of each frame to get a buffer for it.
637 * if pic.reference is set then the frame will be read later by lavc
638 * - encoding: unused
639 * - decoding: set by lavc, user can override
640 */
641 int (*get_buffer)(struct AVCodecContext *c, AVFrame *pic);
642
643 /**
644 * called to release buffers which where allocated with get_buffer.
645 * a released buffer can be reused in get_buffer()
646 * pic.data[*] must be set to NULL
647 * - encoding: unused
648 * - decoding: set by lavc, user can override
649 */
650 void (*release_buffer)(struct AVCodecContext *c, AVFrame *pic);
651
652 /**
653 * is 1 if the decoded stream contains b frames, 0 otherwise.
654 * - encoding: unused
655 * - decoding: set by lavc
656 */
657 int has_b_frames;
658
659 int block_align; ///< used by some WAV based audio codecs
660
661 int parse_only; /* - decoding only: if true, only parsing is done
662 (function avcodec_parse_frame()). The frame
663 data is returned. Only MPEG codecs support this now. */
664
665 /**
666 * 0-> h263 quant 1-> mpeg quant.
667 * - encoding: set by user.
668 * - decoding: unused
669 */
670 int mpeg_quant;
671
672 /**
673 * pass1 encoding statistics output buffer.
674 * - encoding: set by lavc
675 * - decoding: unused
676 */
677 char *stats_out;
678
679 /**
680 * pass2 encoding statistics input buffer.
681 * concatenated stuff from stats_out of pass1 should be placed here
682 * - encoding: allocated/set/freed by user
683 * - decoding: unused
684 */
685 char *stats_in;
686
687 /**
688 * ratecontrol qmin qmax limiting method.
689 * 0-> clipping, 1-> use a nice continous function to limit qscale wthin qmin/qmax
690 * - encoding: set by user.
691 * - decoding: unused
692 */
693 float rc_qsquish;
694
695 float rc_qmod_amp;
696 int rc_qmod_freq;
697
698 /**
699 * ratecontrol override, see RcOverride.
700 * - encoding: allocated/set/freed by user.
701 * - decoding: unused
702 */
703 RcOverride *rc_override;
704 int rc_override_count;
705
706 /**
707 * rate control equation.
708 * - encoding: set by user
709 * - decoding: unused
710 */
711 char *rc_eq;
712
713 /**
714 * maximum bitrate.
715 * - encoding: set by user.
716 * - decoding: unused
717 */
718 int rc_max_rate;
719
720 /**
721 * minimum bitrate.
722 * - encoding: set by user.
723 * - decoding: unused
724 */
725 int rc_min_rate;
726
727 /**
728 * decoder bitstream buffer size.
729 * - encoding: set by user.
730 * - decoding: unused
731 */
732 int rc_buffer_size;
733 float rc_buffer_aggressivity;
734
735 /**
736 * qscale factor between p and i frames.
737 * - encoding: set by user.
738 * - decoding: unused
739 */
740 float i_quant_factor;
741
742 /**
743 * qscale offset between p and i frames.
744 * if > 0 then the last p frame quantizer will be used (q= lastp_q*factor+offset)
745 * if < 0 then normal ratecontrol will be done (q= -normal_q*factor+offset)
746 * - encoding: set by user.
747 * - decoding: unused
748 */
749 float i_quant_offset;
750
751 /**
752 * initial complexity for pass1 ratecontrol.
753 * - encoding: set by user.
754 * - decoding: unused
755 */
756 float rc_initial_cplx;
757
758 /**
759 * dct algorithm, see FF_DCT_* below.
760 * - encoding: set by user
761 * - decoding: unused
762 */
763 int dct_algo;
764 #define FF_DCT_AUTO 0
765 #define FF_DCT_FASTINT 1
766 #define FF_DCT_INT 2
767 #define FF_DCT_MMX 3
768 #define FF_DCT_MLIB 4
769 #define FF_DCT_ALTIVEC 5
770
771 /**
772 * luminance masking (0-> disabled).
773 * - encoding: set by user
774 * - decoding: unused
775 */
776 float lumi_masking;
777
778 /**
779 * temporary complexity masking (0-> disabled).
780 * - encoding: set by user
781 * - decoding: unused
782 */
783 float temporal_cplx_masking;
784
785 /**
786 * spatial complexity masking (0-> disabled).
787 * - encoding: set by user
788 * - decoding: unused
789 */
790 float spatial_cplx_masking;
791
792 /**
793 * p block masking (0-> disabled).
794 * - encoding: set by user
795 * - decoding: unused
796 */
797 float p_masking;
798
799 /**
800 * darkness masking (0-> disabled).
801 * - encoding: set by user
802 * - decoding: unused
803 */
804 float dark_masking;
805
806
807 /* for binary compatibility */
808 int unused;
809
810 /**
811 * idct algorithm, see FF_IDCT_* below.
812 * - encoding: set by user
813 * - decoding: set by user
814 */
815 int idct_algo;
816 #define FF_IDCT_AUTO 0
817 #define FF_IDCT_INT 1
818 #define FF_IDCT_SIMPLE 2
819 #define FF_IDCT_SIMPLEMMX 3
820 #define FF_IDCT_LIBMPEG2MMX 4
821 #define FF_IDCT_PS2 5
822 #define FF_IDCT_MLIB 6
823 #define FF_IDCT_ARM 7
824 #define FF_IDCT_ALTIVEC 8
825
826 /**
827 * slice count.
828 * - encoding: set by lavc
829 * - decoding: set by user (or 0)
830 */
831 int slice_count;
832 /**
833 * slice offsets in the frame in bytes.
834 * - encoding: set/allocated by lavc
835 * - decoding: set/allocated by user (or NULL)
836 */
837 int *slice_offset;
838
839 /**
840 * error concealment flags.
841 * - encoding: unused
842 * - decoding: set by user
843 */
844 int error_concealment;
845 #define FF_EC_GUESS_MVS 1
846 #define FF_EC_DEBLOCK 2
847
848 /**
849 * dsp_mask could be add used to disable unwanted CPU features
850 * CPU features (i.e. MMX, SSE. ...)
851 *
852 * with FORCE flag you may instead enable given CPU features
853 * (Dangerous: usable in case of misdetection, improper usage however will
854 * result into program crash)
855 */
856 unsigned dsp_mask;
857 #define FF_MM_FORCE 0x80000000 /* force usage of selected flags (OR) */
858 /* lower 16 bits - CPU features */
859 #ifdef HAVE_MMX
860 #define FF_MM_MMX 0x0001 /* standard MMX */
861 #define FF_MM_3DNOW 0x0004 /* AMD 3DNOW */
862 #define FF_MM_MMXEXT 0x0002 /* SSE integer functions or AMD MMX ext */
863 #define FF_MM_SSE 0x0008 /* SSE functions */
864 #define FF_MM_SSE2 0x0010 /* PIV SSE2 functions */
865 #endif /* HAVE_MMX */
866
867 /**
868 * bits per sample/pixel from the demuxer (needed for huffyuv).
869 * - encoding: set by lavc
870 * - decoding: set by user
871 */
872 int bits_per_sample;
873
874 /**
875 * prediction method (needed for huffyuv).
876 * - encoding: set by user
877 * - decoding: unused
878 */
879 int prediction_method;
880 #define FF_PRED_LEFT 0
881 #define FF_PRED_PLANE 1
882 #define FF_PRED_MEDIAN 2
883
884 /**
885 * aspect ratio (0 if unknown).
886 * - encoding: set by user.
887 * - decoding: set by lavc.
888 */
889 float aspect_ratio;
890
891 /**
892 * the picture in the bitstream.
893 * - encoding: set by lavc
894 * - decoding: set by lavc
895 */
896 AVFrame *coded_frame;
897
898 /**
899 * debug.
900 * - encoding: set by user.
901 * - decoding: set by user.
902 */
903 int debug;
904 #define FF_DEBUG_PICT_INFO 1
905 #define FF_DEBUG_RC 2
906 #define FF_DEBUG_BITSTREAM 4
907 #define FF_DEBUG_MB_TYPE 8
908 #define FF_DEBUG_QP 16
909 #define FF_DEBUG_MV 32
910 #define FF_DEBUG_VIS_MV 0x00000040
911 #define FF_DEBUG_SKIP 0x00000080
912 #define FF_DEBUG_STARTCODE 0x00000100
913 #define FF_DEBUG_PTS 0x00000200
914 #define FF_DEBUG_ER 0x00000400
915 #define FF_DEBUG_MMCO 0x00000800
916
917 /**
918 * error.
919 * - encoding: set by lavc if flags&CODEC_FLAG_PSNR
920 * - decoding: unused
921 */
922 uint64_t error[4];
923
924 /**
925 * minimum MB quantizer.
926 * - encoding: set by user.
927 * - decoding: unused
928 */
929 int mb_qmin;
930
931 /**
932 * maximum MB quantizer.
933 * - encoding: set by user.
934 * - decoding: unused
935 */
936 int mb_qmax;
937
938 /**
939 * motion estimation compare function.
940 * - encoding: set by user.
941 * - decoding: unused
942 */
943 int me_cmp;
944 /**
945 * subpixel motion estimation compare function.
946 * - encoding: set by user.
947 * - decoding: unused
948 */
949 int me_sub_cmp;
950 /**
951 * macroblock compare function (not supported yet).
952 * - encoding: set by user.
953 * - decoding: unused
954 */
955 int mb_cmp;
956 #define FF_CMP_SAD 0
957 #define FF_CMP_SSE 1
958 #define FF_CMP_SATD 2
959 #define FF_CMP_DCT 3
960 #define FF_CMP_PSNR 4
961 #define FF_CMP_BIT 5
962 #define FF_CMP_RD 6
963 #define FF_CMP_ZERO 7
964 #define FF_CMP_CHROMA 256
965
966 /**
967 * ME diamond size & shape.
968 * - encoding: set by user.
969 * - decoding: unused
970 */
971 int dia_size;
972
973 /**
974 * amount of previous MV predictors (2a+1 x 2a+1 square).
975 * - encoding: set by user.
976 * - decoding: unused
977 */
978 int last_predictor_count;
979
980 /**
981 * pre pass for motion estimation.
982 * - encoding: set by user.
983 * - decoding: unused
984 */
985 int pre_me;
986
987 /**
988 * motion estimation pre pass compare function.
989 * - encoding: set by user.
990 * - decoding: unused
991 */
992 int me_pre_cmp;
993
994 /**
995 * ME pre pass diamond size & shape.
996 * - encoding: set by user.
997 * - decoding: unused
998 */
999 int pre_dia_size;
1000
1001 /**
1002 * subpel ME quality.
1003 * - encoding: set by user.
1004 * - decoding: unused
1005 */
1006 int me_subpel_quality;
1007
1008 /**
1009 * callback to negotiate the pixelFormat.
1010 * @param fmt is the list of formats which are supported by the codec,
1011 * its terminated by -1 as 0 is a valid format, the formats are ordered by quality
1012 * the first is allways the native one
1013 * @return the choosen format
1014 * - encoding: unused
1015 * - decoding: set by user, if not set then the native format will always be choosen
1016 */
1017 enum PixelFormat (*get_format)(struct AVCodecContext *s, enum PixelFormat * fmt);
1018
1019 /**
1020 * DTG active format information (additionnal aspect ratio
1021 * information only used in DVB MPEG2 transport streams). 0 if
1022 * not set.
1023 *
1024 * - encoding: unused.
1025 * - decoding: set by decoder
1026 */
1027 int dtg_active_format;
1028 #define FF_DTG_AFD_SAME 8
1029 #define FF_DTG_AFD_4_3 9
1030 #define FF_DTG_AFD_16_9 10
1031 #define FF_DTG_AFD_14_9 11
1032 #define FF_DTG_AFD_4_3_SP_14_9 13
1033 #define FF_DTG_AFD_16_9_SP_14_9 14
1034 #define FF_DTG_AFD_SP_4_3 15
1035
1036 /**
1037 * Maximum motion estimation search range in subpel units.
1038 * if 0 then no limit
1039 *
1040 * - encoding: set by user.
1041 * - decoding: unused.
1042 */
1043 int me_range;
1044
1045 /**
1046 * frame_rate_base.
1047 * for variable fps this is 1
1048 * - encoding: set by user.
1049 * - decoding: set by lavc.
1050 * @todo move this after frame_rate
1051 */
1052
1053 int frame_rate_base;
1054 /**
1055 * intra quantizer bias.
1056 * - encoding: set by user.
1057 * - decoding: unused
1058 */
1059 int intra_quant_bias;
1060 #define FF_DEFAULT_QUANT_BIAS 999999
1061
1062 /**
1063 * inter quantizer bias.
1064 * - encoding: set by user.
1065 * - decoding: unused
1066 */
1067 int inter_quant_bias;
1068
1069 } AVCodecContext;
1070
1071
1072 /**
1073 * AVOption.
1074 */
1075 typedef struct AVOption {
1076 /** options' name */
1077 const char *name; /* if name is NULL, it indicates a link to next */
1078 /** short English text help or const struct AVOption* subpointer */
1079 const char *help; // const struct AVOption* sub;
1080 /** offset to context structure where the parsed value should be stored */
1081 int offset;
1082 /** options' type */
1083 int type;
1084 #define FF_OPT_TYPE_BOOL 1 ///< boolean - true,1,on (or simply presence)
1085 #define FF_OPT_TYPE_DOUBLE 2 ///< double
1086 #define FF_OPT_TYPE_INT 3 ///< integer
1087 #define FF_OPT_TYPE_STRING 4 ///< string (finished with \0)
1088 #define FF_OPT_TYPE_MASK 0x1f ///< mask for types - upper bits are various flags
1089 //#define FF_OPT_TYPE_EXPERT 0x20 // flag for expert option
1090 #define FF_OPT_TYPE_FLAG (FF_OPT_TYPE_BOOL | 0x40)
1091 #define FF_OPT_TYPE_RCOVERRIDE (FF_OPT_TYPE_STRING | 0x80)
1092 /** min value (min == max -> no limits) */
1093 double min;
1094 /** maximum value for double/int */
1095 double max;
1096 /** default boo [0,1]l/double/int value */
1097 double defval;
1098 /**
1099 * default string value (with optional semicolon delimited extra option-list
1100 * i.e. option1;option2;option3
1101 * defval might select other then first argument as default
1102 */
1103 const char *defstr;
1104 #define FF_OPT_MAX_DEPTH 10
1105 } AVOption;
1106
1107 /**
1108 * Parse option(s) and sets fields in passed structure
1109 * @param strct structure where the parsed results will be written
1110 * @param list list with AVOptions
1111 * @param opts string with options for parsing
1112 */
1113 int avoption_parse(void* strct, const AVOption* list, const char* opts);
1114
1115
1116 /**
1117 * AVCodec.
1118 */
1119 typedef struct AVCodec {
1120 const char *name;
1121 int type;
1122 int id;
1123 int priv_data_size;
1124 int (*init)(AVCodecContext *);
1125 int (*encode)(AVCodecContext *, uint8_t *buf, int buf_size, void *data);
1126 int (*close)(AVCodecContext *);
1127 int (*decode)(AVCodecContext *, void *outdata, int *outdata_size,
1128 uint8_t *buf, int buf_size);
1129 int capabilities;
1130 const AVOption *options;
1131 struct AVCodec *next;
1132 } AVCodec;
1133
1134 /**
1135 * four components are given, that's all.
1136 * the last component is alpha
1137 */
1138 typedef struct AVPicture {
1139 uint8_t *data[4];
1140 int linesize[4];
1141 } AVPicture;
1142
1143 extern AVCodec ac3_encoder;
1144 extern AVCodec mp2_encoder;
1145 extern AVCodec mp3lame_encoder;
1146 extern AVCodec oggvorbis_encoder;
1147 extern AVCodec mpeg1video_encoder;
1148 extern AVCodec h263_encoder;
1149 extern AVCodec h263p_encoder;
1150 extern AVCodec rv10_encoder;
1151 extern AVCodec mjpeg_encoder;
1152 extern AVCodec mpeg4_encoder;
1153 extern AVCodec msmpeg4v1_encoder;
1154 extern AVCodec msmpeg4v2_encoder;
1155 extern AVCodec msmpeg4v3_encoder;
1156 extern AVCodec wmv1_encoder;
1157 extern AVCodec wmv2_encoder;
1158 extern AVCodec huffyuv_encoder;
1159 extern AVCodec h264_encoder;
1160
1161 extern AVCodec h263_decoder;
1162 extern AVCodec mpeg4_decoder;
1163 extern AVCodec msmpeg4v1_decoder;
1164 extern AVCodec msmpeg4v2_decoder;
1165 extern AVCodec msmpeg4v3_decoder;
1166 extern AVCodec wmv1_decoder;
1167 extern AVCodec wmv2_decoder;
1168 extern AVCodec mpeg_decoder;
1169 extern AVCodec h263i_decoder;
1170 extern AVCodec rv10_decoder;
1171 extern AVCodec svq1_decoder;
1172 extern AVCodec dvvideo_decoder;
1173 extern AVCodec dvaudio_decoder;
1174 extern AVCodec wmav1_decoder;
1175 extern AVCodec wmav2_decoder;
1176 extern AVCodec mjpeg_decoder;
1177 extern AVCodec mjpegb_decoder;
1178 extern AVCodec mp2_decoder;
1179 extern AVCodec mp3_decoder;
1180 extern AVCodec mace3_decoder;
1181 extern AVCodec mace6_decoder;
1182 extern AVCodec huffyuv_decoder;
1183 extern AVCodec oggvorbis_decoder;
1184 extern AVCodec cyuv_decoder;
1185 extern AVCodec h264_decoder;
1186
1187 /* pcm codecs */
1188 #define PCM_CODEC(id, name) \
1189 extern AVCodec name ## _decoder; \
1190 extern AVCodec name ## _encoder
1191
1192 PCM_CODEC(CODEC_ID_PCM_S16LE, pcm_s16le);
1193 PCM_CODEC(CODEC_ID_PCM_S16BE, pcm_s16be);
1194 PCM_CODEC(CODEC_ID_PCM_U16LE, pcm_u16le);
1195 PCM_CODEC(CODEC_ID_PCM_U16BE, pcm_u16be);
1196 PCM_CODEC(CODEC_ID_PCM_S8, pcm_s8);
1197 PCM_CODEC(CODEC_ID_PCM_U8, pcm_u8);
1198 PCM_CODEC(CODEC_ID_PCM_ALAW, pcm_alaw);
1199 PCM_CODEC(CODEC_ID_PCM_MULAW, pcm_mulaw);
1200
1201 /* adpcm codecs */
1202
1203 PCM_CODEC(CODEC_ID_ADPCM_IMA_QT, adpcm_ima_qt);
1204 PCM_CODEC(CODEC_ID_ADPCM_IMA_WAV, adpcm_ima_wav);
1205 PCM_CODEC(CODEC_ID_ADPCM_MS, adpcm_ms);
1206
1207 #undef PCM_CODEC
1208
1209 /* dummy raw video codec */
1210 extern AVCodec rawvideo_codec;
1211
1212 /* the following codecs use external GPL libs */
1213 extern AVCodec ac3_decoder;
1214
1215 /* resample.c */
1216
1217 struct ReSampleContext;
1218
1219 typedef struct ReSampleContext ReSampleContext;
1220
1221 ReSampleContext *audio_resample_init(int output_channels, int input_channels,
1222 int output_rate, int input_rate);
1223 int audio_resample(ReSampleContext *s, short *output, short *input, int nb_samples);
1224 void audio_resample_close(ReSampleContext *s);
1225
1226 /* YUV420 format is assumed ! */
1227
1228 struct ImgReSampleContext;
1229
1230 typedef struct ImgReSampleContext ImgReSampleContext;
1231
1232 ImgReSampleContext *img_resample_init(int output_width, int output_height,
1233 int input_width, int input_height);
1234
1235 ImgReSampleContext *img_resample_full_init(int owidth, int oheight,
1236 int iwidth, int iheight,
1237 int topBand, int bottomBand,
1238 int leftBand, int rightBand);
1239
1240 void img_resample(ImgReSampleContext *s,
1241 AVPicture *output, AVPicture *input);
1242
1243 void img_resample_close(ImgReSampleContext *s);
1244
1245 int avpicture_fill(AVPicture *picture, uint8_t *ptr,
1246 int pix_fmt, int width, int height);
1247 int avpicture_get_size(int pix_fmt, int width, int height);
1248 void avcodec_get_chroma_sub_sample(int pix_fmt, int *h_shift, int *v_shift);
1249 const char *avcodec_get_pix_fmt_name(int pix_fmt);
1250
1251 /* convert among pixel formats */
1252 int img_convert(AVPicture *dst, int dst_pix_fmt,
1253 AVPicture *src, int pix_fmt,
1254 int width, int height);
1255
1256 /* deinterlace a picture */
1257 int avpicture_deinterlace(AVPicture *dst, AVPicture *src,
1258 int pix_fmt, int width, int height);
1259
1260 /* external high level API */
1261
1262 extern AVCodec *first_avcodec;
1263
1264 /* returns LIBAVCODEC_VERSION_INT constant */
1265 unsigned avcodec_version(void);
1266 /* returns LIBAVCODEC_BUILD constant */
1267 unsigned avcodec_build(void);
1268 void avcodec_init(void);
1269
1270 void register_avcodec(AVCodec *format);
1271 AVCodec *avcodec_find_encoder(enum CodecID id);
1272 AVCodec *avcodec_find_encoder_by_name(const char *name);
1273 AVCodec *avcodec_find_decoder(enum CodecID id);
1274 AVCodec *avcodec_find_decoder_by_name(const char *name);
1275 void avcodec_string(char *buf, int buf_size, AVCodecContext *enc, int encode);
1276
1277 void avcodec_get_context_defaults(AVCodecContext *s);
1278 AVCodecContext *avcodec_alloc_context(void);
1279 AVFrame *avcodec_alloc_frame(void);
1280
1281 int avcodec_default_get_buffer(AVCodecContext *s, AVFrame *pic);
1282 void avcodec_default_release_buffer(AVCodecContext *s, AVFrame *pic);
1283
1284 int avcodec_open(AVCodecContext *avctx, AVCodec *codec);
1285 int avcodec_decode_audio(AVCodecContext *avctx, int16_t *samples,
1286 int *frame_size_ptr,
1287 uint8_t *buf, int buf_size);
1288 int avcodec_decode_video(AVCodecContext *avctx, AVFrame *picture,
1289 int *got_picture_ptr,
1290 uint8_t *buf, int buf_size);
1291 int avcodec_parse_frame(AVCodecContext *avctx, uint8_t **pdata,
1292 int *data_size_ptr,
1293 uint8_t *buf, int buf_size);
1294 int avcodec_encode_audio(AVCodecContext *avctx, uint8_t *buf, int buf_size,
1295 const short *samples);
1296 int avcodec_encode_video(AVCodecContext *avctx, uint8_t *buf, int buf_size,
1297 const AVFrame *pict);
1298
1299 int avcodec_close(AVCodecContext *avctx);
1300
1301 void avcodec_register_all(void);
1302
1303 void avcodec_flush_buffers(AVCodecContext *avctx);
1304
1305 /* misc usefull functions */
1306 /**
1307 * reduce a fraction.
1308 * this is usefull for framerate calculations
1309 * @param max the maximum allowed for dst_nom & dst_den
1310 * @return 1 if exact, 0 otherwise
1311 */
1312 int av_reduce(int *dst_nom, int *dst_den, int64_t nom, int64_t den, int64_t max);
1313
1314 /**
1315 * rescale a 64bit integer.
1316 * a simple a*b/c isnt possible as it can overflow
1317 */
1318 int64_t av_rescale(int64_t a, int b, int c);
1319
1320
1321 /**
1322 * Interface for 0.5.0 version
1323 *
1324 * do not even think about it's usage for this moment
1325 */
1326
1327 typedef struct {
1328 /// compressed size used from given memory buffer
1329 int size;
1330 /// I/P/B frame type
1331 int frame_type;
1332 } avc_enc_result_t;
1333
1334 /**
1335 * Commands
1336 * order can't be changed - once it was defined
1337 */
1338 typedef enum {
1339 // general commands
1340 AVC_OPEN_BY_NAME = 0xACA000,
1341 AVC_OPEN_BY_CODEC_ID,
1342 AVC_OPEN_BY_FOURCC,
1343 AVC_CLOSE,
1344
1345 AVC_FLUSH,
1346 // pin - struct { uint8_t* src, uint_t src_size }
1347 // pout - struct { AVPicture* img, consumed_bytes,
1348 AVC_DECODE,
1349 // pin - struct { AVPicture* img, uint8_t* dest, uint_t dest_size }
1350 // pout - uint_t used_from_dest_size
1351 AVC_ENCODE,
1352
1353 // query/get video commands
1354 AVC_GET_VERSION = 0xACB000,
1355 AVC_GET_WIDTH,
1356 AVC_GET_HEIGHT,
1357 AVC_GET_DELAY,
1358 AVC_GET_QUANT_TABLE,
1359 // ...
1360
1361 // query/get audio commands
1362 AVC_GET_FRAME_SIZE = 0xABC000,
1363
1364 // maybe define some simple structure which
1365 // might be passed to the user - but they can't
1366 // contain any codec specific parts and these
1367 // calls are usualy necessary only few times
1368
1369 // set video commands
1370 AVC_SET_WIDTH = 0xACD000,
1371 AVC_SET_HEIGHT,
1372
1373 // set video encoding commands
1374 AVC_SET_FRAME_RATE = 0xACD800,
1375 AVC_SET_QUALITY,
1376 AVC_SET_HURRY_UP,
1377
1378 // set audio commands
1379 AVC_SET_SAMPLE_RATE = 0xACE000,
1380 AVC_SET_CHANNELS,
1381
1382 } avc_cmd_t;
1383
1384 /**
1385 * \param handle allocated private structure by libavcodec
1386 * for initialization pass NULL - will be returned pout
1387 * user is supposed to know nothing about its structure
1388 * \param cmd type of operation to be performed
1389 * \param pint input parameter
1390 * \param pout output parameter
1391 *
1392 * \returns command status - eventually for query command it might return
1393 * integer resulting value
1394 */
1395 int avcodec(void* handle, avc_cmd_t cmd, void* pin, void* pout);
1396
1397 /* memory */
1398 void *av_malloc(unsigned int size);
1399 void *av_mallocz(unsigned int size);
1400 void *av_realloc(void *ptr, unsigned int size);
1401 void av_free(void *ptr);
1402 char *av_strdup(const char *s);
1403 void __av_freep(void **ptr);
1404 #define av_freep(p) __av_freep((void **)(p))
1405 void *av_fast_realloc(void *ptr, unsigned int *size, unsigned int min_size);
1406 /* for static data only */
1407 /* call av_free_static to release all staticaly allocated tables */
1408 void av_free_static(void);
1409 void *__av_mallocz_static(void** location, unsigned int size);
1410 #define av_mallocz_static(p, s) __av_mallocz_static((void **)(p), s)
1411
1412 #ifdef __cplusplus
1413 }
1414 #endif
1415
1416 #endif /* AVCODEC_H */