native Indeo3 decoder implementation
[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 CODEC_ID_INDEO3,
54
55 /* various pcm "codecs" */
56 CODEC_ID_PCM_S16LE,
57 CODEC_ID_PCM_S16BE,
58 CODEC_ID_PCM_U16LE,
59 CODEC_ID_PCM_U16BE,
60 CODEC_ID_PCM_S8,
61 CODEC_ID_PCM_U8,
62 CODEC_ID_PCM_MULAW,
63 CODEC_ID_PCM_ALAW,
64
65 /* various adpcm codecs */
66 CODEC_ID_ADPCM_IMA_QT,
67 CODEC_ID_ADPCM_IMA_WAV,
68 CODEC_ID_ADPCM_MS,
69 };
70
71 enum CodecType {
72 CODEC_TYPE_UNKNOWN = -1,
73 CODEC_TYPE_VIDEO,
74 CODEC_TYPE_AUDIO,
75 };
76
77 /**
78 * Pixel format.
79 */
80 enum PixelFormat {
81 PIX_FMT_YUV420P,
82 PIX_FMT_YUV422,
83 PIX_FMT_RGB24, ///< 3 bytes, R is first
84 PIX_FMT_BGR24, ///< 3 bytes, B is first
85 PIX_FMT_YUV422P,
86 PIX_FMT_YUV444P,
87 PIX_FMT_RGBA32, ///< always stored in cpu endianness
88 PIX_FMT_YUV410P,
89 PIX_FMT_YUV411P,
90 PIX_FMT_RGB565, ///< always stored in cpu endianness
91 PIX_FMT_RGB555, ///< always stored in cpu endianness, most significant bit to 1
92 PIX_FMT_GRAY8,
93 PIX_FMT_MONOWHITE, ///< 0 is white
94 PIX_FMT_MONOBLACK, ///< 0 is black
95 PIX_FMT_PAL8, ///< 8 bit with RGBA palette
96 PIX_FMT_NB,
97 };
98
99 /* currently unused, may be used if 24/32 bits samples ever supported */
100 enum SampleFormat {
101 SAMPLE_FMT_S16 = 0, ///< signed 16 bits
102 };
103
104 /* in bytes */
105 #define AVCODEC_MAX_AUDIO_FRAME_SIZE 131072
106
107 /**
108 * Required number of zero bytes at the end of the input bitstream for decoding.
109 * to avoid overreading (and possibly segfaulting)
110 */
111 #define FF_INPUT_BUFFER_PADDING_SIZE 8
112
113 /* motion estimation type, EPZS by default */
114 enum Motion_Est_ID {
115 ME_ZERO = 1,
116 ME_FULL,
117 ME_LOG,
118 ME_PHODS,
119 ME_EPZS,
120 ME_X1
121 };
122
123 typedef struct RcOverride{
124 int start_frame;
125 int end_frame;
126 int qscale; // if this is 0 then quality_factor will be used instead
127 float quality_factor;
128 } RcOverride;
129
130 /* only for ME compatiblity with old apps */
131 extern int motion_estimation_method;
132
133 /* ME algos sorted by quality */
134 static const int Motion_Est_QTab[] = { ME_ZERO, ME_PHODS, ME_LOG,
135 ME_X1, ME_EPZS, ME_FULL };
136
137
138 #define FF_MAX_B_FRAMES 8
139
140 /* encoding support
141 these flags can be passed in AVCodecContext.flags before initing
142 Note: note not everything is supported yet
143 */
144
145 #define CODEC_FLAG_HQ 0x0001 ///< brute force MB-type decission mode (slow)
146 #define CODEC_FLAG_QSCALE 0x0002 ///< use fixed qscale
147 #define CODEC_FLAG_4MV 0x0004 ///< 4 MV per MB allowed
148 #define CODEC_FLAG_QPEL 0x0010 ///< use qpel MC
149 #define CODEC_FLAG_GMC 0x0020 ///< use GMC
150 #define CODEC_FLAG_PART 0x0080 ///< use data partitioning
151 /* parent program gurantees that the input for b-frame containing streams is not written to
152 for at least s->max_b_frames+1 frames, if this is not set than the input will be copied */
153 #define CODEC_FLAG_INPUT_PRESERVED 0x0100
154 #define CODEC_FLAG_PASS1 0x0200 ///< use internal 2pass ratecontrol in first pass mode
155 #define CODEC_FLAG_PASS2 0x0400 ///< use internal 2pass ratecontrol in second pass mode
156 #define CODEC_FLAG_EXTERN_HUFF 0x1000 ///< use external huffman table (for mjpeg)
157 #define CODEC_FLAG_GRAY 0x2000 ///< only decode/encode grayscale
158 #define CODEC_FLAG_EMU_EDGE 0x4000///< dont draw edges
159 #define CODEC_FLAG_PSNR 0x8000 ///< error[?] variables will be set during encoding
160 #define CODEC_FLAG_TRUNCATED 0x00010000 /** input bitstream might be truncated at a random location instead
161 of only at frame boundaries */
162 #define CODEC_FLAG_NORMALIZE_AQP 0x00020000 ///< normalize adaptive quantization
163 #define CODEC_FLAG_INTERLACED_DCT 0x00040000 ///< use interlaced dct
164 #define CODEC_FLAG_LOW_DELAY 0x00080000 ///< force low delay / will fail on b frames
165 #define CODEC_FLAG_ALT_SCAN 0x00100000 ///< use alternate scan
166 #define CODEC_FLAG_TRELLIS_QUANT 0x00200000 ///< use trellis quantization
167 #define CODEC_FLAG_GLOBAL_HEADER 0x00400000 ///< place global headers in extradata instead of every keyframe
168 #define CODEC_FLAG_BITEXACT 0x00800000 ///< use only bitexact stuff (except (i)dct)
169 /* Fx : Flag for h263+ extra options */
170 #define CODEC_FLAG_H263P_AIC 0x01000000 ///< Advanced intra coding
171 #define CODEC_FLAG_H263P_UMV 0x02000000 ///< Unlimited motion vector
172 /* For advanced prediction mode, we reuse the 4MV flag */
173 /* Unsupported options :
174 * Syntax Arithmetic coding (SAC)
175 * Deblocking filter internal loop
176 * Slice structured
177 * Reference Picture Selection
178 * Independant Segment Decoding
179 * Alternative Inter * VLC
180 * Modified Quantization */
181 /* /Fx */
182 /* codec capabilities */
183
184 #define CODEC_CAP_DRAW_HORIZ_BAND 0x0001 ///< decoder can use draw_horiz_band callback
185 #define CODEC_CAP_DR1 0x0002 ///< direct rendering method 1
186 /* if 'parse_only' field is true, then avcodec_parse_frame() can be
187 used */
188 #define CODEC_CAP_PARSE_ONLY 0x0004
189 #define CODEC_CAP_TRUNCATED 0x0008
190
191 #define FF_COMMON_FRAME \
192 /**\
193 * pointer to the picture planes.\
194 * this might be different from the first allocated byte\
195 * - encoding: \
196 * - decoding: \
197 */\
198 uint8_t *data[4];\
199 int linesize[4];\
200 /**\
201 * pointer to the first allocated byte of the picture. can be used in get_buffer/release_buffer\
202 * this isnt used by lavc unless the default get/release_buffer() is used\
203 * - encoding: \
204 * - decoding: \
205 */\
206 uint8_t *base[4];\
207 /**\
208 * 1 -> keyframe, 0-> not\
209 * - encoding: set by lavc\
210 * - decoding: set by lavc\
211 */\
212 int key_frame;\
213 \
214 /**\
215 * picture type of the frame, see ?_TYPE below.\
216 * - encoding: set by lavc for coded_picture (and set by user for input)\
217 * - decoding: set by lavc\
218 */\
219 int pict_type;\
220 \
221 /**\
222 * presentation timestamp in micro seconds (time when frame should be shown to user)\
223 * if 0 then the frame_rate will be used as reference\
224 * - encoding: MUST be set by user\
225 * - decoding: set by lavc\
226 */\
227 int64_t pts;\
228 \
229 /**\
230 * picture number in bitstream order.\
231 * - encoding: set by\
232 * - decoding: set by lavc\
233 */\
234 int coded_picture_number;\
235 /**\
236 * picture number in display order.\
237 * - encoding: set by\
238 * - decoding: set by lavc\
239 */\
240 int display_picture_number;\
241 \
242 /**\
243 * quality (between 1 (good) and 31 (bad)) \
244 * - encoding: set by lavc for coded_picture (and set by user for input)\
245 * - decoding: set by lavc\
246 */\
247 float quality; \
248 \
249 /**\
250 * buffer age (1->was last buffer and dint change, 2->..., ...).\
251 * set to something large if the buffer has not been used yet \
252 * - encoding: unused\
253 * - decoding: MUST be set by get_buffer()\
254 */\
255 int age;\
256 \
257 /**\
258 * is this picture used as reference\
259 * - encoding: unused\
260 * - decoding: set by lavc (before get_buffer() call))\
261 */\
262 int reference;\
263 \
264 /**\
265 * QP table\
266 * - encoding: unused\
267 * - decoding: set by lavc\
268 */\
269 int8_t *qscale_table;\
270 /**\
271 * QP store stride\
272 * - encoding: unused\
273 * - decoding: set by lavc\
274 */\
275 int qstride;\
276 \
277 /**\
278 * mbskip_table[mb]>=1 if MB didnt change\
279 * stride= mb_width = (width+15)>>4\
280 * - encoding: unused\
281 * - decoding: set by lavc\
282 */\
283 uint8_t *mbskip_table;\
284 \
285 /**\
286 * for some private data of the user\
287 * - encoding: unused\
288 * - decoding: set by user\
289 */\
290 void *opaque;\
291 \
292 /**\
293 * error\
294 * - encoding: set by lavc if flags&CODEC_FLAG_PSNR\
295 * - decoding: unused\
296 */\
297 uint64_t error[4];\
298 \
299 /**\
300 * type of the buffer (to keep track of who has to dealloc data[*])\
301 * - encoding: set by the one who allocs it\
302 * - decoding: set by the one who allocs it\
303 * Note: user allocated (direct rendering) & internal buffers can not coexist currently\
304 */\
305 int type;\
306 \
307 /**\
308 * when decoding, this signal how much the picture must be delayed.\
309 * extra_delay = repeat_pict / (2*fps)\
310 * - encoding: unused\
311 * - decoding: set by lavc\
312 */\
313 int repeat_pict;
314
315
316 #define FF_BUFFER_TYPE_INTERNAL 1
317 #define FF_BUFFER_TYPE_USER 2 ///< Direct rendering buffers
318 #define FF_BUFFER_TYPE_SHARED 4 ///< buffer from somewher else, dont dealloc
319
320
321 #define FF_I_TYPE 1 // Intra
322 #define FF_P_TYPE 2 // Predicted
323 #define FF_B_TYPE 3 // Bi-dir predicted
324 #define FF_S_TYPE 4 // S(GMC)-VOP MPEG4
325 #define FF_SI_TYPE 5
326 #define FF_SP_TYPE 6
327
328 /**
329 * Audio Video Frame.
330 */
331 typedef struct AVFrame {
332 FF_COMMON_FRAME
333 } AVFrame;
334
335 #define DEFAULT_FRAME_RATE_BASE 1001000
336
337 /**
338 * main external api structure.
339 */
340 typedef struct AVCodecContext {
341 /**
342 * the average bitrate.
343 * - encoding: set by user. unused for constant quantizer encoding
344 * - decoding: set by lavc. 0 or some bitrate if this info is available in the stream
345 */
346 int bit_rate;
347
348 /**
349 * number of bits the bitstream is allowed to diverge from the reference.
350 * the reference can be CBR (for CBR pass1) or VBR (for pass2)
351 * - encoding: set by user. unused for constant quantizer encoding
352 * - decoding: unused
353 */
354 int bit_rate_tolerance;
355
356 /**
357 * CODEC_FLAG_*.
358 * - encoding: set by user.
359 * - decoding: set by user.
360 */
361 int flags;
362
363 /**
364 * some codecs needs additionnal format info. It is stored here
365 * - encoding: set by user.
366 * - decoding: set by lavc. (FIXME is this ok?)
367 */
368 int sub_id;
369
370 /**
371 * motion estimation algorithm used for video coding.
372 * - encoding: MUST be set by user.
373 * - decoding: unused
374 */
375 int me_method;
376
377 /**
378 * some codecs need / can use extra-data like huffman tables.
379 * mjpeg: huffman tables
380 * rv10: additional flags
381 * mpeg4: global headers (they can be in the bitstream or here)
382 * - encoding: set/allocated/freed by lavc.
383 * - decoding: set/allocated/freed by user.
384 */
385 void *extradata;
386 int extradata_size;
387
388 /* video only */
389 /**
390 * frames per sec multiplied by frame_rate_base.
391 * for variable fps this is the precission, so if the timestamps
392 * can be specified in msec precssion then this is 1000*frame_rate_base
393 * - encoding: MUST be set by user
394 * - decoding: set by lavc. 0 or the frame_rate if available
395 */
396 int frame_rate;
397
398 /**
399 * width / height.
400 * - encoding: MUST be set by user.
401 * - decoding: set by user, some codecs might override / change it during playback
402 */
403 int width, height;
404
405 #define FF_ASPECT_SQUARE 1
406 #define FF_ASPECT_4_3_625 2
407 #define FF_ASPECT_4_3_525 3
408 #define FF_ASPECT_16_9_625 4
409 #define FF_ASPECT_16_9_525 5
410 #define FF_ASPECT_EXTENDED 15
411
412 /**
413 * the number of pictures in a group of pitures, or 0 for intra_only.
414 * - encoding: set by user.
415 * - decoding: unused
416 */
417 int gop_size;
418
419 /**
420 * pixel format, see PIX_FMT_xxx.
421 * - encoding: unused
422 * - decoding: set by lavc.
423 */
424 enum PixelFormat pix_fmt;
425
426 /**
427 * Frame rate emulation. If not zero lower layer (i.e. format handler)
428 * has to read frames at native frame rate.
429 * - encoding: set by user.
430 * - decoding: unused.
431 */
432 int rate_emu;
433
434 /**
435 * if non NULL, 'draw_horiz_band' is called by the libavcodec
436 * decoder to draw an horizontal band. It improve cache usage. Not
437 * all codecs can do that. You must check the codec capabilities
438 * before
439 * - encoding: unused
440 * - decoding: set by user.
441 */
442 void (*draw_horiz_band)(struct AVCodecContext *s,
443 uint8_t **src_ptr, int linesize,
444 int y, int width, int height);
445
446 /* audio only */
447 int sample_rate; ///< samples per sec
448 int channels;
449 int sample_fmt; ///< sample format, currenly unused
450
451 /* the following data should not be initialized */
452 int frame_size; ///< in samples, initialized when calling 'init'
453 int frame_number; ///< audio or video frame number
454 int real_pict_num; ///< returns the real picture number of previous encoded frame
455
456 /**
457 * number of frames the decoded output will be delayed relative to
458 * the encoded input.
459 * - encoding: set by lavc.
460 * - decoding: unused
461 */
462 int delay;
463
464 /* - encoding parameters */
465 float qcompress; ///< amount of qscale change between easy & hard scenes (0.0-1.0)
466 float qblur; ///< amount of qscale smoothing over time (0.0-1.0)
467
468 /**
469 * minimum quantizer.
470 * - encoding: set by user.
471 * - decoding: unused
472 */
473 int qmin;
474
475 /**
476 * maximum quantizer.
477 * - encoding: set by user.
478 * - decoding: unused
479 */
480 int qmax;
481
482 /**
483 * maximum quantizer difference etween frames.
484 * - encoding: set by user.
485 * - decoding: unused
486 */
487 int max_qdiff;
488
489 /**
490 * maximum number of b frames between non b frames.
491 * note: the output will be delayed by max_b_frames+1 relative to the input
492 * - encoding: set by user.
493 * - decoding: unused
494 */
495 int max_b_frames;
496
497 /**
498 * qscale factor between ip and b frames.
499 * - encoding: set by user.
500 * - decoding: unused
501 */
502 float b_quant_factor;
503
504 /** obsolete FIXME remove */
505 int rc_strategy;
506 int b_frame_strategy;
507
508 /**
509 * hurry up amount.
510 * - encoding: unused
511 * - decoding: set by user. 1-> skip b frames, 2-> skip idct/dequant too, 5-> skip everything except header
512 */
513 int hurry_up;
514
515 struct AVCodec *codec;
516
517 void *priv_data;
518
519 /* The following data is for RTP friendly coding */
520 /* By now only H.263/H.263+/MPEG4 coder honours this */
521 int rtp_mode; /* 1 for activate RTP friendly-mode */
522 /* highers numbers represent more error-prone */
523 /* enviroments, by now just "1" exist */
524
525 int rtp_payload_size; /* The size of the RTP payload, the coder will */
526 /* do it's best to deliver a chunk with size */
527 /* below rtp_payload_size, the chunk will start */
528 /* with a start code on some codecs like H.263 */
529 /* This doesn't take account of any particular */
530 /* headers inside the transmited RTP payload */
531
532
533 /* The RTP callcack: This function is called */
534 /* every time the encoder as a packet to send */
535 /* Depends on the encoder if the data starts */
536 /* with a Start Code (it should) H.263 does */
537 void (*rtp_callback)(void *data, int size, int packet_number);
538
539 /* statistics, used for 2-pass encoding */
540 int mv_bits;
541 int header_bits;
542 int i_tex_bits;
543 int p_tex_bits;
544 int i_count;
545 int p_count;
546 int skip_count;
547 int misc_bits;
548
549 /**
550 * number of bits used for the previously encoded frame.
551 * - encoding: set by lavc
552 * - decoding: unused
553 */
554 int frame_bits;
555
556 /**
557 * private data of the user, can be used to carry app specific stuff.
558 * - encoding: set by user
559 * - decoding: set by user
560 */
561 void *opaque;
562
563 char codec_name[32];
564 enum CodecType codec_type; /* see CODEC_TYPE_xxx */
565 enum CodecID codec_id; /* see CODEC_ID_xxx */
566
567 /**
568 * fourcc (LSB first, so "ABCD" -> ('D'<<24) + ('C'<<16) + ('B'<<8) + 'A').
569 * this is used to workaround some encoder bugs
570 * - encoding: unused
571 * - decoding: set by user, will be converted to upper case by lavc during init
572 */
573 unsigned int codec_tag;
574
575 /**
576 * workaround bugs in encoders which sometimes cannot be detected automatically.
577 * - encoding: unused
578 * - decoding: set by user
579 */
580 int workaround_bugs;
581 #define FF_BUG_AUTODETECT 1 ///< autodetection
582 #define FF_BUG_OLD_MSMPEG4 2
583 #define FF_BUG_XVID_ILACE 4
584 #define FF_BUG_UMP4 8
585 #define FF_BUG_NO_PADDING 16
586 #define FF_BUG_AC_VLC 32
587 #define FF_BUG_QPEL_CHROMA 64
588 #define FF_BUG_STD_QPEL 128
589 #define FF_BUG_QPEL_CHROMA2 256
590 #define FF_BUG_DIRECT_BLOCKSIZE 512
591 #define FF_BUG_EDGE 1024
592 //#define FF_BUG_FAKE_SCALABILITY 16 //autodetection should work 100%
593
594 /**
595 * luma single coeff elimination threshold.
596 * - encoding: set by user
597 * - decoding: unused
598 */
599 int luma_elim_threshold;
600
601 /**
602 * chroma single coeff elimination threshold.
603 * - encoding: set by user
604 * - decoding: unused
605 */
606 int chroma_elim_threshold;
607
608 /**
609 * strictly follow the std (MPEG4, ...).
610 * - encoding: set by user
611 * - decoding: unused
612 */
613 int strict_std_compliance;
614
615 /**
616 * qscale offset between ip and b frames.
617 * if > 0 then the last p frame quantizer will be used (q= lastp_q*factor+offset)
618 * if < 0 then normal ratecontrol will be done (q= -normal_q*factor+offset)
619 * - encoding: set by user.
620 * - decoding: unused
621 */
622 float b_quant_offset;
623
624 /**
625 * error resilience higher values will detect more errors but may missdetect
626 * some more or less valid parts as errors.
627 * - encoding: unused
628 * - decoding: set by user
629 */
630 int error_resilience;
631 #define FF_ER_CAREFULL 1
632 #define FF_ER_COMPLIANT 2
633 #define FF_ER_AGGRESSIVE 3
634 #define FF_ER_VERY_AGGRESSIVE 4
635
636 /**
637 * called at the beginning of each frame to get a buffer for it.
638 * if pic.reference is set then the frame will be read later by lavc
639 * - encoding: unused
640 * - decoding: set by lavc, user can override
641 */
642 int (*get_buffer)(struct AVCodecContext *c, AVFrame *pic);
643
644 /**
645 * called to release buffers which where allocated with get_buffer.
646 * a released buffer can be reused in get_buffer()
647 * pic.data[*] must be set to NULL
648 * - encoding: unused
649 * - decoding: set by lavc, user can override
650 */
651 void (*release_buffer)(struct AVCodecContext *c, AVFrame *pic);
652
653 /**
654 * is 1 if the decoded stream contains b frames, 0 otherwise.
655 * - encoding: unused
656 * - decoding: set by lavc
657 */
658 int has_b_frames;
659
660 int block_align; ///< used by some WAV based audio codecs
661
662 int parse_only; /* - decoding only: if true, only parsing is done
663 (function avcodec_parse_frame()). The frame
664 data is returned. Only MPEG codecs support this now. */
665
666 /**
667 * 0-> h263 quant 1-> mpeg quant.
668 * - encoding: set by user.
669 * - decoding: unused
670 */
671 int mpeg_quant;
672
673 /**
674 * pass1 encoding statistics output buffer.
675 * - encoding: set by lavc
676 * - decoding: unused
677 */
678 char *stats_out;
679
680 /**
681 * pass2 encoding statistics input buffer.
682 * concatenated stuff from stats_out of pass1 should be placed here
683 * - encoding: allocated/set/freed by user
684 * - decoding: unused
685 */
686 char *stats_in;
687
688 /**
689 * ratecontrol qmin qmax limiting method.
690 * 0-> clipping, 1-> use a nice continous function to limit qscale wthin qmin/qmax
691 * - encoding: set by user.
692 * - decoding: unused
693 */
694 float rc_qsquish;
695
696 float rc_qmod_amp;
697 int rc_qmod_freq;
698
699 /**
700 * ratecontrol override, see RcOverride.
701 * - encoding: allocated/set/freed by user.
702 * - decoding: unused
703 */
704 RcOverride *rc_override;
705 int rc_override_count;
706
707 /**
708 * rate control equation.
709 * - encoding: set by user
710 * - decoding: unused
711 */
712 char *rc_eq;
713
714 /**
715 * maximum bitrate.
716 * - encoding: set by user.
717 * - decoding: unused
718 */
719 int rc_max_rate;
720
721 /**
722 * minimum bitrate.
723 * - encoding: set by user.
724 * - decoding: unused
725 */
726 int rc_min_rate;
727
728 /**
729 * decoder bitstream buffer size.
730 * - encoding: set by user.
731 * - decoding: unused
732 */
733 int rc_buffer_size;
734 float rc_buffer_aggressivity;
735
736 /**
737 * qscale factor between p and i frames.
738 * - encoding: set by user.
739 * - decoding: unused
740 */
741 float i_quant_factor;
742
743 /**
744 * qscale offset between p and i frames.
745 * if > 0 then the last p frame quantizer will be used (q= lastp_q*factor+offset)
746 * if < 0 then normal ratecontrol will be done (q= -normal_q*factor+offset)
747 * - encoding: set by user.
748 * - decoding: unused
749 */
750 float i_quant_offset;
751
752 /**
753 * initial complexity for pass1 ratecontrol.
754 * - encoding: set by user.
755 * - decoding: unused
756 */
757 float rc_initial_cplx;
758
759 /**
760 * dct algorithm, see FF_DCT_* below.
761 * - encoding: set by user
762 * - decoding: unused
763 */
764 int dct_algo;
765 #define FF_DCT_AUTO 0
766 #define FF_DCT_FASTINT 1
767 #define FF_DCT_INT 2
768 #define FF_DCT_MMX 3
769 #define FF_DCT_MLIB 4
770 #define FF_DCT_ALTIVEC 5
771
772 /**
773 * luminance masking (0-> disabled).
774 * - encoding: set by user
775 * - decoding: unused
776 */
777 float lumi_masking;
778
779 /**
780 * temporary complexity masking (0-> disabled).
781 * - encoding: set by user
782 * - decoding: unused
783 */
784 float temporal_cplx_masking;
785
786 /**
787 * spatial complexity masking (0-> disabled).
788 * - encoding: set by user
789 * - decoding: unused
790 */
791 float spatial_cplx_masking;
792
793 /**
794 * p block masking (0-> disabled).
795 * - encoding: set by user
796 * - decoding: unused
797 */
798 float p_masking;
799
800 /**
801 * darkness masking (0-> disabled).
802 * - encoding: set by user
803 * - decoding: unused
804 */
805 float dark_masking;
806
807
808 /* for binary compatibility */
809 int unused;
810
811 /**
812 * idct algorithm, see FF_IDCT_* below.
813 * - encoding: set by user
814 * - decoding: set by user
815 */
816 int idct_algo;
817 #define FF_IDCT_AUTO 0
818 #define FF_IDCT_INT 1
819 #define FF_IDCT_SIMPLE 2
820 #define FF_IDCT_SIMPLEMMX 3
821 #define FF_IDCT_LIBMPEG2MMX 4
822 #define FF_IDCT_PS2 5
823 #define FF_IDCT_MLIB 6
824 #define FF_IDCT_ARM 7
825 #define FF_IDCT_ALTIVEC 8
826
827 /**
828 * slice count.
829 * - encoding: set by lavc
830 * - decoding: set by user (or 0)
831 */
832 int slice_count;
833 /**
834 * slice offsets in the frame in bytes.
835 * - encoding: set/allocated by lavc
836 * - decoding: set/allocated by user (or NULL)
837 */
838 int *slice_offset;
839
840 /**
841 * error concealment flags.
842 * - encoding: unused
843 * - decoding: set by user
844 */
845 int error_concealment;
846 #define FF_EC_GUESS_MVS 1
847 #define FF_EC_DEBLOCK 2
848
849 /**
850 * dsp_mask could be add used to disable unwanted CPU features
851 * CPU features (i.e. MMX, SSE. ...)
852 *
853 * with FORCE flag you may instead enable given CPU features
854 * (Dangerous: usable in case of misdetection, improper usage however will
855 * result into program crash)
856 */
857 unsigned dsp_mask;
858 #define FF_MM_FORCE 0x80000000 /* force usage of selected flags (OR) */
859 /* lower 16 bits - CPU features */
860 #ifdef HAVE_MMX
861 #define FF_MM_MMX 0x0001 /* standard MMX */
862 #define FF_MM_3DNOW 0x0004 /* AMD 3DNOW */
863 #define FF_MM_MMXEXT 0x0002 /* SSE integer functions or AMD MMX ext */
864 #define FF_MM_SSE 0x0008 /* SSE functions */
865 #define FF_MM_SSE2 0x0010 /* PIV SSE2 functions */
866 #endif /* HAVE_MMX */
867
868 /**
869 * bits per sample/pixel from the demuxer (needed for huffyuv).
870 * - encoding: set by lavc
871 * - decoding: set by user
872 */
873 int bits_per_sample;
874
875 /**
876 * prediction method (needed for huffyuv).
877 * - encoding: set by user
878 * - decoding: unused
879 */
880 int prediction_method;
881 #define FF_PRED_LEFT 0
882 #define FF_PRED_PLANE 1
883 #define FF_PRED_MEDIAN 2
884
885 /**
886 * aspect ratio (0 if unknown).
887 * - encoding: set by user.
888 * - decoding: set by lavc.
889 */
890 float aspect_ratio;
891
892 /**
893 * the picture in the bitstream.
894 * - encoding: set by lavc
895 * - decoding: set by lavc
896 */
897 AVFrame *coded_frame;
898
899 /**
900 * debug.
901 * - encoding: set by user.
902 * - decoding: set by user.
903 */
904 int debug;
905 #define FF_DEBUG_PICT_INFO 1
906 #define FF_DEBUG_RC 2
907 #define FF_DEBUG_BITSTREAM 4
908 #define FF_DEBUG_MB_TYPE 8
909 #define FF_DEBUG_QP 16
910 #define FF_DEBUG_MV 32
911 #define FF_DEBUG_VIS_MV 0x00000040
912 #define FF_DEBUG_SKIP 0x00000080
913 #define FF_DEBUG_STARTCODE 0x00000100
914 #define FF_DEBUG_PTS 0x00000200
915 #define FF_DEBUG_ER 0x00000400
916 #define FF_DEBUG_MMCO 0x00000800
917
918 /**
919 * error.
920 * - encoding: set by lavc if flags&CODEC_FLAG_PSNR
921 * - decoding: unused
922 */
923 uint64_t error[4];
924
925 /**
926 * minimum MB quantizer.
927 * - encoding: set by user.
928 * - decoding: unused
929 */
930 int mb_qmin;
931
932 /**
933 * maximum MB quantizer.
934 * - encoding: set by user.
935 * - decoding: unused
936 */
937 int mb_qmax;
938
939 /**
940 * motion estimation compare function.
941 * - encoding: set by user.
942 * - decoding: unused
943 */
944 int me_cmp;
945 /**
946 * subpixel motion estimation compare function.
947 * - encoding: set by user.
948 * - decoding: unused
949 */
950 int me_sub_cmp;
951 /**
952 * macroblock compare function (not supported yet).
953 * - encoding: set by user.
954 * - decoding: unused
955 */
956 int mb_cmp;
957 #define FF_CMP_SAD 0
958 #define FF_CMP_SSE 1
959 #define FF_CMP_SATD 2
960 #define FF_CMP_DCT 3
961 #define FF_CMP_PSNR 4
962 #define FF_CMP_BIT 5
963 #define FF_CMP_RD 6
964 #define FF_CMP_ZERO 7
965 #define FF_CMP_CHROMA 256
966
967 /**
968 * ME diamond size & shape.
969 * - encoding: set by user.
970 * - decoding: unused
971 */
972 int dia_size;
973
974 /**
975 * amount of previous MV predictors (2a+1 x 2a+1 square).
976 * - encoding: set by user.
977 * - decoding: unused
978 */
979 int last_predictor_count;
980
981 /**
982 * pre pass for motion estimation.
983 * - encoding: set by user.
984 * - decoding: unused
985 */
986 int pre_me;
987
988 /**
989 * motion estimation pre pass compare function.
990 * - encoding: set by user.
991 * - decoding: unused
992 */
993 int me_pre_cmp;
994
995 /**
996 * ME pre pass diamond size & shape.
997 * - encoding: set by user.
998 * - decoding: unused
999 */
1000 int pre_dia_size;
1001
1002 /**
1003 * subpel ME quality.
1004 * - encoding: set by user.
1005 * - decoding: unused
1006 */
1007 int me_subpel_quality;
1008
1009 /**
1010 * callback to negotiate the pixelFormat.
1011 * @param fmt is the list of formats which are supported by the codec,
1012 * its terminated by -1 as 0 is a valid format, the formats are ordered by quality
1013 * the first is allways the native one
1014 * @return the choosen format
1015 * - encoding: unused
1016 * - decoding: set by user, if not set then the native format will always be choosen
1017 */
1018 enum PixelFormat (*get_format)(struct AVCodecContext *s, enum PixelFormat * fmt);
1019
1020 /**
1021 * DTG active format information (additionnal aspect ratio
1022 * information only used in DVB MPEG2 transport streams). 0 if
1023 * not set.
1024 *
1025 * - encoding: unused.
1026 * - decoding: set by decoder
1027 */
1028 int dtg_active_format;
1029 #define FF_DTG_AFD_SAME 8
1030 #define FF_DTG_AFD_4_3 9
1031 #define FF_DTG_AFD_16_9 10
1032 #define FF_DTG_AFD_14_9 11
1033 #define FF_DTG_AFD_4_3_SP_14_9 13
1034 #define FF_DTG_AFD_16_9_SP_14_9 14
1035 #define FF_DTG_AFD_SP_4_3 15
1036
1037 /**
1038 * Maximum motion estimation search range in subpel units.
1039 * if 0 then no limit
1040 *
1041 * - encoding: set by user.
1042 * - decoding: unused.
1043 */
1044 int me_range;
1045
1046 /**
1047 * frame_rate_base.
1048 * for variable fps this is 1
1049 * - encoding: set by user.
1050 * - decoding: set by lavc.
1051 * @todo move this after frame_rate
1052 */
1053
1054 int frame_rate_base;
1055 /**
1056 * intra quantizer bias.
1057 * - encoding: set by user.
1058 * - decoding: unused
1059 */
1060 int intra_quant_bias;
1061 #define FF_DEFAULT_QUANT_BIAS 999999
1062
1063 /**
1064 * inter quantizer bias.
1065 * - encoding: set by user.
1066 * - decoding: unused
1067 */
1068 int inter_quant_bias;
1069
1070 } AVCodecContext;
1071
1072
1073 /**
1074 * AVOption.
1075 */
1076 typedef struct AVOption {
1077 /** options' name */
1078 const char *name; /* if name is NULL, it indicates a link to next */
1079 /** short English text help or const struct AVOption* subpointer */
1080 const char *help; // const struct AVOption* sub;
1081 /** offset to context structure where the parsed value should be stored */
1082 int offset;
1083 /** options' type */
1084 int type;
1085 #define FF_OPT_TYPE_BOOL 1 ///< boolean - true,1,on (or simply presence)
1086 #define FF_OPT_TYPE_DOUBLE 2 ///< double
1087 #define FF_OPT_TYPE_INT 3 ///< integer
1088 #define FF_OPT_TYPE_STRING 4 ///< string (finished with \0)
1089 #define FF_OPT_TYPE_MASK 0x1f ///< mask for types - upper bits are various flags
1090 //#define FF_OPT_TYPE_EXPERT 0x20 // flag for expert option
1091 #define FF_OPT_TYPE_FLAG (FF_OPT_TYPE_BOOL | 0x40)
1092 #define FF_OPT_TYPE_RCOVERRIDE (FF_OPT_TYPE_STRING | 0x80)
1093 /** min value (min == max -> no limits) */
1094 double min;
1095 /** maximum value for double/int */
1096 double max;
1097 /** default boo [0,1]l/double/int value */
1098 double defval;
1099 /**
1100 * default string value (with optional semicolon delimited extra option-list
1101 * i.e. option1;option2;option3
1102 * defval might select other then first argument as default
1103 */
1104 const char *defstr;
1105 #define FF_OPT_MAX_DEPTH 10
1106 } AVOption;
1107
1108 /**
1109 * Parse option(s) and sets fields in passed structure
1110 * @param strct structure where the parsed results will be written
1111 * @param list list with AVOptions
1112 * @param opts string with options for parsing
1113 */
1114 int avoption_parse(void* strct, const AVOption* list, const char* opts);
1115
1116
1117 /**
1118 * AVCodec.
1119 */
1120 typedef struct AVCodec {
1121 const char *name;
1122 int type;
1123 int id;
1124 int priv_data_size;
1125 int (*init)(AVCodecContext *);
1126 int (*encode)(AVCodecContext *, uint8_t *buf, int buf_size, void *data);
1127 int (*close)(AVCodecContext *);
1128 int (*decode)(AVCodecContext *, void *outdata, int *outdata_size,
1129 uint8_t *buf, int buf_size);
1130 int capabilities;
1131 const AVOption *options;
1132 struct AVCodec *next;
1133 } AVCodec;
1134
1135 /**
1136 * four components are given, that's all.
1137 * the last component is alpha
1138 */
1139 typedef struct AVPicture {
1140 uint8_t *data[4];
1141 int linesize[4];
1142 } AVPicture;
1143
1144 extern AVCodec ac3_encoder;
1145 extern AVCodec mp2_encoder;
1146 extern AVCodec mp3lame_encoder;
1147 extern AVCodec oggvorbis_encoder;
1148 extern AVCodec mpeg1video_encoder;
1149 extern AVCodec h263_encoder;
1150 extern AVCodec h263p_encoder;
1151 extern AVCodec rv10_encoder;
1152 extern AVCodec mjpeg_encoder;
1153 extern AVCodec mpeg4_encoder;
1154 extern AVCodec msmpeg4v1_encoder;
1155 extern AVCodec msmpeg4v2_encoder;
1156 extern AVCodec msmpeg4v3_encoder;
1157 extern AVCodec wmv1_encoder;
1158 extern AVCodec wmv2_encoder;
1159 extern AVCodec huffyuv_encoder;
1160 extern AVCodec h264_encoder;
1161
1162 extern AVCodec h263_decoder;
1163 extern AVCodec mpeg4_decoder;
1164 extern AVCodec msmpeg4v1_decoder;
1165 extern AVCodec msmpeg4v2_decoder;
1166 extern AVCodec msmpeg4v3_decoder;
1167 extern AVCodec wmv1_decoder;
1168 extern AVCodec wmv2_decoder;
1169 extern AVCodec mpeg_decoder;
1170 extern AVCodec h263i_decoder;
1171 extern AVCodec rv10_decoder;
1172 extern AVCodec svq1_decoder;
1173 extern AVCodec dvvideo_decoder;
1174 extern AVCodec dvaudio_decoder;
1175 extern AVCodec wmav1_decoder;
1176 extern AVCodec wmav2_decoder;
1177 extern AVCodec mjpeg_decoder;
1178 extern AVCodec mjpegb_decoder;
1179 extern AVCodec mp2_decoder;
1180 extern AVCodec mp3_decoder;
1181 extern AVCodec mace3_decoder;
1182 extern AVCodec mace6_decoder;
1183 extern AVCodec huffyuv_decoder;
1184 extern AVCodec oggvorbis_decoder;
1185 extern AVCodec cyuv_decoder;
1186 extern AVCodec h264_decoder;
1187 extern AVCodec indeo3_decoder;
1188
1189 /* pcm codecs */
1190 #define PCM_CODEC(id, name) \
1191 extern AVCodec name ## _decoder; \
1192 extern AVCodec name ## _encoder
1193
1194 PCM_CODEC(CODEC_ID_PCM_S16LE, pcm_s16le);
1195 PCM_CODEC(CODEC_ID_PCM_S16BE, pcm_s16be);
1196 PCM_CODEC(CODEC_ID_PCM_U16LE, pcm_u16le);
1197 PCM_CODEC(CODEC_ID_PCM_U16BE, pcm_u16be);
1198 PCM_CODEC(CODEC_ID_PCM_S8, pcm_s8);
1199 PCM_CODEC(CODEC_ID_PCM_U8, pcm_u8);
1200 PCM_CODEC(CODEC_ID_PCM_ALAW, pcm_alaw);
1201 PCM_CODEC(CODEC_ID_PCM_MULAW, pcm_mulaw);
1202
1203 /* adpcm codecs */
1204
1205 PCM_CODEC(CODEC_ID_ADPCM_IMA_QT, adpcm_ima_qt);
1206 PCM_CODEC(CODEC_ID_ADPCM_IMA_WAV, adpcm_ima_wav);
1207 PCM_CODEC(CODEC_ID_ADPCM_MS, adpcm_ms);
1208
1209 #undef PCM_CODEC
1210
1211 /* dummy raw video codec */
1212 extern AVCodec rawvideo_codec;
1213
1214 /* the following codecs use external GPL libs */
1215 extern AVCodec ac3_decoder;
1216
1217 /* resample.c */
1218
1219 struct ReSampleContext;
1220
1221 typedef struct ReSampleContext ReSampleContext;
1222
1223 ReSampleContext *audio_resample_init(int output_channels, int input_channels,
1224 int output_rate, int input_rate);
1225 int audio_resample(ReSampleContext *s, short *output, short *input, int nb_samples);
1226 void audio_resample_close(ReSampleContext *s);
1227
1228 /* YUV420 format is assumed ! */
1229
1230 struct ImgReSampleContext;
1231
1232 typedef struct ImgReSampleContext ImgReSampleContext;
1233
1234 ImgReSampleContext *img_resample_init(int output_width, int output_height,
1235 int input_width, int input_height);
1236
1237 ImgReSampleContext *img_resample_full_init(int owidth, int oheight,
1238 int iwidth, int iheight,
1239 int topBand, int bottomBand,
1240 int leftBand, int rightBand);
1241
1242 void img_resample(ImgReSampleContext *s,
1243 AVPicture *output, AVPicture *input);
1244
1245 void img_resample_close(ImgReSampleContext *s);
1246
1247 int avpicture_fill(AVPicture *picture, uint8_t *ptr,
1248 int pix_fmt, int width, int height);
1249 int avpicture_get_size(int pix_fmt, int width, int height);
1250 void avcodec_get_chroma_sub_sample(int pix_fmt, int *h_shift, int *v_shift);
1251 const char *avcodec_get_pix_fmt_name(int pix_fmt);
1252
1253 /* convert among pixel formats */
1254 int img_convert(AVPicture *dst, int dst_pix_fmt,
1255 AVPicture *src, int pix_fmt,
1256 int width, int height);
1257
1258 /* deinterlace a picture */
1259 int avpicture_deinterlace(AVPicture *dst, AVPicture *src,
1260 int pix_fmt, int width, int height);
1261
1262 /* external high level API */
1263
1264 extern AVCodec *first_avcodec;
1265
1266 /* returns LIBAVCODEC_VERSION_INT constant */
1267 unsigned avcodec_version(void);
1268 /* returns LIBAVCODEC_BUILD constant */
1269 unsigned avcodec_build(void);
1270 void avcodec_init(void);
1271
1272 void register_avcodec(AVCodec *format);
1273 AVCodec *avcodec_find_encoder(enum CodecID id);
1274 AVCodec *avcodec_find_encoder_by_name(const char *name);
1275 AVCodec *avcodec_find_decoder(enum CodecID id);
1276 AVCodec *avcodec_find_decoder_by_name(const char *name);
1277 void avcodec_string(char *buf, int buf_size, AVCodecContext *enc, int encode);
1278
1279 void avcodec_get_context_defaults(AVCodecContext *s);
1280 AVCodecContext *avcodec_alloc_context(void);
1281 AVFrame *avcodec_alloc_frame(void);
1282
1283 int avcodec_default_get_buffer(AVCodecContext *s, AVFrame *pic);
1284 void avcodec_default_release_buffer(AVCodecContext *s, AVFrame *pic);
1285
1286 int avcodec_open(AVCodecContext *avctx, AVCodec *codec);
1287 int avcodec_decode_audio(AVCodecContext *avctx, int16_t *samples,
1288 int *frame_size_ptr,
1289 uint8_t *buf, int buf_size);
1290 int avcodec_decode_video(AVCodecContext *avctx, AVFrame *picture,
1291 int *got_picture_ptr,
1292 uint8_t *buf, int buf_size);
1293 int avcodec_parse_frame(AVCodecContext *avctx, uint8_t **pdata,
1294 int *data_size_ptr,
1295 uint8_t *buf, int buf_size);
1296 int avcodec_encode_audio(AVCodecContext *avctx, uint8_t *buf, int buf_size,
1297 const short *samples);
1298 int avcodec_encode_video(AVCodecContext *avctx, uint8_t *buf, int buf_size,
1299 const AVFrame *pict);
1300
1301 int avcodec_close(AVCodecContext *avctx);
1302
1303 void avcodec_register_all(void);
1304
1305 void avcodec_flush_buffers(AVCodecContext *avctx);
1306
1307 /* misc usefull functions */
1308 /**
1309 * reduce a fraction.
1310 * this is usefull for framerate calculations
1311 * @param max the maximum allowed for dst_nom & dst_den
1312 * @return 1 if exact, 0 otherwise
1313 */
1314 int av_reduce(int *dst_nom, int *dst_den, int64_t nom, int64_t den, int64_t max);
1315
1316 /**
1317 * rescale a 64bit integer.
1318 * a simple a*b/c isnt possible as it can overflow
1319 */
1320 int64_t av_rescale(int64_t a, int b, int c);
1321
1322
1323 /**
1324 * Interface for 0.5.0 version
1325 *
1326 * do not even think about it's usage for this moment
1327 */
1328
1329 typedef struct {
1330 /// compressed size used from given memory buffer
1331 int size;
1332 /// I/P/B frame type
1333 int frame_type;
1334 } avc_enc_result_t;
1335
1336 /**
1337 * Commands
1338 * order can't be changed - once it was defined
1339 */
1340 typedef enum {
1341 // general commands
1342 AVC_OPEN_BY_NAME = 0xACA000,
1343 AVC_OPEN_BY_CODEC_ID,
1344 AVC_OPEN_BY_FOURCC,
1345 AVC_CLOSE,
1346
1347 AVC_FLUSH,
1348 // pin - struct { uint8_t* src, uint_t src_size }
1349 // pout - struct { AVPicture* img, consumed_bytes,
1350 AVC_DECODE,
1351 // pin - struct { AVPicture* img, uint8_t* dest, uint_t dest_size }
1352 // pout - uint_t used_from_dest_size
1353 AVC_ENCODE,
1354
1355 // query/get video commands
1356 AVC_GET_VERSION = 0xACB000,
1357 AVC_GET_WIDTH,
1358 AVC_GET_HEIGHT,
1359 AVC_GET_DELAY,
1360 AVC_GET_QUANT_TABLE,
1361 // ...
1362
1363 // query/get audio commands
1364 AVC_GET_FRAME_SIZE = 0xABC000,
1365
1366 // maybe define some simple structure which
1367 // might be passed to the user - but they can't
1368 // contain any codec specific parts and these
1369 // calls are usualy necessary only few times
1370
1371 // set video commands
1372 AVC_SET_WIDTH = 0xACD000,
1373 AVC_SET_HEIGHT,
1374
1375 // set video encoding commands
1376 AVC_SET_FRAME_RATE = 0xACD800,
1377 AVC_SET_QUALITY,
1378 AVC_SET_HURRY_UP,
1379
1380 // set audio commands
1381 AVC_SET_SAMPLE_RATE = 0xACE000,
1382 AVC_SET_CHANNELS,
1383
1384 } avc_cmd_t;
1385
1386 /**
1387 * \param handle allocated private structure by libavcodec
1388 * for initialization pass NULL - will be returned pout
1389 * user is supposed to know nothing about its structure
1390 * \param cmd type of operation to be performed
1391 * \param pint input parameter
1392 * \param pout output parameter
1393 *
1394 * \returns command status - eventually for query command it might return
1395 * integer resulting value
1396 */
1397 int avcodec(void* handle, avc_cmd_t cmd, void* pin, void* pout);
1398
1399 /* memory */
1400 void *av_malloc(unsigned int size);
1401 void *av_mallocz(unsigned int size);
1402 void *av_realloc(void *ptr, unsigned int size);
1403 void av_free(void *ptr);
1404 char *av_strdup(const char *s);
1405 void __av_freep(void **ptr);
1406 #define av_freep(p) __av_freep((void **)(p))
1407 void *av_fast_realloc(void *ptr, unsigned int *size, unsigned int min_size);
1408 /* for static data only */
1409 /* call av_free_static to release all staticaly allocated tables */
1410 void av_free_static(void);
1411 void *__av_mallocz_static(void** location, unsigned int size);
1412 #define av_mallocz_static(p, s) __av_mallocz_static((void **)(p), s)
1413
1414 #ifdef __cplusplus
1415 }
1416 #endif
1417
1418 #endif /* AVCODEC_H */