d9775db527803e905faaebcf657dde8e9e9bc6f1
[libav.git] / libavcodec / avcodec.h
1 #ifndef AVCODEC_H
2 #define AVCODEC_H
3
4 #include "common.h"
5
6 #define LIBAVCODEC_VERSION_INT 0x000406
7 #define LIBAVCODEC_VERSION "0.4.6"
8 #define LIBAVCODEC_BUILD 4614
9 #define LIBAVCODEC_BUILD_STR "4614"
10
11 enum CodecID {
12 CODEC_ID_NONE,
13 CODEC_ID_MPEG1VIDEO,
14 CODEC_ID_H263,
15 CODEC_ID_RV10,
16 CODEC_ID_MP2,
17 CODEC_ID_MP3LAME,
18 CODEC_ID_AC3,
19 CODEC_ID_MJPEG,
20 CODEC_ID_MPEG4,
21 CODEC_ID_RAWVIDEO,
22 CODEC_ID_MSMPEG4V1,
23 CODEC_ID_MSMPEG4V2,
24 CODEC_ID_MSMPEG4V3,
25 CODEC_ID_WMV1,
26 CODEC_ID_WMV2,
27 CODEC_ID_H263P,
28 CODEC_ID_H263I,
29
30 /* various pcm "codecs" */
31 CODEC_ID_PCM_S16LE,
32 CODEC_ID_PCM_S16BE,
33 CODEC_ID_PCM_U16LE,
34 CODEC_ID_PCM_U16BE,
35 CODEC_ID_PCM_S8,
36 CODEC_ID_PCM_U8,
37 CODEC_ID_PCM_MULAW,
38 CODEC_ID_PCM_ALAW,
39 };
40 #define CODEC_ID_MSMPEG4 CODEC_ID_MSMPEG4V3
41
42 enum CodecType {
43 CODEC_TYPE_UNKNOWN = -1,
44 CODEC_TYPE_VIDEO,
45 CODEC_TYPE_AUDIO,
46 };
47
48 enum PixelFormat {
49 PIX_FMT_ANY = -1,
50 PIX_FMT_YUV420P,
51 PIX_FMT_YUV422,
52 PIX_FMT_RGB24,
53 PIX_FMT_BGR24,
54 PIX_FMT_YUV422P,
55 PIX_FMT_YUV444P,
56 };
57
58 /* currently unused, may be used if 24/32 bits samples ever supported */
59 enum SampleFormat {
60 SAMPLE_FMT_S16 = 0, /* signed 16 bits */
61 };
62
63 /* in bytes */
64 #define AVCODEC_MAX_AUDIO_FRAME_SIZE 18432
65
66 /* motion estimation type, EPZS by default */
67 enum Motion_Est_ID {
68 ME_ZERO = 1,
69 ME_FULL,
70 ME_LOG,
71 ME_PHODS,
72 ME_EPZS,
73 ME_X1
74 };
75
76 /* only for ME compatiblity with old apps */
77 extern int motion_estimation_method;
78
79 /* ME algos sorted by quality */
80 static const int Motion_Est_QTab[] = { ME_ZERO, ME_PHODS, ME_LOG,
81 ME_X1, ME_EPZS, ME_FULL };
82
83 #define FF_MAX_B_FRAMES 4
84
85 /* encoding support */
86 /* note not everything is supported yet */
87
88 #define CODEC_FLAG_HQ 0x0001 /* high quality (non real time) encoding */
89 #define CODEC_FLAG_QSCALE 0x0002 /* use fixed qscale */
90 #define CODEC_FLAG_4MV 0x0004 /* 4 MV per MB allowed */
91 #define CODEC_FLAG_QPEL 0x0010 /* use qpel MC */
92 #define CODEC_FLAG_GMC 0x0020 /* use GMC */
93 #define CODEC_FLAG_TYPE 0x0040 /* fixed I/P frame type, from avctx->key_frame */
94 #define CODEC_FLAG_PART 0x0080 /* use data partitioning */
95 /* parent program gurantees that the input for b-frame containing streams is not written to
96 for at least s->max_b_frames+1 frames, if this is not set than the input will be copied */
97 #define CODEC_FLAG_INPUT_PRESERVED 0x0100
98 #define CODEC_FLAG_PASS1 0x0200 /* use internal 2pass ratecontrol in first pass mode */
99 #define CODEC_FLAG_PASS2 0x0400 /* use internal 2pass ratecontrol in second pass mode */
100 #define CODEC_FLAG_EXTERN_HUFF 0x1000 /* use external huffman table (for mjpeg) */
101 #define CODEC_FLAG_GRAY 0x2000 /* only decode/encode grayscale */
102
103 /* codec capabilities */
104
105 /* decoder can use draw_horiz_band callback */
106 #define CODEC_CAP_DRAW_HORIZ_BAND 0x0001
107
108 #define FRAME_RATE_BASE 10000
109
110 typedef struct AVCodecContext {
111 int bit_rate;
112 int bit_rate_tolerance; /* amount of +- bits (>0)*/
113 int flags;
114 int sub_id; /* some codecs needs additionnal format info. It is
115 stored there */
116
117 int me_method; /* ME algorithm used for video coding */
118
119 /* extra data from parent application to codec, e.g. huffman table
120 for mjpeg */
121 /* the parent should allocate and free this buffer */
122 void *extradata;
123 int extradata_size;
124
125 /* video only */
126 int frame_rate; /* frames per sec multiplied by FRAME_RATE_BASE */
127 int width, height;
128 int aspect_ratio_info;
129 #define FF_ASPECT_SQUARE 1
130 #define FF_ASPECT_4_3_625 2
131 #define FF_ASPECT_4_3_525 3
132 #define FF_ASPECT_16_9_625 4
133 #define FF_ASPECT_16_9_525 5
134 int gop_size; /* 0 = intra only */
135 enum PixelFormat pix_fmt; /* pixel format, see PIX_FMT_xxx */
136 int repeat_pict; /* when decoding, this signal how much the picture */
137 /* must be delayed. */
138 /* extra_delay = (repeat_pict / 2) * (1/fps) */
139 /* if non NULL, 'draw_horiz_band' is called by the libavcodec
140 decoder to draw an horizontal band. It improve cache usage. Not
141 all codecs can do that. You must check the codec capabilities
142 before */
143 void (*draw_horiz_band)(struct AVCodecContext *s,
144 UINT8 **src_ptr, int linesize,
145 int y, int width, int height);
146
147 /* audio only */
148 int sample_rate; /* samples per sec */
149 int channels;
150 int sample_fmt; /* sample format, currenly unused */
151
152 /* the following data should not be initialized */
153 int frame_size; /* in samples, initialized when calling 'init' */
154 int frame_number; /* audio or video frame number */
155 int real_pict_num; /* returns the real picture number of
156 previous encoded frame */
157 int key_frame; /* true if the previous compressed frame was
158 a key frame (intra, or seekable) */
159 int pict_type; /* picture type of the previous
160 encoded frame */
161 /* FIXME: these should have FF_ */
162 #define I_TYPE 1 // Intra
163 #define P_TYPE 2 // Predicted
164 #define B_TYPE 3 // Bi-dir predicted
165 #define S_TYPE 4 // S(GMC)-VOP MPEG4
166
167 int delay; /* number of frames the decoded output
168 will be delayed relative to the encoded input */
169 uint8_t *mbskip_table; /* =1 if MB didnt change, is only valid for I/P frames
170 stride= mb_width = (width+15)>>4 */
171
172 /* encoding parameters */
173 int quality; /* quality of the previous encoded frame
174 (between 1 (good) and 31 (bad))
175 this is allso used to set the quality in vbr mode
176 and the per frame quality in CODEC_FLAG_TYPE (second pass mode) */
177 float qcompress; /* amount of qscale change between easy & hard scenes (0.0-1.0)*/
178 float qblur; /* amount of qscale smoothing over time (0.0-1.0) */
179 int qmin; /* min qscale */
180 int qmax; /* max qscale */
181 int max_qdiff; /* max qscale difference between frames */
182 int max_b_frames; /* maximum b frames, the output will be delayed by max_b_frames+1 relative to the input */
183 float b_quant_factor;/* qscale factor between ips and b frames */
184 int rc_strategy;
185 int b_frame_strategy;
186
187 int hurry_up; /* when set to 1 during decoding, b frames will be skiped
188 when set to 2 idct/dequant will be skipped too */
189
190 struct AVCodec *codec;
191 void *priv_data;
192
193 /* The following data is for RTP friendly coding */
194 /* By now only H.263/H.263+/MPEG4 coder honours this */
195 int rtp_mode; /* 1 for activate RTP friendly-mode */
196 /* highers numbers represent more error-prone */
197 /* enviroments, by now just "1" exist */
198
199 int rtp_payload_size; /* The size of the RTP payload, the coder will */
200 /* do it's best to deliver a chunk with size */
201 /* below rtp_payload_size, the chunk will start */
202 /* with a start code on some codecs like H.263 */
203 /* This doesn't take account of any particular */
204 /* headers inside the transmited RTP payload */
205
206
207 /* The RTP callcack: This function is called */
208 /* every time the encoder as a packet to send */
209 /* Depends on the encoder if the data starts */
210 /* with a Start Code (it should) H.263 does */
211 void (*rtp_callback)(void *data, int size, int packet_number);
212
213 /* These are for PSNR calculation, if you set get_psnr to 1 */
214 /* after encoding you will have the PSNR on psnr_y/cb/cr */
215 int get_psnr;
216 float psnr_y;
217 float psnr_cb;
218 float psnr_cr;
219
220 /* statistics, used for 2-pass encoding */
221 int mv_bits;
222 int header_bits;
223 int i_tex_bits;
224 int p_tex_bits;
225 int i_count;
226 int p_count;
227 int skip_count;
228 int misc_bits; // cbp, mb_type
229 int frame_bits;
230
231 /* the following fields are ignored */
232 void *opaque; /* can be used to carry app specific stuff */
233 char codec_name[32];
234 enum CodecType codec_type; /* see CODEC_TYPE_xxx */
235 enum CodecID codec_id; /* see CODEC_ID_xxx */
236 unsigned int codec_tag; /* codec tag, only used if unknown codec */
237
238 int workaround_bugs; /* workaround bugs in encoders which cannot be detected automatically */
239 int luma_elim_threshold;
240 int chroma_elim_threshold;
241 int strict_std_compliance; /* strictly follow the std (MPEG4, ...) */
242 float b_quant_offset;/* qscale offset between ips and b frames, not implemented yet */
243 int error_resilience;
244
245 #ifndef MBC
246 #define MBC 128
247 #define MBR 96
248 #endif
249 int *quant_store; /* field for communicating with external postprocessing */
250 unsigned qstride;
251 //FIXME this should be reordered after kabis API is finished ...
252 /*
253 Note: Below are located reserved fields for further usage
254 It requires for ABI !!!
255 If you'll perform some changes then borrow new space from these fields
256 (void * can be safety replaced with struct * ;)
257 P L E A S E ! ! !
258 IMPORTANT: Never change order of already declared fields!!!
259 */
260 unsigned long long int
261 ull_res0,ull_res1,ull_res2,ull_res3,ull_res4,ull_res5,
262 ull_res6,ull_res7,ull_res8,ull_res9,ull_res10,ull_res11,ull_res12;
263 float
264 flt_res0,flt_res1,flt_res2,flt_res3,flt_res4,flt_res5,
265 flt_res6,flt_res7,flt_res8,flt_res9,flt_res10,flt_res11;
266 void
267 *ptr_res0,*ptr_res1,*ptr_res2,*ptr_res3,*ptr_res4,*ptr_res5,
268 *ptr_res6,*ptr_res7,*ptr_res8,*ptr_res9,*ptr_res10,*ptr_res11;
269 unsigned long int
270 ul_res0,ul_res1,ul_res2,ul_res3,ul_res4,ul_res5,
271 ul_res6,ul_res7,ul_res8,ul_res9,ul_res10,ul_res11,ul_res12;
272 unsigned int
273 ui_res0,ui_res1,ui_res2,ui_res3,ui_res4,ui_res5,
274 ui_res6;
275 unsigned short int
276 us_res0,us_res1,us_res2,us_res3,us_res4,us_res5,
277 us_res6,us_res7,us_res8,us_res9,us_res10,us_res11,us_res12;
278 unsigned char
279 uc_res0,uc_res1,uc_res2,uc_res3,uc_res4,uc_res5,
280 uc_res6,uc_res7,uc_res8,uc_res9,uc_res10,uc_res11,uc_res12;
281 } AVCodecContext;
282
283 typedef struct AVCodec {
284 char *name;
285 int type;
286 int id;
287 int priv_data_size;
288 int (*init)(AVCodecContext *);
289 int (*encode)(AVCodecContext *, UINT8 *buf, int buf_size, void *data);
290 int (*close)(AVCodecContext *);
291 int (*decode)(AVCodecContext *, void *outdata, int *outdata_size,
292 UINT8 *buf, int buf_size);
293 int capabilities;
294 struct AVCodec *next;
295 /*
296 Note: Below are located reserved fields for further usage
297 It requires for ABI !!!
298 If you'll perform some changes then borrow new space from these fields
299 (void * can be safety replaced with struct * ;)
300 P L E A S E ! ! !
301 IMPORTANT: Never change order of already declared fields!!!
302 */
303 unsigned long long int
304 ull_res0,ull_res1,ull_res2,ull_res3,ull_res4,ull_res5,
305 ull_res6,ull_res7,ull_res8,ull_res9,ull_res10,ull_res11,ull_res12;
306 float
307 flt_res0,flt_res1,flt_res2,flt_res3,flt_res4,flt_res5,
308 flt_res6,flt_res7,flt_res8,flt_res9,flt_res10,flt_res11,flt_res12;
309 void
310 *ptr_res0,*ptr_res1,*ptr_res2,*ptr_res3,*ptr_res4,*ptr_res5,
311 *ptr_res6,*ptr_res7,*ptr_res8,*ptr_res9,*ptr_res10,*ptr_res11,*ptr_res12;
312 } AVCodec;
313
314 /* three components are given, that's all */
315 typedef struct AVPicture {
316 UINT8 *data[3];
317 int linesize[3];
318 } AVPicture;
319
320 extern AVCodec ac3_encoder;
321 extern AVCodec mp2_encoder;
322 extern AVCodec mp3lame_encoder;
323 extern AVCodec mpeg1video_encoder;
324 extern AVCodec h263_encoder;
325 extern AVCodec h263p_encoder;
326 extern AVCodec rv10_encoder;
327 extern AVCodec mjpeg_encoder;
328 extern AVCodec mpeg4_encoder;
329 extern AVCodec msmpeg4v1_encoder;
330 extern AVCodec msmpeg4v2_encoder;
331 extern AVCodec msmpeg4v3_encoder;
332 extern AVCodec wmv1_encoder;
333 extern AVCodec wmv2_encoder;
334
335 extern AVCodec h263_decoder;
336 extern AVCodec mpeg4_decoder;
337 extern AVCodec msmpeg4v1_decoder;
338 extern AVCodec msmpeg4v2_decoder;
339 extern AVCodec msmpeg4v3_decoder;
340 extern AVCodec wmv1_decoder;
341 extern AVCodec wmv2_decoder;
342 extern AVCodec mpeg_decoder;
343 extern AVCodec h263i_decoder;
344 extern AVCodec rv10_decoder;
345 extern AVCodec mjpeg_decoder;
346 extern AVCodec mp2_decoder;
347 extern AVCodec mp3_decoder;
348
349 /* pcm codecs */
350 #define PCM_CODEC(id, name) \
351 extern AVCodec name ## _decoder; \
352 extern AVCodec name ## _encoder;
353
354 PCM_CODEC(CODEC_ID_PCM_S16LE, pcm_s16le);
355 PCM_CODEC(CODEC_ID_PCM_S16BE, pcm_s16be);
356 PCM_CODEC(CODEC_ID_PCM_U16LE, pcm_u16le);
357 PCM_CODEC(CODEC_ID_PCM_U16BE, pcm_u16be);
358 PCM_CODEC(CODEC_ID_PCM_S8, pcm_s8);
359 PCM_CODEC(CODEC_ID_PCM_U8, pcm_u8);
360 PCM_CODEC(CODEC_ID_PCM_ALAW, pcm_alaw);
361 PCM_CODEC(CODEC_ID_PCM_MULAW, pcm_mulaw);
362
363 #undef PCM_CODEC
364
365 /* dummy raw video codec */
366 extern AVCodec rawvideo_codec;
367
368 /* the following codecs use external GPL libs */
369 extern AVCodec ac3_decoder;
370
371 /* resample.c */
372
373 struct ReSampleContext;
374
375 typedef struct ReSampleContext ReSampleContext;
376
377 ReSampleContext *audio_resample_init(int output_channels, int input_channels,
378 int output_rate, int input_rate);
379 int audio_resample(ReSampleContext *s, short *output, short *input, int nb_samples);
380 void audio_resample_close(ReSampleContext *s);
381
382 /* YUV420 format is assumed ! */
383
384 struct ImgReSampleContext;
385
386 typedef struct ImgReSampleContext ImgReSampleContext;
387
388 ImgReSampleContext *img_resample_init(int output_width, int output_height,
389 int input_width, int input_height);
390 void img_resample(ImgReSampleContext *s,
391 AVPicture *output, AVPicture *input);
392
393 void img_resample_close(ImgReSampleContext *s);
394
395 void avpicture_fill(AVPicture *picture, UINT8 *ptr,
396 int pix_fmt, int width, int height);
397 int avpicture_get_size(int pix_fmt, int width, int height);
398
399 /* convert among pixel formats */
400 int img_convert(AVPicture *dst, int dst_pix_fmt,
401 AVPicture *src, int pix_fmt,
402 int width, int height);
403
404 /* deinterlace a picture */
405 int avpicture_deinterlace(AVPicture *dst, AVPicture *src,
406 int pix_fmt, int width, int height);
407
408 /* external high level API */
409
410 extern AVCodec *first_avcodec;
411
412 /* returns LIBAVCODEC_VERSION_INT constant */
413 unsigned avcodec_version(void);
414 /* returns LIBAVCODEC_BUILD constant */
415 unsigned avcodec_build(void);
416 void avcodec_init(void);
417
418 void avcodec_set_bit_exact(void);
419
420 void register_avcodec(AVCodec *format);
421 AVCodec *avcodec_find_encoder(enum CodecID id);
422 AVCodec *avcodec_find_encoder_by_name(const char *name);
423 AVCodec *avcodec_find_decoder(enum CodecID id);
424 AVCodec *avcodec_find_decoder_by_name(const char *name);
425 void avcodec_string(char *buf, int buf_size, AVCodecContext *enc, int encode);
426
427 int avcodec_open(AVCodecContext *avctx, AVCodec *codec);
428 int avcodec_decode_audio(AVCodecContext *avctx, INT16 *samples,
429 int *frame_size_ptr,
430 UINT8 *buf, int buf_size);
431 int avcodec_decode_video(AVCodecContext *avctx, AVPicture *picture,
432 int *got_picture_ptr,
433 UINT8 *buf, int buf_size);
434 int avcodec_encode_audio(AVCodecContext *avctx, UINT8 *buf, int buf_size,
435 const short *samples);
436 int avcodec_encode_video(AVCodecContext *avctx, UINT8 *buf, int buf_size,
437 const AVPicture *pict);
438
439 int avcodec_close(AVCodecContext *avctx);
440
441 void avcodec_register_all(void);
442
443 void avcodec_flush_buffers(AVCodecContext *avctx);
444
445 #ifdef FF_POSTPROCESS
446 extern int quant_store[MBR+1][MBC+1]; // [Review]
447 #endif
448
449
450 /**
451 * Interface for 0.5.0 version
452 *
453 * do not even think about it's usage for this moment
454 */
455
456 typedef struct {
457 // compressed size used from given memory buffer
458 int size;
459 /// I/P/B frame type
460 int frame_type;
461 } avc_enc_result_t;
462
463 /**
464 * Commands
465 * order can't be changed - once it was defined
466 */
467 typedef enum {
468 // general commands
469 AVC_OPEN_BY_NAME = 0xACA000,
470 AVC_OPEN_BY_CODEC_ID,
471 AVC_OPEN_BY_FOURCC,
472 AVC_CLOSE,
473
474 AVC_FLUSH,
475 // pin - struct { uint8_t* src, uint_t src_size }
476 // pout - struct { AVPicture* img, consumed_bytes,
477 AVC_DECODE,
478 // pin - struct { AVPicture* img, uint8_t* dest, uint_t dest_size }
479 // pout - uint_t used_from_dest_size
480 AVC_ENCODE,
481
482 // query/get video commands
483 AVC_GET_VERSION = 0xACB000,
484 AVC_GET_WIDTH,
485 AVC_GET_HEIGHT,
486 AVC_GET_DELAY,
487 AVC_GET_QUANT_TABLE,
488 // ...
489
490 // query/get audio commands
491 AVC_GET_FRAME_SIZE = 0xABC000,
492
493 // maybe define some simple structure which
494 // might be passed to the user - but they can't
495 // contain any codec specific parts and these
496 // calls are usualy necessary only few times
497
498 // set video commands
499 AVC_SET_WIDTH = 0xACD000,
500 AVC_SET_HEIGHT,
501
502 // set video encoding commands
503 AVC_SET_FRAME_RATE = 0xACD800,
504 AVC_SET_QUALITY,
505 AVC_SET_HURRY_UP,
506
507 // set audio commands
508 AVC_SET_SAMPLE_RATE = 0xACE000,
509 AVC_SET_CHANNELS,
510
511 } avc_cmd_t;
512
513 /**
514 * \param handle allocated private structure by libavcodec
515 * for initialization pass NULL - will be returned pout
516 * user is supposed to know nothing about its structure
517 * \param cmd type of operation to be performed
518 * \param pint input parameter
519 * \param pout output parameter
520 *
521 * \returns command status - eventually for query command it might return
522 * integer resulting value
523 */
524 int avcodec(void* handle, avc_cmd_t cmd, void* pin, void* pout);
525
526 /* memory */
527 void *av_malloc(int size);
528 void *av_mallocz(int size);
529 void av_free(void *ptr);
530 void __av_freep(void **ptr);
531 #define av_freep(p) __av_freep((void **)(p))
532
533 #endif /* AVCODEC_H */