new codec: Sorenson v1
[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 4615
9 #define LIBAVCODEC_BUILD_STR "4615"
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 CODEC_ID_SVQ1,
30
31 /* various pcm "codecs" */
32 CODEC_ID_PCM_S16LE,
33 CODEC_ID_PCM_S16BE,
34 CODEC_ID_PCM_U16LE,
35 CODEC_ID_PCM_U16BE,
36 CODEC_ID_PCM_S8,
37 CODEC_ID_PCM_U8,
38 CODEC_ID_PCM_MULAW,
39 CODEC_ID_PCM_ALAW,
40 };
41 #define CODEC_ID_MSMPEG4 CODEC_ID_MSMPEG4V3
42
43 enum CodecType {
44 CODEC_TYPE_UNKNOWN = -1,
45 CODEC_TYPE_VIDEO,
46 CODEC_TYPE_AUDIO,
47 };
48
49 enum PixelFormat {
50 PIX_FMT_ANY = -1,
51 PIX_FMT_YUV420P,
52 PIX_FMT_YUV422,
53 PIX_FMT_RGB24,
54 PIX_FMT_BGR24,
55 PIX_FMT_YUV422P,
56 PIX_FMT_YUV444P,
57 PIX_FMT_YUV410P
58 };
59
60 /* currently unused, may be used if 24/32 bits samples ever supported */
61 enum SampleFormat {
62 SAMPLE_FMT_S16 = 0, /* signed 16 bits */
63 };
64
65 /* in bytes */
66 #define AVCODEC_MAX_AUDIO_FRAME_SIZE 18432
67
68 /* motion estimation type, EPZS by default */
69 enum Motion_Est_ID {
70 ME_ZERO = 1,
71 ME_FULL,
72 ME_LOG,
73 ME_PHODS,
74 ME_EPZS,
75 ME_X1
76 };
77
78 /* only for ME compatiblity with old apps */
79 extern int motion_estimation_method;
80
81 /* ME algos sorted by quality */
82 static const int Motion_Est_QTab[] = { ME_ZERO, ME_PHODS, ME_LOG,
83 ME_X1, ME_EPZS, ME_FULL };
84
85 #define FF_MAX_B_FRAMES 4
86
87 /* encoding support */
88 /* note not everything is supported yet */
89
90 #define CODEC_FLAG_HQ 0x0001 /* high quality (non real time) encoding */
91 #define CODEC_FLAG_QSCALE 0x0002 /* use fixed qscale */
92 #define CODEC_FLAG_4MV 0x0004 /* 4 MV per MB allowed */
93 #define CODEC_FLAG_QPEL 0x0010 /* use qpel MC */
94 #define CODEC_FLAG_GMC 0x0020 /* use GMC */
95 #define CODEC_FLAG_TYPE 0x0040 /* fixed I/P frame type, from avctx->key_frame */
96 #define CODEC_FLAG_PART 0x0080 /* use data partitioning */
97 /* parent program gurantees that the input for b-frame containing streams is not written to
98 for at least s->max_b_frames+1 frames, if this is not set than the input will be copied */
99 #define CODEC_FLAG_INPUT_PRESERVED 0x0100
100 #define CODEC_FLAG_PASS1 0x0200 /* use internal 2pass ratecontrol in first pass mode */
101 #define CODEC_FLAG_PASS2 0x0400 /* use internal 2pass ratecontrol in second pass mode */
102 #define CODEC_FLAG_EXTERN_HUFF 0x1000 /* use external huffman table (for mjpeg) */
103 #define CODEC_FLAG_GRAY 0x2000 /* only decode/encode grayscale */
104
105 /* codec capabilities */
106
107 /* decoder can use draw_horiz_band callback */
108 #define CODEC_CAP_DRAW_HORIZ_BAND 0x0001
109
110 #define FRAME_RATE_BASE 10000
111
112 typedef struct AVCodecContext {
113 int bit_rate;
114 int bit_rate_tolerance; /* amount of +- bits (>0)*/
115 int flags;
116 int sub_id; /* some codecs needs additionnal format info. It is
117 stored there */
118
119 int me_method; /* ME algorithm used for video coding */
120
121 /* extra data from parent application to codec, e.g. huffman table
122 for mjpeg */
123 /* the parent should allocate and free this buffer */
124 void *extradata;
125 int extradata_size;
126
127 /* video only */
128 int frame_rate; /* frames per sec multiplied by FRAME_RATE_BASE */
129 int width, height;
130 int aspect_ratio_info;
131 #define FF_ASPECT_SQUARE 1
132 #define FF_ASPECT_4_3_625 2
133 #define FF_ASPECT_4_3_525 3
134 #define FF_ASPECT_16_9_625 4
135 #define FF_ASPECT_16_9_525 5
136 int gop_size; /* 0 = intra only */
137 enum PixelFormat pix_fmt; /* pixel format, see PIX_FMT_xxx */
138 int repeat_pict; /* when decoding, this signal how much the picture */
139 /* must be delayed. */
140 /* extra_delay = (repeat_pict / 2) * (1/fps) */
141 /* if non NULL, 'draw_horiz_band' is called by the libavcodec
142 decoder to draw an horizontal band. It improve cache usage. Not
143 all codecs can do that. You must check the codec capabilities
144 before */
145 void (*draw_horiz_band)(struct AVCodecContext *s,
146 UINT8 **src_ptr, int linesize,
147 int y, int width, int height);
148
149 /* audio only */
150 int sample_rate; /* samples per sec */
151 int channels;
152 int sample_fmt; /* sample format, currenly unused */
153
154 /* the following data should not be initialized */
155 int frame_size; /* in samples, initialized when calling 'init' */
156 int frame_number; /* audio or video frame number */
157 int real_pict_num; /* returns the real picture number of
158 previous encoded frame */
159 int key_frame; /* true if the previous compressed frame was
160 a key frame (intra, or seekable) */
161 int pict_type; /* picture type of the previous
162 encoded frame */
163 /* FIXME: these should have FF_ */
164 #define I_TYPE 1 // Intra
165 #define P_TYPE 2 // Predicted
166 #define B_TYPE 3 // Bi-dir predicted
167 #define S_TYPE 4 // S(GMC)-VOP MPEG4
168
169 int delay; /* number of frames the decoded output
170 will be delayed relative to the encoded input */
171 uint8_t *mbskip_table; /* =1 if MB didnt change, is only valid for I/P frames
172 stride= mb_width = (width+15)>>4 */
173
174 /* encoding parameters */
175 int quality; /* quality of the previous encoded frame
176 (between 1 (good) and 31 (bad))
177 this is allso used to set the quality in vbr mode
178 and the per frame quality in CODEC_FLAG_TYPE (second pass mode) */
179 float qcompress; /* amount of qscale change between easy & hard scenes (0.0-1.0)*/
180 float qblur; /* amount of qscale smoothing over time (0.0-1.0) */
181 int qmin; /* min qscale */
182 int qmax; /* max qscale */
183 int max_qdiff; /* max qscale difference between frames */
184 int max_b_frames; /* maximum b frames, the output will be delayed by max_b_frames+1 relative to the input */
185 float b_quant_factor;/* qscale factor between ips and b frames */
186 int rc_strategy;
187 int b_frame_strategy;
188
189 int hurry_up; /* when set to 1 during decoding, b frames will be skiped
190 when set to 2 idct/dequant will be skipped too */
191
192 struct AVCodec *codec;
193 void *priv_data;
194
195 /* The following data is for RTP friendly coding */
196 /* By now only H.263/H.263+/MPEG4 coder honours this */
197 int rtp_mode; /* 1 for activate RTP friendly-mode */
198 /* highers numbers represent more error-prone */
199 /* enviroments, by now just "1" exist */
200
201 int rtp_payload_size; /* The size of the RTP payload, the coder will */
202 /* do it's best to deliver a chunk with size */
203 /* below rtp_payload_size, the chunk will start */
204 /* with a start code on some codecs like H.263 */
205 /* This doesn't take account of any particular */
206 /* headers inside the transmited RTP payload */
207
208
209 /* The RTP callcack: This function is called */
210 /* every time the encoder as a packet to send */
211 /* Depends on the encoder if the data starts */
212 /* with a Start Code (it should) H.263 does */
213 void (*rtp_callback)(void *data, int size, int packet_number);
214
215 /* These are for PSNR calculation, if you set get_psnr to 1 */
216 /* after encoding you will have the PSNR on psnr_y/cb/cr */
217 int get_psnr;
218 float psnr_y;
219 float psnr_cb;
220 float psnr_cr;
221
222 /* statistics, used for 2-pass encoding */
223 int mv_bits;
224 int header_bits;
225 int i_tex_bits;
226 int p_tex_bits;
227 int i_count;
228 int p_count;
229 int skip_count;
230 int misc_bits; // cbp, mb_type
231 int frame_bits;
232
233 /* the following fields are ignored */
234 void *opaque; /* can be used to carry app specific stuff */
235 char codec_name[32];
236 enum CodecType codec_type; /* see CODEC_TYPE_xxx */
237 enum CodecID codec_id; /* see CODEC_ID_xxx */
238 unsigned int codec_tag; /* codec tag, only used if unknown codec */
239
240 int workaround_bugs; /* workaround bugs in encoders which cannot be detected automatically */
241 int luma_elim_threshold;
242 int chroma_elim_threshold;
243 int strict_std_compliance; /* strictly follow the std (MPEG4, ...) */
244 float b_quant_offset;/* qscale offset between ips and b frames, not implemented yet */
245 int error_resilience;
246
247 #ifndef MBC
248 #define MBC 128
249 #define MBR 96
250 #endif
251 int *quant_store; /* field for communicating with external postprocessing */
252 unsigned qstride;
253 //FIXME this should be reordered after kabis API is finished ...
254 /*
255 Note: Below are located reserved fields for further usage
256 It requires for ABI !!!
257 If you'll perform some changes then borrow new space from these fields
258 (void * can be safety replaced with struct * ;)
259 P L E A S E ! ! !
260 IMPORTANT: Never change order of already declared fields!!!
261 */
262 unsigned long long int
263 ull_res0,ull_res1,ull_res2,ull_res3,ull_res4,ull_res5,
264 ull_res6,ull_res7,ull_res8,ull_res9,ull_res10,ull_res11,ull_res12;
265 float
266 flt_res0,flt_res1,flt_res2,flt_res3,flt_res4,flt_res5,
267 flt_res6,flt_res7,flt_res8,flt_res9,flt_res10,flt_res11;
268 void
269 *ptr_res0,*ptr_res1,*ptr_res2,*ptr_res3,*ptr_res4,*ptr_res5,
270 *ptr_res6,*ptr_res7,*ptr_res8,*ptr_res9,*ptr_res10,*ptr_res11;
271 unsigned long int
272 ul_res0,ul_res1,ul_res2,ul_res3,ul_res4,ul_res5,
273 ul_res6,ul_res7,ul_res8,ul_res9,ul_res10,ul_res11,ul_res12;
274 unsigned int
275 ui_res0,ui_res1,ui_res2,ui_res3,ui_res4,ui_res5,
276 ui_res6;
277 unsigned short int
278 us_res0,us_res1,us_res2,us_res3,us_res4,us_res5,
279 us_res6,us_res7,us_res8,us_res9,us_res10,us_res11,us_res12;
280 unsigned char
281 uc_res0,uc_res1,uc_res2,uc_res3,uc_res4,uc_res5,
282 uc_res6,uc_res7,uc_res8,uc_res9,uc_res10,uc_res11,uc_res12;
283 } AVCodecContext;
284
285 typedef struct AVCodec {
286 char *name;
287 int type;
288 int id;
289 int priv_data_size;
290 int (*init)(AVCodecContext *);
291 int (*encode)(AVCodecContext *, UINT8 *buf, int buf_size, void *data);
292 int (*close)(AVCodecContext *);
293 int (*decode)(AVCodecContext *, void *outdata, int *outdata_size,
294 UINT8 *buf, int buf_size);
295 int capabilities;
296 struct AVCodec *next;
297 /*
298 Note: Below are located reserved fields for further usage
299 It requires for ABI !!!
300 If you'll perform some changes then borrow new space from these fields
301 (void * can be safety replaced with struct * ;)
302 P L E A S E ! ! !
303 IMPORTANT: Never change order of already declared fields!!!
304 */
305 unsigned long long int
306 ull_res0,ull_res1,ull_res2,ull_res3,ull_res4,ull_res5,
307 ull_res6,ull_res7,ull_res8,ull_res9,ull_res10,ull_res11,ull_res12;
308 float
309 flt_res0,flt_res1,flt_res2,flt_res3,flt_res4,flt_res5,
310 flt_res6,flt_res7,flt_res8,flt_res9,flt_res10,flt_res11,flt_res12;
311 void
312 *ptr_res0,*ptr_res1,*ptr_res2,*ptr_res3,*ptr_res4,*ptr_res5,
313 *ptr_res6,*ptr_res7,*ptr_res8,*ptr_res9,*ptr_res10,*ptr_res11,*ptr_res12;
314 } AVCodec;
315
316 /* three components are given, that's all */
317 typedef struct AVPicture {
318 UINT8 *data[3];
319 int linesize[3];
320 } AVPicture;
321
322 extern AVCodec ac3_encoder;
323 extern AVCodec mp2_encoder;
324 extern AVCodec mp3lame_encoder;
325 extern AVCodec mpeg1video_encoder;
326 extern AVCodec h263_encoder;
327 extern AVCodec h263p_encoder;
328 extern AVCodec rv10_encoder;
329 extern AVCodec mjpeg_encoder;
330 extern AVCodec mpeg4_encoder;
331 extern AVCodec msmpeg4v1_encoder;
332 extern AVCodec msmpeg4v2_encoder;
333 extern AVCodec msmpeg4v3_encoder;
334 extern AVCodec wmv1_encoder;
335 extern AVCodec wmv2_encoder;
336
337 extern AVCodec h263_decoder;
338 extern AVCodec mpeg4_decoder;
339 extern AVCodec msmpeg4v1_decoder;
340 extern AVCodec msmpeg4v2_decoder;
341 extern AVCodec msmpeg4v3_decoder;
342 extern AVCodec wmv1_decoder;
343 extern AVCodec wmv2_decoder;
344 extern AVCodec mpeg_decoder;
345 extern AVCodec h263i_decoder;
346 extern AVCodec rv10_decoder;
347 extern AVCodec svq1_decoder;
348 extern AVCodec mjpeg_decoder;
349 extern AVCodec mp2_decoder;
350 extern AVCodec mp3_decoder;
351
352 /* pcm codecs */
353 #define PCM_CODEC(id, name) \
354 extern AVCodec name ## _decoder; \
355 extern AVCodec name ## _encoder;
356
357 PCM_CODEC(CODEC_ID_PCM_S16LE, pcm_s16le);
358 PCM_CODEC(CODEC_ID_PCM_S16BE, pcm_s16be);
359 PCM_CODEC(CODEC_ID_PCM_U16LE, pcm_u16le);
360 PCM_CODEC(CODEC_ID_PCM_U16BE, pcm_u16be);
361 PCM_CODEC(CODEC_ID_PCM_S8, pcm_s8);
362 PCM_CODEC(CODEC_ID_PCM_U8, pcm_u8);
363 PCM_CODEC(CODEC_ID_PCM_ALAW, pcm_alaw);
364 PCM_CODEC(CODEC_ID_PCM_MULAW, pcm_mulaw);
365
366 #undef PCM_CODEC
367
368 /* dummy raw video codec */
369 extern AVCodec rawvideo_codec;
370
371 /* the following codecs use external GPL libs */
372 extern AVCodec ac3_decoder;
373
374 /* resample.c */
375
376 struct ReSampleContext;
377
378 typedef struct ReSampleContext ReSampleContext;
379
380 ReSampleContext *audio_resample_init(int output_channels, int input_channels,
381 int output_rate, int input_rate);
382 int audio_resample(ReSampleContext *s, short *output, short *input, int nb_samples);
383 void audio_resample_close(ReSampleContext *s);
384
385 /* YUV420 format is assumed ! */
386
387 struct ImgReSampleContext;
388
389 typedef struct ImgReSampleContext ImgReSampleContext;
390
391 ImgReSampleContext *img_resample_init(int output_width, int output_height,
392 int input_width, int input_height);
393 void img_resample(ImgReSampleContext *s,
394 AVPicture *output, AVPicture *input);
395
396 void img_resample_close(ImgReSampleContext *s);
397
398 void avpicture_fill(AVPicture *picture, UINT8 *ptr,
399 int pix_fmt, int width, int height);
400 int avpicture_get_size(int pix_fmt, int width, int height);
401
402 /* convert among pixel formats */
403 int img_convert(AVPicture *dst, int dst_pix_fmt,
404 AVPicture *src, int pix_fmt,
405 int width, int height);
406
407 /* deinterlace a picture */
408 int avpicture_deinterlace(AVPicture *dst, AVPicture *src,
409 int pix_fmt, int width, int height);
410
411 /* external high level API */
412
413 extern AVCodec *first_avcodec;
414
415 /* returns LIBAVCODEC_VERSION_INT constant */
416 unsigned avcodec_version(void);
417 /* returns LIBAVCODEC_BUILD constant */
418 unsigned avcodec_build(void);
419 void avcodec_init(void);
420
421 void avcodec_set_bit_exact(void);
422
423 void register_avcodec(AVCodec *format);
424 AVCodec *avcodec_find_encoder(enum CodecID id);
425 AVCodec *avcodec_find_encoder_by_name(const char *name);
426 AVCodec *avcodec_find_decoder(enum CodecID id);
427 AVCodec *avcodec_find_decoder_by_name(const char *name);
428 void avcodec_string(char *buf, int buf_size, AVCodecContext *enc, int encode);
429
430 int avcodec_open(AVCodecContext *avctx, AVCodec *codec);
431 int avcodec_decode_audio(AVCodecContext *avctx, INT16 *samples,
432 int *frame_size_ptr,
433 UINT8 *buf, int buf_size);
434 int avcodec_decode_video(AVCodecContext *avctx, AVPicture *picture,
435 int *got_picture_ptr,
436 UINT8 *buf, int buf_size);
437 int avcodec_encode_audio(AVCodecContext *avctx, UINT8 *buf, int buf_size,
438 const short *samples);
439 int avcodec_encode_video(AVCodecContext *avctx, UINT8 *buf, int buf_size,
440 const AVPicture *pict);
441
442 int avcodec_close(AVCodecContext *avctx);
443
444 void avcodec_register_all(void);
445
446 void avcodec_flush_buffers(AVCodecContext *avctx);
447
448 #ifdef FF_POSTPROCESS
449 extern int quant_store[MBR+1][MBC+1]; // [Review]
450 #endif
451
452
453 /**
454 * Interface for 0.5.0 version
455 *
456 * do not even think about it's usage for this moment
457 */
458
459 typedef struct {
460 // compressed size used from given memory buffer
461 int size;
462 /// I/P/B frame type
463 int frame_type;
464 } avc_enc_result_t;
465
466 /**
467 * Commands
468 * order can't be changed - once it was defined
469 */
470 typedef enum {
471 // general commands
472 AVC_OPEN_BY_NAME = 0xACA000,
473 AVC_OPEN_BY_CODEC_ID,
474 AVC_OPEN_BY_FOURCC,
475 AVC_CLOSE,
476
477 AVC_FLUSH,
478 // pin - struct { uint8_t* src, uint_t src_size }
479 // pout - struct { AVPicture* img, consumed_bytes,
480 AVC_DECODE,
481 // pin - struct { AVPicture* img, uint8_t* dest, uint_t dest_size }
482 // pout - uint_t used_from_dest_size
483 AVC_ENCODE,
484
485 // query/get video commands
486 AVC_GET_VERSION = 0xACB000,
487 AVC_GET_WIDTH,
488 AVC_GET_HEIGHT,
489 AVC_GET_DELAY,
490 AVC_GET_QUANT_TABLE,
491 // ...
492
493 // query/get audio commands
494 AVC_GET_FRAME_SIZE = 0xABC000,
495
496 // maybe define some simple structure which
497 // might be passed to the user - but they can't
498 // contain any codec specific parts and these
499 // calls are usualy necessary only few times
500
501 // set video commands
502 AVC_SET_WIDTH = 0xACD000,
503 AVC_SET_HEIGHT,
504
505 // set video encoding commands
506 AVC_SET_FRAME_RATE = 0xACD800,
507 AVC_SET_QUALITY,
508 AVC_SET_HURRY_UP,
509
510 // set audio commands
511 AVC_SET_SAMPLE_RATE = 0xACE000,
512 AVC_SET_CHANNELS,
513
514 } avc_cmd_t;
515
516 /**
517 * \param handle allocated private structure by libavcodec
518 * for initialization pass NULL - will be returned pout
519 * user is supposed to know nothing about its structure
520 * \param cmd type of operation to be performed
521 * \param pint input parameter
522 * \param pout output parameter
523 *
524 * \returns command status - eventually for query command it might return
525 * integer resulting value
526 */
527 int avcodec(void* handle, avc_cmd_t cmd, void* pin, void* pout);
528
529 /* memory */
530 void *av_malloc(int size);
531 void *av_mallocz(int size);
532 void av_free(void *ptr);
533 void __av_freep(void **ptr);
534 #define av_freep(p) __av_freep((void **)(p))
535
536 #endif /* AVCODEC_H */