avcodec.h

最全面的音频/视频的编码和解码库

#ifndef AVCODEC_H
#define AVCODEC_H

/**
 * @file avcodec.h
 * external api header.
 */


#ifdef __cplusplus
extern "C" {
#endif

#include "avutil.h"
#include <sys/types.h> /* size_t */

//FIXME the following 2 really dont belong in here
#define FFMPEG_VERSION_INT      0x000409
#define FFMPEG_VERSION          "CVS"

#define AV_STRINGIFY(s)         AV_TOSTRING(s)
#define AV_TOSTRING(s) #s

#define LIBAVCODEC_VERSION_INT  ((51<<16)+(9<<8)+0)
#define LIBAVCODEC_VERSION      51.9.0
#define LIBAVCODEC_BUILD        LIBAVCODEC_VERSION_INT

#define LIBAVCODEC_IDENT        "Lavc" AV_STRINGIFY(LIBAVCODEC_VERSION)

#define AV_NOPTS_VALUE          int64_t_C(0x8000000000000000)
#define AV_TIME_BASE            1000000
#define AV_TIME_BASE_Q          (AVRational){1, AV_TIME_BASE}

enum CodecID {
    CODEC_ID_NONE,
    CODEC_ID_MPEG1VIDEO,
    CODEC_ID_MPEG2VIDEO, /* prefered ID for MPEG Video 1 or 2 decoding */
    CODEC_ID_MPEG2VIDEO_XVMC,
    CODEC_ID_H261,
    CODEC_ID_H263,
    CODEC_ID_RV10,
    CODEC_ID_RV20,
    CODEC_ID_MJPEG,
    CODEC_ID_MJPEGB,
    CODEC_ID_LJPEG,
    CODEC_ID_SP5X,
    CODEC_ID_JPEGLS,
    CODEC_ID_MPEG4,
    CODEC_ID_RAWVIDEO,
    CODEC_ID_MSMPEG4V1,
    CODEC_ID_MSMPEG4V2,
    CODEC_ID_MSMPEG4V3,
    CODEC_ID_WMV1,
    CODEC_ID_WMV2,
    CODEC_ID_H263P,
    CODEC_ID_H263I,
    CODEC_ID_FLV1,
    CODEC_ID_SVQ1,
    CODEC_ID_SVQ3,
    CODEC_ID_DVVIDEO,
    CODEC_ID_HUFFYUV,
    CODEC_ID_CYUV,
    CODEC_ID_H264,
    CODEC_ID_INDEO3,
    CODEC_ID_VP3,
    CODEC_ID_THEORA,
    CODEC_ID_ASV1,
    CODEC_ID_ASV2,
    CODEC_ID_FFV1,
    CODEC_ID_4XM,
    CODEC_ID_VCR1,
    CODEC_ID_CLJR,
    CODEC_ID_MDEC,
    CODEC_ID_ROQ,
    CODEC_ID_INTERPLAY_VIDEO,
    CODEC_ID_XAN_WC3,
    CODEC_ID_XAN_WC4,
    CODEC_ID_RPZA,
    CODEC_ID_CINEPAK,
    CODEC_ID_WS_VQA,
    CODEC_ID_MSRLE,
    CODEC_ID_MSVIDEO1,
    CODEC_ID_IDCIN,
    CODEC_ID_8BPS,
    CODEC_ID_SMC,
    CODEC_ID_FLIC,
    CODEC_ID_TRUEMOTION1,
    CODEC_ID_VMDVIDEO,
    CODEC_ID_MSZH,
    CODEC_ID_ZLIB,
    CODEC_ID_QTRLE,
    CODEC_ID_SNOW,
    CODEC_ID_TSCC,
    CODEC_ID_ULTI,
    CODEC_ID_QDRAW,
    CODEC_ID_VIXL,
    CODEC_ID_QPEG,
    CODEC_ID_XVID,
    CODEC_ID_PNG,
    CODEC_ID_PPM,
    CODEC_ID_PBM,
    CODEC_ID_PGM,
    CODEC_ID_PGMYUV,
    CODEC_ID_PAM,
    CODEC_ID_FFVHUFF,
    CODEC_ID_RV30,
    CODEC_ID_RV40,
    CODEC_ID_VC9,
    CODEC_ID_WMV3,
    CODEC_ID_LOCO,
    CODEC_ID_WNV1,
    CODEC_ID_AASC,
    CODEC_ID_INDEO2,
    CODEC_ID_FRAPS,
    CODEC_ID_TRUEMOTION2,
    CODEC_ID_BMP,
    CODEC_ID_CSCD,
    CODEC_ID_MMVIDEO,
    CODEC_ID_ZMBV,
    CODEC_ID_AVS,
    CODEC_ID_SMACKVIDEO,
    CODEC_ID_NUV,
    CODEC_ID_KMVC,

    /* various pcm "codecs" */
    CODEC_ID_PCM_S16LE= 0x10000,
    CODEC_ID_PCM_S16BE,
    CODEC_ID_PCM_U16LE,
    CODEC_ID_PCM_U16BE,
    CODEC_ID_PCM_S8,
    CODEC_ID_PCM_U8,
    CODEC_ID_PCM_MULAW,
    CODEC_ID_PCM_ALAW,
    CODEC_ID_PCM_S32LE,
    CODEC_ID_PCM_S32BE,
    CODEC_ID_PCM_U32LE,
    CODEC_ID_PCM_U32BE,
    CODEC_ID_PCM_S24LE,
    CODEC_ID_PCM_S24BE,
    CODEC_ID_PCM_U24LE,
    CODEC_ID_PCM_U24BE,
    CODEC_ID_PCM_S24DAUD,

    /* various adpcm codecs */
    CODEC_ID_ADPCM_IMA_QT= 0x11000,
    CODEC_ID_ADPCM_IMA_WAV,
    CODEC_ID_ADPCM_IMA_DK3,
    CODEC_ID_ADPCM_IMA_DK4,
    CODEC_ID_ADPCM_IMA_WS,
    CODEC_ID_ADPCM_IMA_SMJPEG,
    CODEC_ID_ADPCM_MS,
    CODEC_ID_ADPCM_4XM,
    CODEC_ID_ADPCM_XA,
    CODEC_ID_ADPCM_ADX,
    CODEC_ID_ADPCM_EA,
    CODEC_ID_ADPCM_G726,
    CODEC_ID_ADPCM_CT,
    CODEC_ID_ADPCM_SWF,
    CODEC_ID_ADPCM_YAMAHA,
    CODEC_ID_ADPCM_SBPRO_4,
    CODEC_ID_ADPCM_SBPRO_3,
    CODEC_ID_ADPCM_SBPRO_2,

    /* AMR */
    CODEC_ID_AMR_NB= 0x12000,
    CODEC_ID_AMR_WB,

    /* RealAudio codecs*/
    CODEC_ID_RA_144= 0x13000,
    CODEC_ID_RA_288,

    /* various DPCM codecs */
    CODEC_ID_ROQ_DPCM= 0x14000,
    CODEC_ID_INTERPLAY_DPCM,
    CODEC_ID_XAN_DPCM,
    CODEC_ID_SOL_DPCM,

    CODEC_ID_MP2= 0x15000,
    CODEC_ID_MP3, /* prefered ID for MPEG Audio layer 1, 2 or3 decoding */
    CODEC_ID_AAC,
    CODEC_ID_MPEG4AAC,
    CODEC_ID_AC3,
    CODEC_ID_DTS,
    CODEC_ID_VORBIS,
    CODEC_ID_DVAUDIO,
    CODEC_ID_WMAV1,
    CODEC_ID_WMAV2,
    CODEC_ID_MACE3,
    CODEC_ID_MACE6,
    CODEC_ID_VMDAUDIO,
    CODEC_ID_SONIC,
    CODEC_ID_SONIC_LS,
    CODEC_ID_FLAC,
    CODEC_ID_MP3ADU,
    CODEC_ID_MP3ON4,
    CODEC_ID_SHORTEN,
    CODEC_ID_ALAC,
    CODEC_ID_WESTWOOD_SND1,
    CODEC_ID_GSM,
    CODEC_ID_QDM2,
    CODEC_ID_COOK,
    CODEC_ID_TRUESPEECH,
    CODEC_ID_TTA,
    CODEC_ID_SMACKAUDIO,

    CODEC_ID_OGGTHEORA= 0x16000,

    /* subtitle codecs */
    CODEC_ID_DVD_SUBTITLE= 0x17000,
    CODEC_ID_DVB_SUBTITLE,

    CODEC_ID_MPEG2TS= 0x20000, /* _FAKE_ codec to indicate a raw MPEG2 transport
                         stream (only used by libavformat) */
};

/* CODEC_ID_MP3LAME is absolete */
#define CODEC_ID_MP3LAME CODEC_ID_MP3

enum CodecType {
    CODEC_TYPE_UNKNOWN = -1,
    CODEC_TYPE_VIDEO,
    CODEC_TYPE_AUDIO,
    CODEC_TYPE_DATA,
    CODEC_TYPE_SUBTITLE,
};

/* currently unused, may be used if 24/32 bits samples ever supported */
/* all in native endian */
enum SampleFormat {
    SAMPLE_FMT_NONE = -1,
    SAMPLE_FMT_U8,              ///< unsigned 8 bits
    SAMPLE_FMT_S16,             ///< signed 16 bits
    SAMPLE_FMT_S24,             ///< signed 24 bits
    SAMPLE_FMT_S32,             ///< signed 32 bits
    SAMPLE_FMT_FLT,             ///< float
};

/* in bytes */
#define AVCODEC_MAX_AUDIO_FRAME_SIZE 192000 // 1 second of 48khz 32bit audio

/**
 * Required number of additionally allocated bytes at the end of the input bitstream for decoding.
 * this is mainly needed because some optimized bitstream readers read
 * 32 or 64 bit at once and could read over the end<br>
 * Note, if the first 23 bits of the additional bytes are not 0 then damaged
 * MPEG bitstreams could cause overread and segfault
 */
#define FF_INPUT_BUFFER_PADDING_SIZE 8

/**
 * minimum encoding buffer size.
 * used to avoid some checks during header writing
 */
#define FF_MIN_BUFFER_SIZE 16384

/* motion estimation type, EPZS by default */
enum Motion_Est_ID {
    ME_ZERO = 1,
    ME_FULL,
    ME_LOG,
    ME_PHODS,
    ME_EPZS,
    ME_X1,
    ME_HEX,
    ME_UMH,
    ME_ITER,
};

enum AVDiscard{
//we leave some space between them for extensions (drop some keyframes for intra only or drop just some bidir frames)
    AVDISCARD_NONE   =-16, ///< discard nothing
    AVDISCARD_DEFAULT=  0, ///< discard useless packets like 0 size packets in avi
    AVDISCARD_NONREF =  8, ///< discard all non reference
    AVDISCARD_BIDIR  = 16, ///< discard all bidirectional frames
    AVDISCARD_NONKEY = 32, ///< discard all frames except keyframes
    AVDISCARD_ALL    = 48, ///< discard all
};

typedef struct RcOverride{
    int start_frame;
    int end_frame;
    int qscale; // if this is 0 then quality_factor will be used instead
    float quality_factor;
} RcOverride;

/* only for ME compatiblity with old apps */
extern int motion_estimation_method;

#define FF_MAX_B_FRAMES 16

/* encoding support
   these flags can be passed in AVCodecContext.flags before initing
   Note: not everything is supported yet.
*/

#define CODEC_FLAG_QSCALE 0x0002  ///< use fixed qscale
#define CODEC_FLAG_4MV    0x0004  ///< 4 MV per MB allowed / Advanced prediction for H263
#define CODEC_FLAG_QPEL   0x0010  ///< use qpel MC
#define CODEC_FLAG_GMC    0x0020  ///< use GMC
#define CODEC_FLAG_MV0    0x0040  ///< always try a MB with MV=<0,0>
#define CODEC_FLAG_PART   0x0080  ///< use data partitioning
/* parent program gurantees that the input for b-frame containing streams is not written to
   for at least s->max_b_frames+1 frames, if this is not set than the input will be copied */
#define CODEC_FLAG_INPUT_PRESERVED 0x0100
#define CODEC_FLAG_PASS1 0x0200   ///< use internal 2pass ratecontrol in first  pass mode
#define CODEC_FLAG_PASS2 0x0400   ///< use internal 2pass ratecontrol in second pass mode
#define CODEC_FLAG_EXTERN_HUFF 0x1000 ///< use external huffman table (for mjpeg)
#define CODEC_FLAG_GRAY  0x2000   ///< only decode/encode grayscale
#define CODEC_FLAG_EMU_EDGE 0x4000///< don't draw edges
#define CODEC_FLAG_PSNR           0x8000 ///< error[?] variables will be set during encoding
#define CODEC_FLAG_TRUNCATED  0x00010000 /** input bitstream might be truncated at a random location instead
                                            of only at frame boundaries */
#define CODEC_FLAG_NORMALIZE_AQP  0x00020000 ///< normalize adaptive quantization
#define CODEC_FLAG_INTERLACED_DCT 0x00040000 ///< use interlaced dct
#define CODEC_FLAG_LOW_DELAY      0x00080000 ///< force low delay
#define CODEC_FLAG_ALT_SCAN       0x00100000 ///< use alternate scan
#define CODEC_FLAG_TRELLIS_QUANT  0x00200000 ///< use trellis quantization
#define CODEC_FLAG_GLOBAL_HEADER  0x00400000 ///< place global headers in extradata instead of every keyframe
#define CODEC_FLAG_BITEXACT       0x00800000 ///< use only bitexact stuff (except (i)dct)
/* Fx : Flag for h263+ extra options */
#define CODEC_FLAG_H263P_AIC      0x01000000 ///< H263 Advanced intra coding / MPEG4 AC prediction (remove this)
#define CODEC_FLAG_AC_PRED        0x01000000 ///< H263 Advanced intra coding / MPEG4 AC prediction
#define CODEC_FLAG_H263P_UMV      0x02000000 ///< Unlimited motion vector
#define CODEC_FLAG_CBP_RD         0x04000000 ///< use rate distortion optimization for cbp
#define CODEC_FLAG_QP_RD          0x08000000 ///< use rate distortion optimization for qp selectioon
#define CODEC_FLAG_H263P_AIV      0x00000008 ///< H263 Alternative inter vlc
#define CODEC_FLAG_OBMC           0x00000001 ///< OBMC
#define CODEC_FLAG_LOOP_FILTER    0x00000800 ///< loop filter
#define CODEC_FLAG_H263P_SLICE_STRUCT 0x10000000
#define CODEC_FLAG_INTERLACED_ME  0x20000000 ///< interlaced motion estimation
#define CODEC_FLAG_SVCD_SCAN_OFFSET 0x40000000 ///< will reserve space for SVCD scan offset user data
#define CODEC_FLAG_CLOSED_GOP     0x80000000
#define CODEC_FLAG2_FAST          0x00000001 ///< allow non spec compliant speedup tricks
#define CODEC_FLAG2_STRICT_GOP    0x00000002 ///< strictly enforce GOP size
#define CODEC_FLAG2_NO_OUTPUT     0x00000004 ///< skip bitstream encoding
#define CODEC_FLAG2_LOCAL_HEADER  0x00000008 ///< place global headers at every keyframe instead of in extradata
#define CODEC_FLAG2_BPYRAMID      0x00000010 ///< H.264 allow b-frames to be used as references
#define CODEC_FLAG2_WPRED         0x00000020 ///< H.264 weighted biprediction for b-frames
#define CODEC_FLAG2_MIXED_REFS    0x00000040 ///< H.264 multiple references per partition
#define CODEC_FLAG2_8X8DCT        0x00000080 ///< H.264 high profile 8x8 transform
#define CODEC_FLAG2_FASTPSKIP     0x00000100 ///< H.264 fast pskip
#define CODEC_FLAG2_AUD           0x00000200 ///< H.264 access unit delimiters
#define CODEC_FLAG2_BRDO          0x00000400 ///< b-frame rate-distortion optimization

/* Unsupported options :
 *              Syntax Arithmetic coding (SAC)
 *              Reference Picture Selection
 *              Independant Segment Decoding */
/* /Fx */
/* codec capabilities */

#define CODEC_CAP_DRAW_HORIZ_BAND 0x0001 ///< decoder can use draw_horiz_band callback
/**
 * Codec uses get_buffer() for allocating buffers.
 * direct rendering method 1
 */
#define CODEC_CAP_DR1             0x0002
/* if 'parse_only' field is true, then avcodec_parse_frame() can be
   used */
#define CODEC_CAP_PARSE_ONLY      0x0004
#define CODEC_CAP_TRUNCATED       0x0008
/* codec can export data for HW decoding (XvMC) */
#define CODEC_CAP_HWACCEL         0x0010
/**
 * codec has a non zero delay and needs to be feeded with NULL at the end to get the delayed data.
 * if this is not set, the codec is guranteed to never be feeded with NULL data
 */
#define CODEC_CAP_DELAY           0x0020

//the following defines may change, don't expect compatibility if you use them
#define MB_TYPE_INTRA4x4   0x0001
#define MB_TYPE_INTRA16x16 0x0002 //FIXME h264 specific
#define MB_TYPE_INTRA_PCM  0x0004 //FIXME h264 specific
#define MB_TYPE_16x16      0x0008
#define MB_TYPE_16x8       0x0010
#define MB_TYPE_8x16       0x0020
#define MB_TYPE_8x8        0x0040
#define MB_TYPE_INTERLACED 0x0080
#define MB_TYPE_DIRECT2     0x0100 //FIXME
#define MB_TYPE_ACPRED     0x0200
#define MB_TYPE_GMC        0x0400
#define MB_TYPE_SKIP       0x0800
#define MB_TYPE_P0L0       0x1000
#define MB_TYPE_P1L0       0x2000
#define MB_TYPE_P0L1       0x4000
#define MB_TYPE_P1L1       0x8000
#define MB_TYPE_L0         (MB_TYPE_P0L0 | MB_TYPE_P1L0)
#define MB_TYPE_L1         (MB_TYPE_P0L1 | MB_TYPE_P1L1)
#define MB_TYPE_L0L1       (MB_TYPE_L0   | MB_TYPE_L1)
#define MB_TYPE_QUANT      0x00010000
#define MB_TYPE_CBP        0x00020000
//Note bits 24-31 are reserved for codec specific use (h264 ref0, mpeg1 0mv, ...)

/**
 * Pan Scan area.
 * this specifies the area which should be displayed. Note there may be multiple such areas for one frame
 */
typedef struct AVPanScan{
    /**
     * id.
     * - encoding: set by user.
     * - decoding: set by lavc
     */
    int id;

    /**
     * width and height in 1/16 pel
     * - encoding: set by user.
     * - decoding: set by lavc
     */
    int width;
    int height;

    /**
     * position of the top left corner in 1/16 pel for up to 3 fields/frames.
     * - encoding: set by user.
     * - decoding: set by lavc
     */
    int16_t position[3][2];
}AVPanScan;

#define FF_COMMON_FRAME \
    /**\
     * pointer to the picture planes.\
     * this might be different from the first allocated byte\
     * - encoding: \
     * - decoding: \
     */\
    uint8_t *data[4];\
    int linesize[4];\
    /**\
     * pointer to the first allocated byte of the picture. can be used in get_buffer/release_buffer\
     * this isn't used by lavc unless the default get/release_buffer() is used\
     * - encoding: \
     * - decoding: \
     */\
    uint8_t *base[4];\
    /**\
     * 1 -> keyframe, 0-> not\
     * - encoding: set by lavc\
     * - decoding: set by lavc\
     */\
    int key_frame;\
\
    /**\
     * picture type of the frame, see ?_TYPE below.\
     * - encoding: set by lavc for coded_picture (and set by user for input)\
     * - decoding: set by lavc\
     */\