FFMPEG结构体分析 AVFormatContext

分享一下我老师大神的人工智能教程！零基础，通俗易懂！http://blog.csdn.net/jiangjunshow

也欢迎大家转载本篇文章。分享知识，造福人民，实现我们中华民族伟大复兴！

               

注：写了一系列的结构体的分析的文章，在这里列一个列表：

FFMPEG结构体分析：AVFrame

FFMPEG结构体分析：AVFormatContext

FFMPEG结构体分析：AVCodecContext

FFMPEG结构体分析：AVIOContext

FFMPEG结构体分析：AVCodec

FFMPEG结构体分析：AVStream
FFMPEG结构体分析：AVPacket

FFMPEG有几个最重要的结构体，包含了解协议，解封装，解码操作，此前已经进行过分析：

FFMPEG中最关键的结构体之间的关系

在此不再详述，其中AVFormatContext是包含码流参数较多的结构体。本文将会详细分析一下该结构体里每个变量的含义和作用。

首先看一下结构体的定义（位于avformat.h）：

/* 雷霄骅 * 中国传媒大学/数字电视技术 * leixiaohua1020@126.com * *//** * Format I/O context. * New fields can be added to the end with minor version bumps. * Removal, reordering and changes to existing fields require a major * version bump. * sizeof(AVFormatContext) must not be used outside libav*, use * avformat_alloc_context() to create an AVFormatContext. */typedef struct AVFormatContext {    /**     * A class for logging and AVOptions. Set by avformat_alloc_context().     * Exports (de)muxer private options if they exist.     */    const AVClass *av_class;    /**     * Can only be iformat or oformat, not both at the same time.     *     * decoding: set by avformat_open_input().     * encoding: set by the user.     */    struct AVInputFormat *iformat;    struct AVOutputFormat *oformat;    /**     * Format private data. This is an AVOptions-enabled struct     * if and only if iformat/oformat.priv_class is not NULL.     */    void *priv_data;    /*     * I/O context.     *     * decoding: either set by the user before avformat_open_input() (then     * the user must close it manually) or set by avformat_open_input().     * encoding: set by the user.     *     * Do NOT set this field if AVFMT_NOFILE flag is set in     * iformat/oformat.flags. In such a case, the (de)muxer will handle     * I/O in some other way and this field will be NULL.     */    AVIOContext *pb;    /* stream info */    int ctx_flags; /**< Format-specific flags, see AVFMTCTX_xx */    /**     * A list of all streams in the file. New streams are created with     * avformat_new_stream().     *     * decoding: streams are created by libavformat in avformat_open_input().     * If AVFMTCTX_NOHEADER is set in ctx_flags, then new streams may also     * appear in av_read_frame().     * encoding: streams are created by the user before avformat_write_header().     */    unsigned int nb_streams;    AVStream **streams;    char filename[1024]; /**< input or output filename */    /**     * Decoding: position of the first frame of the component, in     * AV_TIME_BASE fractional seconds. NEVER set this value directly:     * It is deduced from the AVStream values.     */    int64_t start_time;    /**     * Decoding: duration of the stream, in AV_TIME_BASE fractional     * seconds. Only set this value if you know none of the individual stream     * durations and also do not set any of them. This is deduced from the     * AVStream values if not set.     */    int64_t duration;    /**     * Decoding: total stream bitrate in bit/s, 0 if not     * available. Never set it directly if the file_size and the     * duration are known as FFmpeg can compute it automatically.     */    int bit_rate;    unsigned int packet_size;    int max_delay;    int flags;#define AVFMT_FLAG_GENPTS       0x0001 ///< Generate missing pts even if it requires parsing future frames.#define AVFMT_FLAG_IGNIDX       0x0002 ///< Ignore index.#define AVFMT_FLAG_NONBLOCK     0x0004 ///< Do not block when reading packets from input.#define AVFMT_FLAG_IGNDTS       0x0008 ///< Ignore DTS on frames that contain both DTS & PTS#define AVFMT_FLAG_NOFILLIN     0x0010 ///< Do not infer any values from other values, just return what is stored in the container#define AVFMT_FLAG_NOPARSE      0x0020 ///< Do not use AVParsers, you also must set AVFMT_FLAG_NOFILLIN as the fillin code works on frames and no parsing -> no frames. Also seeking to frames can not work if parsing to find frame boundaries has been disabled#define AVFMT_FLAG_CUSTOM_IO    0x0080 ///< The caller has supplied a custom AVIOContext, don't avio_close() it.#define AVFMT_FLAG_DISCARD_CORRUPT  0x0100 ///< Discard frames marked corrupted#define AVFMT_FLAG_MP4A_LATM    0x8000 ///< Enable RTP MP4A-LATM payload#define AVFMT_FLAG_SORT_DTS    0x10000 ///< try to interleave outputted packets by dts (using this flag can slow demuxing down)#define AVFMT_FLAG_PRIV_OPT    0x20000 ///< Enable use of private options by delaying codec open (this could be made default once all code is converted)#define AVFMT_FLAG_KEEP_SIDE_DATA 0x40000 ///< Dont merge side data but keep it separate.    /**     * decoding: size of data to probe; encoding: unused.     */    unsigned int probesize;    /**     * decoding: maximum time (in AV_TIME_BASE units) during which the input should     * be analyzed in avformat_find_stream_info().     */    int max_analyze_duration;    const uint8_t *key;    int keylen;    unsigned int nb_programs;    AVProgram **programs;    /**     * Forced video codec_id.     * Demuxing: Set by user.     */    enum CodecID video_codec_id;    /**     * Forced audio codec_id.     * Demuxing: Set by user.     */    enum CodecID audio_codec_id;    /**     * Forced subtitle codec_id.     * Demuxing: Set by user.     */    enum CodecID subtitle_codec_id;    /**     * Maximum amount of memory in bytes to use for the index of each stream.     * If the index exceeds this size, entries will be discarded as     * needed to maintain a smaller size. This can lead to slower or less     * accurate seeking (depends on demuxer).     * Demuxers for which a full in-memory index is mandatory will ignore     * this.     * muxing  : unused     * demuxing: set by user     */    unsigned int max_index_size;    /**     * Maximum amount of memory in bytes to use for buffering frames     * obtained from realtime capture devices.     */    unsigned int max_picture_buffer;    unsigned int nb_chapters;    AVChapter **chapters;    AVDictionary *metadata;    /**     * Start time of the stream in real world time, in microseconds     * since the unix epoch (00:00 1st January 1970). That is, pts=0     * in the stream was captured at this real world time.     * - encoding: Set by user.     * - decoding: Unused.     */    int64_t start_time_realtime;    /**     * decoding: number of frames used to probe fps     */    int fps_probe_size;    /**     * Error recognition; higher values will detect more errors but may     * misdetect some more or less valid parts as errors.     * - encoding: unused     * - decoding: Set by user.     */    int error_recognition;    /**     * Custom interrupt callbacks for the I/O layer.     *     * decoding: set by the user before avformat_open_input().     * encoding: set by the user before avformat_write_header()     * (mainly useful for AVFMT_NOFILE formats). The callback     * should also be passed to avio_open2() if it's used to     * open the file.     */    AVIOInterruptCB interrupt_callback;    /**     * Flags to enable debugging.     */    int debug;#define FF_FDEBUG_TS        0x0001    /**     * Transport stream id.     * This will be moved into demuxer private options. Thus no API/ABI compatibility     */    int ts_id;    /**     * Audio preload in microseconds.     * Note, not all formats support this and unpredictable things may happen if it is used when not supported.     * - encoding: Set by user via AVOptions (NO direct access)     * - decoding: unused     */    int audio_preload;    /**     * Max chunk time in microseconds.     * Note, not all formats support this and unpredictable things may happen if it is used when not supported.     * - encoding: Set by user via AVOptions (NO direct access)     * - decoding: unused     */    int max_chunk_duration;    /**     * Max chunk size in bytes     * Note, not all formats support this and unpredictable things may happen if it is used when not supported.     * - encoding: Set by user via AVOptions (NO direct access)     * - decoding: unused     */    int max_chunk_size;    /*****************************************************************     * All fields below this line are not part of the public API. They     * may not be used outside of libavformat and can be changed and     * removed at will.     * New public fields should be added right above.     *****************************************************************     */    /**     * This buffer is only needed when packets were already buffered but     * not decoded, for example to get the codec parameters in MPEG     * streams.     */    struct AVPacketList *packet_buffer;    struct AVPacketList *packet_buffer_end;    /* av_seek_frame() support */    int64_t data_offset; /**< offset of the first packet */    /**     * Raw packets from the demuxer, prior to parsing and decoding.     * This buffer is used for buffering packets until the codec can     * be identified, as parsing cannot be done without knowing the     * codec.     */    struct AVPacketList *raw_packet_buffer;    struct AVPacketList *raw_packet_buffer_end;    /**     * Packets split by the parser get queued here.     */    struct AVPacketList *parse_queue;    struct AVPacketList *parse_queue_end;    /**     * Remaining size available for raw_packet_buffer, in bytes.     */#define RAW_PACKET_BUFFER_SIZE 2500000    int raw_packet_buffer_remaining_size;} AVFormatContext;

在使用FFMPEG进行开发的时候，AVFormatContext是一个贯穿始终的数据结构，很多函数都要用到它作为参数。它是FFMPEG解封装（flv，mp4，rmvb，avi）功能的结构体。下面看几个主要变量的作用（在这里考虑解码的情况）：

struct AVInputFormat *iformat：输入数据的封装格式

AVIOContext *pb：输入数据的缓存

unsigned int nb_streams：视音频流的个数

AVStream **streams：视音频流

char filename[1024]：文件名

int64_t duration：时长（单位：微秒us，转换为秒需要除以1000000）

int bit_rate：比特率（单位bps，转换为kbps需要除以1000）

AVDictionary *metadata：元数据

视频的时长可以转换成HH:MM:SS的形式，示例代码如下：

AVFormatContext *pFormatCtx;CString timelong;...//duration是以微秒为单位//转换成hh:mm:ss形式int tns, thh, tmm, tss;tns  = (pFormatCtx->duration)/1000000;thh  = tns / 3600;tmm  = (tns % 3600) / 60;tss  = (tns % 60);timelong.Format("%02d:%02d:%02d",thh,tmm,tss);

视频的原数据（metadata）信息可以通过AVDictionary获取。元数据存储在AVDictionaryEntry结构体中，如下所示

typedef struct AVDictionaryEntry {    char *key;    char *value;} AVDictionaryEntry;

每一条元数据分为key和value两个属性。

在ffmpeg中通过av_dict_get()函数获得视频的原数据。

下列代码显示了获取元数据并存入meta字符串变量的过程，注意每一条key和value之间有一个"\t:"，value之后有一个"\r\n"

//MetaData------------------------------------------------------------//从AVDictionary获得//需要用到AVDictionaryEntry对象//CString author,copyright,description;CString meta=NULL,key,value;AVDictionaryEntry *m = NULL;//不用一个一个找出来/* m=av_dict_get(pFormatCtx->metadata,"author",m,0);author.Format("作者：%s",m->value);m=av_dict_get(pFormatCtx->metadata,"copyright",m,0);copyright.Format("版权：%s",m->value);m=av_dict_get(pFormatCtx->metadata,"description",m,0);description.Format("描述：%s",m->value);*///使用循环读出//(需要读取的数据，字段名称，前一条字段（循环时使用），参数)while(m=av_dict_get(pFormatCtx->metadata,"",m,AV_DICT_IGNORE_SUFFIX)){ key.Format(m->key); value.Format(m->value); meta+=key+"\t:"+value+"\r\n" ;}

           

给我老师的人工智能教程打call！http://blog.csdn.net/jiangjunshow