ffmpeg 中 rtsp相关源码解析(五)

谁践踏了优雅 2021-09-07 06:10 561阅读 0赞

# [<<ffmpeg 中 rtsp相关源码解析(四)>>][ffmpeg _ rtsp]  #

### 简单介绍了rtp包是通过ff\_rtsp\_fetch\_packet来获取的 ###

我们更深入地问一个问题，那么对于vp8 、h264等音视频编解码标准，这些packet又是如何处理的呢？

这边就涉及一个非常重要的结构体

RTPDynamicProtocolHandler

位于rtpdec.h文件中定义如下

struct RTPDynamicProtocolHandler {
        const char *enc_name;
        enum AVMediaType codec_type;
        enum AVCodecID codec_id;
        enum AVStreamParseType need_parsing;
        int static_payload_id; /* 0 means no payload id is set. 0 is a valid
                                * payload ID (PCMU), too, but that format doesn't
                                * require any custom depacketization code. */
        int priv_data_size;
    
        /** Initialize dynamic protocol handler, called after the full rtpmap line is parsed, may be null */
        int (*init)(AVFormatContext *s, int st_index, PayloadContext *priv_data);
        /** Parse the a= line from the sdp field */
        int (*parse_sdp_a_line)(AVFormatContext *s, int st_index,
                                PayloadContext *priv_data, const char *line);
        /** Free any data needed by the rtp parsing for this dynamic data.
          * Don't free the protocol_data pointer itself, that is freed by the
          * caller. This is called even if the init method failed. */
        void (*close)(PayloadContext *protocol_data);
        /** Parse handler for this dynamic packet */
        DynamicPayloadPacketHandlerProc parse_packet;
        int (*need_keyframe)(PayloadContext *context);
    
        struct RTPDynamicProtocolHandler *next;
    };

其中DynamicPayloadPacketHandlerProc parse\_packet; 就是我们关注的解析包的方法

以h264协议为例，ffmpeg 在 rtpdec\_h264.c中定义了ff\_h264\_dynamic\_handler，并且在rtpdec.c 中

通过ff\_register\_dynamic\_payload\_handler(&ff\_h264\_dynamic\_handler);注册了这个handler

RTPDynamicProtocolHandler ff_h264_dynamic_handler = {
        .enc_name         = "H264",
        .codec_type       = AVMEDIA_TYPE_VIDEO,
        .codec_id         = AV_CODEC_ID_H264,
        .need_parsing     = AVSTREAM_PARSE_FULL,
        .priv_data_size   = sizeof(PayloadContext),
        .parse_sdp_a_line = parse_h264_sdp_line,
        .close            = h264_close_context,
        .parse_packet     = h264_handle_packet,
    };

而它的调用方式如下

如果有buffer

ff\_rtsp\_fetch\_packet ---> ff\_rtp\_parse\_packet  ----> rtp\_parse\_one\_packet ---> rtp\_parse\_packet\_internal ---> parse\_packet

如果没有buffer

ff\_rtsp\_fetch\_packet ---> ff\_rtp\_parse\_packet  ----> rtp\_parse\_one\_packet  ---> parse\_packet

[ffmpeg _ rtsp]: https://blog.csdn.net/u011337602/article/details/113773210