FFmpeg的AVcodecParser

news/2024/7/24 8:47:27 标签: ffmpeg

文章目录

    • 结构体
    • 操作函数
    • 支持的AVCodecParser

这个模块是AVCodec中的子模块,专门用来提前解析码流的元数据,为后面的解码做准备,这一点对cuda-NVdec非常明显,英伟达解码器的元数据解析是放在CPU上的,所以就非常依赖这个解析模块。
FFmpeg中有很多现成的parser
在这里插入图片描述

结构体

下面是它的上下文


typedef struct AVCodecParserContext {
    void *priv_data;
    const struct AVCodecParser *parser;
    int64_t frame_offset; /* offset of the current frame */
    int64_t cur_offset; /* current offset
                           (incremented by each av_parser_parse()) */
    int64_t next_frame_offset; /* offset of the next frame */
    /* video info */
    int pict_type; /* XXX: Put it back in AVCodecContext. */
    /**
     * This field is used for proper frame duration computation in lavf.
     * It signals, how much longer the frame duration of the current frame
     * is compared to normal frame duration.
     *
     * frame_duration = (1 + repeat_pict) * time_base
     *
     * It is used by codecs like H.264 to display telecined material.
     */
    int repeat_pict; /* XXX: Put it back in AVCodecContext. */
    int64_t pts;     /* pts of the current frame */
    int64_t dts;     /* dts of the current frame */

    /* private data */
    int64_t last_pts;
    int64_t last_dts;
    int fetch_timestamp;

#define AV_PARSER_PTS_NB 4
    int cur_frame_start_index;
    int64_t cur_frame_offset[AV_PARSER_PTS_NB];
    int64_t cur_frame_pts[AV_PARSER_PTS_NB];
    int64_t cur_frame_dts[AV_PARSER_PTS_NB];

    int flags;
#define PARSER_FLAG_COMPLETE_FRAMES           0x0001
#define PARSER_FLAG_ONCE                      0x0002
/// Set if the parser has a valid file offset
#define PARSER_FLAG_FETCHED_OFFSET            0x0004
#define PARSER_FLAG_USE_CODEC_TS              0x1000

    int64_t offset;      ///< byte offset from starting packet start
    int64_t cur_frame_end[AV_PARSER_PTS_NB];

    /**
     * Set by parser to 1 for key frames and 0 for non-key frames.
     * It is initialized to -1, so if the parser doesn't set this flag,
     * old-style fallback using AV_PICTURE_TYPE_I picture type as key frames
     * will be used.
     */
    int key_frame;

    // Timestamp generation support:
    /**
     * Synchronization point for start of timestamp generation.
     *
     * Set to >0 for sync point, 0 for no sync point and <0 for undefined
     * (default).
     *
     * For example, this corresponds to presence of H.264 buffering period
     * SEI message.
     */
    int dts_sync_point;

    /**
     * Offset of the current timestamp against last timestamp sync point in
     * units of AVCodecContext.time_base.
     *
     * Set to INT_MIN when dts_sync_point unused. Otherwise, it must
     * contain a valid timestamp offset.
     *
     * Note that the timestamp of sync point has usually a nonzero
     * dts_ref_dts_delta, which refers to the previous sync point. Offset of
     * the next frame after timestamp sync point will be usually 1.
     *
     * For example, this corresponds to H.264 cpb_removal_delay.
     */
    int dts_ref_dts_delta;

    /**
     * Presentation delay of current frame in units of AVCodecContext.time_base.
     *
     * Set to INT_MIN when dts_sync_point unused. Otherwise, it must
     * contain valid non-negative timestamp delta (presentation time of a frame
     * must not lie in the past).
     *
     * This delay represents the difference between decoding and presentation
     * time of the frame.
     *
     * For example, this corresponds to H.264 dpb_output_delay.
     */
    int pts_dts_delta;

    /**
     * Position of the packet in file.
     *
     * Analogous to cur_frame_pts/dts
     */
    int64_t cur_frame_pos[AV_PARSER_PTS_NB];

    /**
     * Byte position of currently parsed frame in stream.
     */
    int64_t pos;

    /**
     * Previous frame byte position.
     */
    int64_t last_pos;

    /**
     * Duration of the current frame.
     * For audio, this is in units of 1 / AVCodecContext.sample_rate.
     * For all other types, this is in units of AVCodecContext.time_base.
     */
    int duration;

    enum AVFieldOrder field_order;

    /**
     * Indicate whether a picture is coded as a frame, top field or bottom field.
     *
     * For example, H.264 field_pic_flag equal to 0 corresponds to
     * AV_PICTURE_STRUCTURE_FRAME. An H.264 picture with field_pic_flag
     * equal to 1 and bottom_field_flag equal to 0 corresponds to
     * AV_PICTURE_STRUCTURE_TOP_FIELD.
     */
    enum AVPictureStructure picture_structure;

    /**
     * Picture number incremented in presentation or output order.
     * This field may be reinitialized at the first picture of a new sequence.
     *
     * For example, this corresponds to H.264 PicOrderCnt.
     */
    int output_picture_number;

    /**
     * Dimensions of the decoded video intended for presentation.
     */
    int width;
    int height;

    /**
     * Dimensions of the coded video.
     */
    int coded_width;
    int coded_height;

    /**
     * The format of the coded data, corresponds to enum AVPixelFormat for video
     * and for enum AVSampleFormat for audio.
     *
     * Note that a decoder can have considerable freedom in how exactly it
     * decodes the data, so the format reported here might be different from the
     * one returned by a decoder.
     */
    int format;
} AVCodecParserContext;

下面是插件接口的入口。
如果你要实现一个自己的parse,只要简单的重写下面四个函数就可以了。

typedef struct AVCodecParser {
    int codec_ids[7]; /* several codec IDs are permitted */
    int ;
    int (*parser_init)(AVCodecParserContext *s);
    /* This callback never returns an error, a negative value means that
     * the frame start was in a previous packet. */
    int (*parser_parse)(AVCodecParserContext *s,
                        AVCodecContext *avctx,
                        const uint8_t **poutbuf, int *poutbuf_size,
                        const uint8_t *buf, int buf_size);
    void (*parser_close)(AVCodecParserContext *s);
    int (*split)(AVCodecContext *avctx, const uint8_t *buf, int buf_size);
} AVCodecParser;

特别注意上面的priv_data_size是AVCodecParserContext中的priv_data,这里面就是保存各种元信息。比如说对于h264来说,里面保存的就是:

static av_cold int init(AVCodecParserContext *s)
{
    H264ParseContext *p = s->priv_data;
	...
    return 0;
}

其实就是sps,pps,sei的值

typedef struct H264ParseContext {
    ParseContext pc;
    H264ParamSets ps;
    H264DSPContext h264dsp;
    H264POCContext poc;
    H264SEIContext sei;
    int is_avc;
    int nal_length_size;
    int got_first;
    int picture_structure;
    uint8_t parse_history[6];
    int parse_history_count;
    int parse_last_mb;
    int64_t reference_dts;
    int last_frame_num, last_picture_structure;
} H264ParseContext;

比如h264就是重写了下面三个函数接口。

const AVCodecParser ff_h264_parser = {
    .codec_ids      = { AV_CODEC_ID_H264 },
    .priv_data_size = sizeof(H264ParseContext),
    .parser_init    = init,
    .parser_parse   = h264_parse,
    .parser_close   = h264_close,
};

操作函数

下面是操作函数,第一个是列出所有的parser

/**
 * Iterate over all registered codec parsers.
 *
 * @param opaque a pointer where libavcodec will store the iteration state. Must
 *               point to NULL to start the iteration.
 *
 * @return the next registered codec parser or NULL when the iteration is
 *         finished
 */
const AVCodecParser *av_parser_iterate(void **opaque);

第二个是初始化parser

AVCodecParserContext *av_parser_init(int codec_id);

第三个是最重要的,也就是解析pkt

/**
 * Parse a packet.
 *
 * @param s             parser context.
 * @param avctx         codec context.
 * @param poutbuf       set to pointer to parsed buffer or NULL if not yet finished.
 * @param poutbuf_size  set to size of parsed buffer or zero if not yet finished.
 * @param buf           input buffer.
 * @param buf_size      buffer size in bytes without the padding. I.e. the full buffer
                        size is assumed to be buf_size + AV_INPUT_BUFFER_PADDING_SIZE.
                        To signal EOF, this should be 0 (so that the last frame
                        can be output).
 * @param pts           input presentation timestamp.
 * @param dts           input decoding timestamp.
 * @param pos           input byte position in stream.
 * @return the number of bytes of the input bitstream used.
 *
 * Example:
 * @code
 *   while(in_len){
 *       len = av_parser_parse2(myparser, AVCodecContext, &data, &size,
 *                                        in_data, in_len,
 *                                        pts, dts, pos);
 *       in_data += len;
 *       in_len  -= len;
 *
 *       if(size)
 *          decode_frame(data, size);
 *   }
 * @endcode
 */
int av_parser_parse2(AVCodecParserContext *s,
                     AVCodecContext *avctx,
                     uint8_t **poutbuf, int *poutbuf_size,
                     const uint8_t *buf, int buf_size,
                     int64_t pts, int64_t dts,
                     int64_t pos);


最后一个是close parser

void av_parser_close(AVCodecParserContext *s);

下面是一个用例


int main(int argc, char **argv)
{
    const char *filename, *outfilename;
    const AVCodec *codec;
    AVCodecParserContext *parser;
    AVCodecContext *c= NULL;
    FILE *f;
    AVFrame *frame;
    uint8_t inbuf[INBUF_SIZE + AV_INPUT_BUFFER_PADDING_SIZE];
    uint8_t *data;
    size_t   data_size;
    int ret;
    AVPacket *pkt;

    if (argc <= 2) {
        fprintf(stderr, "Usage: %s <input file> <output file>\n"
                "And check your input file is encoded by mpeg1video please.\n", argv[0]);
        exit(0);
    }
    filename    = argv[1];
    outfilename = argv[2];

    pkt = av_packet_alloc();
    if (!pkt)
        exit(1);

    /* set end of buffer to 0 (this ensures that no overreading happens for damaged MPEG streams) */
    memset(inbuf + INBUF_SIZE, 0, AV_INPUT_BUFFER_PADDING_SIZE);

    /* find the MPEG-1 video decoder */
    codec = avcodec_find_decoder(AV_CODEC_ID_MPEG1VIDEO);
    if (!codec) {
        fprintf(stderr, "Codec not found\n");
        exit(1);
    }

    parser = av_parser_init(codec->id);
    if (!parser) {
        fprintf(stderr, "parser not found\n");
        exit(1);
    }

    c = avcodec_alloc_context3(codec);
    if (!c) {
        fprintf(stderr, "Could not allocate video codec context\n");
        exit(1);
    }

    /* For some codecs, such as msmpeg4 and mpeg4, width and height
       MUST be initialized there because this information is not
       available in the bitstream. */

    /* open it */
    if (avcodec_open2(c, codec, NULL) < 0) {
        fprintf(stderr, "Could not open codec\n");
        exit(1);
    }

    f = fopen(filename, "rb");
    if (!f) {
        fprintf(stderr, "Could not open %s\n", filename);
        exit(1);
    }

    frame = av_frame_alloc();
    if (!frame) {
        fprintf(stderr, "Could not allocate video frame\n");
        exit(1);
    }

    while (!feof(f)) {
        /* read raw data from the input file */
        data_size = fread(inbuf, 1, INBUF_SIZE, f);
        if (!data_size)
            break;

        /* use the parser to split the data into frames */
        data = inbuf;
        while (data_size > 0) {
            ret = av_parser_parse2(parser, c, &pkt->data, &pkt->size,
                                   data, data_size, AV_NOPTS_VALUE, AV_NOPTS_VALUE, 0);
            if (ret < 0) {
                fprintf(stderr, "Error while parsing\n");
                exit(1);
            }
            data      += ret;
            data_size -= ret;

            if (pkt->size)
                decode(c, frame, pkt, outfilename);
        }
    }

    /* flush the decoder */
    decode(c, frame, NULL, outfilename);

    fclose(f);

    av_parser_close(parser);
    avcodec_free_context(&c);
    av_frame_free(&frame);
    av_packet_free(&pkt);

    return 0;
}

上面的例子展示了一个从文件中读取码流,然后通过av_parser_parse2将码流解析为一个个分离的NALU单元,然后送去解码,我们可以看到,第二个参数是c,也就是说将来要把sps,pps,sei这些信息保存在c中的extradata中。

下面第一个函数是av_parser_init,看到除了分配空间外,就是对parser->parser_init(s)的封装。


AVCodecParserContext *av_parser_init(int codec_id)
{
    AVCodecParserContext *s = NULL;
    const AVCodecParser *parser;
    void *i = 0;
    int ret;

    if (codec_id == AV_CODEC_ID_NONE)
        return NULL;

    while ((parser = av_parser_iterate(&i))) {
        if (parser->codec_ids[0] == codec_id ||
            parser->codec_ids[1] == codec_id ||
            parser->codec_ids[2] == codec_id ||
            parser->codec_ids[3] == codec_id ||
            parser->codec_ids[4] == codec_id ||
            parser->codec_ids[5] == codec_id ||
            parser->codec_ids[6] == codec_id)
            goto found;
    }
    return NULL;

found:
    s = av_mallocz(sizeof(AVCodecParserContext));
    if (!s)
        goto err_out;
    s->parser = parser;
    s->priv_data = av_mallocz(parser->priv_data_size);
    if (!s->priv_data)
        goto err_out;
    s->fetch_timestamp=1;
    s->pict_type = AV_PICTURE_TYPE_I;
    if (parser->parser_init) {
        ret = parser->parser_init(s);
        if (ret != 0)
            goto err_out;
    }
    s->key_frame            = -1;
    s->dts_sync_point       = INT_MIN;
    s->dts_ref_dts_delta    = INT_MIN;
    s->pts_dts_delta        = INT_MIN;
    s->format               = -1;

    return s;

err_out:
    if (s)
        av_freep(&s->priv_data);
    av_free(s);
    return NULL;
}

在这里插入图片描述
第二个函数也很简单av_parser_parse2,内部其实也是对parser_parse的封装。

int av_parser_parse2(AVCodecParserContext *s, AVCodecContext *avctx,
                     uint8_t **poutbuf, int *poutbuf_size,
                     const uint8_t *buf, int buf_size,
                     int64_t pts, int64_t dts, int64_t pos){
   ...
    index = s->parser->parser_parse(s, avctx, (const uint8_t **) poutbuf,
                                    poutbuf_size, buf, buf_size);
                                   
  ...
  }

第三个函数av_parser_close就更加简单了

void av_parser_close(AVCodecParserContext *s)
{
    if (s) {
        if (s->parser->parser_close)
            s->parser->parser_close(s);
        av_freep(&s->priv_data);
        av_free(s);
    }
}

上面全部是架构代码,如果想要查看详细的解析过程,可以单独看每个插件的内容,也就是那三个函数。

支持的AVCodecParser


extern const AVCodecParser ff_aac_parser;
extern const AVCodecParser ff_aac_latm_parser;
extern const AVCodecParser ff_ac3_parser;
extern const AVCodecParser ff_adx_parser;
extern const AVCodecParser ff_amr_parser;
extern const AVCodecParser ff_av1_parser;
extern const AVCodecParser ff_avs2_parser;
extern const AVCodecParser ff_avs3_parser;
extern const AVCodecParser ff_bmp_parser;
extern const AVCodecParser ff_cavsvideo_parser;
extern const AVCodecParser ff_cook_parser;
extern const AVCodecParser ff_cri_parser;
extern const AVCodecParser ff_dca_parser;
extern const AVCodecParser ff_dirac_parser;
extern const AVCodecParser ff_dnxhd_parser;
extern const AVCodecParser ff_dolby_e_parser;
extern const AVCodecParser ff_dpx_parser;
extern const AVCodecParser ff_dvaudio_parser;
extern const AVCodecParser ff_dvbsub_parser;
extern const AVCodecParser ff_dvdsub_parser;
extern const AVCodecParser ff_dvd_nav_parser;
extern const AVCodecParser ff_flac_parser;
extern const AVCodecParser ff_g723_1_parser;
extern const AVCodecParser ff_g729_parser;
extern const AVCodecParser ff_gif_parser;
extern const AVCodecParser ff_gsm_parser;
extern const AVCodecParser ff_h261_parser;
extern const AVCodecParser ff_h263_parser;
extern const AVCodecParser ff_h264_parser;
extern const AVCodecParser ff_hevc_parser;
extern const AVCodecParser ff_ipu_parser;
extern const AVCodecParser ff_jpeg2000_parser;
extern const AVCodecParser ff_mjpeg_parser;
extern const AVCodecParser ff_mlp_parser;
extern const AVCodecParser ff_mpeg4video_parser;
extern const AVCodecParser ff_mpegaudio_parser;
extern const AVCodecParser ff_mpegvideo_parser;
extern const AVCodecParser ff_opus_parser;
extern const AVCodecParser ff_png_parser;
extern const AVCodecParser ff_pnm_parser;
extern const AVCodecParser ff_rv30_parser;
extern const AVCodecParser ff_rv40_parser;
extern const AVCodecParser ff_sbc_parser;
extern const AVCodecParser ff_sipr_parser;
extern const AVCodecParser ff_tak_parser;
extern const AVCodecParser ff_vc1_parser;
extern const AVCodecParser ff_vorbis_parser;
extern const AVCodecParser ff_vp3_parser;
extern const AVCodecParser ff_vp8_parser;
extern const AVCodecParser ff_vp9_parser;
extern const AVCodecParser ff_webp_parser;
extern const AVCodecParser ff_xbm_parser;
extern const AVCodecParser ff_xma_parser;

http://www.niftyadmin.cn/n/5255629.html

相关文章

eclipse乱码怎么解决?

在使用eclipse的时候遇到了乱码这种情况&#xff0c;这个时候我们该怎么办呢?接下来小编就为大家带来了eclipse乱码解决办法&#xff0c;还不清楚如何解决的小伙伴千万不要错过这个教程哦。 eclipse乱码解决方法 首先&#xff0c;打开Eclipse软件&#xff0c;选中项目并点击…

Spring 的缓存机制【记录】

一、背景 在最近的业务需求开发过程中遇到了“传说中”的循环依赖问题&#xff0c;在之前学习Spring的时候经常会看到Spring是如何解决循环依赖问题的&#xff0c;所谓循环依赖即形成了一个环状的依赖关系&#xff0c;这个环中的某一个点产生不稳定变化都会导致整个链路产生不…

java工程(ajax/axios/postman)向请求头中添加消息

1、问题概述 在项目中我们经常会遇到需要向请求头中添加消息的场景&#xff0c;然后后端通过request.getRequest()或者RequestHeader获取请求头中的消息。 下面提供几种前端向请求头添加消息的方式 2、创建一个springmvc工程用于测试 2.1、创建工程并引入相关包信息 sprin…

第三十一章 控制到 XML 模式的映射 - %ListOfDataTypes

文章目录 第三十一章 控制到 XML 模式的映射 - %ListOfDataTypes%ListOfDataTypes%ArrayOfDataTypes 第三十一章 控制到 XML 模式的映射 - %ListOfDataTypes %ListOfDataTypes 本部分显示从支持 XML 的类生成的 XML 架构的一部分&#xff0c;该类包含中定义为%ListOfDataType…

PyCharm连接远程服务器上Docker容器,使用远程服务器的python intercepter解释器和GPU资源 [本地调试深度学习代码]

概述 在编写常规深度学习代码时&#xff0c;总是需要使用服务器上的GPU资源&#xff0c;所以一般要写完代码&#xff0c;放到服务器&#xff0c;然后使用GPU运行。但是由于之前的习惯&#xff0c;总想本地调试一下或者本地直接跑测试结果&#xff0c;再放到服务器去跑。 网上…

Spark与PySpark(1.概述、框架、模块)

目录 1.Spark 概念 2. Hadoop和Spark的对比 3. Spark特点 3.1 运行速度快 3.2 简单易用 3.3 通用性强 3.4 可以允许运行在很多地方 4. Spark框架模块 4.1 Spark Core 4.2 SparkSQL 4.3 SparkStreaming 4.4 MLlib 4.5 GraphX 5. Spark的运行模式 5.1 本地模式(单机) Local运行模…

[cleanrl] ppo_continuous_action源码解析

1 import库&#xff08;略&#xff09; import os import random import time from dataclasses import dataclassimport gymnasium as gym import numpy as np import torch import torch.nn as nn import torch.optim as optim import tyro from torch.distributions.normal…

Oracle-查找函数session,并结束此函数的session执行,查看错误日志

select * from V$DB_OBJECT_CASHE WHERE locks >0 and pins >0 and TYPE like FUNCTION;--查找执行的函数&#xff0c;如果查存储过程则PROCEDURE,找到对应的name值了。 SELECT * FROM V$ACCESS WHERE OWNER过程的所属用户 AND NAME刚才查到的过程名;--有的版本需查表GV…