音視頻解封裝demo：使用libmp4v2解封裝(demux)出mp4文件中的h264視頻數據和aac語音數據

1、README

前言

本demo是使用的mp4v2來將mp4文件解封裝得到h264、aac的，目前demo提供的.a靜態庫文件是在x86_64架構的Ubuntu16.04編譯得到的，如果想在其他環境下測試demo，可以自行編譯mp4v2并替換相應的庫文件（libmp4v2.a）。

a. 編譯

$ make # 或者`make DEBUG=1`打開調試打印信息

如果想編譯mp4v2，則可以參考以下步驟：

mp4v2源碼下載地址：https://github.com/TechSmith/mp4v2

$ tar xjf mp4v2-2.0.0.tar.bz2
$ cd mp4v2-2.0.0/
$ ./configure --prefix=$PWD/_install # 交叉編譯可參考加上選項: --host=arm-linux-gnueabihf
$ make -j96
$ make install

b. 使用

注：示例2中的音視頻測試源文件是不同步的，不影響本demo的解封裝。

$ ./mp4v2_unpack_demo 
Usage: ./mp4v2_unpack_demo ./avfile/test1.mp4 ./test1_out.h264 ./test1_out.aac./mp4v2_unpack_demo ./avfile/test2.mp4 ./test2_out.h264 ./test2_out.aac

c. 參考文章

01.mp4v2應用—mp4轉h264 - wade_linux - 博客園.mhtml
mp4文件格式解析 - 簡書
MP4格式詳解_DONGHONGBAI的專欄-CSDN博客
使用mp4v2解碼mp4轉成h264碼流和aac碼流_lq496387202的博客-CSDN博客_mp4v2解碼

d. demo目錄結構

.
├── avfile
│   ├── test1.mp4
│   ├── test1_out.aac
│   ├── test1_out.h264
│   ├── test2.mp4
│   ├── test2_out.aac
│   └── test2_out.h264
├── docs
│   ├── 01.mp4v2應用—mp4轉h264 - wade_linux - 博客園.mhtml
│   ├── mp4文件格式解析 - 簡書.mhtml
│   ├── MP4格式詳解_DONGHONGBAI的專欄-CSDN博客.mhtml
│   └── 使用mp4v2解碼mp4轉成h264碼流和aac碼流_lq496387202的博客-CSDN博客_mp4v2解碼.mhtml
├── include
│   └── mp4v2
│       ├── chapter.h
│       ├── file.h
│       ├── file_prop.h
│       ├── general.h
│       ├── isma.h
│       ├── itmf_generic.h
│       ├── itmf_tags.h
│       ├── mp4v2.h
│       ├── platform.h
│       ├── project.h
│       ├── sample.h
│       ├── streaming.h
│       ├── track.h
│       └── track_prop.h
├── lib
│   └── libmp4v2.a
├── main.c
├── Makefile
└── README.md

2、主要代碼片段

main.c

#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <unistd.h>#include "mp4v2/mp4v2.h"// 編譯時Makefile里控制
#ifdef ENABLE_DEBUG#define DEBUG(fmt, args...) 	printf(fmt, ##args)
#else#define DEBUG(fmt, args...)
#endifint unpackMp4File(char *mp4FileName, char *videoFileName, char *audioFileName);unsigned char g_sps[64] = {0};
unsigned char g_pps[64] = {0};
unsigned int  g_spslen  = 0;
unsigned int  g_ppslen  = 0;int main(int argc, char **argv)
{if(argc < 2){printf("Usage: \n""   %s ./avfile/test1.mp4 ./test1_out.h264 ./test1_out.aac\n""   %s ./avfile/test2.mp4 ./test2_out.h264 ./test2_out.aac\n",argv[0], argv[0]);return -1;}int ret = unpackMp4File(argv[1], argv[2], argv[3]);if(ret == 0){printf("\033[32mSuccess!\033[0m\n");}else{printf("\033[31mFailed!\033[0m\n");}return 0;
}int getH264Stream(MP4FileHandle mp4Handler, int videoTrackId, int totalSamples, char *saveFileName)
{// 調用的接口要傳的參數uint32_t curFrameIndex = 1; // `MP4ReadSample`函數的參數要求是從1開始，但我們打印幀下標還是從0開始uint8_t *pData = NULL;uint32_t nSize = 0;MP4Timestamp pStartTime;MP4Duration pDuration;MP4Duration pRenderingOffset;bool pIsSyncSample = 0;// 寫文件要用的參數char naluHeader[4] = {0x00, 0x00, 0x00, 0x01};FILE *fpVideo = NULL;if(!mp4Handler)return -1;fpVideo = fopen(saveFileName, "wb"); if (fpVideo == NULL){printf("open file(%s) error!\n", saveFileName);return -1;}while(curFrameIndex <= totalSamples){   // 如果傳入MP4ReadSample的視頻pData是null，它內部就會new 一個內存// 如果傳入的是已知的內存區域，則需要保證空間bigger then max frames size.MP4ReadSample(mp4Handler, videoTrackId, curFrameIndex, &pData, &nSize, &pStartTime, &pDuration, &pRenderingOffset, &pIsSyncSample);DEBUG("[\033[35mvideo\033[0m] ");if(pIsSyncSample){DEBUG("IDR\t");fwrite(naluHeader, 4, 1, fpVideo);fwrite(g_sps, g_spslen, 1, fpVideo);fwrite(naluHeader, 4, 1, fpVideo);fwrite(g_pps, g_ppslen, 1, fpVideo);}else{DEBUG("SLICE\t");}if(pData && nSize > 4){// `MP4ReadSample`函數的參數要求是從1開始，但我們打印幀下標還是從0開始；而大小已經包含了4字節的start code長度DEBUG("frame index: %d\t size: %d\n", curFrameIndex - 1, nSize);fwrite(naluHeader, 4, 1, fpVideo);fwrite(pData + 4, nSize - 4, 1, fpVideo); // pData+4了，那nSize就要-4}free(pData);pData = NULL;curFrameIndex++;}       fflush(fpVideo);fclose(fpVideo);  return 0;
}int getAACStream(MP4FileHandle mp4Handler, int audioTrackId, int totalSamples, char *saveFileName)
{// 調用的接口要傳的參數uint32_t curFrameIndex = 1; // `MP4ReadSample`函數的參數要求是從1開始，但我們打印幀下標還是從0開始uint8_t *pData = NULL;uint32_t nSize = 0;// 寫文件要用的參數FILE *fpAudio = NULL;if(!mp4Handler)return -1;fpAudio = fopen(saveFileName, "wb");if (fpAudio == NULL){printf("open file(%s) error!\n", saveFileName);return -1;}while(curFrameIndex <= totalSamples){// 如果傳入MP4ReadSample的音頻pData是null，它內部就會new 一個內存// 如果傳入的是已知的內存區域，則需要保證空間bigger then max frames size.MP4ReadSample(mp4Handler, audioTrackId, curFrameIndex, &pData, &nSize, NULL, NULL, NULL, NULL);DEBUG("[\033[36maudio\033[0m] ");if(pData){			DEBUG("frame index: %d\t size: %d\n", curFrameIndex - 1, nSize);fwrite(pData, nSize, 1, fpAudio);}free(pData);pData = NULL;curFrameIndex++;}		fflush(fpAudio);fclose(fpAudio);  return 0;
}int unpackMp4File(char *mp4FileName, char *videoFileName, char *audioFileName)
{MP4FileHandle mp4Handler = 0;uint32_t trackCnt = 0;	int videoTrackId = -1;int audioTrackId = -1;unsigned int videoSampleCnt = 0;unsigned int audioSampleCnt = 0;mp4Handler = MP4Read(mp4FileName);if (mp4Handler <= 0){printf("MP4Read(%s) error!\n", mp4FileName);return -1;}trackCnt = MP4GetNumberOfTracks(mp4Handler, NULL, 0); //獲取音視頻軌道數printf("****************************\n");printf("trackCnt: %d\n", trackCnt);for (int i = 0; i < trackCnt; i++){// 獲取trackId，判斷獲取數據類型: 1-獲取視頻數據，2-獲取音頻數據MP4TrackId trackId = MP4FindTrackId(mp4Handler, i, NULL, 0);const char* trackType = MP4GetTrackType(mp4Handler, trackId);if (MP4_IS_VIDEO_TRACK_TYPE(trackType)){// 不關心，只是打印出來看看MP4Duration duration = 0;uint32_t timescale = 0;videoTrackId = trackId;duration = MP4GetTrackDuration(mp4Handler, videoTrackId);timescale = MP4GetTrackTimeScale(mp4Handler, videoTrackId);videoSampleCnt = MP4GetTrackNumberOfSamples(mp4Handler, videoTrackId);printf("video params: \n"" - trackId: %d\n"" - duration: %lu\n"" - timescale: %d\n"" - samples count: %d\n",videoTrackId, duration, timescale, videoSampleCnt);// 讀取 sps/pps uint8_t **seqheader;			uint32_t *seqheadersize;uint8_t **pictheader;uint32_t *pictheadersize;MP4GetTrackH264SeqPictHeaders(mp4Handler, videoTrackId, &seqheader, &seqheadersize, &pictheader, &pictheadersize);// 獲取spsfor (int ix = 0; seqheadersize[ix] != 0; ix++){memcpy(g_sps, seqheader[ix], seqheadersize[ix]);g_spslen = seqheadersize[ix];free(seqheader[ix]);}free(seqheader);free(seqheadersize);// 獲取ppsfor (int ix = 0; pictheader[ix] != 0; ix++){memcpy(g_pps, pictheader[ix], pictheadersize[ix]);g_ppslen = pictheadersize[ix];free(pictheader[ix]);}free(pictheader);free(pictheadersize);}else if (MP4_IS_AUDIO_TRACK_TYPE(trackType)){audioTrackId = trackId;audioSampleCnt = MP4GetTrackNumberOfSamples(mp4Handler, audioTrackId);printf("audio params: \n"" - trackId: %d\n"" - samples count: %d\n",audioTrackId, audioSampleCnt);}}printf("****************************\n");// 解析完了mp4，主要是為了獲取sps pps 還有video的trackIDif(videoTrackId >= 0){getH264Stream(mp4Handler, videoTrackId, videoSampleCnt, videoFileName);  }if(audioTrackId >= 0){getAACStream(mp4Handler, audioTrackId, audioSampleCnt, audioFileName);}// 需要mp4close 否則在嵌入式設備打開mp4上多了會內存泄露掛掉.MP4Close(mp4Handler, 0);return 0;
}