ffmpeg h264文件转mp4

Posted 2023-04-08 tusong86

h264文件不能直接在网页上播放，比如在浏览器上输入http://10.0.0.2/2022-01-08T22-32-58.h264，变成了下载。

若在浏览器上输入http://10.0.0.2/2022-01-08T22-32-58.mp4，则可以播放。

本文讲解用ffmpeg将h264文件转换成mp4。

首先，准备h264文件，这个可以用ffmpeg将一个mp4的视频部分转成h264，命令如下：

ffmpeg -i 2022-01-08T22-32-58.mp4 -an -vcodec copy 2022-01-08T22-32-58.h264

注意，我这里mp4里面的视频编码格式是h264，故用的vcodec copy，否则用vcodec libx264。

下面是将这个2022-01-08T22-32-58.h264封装成mp4，注意，没有解码编码过程，速度会很快。
大概过程如下：
1.启动两个线程
2.第一个线程av_read_frame读到packet，然后av_packet_clone拷贝出一个packet，塞入队列
3.第二个线程从队列中读取packet，进行pts,dts,duration设置，然后调用av_interleaved_write_frame写入。

这里面需要注意的地方在于pts,dts的设置，从ffmpeg源码上看，对于pts,dts的设置，也是很麻烦的，源码中，很多地方，动不动就看pts,dts是否需要设置。

本人之前采取下列方式进行dts和pts的设置，即由于读的context和写入的context的时间基(time_base)不一致，需要转换成目标文件context的时间基。

pPacketVideoA->dts = av_rescale_q_rnd(pPacketVideoA->dts, m_pFormatCtx_File->streams[m_iVideoIndex]->time_base, m_pFormatCtx_Out->streams[m_iVideoIndex]->time_base, AVRounding(1));

但实际运行时，出现了问题，读取出来的packet的pts和dts都是AV_NOPTS_VALUE，经过转换还是AV_NOPTS_VALUE，然后调用av_interleaved_write_frame出错。

最后调试ffmpeg源码，得到pts和dts的处理如下：

pPacketVideoA->duration = av_rescale_q_rnd(pPacketVideoA->duration, m_pFormatCtx_File->streams[m_iVideoIndex]->time_base, m_pFormatCtx_Out->streams[m_iVideoIndex]->time_base, AVRounding(1));
if (pPacketVideoA->dts != AV_NOPTS_VALUE)

	pPacketVideoA->dts = av_rescale_q_rnd(pPacketVideoA->dts, m_pFormatCtx_File->streams[m_iVideoIndex]->time_base, m_pFormatCtx_Out->streams[m_iVideoIndex]->time_base, AVRounding(1));
	pPacketVideoA->pts = pPacketVideoA->dts;

else

	pPacketVideoA->pts = pPacketVideoA->dts = m_iTotalDuration;

m_iTotalDuration += pPacketVideoA->duration;

m_iTotalDuration的初始值，本人设置的是0，即第一帧的pts和dts都设置成0，下一帧的dts就是前面总帧的duration和，这个是ffmpeg里面的逻辑。

当然，ffmpeg里面判断复杂多了，比如参考了是否有B帧，本人所写的没考虑B帧，只考虑一个文件里面只有I帧，P帧的情况。

下面是代码结构

其中FfmpegH264ToMp4.cpp的内容如下：

// FfmpegH264ToMp4.cpp : 此文件包含 "main" 函数。程序执行将在此处开始并结束。
//

#include <iostream>
#include "H264ToMp4.h"

int main()

	CVideoConvert cVideoConvert;

	std::string strVideoFile = "2022-01-08T22-32-58.h264";
	std::string strOutFile = "out.mp4";
	cVideoConvert.StartConvert(strVideoFile, strOutFile);
	cVideoConvert.WaitFinish();
    return 0;

H264ToMp4.h里面的内容如下：

#pragma once
#include <string>

#include <Windows.h>
#include <deque>

#define MAX_VIDEOPACKET_NUM 100

#ifdef	__cplusplus
extern "C"

#endif
#include "libavcodec/avcodec.h"
#include "libavformat/avformat.h"
#include "libswscale/swscale.h"
#include "libswresample/swresample.h"
#include "libavdevice/avdevice.h"
#include "libavutil/audio_fifo.h"
#include "libavutil/avutil.h"
#include "libavutil/fifo.h"
#include "libavutil/frame.h"
#include "libavutil/imgutils.h"

#include "libavfilter/avfilter.h"
#include "libavfilter/buffersink.h"
#include "libavfilter/buffersrc.h"


#ifdef __cplusplus
;
#endif

class CVideoConvert

public:
	CVideoConvert();
	~CVideoConvert();
public:
	int StartConvert(std::string strFileName, std::string strFileOut);
	void StopConvert();
public:
	int OpenFile(const char *pFile);
	int OpenOutPut(const char *pFileOut);
private:
	static DWORD WINAPI FileReadProc(LPVOID lpParam);
	void FileRead();

	static DWORD WINAPI FileWriteProc(LPVOID lpParam);
	void FileWrite();

public:
	void WaitFinish();
public:
	AVFormatContext *m_pFormatCtx_File = NULL;
	AVFormatContext *m_pFormatCtx_Out = NULL;

	std::deque<AVPacket *> m_vecReadVideo;


	int m_iVideoWidth = 1920;
	int m_iVideoHeight = 1080;
	int m_iYuv420FrameSize = 0;
	int m_iVideoFrameNum = 0;

	int m_iVideoIndex = -1;
	bool m_bStartConvert = false;
	int64_t m_iTotalDuration = 0;
private:
	CRITICAL_SECTION m_csVideoReadSection;
	HANDLE m_hFileReadThread = NULL;
	HANDLE m_hFileWriteThread = NULL;
;

H264ToMp4.cpp里面的内容如下：

#include "H264ToMp4.h"


#ifdef	__cplusplus
extern "C"

#endif

#pragma comment(lib, "avcodec.lib")
#pragma comment(lib, "avformat.lib")
#pragma comment(lib, "avutil.lib")
#pragma comment(lib, "avdevice.lib")
#pragma comment(lib, "avfilter.lib")
#pragma comment(lib, "postproc.lib")
#pragma comment(lib, "swresample.lib")
#pragma comment(lib, "swscale.lib")


#ifdef __cplusplus
;
#endif

CVideoConvert::CVideoConvert()

	InitializeCriticalSection(&m_csVideoReadSection);


CVideoConvert::~CVideoConvert()

	DeleteCriticalSection(&m_csVideoReadSection);


int CVideoConvert::StartConvert(std::string strFileName, std::string strFileOut)

	int ret = -1;
	do
	
		ret = OpenFile(strFileName.c_str());
		if (ret != 0)
		
			break;
		

		ret = OpenOutPut(strFileOut.c_str());
		if (ret != 0)
		
			break;
		

		if (ret < 0)
		
			break;
		

		m_bStartConvert = true;
		m_hFileReadThread = CreateThread(NULL, 0, FileReadProc, this, 0, NULL);
		m_hFileWriteThread = CreateThread(NULL, 0, FileWriteProc, this, 0, NULL);
	 while (0);

	return ret;


void CVideoConvert::StopConvert()

	m_bStartConvert = false;


int CVideoConvert::OpenFile(const char *pFile)

	int ret = -1;

	do
	
		if ((ret = avformat_open_input(&m_pFormatCtx_File, pFile, 0, 0)) < 0) 
			printf("Could not open input file.");
			break;
		
		if ((ret = avformat_find_stream_info(m_pFormatCtx_File, 0)) < 0) 
			printf("Failed to retrieve input stream information");
			break;
		
		if (m_pFormatCtx_File->nb_streams > 2)
		
			break;
		
		for (int i = 0; i < m_pFormatCtx_File->nb_streams; i++)
		
			if (m_pFormatCtx_File->streams[i]->codecpar->codec_type == AVMEDIA_TYPE_VIDEO)
			
				m_iVideoIndex = i;
			
		

		if (m_iVideoIndex == -1)
		
			break;
		

		ret = 0;
	 while (0);


	return ret;



int CVideoConvert::OpenOutPut(const char *pFileOut)

	int iRet = -1;

	AVStream *pVideoStream = NULL;
	do
	
		avformat_alloc_output_context2(&m_pFormatCtx_Out, NULL, NULL, pFileOut);
		if (m_iVideoIndex != -1)
		
			AVCodecID eTargetVideoCodecId = m_pFormatCtx_File->streams[m_iVideoIndex]->codecpar->codec_id;
			AVCodec* pCodecEncode_Video = (AVCodec *)avcodec_find_encoder(eTargetVideoCodecId);
			pVideoStream = avformat_new_stream(m_pFormatCtx_Out, pCodecEncode_Video);
			if (!pVideoStream)
			
				break;
			

			avcodec_parameters_copy(pVideoStream->codecpar, m_pFormatCtx_File->streams[m_iVideoIndex]->codecpar);
			pVideoStream->codecpar->codec_tag = 0;
		

		if (!(m_pFormatCtx_Out->oformat->flags & AVFMT_NOFILE))
		
			if (avio_open(&m_pFormatCtx_Out->pb, pFileOut, AVIO_FLAG_WRITE) < 0)
			
				break;
			
		

		if (avformat_write_header(m_pFormatCtx_Out, NULL) < 0)
		
			break;
		

		iRet = 0;
	 while (0);


	if (iRet != 0)
	
		if (m_pFormatCtx_Out != NULL)
		
			avformat_free_context(m_pFormatCtx_Out);
			m_pFormatCtx_Out = NULL;
		
	

	return iRet;


DWORD WINAPI CVideoConvert::FileReadProc(LPVOID lpParam)

	CVideoConvert *pVideoConvert = (CVideoConvert *)lpParam;
	if (pVideoConvert != NULL)
	
		pVideoConvert->FileRead();
	
	return 0;


void CVideoConvert::FileRead()

	AVPacket packet =  0 ;
	int ret = 0;
	int iPicCount = 0;
	const int genpts = m_pFormatCtx_File->flags & AVFMT_FLAG_GENPTS;
	while (m_bStartConvert)
	
		av_packet_unref(&packet);

		ret = av_read_frame(m_pFormatCtx_File, &packet);
		if (ret == AVERROR(EAGAIN))
		
			continue;
		
		else if (ret == AVERROR_EOF)
		
			break;
		
		else if (ret < 0)
		
			break;
		

		if (packet.stream_index == m_iVideoIndex)
		
			while (1)
			
				int iAudioPacketNum = m_vecReadVideo.size();
				if (iAudioPacketNum >= MAX_VIDEOPACKET_NUM)
				
					Sleep(10);
					continue;
				
				else
				
					AVPacket *pAudioPacket = av_packet_clone(&packet);
					if (pAudioPacket != NULL)
					
						EnterCriticalSection(&m_csVideoReadSection);
						m_vecReadVideo.push_back(pAudioPacket);
						LeaveCriticalSection(&m_csVideoReadSection);
					
				
				break;
			
		
		
	
	//FlushVideoDecoder(iPicCount);




DWORD WINAPI CVideoConvert::FileWriteProc(LPVOID lpParam)

	CVideoConvert *pVideoConvert = (CVideoConvert *)lpParam;
	if (pVideoConvert != NULL)
	
		pVideoConvert->FileWrite();
	
	return 0;



void CVideoConvert::FileWrite()

	int ret = 0;


	AVPacket packet =  0 ;
	int iRealPicCount = 0;

	while (m_bStartConvert)
	

		int iVideoSize = m_vecReadVideo.size();

		if (iVideoSize > 0)
		
			AVPacket *pPacketVideoA = NULL;
			EnterCriticalSection(&m_csVideoReadSection);
			if (!m_vecReadVideo.empty())
			
				pPacketVideoA = m_vecReadVideo.front();
				m_vecReadVideo.pop_front();
			
			LeaveCriticalSection(&m_csVideoReadSection);

			pPacketVideoA->stream_index = m_iVideoIndex;
			pPacketVideoA->duration = av_rescale_q_rnd(pPacketVideoA->duration, m_pFormatCtx_File->streams[m_iVideoIndex]->time_base, m_pFormatCtx_Out->streams[m_iVideoIndex]->time_base, AVRounding(1));
			if (pPacketVideoA->dts != AV_NOPTS_VALUE)
			
				pPacketVideoA->dts = av_rescale_q_rnd(pPacketVideoA->dts, m_pFormatCtx_File->streams[m_iVideoIndex]->time_base, m_pFormatCtx_Out->streams[m_iVideoIndex]->time_base, AVRounding(1));
				pPacketVideoA->pts = pPacketVideoA->dts;
			
			else
			
				pPacketVideoA->pts = pPacketVideoA->dts = m_iTotalDuration;
			
			m_iTotalDuration += pPacketVideoA->duration;

			//ret = av_write_frame(m_pFormatCtx_Out, pPacketVideoA);
			ret = av_interleaved_write_frame(m_pFormatCtx_Out, pPacketVideoA);
			av_packet_free(&pPacketVideoA);
			if (ret == AVERROR(EAGAIN))
			
				continue;
			
			else if (ret == AVERROR_EOF)
			
				break;
			
			iRealPicCount++;
		
		else
		
			Sleep(10);
		
	
	///FlushEncoder(iRealPicCount);
	av_write_trailer(m_pFormatCtx_Out);
	avio_close(m_pFormatCtx_Out->pb);


void CVideoConvert::WaitFinish()

	int ret = 0;
	do
	
		if (NULL == m_hFileReadThread)
		
			break;
		
		WaitForSingleObject(m_hFileReadThread, INFINITE);

		CloseHandle(m_hFileReadThread);
		m_hFileReadThread = NULL;

		while (m_vecReadVideo.size() > 0)
		
			Sleep(1000);
		
		m_bStartConvert = false;
		WaitForSingleObject(m_hFileWriteThread, INFINITE);
		CloseHandle(m_hFileWriteThread);
		m_hFileWriteThread = NULL;
	 while (0);

ffmpeg常用的命令记录

在使用ffmpeg工具的时候有些常用的命令需要记录一下：

• 进行mp4转yuv
  命令如下：

ffmpeg -i 1.mp4 1.yuv

• 播放yuv图像命令如下：

ffplay -f rawvideo -video_size 1920x1080 -i 1.yuv

将mp4文件解码成h264文件命令如下：

ffmpeg  -i 1.mp4 -vcodec copy -bsf h264_mp4toannexb -f h264 1.h264

• 从视频前10s中提取图像，1s提取一帧,命令如下：

ffmpeg -i 1.mp4 -t 10 -r 1 -f image2 pic-%03d.jpg

• jpeg转rgba的命令如下：

ffmpeg -i