ffmpeg 本地麦克风声音和系统声音混音后，再混合本地桌面成最终的mp4文件-修正
作者：互联网
之前本人写过一篇博客：
ffmpeg 本地麦克风声音和系统声音混音后，再混合本地桌面成最终的mp4文件
但是存在着下面两个问题：
1.系统声音和麦克风对应的设备的采样率不一样，没有进行重采样，比如系统声音设备的采样率是 48000，若不进行重采样，则最终系统声音播放出来，会发现播放变慢。
2.av_read_frame采集的视频图像，在用下面两个函数编码时，avcodec_receive_packet经常返回 AVERROR(EAGAIN)，若单纯写一个桌面录制的功能，在主线程中进行抓图，编码，则大概率发现是正常的，但是如果创建一个子线程，然后在这个线程里面进行抓图，编码，就会发现出错概率很高。
我曾经将ffmpeg命令行在子线程中调用，发现录制1分钟时，生成的视频文件只有2M。
ret = avcodec_send_frame(pCodecEncodeCtx_Video, pFrameYUV);
if (ret == AVERROR(EAGAIN))
{
continue;
}
ret = avcodec_receive_packet(pCodecEncodeCtx_Video, &packet);
if (ret == AVERROR(EAGAIN))
{
continue;
}
针对第一个问题：系统声音设备的采样率是48000，麦克风的是44100，最终编码的声音要求是44100，则先将系统声音重采样成44100，再和麦克风的混音即可。
针对第二个问题：估计是自己目前学艺不精的问题，av_read_frame出现了大问题，故本人自己采取gdi抓图，不用ffmpeg的库函数av_read_frame，关于这个的具体细节，读者可以参考我写的一篇博客：ffmpeg录制桌面(自己用gdi抓图)
下面讲讲本人录制桌面视频，系统音频，麦克风音频的大致方法。
m_hAudioInnerCapture = CreateThread(NULL, 0, AudioInnerCaptureProc, this, 0, NULL);
m_hAudioInnerResample = CreateThread(NULL, 0, AudioInnerResampleProc, this, 0, NULL);
m_hAudioMicCapture = CreateThread(NULL, 0, AudioMicCaptureProc, this, 0, NULL);
m_hAudioMix = CreateThread(NULL, 0, AudioMixProc, this, 0, NULL);
m_hScreenCapture = CreateThread(NULL, 0, ScreenCaptureProc, this, 0, NULL);
m_hScreenAudioMix = CreateThread(NULL, 0, ScreenAudioMixProc, this, 0, NULL);
上面代码中，一共创建了6个线程，其中：
m_hAudioInnerCapture代码的是系统声音抓取。
m_hAudioInnerResample代表的是系统声音重采样。
m_hAudioMicCapture代表的是麦克风声音抓取。
m_hAudioMix代表的是混音，是将重采样后的系统声音和麦克风声音混合。
m_hScreenCapture代表的是桌面图像抓取
m_hScreenAudioMix代表的是桌面图像和混合后的音频进行混合，生成最终的mp4文件。
如下所示，本人定义了五个队列，其中m_pVideoFifo代表的是桌面图像，m_hScreenCapture线程往这个队列里面写数据，m_hScreenAudioMix从这个队列里面读取数据。
m_pAudioInnerFifo代表的系统声音，原始的，m_hAudioInnerCapture线程往这个队列里面写数据，m_hAudioInnerResample线程从这个队列里面读取数据，进行重采样，然后将采样后的结果放入队列
m_pAudioInnerResampleFifo。
AVFifoBuffer *m_pVideoFifo = NULL;
AVAudioFifo *m_pAudioInnerFifo = NULL;
AVAudioFifo *m_pAudioInnerResampleFifo = NULL;
AVAudioFifo *m_pAudioMicFifo = NULL;
AVAudioFifo *m_pAudioMixFifo = NULL;
下面我给出自己的代码结构：
其中appfun和log两个文件夹，没有具体业务含义，大家不用管。
CaptureScreen.cpp用于gdi抓取桌面图像。
ULinkRecord.cpp里面完成了音视频的抓取，混音，以及声音和视频混合。
FfmpegVideoAudioAndMicOneFileTest.cpp是调用方，其代码如下：
#include "ULinkRecord.h"
#include <stdio.h>
#include <conio.h>
int main()
{
ULinkRecord cULinkRecord;
cULinkRecord.SetMicName(L"麦克风 (2- Synaptics HD Audio)");
cULinkRecord.SetRecordPath("E:\\learn\\ffmpeg\\FfmpegTest\\x64\\Release");
RECT rect;
rect.left = 0;
rect.top = 0;
rect.right = 1920;
rect.bottom = 1080;
cULinkRecord.SetRecordRect(rect);
cULinkRecord.StartRecord();
Sleep(60000);
printf("begin StopRecord\n");
cULinkRecord.StopRecord();
printf("end StopRecord\n");
return 0;
}
可以看出，本次抓取的时长为1分钟。
下面再给出其他4个主要文件的内容，尽管很长，我还是觉得贴出来会比较好。
CaptureScreen.h内容如下：
#ifndef _CCAPTURE_SCREEN_HH
#define _CCAPTURE_SCREEN_HH
#include<time.h>
#include <d3d9.h>
#include <stdio.h>
#include <stdlib.h>
#include <stdarg.h>
#include <windows.h>
#include <tchar.h>
#include <winbase.h>
#include <winreg.h>
#include <Strsafe.h>
//
// ---抓屏类----
//
class CCaptureScreen
{
public:
CCaptureScreen(void);
~CCaptureScreen(void);
public:
/*-----------定义外部调用函数-----------*/
int Init(int&, int&);//初始化
BYTE* CaptureImage(); //抓取屏幕
private:
/*-----------定义内部调用函数-----------*/
void* CaptureScreenFrame(int, int, int, int);//抓屏
HCURSOR FetchCursorHandle(); //获取鼠标光标
private:
/*-----------定义私有变量-----------*/
int m_width;
int m_height;
UINT wLineLen;
DWORD dwSize;
DWORD wColSize;
//设备句柄
HDC hScreenDC;
HDC hMemDC;
//图像RGB内存缓存
PRGBTRIPLE m_hdib;
//位图头信息结构体
BITMAPINFO pbi;
HBITMAP hbm;
//鼠标光标
HCURSOR m_hSavedCursor;
};
#endif //--_CCAPTURE_SCREEN_HH
CaptureScreen.cpp内容如下：
//#include "stdafx.h"
#include "CaptureScreen.h"
CCaptureScreen::CCaptureScreen(void)
{
m_hdib = NULL;
m_hSavedCursor = NULL;
hScreenDC = NULL;
hMemDC = NULL;
hbm = NULL;
m_width = 1920;
m_height = 1080;
FetchCursorHandle();
}
//
// 释放资源
//
CCaptureScreen::~CCaptureScreen(void)
{
DeleteObject(hbm);
if (m_hdib){
free(m_hdib);
m_hdib = NULL;
}
if (hScreenDC){
::ReleaseDC(NULL, hScreenDC);
}
if (hMemDC) {
DeleteDC(hMemDC);
}
if (hbm)
{
DeleteObject(hbm);
}
}
//
// 初始化
//
int CCaptureScreen::Init(int& src_VideoWidth, int& src_VideoHeight)
{
hScreenDC = ::GetDC(GetDesktopWindow());
if (hScreenDC == NULL) return 0;
int m_nMaxxScreen = GetDeviceCaps(hScreenDC, HORZRES);
int m_nMaxyScreen = GetDeviceCaps(hScreenDC, VERTRES);
hMemDC = ::CreateCompatibleDC(hScreenDC);
if (hMemDC == NULL) return 0;
m_width = m_nMaxxScreen;
m_height = m_nMaxyScreen;
if (!m_hdib){
m_hdib = (PRGBTRIPLE)malloc(m_width * m_height * 3);//24位图像大小
}
//位图头信息结构体
pbi.bmiHeader.biSize = sizeof(BITMAPINFOHEADER);
pbi.bmiHeader.biWidth = m_width;
pbi.bmiHeader.biHeight = m_height;
pbi.bmiHeader.biPlanes = 1;
pbi.bmiHeader.biBitCount = 24;
pbi.bmiHeader.biCompression = BI_RGB;
src_VideoWidth = m_width;
src_VideoHeight = m_height;
hbm = CreateCompatibleBitmap(hScreenDC, m_width, m_height);
SelectObject(hMemDC, hbm);
wLineLen = ((m_width * 24 + 31) & 0xffffffe0) / 8;
wColSize = sizeof(RGBQUAD)* ((24 <= 8) ? 1 << 24 : 0);
dwSize = (DWORD)(UINT)wLineLen * (DWORD)(UINT)m_height;
return 1;
}
//抓取屏幕数据
BYTE* CCaptureScreen::CaptureImage()
{
VOID* alpbi = CaptureScreenFrame(0, 0, m_width, m_height);
return (BYTE*)(alpbi);
}
void* CCaptureScreen::CaptureScreenFrame(int left, int top, int width, int height)
{
if (hbm == NULL || hMemDC == NULL || hScreenDC == NULL) return NULL;
BitBlt(hMemDC, 0, 0, width, height, hScreenDC, left, top, SRCCOPY);
/*-------------------------捕获鼠标-------------------------------*/
{
POINT xPoint;
GetCursorPos(&xPoint);
HCURSOR hcur = FetchCursorHandle();
xPoint.x -= left;
xPoint.y -= top;
ICONINFO iconinfo;
BOOL ret;
ret = GetIconInfo(hcur, &iconinfo);
if (ret){
xPoint.x -= iconinfo.xHotspot;
xPoint.y -= iconinfo.yHotspot;
if (iconinfo.hbmMask) DeleteObject(iconinfo.hbmMask);
if (iconinfo.hbmColor) DeleteObject(iconinfo.hbmColor);
}
/*画鼠标*/
::DrawIcon(hMemDC, xPoint.x, xPoint.y, hcur);
}
//动态分配的内存
PRGBTRIPLE hdib = m_hdib;
if (!hdib)
return hdib;
GetDIBits(hMemDC, hbm, 0, m_height, hdib, (LPBITMAPINFO)&pbi, DIB_RGB_COLORS);
return hdib;
}
//
// 获取窗体鼠标光标
//
HCURSOR CCaptureScreen::FetchCursorHandle()
{
if (m_hSavedCursor == NULL)
{
m_hSavedCursor = GetCursor();
}
return m_hSavedCursor;
}
ULinkRecord.h的内容如下：
#pragma once
#include <string>
#include <Windows.h>
#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"
#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
class ULinkRecord
{
public:
ULinkRecord();
~ULinkRecord();
public:
void SetMicName(const wchar_t* pMicName);
void SetRecordPath(const char* pRecordPath);
void SetRecordRect(RECT rectRecord);
int StartRecord();
void StopRecord();
private:
int OpenAudioInnerCapture();
int OpenAudioMicCapture();
int OpenOutPut();
int InitFilter(const char* filter_desc);
void Clear();
private:
static DWORD WINAPI AudioInnerCaptureProc(LPVOID lpParam);
void AudioInnerCapture();
static DWORD WINAPI AudioInnerResampleProc(LPVOID lpParam);
void AudioInnerResample();
static DWORD WINAPI AudioMicCaptureProc(LPVOID lpParam);
void AudioMicCapture();
static DWORD WINAPI AudioMixProc(LPVOID lpParam);
void AudioMix();
static DWORD WINAPI ScreenCaptureProc(LPVOID lpParam);
void ScreenCapture();
static DWORD WINAPI ScreenAudioMixProc(LPVOID lpParam);
void ScreenAudioMix();
private:
std::wstring m_wstrMicName;
std::string m_strRecordPath;
std::string m_strFilePrefix;
private:
CRITICAL_SECTION m_csVideoSection;
CRITICAL_SECTION m_csAudioInnerSection;
CRITICAL_SECTION m_csAudioInnerResampleSection;
CRITICAL_SECTION m_csAudioMicSection;
CRITICAL_SECTION m_csAudioMixSection;
AVFifoBuffer *m_pVideoFifo = NULL;
AVAudioFifo *m_pAudioInnerFifo = NULL;
AVAudioFifo *m_pAudioInnerResampleFifo = NULL;
AVAudioFifo *m_pAudioMicFifo = NULL;
AVAudioFifo *m_pAudioMixFifo = NULL;
AVFormatContext *m_pFormatCtx_Out = NULL;
AVFormatContext *m_pFormatCtx_AudioInner = NULL;
AVFormatContext *m_pFormatCtx_AudioMic = NULL;
AVCodecContext *m_pReadCodecCtx_AudioInner = NULL;
AVCodecContext *m_pReadCodecCtx_AudioMic = NULL;
AVCodec *m_pReadCodec_Video = NULL;
AVCodecContext *m_pCodecEncodeCtx_Video = NULL;
AVCodecContext *m_pCodecEncodeCtx_Audio = NULL;
AVCodec *m_pCodecEncode_Audio = NULL;
SwsContext *m_pImgConvertCtx = NULL;
SwrContext *m_pAudioInnerResampleCtx = NULL;
SwrContext *m_pAudioConvertCtx = NULL;
AVFilterGraph* m_pFilterGraph = NULL;
AVFilterContext* m_pFilterCtxSrcInner = NULL;
AVFilterContext* m_pFilterCtxSrcMic = NULL;
AVFilterContext* m_pFilterCtxSink = NULL;
int m_iVideoStreamIndex = 0;
int m_iAudioStreamIndex = 0;
bool m_bRecord;
HANDLE m_hAudioInnerCapture = NULL;
HANDLE m_hAudioInnerResample = NULL;
HANDLE m_hAudioMicCapture = NULL;
HANDLE m_hAudioMix = NULL;
HANDLE m_hScreenCapture = NULL;
HANDLE m_hScreenAudioMix = NULL;
int m_iYuv420FrameSize = 0;
int m_iRecordPosX = 0;
int m_iRecordPosY = 0;
int m_iRecordWidth = 0;
int m_iRecordHeight = 0;
int m_iFrameNumber = 0;
};
ULinkRecord.cpp的内容如下：
#include "ULinkRecord.h"
#include "log/log.h"
#include "appfun/appfun.h"
#include "LocalRecord.h"
#include "CaptureScreen.h"
typedef struct BufferSourceContext {
const AVClass *bscclass;
AVFifoBuffer *fifo;
AVRational time_base; ///< time_base to set in the output link
AVRational frame_rate; ///< frame_rate to set in the output link
unsigned nb_failed_requests;
unsigned warning_limit;
/* video only */
int w, h;
enum AVPixelFormat pix_fmt;
AVRational pixel_aspect;
char *sws_param;
AVBufferRef *hw_frames_ctx;
/* audio only */
int sample_rate;
enum AVSampleFormat sample_fmt;
int channels;
uint64_t channel_layout;
char *channel_layout_str;
int got_format_from_params;
int eof;
} BufferSourceContext;
static char *dup_wchar_to_utf8(const wchar_t *w)
{
char *s = NULL;
int l = WideCharToMultiByte(CP_UTF8, 0, w, -1, 0, 0, 0, 0);
s = (char *)av_malloc(l);
if (s)
WideCharToMultiByte(CP_UTF8, 0, w, -1, s, l, 0, 0);
return s;
}
/* just pick the highest supported samplerate */
static int select_sample_rate(const AVCodec *codec)
{
const int *p;
int best_samplerate = 0;
if (!codec->supported_samplerates)
return 44100;
p = codec->supported_samplerates;
while (*p) {
if (!best_samplerate || abs(44100 - *p) < abs(44100 - best_samplerate))
best_samplerate = *p;
p++;
}
return best_samplerate;
}
/* select layout with the highest channel count */
static int select_channel_layout(const AVCodec *codec)
{
const uint64_t *p;
uint64_t best_ch_layout = 0;
int best_nb_channels = 0;
if (!codec->channel_layouts)
return AV_CH_LAYOUT_STEREO;
p = codec->channel_layouts;
while (*p) {
int nb_channels = av_get_channel_layout_nb_channels(*p);
if (nb_channels > best_nb_channels) {
best_ch_layout = *p;
best_nb_channels = nb_channels;
}
p++;
}
return best_ch_layout;
}
unsigned char clip_value(unsigned char x, unsigned char min_val, unsigned char max_val) {
if (x > max_val) {
return max_val;
}
else if (x < min_val) {
return min_val;
}
else {
return x;
}
}
//RGB to YUV420
bool RGB24_TO_YUV420(unsigned char *RgbBuf, int w, int h, unsigned char *yuvBuf)
{
unsigned char*ptrY, *ptrU, *ptrV, *ptrRGB;
memset(yuvBuf, 0, w*h * 3 / 2);
ptrY = yuvBuf;
ptrU = yuvBuf + w * h;
ptrV = ptrU + (w*h * 1 / 4);
unsigned char y, u, v, r, g, b;
for (int j = h - 1; j >= 0; j--) {
ptrRGB = RgbBuf + w * j * 3;
for (int i = 0; i < w; i++) {
b = *(ptrRGB++);
g = *(ptrRGB++);
r = *(ptrRGB++);
y = (unsigned char)((66 * r + 129 * g + 25 * b + 128) >> 8) + 16;
u = (unsigned char)((-38 * r - 74 * g + 112 * b + 128) >> 8) + 128;
v = (unsigned char)((112 * r - 94 * g - 18 * b + 128) >> 8) + 128;
*(ptrY++) = clip_value(y, 0, 255);
if (j % 2 == 0 && i % 2 == 0) {
*(ptrU++) = clip_value(u, 0, 255);
}
else {
if (i % 2 == 0) {
*(ptrV++) = clip_value(v, 0, 255);
}
}
}
}
return true;
}
ULinkRecord::ULinkRecord()
{
InitializeCriticalSection(&m_csVideoSection);
InitializeCriticalSection(&m_csAudioInnerSection);
InitializeCriticalSection(&m_csAudioInnerResampleSection);
InitializeCriticalSection(&m_csAudioMicSection);
InitializeCriticalSection(&m_csAudioMixSection);
avdevice_register_all();
}
ULinkRecord::~ULinkRecord()
{
DeleteCriticalSection(&m_csVideoSection);
DeleteCriticalSection(&m_csAudioInnerSection);
DeleteCriticalSection(&m_csAudioInnerResampleSection);
DeleteCriticalSection(&m_csAudioMicSection);
DeleteCriticalSection(&m_csAudioMixSection);
}
void ULinkRecord::SetMicName(const wchar_t* pMicName)
{
m_wstrMicName = pMicName;
}
void ULinkRecord::SetRecordPath(const char* pRecordPath)
{
m_strRecordPath = pRecordPath;
if (!m_strRecordPath.empty())
{
if (m_strRecordPath[m_strRecordPath.length() - 1] != '\\')
{
m_strRecordPath = m_strRecordPath + "\\";
}
}
}
void ULinkRecord::SetRecordRect(RECT rectRecord)
{
m_iRecordPosX = rectRecord.left;
m_iRecordPosY = rectRecord.top;
m_iRecordWidth = rectRecord.right - rectRecord.left;
m_iRecordHeight = rectRecord.bottom - rectRecord.top;
}
int ULinkRecord::StartRecord()
{
int iRet = -1;
do
{
m_pAudioConvertCtx = swr_alloc();
av_opt_set_channel_layout(m_pAudioConvertCtx, "in_channel_layout", AV_CH_LAYOUT_STEREO, 0);
av_opt_set_channel_layout(m_pAudioConvertCtx, "out_channel_layout", AV_CH_LAYOUT_STEREO, 0);
av_opt_set_int(m_pAudioConvertCtx, "in_sample_rate", 44100, 0);
av_opt_set_int(m_pAudioConvertCtx, "out_sample_rate", 44100, 0);
av_opt_set_sample_fmt(m_pAudioConvertCtx, "in_sample_fmt", AV_SAMPLE_FMT_S16, 0);
//av_opt_set_sample_fmt(audio_convert_ctx, "in_sample_fmt", AV_SAMPLE_FMT_FLTP, 0);
av_opt_set_sample_fmt(m_pAudioConvertCtx, "out_sample_fmt", AV_SAMPLE_FMT_FLTP, 0);
iRet = swr_init(m_pAudioConvertCtx);
m_pAudioInnerResampleCtx = swr_alloc();
av_opt_set_channel_layout(m_pAudioInnerResampleCtx, "in_channel_layout", AV_CH_LAYOUT_STEREO, 0);
av_opt_set_channel_layout(m_pAudioInnerResampleCtx, "out_channel_layout", AV_CH_LAYOUT_STEREO, 0);
av_opt_set_int(m_pAudioInnerResampleCtx, "in_sample_rate", 48000, 0);
av_opt_set_int(m_pAudioInnerResampleCtx, "out_sample_rate", 44100, 0);
av_opt_set_sample_fmt(m_pAudioInnerResampleCtx, "in_sample_fmt", AV_SAMPLE_FMT_S16, 0);
av_opt_set_sample_fmt(m_pAudioInnerResampleCtx, "out_sample_fmt", AV_SAMPLE_FMT_S16, 0);
iRet = swr_init(m_pAudioInnerResampleCtx);
if (OpenAudioInnerCapture() < 0)
{
break;
}
if (OpenAudioMicCapture() < 0)
{
break;
}
if (OpenOutPut() < 0)
{
break;
}
const char* filter_desc = "[in0][in1]amix=inputs=2[out]";
iRet = InitFilter(filter_desc);
if (iRet < 0)
{
break;
}
if (NULL == m_pAudioInnerResampleFifo)
{
m_pAudioInnerResampleFifo = av_audio_fifo_alloc((AVSampleFormat)m_pFormatCtx_AudioInner->streams[0]->codecpar->format,
m_pFormatCtx_AudioInner->streams[0]->codecpar->channels, 3000 * 1024);
}
if (NULL == m_pAudioInnerFifo)
{
m_pAudioInnerFifo = av_audio_fifo_alloc((AVSampleFormat)m_pFormatCtx_AudioInner->streams[0]->codecpar->format,
m_pFormatCtx_AudioInner->streams[0]->codecpar->channels, 3000 * 1024);
}
m_iYuv420FrameSize = av_image_get_buffer_size(AV_PIX_FMT_YUV420P, m_iRecordWidth, m_iRecordHeight, 1);
//申请30帧缓存
m_pVideoFifo = av_fifo_alloc(30 * m_iYuv420FrameSize);
m_hAudioInnerCapture = CreateThread(NULL, 0, AudioInnerCaptureProc, this, 0, NULL);
m_hAudioInnerResample = CreateThread(NULL, 0, AudioInnerResampleProc, this, 0, NULL);
m_hAudioMicCapture = CreateThread(NULL, 0, AudioMicCaptureProc, this, 0, NULL);
m_hAudioMix = CreateThread(NULL, 0, AudioMixProc, this, 0, NULL);
m_hScreenCapture = CreateThread(NULL, 0, ScreenCaptureProc, this, 0, NULL);
m_hScreenAudioMix = CreateThread(NULL, 0, ScreenAudioMixProc, this, 0, NULL);
iRet = 0;
} while (0);
if (0 == iRet)
{
m_bRecord = true;
}
else
{
Clear();
}
return iRet;
}
void ULinkRecord::StopRecord()
{
m_bRecord = false;
if (m_hAudioInnerCapture == NULL && m_hAudioInnerResample == NULL && m_hAudioMicCapture == NULL && m_hAudioMix == NULL && m_hScreenCapture == NULL && m_hScreenAudioMix == NULL)
{
return;
}
Sleep(1000);
HANDLE hThreads[6];
hThreads[0] = m_hAudioInnerCapture;
hThreads[1] = m_hAudioInnerResample;
hThreads[2] = m_hAudioMicCapture;
hThreads[3] = m_hAudioMix;
hThreads[4] = m_hScreenCapture;
WaitForMultipleObjects(6, hThreads, TRUE, INFINITE);
CloseHandle(m_hAudioInnerCapture);
m_hAudioInnerCapture = NULL;
CloseHandle(m_hAudioInnerResample);
m_hAudioInnerResample = NULL;
CloseHandle(m_hAudioMicCapture);
m_hAudioMicCapture = NULL;
CloseHandle(m_hAudioMix);
m_hAudioMix = NULL;
if (m_hScreenCapture != NULL)
{
CloseHandle(m_hScreenCapture);
m_hScreenCapture = NULL;
}
if (m_hScreenAudioMix != NULL)
{
CloseHandle(m_hScreenAudioMix);
m_hScreenAudioMix = NULL;
}
Clear();
}
int ULinkRecord::OpenAudioInnerCapture()
{
int iRet = -1;
do
{
//查找输入方式
const AVInputFormat *pAudioInputFmt = av_find_input_format("dshow");
//以Direct Show的方式打开设备，并将 输入方式 关联到格式上下文
//const char * psDevName = dup_wchar_to_utf8(L"audio=麦克风 (2- Synaptics HD Audio)");
char * psDevName = dup_wchar_to_utf8(L"audio=virtual-audio-capturer");
if (avformat_open_input(&m_pFormatCtx_AudioInner, psDevName, pAudioInputFmt, NULL) < 0)
{
LOG_ERROR("avformat_open_input failed, m_pFormatCtx_AudioInner");
break;
}
if (avformat_find_stream_info(m_pFormatCtx_AudioInner, NULL) < 0)
{
break;
}
if (m_pFormatCtx_AudioInner->streams[0]->codecpar->codec_type != AVMEDIA_TYPE_AUDIO)
{
LOG_ERROR("Couldn't find audio stream information");
break;
}
const AVCodec *tmpCodec = avcodec_find_decoder(m_pFormatCtx_AudioInner->streams[0]->codecpar->codec_id);
m_pReadCodecCtx_AudioInner = avcodec_alloc_context3(tmpCodec);
m_pReadCodecCtx_AudioInner->sample_rate = m_pFormatCtx_AudioInner->streams[0]->codecpar->sample_rate;
m_pReadCodecCtx_AudioInner->channel_layout = select_channel_layout(tmpCodec);
m_pReadCodecCtx_AudioInner->channels = av_get_channel_layout_nb_channels(m_pReadCodecCtx_AudioInner->channel_layout);
m_pReadCodecCtx_AudioInner->sample_fmt = (AVSampleFormat)m_pFormatCtx_AudioInner->streams[0]->codecpar->format;
//pReadCodecCtx_Audio->sample_fmt = AV_SAMPLE_FMT_FLTP;
if (0 > avcodec_open2(m_pReadCodecCtx_AudioInner, tmpCodec, NULL))
{
LOG_ERROR("avcodec_open2 failed, m_pReadCodecCtx_AudioInner");
break;
}
avcodec_parameters_from_context(m_pFormatCtx_AudioInner->streams[0]->codecpar, m_pReadCodecCtx_AudioInner);
iRet = 0;
} while (0);
if (iRet != 0)
{
if (m_pReadCodecCtx_AudioInner != NULL)
{
avcodec_free_context(&m_pReadCodecCtx_AudioInner);
m_pReadCodecCtx_AudioInner = NULL;
}
if (m_pFormatCtx_AudioInner != NULL)
{
avformat_close_input(&m_pFormatCtx_AudioInner);
m_pFormatCtx_AudioInner = NULL;
}
}
return iRet;
}
int ULinkRecord::OpenAudioMicCapture()
{
int iRet = -1;
do
{
//查找输入方式
const AVInputFormat *pAudioInputFmt = av_find_input_format("dshow");
//以Direct Show的方式打开设备，并将 输入方式 关联到格式上下文
std::wstring strDeviceName = L"audio=";
strDeviceName += m_wstrMicName;
const char * psDevName = dup_wchar_to_utf8(strDeviceName.c_str());
if (avformat_open_input(&m_pFormatCtx_AudioMic, psDevName, pAudioInputFmt, NULL) < 0)
{
LOG_ERROR("avformat_open_input failed, m_pFormatCtx_AudioMic");
break;
}
if (m_pFormatCtx_AudioMic->streams[0]->codecpar->codec_type != AVMEDIA_TYPE_AUDIO)
{
LOG_ERROR("Couldn't find audio stream information");
break;
}
const AVCodec *tmpCodec = avcodec_find_decoder(m_pFormatCtx_AudioMic->streams[0]->codecpar->codec_id);
m_pReadCodecCtx_AudioMic = avcodec_alloc_context3(tmpCodec);
m_pReadCodecCtx_AudioMic->sample_rate = select_sample_rate(tmpCodec);
m_pReadCodecCtx_AudioMic->channel_layout = select_channel_layout(tmpCodec);
m_pReadCodecCtx_AudioMic->channels = av_get_channel_layout_nb_channels(m_pReadCodecCtx_AudioMic->channel_layout);
m_pReadCodecCtx_AudioMic->sample_fmt = (AVSampleFormat)m_pFormatCtx_AudioMic->streams[0]->codecpar->format;
//pReadCodecCtx_Audio->sample_fmt = AV_SAMPLE_FMT_FLTP;
if (0 > avcodec_open2(m_pReadCodecCtx_AudioMic, tmpCodec, NULL))
{
LOG_ERROR("avcodec_open2 failed, m_pReadCodecCtx_AudioMic");
break;
}
avcodec_parameters_from_context(m_pFormatCtx_AudioMic->streams[0]->codecpar, m_pReadCodecCtx_AudioMic);
iRet = 0;
} while (0);
if (iRet != 0)
{
if (m_pReadCodecCtx_AudioMic != NULL)
{
avcodec_free_context(&m_pReadCodecCtx_AudioMic);
m_pReadCodecCtx_AudioMic = NULL;
}
if (m_pFormatCtx_AudioMic != NULL)
{
avformat_close_input(&m_pFormatCtx_AudioMic);
m_pFormatCtx_AudioMic = NULL;
}
}
return iRet;
}
int ULinkRecord::OpenOutPut()
{
std::string strFileName = m_strRecordPath;
m_strFilePrefix = strFileName + time_tSimpleString(time(NULL));
int iRet = -1;
AVStream *pAudioStream = NULL;
AVStream *pVideoStream = NULL;
do
{
std::string strFileName = m_strRecordPath;
strFileName += time_tSimpleString(time(NULL));
strFileName += ".mp4";
const char *outFileName = strFileName.c_str();
avformat_alloc_output_context2(&m_pFormatCtx_Out, NULL, NULL, outFileName);
{
m_iVideoStreamIndex = 0;
AVCodec* pCodecEncode_Video = (AVCodec *)avcodec_find_encoder(m_pFormatCtx_Out->oformat->video_codec);
m_pCodecEncodeCtx_Video = avcodec_alloc_context3(pCodecEncode_Video);
if (!m_pCodecEncodeCtx_Video)
{
LOG_ERROR("avcodec_alloc_context3 failed, m_pCodecEncodeCtx_Audio");
break;
}
pVideoStream = avformat_new_stream(m_pFormatCtx_Out, pCodecEncode_Video);
if (!pVideoStream)
{
break;
}
int frameRate = 10;
m_pCodecEncodeCtx_Video->flags |= AV_CODEC_FLAG_QSCALE;
m_pCodecEncodeCtx_Video->bit_rate = 4000000;
m_pCodecEncodeCtx_Video->rc_min_rate = 4000000;
m_pCodecEncodeCtx_Video->rc_max_rate = 4000000;
m_pCodecEncodeCtx_Video->bit_rate_tolerance = 4000000;
m_pCodecEncodeCtx_Video->time_base.den = frameRate;
m_pCodecEncodeCtx_Video->time_base.num = 1;
m_pCodecEncodeCtx_Video->width = m_iRecordWidth;
m_pCodecEncodeCtx_Video->height = m_iRecordHeight;
//pH264Encoder->pCodecCtx->frame_number = 1;
m_pCodecEncodeCtx_Video->gop_size = 12;
m_pCodecEncodeCtx_Video->max_b_frames = 0;
m_pCodecEncodeCtx_Video->thread_count = 4;
m_pCodecEncodeCtx_Video->pix_fmt = AV_PIX_FMT_YUV420P;
m_pCodecEncodeCtx_Video->codec_id = AV_CODEC_ID_H264;
m_pCodecEncodeCtx_Video->codec_type = AVMEDIA_TYPE_VIDEO;
av_opt_set(m_pCodecEncodeCtx_Video->priv_data, "b-pyramid", "none", 0);
av_opt_set(m_pCodecEncodeCtx_Video->priv_data, "preset", "superfast", 0);
av_opt_set(m_pCodecEncodeCtx_Video->priv_data, "tune", "zerolatency", 0);
if (m_pFormatCtx_Out->oformat->flags & AVFMT_GLOBALHEADER)
m_pCodecEncodeCtx_Video->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
if (avcodec_open2(m_pCodecEncodeCtx_Video, pCodecEncode_Video, 0) < 0)
{
//编码器打开失败，退出程序
LOG_ERROR("avcodec_open2 failed, m_pCodecEncodeCtx_Audio");
break;
}
}
if (m_pFormatCtx_AudioInner->streams[0]->codecpar->codec_type == AVMEDIA_TYPE_AUDIO && m_pFormatCtx_AudioMic->streams[0]->codecpar->codec_type == AVMEDIA_TYPE_AUDIO)
{
pAudioStream = avformat_new_stream(m_pFormatCtx_Out, NULL);
m_iAudioStreamIndex = 1;
m_pCodecEncode_Audio = (AVCodec *)avcodec_find_encoder(m_pFormatCtx_Out->oformat->audio_codec);
m_pCodecEncodeCtx_Audio = avcodec_alloc_context3(m_pCodecEncode_Audio);
if (!m_pCodecEncodeCtx_Audio)
{
LOG_ERROR("avcodec_alloc_context3 failed, m_pCodecEncodeCtx_Audio");
break;
}
//pCodecEncodeCtx_Audio->codec_id = pFormatCtx_Out->oformat->audio_codec;
m_pCodecEncodeCtx_Audio->sample_fmt = m_pCodecEncode_Audio->sample_fmts ? m_pCodecEncode_Audio->sample_fmts[0] : AV_SAMPLE_FMT_FLTP;
m_pCodecEncodeCtx_Audio->bit_rate = 64000;
m_pCodecEncodeCtx_Audio->sample_rate = 44100;
m_pCodecEncodeCtx_Audio->channel_layout = AV_CH_LAYOUT_STEREO;
m_pCodecEncodeCtx_Audio->channels = av_get_channel_layout_nb_channels(m_pCodecEncodeCtx_Audio->channel_layout);
AVRational timeBase;
timeBase.den = m_pCodecEncodeCtx_Audio->sample_rate;
timeBase.num = 1;
pAudioStream->time_base = timeBase;
if (avcodec_open2(m_pCodecEncodeCtx_Audio, m_pCodecEncode_Audio, 0) < 0)
{
//编码器打开失败，退出程序
LOG_ERROR("avcodec_open2 failed, m_pCodecEncodeCtx_Audio");
break;
}
}
if (!(m_pFormatCtx_Out->oformat->flags & AVFMT_NOFILE))
{
if (avio_open(&m_pFormatCtx_Out->pb, outFileName, AVIO_FLAG_WRITE) < 0)
{
LOG_ERROR("avio_open failed, m_pFormatCtx_Out->pb");
break;
}
}
avcodec_parameters_from_context(pVideoStream->codecpar, m_pCodecEncodeCtx_Video);
avcodec_parameters_from_context(pAudioStream->codecpar, m_pCodecEncodeCtx_Audio);
if (avformat_write_header(m_pFormatCtx_Out, NULL) < 0)
{
LOG_ERROR("avio_open avformat_write_header,m_pFormatCtx_Out");
break;
}
iRet = 0;
} while (0);
if (iRet != 0)
{
if (m_pCodecEncodeCtx_Video != NULL)
{
avcodec_free_context(&m_pCodecEncodeCtx_Video);
m_pCodecEncodeCtx_Video = NULL;
}
if (m_pCodecEncodeCtx_Audio != NULL)
{
avcodec_free_context(&m_pCodecEncodeCtx_Audio);
m_pCodecEncodeCtx_Audio = NULL;
}
if (m_pFormatCtx_Out != NULL)
{
avformat_free_context(m_pFormatCtx_Out);
m_pFormatCtx_Out = NULL;
}
}
return iRet;
}
int ULinkRecord::InitFilter(const char* filter_desc)
{
char args_inner[512];
const char* pad_name_inner = "in0";
char args_mic[512];
const char* pad_name_mic = "in1";
AVFilter* filter_src_spk = (AVFilter *)avfilter_get_by_name("abuffer");
AVFilter* filter_src_mic = (AVFilter *)avfilter_get_by_name("abuffer");
AVFilter* filter_sink = (AVFilter *)avfilter_get_by_name("abuffersink");
AVFilterInOut* filter_output_inner = avfilter_inout_alloc();
AVFilterInOut* filter_output_mic = avfilter_inout_alloc();
AVFilterInOut* filter_input = avfilter_inout_alloc();
m_pFilterGraph = avfilter_graph_alloc();
/*sprintf_s(args_inner, sizeof(args_inner), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%I64x",
m_pReadCodecCtx_AudioInner->time_base.num,
m_pReadCodecCtx_AudioInner->time_base.den,
m_pReadCodecCtx_AudioInner->sample_rate,
av_get_sample_fmt_name((AVSampleFormat)m_pReadCodecCtx_AudioInner->sample_fmt),
m_pReadCodecCtx_AudioInner->channel_layout);*/
sprintf_s(args_inner, sizeof(args_inner), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%I64x",
m_pReadCodecCtx_AudioMic->time_base.num,
m_pReadCodecCtx_AudioMic->time_base.den,
m_pReadCodecCtx_AudioMic->sample_rate,
av_get_sample_fmt_name((AVSampleFormat)m_pReadCodecCtx_AudioMic->sample_fmt),
m_pReadCodecCtx_AudioMic->channel_layout);
sprintf_s(args_mic, sizeof(args_mic), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%I64x",
m_pReadCodecCtx_AudioMic->time_base.num,
m_pReadCodecCtx_AudioMic->time_base.den,
m_pReadCodecCtx_AudioMic->sample_rate,
av_get_sample_fmt_name((AVSampleFormat)m_pReadCodecCtx_AudioMic->sample_fmt),
m_pReadCodecCtx_AudioMic->channel_layout);
//sprintf_s(args_spk, sizeof(args_spk), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%I64x", _fmt_ctx_out->streams[_index_a_out]->codec->time_base.num, _fmt_ctx_out->streams[_index_a_out]->codec->time_base.den, _fmt_ctx_out->streams[_index_a_out]->codec->sample_rate, av_get_sample_fmt_name(_fmt_ctx_out->streams[_index_a_out]->codec->sample_fmt), _fmt_ctx_out->streams[_index_a_out]->codec->channel_layout);
//sprintf_s(args_mic, sizeof(args_mic), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%I64x", _fmt_ctx_out->streams[_index_a_out]->codec->time_base.num, _fmt_ctx_out->streams[_index_a_out]->codec->time_base.den, _fmt_ctx_out->streams[_index_a_out]->codec->sample_rate, av_get_sample_fmt_name(_fmt_ctx_out->streams[_index_a_out]->codec->sample_fmt), _fmt_ctx_out->streams[_index_a_out]->codec->channel_layout);
int ret = 0;
ret = avfilter_graph_create_filter(&m_pFilterCtxSrcInner, filter_src_spk, pad_name_inner, args_inner, NULL, m_pFilterGraph);
if (ret < 0)
{
printf("Filter: failed to call avfilter_graph_create_filter -- src inner\n");
return -1;
}
ret = avfilter_graph_create_filter(&m_pFilterCtxSrcMic, filter_src_mic, pad_name_mic, args_mic, NULL, m_pFilterGraph);
if (ret < 0)
{
printf("Filter: failed to call avfilter_graph_create_filter -- src mic\n");
return -1;
}
ret = avfilter_graph_create_filter(&m_pFilterCtxSink, filter_sink, "out", NULL, NULL, m_pFilterGraph);
if (ret < 0)
{
printf("Filter: failed to call avfilter_graph_create_filter -- sink\n");
return -1;
}
AVCodecContext* encodec_ctx = m_pCodecEncodeCtx_Audio;
//ret = av_opt_set_bin(_filter_ctx_sink, "sample_fmts", (uint8_t*)&encodec_ctx->sample_fmt, sizeof(encodec_ctx->sample_fmt), AV_OPT_SEARCH_CHILDREN);
ret = av_opt_set_bin(m_pFilterCtxSink, "sample_fmts", (uint8_t*)&m_pReadCodecCtx_AudioInner->sample_fmt, sizeof(m_pReadCodecCtx_AudioInner->sample_fmt), AV_OPT_SEARCH_CHILDREN);
if (ret < 0)
{
printf("Filter: failed to call av_opt_set_bin -- sample_fmts\n");
return -1;
}
ret = av_opt_set_bin(m_pFilterCtxSink, "channel_layouts", (uint8_t*)&encodec_ctx->channel_layout, sizeof(encodec_ctx->channel_layout), AV_OPT_SEARCH_CHILDREN);
if (ret < 0)
{
printf("Filter: failed to call av_opt_set_bin -- channel_layouts\n");
return -1;
}
ret = av_opt_set_bin(m_pFilterCtxSink, "sample_rates", (uint8_t*)&encodec_ctx->sample_rate, sizeof(encodec_ctx->sample_rate), AV_OPT_SEARCH_CHILDREN);
if (ret < 0)
{
printf("Filter: failed to call av_opt_set_bin -- sample_rates\n");
return -1;
}
filter_output_inner->name = av_strdup(pad_name_inner);
filter_output_inner->filter_ctx = m_pFilterCtxSrcInner;
filter_output_inner->pad_idx = 0;
filter_output_inner->next = filter_output_mic;
filter_output_mic->name = av_strdup(pad_name_mic);
filter_output_mic->filter_ctx = m_pFilterCtxSrcMic;
filter_output_mic->pad_idx = 0;
filter_output_mic->next = NULL;
filter_input->name = av_strdup("out");
filter_input->filter_ctx = m_pFilterCtxSink;
filter_input->pad_idx = 0;
filter_input->next = NULL;
AVFilterInOut* filter_outputs[2];
filter_outputs[0] = filter_output_inner;
filter_outputs[1] = filter_output_mic;
ret = avfilter_graph_parse_ptr(m_pFilterGraph, filter_desc, &filter_input, filter_outputs, NULL);
if (ret < 0)
{
printf("Filter: failed to call avfilter_graph_parse_ptr\n");
return -1;
}
ret = avfilter_graph_config(m_pFilterGraph, NULL);
if (ret < 0)
{
printf("Filter: failed to call avfilter_graph_config\n");
return -1;
}
avfilter_inout_free(&filter_input);
av_free(filter_src_spk);
av_free(filter_src_mic);
avfilter_inout_free(filter_outputs);
//av_free(filter_outputs);
char* temp = avfilter_graph_dump(m_pFilterGraph, NULL);
printf("%s\n", temp);
return 0;
}
void ULinkRecord::Clear()
{
if (m_pReadCodecCtx_AudioInner)
{
avcodec_free_context(&m_pReadCodecCtx_AudioInner);
m_pReadCodecCtx_AudioInner = NULL;
}
if (m_pReadCodecCtx_AudioMic != NULL)
{
avcodec_free_context(&m_pReadCodecCtx_AudioMic);
m_pReadCodecCtx_AudioMic = NULL;
}
if (m_pCodecEncodeCtx_Audio != NULL)
{
avcodec_free_context(&m_pCodecEncodeCtx_Audio);
m_pCodecEncodeCtx_Audio = NULL;
}
if (m_pAudioConvertCtx != NULL)
{
swr_free(&m_pAudioConvertCtx);
m_pAudioConvertCtx = NULL;
}
if (m_pAudioInnerFifo != NULL)
{
av_audio_fifo_free(m_pAudioInnerFifo);
m_pAudioInnerFifo = NULL;
}
if (m_pAudioInnerResampleFifo != NULL)
{
av_audio_fifo_free(m_pAudioInnerResampleFifo);
m_pAudioInnerResampleFifo = NULL;
}
if (m_pAudioMicFifo != NULL)
{
av_audio_fifo_free(m_pAudioMicFifo);
m_pAudioMicFifo = NULL;
}
if (m_pAudioMixFifo != NULL)
{
av_audio_fifo_free(m_pAudioMixFifo);
m_pAudioMixFifo = NULL;
}
if (m_pFormatCtx_AudioInner != NULL)
{
avformat_close_input(&m_pFormatCtx_AudioInner);
m_pFormatCtx_AudioInner = NULL;
}
if (m_pFormatCtx_AudioMic != NULL)
{
avformat_close_input(&m_pFormatCtx_AudioMic);
m_pFormatCtx_AudioMic = NULL;
}
if (m_pFormatCtx_Out != NULL)
{
avformat_free_context(m_pFormatCtx_Out);
m_pFormatCtx_Out = NULL;
}
if (m_hScreenCapture != NULL)
{
CloseHandle(m_hScreenCapture);
m_hScreenCapture = NULL;
}
if (m_hScreenAudioMix != NULL)
{
CloseHandle(m_hScreenAudioMix);
m_hScreenAudioMix = NULL;
}
}
DWORD WINAPI ULinkRecord::AudioInnerCaptureProc(LPVOID lpParam)
{
ULinkRecord *pULinkRecord = (ULinkRecord *)lpParam;
if (pULinkRecord != NULL)
{
pULinkRecord->AudioInnerCapture();
}
return 0;
}
void ULinkRecord::AudioInnerCapture()
{
AVFrame *pFrame;
pFrame = av_frame_alloc();
AVPacket packet = { 0 };
int ret = 0;
while (m_bRecord)
{
av_packet_unref(&packet);
if (av_read_frame(m_pFormatCtx_AudioInner, &packet) < 0)
{
continue;
}
ret = avcodec_send_packet(m_pReadCodecCtx_AudioInner, &packet);
if (ret >= 0)
{
ret = avcodec_receive_frame(m_pReadCodecCtx_AudioInner, pFrame);
if (ret == AVERROR(EAGAIN))
{
continue;
}
else if (ret == AVERROR_EOF)
{
break;
}
else if (ret < 0) {
fprintf(stderr, "Error during decoding\n");
exit(1);
}
int buf_space = av_audio_fifo_space(m_pAudioInnerFifo);
if (buf_space >= pFrame->nb_samples)
{
//AudioSection
EnterCriticalSection(&m_csAudioInnerSection);
ret = av_audio_fifo_write(m_pAudioInnerFifo, (void **)pFrame->data, pFrame->nb_samples);
LeaveCriticalSection(&m_csAudioInnerSection);
}
av_packet_unref(&packet);
}
}
av_frame_free(&pFrame);
}
DWORD WINAPI ULinkRecord::AudioInnerResampleProc(LPVOID lpParam)
{
ULinkRecord *pULinkRecord = (ULinkRecord *)lpParam;
if (pULinkRecord != NULL)
{
pULinkRecord->AudioInnerResample();
}
return 0;
}
void ULinkRecord::AudioInnerResample()
{
int ret = 0;
while (1)
{
if (av_audio_fifo_size(m_pAudioInnerFifo) >=
(m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size > 0 ? m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size : 1024))
{
AVFrame *frame_audio_inner = NULL;
frame_audio_inner = av_frame_alloc();
frame_audio_inner->nb_samples = m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size > 0 ? m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size : 1024;
frame_audio_inner->channel_layout = m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->channel_layout;
frame_audio_inner->format = m_pFormatCtx_AudioInner->streams[0]->codecpar->format;
frame_audio_inner->sample_rate = m_pFormatCtx_AudioInner->streams[0]->codecpar->sample_rate;
av_frame_get_buffer(frame_audio_inner, 0);
EnterCriticalSection(&m_csAudioInnerSection);
int readcount = av_audio_fifo_read(m_pAudioInnerFifo, (void **)frame_audio_inner->data,
(m_pFormatCtx_AudioInner->streams[0]->codecpar->frame_size > 0 ? m_pFormatCtx_AudioInner->streams[0]->codecpar->frame_size : 1024));
LeaveCriticalSection(&m_csAudioInnerSection);
AVFrame *frame_audio_inner_resample = NULL;
frame_audio_inner_resample = av_frame_alloc();
//int iDelaySamples = swr_get_delay(audio_convert_ctx, frame_mic->sample_rate);
int iDelaySamples = 0;
//int dst_nb_samples = av_rescale_rnd(iDelaySamples + frame_mic->nb_samples, frame_mic->sample_rate, pCodecEncodeCtx_Audio->sample_rate, AVRounding(1));
int dst_nb_samples = av_rescale_rnd(iDelaySamples + frame_audio_inner->nb_samples, m_pCodecEncodeCtx_Audio->sample_rate, frame_audio_inner->sample_rate, AV_ROUND_UP);
frame_audio_inner_resample->nb_samples = m_pCodecEncodeCtx_Audio->frame_size;
frame_audio_inner_resample->channel_layout = m_pCodecEncodeCtx_Audio->channel_layout;
frame_audio_inner_resample->format = m_pFormatCtx_AudioInner->streams[0]->codecpar->format;
frame_audio_inner_resample->sample_rate = m_pCodecEncodeCtx_Audio->sample_rate;
av_frame_get_buffer(frame_audio_inner_resample, 0);
uint8_t* out_buffer = (uint8_t*)frame_audio_inner_resample->data[0];
int nb = swr_convert(m_pAudioInnerResampleCtx, &out_buffer, dst_nb_samples, (const uint8_t**)frame_audio_inner->data, frame_audio_inner->nb_samples);
//if (av_audio_fifo_space(fifo_audio_resample) >= pFrame->nb_samples)
{
EnterCriticalSection(&m_csAudioInnerResampleSection);
ret = av_audio_fifo_write(m_pAudioInnerResampleFifo, (void **)frame_audio_inner_resample->data, dst_nb_samples);
LeaveCriticalSection(&m_csAudioInnerResampleSection);
}
av_frame_free(&frame_audio_inner);
av_frame_free(&frame_audio_inner_resample);
if (!m_bRecord)
{
if (av_audio_fifo_size(m_pAudioInnerFifo) < 1024)
{
break;
}
}
}
else
{
if (!m_bRecord)
{
break;
}
}
}
}
DWORD WINAPI ULinkRecord::AudioMicCaptureProc(LPVOID lpParam)
{
ULinkRecord *pULinkRecord = (ULinkRecord *)lpParam;
if (pULinkRecord != NULL)
{
pULinkRecord->AudioMicCapture();
}
return 0;
}
void ULinkRecord::AudioMicCapture()
{
AVFrame *pFrame;
pFrame = av_frame_alloc();
AVPacket packet = { 0 };
int ret = 0;
while (m_bRecord)
{
av_packet_unref(&packet);
if (av_read_frame(m_pFormatCtx_AudioMic, &packet) < 0)
{
continue;
}
ret = avcodec_send_packet(m_pReadCodecCtx_AudioMic, &packet);
if (ret >= 0)
{
ret = avcodec_receive_frame(m_pReadCodecCtx_AudioMic, pFrame);
if (ret == AVERROR(EAGAIN))
{
continue;
}
else if (ret == AVERROR_EOF)
{
break;
}
else if (ret < 0) {
break;
}
if (NULL == m_pAudioMicFifo)
{
m_pAudioMicFifo = av_audio_fifo_alloc((AVSampleFormat)m_pFormatCtx_AudioMic->streams[0]->codecpar->format,
m_pFormatCtx_AudioMic->streams[0]->codecpar->channels, 3000 * pFrame->nb_samples);
}
int buf_space = av_audio_fifo_space(m_pAudioMicFifo);
if (buf_space >= pFrame->nb_samples)
{
EnterCriticalSection(&m_csAudioMicSection);
ret = av_audio_fifo_write(m_pAudioMicFifo, (void **)pFrame->data, pFrame->nb_samples);
LeaveCriticalSection(&m_csAudioMicSection);
}
av_packet_unref(&packet);
}
}
av_frame_free(&pFrame);
}
DWORD WINAPI ULinkRecord::AudioMixProc(LPVOID lpParam)
{
ULinkRecord *pULinkRecord = (ULinkRecord *)lpParam;
if (pULinkRecord != NULL)
{
pULinkRecord->AudioMix();
}
return 0;
}
void ULinkRecord::AudioMix()
{
int ret = 0;
int iAudioFrameMixedIndex = 0;
AVFrame *frame_audio_inner_resample = NULL;
frame_audio_inner_resample = av_frame_alloc();
AVFrame *frame_audio_mic = NULL;
frame_audio_mic = av_frame_alloc();
if (NULL == m_pAudioMixFifo)
{
m_pAudioMixFifo = av_audio_fifo_alloc((AVSampleFormat)m_pFormatCtx_AudioInner->streams[0]->codecpar->format,
m_pFormatCtx_AudioInner->streams[0]->codecpar->channels, 3000 * 1024);
}
while (m_bRecord)
{
if (NULL == m_pAudioInnerResampleFifo)
{
continue;
}
if (NULL == m_pAudioMicFifo)
{
continue;
}
int fifo_inner_resample_size = av_audio_fifo_size(m_pAudioInnerResampleFifo);
int fifo_mic_size = av_audio_fifo_size(m_pAudioMicFifo);
//int frame_inner_min_size = pReadCodecCtx_Audio->frame_size;
//int frame_mic_min_size = pReadCodecCtx_AudioMic->frame_size;
int frame_inner_resample_min_size = 1024;
int frame_mic_min_size = 1024;
if (fifo_inner_resample_size >= frame_inner_resample_min_size && fifo_mic_size >= frame_mic_min_size)
{
frame_audio_inner_resample->nb_samples = frame_inner_resample_min_size;
frame_audio_inner_resample->channel_layout = 3;
frame_audio_inner_resample->format = m_pFormatCtx_AudioInner->streams[0]->codecpar->format;
frame_audio_inner_resample->sample_rate = m_pFormatCtx_AudioMic->streams[0]->codecpar->sample_rate;
av_frame_get_buffer(frame_audio_inner_resample, 0);
frame_audio_mic->nb_samples = frame_mic_min_size;
frame_audio_mic->channel_layout = 3;
frame_audio_mic->format = m_pFormatCtx_AudioMic->streams[0]->codecpar->format;
frame_audio_mic->sample_rate = m_pFormatCtx_AudioMic->streams[0]->codecpar->sample_rate;
av_frame_get_buffer(frame_audio_mic, 0);
EnterCriticalSection(&m_csAudioInnerResampleSection);
int readcount = av_audio_fifo_read(m_pAudioInnerResampleFifo, (void **)frame_audio_inner_resample->data, frame_inner_resample_min_size);
LeaveCriticalSection(&m_csAudioInnerResampleSection);
EnterCriticalSection(&m_csAudioMicSection);
readcount = av_audio_fifo_read(m_pAudioMicFifo, (void**)frame_audio_mic->data, frame_mic_min_size);
LeaveCriticalSection(&m_csAudioMicSection);
frame_audio_inner_resample->pts = iAudioFrameMixedIndex * m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size;
frame_audio_mic->pts = iAudioFrameMixedIndex * m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size;
BufferSourceContext* s = (BufferSourceContext*)m_pFilterCtxSrcInner->priv;
bool b1 = (s->sample_fmt != frame_audio_inner_resample->format);
bool b2 = (s->sample_rate != frame_audio_inner_resample->sample_rate);
bool b3 = (s->channel_layout != frame_audio_inner_resample->channel_layout);
bool b4 = (s->channels != frame_audio_inner_resample->channels);
ret = av_buffersrc_add_frame(m_pFilterCtxSrcInner, frame_audio_inner_resample);
if (ret < 0)
{
printf("Mixer: failed to call av_buffersrc_add_frame (speaker)\n");
break;
}
ret = av_buffersrc_add_frame(m_pFilterCtxSrcMic, frame_audio_mic);
if (ret < 0)
{
printf("Mixer: failed to call av_buffersrc_add_frame (microphone)\n");
break;
}
while (1)
{
AVFrame* pFrame_out = av_frame_alloc();
ret = av_buffersink_get_frame_flags(m_pFilterCtxSink, pFrame_out, 0);
if (ret < 0)
{
av_frame_free(&pFrame_out);
//printf("Mixer: failed to call av_buffersink_get_frame_flags\n");
break;
}
iAudioFrameMixedIndex++;
EnterCriticalSection(&m_csAudioMixSection);
ret = av_audio_fifo_write(m_pAudioMixFifo, (void **)pFrame_out->data, pFrame_out->nb_samples);
LeaveCriticalSection(&m_csAudioMixSection);
av_frame_free(&pFrame_out);
}
}
}
av_frame_free(&frame_audio_inner_resample);
av_frame_free(&frame_audio_mic);
}
DWORD WINAPI ULinkRecord::ScreenCaptureProc(LPVOID lpParam)
{
ULinkRecord *pULinkRecord = (ULinkRecord *)lpParam;
if (pULinkRecord != NULL)
{
pULinkRecord->ScreenCapture();
}
return 0;
}
void ULinkRecord::ScreenCapture()
{
CCaptureScreen* ccs = new CCaptureScreen();
int width = 0;
int height = 0;
ccs->Init(width, height);
AVFrame *pFrameYUV = av_frame_alloc();
pFrameYUV->format = AV_PIX_FMT_YUV420P;
pFrameYUV->width = width;
pFrameYUV->height = height;
int y_size = m_iRecordWidth * m_iRecordHeight;
int frame_size = av_image_get_buffer_size(AV_PIX_FMT_YUV420P, m_iRecordWidth, m_iRecordHeight, 1);
BYTE* out_buffer_yuv420 = new BYTE[frame_size];
av_image_fill_arrays(pFrameYUV->data, pFrameYUV->linesize, out_buffer_yuv420, AV_PIX_FMT_YUV420P, m_iRecordWidth, m_iRecordHeight, 1);
AVPacket* packet = av_packet_alloc();
m_iFrameNumber = 0;
int ret = 0;
DWORD dwBeginTime = ::GetTickCount();
while(m_bRecord)
{
BYTE* frameimage = ccs->CaptureImage();
RGB24_TO_YUV420(frameimage, width, height, out_buffer_yuv420);
//Sframe->pkt_dts = frame->pts = frameNumber * avCodecCtx_Out->time_base.num * avStream->time_base.den / (avCodecCtx_Out->time_base.den * avStream->time_base.num);
pFrameYUV->pkt_dts = pFrameYUV->pts = av_rescale_q_rnd(m_iFrameNumber, m_pCodecEncodeCtx_Video->time_base, m_pFormatCtx_Out->streams[m_iVideoStreamIndex]->time_base, (AVRounding)(AV_ROUND_NEAR_INF | AV_ROUND_PASS_MINMAX));
pFrameYUV->pkt_duration = 0;
pFrameYUV->pkt_pos = -1;
if (av_fifo_space(m_pVideoFifo) >= frame_size)
{
EnterCriticalSection(&m_csVideoSection);
av_fifo_generic_write(m_pVideoFifo, pFrameYUV->data[0], y_size, NULL);
av_fifo_generic_write(m_pVideoFifo, pFrameYUV->data[1], y_size / 4, NULL);
av_fifo_generic_write(m_pVideoFifo, pFrameYUV->data[2], y_size / 4, NULL);
LeaveCriticalSection(&m_csVideoSection);
m_iFrameNumber++;
DWORD dwCurrentTime = ::GetTickCount();
int dwPassedMillSeconds = dwCurrentTime - dwBeginTime;
int dwDiff = m_iFrameNumber * 100 - dwPassedMillSeconds;
if (dwDiff > 0)
{
Sleep(dwDiff);
}
}
}
av_packet_free(&packet);
av_frame_free(&pFrameYUV);
delete[] out_buffer_yuv420;
}
DWORD WINAPI ULinkRecord::ScreenAudioMixProc(LPVOID lpParam)
{
ULinkRecord *pULinkRecord = (ULinkRecord *)lpParam;
if (pULinkRecord != NULL)
{
pULinkRecord->ScreenAudioMix();
}
return 0;
}
void ULinkRecord::ScreenAudioMix()
{
int ret = 0;
int cur_pts_v = 0;
int cur_pts_a = 0;
AVFrame *pFrameYUVInMain = av_frame_alloc();
uint8_t *out_buffer_yuv420 = (uint8_t *)av_malloc(m_iYuv420FrameSize);
av_image_fill_arrays(pFrameYUVInMain->data, pFrameYUVInMain->linesize, out_buffer_yuv420, AV_PIX_FMT_YUV420P, m_iRecordWidth, m_iRecordHeight, 1);
int AudioFrameIndex_mic = 1;
AVPacket packet = { 0 };
int iPicCount = 0;
while (m_bRecord)
{
if (NULL == m_pVideoFifo)
{
continue;
}
if (NULL == m_pAudioMixFifo)
{
continue;
}
if (av_compare_ts(cur_pts_v, m_pFormatCtx_Out->streams[m_iVideoStreamIndex]->time_base,
cur_pts_a, m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->time_base) <= 0)
{
if (av_fifo_size(m_pVideoFifo) >= m_iYuv420FrameSize)
{
EnterCriticalSection(&m_csVideoSection);
av_fifo_generic_read(m_pVideoFifo, out_buffer_yuv420, m_iYuv420FrameSize, NULL);
LeaveCriticalSection(&m_csVideoSection);
pFrameYUVInMain->pkt_dts = pFrameYUVInMain->pts = av_rescale_q_rnd(iPicCount, m_pCodecEncodeCtx_Video->time_base, m_pFormatCtx_Out->streams[m_iVideoStreamIndex]->time_base, (AVRounding)(AV_ROUND_NEAR_INF | AV_ROUND_PASS_MINMAX));
pFrameYUVInMain->pkt_duration = 0;
pFrameYUVInMain->pkt_pos = -1;
pFrameYUVInMain->width = m_iRecordWidth;
pFrameYUVInMain->height = m_iRecordHeight;
pFrameYUVInMain->format = AV_PIX_FMT_YUV420P;
cur_pts_v = packet.pts;
ret = avcodec_send_frame(m_pCodecEncodeCtx_Video, pFrameYUVInMain);
ret = avcodec_receive_packet(m_pCodecEncodeCtx_Video, &packet);
if (packet.size > 0)
{
//av_packet_rescale_ts(packet, avCodecCtx_Out->time_base, avStream->time_base);
av_write_frame(m_pFormatCtx_Out, &packet);
iPicCount++;
}
}
}
else
{
if (av_audio_fifo_size(m_pAudioMixFifo) >=
(m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size > 0 ? m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size : 1024))
{
AVFrame *frame_mix = NULL;
frame_mix = av_frame_alloc();
frame_mix->nb_samples = m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size > 0 ? m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size : 1024;
frame_mix->channel_layout = m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->channel_layout;
//frame_mix->format = pFormatCtx_Out->streams[iAudioStreamIndex]->codecpar->format;
frame_mix->format = 1;
frame_mix->sample_rate = m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->sample_rate;
av_frame_get_buffer(frame_mix, 0);
EnterCriticalSection(&m_csAudioMixSection);
int readcount = av_audio_fifo_read(m_pAudioMixFifo, (void **)frame_mix->data,
(m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size > 0 ? m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size : 1024));
LeaveCriticalSection(&m_csAudioMixSection);
AVPacket pkt_out_mic = { 0 };
pkt_out_mic.data = NULL;
pkt_out_mic.size = 0;
frame_mix->pts = AudioFrameIndex_mic * m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size;
AVFrame *frame_mic_encode = NULL;
frame_mic_encode = av_frame_alloc();
frame_mic_encode->nb_samples = m_pCodecEncodeCtx_Audio->frame_size;
frame_mic_encode->channel_layout = m_pCodecEncodeCtx_Audio->channel_layout;
frame_mic_encode->format = m_pCodecEncodeCtx_Audio->sample_fmt;
frame_mic_encode->sample_rate = m_pCodecEncodeCtx_Audio->sample_rate;
av_frame_get_buffer(frame_mic_encode, 0);
int dst_nb_samples = av_rescale_rnd(swr_get_delay(m_pAudioConvertCtx, frame_mix->sample_rate) + frame_mix->nb_samples, frame_mix->sample_rate, frame_mix->sample_rate, AVRounding(1));
//uint8_t *audio_buf = NULL;
uint8_t *audio_buf[2] = { 0 };
audio_buf[0] = (uint8_t *)frame_mic_encode->data[0];
audio_buf[1] = (uint8_t *)frame_mic_encode->data[1];
int nb = swr_convert(m_pAudioConvertCtx, audio_buf, dst_nb_samples, (const uint8_t**)frame_mix->data, frame_mix->nb_samples);
ret = avcodec_send_frame(m_pCodecEncodeCtx_Audio, frame_mic_encode);
ret = avcodec_receive_packet(m_pCodecEncodeCtx_Audio, &pkt_out_mic);
if (ret == AVERROR(EAGAIN))
{
continue;
}
av_frame_free(&frame_mix);
av_frame_free(&frame_mic_encode);
{
pkt_out_mic.stream_index = m_iAudioStreamIndex;
pkt_out_mic.pts = AudioFrameIndex_mic * m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size;
pkt_out_mic.dts = AudioFrameIndex_mic * m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size;
pkt_out_mic.duration = m_pFormatCtx_Out->streams[m_iAudioStreamIndex]->codecpar->frame_size;
cur_pts_a = pkt_out_mic.pts;
av_write_frame(m_pFormatCtx_Out, &pkt_out_mic);
//int ret2 = av_interleaved_write_frame(m_pFormatCtx_Out, &pkt_out_mic);
av_packet_unref(&pkt_out_mic);
}
AudioFrameIndex_mic++;
}
}
}
Sleep(100);
av_write_trailer(m_pFormatCtx_Out);
avio_close(m_pFormatCtx_Out->pb);
av_frame_free(&pFrameYUVInMain);
}
标签：audio,ffmpeg,int,frame,mp4,本地,av,NULL,pFormatCtx 来源： https://blog.csdn.net/tusong86/article/details/121863390