我是一名大学生，现在正在学习FFmpeg。
我写了一个软件，可以用FFmpeg录制桌面和音频（'virtual-audio-capturer'），现在正在编写音频和视频同步。 我遇到了录像播放速度过快的一些问题。
当我在互联网上寻求音频和视频同步帮助时，我发现了一个计算PTS的公式：
使用此公式时，我发现一种现象。
1。帧速率越高，视频播放速度越快。
2。帧频越慢，视频播放越快。
我还发现帧速率为10时，视频播放速度将是正确的。
我已经考虑了这个问题三天了。我真的希望有人能帮助我解决这个问题。
我非常感谢您的帮助。
#include "stdafx.h"
#ifdef __cplusplus
extern "C"
{
#endif
#include "libavcodec/avcodec.h"
#include "libavformat/avformat.h"
#include "libswscale/swscale.h"
#include "libavdevice/avdevice.h"
#include "libavutil/audio_fifo.h"
#include "libavfilter/buffersink.h"
#include "libavfilter/buffersrc.h"
#include "libavutil/imgutils.h"
#include "libavutil/mathematics.h"
#include "libavutil/samplefmt.h"
#include "libavutil/time.h"
#include "libavutil/opt.h"
#include "libavutil/pixdesc.h"
#include "libavutil/file.h"
#include "libavutil/mem.h"
#include "libavutil/frame.h"
#include "libavfilter/avfilter.h"
#include "libswresample/swresample.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, "avfilter.lib")
#pragma comment(lib, "postproc.lib")
#pragma comment(lib, "swresample.lib")
#pragma comment(lib, "swscale.lib")
#ifdef __cplusplus
};
#endif
AVFormatContext *pFormatCtx_Video = NULL, *pFormatCtx_Audio = NULL, *pFormatCtx_Out = NULL;
AVCodecContext *outVideoCodecCtx = NULL;
AVCodecContext *outAudioCodecCtx = NULL;
AVStream *pVideoStream = NULL, *pAudioStream = NULL;
AVCodec *outAVCodec;
AVCodec *outAudioCodec;
AVCodecContext *pCodecCtx_Video;
AVCodec *pCodec_Video;
AVFifoBuffer *fifo_video = NULL;
AVAudioFifo *fifo_audio = NULL;
int VideoIndex, AudioIndex;
int codec_id;
CRITICAL_SECTION AudioSection, VideoSection;
SwsContext *img_convert_ctx;
int frame_size = 0;
uint8_t *picture_buf = NULL, *frame_buf = NULL;
bool bCap = true;
DWORD WINAPI ScreenCapThreadProc( LPVOID lpParam );
DWORD WINAPI AudioCapThreadProc( LPVOID lpParam );
int OpenVideoCapture()
{
AVInputFormat *ifmt=av_find_input_format("gdigrab");
AVDictionary *options = NULL;
av_dict_set(&options, "framerate", "60", NULL);
if(avformat_open_input(&pFormatCtx_Video, "desktop", ifmt, &options)!=0)
{
printf("Couldn't open input stream.（无法打开视频输入流）\n");
return -1;
}
if(avformat_find_stream_info(pFormatCtx_Video,NULL)<0)
{
printf("Couldn't find stream information.（无法获取视频流信息）\n");
return -1;
}
if (pFormatCtx_Video->streams[0]->codec->codec_type != AVMEDIA_TYPE_VIDEO)
{
printf("Couldn't find video stream information.（无法获取视频流信息）\n");
return -1;
}
pCodecCtx_Video = pFormatCtx_Video->streams[0]->codec;
pCodec_Video = avcodec_find_decoder(pCodecCtx_Video->codec_id);
if(pCodec_Video == NULL)
{
printf("Codec not found.（没有找到解码器）\n");
return -1;
}
if(avcodec_open2(pCodecCtx_Video, pCodec_Video, NULL) < 0)
{
printf("Could not open codec.（无法打开解码器）\n");
return -1;
}
av_dump_format(pFormatCtx_Video, 0, NULL, 0);
img_convert_ctx = sws_getContext(pCodecCtx_Video->width, pCodecCtx_Video->height, pCodecCtx_Video->pix_fmt,
pCodecCtx_Video->width, pCodecCtx_Video->height, PIX_FMT_YUV420P, SWS_BICUBIC, NULL, NULL, NULL);
frame_size = avpicture_get_size(pCodecCtx_Video->pix_fmt, pCodecCtx_Video->width, pCodecCtx_Video->height);
fifo_video = av_fifo_alloc(30 * avpicture_get_size(AV_PIX_FMT_YUV420P, pCodecCtx_Video->width, pCodecCtx_Video->height));
return 0;
}
static char *dup_wchar_to_utf8(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;
}
int OpenAudioCapture()
{
AVInputFormat *pAudioInputFmt = av_find_input_format("dshow");
char * psDevName = dup_wchar_to_utf8(L"audio=virtual-audio-capturer");
if (avformat_open_input(&pFormatCtx_Audio, psDevName, pAudioInputFmt,NULL) < 0)
{
printf("Couldn't open input stream.（无法打开音频输入流）\n");
return -1;
}
if(avformat_find_stream_info(pFormatCtx_Audio,NULL)<0)
return -1;
if(pFormatCtx_Audio->streams[0]->codec->codec_type != AVMEDIA_TYPE_AUDIO)
{
printf("Couldn't find video stream information.（无法获取音频流信息）\n");
return -1;
}
AVCodec *tmpCodec = avcodec_find_decoder(pFormatCtx_Audio->streams[0]->codec->codec_id);
if(0 > avcodec_open2(pFormatCtx_Audio->streams[0]->codec, tmpCodec, NULL))
{
printf("can not find or open audio decoder!\n");
}
av_dump_format(pFormatCtx_Audio, 0, NULL, 0);
return 0;
}
int OpenOutPut()
{
AVStream *pVideoStream = NULL, *pAudioStream = NULL;
const char *outFileName = "test.mp4";
avformat_alloc_output_context2(&pFormatCtx_Out, NULL, NULL, outFileName);
if (pFormatCtx_Video->streams[0]->codec->codec_type == AVMEDIA_TYPE_VIDEO)
{
VideoIndex = 0;
pVideoStream = avformat_new_stream(pFormatCtx_Out, NULL);
if (!pVideoStream)
{
printf("can not new stream for output!\n");
return -1;
}
outVideoCodecCtx = avcodec_alloc_context3(outAVCodec);
if ( !outVideoCodecCtx )
{
printf("Error : avcodec_alloc_context3()\n");
return -1;
}
//set codec context param
outVideoCodecCtx = pVideoStream->codec;
outVideoCodecCtx->codec_id = AV_CODEC_ID_MPEG4;
outVideoCodecCtx->width = pFormatCtx_Video->streams[0]->codec->width;
outVideoCodecCtx->height = pFormatCtx_Video->streams[0]->codec->height;
outVideoCodecCtx->time_base = pFormatCtx_Video->streams[0]->codec->time_base;
outVideoCodecCtx->pix_fmt = AV_PIX_FMT_YUV420P;
outVideoCodecCtx->codec_type = AVMEDIA_TYPE_VIDEO;
if (codec_id == AV_CODEC_ID_H264)
{
av_opt_set(outVideoCodecCtx->priv_data, "preset", "slow", 0);
}
outAVCodec = avcodec_find_encoder(AV_CODEC_ID_MPEG4);
if( !outAVCodec )
{
printf("\n\nError : avcodec_find_encoder()");
return -1;
}
if (pFormatCtx_Out->oformat->flags & AVFMT_GLOBALHEADER)
outVideoCodecCtx->flags |=CODEC_FLAG_GLOBAL_HEADER;
if ((avcodec_open2(outVideoCodecCtx,outAVCodec, NULL)) < 0)
{
printf("can not open the encoder\n");
return -1;
}
}
if(pFormatCtx_Audio->streams[0]->codec->codec_type == AVMEDIA_TYPE_AUDIO)
{
AVCodecContext *pOutputCodecCtx;
AudioIndex = 1;
pAudioStream = avformat_new_stream(pFormatCtx_Out, NULL);
pAudioStream->codec->codec = avcodec_find_encoder(pFormatCtx_Out->oformat->audio_codec);
pOutputCodecCtx = pAudioStream->codec;
pOutputCodecCtx->sample_rate = pFormatCtx_Audio->streams[0]->codec->sample_rate;
pOutputCodecCtx->channel_layout = pFormatCtx_Out->streams[0]->codec->channel_layout;
pOutputCodecCtx->channels = av_get_channel_layout_nb_channels(pAudioStream->codec->channel_layout);
if(pOutputCodecCtx->channel_layout == 0)
{
pOutputCodecCtx->channel_layout = AV_CH_LAYOUT_STEREO;
pOutputCodecCtx->channels = av_get_channel_layout_nb_channels(pOutputCodecCtx->channel_layout);
}
pOutputCodecCtx->sample_fmt = pAudioStream->codec->codec->sample_fmts[0];
AVRational time_base={1, pAudioStream->codec->sample_rate};
pAudioStream->time_base = time_base;
//audioCodecCtx->time_base = time_base;
pOutputCodecCtx->codec_tag = 0;
if (pFormatCtx_Out->oformat->flags & AVFMT_GLOBALHEADER)
pOutputCodecCtx->flags |= CODEC_FLAG_GLOBAL_HEADER;
if (avcodec_open2(pOutputCodecCtx, pOutputCodecCtx->codec, 0) < 0)
{
printf("编码器打开失败，退出程序\n");
return -1;
}
}
if (!(pFormatCtx_Out->oformat->flags & AVFMT_NOFILE))
{
if(avio_open(&pFormatCtx_Out->pb, outFileName, AVIO_FLAG_WRITE) < 0)
{
printf("can not open output file handle!\n");
return -1;
}
}
if(avformat_write_header(pFormatCtx_Out, NULL) < 0)
{
printf("can not write the header of the output file!\n");
return -1;
}
return 0;
}
int _tmain(int argc, _TCHAR* argv[])
{
av_register_all();
avdevice_register_all();
if (OpenVideoCapture() < 0)
{
return -1;
}
if (OpenAudioCapture() < 0)
{
return -1;
}
if (OpenOutPut() < 0)
{
return -1;
}
// int fps;
/*printf("输入帧率：");
scanf_s("%d",&fps);
if ( NULL == fps)
{
fps = 10;
}*/
InitializeCriticalSection(&VideoSection);
InitializeCriticalSection(&AudioSection);
AVFrame *picture = av_frame_alloc();
int size = avpicture_get_size(pFormatCtx_Out->streams[VideoIndex]->codec->pix_fmt,
pFormatCtx_Out->streams[VideoIndex]->codec->width, pFormatCtx_Out->streams[VideoIndex]->codec->height);
picture_buf = new uint8_t[size];
avpicture_fill((AVPicture *)picture, picture_buf,
pFormatCtx_Out->streams[VideoIndex]->codec->pix_fmt,
pFormatCtx_Out->streams[VideoIndex]->codec->width,
pFormatCtx_Out->streams[VideoIndex]->codec->height);
//star cap screen thread
CreateThread( NULL, 0, ScreenCapThreadProc, 0, 0, NULL);
//star cap audio thread
CreateThread( NULL, 0, AudioCapThreadProc, 0, 0, NULL);
int64_t cur_pts_v=0,cur_pts_a=0;
int VideoFrameIndex = 0, AudioFrameIndex = 0;
while(1)
{
if (_kbhit() != 0 && bCap)
{
bCap = false;
Sleep(2000);
}
if (fifo_audio && fifo_video)
{
int sizeAudio = av_audio_fifo_size(fifo_audio);
int sizeVideo = av_fifo_size(fifo_video);
//缓存数据写完就结束循环
if (av_audio_fifo_size(fifo_audio) <= pFormatCtx_Out->streams[AudioIndex]->codec->frame_size &&
av_fifo_size(fifo_video) <= frame_size && !bCap)
{
break;
}
}
if(av_compare_ts(cur_pts_v, pFormatCtx_Out->streams[VideoIndex]->time_base,
cur_pts_a,pFormatCtx_Out->streams[AudioIndex]->time_base) <= 0)
{
if (av_fifo_size(fifo_video) < frame_size && !bCap)
{
cur_pts_v = 0x7fffffffffffffff;
}
if(av_fifo_size(fifo_video) >= size)
{
EnterCriticalSection(&VideoSection);
av_fifo_generic_read(fifo_video, picture_buf, size, NULL); //将数据从avfifobuffer馈送到用户提供的回调。
LeaveCriticalSection(&VideoSection);
avpicture_fill((AVPicture *)picture, picture_buf,
pFormatCtx_Out->streams[VideoIndex]->codec->pix_fmt,
pFormatCtx_Out->streams[VideoIndex]->codec->width,
pFormatCtx_Out->streams[VideoIndex]->codec->height); //根据指定的图像参数和提供的图像数据缓冲区设置图片字段。
//pts = n * (（1 / timbase）/ fps);
//picture->pts = VideoFrameIndex * ((pFormatCtx_Video->streams[0]->time_base.den / pFormatCtx_Video->streams[0]->time_base.num) / 24);
picture->pts = VideoFrameIndex * ((outVideoCodecCtx->time_base.den * 100000 / outVideoCodecCtx->time_base.num) / 180);
int got_picture = 0;
AVPacket pkt;
av_init_packet(&pkt);
pkt.data = NULL;
pkt.size = 0;
//从帧中获取输入的原始视频数据
int ret = avcodec_encode_video2(pFormatCtx_Out->streams[VideoIndex]->codec, &pkt, picture, &got_picture);
if(ret < 0)
{
continue;
}
if (got_picture==1)
{
pkt.stream_index = VideoIndex;
/*int count = 1;
pkt.pts = pkt.dts = count * ((pFormatCtx_Video->streams[0]->time_base.den / pFormatCtx_Video->streams[0]->time_base.num) / 15);
count++;*/
//x = pts * (timebase1.num / timebase1.den )* (timebase2.den / timebase2.num);
pkt.pts = av_rescale_q_rnd(pkt.pts, pFormatCtx_Video->streams[0]->time_base,
pFormatCtx_Out->streams[VideoIndex]->time_base, (AVRounding)(AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX));
pkt.dts = av_rescale_q_rnd(pkt.dts, pFormatCtx_Video->streams[0]->time_base,
pFormatCtx_Out->streams[VideoIndex]->time_base, (AVRounding)(AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX));
pkt.duration = ((pFormatCtx_Out->streams[0]->time_base.den / pFormatCtx_Out->streams[0]->time_base.num) / 60);
//pkt.duration = 1000/60;
//pkt.pts = pkt.dts = Count * (ofmt_ctx->streams[stream_index]->time_base.den) /ofmt_ctx->streams[stream_index]->time_base.num / 10;
//Count++;
cur_pts_v = pkt.pts;
ret = av_interleaved_write_frame(pFormatCtx_Out, &pkt);
//delete[] pkt.data;
av_free_packet(&pkt);
}
VideoFrameIndex++;
}
}
else
{
if (NULL == fifo_audio)
{
continue;//还未初始化fifo
}
if (av_audio_fifo_size(fifo_audio) < pFormatCtx_Out->streams[AudioIndex]->codec->frame_size && !bCap)
{
cur_pts_a = 0x7fffffffffffffff;
}
if(av_audio_fifo_size(fifo_audio) >=
(pFormatCtx_Out->streams[AudioIndex]->codec->frame_size > 0 ? pFormatCtx_Out->streams[AudioIndex]->codec->frame_size : 1024))
{
AVFrame *frame;
frame = av_frame_alloc();
frame->nb_samples = pFormatCtx_Out->streams[AudioIndex]->codec->frame_size>0 ? pFormatCtx_Out->streams[AudioIndex]->codec->frame_size: 1024;
frame->channel_layout = pFormatCtx_Out->streams[AudioIndex]->codec->channel_layout;
frame->format = pFormatCtx_Out->streams[AudioIndex]->codec->sample_fmt;
frame->sample_rate = pFormatCtx_Out->streams[AudioIndex]->codec->sample_rate;
av_frame_get_buffer(frame, 0);
EnterCriticalSection(&AudioSection);
av_audio_fifo_read(fifo_audio, (void **)frame->data,
(pFormatCtx_Out->streams[AudioIndex]->codec->frame_size > 0 ? pFormatCtx_Out->streams[AudioIndex]->codec->frame_size : 1024));
LeaveCriticalSection(&AudioSection);
AVPacket pkt_out;
av_init_packet(&pkt_out);
int got_picture = -1;
pkt_out.data = NULL;
pkt_out.size = 0;
frame->pts = AudioFrameIndex * pFormatCtx_Out->streams[AudioIndex]->codec->frame_size;
if (avcodec_encode_audio2(pFormatCtx_Out->streams[AudioIndex]->codec, &pkt_out, frame, &got_picture) < 0)
{
printf("can not decoder a frame");
}
av_frame_free(&frame);
if (got_picture)
{
pkt_out.stream_index = AudioIndex;
pkt_out.pts = AudioFrameIndex * pFormatCtx_Out->streams[AudioIndex]->codec->frame_size;
pkt_out.dts = AudioFrameIndex * pFormatCtx_Out->streams[AudioIndex]->codec->frame_size;
pkt_out.duration = pFormatCtx_Out->streams[AudioIndex]->codec->frame_size;
cur_pts_a = pkt_out.pts;
int ret = av_interleaved_write_frame(pFormatCtx_Out, &pkt_out);
av_free_packet(&pkt_out);
}
AudioFrameIndex++;
}
}
}
delete[] picture_buf;
av_fifo_free(fifo_video);
av_audio_fifo_free(fifo_audio);
av_write_trailer(pFormatCtx_Out);
avio_close(pFormatCtx_Out->pb);
avformat_free_context(pFormatCtx_Out);
if (pFormatCtx_Video != NULL)
{
avformat_close_input(&pFormatCtx_Video);
pFormatCtx_Video = NULL;
}
if (pFormatCtx_Audio != NULL)
{
avformat_close_input(&pFormatCtx_Audio);
pFormatCtx_Audio = NULL;
}
return 0;
}
DWORD WINAPI ScreenCapThreadProc( LPVOID lpParam )
{
AVPacket packet;
int got_picture;
AVFrame *pFrame;
pFrame=av_frame_alloc();
AVFrame *picture = av_frame_alloc();
int size = avpicture_get_size(pFormatCtx_Out->streams[VideoIndex]->codec->pix_fmt,
pFormatCtx_Out->streams[VideoIndex]->codec->width,
pFormatCtx_Out->streams[VideoIndex]->codec->height);
avpicture_fill((AVPicture *)picture, picture_buf,
pFormatCtx_Out->streams[VideoIndex]->codec->pix_fmt,
pFormatCtx_Out->streams[VideoIndex]->codec->width,
pFormatCtx_Out->streams[VideoIndex]->codec->height);
FILE *p = NULL;
p = fopen("proc_test.yuv", "wb+");
av_init_packet(&packet);
int height = pFormatCtx_Out->streams[VideoIndex]->codec->height;
int width = pFormatCtx_Out->streams[VideoIndex]->codec->width;
int y_size=height*width;
while(bCap)
{
packet.data = NULL;
packet.size = 0;
if (av_read_frame(pFormatCtx_Video, &packet) < 0)
{
continue;
}
if(packet.stream_index == 0)
{
if (avcodec_decode_video2(pCodecCtx_Video, pFrame, &got_picture, &packet) < 0)
{
printf("Decode Error.（解码错误）\n");
continue;
}
if (got_picture)
{
sws_scale(img_convert_ctx,
(const uint8_t* const*)pFrame->data,
pFrame->linesize,
0,
pFormatCtx_Out->streams[VideoIndex]->codec->height,
picture->data,
picture->linesize);
if (av_fifo_space(fifo_video) >= size)
{
EnterCriticalSection(&VideoSection);
av_fifo_generic_write(fifo_video, picture->data[0], y_size, NULL);
av_fifo_generic_write(fifo_video, picture->data[1], y_size/4, NULL);
av_fifo_generic_write(fifo_video, picture->data[2], y_size/4, NULL);
LeaveCriticalSection(&VideoSection);
}
}
}
av_free_packet(&packet);
}
av_frame_free(&pFrame);
av_frame_free(&picture);
return 0;
}
DWORD WINAPI AudioCapThreadProc( LPVOID lpParam )
{
AVPacket pkt;
AVFrame *frame;
frame = av_frame_alloc();
int gotframe;
while(bCap)
{
pkt.data = NULL;
pkt.size = 0;
if(av_read_frame(pFormatCtx_Audio,&pkt) < 0)
{
continue;
}
if (avcodec_decode_audio4(pFormatCtx_Audio->streams[0]->codec, frame, &gotframe, &pkt) < 0)
{
av_frame_free(&frame);
printf("can not decoder a frame");
break;
}
av_free_packet(&pkt);
if (!gotframe)
{
printf("没有获取到数据，继续下一次");
continue;
}
if (NULL == fifo_audio)
{
fifo_audio = av_audio_fifo_alloc(pFormatCtx_Audio->streams[0]->codec->sample_fmt,
pFormatCtx_Audio->streams[0]->codec->channels, 30 * frame->nb_samples);
}
int buf_space = av_audio_fifo_space(fifo_audio);
if (av_audio_fifo_space(fifo_audio) >= frame->nb_samples)
{
EnterCriticalSection(&AudioSection);
av_audio_fifo_write(fifo_audio, (void **)frame->data, frame->nb_samples);
LeaveCriticalSection(&AudioSection);
}
}
av_frame_free(&frame);
return 0;
}
也许还有另一种计算PTS和DTS的方法
我希望无论帧速率如何，视频播放速度都是正确的。不要太快或太慢。
答案 0 :(得分：2)
最后，我找出了问题的原因。
由视频录制生成的视频的帧频取决于录制的视频流。我的计算机的gdigrab的最大帧频为10帧，因此如果将其设置为10帧以上，则播放速度会很快。我将其设置为少于10帧，则播放速度会很慢。但是我在我玩游戏的计算机上运行代码。无论我选择10帧还是15帧，回溯速度都是正确的。
不过，我仍然不知道我的gdigrab的帧速率仅为0-10帧的原因。CPU显卡，卡，显示器和内存等许多因素会影响视频记录的帧速率。 / p>
这是最终代码 使用FFMPEG捕获屏幕和音频
#include "stdafx.h"
#ifdef __cplusplus
extern "C"
{
#endif
#include "libavcodec/avcodec.h"
#include "libavformat/avformat.h"
#include "libswscale/swscale.h"
#include "libavdevice/avdevice.h"
#include "libavutil/audio_fifo.h"
#include "libavfilter/buffersink.h"
#include "libavfilter/buffersrc.h"
#include "libavutil/imgutils.h"
#include "libavutil/mathematics.h"
#include "libavutil/samplefmt.h"
#include "libavutil/time.h"
#include "libavutil/opt.h"
#include "libavutil/pixdesc.h"
#include "libavutil/file.h"
#include "libavutil/mem.h"
#include "libavutil/frame.h"
#include "libavfilter/avfilter.h"
#include "libswresample/swresample.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, "avfilter.lib")
#pragma comment(lib, "postproc.lib")
#pragma comment(lib, "swresample.lib")
#pragma comment(lib, "swscale.lib")
#ifdef __cplusplus
};
#endif
AVFormatContext *pFormatCtx_Video = NULL, *pFormatCtx_Audio = NULL, *pFormatCtx_Out = NULL;
AVCodecContext *outVideoCodecCtx = NULL;
AVCodecContext *outAudioCodecCtx = NULL;
AVStream *pVideoStream = NULL, *pAudioStream = NULL;
AVCodec *outAVCodec;
AVCodec *outAudioCodec;
AVCodecContext *pCodecCtx_Video;
AVCodec *pCodec_Video;
AVFifoBuffer *fifo_video = NULL;
AVAudioFifo *fifo_audio = NULL;
int VideoIndex, AudioIndex;
int codec_id;
CRITICAL_SECTION AudioSection, VideoSection;
SwsContext *img_convert_ctx;
int frame_size = 0;
uint8_t *picture_buf = NULL;
bool bCap = true;
DWORD WINAPI ScreenCapThreadProc( LPVOID lpParam );
DWORD WINAPI AudioCapThreadProc( LPVOID lpParam );
static char *dup_wchar_to_utf8(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;
}
int OpenVideoCapture()
{
int fps = 10;
char opt;
printf("选择视频播放帧率：\n");
printf("A 5帧\n");
printf("B 10帧\n");
printf("C 15帧\n");
printf("D 20帧\n");
printf("E 25帧\n");
printf("F 30帧\n");
opt = getchar();
AVDictionary *options = NULL;
switch (opt)
{
case 'A':
fps = 5;
av_dict_set(&options, "framerate", "5", 0);
break;
case 'a':
fps = 5;
av_dict_set(&options, "framerate", "5", 0);
break;
case 'B':
fps = 10;
av_dict_set(&options, "framerate", "10", 0);
break;
case 'b':
fps = 10;
av_dict_set(&options, "framerate", "10", 0);
break;
case 'C':
fps = 15;
av_dict_set(&options, "framerate", "15", 0);
break;
case 'c':
fps = 15;
av_dict_set(&options, "framerate", "15", 0);
break;
case 'D':
fps = 20;
av_dict_set(&options, "framerate", "20", 0);
break;
case 'd':
fps = 20;
av_dict_set(&options, "framerate", "20", 0);
break;
case 'E':
fps = 25;
av_dict_set(&options, "framerate", "25", 0);
break;
case 'e':
fps = 25;
av_dict_set(&options, "framerate", "25", 0);
break;
case 'F':
fps = 30;
av_dict_set(&options, "framerate", "30", 0);
break;
case 'f':
fps = 30;
av_dict_set(&options, "framerate", "30", 0);
break;
default:
printf("选项输入错误\n");
return -1;
}
//AVDictionary *options = NULL;
//av_dict_set(&options, "framerate", "15", 0);
AVInputFormat *ifmt=av_find_input_format("gdigrab");
if(avformat_open_input(&pFormatCtx_Video, "desktop", ifmt, &options) != 0)
{
printf("Couldn't open input stream.（无法打开视频输入流）\n");
return -1;
}
if(avformat_find_stream_info(pFormatCtx_Video,NULL)<0)
{
printf("Couldn't find stream information.（无法获取视频流信息）\n");
return -1;
}
if (pFormatCtx_Video->streams[0]->codec->codec_type != AVMEDIA_TYPE_VIDEO)
{
printf("Couldn't find video stream information.（无法获取视频流信息）\n");
return -1;
}
pCodecCtx_Video = pFormatCtx_Video->streams[0]->codec;
pCodec_Video = avcodec_find_decoder(pCodecCtx_Video->codec_id);
if(pCodec_Video == NULL)
{
printf("Codec not found.（没有找到解码器）\n");
return -1;
}
if(avcodec_open2(pCodecCtx_Video, pCodec_Video, NULL) < 0)
{
printf("Could not open codec.（无法打开解码器）\n");
return -1;
}
av_dump_format(pFormatCtx_Video, 0, NULL, 0);
img_convert_ctx = sws_getContext(pCodecCtx_Video->width,
pCodecCtx_Video->height,
pCodecCtx_Video->pix_fmt,
pCodecCtx_Video->width,
pCodecCtx_Video->height,
PIX_FMT_YUV420P,
SWS_BICUBIC, NULL, NULL, NULL);
frame_size = avpicture_get_size(pCodecCtx_Video->pix_fmt, pCodecCtx_Video->width, pCodecCtx_Video->height);
fifo_video = av_fifo_alloc(60 * avpicture_get_size(AV_PIX_FMT_YUV420P, pCodecCtx_Video->width, pCodecCtx_Video->height));
av_dict_free(&options);
return 0;
}
int OpenAudioCapture()
{
AVInputFormat *pAudioInputFmt = av_find_input_format("dshow");
AVDictionary *opt = NULL;
char * psDevName = dup_wchar_to_utf8(L"audio=virtual-audio-capturer");
if (avformat_open_input(&pFormatCtx_Audio, psDevName, pAudioInputFmt, &opt) < 0)
{
printf("Couldn't open input stream.（无法打开音频输入流）\n");
return -1;
}
if(avformat_find_stream_info(pFormatCtx_Audio,NULL)<0)
return -1;
if(pFormatCtx_Audio->streams[0]->codec->codec_type != AVMEDIA_TYPE_AUDIO)
{
printf("Couldn't find video stream information.（无法获取音频流信息）\n");
return -1;
}
AVCodec *tmpCodec = avcodec_find_decoder(pFormatCtx_Audio->streams[0]->codec->codec_id);
if(0 > avcodec_open2(pFormatCtx_Audio->streams[0]->codec, tmpCodec, NULL))
{
printf("can not find or open audio decoder!\n");
}
av_dump_format(pFormatCtx_Audio, 0, NULL, 0);
return 0;
}
int OpenOutPut()
{
AVStream *pVideoStream = NULL, *pAudioStream = NULL;
const char *outFileName = "test.mp4";
avformat_alloc_output_context2(&pFormatCtx_Out, NULL, NULL, outFileName);
if (pFormatCtx_Video->streams[0]->codec->codec_type == AVMEDIA_TYPE_VIDEO)
{
VideoIndex = 0;
pVideoStream = avformat_new_stream(pFormatCtx_Out, NULL);
if (!pVideoStream)
{
printf("can not new stream for output!\n");
return -1;
}
outVideoCodecCtx = avcodec_alloc_context3(outAVCodec);
if ( !outVideoCodecCtx )
{
printf("Error : avcodec_alloc_context3()\n");
return -1;
}
outVideoCodecCtx = pVideoStream->codec;
outVideoCodecCtx->codec_id = AV_CODEC_ID_MPEG4;
outVideoCodecCtx->width = pFormatCtx_Video->streams[0]->codec->width;
outVideoCodecCtx->height = pFormatCtx_Video->streams[0]->codec->height;
outVideoCodecCtx->time_base = pFormatCtx_Video->streams[0]->codec->time_base;;
outVideoCodecCtx->pix_fmt = AV_PIX_FMT_YUV420P;
outVideoCodecCtx->codec_type = AVMEDIA_TYPE_VIDEO;
if (codec_id == AV_CODEC_ID_H264)
{
av_opt_set(outVideoCodecCtx->priv_data, "preset", "slow", 0);
}
outAVCodec = avcodec_find_encoder(AV_CODEC_ID_MPEG4);
if( !outAVCodec )
{
printf("\n\nError : avcodec_find_encoder()");
return -1;
}
if (pFormatCtx_Out->oformat->flags & AVFMT_GLOBALHEADER)
outVideoCodecCtx->flags |=CODEC_FLAG_GLOBAL_HEADER;
if ((avcodec_open2(outVideoCodecCtx,outAVCodec, NULL)) < 0)
{
printf("can not open the encoder\n");
return -1;
}
}
if(pFormatCtx_Audio->streams[0]->codec->codec_type == AVMEDIA_TYPE_AUDIO)
{
AVCodecContext *pOutputCodecCtx;
AudioIndex = 1;
pAudioStream = avformat_new_stream(pFormatCtx_Out, NULL);
pAudioStream->codec->codec = avcodec_find_encoder(pFormatCtx_Out->oformat->audio_codec);
pOutputCodecCtx = pAudioStream->codec;
pOutputCodecCtx->sample_rate = pFormatCtx_Audio->streams[0]->codec->sample_rate;
pOutputCodecCtx->channel_layout = pFormatCtx_Out->streams[0]->codec->channel_layout;
pOutputCodecCtx->channels = av_get_channel_layout_nb_channels(pAudioStream->codec->channel_layout);
if(pOutputCodecCtx->channel_layout == 0)
{
pOutputCodecCtx->channel_layout = AV_CH_LAYOUT_STEREO;
pOutputCodecCtx->channels = av_get_channel_layout_nb_channels(pOutputCodecCtx->channel_layout);
}
pOutputCodecCtx->sample_fmt = pAudioStream->codec->codec->sample_fmts[0];
AVRational time_base={1, pAudioStream->codec->sample_rate};
pAudioStream->time_base = time_base;
pOutputCodecCtx->codec_tag = 0;
if (pFormatCtx_Out->oformat->flags & AVFMT_GLOBALHEADER)
pOutputCodecCtx->flags |= CODEC_FLAG_GLOBAL_HEADER;
if (avcodec_open2(pOutputCodecCtx, pOutputCodecCtx->codec, 0) < 0)
{
printf("编码器打开失败，退出程序\n");
return -1;
}
}
if (!(pFormatCtx_Out->oformat->flags & AVFMT_NOFILE))
{
if(avio_open(&pFormatCtx_Out->pb, outFileName, AVIO_FLAG_WRITE) < 0)
{
printf("can not open output file handle!\n");
return -1;
}
}
if(avformat_write_header(pFormatCtx_Out, NULL) < 0)
{
printf("can not write the header of the output file!\n");
return -1;
}
return 0;
}
int _tmain(int argc, _TCHAR* argv[])
{
av_register_all();
avdevice_register_all();
if (OpenVideoCapture() < 0)
{
return -1;
}
if (OpenAudioCapture() < 0)
{
return -1;
}
if (OpenOutPut() < 0)
{
return -1;
}
InitializeCriticalSection(&VideoSection);
InitializeCriticalSection(&AudioSection);
AVFrame *picture = av_frame_alloc();
int size = avpicture_get_size(pFormatCtx_Out->streams[VideoIndex]->codec->pix_fmt,
pFormatCtx_Out->streams[VideoIndex]->codec->width, pFormatCtx_Out->streams[VideoIndex]->codec->height);
picture_buf = new uint8_t[size];
avpicture_fill((AVPicture *)picture, picture_buf,
pFormatCtx_Out->streams[VideoIndex]->codec->pix_fmt,
pFormatCtx_Out->streams[VideoIndex]->codec->width,
pFormatCtx_Out->streams[VideoIndex]->codec->height);
//star cap screen thread
CreateThread( NULL, 0, ScreenCapThreadProc, 0, 0, NULL);
//star cap audio thread
CreateThread( NULL, 0, AudioCapThreadProc, 0, 0, NULL);
int64_t cur_pts_v=0,cur_pts_a=0;
int64_t VideoFrameIndex = 0, AudioFrameIndex = 0;
int64_t count = 1;
int64_t video_pre_pts = 0;
while(1)
{
if (_kbhit() != 0 && bCap)
{
bCap = false;
Sleep(2000);
}
if (fifo_audio && fifo_video)
{
int sizeAudio = av_audio_fifo_size(fifo_audio);
int sizeVideo = av_fifo_size(fifo_video);
//缓存数据写完就结束循环
if (av_audio_fifo_size(fifo_audio) <= pFormatCtx_Out->streams[AudioIndex]->codec->frame_size &&
av_fifo_size(fifo_video) <= frame_size && !bCap)
{
break;
}
}
if(av_compare_ts(cur_pts_v, pFormatCtx_Out->streams[VideoIndex]->time_base, cur_pts_a,pFormatCtx_Out->streams[AudioIndex]->time_base) <= 0)
{
if (av_fifo_size(fifo_video) < frame_size && !bCap)
{
cur_pts_v = 0x7fffffffffffffff;
}
if(av_fifo_size(fifo_video) >= size)
{
//将数据从avfifobuffer馈送到用户提供的回调。
EnterCriticalSection(&VideoSection);
av_fifo_generic_read(fifo_video, picture_buf, size, NULL);
LeaveCriticalSection(&VideoSection);
//根据指定的图像参数和提供的图像数据缓冲区设置图片字段。
avpicture_fill((AVPicture *)picture, picture_buf,
pFormatCtx_Out->streams[VideoIndex]->codec->pix_fmt,
pFormatCtx_Out->streams[VideoIndex]->codec->width,
pFormatCtx_Out->streams[VideoIndex]->codec->height);
//pts = n * (（1 / timbase）/ fps);
//picture->pts = VideoFrameIndex * ((pFormatCtx_Video->streams[0]->time_base.den / pFormatCtx_Video->streams[0]->time_base.num) / 15);
picture->pts = av_rescale_q(VideoFrameIndex,outVideoCodecCtx->time_base,pFormatCtx_Video->streams[0]->time_base);
printf("picture->pts: %d\n",picture->pts);
int got_picture = 0;
AVPacket pkt;
av_init_packet(&pkt);
pkt.data = NULL;
pkt.size = 0;
//从帧中获取输入的原始视频数据
int ret = avcodec_encode_video2(pFormatCtx_Out->streams[VideoIndex]->codec, &pkt, picture, &got_picture);
if(ret < 0)
{
continue;
}
if (got_picture==1)
{
pkt.stream_index = VideoIndex;
//pFormatCtx_Video //pFormatCtx_Out
pkt.pts = av_rescale_q_rnd(pkt.pts, pFormatCtx_Video->streams[0]->time_base, pFormatCtx_Out->streams[VideoIndex]->time_base, (AVRounding)(AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX));
printf("pkt.pts = %d\n",pkt.pts);
pkt.dts = av_rescale_q_rnd(pkt.dts, pFormatCtx_Video->streams[0]->time_base, pFormatCtx_Out->streams[VideoIndex]->time_base, (AVRounding)(AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX));
printf("pkt.dts = %d\n",pkt.dts);
pkt.duration = ((pFormatCtx_Out->streams[0]->time_base.den / pFormatCtx_Out->streams[0]->time_base.num) / outVideoCodecCtx->time_base.den);
//pkt.duration = 1;
//pkt.duration = av_rescale_q(pkt.duration,outVideoCodecCtx->time_base,pFormatCtx_Video->streams[0]->time_base);
printf("pkt.duration = %d\n",pkt.duration);
pkt.pos = -1;
cur_pts_v = pkt.pts;
ret = av_interleaved_write_frame(pFormatCtx_Out, &pkt);
if (ret<0)
{
printf("Error muxing packet\n");
break;
}
av_free_packet(&pkt);
}
VideoFrameIndex++;
}
}
else
{
if (NULL == fifo_audio)
{
continue;//还未初始化fifo
}
if (av_audio_fifo_size(fifo_audio) < pFormatCtx_Out->streams[AudioIndex]->codec->frame_size && !bCap)
{
cur_pts_a = 0x7fffffffffffffff;
}
if(av_audio_fifo_size(fifo_audio) >=
(pFormatCtx_Out->streams[AudioIndex]->codec->frame_size > 0 ? pFormatCtx_Out->streams[AudioIndex]->codec->frame_size : 1024))
{
AVFrame *frame;
frame = av_frame_alloc();
frame->nb_samples = pFormatCtx_Out->streams[AudioIndex]->codec->frame_size>0 ? pFormatCtx_Out->streams[AudioIndex]->codec->frame_size: 1024;
frame->channel_layout = pFormatCtx_Out->streams[AudioIndex]->codec->channel_layout;
frame->format = pFormatCtx_Out->streams[AudioIndex]->codec->sample_fmt;
frame->sample_rate = pFormatCtx_Out->streams[AudioIndex]->codec->sample_rate;
av_frame_get_buffer(frame, 0);
EnterCriticalSection(&AudioSection);
av_audio_fifo_read(fifo_audio, (void **)frame->data,
(pFormatCtx_Out->streams[AudioIndex]->codec->frame_size > 0 ? pFormatCtx_Out->streams[AudioIndex]->codec->frame_size : 1024));
LeaveCriticalSection(&AudioSection);
AVPacket pkt_out;
av_init_packet(&pkt_out);
int got_picture = -1;
pkt_out.data = NULL;
pkt_out.size = 0;
frame->pts = AudioFrameIndex * pFormatCtx_Out->streams[AudioIndex]->codec->frame_size;
if (avcodec_encode_audio2(pFormatCtx_Out->streams[AudioIndex]->codec, &pkt_out, frame, &got_picture) < 0)
{
printf("can not decoder a frame");
}
av_frame_free(&frame);
if (got_picture)
{
pkt_out.stream_index = AudioIndex;
pkt_out.pts = AudioFrameIndex * pFormatCtx_Out->streams[AudioIndex]->codec->frame_size;
pkt_out.dts = AudioFrameIndex * pFormatCtx_Out->streams[AudioIndex]->codec->frame_size;
pkt_out.duration = pFormatCtx_Out->streams[AudioIndex]->codec->frame_size;
cur_pts_a = pkt_out.pts;
int ret = av_interleaved_write_frame(pFormatCtx_Out, &pkt_out);
if (ret<0)
{
printf("Error muxing packet\n");
break;
}
av_free_packet(&pkt_out);
}
AudioFrameIndex++;
}
}
}
delete[] picture_buf;
av_fifo_free(fifo_video);
av_audio_fifo_free(fifo_audio);
av_write_trailer(pFormatCtx_Out);
avio_close(pFormatCtx_Out->pb);
avformat_free_context(pFormatCtx_Out);
if (pFormatCtx_Video != NULL)
{
avformat_close_input(&pFormatCtx_Video);
pFormatCtx_Video = NULL;
}
if (pFormatCtx_Audio != NULL)
{
avformat_close_input(&pFormatCtx_Audio);
pFormatCtx_Audio = NULL;
}
return 0;
}
DWORD WINAPI ScreenCapThreadProc( LPVOID lpParam )
{
AVPacket packet;
int got_picture;
AVFrame *pFrame;
pFrame=av_frame_alloc();
AVFrame *picture = av_frame_alloc();
int size = avpicture_get_size(pFormatCtx_Out->streams[VideoIndex]->codec->pix_fmt,
pFormatCtx_Out->streams[VideoIndex]->codec->width,
pFormatCtx_Out->streams[VideoIndex]->codec->height);
avpicture_fill((AVPicture *)picture, picture_buf,
pFormatCtx_Out->streams[VideoIndex]->codec->pix_fmt,
pFormatCtx_Out->streams[VideoIndex]->codec->width,
pFormatCtx_Out->streams[VideoIndex]->codec->height);
av_init_packet(&packet);
int height = pFormatCtx_Out->streams[VideoIndex]->codec->height;
int width = pFormatCtx_Out->streams[VideoIndex]->codec->width;
int y_size=height*width;
while(bCap)
{
packet.data = NULL;
packet.size = 0;
if (av_read_frame(pFormatCtx_Video, &packet) < 0)
{
continue;
}
if(packet.stream_index == 0)
{
if (avcodec_decode_video2(pCodecCtx_Video, pFrame, &got_picture, &packet) < 0)
{
printf("Decode Error.（解码错误）\n");
continue;
}
if (got_picture)
{
sws_scale(img_convert_ctx,
(const uint8_t* const*)pFrame->data,
pFrame->linesize,
0,
pFormatCtx_Out->streams[VideoIndex]->codec->height,
picture->data,
picture->linesize);
if (av_fifo_space(fifo_video) >= size)
{
EnterCriticalSection(&VideoSection);
av_fifo_generic_write(fifo_video, picture->data[0], y_size, NULL);
av_fifo_generic_write(fifo_video, picture->data[1], y_size/4, NULL);
av_fifo_generic_write(fifo_video, picture->data[2], y_size/4, NULL);
LeaveCriticalSection(&VideoSection);
}
}
}
av_free_packet(&packet);
}
av_frame_free(&pFrame);
av_frame_free(&picture);
return 0;
}
DWORD WINAPI AudioCapThreadProc( LPVOID lpParam )
{
AVPacket pkt;
AVFrame *frame;
frame = av_frame_alloc();
int gotframe;
while(bCap)
{
pkt.data = NULL;
pkt.size = 0;
if(av_read_frame(pFormatCtx_Audio,&pkt) < 0)
{
continue;
}
if (avcodec_decode_audio4(pFormatCtx_Audio->streams[0]->codec, frame, &gotframe, &pkt) < 0)
{
av_frame_free(&frame);
printf("can not decoder a frame");
break;
}
av_free_packet(&pkt);
if (!gotframe)
{
printf("没有获取到数据，继续下一次");
continue;
}
if (NULL == fifo_audio)
{
fifo_audio = av_audio_fifo_alloc(pFormatCtx_Audio->streams[0]->codec->sample_fmt,
pFormatCtx_Audio->streams[0]->codec->channels, 30 * frame->nb_samples);
}
int buf_space = av_audio_fifo_space(fifo_audio);
if (av_audio_fifo_space(fifo_audio) >= frame->nb_samples)
{
EnterCriticalSection(&AudioSection);
av_audio_fifo_write(fifo_audio, (void **)frame->data, frame->nb_samples);
LeaveCriticalSection(&AudioSection);
}
}
av_frame_free(&frame);
return 0;
}