使用 ffmpeg 进行音频/视频编码：
我正在尝试使用 ffmpeg 创建一个带有编码视频和音频的 avi 文件。
首先，我创建文件：
//define BITRATE 10000000
//define GOP 300
//define FPS 60
//define VIDEOTYPE "avi"
if (!encoder_->createFile(QFileInfo(*(videoFile_.data())).absoluteFilePath(), targetRect.width(), targetRect.height(), BITRATE*(1000 / FPS), GOP, 1000))
缓冲区初始化为：
audio_outbuf_size = 44100 * 0.005 * 16; //5ms of audio should be encoded, each time this function is called
audio_outbuf = new uint8_t[audio_outbuf_size];
outbuf_size = getWidth()*getHeight() * 3;
outbuf = new uint8_t[outbuf_size];
然后添加音频和视频流（音频：CODEC_ID_PCM_S16LE，16000 kb/s 和 44100 Hz，视频：PIX_FMT_YUV420P）
void MediaMuxer::addAudioStream(QString fileName, ffmpeg::CodecID codec_id)
{
// Add the audio stream
ffmpeg::AVCodec *encoder = avcodec_find_encoder(codec_id);
pAudioStream_ = ffmpeg::av_new_stream(pOutputFormatCtx_, 0);
if (!pAudioStream_) {
printf("Could not allocate stream\n");
return;
}
pAudioCodecCtx_ = pAudioStream_->codec;
pAudioCodecCtx_->codec_id = codec_id;
pAudioCodecCtx_->codec_type = ffmpeg::AVMEDIA_TYPE_AUDIO;
pAudioCodecCtx_->sample_fmt = ffmpeg::AV_SAMPLE_FMT_S16;
pAudioCodecCtx_->sample_fmt = encoder->sample_fmts[0];
pAudioCodecCtx_->bit_rate = 16000;
//pAudioCodecCtx_->bit_rate = 64000;
pAudioCodecCtx_->sample_rate = N;
pAudioCodecCtx_->channels = 1;
pAudioCodecCtx_->time_base.den = FPS;
pAudioCodecCtx_->time_base.num = 1;
avcodec_thread_init(pAudioCodecCtx_, 10);
// some formats want stream headers to be separate
if (pOutputFormatCtx_->oformat->flags & AVFMT_GLOBALHEADER)
pAudioCodecCtx_->flags |= CODEC_FLAG_GLOBAL_HEADER;
if (av_set_parameters(pOutputFormatCtx_, NULL) < 0)
{
printf("Invalid output format parameters\n");
return;
}
//ffmpeg::dump_format(pOutputFormatCtx_, 0, fileName.toStdString().c_str(), 1);
// open_video
// find the audio encoder
pAudioCodec_ = avcodec_find_encoder(pAudioCodecCtx_->codec_id);
if (!pAudioCodec_)
{
printf("codec not found\n");
return;
}
// open the codec
if (avcodec_open(pAudioCodecCtx_, pAudioCodec_) < 0)
{
printf("could not open codec\n");
return;
}
// Allocate memory for output
if (!initAudioOutputBuf())
{
printf("Can't allocate memory for audio output bitstream\n");
return;
}
// Allocate the audio frame
if (!initAudioFrame())
{
printf("Can't init audio frame\n");
return;
}
if (url_fopen(&pOutputFormatCtx_->pb, fileName.toStdString().c_str(), URL_WRONLY) < 0)
{
printf("Could not open '%s'\n", fileName.toStdString().c_str());
return;
}
av_write_header(pOutputFormatCtx_);
}
void MediaMuxer::addVideoStream(QString fileName)
{
// Add the video stream
pVideoStream_ = ffmpeg::av_new_stream(pOutputFormatCtx_, 0);
if (!pVideoStream_)
{
printf("Could not allocate stream\n");
return;
}
pVideoCodecCtx_ = pVideoStream_->codec;
pVideoCodecCtx_->codec_id = pOutputFormat_->video_codec;
pVideoCodecCtx_->codec_type = ffmpeg::AVMEDIA_TYPE_VIDEO;
pVideoCodecCtx_->bit_rate = Bitrate;
pVideoCodecCtx_->width = getWidth();
pVideoCodecCtx_->height = getHeight();
pVideoCodecCtx_->time_base.den = FPS;
pVideoCodecCtx_->time_base.num = 1;
pVideoCodecCtx_->gop_size = Gop;
pVideoCodecCtx_->pix_fmt = ffmpeg::PIX_FMT_YUV420P;
avcodec_thread_init(pVideoCodecCtx_, 10);
// some formats want stream headers to be separate
if (pOutputFormatCtx_->oformat->flags & AVFMT_GLOBALHEADER)
pVideoCodecCtx_->flags |= CODEC_FLAG_GLOBAL_HEADER;
if (av_set_parameters(pOutputFormatCtx_, NULL) < 0)
{
printf("Invalid output format parameters\n");
return;
}
//ffmpeg::dump_format(pOutputFormatCtx_, 0, fileName.toStdString().c_str(), 1);
// open_video
// find the video encoder
pVideoCodec_ = avcodec_find_encoder(pVideoCodecCtx_->codec_id);
if (!pVideoCodec_)
{
printf("codec not found\n");
return;
}
// open the codec
if (avcodec_open(pVideoCodecCtx_, pVideoCodec_) < 0)
{
printf("could not open codec\n");
return;
}
// Allocate memory for output
if (!initOutputBuf())
{
printf("Can't allocate memory for output bitstream\n");
return;
}
// Allocate the YUV frame
if (!initFrame())
{
printf("Can't init frame\n");
return;
}
if (url_fopen(&pOutputFormatCtx_->pb, fileName.toStdString().c_str(), URL_WRONLY) < 0)
{
printf("Could not open '%s'\n", fileName.toStdString().c_str());
return;
}
av_write_header(pOutputFormatCtx_);
}
最后，我交替调用 encodeVideo/encodeAudio 以在特定录制时间（pts）对视频和 PCM 音频帧进行编码：
int MediaMuxer::encodeVideo(const QImage &img, unsigned pts)
{
convertImage_sws(img); // SWS conversion
pVideoCodecCtx_->coded_frame->pts = pts; // Set the time stamp
int out_size = ffmpeg::avcodec_encode_video(pVideoCodecCtx_, outbuf, outbuf_size, ppicture);
pVideoCodecCtx_->coded_frame->pts = pts; // Set the time stamp
if (out_size > 0)
{
ffmpeg::av_init_packet(&pkt);
if (pVideoCodecCtx_->coded_frame->pts != (0x8000000000000000LL))
pkt.pts = av_rescale_q(pVideoCodecCtx_->coded_frame->pts, pVideoCodecCtx_->time_base, pVideoStream_->time_base);
if (pVideoCodecCtx_->coded_frame->key_frame)
pkt.flags |= AV_PKT_FLAG_KEY;
pkt.stream_index = pVideoStream_->index;
pkt.data = outbuf;
pkt.size = out_size;
int ret = ffmpeg::av_interleaved_write_frame(pOutputFormatCtx_, &pkt);
if (ret<0)
return -1;
}
return out_size;
}
int MediaMuxer::encodeAudio(unsigned pts)
{
pAudioCodecCtx_->coded_frame->pts = pts; // Set the time stamp
// simple sound encoding
int16_t samples[220] = { 0 }; // buffer
int n; // buffer index
double Fs = 44100.0; // sampling frequency
// Generate audio data
for (n = 0; n < 220; ++n) //220 samples (44100*.005sec as the interval between 2 video frames is 10ms)
samples[n] = 16383.0 * sin(n*1000.0*2.0*M_PI / Fs); //sine wav
int out_size = ffmpeg::avcodec_encode_audio(pAudioCodecCtx_, audio_outbuf, audio_outbuf_size, (const short*)samples);
pAudioCodecCtx_->coded_frame->pts = pts; // Set the time stamp
if (out_size>0)
{
// Packet
ffmpeg::AVPacket pkt = { 0 };
av_init_packet(&pkt);
pkt.data = NULL; // packet data will be allocated by the encoder
pkt.size = 0;
if (pAudioCodecCtx_->coded_frame->pts != (0x8000000000000000LL))
pkt.pts = av_rescale_q(pAudioCodecCtx_->coded_frame->pts, pAudioCodecCtx_->time_base, pAudioStream_->time_base);
if (pAudioCodecCtx_->coded_frame->key_frame)
pkt.flags |= AV_PKT_FLAG_KEY;
pkt.stream_index = pAudioStream_->index;
pkt.data = audio_outbuf;
pkt.size = out_size;
int ret = av_interleaved_write_frame(pOutputFormatCtx_, &pkt);
if (ret<0)
return -1;
av_free_packet(&pkt);
}
//end simple sound encoding
return pkt.size;
}
结果是一个不错的视频，后面有一些音频（或者是定期发出的哔哔声，但比视频更早结束，或者是持续时间比视频短的连续较长的声音）。
每次调用函数 encodeAudio() 时，我想以不规则的间隔生成哔声。我尝试修改采样率、缓冲区大小、pkt 大小和样本数，但没有任何成功。我也尝试在不同的时间设置积分，但它并没有让我到达我想去的地方。有人可以帮忙吗？