原文链接：https://blog.csdn.net/weixin_41176628/article/details/104047425
一、实现思路
ffmpeg的滤波filter有amix这个混音器，可以借助它来做音频的多路混音。首先我们需要编译ffmpeg并安装它，可以具备编码功能，考虑可以加mp3或者aac编码库进来，最简单的方式是编码成pcm格式直接输出到文件，用VLC也可以播放。
二、编译ffmpeg
$: git clone https://git.ffmpeg.org/ffmpeg.git
$: cd ffmpeg
$: git tag //打印出git提交标签，我们选择n2.9-dev 这个版本
$: git checkout n2.9-dev -b n2.9-dev //新建一个分支并切换到n2.9-dev这个版本
$: ./configure --disable-yasm --disable-ffplay --disable-ffprobe --disable-debug --enable-avdevice --disable-devices --disable-zlib --disable-bzlib --enable-avfilter --enable-encoders --enable-ffmpeg --enable-static --enable-gpl --enable-small --enable-nonfree --prefix=/home/xxx/ffmpegInstall_linux //编译它使用静态库编译（不编译动态库需要的自行添加即可）
$: make clean
$: make -j4
$: make install
三、简易框架
如图，即将两个mp3文件通过amix filter 混合成一个音频输出。
四、编写程序audioMix.c
1、构建输入封装器上下文input avformat
cd ~/ffmpegInstall_linux
mkdir src
新建audioMix.c文件并拷贝3个mp3文件进来
编译指令：
$: g++ -I../include -L../lib audioMix.cpp -o audioMix -lavfilter -lavformat -lavcodec -lavutil -lmp3lame -lswresample -lswscale -lavdevice -lpostproc -lpthread -lm -ldl
输入MP3格式的文件即可。或者其他编码的也行。
int OpenNetworkInput(char* inputForamt, const char* url)
{
AVInputFormat* ifmt = av_find_input_format(inputForamt);
AVDictionary* opt1 = NULL;
av_dict_set(&opt1, "rtbufsize", "10M", 0);
int ret = 0;
ret = avformat_open_input(&_fmt_ctx_net, url, ifmt, &opt1);
if (ret < 0)
{
printf("Speaker: failed to call avformat_open_input\n");
return -1;
}
ret = avformat_find_stream_info(_fmt_ctx_net, NULL);
if (ret < 0)
{
printf("Speaker: failed to call avformat_find_stream_info\n");
return -1;
}
for (int i = 0; i < _fmt_ctx_net->nb_streams; i++)
{
if (_fmt_ctx_net->streams[i]->codec->codec_type == AVMEDIA_TYPE_AUDIO)
{
_index_net = i;
break;
}
}
if (_index_net < 0)
{
printf("Speaker: negative audio index\n");
return -1;
}
AVCodecContext* codec_ctx = _fmt_ctx_net->streams[_index_net]->codec;
AVCodec* codec = avcodec_find_decoder(codec_ctx->codec_id);
if (codec == NULL)
{
printf("Speaker: null audio decoder\n");
return -1;
}
ret = avcodec_open2(codec_ctx, codec, NULL);
if (ret < 0)
{
printf("Speaker: failed to call avcodec_open2\n");
return -1;
}
av_dump_format(_fmt_ctx_net, _index_net, url, 0);
return 0;
}
2、构建输出封装器上下文
output format这里输出的是44100 s16 pcm 小端序格式的音频。需要其他格式的音频要改AV_CODEC_ID_PCM_S16LE 。并且需要添加编码库，ffmpeg内部不集成mp3或者aac。
这里还会构建几组fifo来存储input raw data。
int OpenFileOutput(char* fileName)
{
int ret = 0;
ret = avformat_alloc_output_context2(&_fmt_ctx_out, NULL, NULL, fileName);
if (ret < 0)
{
printf("Mixer: failed to call avformat_alloc_output_context2\n");
return -1;
}
AVStream* stream_a = NULL;
stream_a = avformat_new_stream(_fmt_ctx_out, NULL);
if (stream_a == NULL)
{
printf("Mixer: failed to call avformat_new_stream\n");
return -1;
}
_index_a_out = 0;
stream_a->codec->codec_type = AVMEDIA_TYPE_AUDIO;
AVCodec* codec_mp3 = NULL;
codec_mp3 = avcodec_find_encoder(AV_CODEC_ID_PCM_S16LE);
if (codec_mp3 == NULL)
{
printf("Mixer: failed to call avcodec_find_encoder ++++\n");
return -1;
}
stream_a->codec->codec = codec_mp3;
stream_a->codec->sample_rate = 44100;
stream_a->codec->channels = 2;
stream_a->codec->channel_layout = av_get_default_channel_layout(2);
stream_a->codec->sample_fmt = codec_mp3->sample_fmts[0];
stream_a->codec->bit_rate = 320000;
stream_a->codec->time_base.num = 1;
stream_a->codec->time_base.den = stream_a->codec->sample_rate;
stream_a->codec->codec_tag = 0;
if (_fmt_ctx_out->oformat->flags & AVFMT_GLOBALHEADER)
stream_a->codec->flags |= CODEC_FLAG_GLOBAL_HEADER;
if (avcodec_open2(stream_a->codec, stream_a->codec->codec, NULL) < 0)
{
printf("Mixer: failed to call avcodec_open2\n");
return -1;
}
if (!(_fmt_ctx_out->oformat->flags & AVFMT_NOFILE))
{
if (avio_open(&_fmt_ctx_out->pb, fileName, AVIO_FLAG_WRITE) < 0)
{
printf("Mixer: failed to call avio_open\n");
return -1;
}
}
if (avformat_write_header(_fmt_ctx_out, NULL) < 0)
{
printf("Mixer: failed to call avformat_write_header\n");
return -1;
}
bool b = (!_fmt_ctx_out->streams[0]->time_base.num && _fmt_ctx_out->streams[0]->codec->time_base.num);
av_dump_format(_fmt_ctx_out, _index_a_out, fileName, 1);
_fifo_net = av_audio_fifo_alloc(_fmt_ctx_net->streams[_index_net]->codec->sample_fmt, _fmt_ctx_net->streams[_index_net]->codec->channels, 30*_fmt_ctx_net->streams[_index_net]->codec->frame_size);
_fifo_spk = av_audio_fifo_alloc(_fmt_ctx_spk->streams[_index_spk]->codec->sample_fmt, _fmt_ctx_spk->streams[_index_spk]->codec->channels, 30*_fmt_ctx_spk->streams[_index_spk]->codec->frame_size);
_fifo_mic = av_audio_fifo_alloc(_fmt_ctx_mic->streams[_index_mic]->codec->sample_fmt, _fmt_ctx_mic->streams[_index_mic]->codec->channels, 30*_fmt_ctx_mic->streams[_index_spk]->codec->frame_size);
return 0;
}
3、构建in filter ，out filter 和filter graph（滤波过程）上下文
int InitFilter(const char* filter_desc)
{
char args_spk[512];
const char* pad_name_spk = "in0";
char args_mic[512];
const char* pad_name_mic = "in1";
char args_net[512];
const char* pad_name_net = "in2";
char args_vol[512] = "-3dB";
const char* pad_name_vol = "vol";
AVFilter* filter_src_spk = NULL;
filter_src_spk = avfilter_get_by_name("abuffer");
if(filter_src_spk == NULL)
{
printf("fail to get avfilter ..\n");
return -1;
}
AVFilter* filter_src_mic = NULL;
filter_src_mic = avfilter_get_by_name("abuffer");
if(filter_src_mic == NULL)
{
printf("fail to get avfilter ..\n");
return -1;
}
AVFilter* filter_src_net = NULL;
filter_src_net = avfilter_get_by_name("abuffer");
if(filter_src_net == NULL)
{
printf("fail to get avfilter ..\n");
return -1;
}
AVFilter* filter_src_vol = NULL;
filter_src_vol = avfilter_get_by_name("volume");
if(filter_src_vol == NULL)
{
printf("fail to get avfilter ..\n");
return -1;
}
AVFilter* filter_sink = avfilter_get_by_name("abuffersink");
AVFilterInOut* filter_output_spk = avfilter_inout_alloc();
AVFilterInOut* filter_output_mic = avfilter_inout_alloc();
AVFilterInOut* filter_output_net = avfilter_inout_alloc();
AVFilterInOut* filter_output_vol = avfilter_inout_alloc();
AVFilterInOut* filter_input = avfilter_inout_alloc();
_filter_graph = avfilter_graph_alloc();
//snprintf(args_spk, sizeof(args_spk), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%I64x",
snprintf(args_spk, sizeof(args_spk), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%"PRIx64,
_fmt_ctx_spk->streams[_index_spk]->codec->time_base.num,
_fmt_ctx_spk->streams[_index_spk]->codec->time_base.den,
_fmt_ctx_spk->streams[_index_spk]->codec->sample_rate,
av_get_sample_fmt_name(_fmt_ctx_spk->streams[_index_spk]->codec->sample_fmt),
_fmt_ctx_spk->streams[_index_spk]->codec->channel_layout);
printf("args spk : %s channel_layout : %d\n",args_spk,_fmt_ctx_spk->streams[_index_spk]->codec->channel_layout);
//snprintf(args_mic, sizeof(args_mic), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%I64x",
snprintf(args_mic, sizeof(args_mic), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%"PRIx64,
_fmt_ctx_mic->streams[_index_mic]->codec->time_base.num,
_fmt_ctx_mic->streams[_index_mic]->codec->time_base.den,
_fmt_ctx_mic->streams[_index_mic]->codec->sample_rate,
av_get_sample_fmt_name(_fmt_ctx_mic->streams[_index_mic]->codec->sample_fmt),
_fmt_ctx_mic->streams[_index_mic]->codec->channel_layout);
printf("args mic : %s channel_layout : %d\n",args_mic,_fmt_ctx_mic->streams[_index_mic]->codec->channel_layout);
//snprintf(args_mic, sizeof(args_mic), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%I64x",
snprintf(args_net, sizeof(args_net), "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=0x%"PRIx64,
_fmt_ctx_net->streams[_index_net]->codec->time_base.num,
_fmt_ctx_net->streams[_index_net]->codec->time_base.den,
_fmt_ctx_net->streams[_index_net]->codec->sample_rate,
av_get_sample_fmt_name(_fmt_ctx_net->streams[_index_net]->codec->sample_fmt),
_fmt_ctx_net->streams[_index_net]->codec->channel_layout);
printf("args net : %s channel_layout : %d\n",args_net,_fmt_ctx_net->streams[_index_net]->codec->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(&_filter_ctx_src_spk, filter_src_spk, pad_name_spk, args_spk, NULL, _filter_graph);
if (ret < 0)
{
printf("Filter: failed to call avfilter_graph_create_filter -- src spk\n");
return -1;
}
ret = avfilter_graph_create_filter(&_filter_ctx_src_mic, filter_src_mic, pad_name_mic, args_mic, NULL, _filter_graph);
if (ret < 0)
{
printf("Filter: failed to call avfilter_graph_create_filter -- src mic\n");
return -1;
}
ret = avfilter_graph_create_filter(&_filter_ctx_src_net, filter_src_net, pad_name_net, args_net, NULL, _filter_graph);
if (ret < 0)
{
printf("Filter: failed to call avfilter_graph_create_filter -- src net\n");
return -1;
}
ret = avfilter_graph_create_filter(&_filter_ctx_sink, filter_sink, "out", NULL, NULL, _filter_graph);
if (ret < 0)
{
printf("Filter: failed to call avfilter_graph_create_filter -- sink\n");
return -1;
}
AVCodecContext* encodec_ctx = _fmt_ctx_out->streams[_index_a_out]->codec;
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);
if (ret < 0)
{
printf("Filter: failed to call av_opt_set_bin -- sample_fmts\n");
return -1;
}
ret = av_opt_set_bin(_filter_ctx_sink, "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(_filter_ctx_sink, "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_spk->name = av_strdup(pad_name_spk);
filter_output_spk->filter_ctx = _filter_ctx_src_spk;
filter_output_spk->pad_idx = 0;
filter_output_spk->next = filter_output_mic;
filter_output_mic->name = av_strdup(pad_name_mic);
filter_output_mic->filter_ctx = _filter_ctx_src_mic;
filter_output_mic->pad_idx = 0;
filter_output_mic->next = filter_output_net; //重點！！！！！
filter_output_net->name = av_strdup(pad_name_net);
filter_output_net->filter_ctx = _filter_ctx_src_net;
filter_output_net->pad_idx = 0;
filter_output_net->next = NULL;
filter_output_vol->name = av_strdup(pad_name_vol);
filter_output_vol->filter_ctx = _filter_ctx_src_vol;
filter_output_vol->pad_idx = 0;
filter_output_vol->next = filter_input;
filter_input->name = av_strdup("out");
filter_input->filter_ctx = _filter_ctx_sink;
filter_input->pad_idx = 0;
filter_input->next = NULL;
AVFilterInOut* filter_outputs[3];
filter_outputs[0] = filter_output_spk;
filter_outputs[1] = filter_output_mic;
filter_outputs[2] = filter_output_net;
AVFilterInOut* filter_volinputs[2];
filter_volinputs[0] = filter_output_vol;
filter_volinputs[1] = filter_input;
ret = avfilter_graph_parse_ptr(_filter_graph, 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(_filter_graph, 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);
//av_free(filter_src_net);
avfilter_inout_free(filter_outputs);
//av_free(filter_outputs);
char* temp = avfilter_graph_dump(_filter_graph, NULL);
printf("%s\n", temp);
return 0;
}
这里需要注意const char* filter_desc 传进来的filter关系描述符
const char* filter_desc = "[in0][in1][in2]amix=inputs=3[out]";
解释一下，[in0] 这个对应上面的const char* pad_name_spk = “in0”;代表某一filter，[in0][in1][in2]代表的是输入节点，amix是混音滤波器=inputs=3这个是参数表示输入是3个，[out]表示输出。更加详细的解释可以参考[ffmpeg] 滤波这篇博客。这里的讲解比较清晰。
注意这段代码
filter_output_spk->name = av_strdup(pad_name_spk);
filter_output_spk->filter_ctx = _filter_ctx_src_spk;
filter_output_spk->pad_idx = 0;
filter_output_spk->next = filter_output_mic;
filter_output_mic->name = av_strdup(pad_name_mic);
filter_output_mic->filter_ctx = _filter_ctx_src_mic;
filter_output_mic->pad_idx = 0;
filter_output_mic->next = filter_output_net; //重點！！！！！
filter_output_net->name = av_strdup(pad_name_net);
filter_output_net->filter_ctx = _filter_ctx_src_net;
filter_output_net->pad_idx = 0;
filter_output_net->next = NULL;
filter_output_vol->name = av_strdup(pad_name_vol);
filter_output_vol->filter_ctx = _filter_ctx_src_vol;
filter_output_vol->pad_idx = 0;
filter_output_vol->next = filter_input;
filter_input->name = av_strdup("out");
filter_input->filter_ctx = _filter_ctx_sink;
filter_input->pad_idx = 0;
filter_input->next = NULL;
AVFilterInOut* filter_outputs[3];
filter_outputs[0] = filter_output_spk;
filter_outputs[1] = filter_output_mic;
filter_outputs[2] = filter_output_net;
这是有填充next的即将输入串联了起来。
4、构建输入线程
void * NetworkCapThreadProc(void *lpParam)
{
AVFrame* pFrame = av_frame_alloc();
AVPacket packet;
av_init_packet(&packet);
int got_sound;
while (_state == RUNNING)
{
packet.data = NULL;
packet.size = 0;
if (av_read_frame(_fmt_ctx_net, &packet) < 0)
{
printf("there is no context spk \n");
break;
}
if (packet.stream_index == _index_net)
{
if (avcodec_decode_audio4(_fmt_ctx_net->streams[_index_net]->codec, pFrame, &got_sound, &packet) < 0)
{
break;
}
av_free_packet(&packet);
if (!got_sound)
{
continue;
}
int fifo_net_space = av_audio_fifo_space(_fifo_net);
while(fifo_net_space < pFrame->nb_samples && _state == RUNNING)
{
usleep(10*1000);
printf("_fifo_spk full !\n");
fifo_net_space = av_audio_fifo_space(_fifo_net);
}
if (fifo_net_space >= pFrame->nb_samples)
{
//EnterCriticalSection(&_section_spk);
pthread_mutex_lock(&mutex_net);
int nWritten = av_audio_fifo_write(_fifo_net, (void**)pFrame->data, pFrame->nb_samples);
//LeaveCriticalSection(&_section_spk);
pthread_mutex_unlock(&mutex_net);
}
}
}
av_frame_free(&pFrame);
}
5、处理数据
while (_state != FINISHED)
{
if (kbhit())
{
_state = STOPPED;
break;
}
else
{
int ret = 0;
AVFrame* pFrame_spk = av_frame_alloc();
AVFrame* pFrame_mic = av_frame_alloc();
AVFrame* pFrame_net = av_frame_alloc();
AVPacket packet_out;
int got_packet_ptr = 0;
int fifo_spk_size = av_audio_fifo_size(_fifo_spk);
int fifo_mic_size = av_audio_fifo_size(_fifo_mic);
int fifo_net_size = av_audio_fifo_size(_fifo_net);
int frame_spk_min_size = _fmt_ctx_spk->streams[_index_spk]->codec->frame_size;
int frame_mic_min_size = _fmt_ctx_mic->streams[_index_mic]->codec->frame_size;
int frame_net_min_size = _fmt_ctx_net->streams[_index_net]->codec->frame_size;
printf("get mini frame size : %d ###################\n",frame_spk_min_size);
if (fifo_spk_size >= frame_spk_min_size && fifo_mic_size >= frame_mic_min_size && fifo_net_size >= frame_net_min_size)
{
tmpFifoFailed = 0;
pFrame_spk->nb_samples = frame_spk_min_size;
pFrame_spk->channel_layout = _fmt_ctx_spk->streams[_index_spk]->codec->channel_layout;
pFrame_spk->format = _fmt_ctx_spk->streams[_index_spk]->codec->sample_fmt;
pFrame_spk->sample_rate = _fmt_ctx_spk->streams[_index_spk]->codec->sample_rate;
av_frame_get_buffer(pFrame_spk, 0);
pFrame_mic->nb_samples = frame_mic_min_size;
pFrame_mic->channel_layout = _fmt_ctx_mic->streams[_index_mic]->codec->channel_layout;
pFrame_mic->format = _fmt_ctx_mic->streams[_index_mic]->codec->sample_fmt;
pFrame_mic->sample_rate = _fmt_ctx_mic->streams[_index_mic]->codec->sample_rate;
av_frame_get_buffer(pFrame_mic, 0);
pFrame_net->nb_samples = frame_net_min_size;
pFrame_net->channel_layout = _fmt_ctx_net->streams[_index_net]->codec->channel_layout;
pFrame_net->format = _fmt_ctx_net->streams[_index_net]->codec->sample_fmt;
pFrame_net->sample_rate = _fmt_ctx_net->streams[_index_net]->codec->sample_rate;
av_frame_get_buffer(pFrame_net, 0);
pthread_mutex_lock(&mutex_spk);
//EnterCriticalSection(&_section_spk);
ret = av_audio_fifo_read(_fifo_spk, (void**)pFrame_spk->data, frame_spk_min_size);
//LeaveCriticalSection(&_section_spk);
pthread_mutex_unlock(&mutex_spk);
pthread_mutex_lock(&mutex_mic);
//EnterCriticalSection(&_section_mic);
ret = av_audio_fifo_read(_fifo_mic, (void**)pFrame_mic->data, frame_mic_min_size);
//LeaveCriticalSection(&_section_mic);
pthread_mutex_unlock(&mutex_mic);
pthread_mutex_lock(&mutex_net);
//EnterCriticalSection(&_section_net);
ret = av_audio_fifo_read(_fifo_net, (void**)pFrame_net->data, frame_net_min_size);
//LeaveCriticalSection(&_section_net);
pthread_mutex_unlock(&mutex_net);
pFrame_spk->pts = av_frame_get_best_effort_timestamp(pFrame_spk);
pFrame_mic->pts = av_frame_get_best_effort_timestamp(pFrame_mic);
pFrame_net->pts = av_frame_get_best_effort_timestamp(pFrame_net);
BufferSourceContext* s = (BufferSourceContext*)_filter_ctx_src_spk->priv;
bool b1 = (s->sample_fmt != pFrame_spk->format);
bool b2 = (s->sample_rate != pFrame_spk->sample_rate);
bool b3 = (s->channel_layout != pFrame_spk->channel_layout);
bool b4 = (s->channels != pFrame_spk->channels);
ret = av_buffersrc_add_frame(_filter_ctx_src_spk, pFrame_spk);
if (ret < 0)
{
printf("Mixer: failed to call av_buffersrc_add_frame (speaker)\n");
break;
}
ret = av_buffersrc_add_frame(_filter_ctx_src_mic, pFrame_mic);
if (ret < 0)
{
printf("Mixer: failed to call av_buffersrc_add_frame (microphone)\n");
break;
}
ret = av_buffersrc_add_frame(_filter_ctx_src_net, pFrame_net);
if (ret < 0)
{
printf("Mixer: failed to call av_buffersrc_add_frame (network)\n");
break;
}
while (1)
{
AVFrame* pFrame_out = av_frame_alloc();
//AVERROR(EAGAIN) 返回这个表示还没转换完成既 不存在帧，则返回AVERROR（EAGAIN）
ret = av_buffersink_get_frame_flags(_filter_ctx_sink, pFrame_out, 0 );
if(ret == AVERROR(EAGAIN) || ret == AVERROR_EOF)
{
//printf("%d %d \n",AVERROR(EAGAIN),AVERROR_EOF);
}
if (ret < 0)
{
printf("Mixer: failed to call av_buffersink_get_frame_flags ret : %d \n",ret);
break;
}
if (pFrame_out->data[0] != NULL)
{
av_init_packet(&packet_out);
packet_out.data = NULL;
packet_out.size = 0;
ret = avcodec_encode_audio2(_fmt_ctx_out->streams[_index_a_out]->codec, &packet_out, pFrame_out, &got_packet_ptr);
if (ret < 0)
{
printf("Mixer: failed to call avcodec_decode_audio4\n");
break;
}
if (got_packet_ptr)
{
packet_out.stream_index = _index_a_out;
packet_out.pts = frame_count * _fmt_ctx_out->streams[_index_a_out]->codec->frame_size;
packet_out.dts = packet_out.pts;
packet_out.duration = _fmt_ctx_out->streams[_index_a_out]->codec->frame_size;
packet_out.pts = av_rescale_q_rnd(packet_out.pts,
_fmt_ctx_out->streams[_index_a_out]->codec->time_base,
_fmt_ctx_out->streams[_index_a_out]->time_base,
(AVRounding)(AV_ROUND_NEAR_INF | AV_ROUND_PASS_MINMAX));
packet_out.dts = packet_out.pts;
packet_out.duration = av_rescale_q_rnd(packet_out.duration,
_fmt_ctx_out->streams[_index_a_out]->codec->time_base,
_fmt_ctx_out->streams[_index_a_out]->time_base,
(AVRounding)(AV_ROUND_NEAR_INF | AV_ROUND_PASS_MINMAX));
frame_count++;
ret = av_interleaved_write_frame(_fmt_ctx_out, &packet_out);
if (ret < 0)
{
printf("Mixer: failed to call av_interleaved_write_frame\n");
}
//printf("Mixer: write frame to file\n");
}
av_free_packet(&packet_out);
}
av_frame_free(&pFrame_out);
}
}
else
{
tmpFifoFailed++;
usleep(20*1000);
if (tmpFifoFailed > 300)
{
_state = STOPPED;
usleep(30*1000);
break;
}
}
av_frame_free(&pFrame_spk);
av_frame_free(&pFrame_mic);
}
}
完整代码请下载audioMix.c
>>> 音视频开发 视频教程： https://ke.qq.com/course/3202131?flowToken=1031864
>>> 音视频开发学习资料、教学视频，免费分享有需要的可以自行添加学习交流群 739729163 领取