diff options
Diffstat (limited to 'rotord/rotor.cpp')
| -rw-r--r-- | rotord/rotor.cpp | 60 |
1 files changed, 55 insertions, 5 deletions
diff --git a/rotord/rotor.cpp b/rotord/rotor.cpp index 2ab5734..53ba33b 100644 --- a/rotord/rotor.cpp +++ b/rotord/rotor.cpp @@ -431,9 +431,31 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process */ - http://www.gamedev.net/topic/624876-how-to-read-an-audio-file-with-ffmpeg-in-c/ + //http://www.gamedev.net/topic/624876-how-to-read-an-audio-file-with-ffmpeg-in-c/ // Initialize FFmpeg + + //this seems to only read a certain percentage of the sound file? - different for wav and mp3 + //peculiar, because the data comes in as frames and we just read them until all are done + //is it possible that thhe problem lies with the drawing process only? + //wav sees 6/10 waves and mp3 sees 3.5/10 + //could this be an aliasing effect i.e. from reading the sample format wrong? + //I think probably not? + + //every format combination seems to miss a different number of samples + //I don't understand how different formats would make the samples seem to truncate? + //unless its an aliasing type thing? + + //wave.mp3 reads 16.409% of 160913 + //wave.wav reads 29.706% of 882044 + + + av_register_all(); + + //?? + //avcodec_init(); + avcodec_register_all(); + //?? AVFrame* frame = avcodec_alloc_frame(); if (!frame) @@ -497,6 +519,12 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process std::cout << "Couldn't open the context with the decoder" << std::endl; return false; } + + // + // why is the file truncated??? + //if(codecContext->codec->capabilities & CODEC_CAP_TRUNCATED) codecContext->codec->capabilities|=CODEC_FLAG_TRUNCATED; + // + // av_dump_format(formatContext, 0, 0, false); //avformat.h line 1256 int samples = ((formatContext->duration + 5000)*codecContext->sample_rate)/AV_TIME_BASE; @@ -514,10 +542,19 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process AVPacket packet; av_init_packet(&packet); int sample_processed=0; + + bool diag=true; // Read the packets in a loop - while (av_read_frame(formatContext, &packet) == 0) + while (true) + //while(sample_processed<samples) { + int ret=av_read_frame(formatContext, &packet); + if (ret<0) { + cerr << "finished with code "<<ret <<(ret==AVERROR_EOF?" ,EOF":"")<<endl; + break; + } + //av_read_frame(formatContext, &packet); //hangs once the packets have been read if (packet.stream_index == audioStream->index) { // Try to decode the packet into a frame @@ -535,12 +572,20 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process // the audio data, so I won't add any junk here that might confuse you. Typically, if I want to find // documentation on an FFmpeg structure or function, I just type "<name> doxygen" into google (like // "AVFrame doxygen" for AVFrame's docs) + + //av_get_channel_layout_string (char *buf, int buf_size, int nb_channels, uint64_t channel_layout) + + if (diag) { + cerr << "first frame: "<<bytes << ", "<<frame->nb_samples<<" samples in "<<av_get_sample_fmt_name(frame->format)<<" format with channel layout "<<frame->channel_layout<< std::endl; + diag=false; + } //std::cout << "Got a frame: bytes " << bytes << ", "<<frame->nb_samples<<" samples"<<std::endl; //now we can pass the data to the processor(s) for (auto p: processors) { sample_processed=p->process_frame(frame->data[0],frame->nb_samples); } + mutex.lock(); progress=((double)sample_processed)/samples; mutex.unlock(); @@ -568,6 +613,9 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process mutex.unlock(); } } + + cerr << "finished processed: "<<sample_processed << " samples of "<<samples<<" , "<<((double)sample_processed*100)/samples<<"%"<< std::endl; + // Clean up! av_free(frame); @@ -638,6 +686,8 @@ void audio_thumbnailer::init(int _channels,int _bits,int _samples) { channels=_channels; bits=_bits; samples=_samples; + + cerr << "init audio thumbnailer with "<<channels<<" channels, "<<samples<<" samples of "<< bits<<" bits."<<endl; samples_per_column=samples/width; @@ -647,10 +697,10 @@ void audio_thumbnailer::init(int _channels,int _bits,int _samples) { scale=1.0/offset; } int audio_thumbnailer::process_frame(uint8_t *_data,int samples_in_frame){ - //begin by processing remaining samplesthread 4 - //samples per column could be larger than a frame! (probably is) + //begin by processing remaining samples + //samples per column could be larger than a frame! (probably is) //but all we are doing is averaging - int bytes=(bits>>3); + int bytes=(bits>>3); int stride=channels*bytes; int in_sample=0; while (in_sample<samples_in_frame&&column<width) { |