working on avcodec interface

3 files changed, 53 insertions, 13 deletions

diff --git a/rotord/rotor.cpp b/rotord/rotor.cpp
index 931bfa9..0a3a502 100644
--- a/rotord/rotor.cpp
+++ b/rotord/rotor.cpp
@@ -32,9 +32,10 @@ void Render_context::runTask() {
 			//audio_analyser.process(audio_filename);
 			//vampHost::runPlugin("","qm-vamp-plugins","qm-tempotracker", "",0, audio_filename, cerr,true);
 			vector<base_audio_processor*> proc;
-			load_audio(audio_filename,proc);
-			
-			state=AUDIO_READY;
+			if (load_audio(audio_filename,proc)) {
+				state=AUDIO_READY;
+			}
+			else state=IDLE;
 		}
 		sleep(100);
 	}
@@ -269,7 +270,7 @@ Command_response Render_context::session_command(const std::vector<std::string>&
 //http://blog.tomaka17.com/2012/03/libavcodeclibavformat-tutorial/
 //great to use c++11 features
 
-bool Render_context::load_audio(string &filename,vector<base_audio_processor*> processors){
+bool Render_context::load_audio(const string &filename,vector<base_audio_processor*> processors){
 	//load audio data from file
 	//what's the best way to use this? the model is background processing, and we want to update a progress bar
 	//could pass a function pointer to call when each chunk of data becomes available?
@@ -279,7 +280,17 @@ bool Render_context::load_audio(string &filename,vector<base_audio_processor*> p
 	//there could even be an array of audio analysis functions to perform simultaneously?
 	//how about a vector of  objects that subclass the base audio processor class?
 	
-	 av_register_all();
+	//1st get parameters and initialise the processors
+	//then begin data loop locking progress variable after each frame
+	
+	//
+	//
+	//the example in ffmpeg works, but it isn't one that identifies a codec- it is hard coded to look for a codec for AV_CODEC_ID_MP2
+	//it also doesn't load through libavformat - which opens containers- it just loads a naked .mp2 stream
+	//
+	//
+	
+	av_register_all();
 	
 	std::shared_ptr<AVFormatContext> avFormat(avformat_alloc_context(), &avformat_free_context);
 	
@@ -294,7 +305,7 @@ bool Render_context::load_audio(string &filename,vector<base_audio_processor*> p
 		return false;
 	}
 	
-	av_dump_format(avFormat.get(), 0, 0, false);
+	av_dump_format(avFormat.get(), 0, 0, false); //avformat.h line 1256
 
 	AVStream* stream = nullptr;
 	for (unsigned int i = 0; i < avFormat->nb_streams; ++i) {
@@ -337,7 +348,6 @@ bool Render_context::load_audio(string &filename,vector<base_audio_processor*> p
 	
 	//avcodec.h line 1026
 	
-	cerr << "audio codec context -  sample rate: "<< audioCodec->sample_rate <<" ,channels: "<<audioCodec->channels<<" ,sample format: "<<audioCodec->sample_fmt<<endl;
 	
 	Packet packet(avFormat.get());
 	if (packet.packet.data == nullptr) {
@@ -345,6 +355,9 @@ bool Render_context::load_audio(string &filename,vector<base_audio_processor*> p
 		return true;
 	}
 	
+	cerr << "audio codec context -  sample rate: "<< audioCodec->sample_rate <<", channels: "<<audioCodec->channels<<", sample format: "<<audioCodec->sample_fmt<<endl;
+	
+	
 	do {
 		packet.reset(avFormat.get());
 		if (packet.packet.stream_index != stream->index)
@@ -360,18 +373,25 @@ bool Render_context::load_audio(string &filename,vector<base_audio_processor*> p
 	size_t offsetInData = 0;
 	
 	
-
-	/*
+	bool foundPacket=false;
+	
 	// the decoding loop, running until EOF
 	while (true) {
 		// reading a packet using libavformat
 		if (offsetInData >= packet.packet.size) {
 			do {
 				packet.reset(avFormat.get());
-				if (packet.stream_index != videoStream->index)
+				if (packet.packet.stream_index != stream->index)
 					continue;
 			} while(0);
+			if (!foundPacket){
+				cerr << "audio codec context -  sample rate: "<< audioCodec->sample_rate <<", channels: "<<audioCodec->channels<<", sample format: "<<audioCodec->sample_fmt<<endl;
+				foundPacket=true;
+			}
+			
 		}
+		
+		
 
 		// preparing the packet that we will send to libavcodec
 		AVPacket packetToSend;
@@ -380,13 +400,17 @@ bool Render_context::load_audio(string &filename,vector<base_audio_processor*> p
 		
 		// sending data to libavcodec
 		int isFrameAvailable = 0;
-		const auto processedLength = avcodec_decode_video2(avVideoCodec.get(), avFrame.get(), &isFrameAvailable, &packetToSend);
+		//const auto processedLength = avcodec_decode_video2(avVideoCodec.get(), avFrame.get(), &isFrameAvailable, &packetToSend);
+		const auto processedLength = avcodec_decode_audio4(audioCodec.get(), avFrame.get(), &isFrameAvailable, &packetToSend);
+
+		
 		if (processedLength < 0) {
 			av_free_packet(&packet);
 			throw std::runtime_error("Error while processing the data");
 		}
 		offsetInData += processedLength;
 
+		/*
 		// processing the image if available
 		if (isFrameAvailable) {
 			// display image on the screen
@@ -395,8 +419,10 @@ bool Render_context::load_audio(string &filename,vector<base_audio_processor*> p
 			const auto msToWait = avVideoContext->ticks_per_frame * 1000 * avVideoContext->time_base.num / avVideoContext->time_base.den;
 			std::this_thread::sleep(std::chrono::milliseconds(msToWait));
 		}
+		*/
+		
 	}
-	*/
+
 
 	return true;
 }
diff --git a/rotord/rotor.h b/rotord/rotor.h
index c7d25b2..5667b62 100755
--- a/rotord/rotor.h
+++ b/rotord/rotor.h
@@ -359,7 +359,7 @@ namespace Rotor {
 			int load_graph(Poco::UUID uid);
 			bool load_graph(string &graph_filename); 		//should eventually be as above
 			UUID save_graph();							//returns UUID of saved graph
-			bool load_audio(string &filename,vector<base_audio_processor*> processors);
+			bool load_audio(const string &filename,vector<base_audio_processor*> processors);
 			Render_requirements get_requirements();
 			int load_video(int num,string &filename);		//can be performance or clip
 
diff --git a/rotord/rotord.cpp b/rotord/rotord.cpp
index f0bcc0d..734e159 100755
--- a/rotord/rotord.cpp
+++ b/rotord/rotord.cpp
@@ -29,6 +29,20 @@ save a movie
 
 how do we deal with frames in libavcodec - can we request a frame by number
 how do we deal with resolution
+
+
+avcodec class
+what does it have to do?
+
+-open files (could be video, audio, or both)
+-process files i.e. thumbnail the audio or do audio analysis, make a no-keyframe proxy of a movie
+-retrieve video frames (with caching- what's the best way to approach this)
+
+just for a quick think through: signal data is trivial: 1 floating point number per frame.
+to store a whole uncompressed video track for a 3 minute @ 720.25p in RGB \= 12GB
+caching the whole thing isn't a goer
+
+
 */
 
 RotorRequestHandler::RotorRequestHandler(const std::string& format): _format(format){