nearly working thumbnailer

3 files changed, 124 insertions, 136 deletions

diff --git a/rotord/rotor.cpp b/rotord/rotor.cpp
index de9f4ea..85382ea 100644
--- a/rotord/rotor.cpp
+++ b/rotord/rotor.cpp
@@ -6,7 +6,7 @@ bool fequal(const float u,const float v){
 	if (abs(u-v)<.001) return true;
 	else return false;
 };
-	
+
 
 using namespace Rotor;
 
@@ -17,7 +17,7 @@ using namespace Rotor;
 	string output="";
 	int outputNo=0;
 	*/
-	
+
 void Render_context::runTask() {
 	while (!isCancelled()) {
 		int cmd=0;
@@ -32,7 +32,7 @@ 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;
-			proc.push_back(new audio_thumbnailer());
+			proc.push_back(audio_thumb);
 			if (load_audio(audio_filename,proc)) {
 				state=AUDIO_READY;
 				//set the response
@@ -83,8 +83,8 @@ Command_response Render_context::session_command(const std::vector<std::string>&
 												//check file exists
 			Poco::File f=Poco::File(command[3]);
 			//std::auto_ptr<std::istream> pStr(URIStreamOpener::defaultOpener().open(command[3]));
-			
-			
+
+
 			if (f.exists()) {
 												//pass to worker thread ??if engine is ready?? ??what if engine has finished but results aren't read??
 			    audio_filename=command[3]; 			//for now, store session variables in memory
@@ -95,7 +95,7 @@ Command_response Render_context::session_command(const std::vector<std::string>&
 			    response.status=HTTPResponse::HTTP_NOT_FOUND;
 			    response.description="<status context='"+command[1]+"'>File "+command[3]+" not found</status>\n";
 			}
-			
+
 		    }
 		    else {
 			response.status=HTTPResponse::HTTP_BAD_REQUEST;
@@ -108,24 +108,20 @@ Command_response Render_context::session_command(const std::vector<std::string>&
 			     response.status=HTTPResponse::HTTP_OK;
 			     response.description="<status context='"+command[1]+"'>Rotor: analysing audio</status>\n";
 			     char c[20];
-			     sprintf(c,"%02f",audio_analyser.get_progress());
+			     sprintf(c,"%02f",progress);
 			     response.description+="<progress>"+string(c)+"</progress>\n";
 		     }
-		     if (state==AUDIO_READY) {
-			     //not sure about this-- should this state be retained?
-			     //can the data  only be read once?
-			     //for now
-			     response.status=HTTPResponse::HTTP_OK;
-			     response.description="<status context='"+command[1]+"'>Rotor: audio ready</status>\n";
-			      response.description+="<beats>";
-			     for (auto& i: audio_analyser.beats) { //is actually giving no data?
-				     char c[20];
-				     sprintf(c,"%02f",i);
-				     response.description+="<beat>"+string(c)+"</beat>";
-				}     
-			      response.description+="\n</beats>";
-			     state=IDLE;
-		     }
+			if (state==AUDIO_READY) {
+				//not sure about this-- should this state be retained?
+				//can the data  only be read once?
+				//for now
+				response.status=HTTPResponse::HTTP_OK;
+				response.description="<status context='"+command[1]+"'>Rotor: audio ready</status>\n";
+				response.description+="<audio>\n";
+				response.description+=audio_thumb->print();
+				response.description+="\n</audio>";
+				state=IDLE;
+			}
 	     }
 	     if (command[0]=="DELETE") {
 		     //for now
@@ -151,7 +147,7 @@ Command_response Render_context::session_command(const std::vector<std::string>&
 				if (state==IDLE) {					//eventually not like this
 					Poco::File f=Poco::File(command[3]);
 					if (f.exists()) {
-						string graph_filename=command[3]; 
+						string graph_filename=command[3];
 						if (load_graph(graph_filename)) {
 							response.status=HTTPResponse::HTTP_OK;
 							//response.description="<status context='"+command[1]+"'>Rotor: loaded graph "+command[3]+"</status>\n";
@@ -241,7 +237,7 @@ Command_response Render_context::session_command(const std::vector<std::string>&
 			response.description="<status>DUMMY RESPONSE 1</status>\n";
 			response.status=HTTPResponse::HTTP_OK;
 	     }
-		
+
 	}
 	if (command[2]=="render") {
 		if (command[0]=="GET") {
@@ -281,22 +277,22 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 	//in the case of audio analysis, the daemon will pass each audio analysis object each chunk of data as it gets it
 	//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?
-	
+
 	//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);
-	
+
 	auto avFormatPtr = avFormat.get();
 	if (avformat_open_input(&avFormatPtr,filename.c_str(),nullptr, nullptr) != 0) {
 		cerr <<"Rotor: Error while calling avformat_open_input (probably invalid file format)" << endl;
@@ -307,11 +303,11 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 		cerr << "Rotor: Error while calling avformat_find_stream_info" << endl;
 		return 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) {
 		if (avFormat->streams[i]->codec->codec_type == AVMEDIA_TYPE_AUDIO) {
@@ -324,16 +320,16 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 		cerr <<"Rotor: Didn't find any audio stream in the file"<< endl;
 		return false;
 	}
-	
+
 	// getting the required codec structure
 	const auto codec = avcodec_find_decoder(stream->codec->codec_id); //returns AVCodec*
 	if (codec == nullptr) {
 		cerr <<"Rotor: Audio codec not available"<< endl;
 		return false;
 	}
-	
-	//AVCodecContext??  avFormat->streams[i]->codec 
-	
+
+	//AVCodecContext??  avFormat->streams[i]->codec
+
 	// allocating a structure
 	std::shared_ptr<AVCodecContext> audioCodec(avcodec_alloc_context3(codec), [](AVCodecContext* c) { avcodec_close(c); av_free(c); });
 
@@ -342,7 +338,7 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 	// make sure that this vector exists as long as the avVideoCodec exists
 	std::vector<uint8_t> codecContextExtraData;
 	codecContextExtraData.assign(stream->codec->extradata, stream->codec->extradata + stream->codec->extradata_size);
-	
+
 	audioCodec->extradata = reinterpret_cast<uint8_t*>(codecContextExtraData.data());
 	audioCodec->extradata_size = codecContextExtraData.size();
 
@@ -351,23 +347,23 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 		cerr <<"Rotor: Could not open codec"<< endl;
 		return false;
 	}
-	
-	
+
+
 	//avcodec.h line 1026
-	
-	
+
+
 	Packet packet;
 
 // decoding code here
 	//cerr << "audio codec context -  sample rate: "<< audioCodec->sample_rate <<", channels: "<<audioCodec->channels<<", sample format: "<<audioCodec->sample_fmt<<endl;
-	
+
 	// suspect why was this in twice 210313
 	//do {
 	//	if (packet.packet.stream_index != stream->index)
 	///		continue;	// the packet is not about the stream we want, let's jump again the start of the loop
 	//} while(0);
 	//
-	
+
 	// allocating an AVFrame
 	std::shared_ptr<AVFrame> avFrame(avcodec_alloc_frame(), &av_free);
 
@@ -375,10 +371,10 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 	//Packet packet;
 	// data in the packet of data already processed
 	size_t offsetInData = 0;
-	
-	
+
+
 	bool foundPacket=false;
-	
+
 	// the decoding loop, running until EOF
 	while (true) {
 		// reading a packet using libavformat
@@ -392,20 +388,20 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 				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;
 		packetToSend.data = packet.packet.data + offsetInData;
 		packetToSend.size = packet.packet.size - offsetInData;
-		
+
 		// sending data to libavcodec
 		int isFrameAvailable = 0;
 		//const auto processedLength = avcodec_decode_video2(avVideoCodec.get(), avFrame.get(), &isFrameAvailable, &packetToSend);
-		
+
 		const auto processedLength = avcodec_decode_audio4(audioCodec.get(), avFrame.get(), &isFrameAvailable, &packetToSend);
 		//utils.c line 2018
 
@@ -414,16 +410,16 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 		//unsigned int numberOfSamplesContainedInTheFrame = stream.frame->nb_samples * stream.codecContext->channels;
 		//unsigned int numberOfSamplesToPlayPerSecond = stream.codecContext->sample_rate;   // usually 44100 or 48000
 		//const void* data = stream.frame->data[0];
-		
-		
+
+
 		if (processedLength < 0) {
 			//av_free_packet(&packet); shouldn't have to because of type safe wrapper
 			cerr <<"Error "<<processedLength<< " while processing the data"<< endl;
 			return false;
 		}
 		offsetInData += processedLength;
-		
-		
+
+
 		// processing the image if available
 		if (isFrameAvailable) {
 			// display image on the screen
@@ -433,8 +429,8 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 			std::this_thread::sleep(std::chrono::milliseconds(msToWait));
 		}
 		*/
-		
-		
+
+
 		http://www.gamedev.net/topic/624876-how-to-read-an-audio-file-with-ffmpeg-in-c/
 		// Initialize FFmpeg
 	av_register_all();
@@ -455,7 +451,7 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 		std::cout << "Error opening the file" << std::endl;
 		return false;
 	}
-	
+
 
 	if (avformat_find_stream_info(formatContext, NULL) < 0)
 	{
@@ -504,22 +500,20 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 
 	av_dump_format(formatContext, 0, 0, false); //avformat.h line 1256
 	int samples = ((formatContext->duration + 5000)*codecContext->sample_rate)/AV_TIME_BASE;
-	
+
 	std::cout << "This stream has " << codecContext->channels << " channels, a sample rate of " << codecContext->sample_rate << "Hz and "<<samples <<" samples" << std::endl;
 	std::cout << "The data is in format " <<codecContext->sample_fmt<< " (aka "<< av_get_sample_fmt_name(codecContext->sample_fmt) << ") "<<std::endl;
-	
+
 	//we can now tell the processors the format
 	//we can work out the number of samples at this point
-	
-	
-	
-	
+
 	for (auto p: processors) {
 		p->init(codecContext->channels,16,samples);
 	}
 
 	AVPacket packet;
 	av_init_packet(&packet);
+	int sample_processed=0;
 
 	// Read the packets in a loop
 	while (av_read_frame(formatContext, &packet) == 0)
@@ -541,18 +535,19 @@ 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)
-				
+
 				//std::cout << "Got a frame: bytes " << bytes << std::endl;
 				//now we can pass the data to the processor(s)
 				for (auto p: processors) {
-					p->process_frame(frame->data[0],frame->nb_samples);
+					sample_processed=p->process_frame(frame->data[0],frame->nb_samples);
 				}
-				
+				mutex.lock();
+				progress=((double)sample_processed)/samples;
+				mutex.unlock();
 			}
 		}
-
 		// You *must* call av_free_packet() after each call to av_read_frame() or else you'll leak memory
-		av_free_packet(&packet);
+		av_free_packet(&packet);  //crashes here!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! SIGSEV In _int_free (av=0xb4600010, p=0xb46025c8, have_lock=0) at malloc.c:4085 ()
 	}
 
 	// Some codecs will cause frames to be buffered up in the decoding process. If the CODEC_CAP_DELAY flag
@@ -568,6 +563,9 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 			for (auto p: processors) {
 				p->process_frame(frame->data[0],frame->nb_samples);
 			}
+			mutex.lock();
+			progress=((double)sample_processed)/samples;
+			mutex.unlock();
 		}
 	}
 
@@ -575,9 +573,6 @@ bool Render_context::load_audio(const string &filename,vector<base_audio_process
 	av_free(frame);
 	avcodec_close(codecContext);
 	av_close_input_file(formatContext);
-	
-	
-
 
 	return true;
 }
@@ -614,7 +609,7 @@ bool Render_context::load_graph(string &filename){
 								}
 								else cerr << "Rotor: linked input " << i2 << " of node '" << nodeID << "' to node '" << fromID << "'" << endl;
 							}
-							else cerr << "Rotor: linking input " << i2 << " of node: '" << nodeID << "', cannot find target '" << fromID << "'" << endl;							
+							else cerr << "Rotor: linking input " << i2 << " of node: '" << nodeID << "', cannot find target '" << fromID << "'" << endl;
 						}
 						xml.popTag();
 					}
@@ -626,44 +621,6 @@ bool Render_context::load_graph(string &filename){
 			}
 			xml.popTag();
 		}
-		/*
-		model_name=XML.getAttribute("map4","model","none",0);
-		model_x=ofToFloat(XML.getAttribute("map4","x","none",0));
-		model_y=ofToFloat(XML.getAttribute("map4","y","none",0));
-		model_z=ofToFloat(XML.getAttribute("map4","z","none",0));
-		if(XML.pushTag("map4")) {
-		numViews=XML.getNumTags("view");
-		if(numViews) {
-		views=new viewpoint[numViews];
-		for (int i=0;i<numViews;i++){
-		    XML.pushTag("view",i);
-			vector<string>keys;
-			XML.getAttributeNames("settings", keys, 0);
-			map<string,string>settings;
-			for (int k=0;k<keys.size();k++) {
-			    settings[keys[k]]=XML.getAttribute("settings",keys[k],"none",0);
-			}
-			views[i].setup(settings);
-		    XML.popTag();
-		}
-		}
-
-		numClips=XML.getNumTags("clip");
-		if (numClips) {
-		clips=new string[numClips];
-		for (int i=0;i< numClips;i++) {
-		    XML.pushTag("clip",i);
-			clips[i]=XML.getAttribute("settings","file","none",0);
-		    XML.popTag();
-		}
-		}
-
-		XML.popTag();
-		for (int i=0;i< numClips;i++) {
-		printf("clip: %s\n",clips[i].c_str());
-		}
-		}
-		*/
 		return true;
 	}
 	else return false;
@@ -675,4 +632,4 @@ Node_factory::Node_factory(){
 	add_type("divide",new Signal_divide());
 	add_type("==",new Is_new_integer());
 	add_type("signal_output",new Signal_output());
-}
\ No newline at end of file
+}
diff --git a/rotord/rotor.h b/rotord/rotor.h
index e16520b..de8e686 100755
--- a/rotord/rotor.h
+++ b/rotord/rotor.h
@@ -335,7 +335,7 @@ namespace Rotor {
 	};
 	class base_audio_processor {
 		public:
-			virtual bool process_frame(uint8_t *data,int samples)=0;
+			virtual int process_frame(uint8_t *data,int samples)=0;
 			void init(int _channels,int _bits,int _samples) {
 				channels=_channels;
 				bits=_bits;
@@ -350,7 +350,7 @@ namespace Rotor {
 		public:
 			audio_thumbnailer(){
 				height=64;
-				width=512;
+				width=128;  //fit 
 				data=new uint8_t[height*width];
 				memset(data,0,height*width);
 			}
@@ -360,39 +360,66 @@ namespace Rotor {
 			void init(int _channels,int _bits,int _samples) {
 				base_audio_processor::init(_channels,_bits,_samples);
 				samples_per_column=samples/width;
-				stored_data=new uint8_t[samples_per_column*channels*(bits>>3)];
-				sample=0;
+				column=0; //point thumbnail bitmap
+				out_sample=0; //sample in whole track
+				offset=(int)(pow(2.0,bits)/2.0);
+				scale=1.0/offset;
 			}
-			bool process_frame(uint8_t *data,int samples_in_frame){
+			int process_frame(uint8_t *data,int samples_in_frame){
 				//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 stride=channels*bytes;
-				int next_sample=0;
-				uint64_t accum;
-				while (next_sample<samples_in_frame) {
-					while (sample<samples_in_column&&next_sample<samples_in_frame) {
-						//do the accumulation
+				int in_sample=0;
+				while (in_sample<samples_in_frame&&column<width) {
+					int sample=0;
+					int samples=0;
+					double accum=0;
+					while (sample<samples_per_column&&in_sample<samples_in_frame) {
+						//accumulate squares for this column
 						for (int i=0;i<channels;i++) {
 							int this_val=0;
 							for (int j=0;j<bytes;j++) {
 								this_val+=data[(sample*stride)+(i*bytes)+j]<<j;
 							}
-							accum+=this_val*this_val;
-								next_sample++;
-								sample++;
-							}
+							//convert from integer data format - i.e s16p - to audio signal in -1..1 range
+							//don't know how many bytes we are dealing with necessarily?
+							double val=(this_val-offset)*scale;
+							accum+=val*val;
+							samples++;
 						}
-					} do;
-					
-				} do;
-				return true;
+						in_sample++;
+						sample++;
+						out_sample++;
+					}
+					//get root-mean
+					double mean=pow(accum/samples,0.5);
+					int colheight=height*mean*0.5;
+					int hh=height>>1;
+					for (int i=0;i<height;i++) {
+						data[i*width+column]=abs(i-hh)<colheight?0xff:0x00;
+					}
+					column++;
+				}
+				return out_sample;
+			}
+			string print(){
+				string output;
+				for (int j=0;j<height;j++) {
+					for (int i=0;i<width;i++) {
+						output+=data[j*width+i]<0x7f?"0":"1";
+					}
+					output +="\n";
+				}
+				return output;
 			}
 			uint8_t *data;
-			uint8_t *stored_data;
 			int height,width,samples_per_column;
-			int sample;
+			int column,out_sample;
+			//for drawing graph
+			int offset;
+			double scale;
 	};
 	class Render_context: public Poco::Task {				//Poco task object
 													//manages a 'patchbay'
@@ -400,7 +427,8 @@ namespace Rotor {
 													//and low level interface onto the graph
 		public:
 			Render_context(const std::string& name): Task(name)  {
-			    state=IDLE;
+				audio_thumb=new audio_thumbnailer();
+				state=IDLE;
 			};
 			void runTask();
 			void add_queue(int item);
@@ -417,11 +445,12 @@ namespace Rotor {
 
 		private:
 			int state;
-			float progress;								//for a locking process: audio analysis or rendering
+			double progress;								//for a locking process: audio analysis or rendering
 													//thread only does one thing at once
 			std::deque<int> work_queue;
 			Poco::Mutex mutex;            					//lock for access from parent thread
 			std::string audio_filename;
+			audio_thumbnailer *audio_thumb;
 			vampHost::QMAnalyser audio_analyser;
 			xmlIO xml;
 			Graph graph;
@@ -431,6 +460,6 @@ namespace Rotor {
 
 /*
 coding style
-Types begin with a capital, use underscore as a seperator
-instances begin with a small letter
+Types begin with capitals 'CamelCase'
+variables/ instances use lower case with underscore as a seperator
 */
diff --git a/rotord/rotord.cbp b/rotord/rotord.cbp
index 4f964e9..f4eafc3 100644
--- a/rotord/rotord.cbp
+++ b/rotord/rotord.cbp
@@ -49,6 +49,8 @@
 			<Add option="-Wall" />
 		</Compiler>
 		<Unit filename="Makefile" />
+		<Unit filename="avCodec.cpp" />
+		<Unit filename="avCodec.h" />
 		<Unit filename="ofUtils.cpp" />
 		<Unit filename="ofUtils.h" />
 		<Unit filename="rotor.cpp" />