Basic stuff to analyse audio (job).

9 files changed, 329 insertions, 7 deletions

diff --git a/src/lib/analyse_audio_job.cc b/src/lib/analyse_audio_job.cc
new file mode 100644
index 000000000..5623fdfcc
--- /dev/null
+++ b/src/lib/analyse_audio_job.cc
@@ -0,0 +1,108 @@
+/*
+    Copyright (C) 2012 Carl Hetherington <cth@carlh.net>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+
+*/
+
+#include "audio_analysis.h"
+#include "analyse_audio_job.h"
+#include "compose.hpp"
+#include "film.h"
+#include "options.h"
+#include "decoder_factory.h"
+#include "audio_decoder.h"
+
+using std::string;
+using std::max;
+using boost::shared_ptr;
+
+int const AnalyseAudioJob::_num_points = 1024;
+
+AnalyseAudioJob::AnalyseAudioJob (shared_ptr<Film> f)
+	: Job (f)
+	, _done_for_this_point (0)
+	, _done (0)
+	, _samples_per_point (1)
+{
+
+}
+
+string
+AnalyseAudioJob::name () const
+{
+	return String::compose ("Analyse audio of %1", _film->name());
+}
+
+void
+AnalyseAudioJob::run ()
+{
+	if (!_film->audio_stream () || !_film->length()) {
+		set_progress (1);
+		set_state (FINISHED_ERROR);
+		return;
+	}
+		
+	DecodeOptions options;
+	options.decode_video = false;
+
+	Decoders decoders = decoder_factory (_film, options);
+	assert (decoders.audio);
+	
+	decoders.audio->set_audio_stream (_film->audio_stream ());
+	decoders.audio->Audio.connect (bind (&AnalyseAudioJob::audio, this, _1));
+
+	int64_t total_audio_frames = video_frames_to_audio_frames (_film->length().get(), _film->audio_stream()->sample_rate(), _film->frames_per_second());
+	_samples_per_point = total_audio_frames / _num_points;
+
+	_current.resize (_film->audio_stream()->channels ());
+	_analysis.reset (new AudioAnalysis (_film->audio_stream()->channels()));
+			 
+	while (!decoders.audio->pass()) {
+		set_progress (float (_done) / total_audio_frames);
+	}
+
+	_analysis->write (_film->audio_analysis_path ());
+	
+	set_progress (1);
+	set_state (FINISHED_OK);
+}
+
+void
+AnalyseAudioJob::audio (shared_ptr<AudioBuffers> b)
+{
+	for (int i = 0; i < b->frames(); ++i) {
+		for (int j = 0; j < b->channels(); ++j) {
+			float const s = b->data(j)[i];
+			_current[j][AudioPoint::RMS] += pow (s, 2);
+			_current[j][AudioPoint::PEAK] = max (_current[j][AudioPoint::PEAK], fabsf (s));
+
+			if (_done_for_this_point == _samples_per_point) {
+				_current[j][AudioPoint::RMS] = 20 * log10 (sqrt (_current[j][AudioPoint::RMS] / _samples_per_point));
+				_current[j][AudioPoint::PEAK] = 20 * log10 (_current[j][AudioPoint::PEAK]);
+				
+				_analysis->add_point (j, _current[j]);
+				
+				_done_for_this_point = 0;
+				_current[j] = AudioPoint ();
+			}
+		}
+		
+		++_done_for_this_point;
+	}
+
+	_done += b->frames ();
+}
+
diff --git a/src/lib/analyse_audio_job.h b/src/lib/analyse_audio_job.h
new file mode 100644
index 000000000..1e7229ce6
--- /dev/null
+++ b/src/lib/analyse_audio_job.h
@@ -0,0 +1,45 @@
+/*
+    Copyright (C) 2012 Carl Hetherington <cth@carlh.net>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+
+*/
+
+#include "job.h"
+#include "audio_analysis.h"
+
+class AudioBuffers;
+
+class AnalyseAudioJob : public Job
+{
+public:
+	AnalyseAudioJob (boost::shared_ptr<Film> f);
+
+	std::string name () const;
+	void run ();
+
+private:
+	void audio (boost::shared_ptr<AudioBuffers>);
+
+	int64_t _done_for_this_point;
+	int64_t _done;
+	int64_t _samples_per_point;
+	std::vector<AudioPoint> _current;
+
+	boost::shared_ptr<AudioAnalysis> _analysis;
+
+	static const int _num_points;
+};
+
diff --git a/src/lib/audio_analysis.cc b/src/lib/audio_analysis.cc
new file mode 100644
index 000000000..4a710f4c1
--- /dev/null
+++ b/src/lib/audio_analysis.cc
@@ -0,0 +1,69 @@
+/*
+    Copyright (C) 2012 Carl Hetherington <cth@carlh.net>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+
+*/
+
+#include <stdint.h>
+#include <cassert>
+#include <fstream>
+#include "audio_analysis.h"
+
+using std::ostream;
+using std::string;
+using std::ofstream;
+using std::vector;
+
+AudioPoint::AudioPoint ()
+{
+	for (int i = 0; i < COUNT; ++i) {
+		_data[i] = 0;
+	}
+}
+
+void
+AudioPoint::write (ostream& s) const
+{
+	for (int i = 0; i < COUNT; ++i) {
+		s << _data[i] << "\n";
+	}
+}
+	
+
+AudioAnalysis::AudioAnalysis (int channels)
+{
+	_data.resize (channels);
+}
+
+void
+AudioAnalysis::add_point (int c, AudioPoint const & p)
+{
+	assert (c < int (_data.size ()));
+	_data[c].push_back (p);
+}
+
+void
+AudioAnalysis::write (string filename)
+{
+	ofstream f (filename.c_str ());
+	f << _data.size() << "\n";
+	for (vector<vector<AudioPoint> >::iterator i = _data.begin(); i != _data.end(); ++i) {
+		f << i->size () << "\n";
+		for (vector<AudioPoint>::iterator j = i->begin(); j != i->end(); ++j) {
+			j->write (f);
+		}
+	}
+}
diff --git a/src/lib/audio_analysis.h b/src/lib/audio_analysis.h
new file mode 100644
index 000000000..1c668b9c2
--- /dev/null
+++ b/src/lib/audio_analysis.h
@@ -0,0 +1,59 @@
+/*
+    Copyright (C) 2012 Carl Hetherington <cth@carlh.net>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+
+*/
+
+#ifndef DVDOMATIC_AUDIO_ANALYSIS_H
+#define DVDOMATIC_AUDIO_ANALYSIS_H
+
+#include <iostream>
+#include <vector>
+
+class AudioPoint
+{
+public:
+	enum Type {
+		PEAK,
+		RMS,
+		COUNT
+	};
+
+	AudioPoint ();
+
+	void write (std::ostream &) const;
+	
+	float& operator[] (Type t) {
+		return _data[t];
+	}
+
+private:
+	float _data[COUNT];
+};
+
+class AudioAnalysis
+{
+public:
+	AudioAnalysis (int c);
+
+	void add_point (int c, AudioPoint const & p);
+	void write (std::string);
+
+private:
+	std::vector<std::vector<AudioPoint> > _data;
+};
+
+#endif
diff --git a/src/lib/ffmpeg_decoder.cc b/src/lib/ffmpeg_decoder.cc
index d4ed76e37..58c7317ac 100644
--- a/src/lib/ffmpeg_decoder.cc
+++ b/src/lib/ffmpeg_decoder.cc
@@ -236,8 +236,10 @@ FFmpegDecoder::pass ()
 
 		int frame_finished;
 
-		while (avcodec_decode_video2 (_video_codec_context, _frame, &frame_finished, &_packet) >= 0 && frame_finished) {
-			filter_and_emit_video (_frame);
+		if (_opt.decode_video) {
+			while (avcodec_decode_video2 (_video_codec_context, _frame, &frame_finished, &_packet) >= 0 && frame_finished) {
+				filter_and_emit_video (_frame);
+			}
 		}
 
 		if (_audio_stream && _opt.decode_audio) {
@@ -258,7 +260,7 @@ FFmpegDecoder::pass ()
 
 	shared_ptr<FFmpegAudioStream> ffa = dynamic_pointer_cast<FFmpegAudioStream> (_audio_stream);
 
-	if (_packet.stream_index == _video_stream) {
+	if (_packet.stream_index == _video_stream && _opt.decode_video) {
 
 		int frame_finished;
 		int const r = avcodec_decode_video2 (_video_codec_context, _frame, &frame_finished, &_packet);
@@ -288,9 +290,9 @@ FFmpegDecoder::pass ()
 			   was before this packet.  Until then audio is thrown away.
 			*/
 				
-			if (_first_video && _first_video.get() <= source_pts_seconds) {
+			if ((_first_video && _first_video.get() <= source_pts_seconds) || !_opt.decode_video) {
 
-				if (!_first_audio) {
+				if (!_first_audio && _opt.decode_video) {
 					_first_audio = source_pts_seconds;
 					
 					/* This is our first audio frame, and if we've arrived here we must have had our
diff --git a/src/lib/film.cc b/src/lib/film.cc
index 1cf161259..fb3fb4cde 100644
--- a/src/lib/film.cc
+++ b/src/lib/film.cc
@@ -51,6 +51,7 @@
 #include "video_decoder.h"
 #include "audio_decoder.h"
 #include "external_audio_decoder.h"
+#include "analyse_audio_job.h"
 
 using std::string;
 using std::stringstream;
@@ -237,6 +238,15 @@ Film::video_mxf_filename () const
 	return video_state_identifier() + ".mxf";
 }
 
+string
+Film::audio_analysis_path () const
+{
+	boost::filesystem::path p;
+	p /= "analysis";
+	p /= content_digest();
+	return file (p.string ());
+}
+
 /** Add suitable Jobs to the JobManager to create a DCP for this Film */
 void
 Film::make_dcp ()
@@ -303,6 +313,19 @@ Film::make_dcp ()
 	}
 }
 
+/** Start a job to analyse the audio of our content file */
+void
+Film::analyse_audio ()
+{
+	if (_analyse_audio_job) {
+		return;
+	}
+
+	_analyse_audio_job.reset (new AnalyseAudioJob (shared_from_this()));
+	_analyse_audio_job->Finished.connect (bind (&Film::analyse_audio_finished, this));
+	JobManager::instance()->add (_analyse_audio_job);
+}
+
 /** Start a job to examine our content file */
 void
 Film::examine_content ()
@@ -317,6 +340,12 @@ Film::examine_content ()
 }
 
 void
+Film::analyse_audio_finished ()
+{
+	_analyse_audio_job.reset ();
+}
+
+void
 Film::examine_content_finished ()
 {
 	_examine_content_job.reset ();
diff --git a/src/lib/film.h b/src/lib/film.h
index 04a483998..c268d3eac 100644
--- a/src/lib/film.h
+++ b/src/lib/film.h
@@ -45,6 +45,7 @@ class Job;
 class Filter;
 class Log;
 class ExamineContentJob;
+class AnalyseAudioJob;
 class ExternalAudioStream;
 
 /** @class Film
@@ -65,8 +66,10 @@ public:
 	std::string info_path (int f) const;
 	std::string video_mxf_dir () const;
 	std::string video_mxf_filename () const;
+	std::string audio_analysis_path () const;
 
 	void examine_content ();
+	void analyse_audio ();
 	void send_dcp_to_tms ();
 
 	void make_dcp ();
@@ -374,9 +377,12 @@ private:
 
 	/** Any running ExamineContentJob, or 0 */
 	boost::shared_ptr<ExamineContentJob> _examine_content_job;
+	/** Any running AnalyseAudioJob, or 0 */
+	boost::shared_ptr<AnalyseAudioJob> _analyse_audio_job;
 
 	void signal_changed (Property);
 	void examine_content_finished ();
+	void analyse_audio_finished ();
 	std::string video_state_identifier () const;
 
 	/** Complete path to directory containing the film metadata;
diff --git a/src/lib/options.h b/src/lib/options.h
index 2cd7dffde..0d2c07fd5 100644
--- a/src/lib/options.h
+++ b/src/lib/options.h
@@ -28,11 +28,13 @@ class DecodeOptions
 {
 public:
 	DecodeOptions ()
-		: decode_audio (true)
+		: decode_video (true)
+		, decode_audio (true)
 		, decode_subtitles (false)
 		, video_sync (true)
 	{}
-	
+
+	bool decode_video;
 	bool decode_audio;
 	bool decode_subtitles;
 	bool video_sync;
diff --git a/src/lib/wscript b/src/lib/wscript
index eee04190c..c2b46112c 100644
--- a/src/lib/wscript
+++ b/src/lib/wscript
@@ -16,6 +16,8 @@ def build(bld):
     obj.source = """
 		 ab_transcode_job.cc
 		 ab_transcoder.cc
+                 analyse_audio_job.cc
+                 audio_analysis.cc
                  audio_decoder.cc
                  audio_source.cc
 		 config.cc