summaryrefslogtreecommitdiff
path: root/src/lib/analyse_audio_job.cc
blob: 5dba56598d9196f78fe2efab0466d587f45d1d65 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
/*
    Copyright (C) 2012-2015 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 "audio_buffers.h"
#include "analyse_audio_job.h"
#include "audio_content.h"
#include "compose.hpp"
#include "film.h"
#include "player.h"
#include "playlist.h"
#include "filter.h"
#include "audio_filter_graph.h"
extern "C" {
#include <libavutil/channel_layout.h>
}
#include <boost/foreach.hpp>
#include <iostream>

#include "i18n.h"

using std::string;
using std::max;
using std::min;
using std::cout;
using boost::shared_ptr;
using boost::dynamic_pointer_cast;

int const AnalyseAudioJob::_num_points = 1024;

extern "C" {
/* I added these functions to the FFmpeg that DCP-o-matic is built with, but there isn't an
   existing header file to put the prototype in.  Cheat by putting it in here.
*/
double* av_ebur128_get_true_peaks         (void* context);
double* av_ebur128_get_sample_peaks       (void* context);
double  av_ebur128_get_integrated_loudness(void* context);
double  av_ebur128_get_loudness_range     (void* context);
}

AnalyseAudioJob::AnalyseAudioJob (shared_ptr<const Film> film, shared_ptr<const Playlist> playlist)
	: Job (film)
	, _playlist (playlist)
	, _done (0)
	, _samples_per_point (1)
	, _current (0)
	, _sample_peak (0)
	, _sample_peak_frame (0)
	, _ebur128 (new AudioFilterGraph (film->audio_frame_rate(), av_get_default_channel_layout(film->audio_channels())))
{
	_filters.push_back (new Filter ("ebur128", "ebur128", "audio", "ebur128=peak=true"));
	_ebur128->setup (_filters);
}

AnalyseAudioJob::~AnalyseAudioJob ()
{
	BOOST_FOREACH (Filter const * i, _filters) {
		delete const_cast<Filter*> (i);
	}
	delete[] _current;
}

string
AnalyseAudioJob::name () const
{
	return _("Analyse audio");
}

string
AnalyseAudioJob::json_name () const
{
	return N_("analyse_audio");
}

void
AnalyseAudioJob::run ()
{
	shared_ptr<Player> player (new Player (_film, _playlist));
	player->set_ignore_video ();
	player->set_fast ();
	player->set_play_referenced ();

	DCPTime const start = _playlist->start().get_value_or (DCPTime ());
	DCPTime const length = _playlist->length ();

	Frame const len = DCPTime (length - start).frames_round (_film->audio_frame_rate());
	_samples_per_point = max (int64_t (1), len / _num_points);

	delete[] _current;
	_current = new AudioPoint[_film->audio_channels ()];
	_analysis.reset (new AudioAnalysis (_film->audio_channels ()));

	bool has_any_audio = false;
	BOOST_FOREACH (shared_ptr<Content> c, _playlist->content ()) {
		if (dynamic_pointer_cast<AudioContent> (c)) {
			has_any_audio = true;
		}
	}

	if (has_any_audio) {
		_done = 0;
		DCPTime const block = DCPTime::from_seconds (1.0 / 8);
		for (DCPTime t = start; t < length; t += block) {
			shared_ptr<const AudioBuffers> audio = player->get_audio (t, block, false);
			_ebur128->process (audio);
			analyse (audio);
			set_progress ((t.seconds() - start.seconds()) / (length.seconds() - start.seconds()));
		}
	}

	_analysis->set_sample_peak (_sample_peak, DCPTime::from_frames (_sample_peak_frame, _film->audio_frame_rate ()));

	void* eb = _ebur128->get("Parsed_ebur128_0")->priv;
	double true_peak = 0;
	for (int i = 0; i < _film->audio_channels(); ++i) {
		true_peak = max (true_peak, av_ebur128_get_true_peaks(eb)[i]);
	}
	_analysis->set_true_peak (true_peak);
	_analysis->set_integrated_loudness (av_ebur128_get_integrated_loudness(eb));
	_analysis->set_loudness_range (av_ebur128_get_loudness_range(eb));

	if (_playlist->content().size() == 1) {
		/* If there was only one piece of content in this analysis we may later need to know what its
		   gain was when we analysed it.
		*/
		shared_ptr<const AudioContent> ac = dynamic_pointer_cast<const AudioContent> (_playlist->content().front ());
		DCPOMATIC_ASSERT (ac);
		_analysis->set_analysis_gain (ac->audio_gain ());
	}

	_analysis->write (_film->audio_analysis_path (_playlist));

	set_progress (1);
	set_state (FINISHED_OK);
}

void
AnalyseAudioJob::analyse (shared_ptr<const AudioBuffers> b)
{
	int const frames = b->frames ();
	int const channels = b->channels ();

	for (int j = 0; j < channels; ++j) {
		float* data = b->data(j);
		for (int i = 0; i < frames; ++i) {
			float s = data[i];
			float as = fabsf (s);
			if (as < 10e-7) {
				/* SafeStringStream can't serialise and recover inf or -inf, so prevent such
				   values by replacing with this (140dB down) */
				s = as = 10e-7;
			}
			_current[j][AudioPoint::RMS] += pow (s, 2);
			_current[j][AudioPoint::PEAK] = max (_current[j][AudioPoint::PEAK], as);

			if (as > _sample_peak) {
				_sample_peak = as;
				_sample_peak_frame = _done + i;
			}

			if (((_done + i) % _samples_per_point) == 0) {
				_current[j][AudioPoint::RMS] = sqrt (_current[j][AudioPoint::RMS] / _samples_per_point);
				_analysis->add_point (j, _current[j]);
				_current[j] = AudioPoint ();
			}
		}
	}

	_done += frames;
}