summaryrefslogtreecommitdiff
path: root/src/lib/audio_decoder.cc
blob: 60b0f07ada2757b0a129f6da128b1d5a9841a405 (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
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
/*
    Copyright (C) 2012-2021 Carl Hetherington <cth@carlh.net>

    This file is part of DCP-o-matic.

    DCP-o-matic 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.

    DCP-o-matic 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 DCP-o-matic.  If not, see <http://www.gnu.org/licenses/>.

*/


#include "audio_decoder.h"
#include "audio_buffers.h"
#include "audio_content.h"
#include "dcpomatic_log.h"
#include "log.h"
#include "resampler.h"
#include "compose.hpp"
#include <iostream>

#include "i18n.h"


using std::cout;
using std::shared_ptr;
using std::make_shared;
using boost::optional;
using namespace dcpomatic;


AudioDecoder::AudioDecoder (Decoder* parent, shared_ptr<const AudioContent> content, bool fast)
	: DecoderPart (parent)
	, _content (content)
	, _fast (fast)
{
	/* Set up _positions so that we have one for each stream */
	for (auto i: content->streams ()) {
		_positions[i] = 0;
	}
}


/** @param time_already_delayed true if the delay should not be added to time */
void
AudioDecoder::emit(shared_ptr<const Film> film, AudioStreamPtr stream, shared_ptr<const AudioBuffers> data, ContentTime time, bool flushing)
{
	if (ignore ()) {
		return;
	}

	int const resampled_rate = _content->resampled_frame_rate(film);
	if (!flushing) {
		time += ContentTime::from_seconds (_content->delay() / 1000.0);
	}

	/* Amount of error we will tolerate on audio timestamps; see comment below.
	 * ffplay seems to use 1 24fps video frame, but I was sent a file that played
	 * fine in VLC but not here until the slack was increased to 4 frames @ 24fps.
	 */
	Frame const slack_frames = 4 * resampled_rate / 24;

	/* first_since_seek is set to true if this is the first data we have
	   received since initialisation or seek.  We'll set the position based
	   on the ContentTime that was given.  After this first time we just
	   count samples unless the timestamp is more than slack_frames away
	   from where we think it should be.  This is because ContentTimes seem
	   to be slightly unreliable from FFmpegDecoder (i.e.  not sample
	   accurate), but we still need to obey them sometimes otherwise we get
	   sync problems such as #1833.
	*/

	auto const first_since_seek = _positions[stream] == 0;
	auto const need_reset = !first_since_seek && (std::abs(_positions[stream] - time.frames_round(resampled_rate)) > slack_frames);

	if (need_reset) {
		LOG_GENERAL (
			"Reset audio position: was %1, new data at %2, slack: %3 frames (more than threshold %4)",
			_positions[stream],
			time.frames_round(resampled_rate),
			std::abs(_positions[stream] - time.frames_round(resampled_rate)),
			slack_frames
			);
	}

	if (first_since_seek || need_reset) {
		_positions[stream] = time.frames_round (resampled_rate);
	}

	if (first_since_seek && _content->delay() > 0) {
		silence (stream, _content->delay());
	}

	shared_ptr<Resampler> resampler;
	auto i = _resamplers.find(stream);
	if (i != _resamplers.end()) {
		resampler = i->second;
	} else {
		if (stream->frame_rate() != resampled_rate) {
			LOG_GENERAL (
				"Creating new resampler from %1 to %2 with %3 channels",
				stream->frame_rate(),
				resampled_rate,
				stream->channels()
				);

			resampler = make_shared<Resampler>(stream->frame_rate(), resampled_rate, stream->channels());
			if (_fast) {
				resampler->set_fast ();
			}
			_resamplers[stream] = resampler;
		}
	}

	if (resampler && !flushing) {
		/* It can be the the data here has a different number of channels than the stream
		 * it comes from (e.g. the files decoded by FFmpegDecoder sometimes have a random
		 * frame, often at the end, with more channels).  Insert silence or discard channels
		 * here.
		 */
		if (resampler->channels() != data->channels()) {
			LOG_WARNING("Received audio data with an unexpected channel count of %1 instead of %2", data->channels(), resampler->channels());
			auto data_copy = data->clone();
			data_copy->set_channels(resampler->channels());
			data = resampler->run(data_copy);
		} else {
			data = resampler->run(data);
		}

		if (data->frames() == 0) {
			return;
		}
	}

	Data(stream, ContentAudio (data, _positions[stream]));
	_positions[stream] += data->frames();
}


/** @return Time just after the last thing that was emitted from a given stream */
ContentTime
AudioDecoder::stream_position (shared_ptr<const Film> film, AudioStreamPtr stream) const
{
	auto i = _positions.find (stream);
	DCPOMATIC_ASSERT (i != _positions.end ());
	return ContentTime::from_frames (i->second, _content->resampled_frame_rate(film));
}


boost::optional<ContentTime>
AudioDecoder::position (shared_ptr<const Film> film) const
{
	optional<ContentTime> p;
	for (auto i: _positions) {
		auto const ct = stream_position (film, i.first);
		if (!p || ct < *p) {
			p = ct;
		}
	}

	return p;
}


void
AudioDecoder::seek ()
{
	for (auto i: _resamplers) {
		i.second->flush ();
		i.second->reset ();
	}

	for (auto& i: _positions) {
		i.second = 0;
	}
}


void
AudioDecoder::flush ()
{
	for (auto const& i: _resamplers) {
		auto ro = i.second->flush ();
		if (ro->frames() > 0) {
			Data (i.first, ContentAudio (ro, _positions[i.first]));
			_positions[i.first] += ro->frames();
		}
	}

	if (_content->delay() < 0) {
		/* Finish off with the gap caused by the delay */
		for (auto stream: _content->streams()) {
			silence (stream, -_content->delay());
		}
	}
}


void
AudioDecoder::silence (AudioStreamPtr stream, int milliseconds)
{
	int const samples = ContentTime::from_seconds(milliseconds / 1000.0).frames_round(stream->frame_rate());
	auto silence = make_shared<AudioBuffers>(stream->channels(), samples);
	silence->make_silent ();
	Data (stream, ContentAudio(silence, _positions[stream]));
}