OLD | NEW |
---|---|
(Empty) | |
1 // Copyright 2015 The Chromium Authors. All rights reserved. | |
2 // Use of this source code is governed by a BSD-style license that can be | |
3 // found in the LICENSE file. | |
4 | |
5 #include "content/renderer/media/audio_repetition_detector.h" | |
6 | |
7 #include "base/logging.h" | |
8 #include "base/macros.h" | |
9 #include "base/metrics/histogram_macros.h" | |
10 | |
11 namespace content { | |
12 | |
13 namespace { | |
14 // Minimum duration of a repetition. | |
15 const int kMinLengthMs = 1; | |
16 | |
17 // The following variables defines the look back time of repetitions that will | |
18 // be logged. The complexity of the detector is proportional to the number of | |
19 // look back times we keep track. | |
20 const int kMinLookbackTimeMS = 10; | |
Henrik Grunell
2015/10/16 07:59:50
Nit: Ms
minyue
2015/10/16 08:34:42
Oh, yes
| |
21 const int kMaxLookbackTimeMS = 200; | |
22 const int kLookbackTimeStepMS = 10; | |
23 | |
24 // This is used for increasing the efficiency of copying data into the buffer. | |
25 // Input longer than |kMaxFrames| won't be a problem, and will be devided into | |
26 // chunks automatically. | |
27 const size_t kMaxFrames = 480; // 10 ms * 48 kHz | |
28 | |
29 } // namespace | |
30 | |
31 AudioRepetitionDetector::AudioRepetitionDetector() | |
32 : max_look_back_ms_(0), | |
33 min_length_ms_(kMinLengthMs), | |
Henrik Grunell
2015/10/16 07:59:50
Can |min_length_ms_| be removed?
minyue
2015/10/16 08:34:42
|min_length_ms_| is important in the algorithm, an
Henrik Grunell
2015/10/16 08:44:11
I meant to use the constant instead since it never
minyue
2015/10/16 09:27:27
Sure it is not really subjected to change in curre
Henrik Grunell
2015/10/16 11:08:11
I'm not sure I understand why the test has a lower
minyue
2015/10/23 12:05:23
Already covered with offline discussion.
Adding a
| |
34 sample_rate_(0), | |
35 buffer_size_frames_(0), | |
36 buffer_end_index_(0), | |
37 max_frames_(kMaxFrames) { | |
Henrik Grunell
2015/10/16 07:59:50
Can |max_frames_| be removed?
minyue
2015/10/16 08:34:42
same as min_length_ms_
minyue
2015/10/16 09:27:28
|kMaxFrames| is set to be the largest possible fra
Henrik Grunell
2015/10/16 11:08:11
Can you explain why we need to test with some othe
| |
38 for (int time = kMaxLookbackTimeMS; time >= kMinLookbackTimeMS; | |
39 time -= kLookbackTimeStepMS) | |
Henrik Grunell
2015/10/16 07:59:50
Use {} (Since total for block is more than 2 lines
minyue
2015/10/16 08:34:42
Ok. But why alligning -=, I am not aware that line
Henrik Grunell
2015/10/16 08:44:11
To align the second line with the (.
for (int tim
minyue
2015/10/16 09:27:28
ok. these ;-separated lines seems to me different
Henrik Grunell
2015/10/16 11:08:11
Align as functions is common practice in Chromium.
ajm
2015/10/16 15:54:47
Just run git cl format and be done with it :)
| |
40 RegisterLookbackTime(time); | |
41 | |
42 // May be created in the main render thread and used in the audio threads. | |
43 thread_checker_.DetachFromThread(); | |
44 } | |
45 | |
46 AudioRepetitionDetector::~AudioRepetitionDetector() = default; | |
47 | |
48 void AudioRepetitionDetector::Detect(const float* data, size_t num_frames, | |
49 size_t num_channels, int sample_rate) { | |
50 DCHECK(thread_checker_.CalledOnValidThread()); | |
51 DCHECK(!states_.empty()); | |
52 | |
53 if (num_channels != num_channels_ || sample_rate != sample_rate_) | |
54 Reset(num_channels, sample_rate); | |
55 | |
56 // The maximum number of frames |audio_buffer_| can take in is |max_frames_|. | |
57 // Therefore, input data with larger frames needs be divided into chunks. | |
58 const size_t chunk_size = max_frames_ * num_channels; | |
59 while (num_frames > max_frames_) { | |
60 Detect(data, max_frames_, num_channels, sample_rate); | |
61 data += chunk_size; | |
62 num_frames -= max_frames_; | |
63 } | |
64 | |
65 if (num_frames == 0) | |
66 return; | |
67 | |
68 AddFramesToBuffer(data, num_frames); | |
69 | |
70 for (size_t idx = num_frames; idx > 0; --idx, data += num_channels) { | |
71 for (State* state : states_) { | |
72 // Look back position depends on the sample rate. It is rounded down to | |
73 // the closest integer. | |
74 const size_t look_back_frames = | |
75 state->look_back_ms() * sample_rate_ / 1000; | |
76 // Equal(data, offset) checks if |data| equals the audio frame located | |
77 // |offset| frames from the end of buffer. Now a full frame has been | |
78 // inserted to the buffer, and thus |offset| should compensate for it. | |
79 if (Equal(data, look_back_frames + idx)) { | |
80 if (!state->reported()) { | |
81 state->Increment(IsZero(data, num_channels)); | |
82 if (HasValidReport(state)) { | |
83 ReportRepetition(state->look_back_ms()); | |
84 state->set_reported(true); | |
85 } | |
86 } | |
87 } else { | |
88 state->Reset(); | |
89 } | |
90 } | |
91 } | |
92 } | |
93 | |
94 AudioRepetitionDetector::State::State(int look_back_ms) | |
95 : look_back_ms_(look_back_ms) { | |
96 Reset(); | |
97 } | |
98 | |
99 void AudioRepetitionDetector::State::Increment(bool zero) { | |
100 if (0 == count_frames_ && zero) { | |
101 // If a repetition starts with zeros, we enter the all zero mode until | |
102 // a non zero is found later. The point is that the beginning zeros should | |
103 // be counted in the length of the repetition as long as the repetition does | |
104 // not comprise only zeros. | |
105 all_zero_ = true; | |
106 } | |
107 ++count_frames_; | |
108 if (!zero) | |
109 all_zero_ = false; | |
110 } | |
111 | |
112 void AudioRepetitionDetector::State::Reset() { | |
113 count_frames_ = 0; | |
114 all_zero_ = true; | |
115 reported_ = false; | |
116 } | |
117 | |
118 void AudioRepetitionDetector::RegisterLookbackTime(int look_back_ms) { | |
119 DCHECK(thread_checker_.CalledOnValidThread()); | |
120 | |
121 // States are added in the order of their look back times. | |
122 auto it = states_.begin(); | |
123 for (; it != states_.end(); ++it) { | |
124 const int it_look_back = (*it)->look_back_ms(); | |
125 if (it_look_back == look_back_ms) | |
126 return; | |
127 if (it_look_back < look_back_ms) | |
128 break; | |
129 } | |
130 states_.insert(it, new State(look_back_ms)); | |
131 if (look_back_ms > max_look_back_ms_) { | |
132 max_look_back_ms_ = look_back_ms; | |
133 } | |
134 } | |
135 | |
136 void AudioRepetitionDetector::Reset(size_t num_channels, int sample_rate) { | |
137 DCHECK(thread_checker_.CalledOnValidThread()); | |
138 num_channels_ = num_channels; | |
139 sample_rate_ = sample_rate; | |
140 | |
141 // |(xxx + 999) / 1000| is an arithmetic way to round up |xxx / 1000|. | |
142 buffer_size_frames_ = | |
143 (max_look_back_ms_ * sample_rate_ + 999) / 1000 + max_frames_; | |
144 | |
145 audio_buffer_.resize(buffer_size_frames_ * num_channels_); | |
146 for (State* state : states_) | |
147 state->Reset(); | |
148 } | |
149 | |
150 void AudioRepetitionDetector::AddFramesToBuffer(const float* data, | |
151 size_t num_frames) { | |
152 DCHECK(thread_checker_.CalledOnValidThread()); | |
153 DCHECK_LE(num_frames, buffer_size_frames_); | |
154 const size_t margin = buffer_size_frames_ - buffer_end_index_; | |
155 const auto it = audio_buffer_.begin() + buffer_end_index_ * num_channels_; | |
156 if (num_frames <= margin) { | |
157 std::copy(data, data + num_frames * num_channels_, it); | |
158 buffer_end_index_ += num_frames; | |
159 } else { | |
160 std::copy(data, data + margin * num_channels_, it); | |
161 std::copy(data + margin * num_channels_, data + num_frames * num_channels_, | |
162 audio_buffer_.begin()); | |
163 buffer_end_index_ = num_frames - margin; | |
164 } | |
165 } | |
166 | |
167 bool AudioRepetitionDetector::Equal(const float* frame, | |
168 int look_back_frames) const { | |
169 DCHECK(thread_checker_.CalledOnValidThread()); | |
170 const size_t look_back_index = | |
171 (buffer_end_index_ + buffer_size_frames_ - look_back_frames) % | |
172 buffer_size_frames_ ; | |
173 auto it = audio_buffer_.begin() + look_back_index * num_channels_; | |
174 for (size_t channel = 0; channel < num_channels_; ++channel, ++frame, ++it) { | |
175 if (*frame != *it) | |
176 return false; | |
177 } | |
178 return true; | |
179 } | |
180 | |
181 bool AudioRepetitionDetector::IsZero(const float* frame, | |
182 size_t num_channels) const { | |
183 for (size_t channel = 0; channel < num_channels; ++channel, ++frame) { | |
184 if (*frame != 0) | |
185 return false; | |
186 } | |
187 return true; | |
188 } | |
189 | |
190 bool AudioRepetitionDetector::HasValidReport(const State* state) const { | |
191 return (!state->all_zero() && state->count_frames() >= | |
192 static_cast<size_t>(min_length_ms_ * sample_rate_ / 1000)); | |
193 } | |
194 | |
195 void AudioRepetitionDetector::ReportRepetition(int look_back_ms) { | |
196 DCHECK(thread_checker_.CalledOnValidThread()); | |
197 UMA_HISTOGRAM_CUSTOM_COUNTS( | |
198 "Media.AudioCapturerRepetition", look_back_ms, | |
199 kMinLookbackTimeMS, kMaxLookbackTimeMS, | |
200 (kMaxLookbackTimeMS - kMinLookbackTimeMS) / kLookbackTimeStepMS + 1); | |
201 } | |
202 | |
203 } // namespace content | |
OLD | NEW |