blob: fc78f07ebb0a2c813b5b8fc28b9a5a67d8bb028e (
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
|
/*
* Copyright (c) 2012 The WebRTC project authors. All Rights Reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
#include "webrtc/modules/audio_processing/agc/agc.h"
#include <cmath>
#include <cstdlib>
#include <algorithm>
#include <vector>
#include "webrtc/base/checks.h"
#include "webrtc/modules/audio_processing/agc/histogram.h"
#include "webrtc/modules/audio_processing/agc/utility.h"
#include "webrtc/modules/include/module_common_types.h"
namespace webrtc {
namespace {
const int kDefaultLevelDbfs = -18;
const int kNumAnalysisFrames = 100;
const double kActivityThreshold = 0.3;
} // namespace
Agc::Agc()
: target_level_loudness_(Dbfs2Loudness(kDefaultLevelDbfs)),
target_level_dbfs_(kDefaultLevelDbfs),
histogram_(Histogram::Create(kNumAnalysisFrames)),
inactive_histogram_(Histogram::Create()) {
}
Agc::~Agc() {}
float Agc::AnalyzePreproc(const int16_t* audio, size_t length) {
assert(length > 0);
size_t num_clipped = 0;
for (size_t i = 0; i < length; ++i) {
if (audio[i] == 32767 || audio[i] == -32768)
++num_clipped;
}
return 1.0f * num_clipped / length;
}
int Agc::Process(const int16_t* audio, size_t length, int sample_rate_hz) {
vad_.ProcessChunk(audio, length, sample_rate_hz);
const std::vector<double>& rms = vad_.chunkwise_rms();
const std::vector<double>& probabilities =
vad_.chunkwise_voice_probabilities();
RTC_DCHECK_EQ(rms.size(), probabilities.size());
for (size_t i = 0; i < rms.size(); ++i) {
histogram_->Update(rms[i], probabilities[i]);
}
return 0;
}
bool Agc::GetRmsErrorDb(int* error) {
if (!error) {
assert(false);
return false;
}
if (histogram_->num_updates() < kNumAnalysisFrames) {
// We haven't yet received enough frames.
return false;
}
if (histogram_->AudioContent() < kNumAnalysisFrames * kActivityThreshold) {
// We are likely in an inactive segment.
return false;
}
double loudness = Linear2Loudness(histogram_->CurrentRms());
*error = std::floor(Loudness2Db(target_level_loudness_ - loudness) + 0.5);
histogram_->Reset();
return true;
}
void Agc::Reset() {
histogram_->Reset();
}
int Agc::set_target_level_dbfs(int level) {
// TODO(turajs): just some arbitrary sanity check. We can come up with better
// limits. The upper limit should be chosen such that the risk of clipping is
// low. The lower limit should not result in a too quiet signal.
if (level >= 0 || level <= -100)
return -1;
target_level_dbfs_ = level;
target_level_loudness_ = Dbfs2Loudness(level);
return 0;
}
} // namespace webrtc
|
