blob: 5d01100eb73a1a4352e9434e356d5cebe30d32f0 (
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
|
/*
* Copyright (c) 2018 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.
*/
#ifndef MODULES_AUDIO_PROCESSING_AGC2_AGC2_COMMON_H_
#define MODULES_AUDIO_PROCESSING_AGC2_AGC2_COMMON_H_
#include <stddef.h>
namespace webrtc {
constexpr float kMinFloatS16Value = -32768.f;
constexpr float kMaxFloatS16Value = 32767.f;
constexpr float kMaxAbsFloatS16Value = 32768.0f;
constexpr size_t kFrameDurationMs = 10;
constexpr size_t kSubFramesInFrame = 20;
constexpr size_t kMaximalNumberOfSamplesPerChannel = 480;
constexpr float kAttackFilterConstant = 0.f;
// Adaptive digital gain applier settings below.
constexpr float kHeadroomDbfs = 1.f;
constexpr float kMaxGainDb = 30.f;
constexpr float kInitialAdaptiveDigitalGainDb = 8.f;
// At what limiter levels should we start decreasing the adaptive digital gain.
constexpr float kLimiterThresholdForAgcGainDbfs = -kHeadroomDbfs;
// This is the threshold for speech. Speech frames are used for updating the
// speech level, measuring the amount of speech, and decide when to allow target
// gain reduction.
constexpr float kVadConfidenceThreshold = 0.9f;
// The amount of 'memory' of the Level Estimator. Decides leak factors.
constexpr size_t kFullBufferSizeMs = 1200;
constexpr float kFullBufferLeakFactor = 1.f - 1.f / kFullBufferSizeMs;
constexpr float kInitialSpeechLevelEstimateDbfs = -30.f;
// Robust VAD probability and speech decisions.
constexpr float kDefaultSmoothedVadProbabilityAttack = 1.f;
constexpr int kDefaultLevelEstimatorAdjacentSpeechFramesThreshold = 1;
// Saturation Protector settings.
constexpr float kDefaultInitialSaturationMarginDb = 20.f;
constexpr float kDefaultExtraSaturationMarginDb = 2.f;
constexpr size_t kPeakEnveloperSuperFrameLengthMs = 400;
static_assert(kFullBufferSizeMs % kPeakEnveloperSuperFrameLengthMs == 0,
"Full buffer size should be a multiple of super frame length for "
"optimal Saturation Protector performance.");
constexpr size_t kPeakEnveloperBufferSize =
kFullBufferSizeMs / kPeakEnveloperSuperFrameLengthMs + 1;
// This value is 10 ** (-1/20 * frame_size_ms / satproc_attack_ms),
// where satproc_attack_ms is 5000.
constexpr float kSaturationProtectorAttackConstant = 0.9988493699365052f;
// This value is 10 ** (-1/20 * frame_size_ms / satproc_decay_ms),
// where satproc_decay_ms is 1000.
constexpr float kSaturationProtectorDecayConstant = 0.9997697679981565f;
// This is computed from kDecayMs by
// 10 ** (-1/20 * subframe_duration / kDecayMs).
// |subframe_duration| is |kFrameDurationMs / kSubFramesInFrame|.
// kDecayMs is defined in agc2_testing_common.h
constexpr float kDecayFilterConstant = 0.9998848773724686f;
// Number of interpolation points for each region of the limiter.
// These values have been tuned to limit the interpolated gain curve error given
// the limiter parameters and allowing a maximum error of +/- 32768^-1.
constexpr size_t kInterpolatedGainCurveKneePoints = 22;
constexpr size_t kInterpolatedGainCurveBeyondKneePoints = 10;
constexpr size_t kInterpolatedGainCurveTotalPoints =
kInterpolatedGainCurveKneePoints + kInterpolatedGainCurveBeyondKneePoints;
} // namespace webrtc
#endif // MODULES_AUDIO_PROCESSING_AGC2_AGC2_COMMON_H_
|
