OLD | NEW |
1 /* | 1 /* |
2 * Copyright (c) 2014 The WebRTC project authors. All Rights Reserved. | 2 * Copyright (c) 2014 The WebRTC project authors. All Rights Reserved. |
3 * | 3 * |
4 * Use of this source code is governed by a BSD-style license | 4 * Use of this source code is governed by a BSD-style license |
5 * that can be found in the LICENSE file in the root of the source | 5 * that can be found in the LICENSE file in the root of the source |
6 * tree. An additional intellectual property rights grant can be found | 6 * tree. An additional intellectual property rights grant can be found |
7 * in the file PATENTS. All contributing project authors may | 7 * in the file PATENTS. All contributing project authors may |
8 * be found in the AUTHORS file in the root of the source tree. | 8 * be found in the AUTHORS file in the root of the source tree. |
9 */ | 9 */ |
10 | 10 |
(...skipping 54 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
65 kCaptureStream, // Noise stream. | 65 kCaptureStream, // Noise stream. |
66 }; | 66 }; |
67 | 67 |
68 // Provides access point to the frequency domain. | 68 // Provides access point to the frequency domain. |
69 class TransformCallback : public LappedTransform::Callback { | 69 class TransformCallback : public LappedTransform::Callback { |
70 public: | 70 public: |
71 TransformCallback(IntelligibilityEnhancer* parent, AudioSource source); | 71 TransformCallback(IntelligibilityEnhancer* parent, AudioSource source); |
72 | 72 |
73 // All in frequency domain, receives input |in_block|, applies | 73 // All in frequency domain, receives input |in_block|, applies |
74 // intelligibility enhancement, and writes result to |out_block|. | 74 // intelligibility enhancement, and writes result to |out_block|. |
75 virtual void ProcessAudioBlock(const std::complex<float>* const* in_block, | 75 void ProcessAudioBlock(const std::complex<float>* const* in_block, |
76 int in_channels, | 76 int in_channels, |
77 int frames, | 77 int frames, |
78 int out_channels, | 78 int out_channels, |
79 std::complex<float>* const* out_block); | 79 std::complex<float>* const* out_block) override; |
80 | 80 |
81 private: | 81 private: |
82 IntelligibilityEnhancer* parent_; | 82 IntelligibilityEnhancer* parent_; |
83 AudioSource source_; | 83 AudioSource source_; |
84 }; | 84 }; |
85 friend class TransformCallback; | 85 friend class TransformCallback; |
86 | 86 |
87 // Sends streams to ProcessClearBlock or ProcessNoiseBlock based on source. | 87 // Sends streams to ProcessClearBlock or ProcessNoiseBlock based on source. |
88 void DispatchAudio(AudioSource source, | 88 void DispatchAudio(AudioSource source, |
89 const std::complex<float>* in_block, | 89 const std::complex<float>* in_block, |
(...skipping 21 matching lines...) Expand all Loading... |
111 // Negative gains are set to 0. Stores the results in |sols|. | 111 // Negative gains are set to 0. Stores the results in |sols|. |
112 void SolveForGainsGivenLambda(float lambda, int start_freq, float* sols); | 112 void SolveForGainsGivenLambda(float lambda, int start_freq, float* sols); |
113 | 113 |
114 // Computes variance across ERB filters from freq variance |var|. | 114 // Computes variance across ERB filters from freq variance |var|. |
115 // Stores in |result|. | 115 // Stores in |result|. |
116 void FilterVariance(const float* var, float* result); | 116 void FilterVariance(const float* var, float* result); |
117 | 117 |
118 // Returns dot product of vectors specified by size |length| arrays |a|,|b|. | 118 // Returns dot product of vectors specified by size |length| arrays |a|,|b|. |
119 static float DotProduct(const float* a, const float* b, int length); | 119 static float DotProduct(const float* a, const float* b, int length); |
120 | 120 |
121 static const int kErbResolution; | |
122 static const int kWindowSizeMs; | 121 static const int kWindowSizeMs; |
123 static const int kChunkSizeMs; | 122 static const int kChunkSizeMs; |
124 static const int kAnalyzeRate; // Default for |analysis_rate_|. | 123 static const int kAnalyzeRate; // Default for |analysis_rate_|. |
125 static const int kVarianceRate; // Default for |variance_rate_|. | 124 static const int kVarianceRate; // Default for |variance_rate_|. |
126 static const float kClipFreq; | 125 static const float kClipFreq; |
127 static const float kConfigRho; // Default production and interpretation SNR. | 126 static const float kConfigRho; // Default production and interpretation SNR. |
128 static const float kKbdAlpha; | 127 static const float kKbdAlpha; |
129 static const float kGainChangeLimit; | 128 static const float kGainChangeLimit; |
130 | 129 |
131 const int freqs_; // Num frequencies in frequency domain. | 130 const int freqs_; // Num frequencies in frequency domain. |
(...skipping 37 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
169 // Note: VAD currently does not affect anything in IntelligibilityEnhancer. | 168 // Note: VAD currently does not affect anything in IntelligibilityEnhancer. |
170 VadInst* vad_high_; | 169 VadInst* vad_high_; |
171 VadInst* vad_low_; | 170 VadInst* vad_low_; |
172 rtc::scoped_ptr<int16_t[]> vad_tmp_buffer_; | 171 rtc::scoped_ptr<int16_t[]> vad_tmp_buffer_; |
173 bool has_voice_low_; // Whether voice detected in speech stream. | 172 bool has_voice_low_; // Whether voice detected in speech stream. |
174 }; | 173 }; |
175 | 174 |
176 } // namespace webrtc | 175 } // namespace webrtc |
177 | 176 |
178 #endif // WEBRTC_MODULES_AUDIO_PROCESSING_INTELLIGIBILITY_INTELLIGIBILITY_ENHAN
CER_H_ | 177 #endif // WEBRTC_MODULES_AUDIO_PROCESSING_INTELLIGIBILITY_INTELLIGIBILITY_ENHAN
CER_H_ |
OLD | NEW |