OLD | NEW |
| (Empty) |
1 /* | |
2 * Copyright (c) 2012 The WebRTC project authors. All Rights Reserved. | |
3 * | |
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 | |
6 * tree. An additional intellectual property rights grant can be found | |
7 * in the file PATENTS. All contributing project authors may | |
8 * be found in the AUTHORS file in the root of the source tree. | |
9 */ | |
10 | |
11 // This sub-API supports the following functionalities: | |
12 // | |
13 // - Noise Suppression (NS). | |
14 // - Automatic Gain Control (AGC). | |
15 // - Echo Control (EC). | |
16 // - Receiving side VAD, NS and AGC. | |
17 // - Measurements of instantaneous speech, noise and echo levels. | |
18 // - Generation of AP debug recordings. | |
19 // - Detection of keyboard typing which can disrupt a voice conversation. | |
20 // | |
21 // Usage example, omitting error checking: | |
22 // | |
23 // using namespace webrtc; | |
24 // VoiceEngine* voe = VoiceEngine::Create(); | |
25 // VoEBase* base = VoEBase::GetInterface(); | |
26 // VoEAudioProcessing* ap = VoEAudioProcessing::GetInterface(voe); | |
27 // base->Init(); | |
28 // ap->SetEcStatus(true, kAgcAdaptiveAnalog); | |
29 // ... | |
30 // base->Terminate(); | |
31 // base->Release(); | |
32 // ap->Release(); | |
33 // VoiceEngine::Delete(voe); | |
34 // | |
35 #ifndef WEBRTC_VOICE_ENGINE_VOE_AUDIO_PROCESSING_H | |
36 #define WEBRTC_VOICE_ENGINE_VOE_AUDIO_PROCESSING_H | |
37 | |
38 #include <stdio.h> | |
39 | |
40 #include "webrtc/common_types.h" | |
41 | |
42 namespace webrtc { | |
43 | |
44 class VoiceEngine; | |
45 | |
46 // VoEAudioProcessing | |
47 class WEBRTC_DLLEXPORT VoEAudioProcessing { | |
48 public: | |
49 // Factory for the VoEAudioProcessing sub-API. Increases an internal | |
50 // reference counter if successful. Returns NULL if the API is not | |
51 // supported or if construction fails. | |
52 static VoEAudioProcessing* GetInterface(VoiceEngine* voiceEngine); | |
53 | |
54 // Releases the VoEAudioProcessing sub-API and decreases an internal | |
55 // reference counter. Returns the new reference count. This value should | |
56 // be zero for all sub-API:s before the VoiceEngine object can be safely | |
57 // deleted. | |
58 virtual int Release() = 0; | |
59 | |
60 // Sets Noise Suppression (NS) status and mode. | |
61 // The NS reduces noise in the microphone signal. | |
62 virtual int SetNsStatus(bool enable, NsModes mode = kNsUnchanged) = 0; | |
63 | |
64 // Gets the NS status and mode. | |
65 virtual int GetNsStatus(bool& enabled, NsModes& mode) = 0; | |
66 | |
67 // Sets the Automatic Gain Control (AGC) status and mode. | |
68 // The AGC adjusts the microphone signal to an appropriate level. | |
69 virtual int SetAgcStatus(bool enable, AgcModes mode = kAgcUnchanged) = 0; | |
70 | |
71 // Gets the AGC status and mode. | |
72 virtual int GetAgcStatus(bool& enabled, AgcModes& mode) = 0; | |
73 | |
74 // Sets the AGC configuration. | |
75 // Should only be used in situations where the working environment | |
76 // is well known. | |
77 virtual int SetAgcConfig(AgcConfig config) = 0; | |
78 | |
79 // Gets the AGC configuration. | |
80 virtual int GetAgcConfig(AgcConfig& config) = 0; | |
81 | |
82 // Sets the Echo Control (EC) status and mode. | |
83 // The EC mitigates acoustic echo where a user can hear their own | |
84 // speech repeated back due to an acoustic coupling between the | |
85 // speaker and the microphone at the remote end. | |
86 virtual int SetEcStatus(bool enable, EcModes mode = kEcUnchanged) = 0; | |
87 | |
88 // Gets the EC status and mode. | |
89 virtual int GetEcStatus(bool& enabled, EcModes& mode) = 0; | |
90 | |
91 // Enables the compensation of clock drift between the capture and render | |
92 // streams by the echo canceller (i.e. only using EcMode==kEcAec). It will | |
93 // only be enabled if supported on the current platform; otherwise an error | |
94 // will be returned. Check if the platform is supported by calling | |
95 // |DriftCompensationSupported()|. | |
96 virtual int EnableDriftCompensation(bool enable) = 0; | |
97 virtual bool DriftCompensationEnabled() = 0; | |
98 static bool DriftCompensationSupported(); | |
99 | |
100 // Sets a delay |offset| in ms to add to the system delay reported by the | |
101 // OS, which is used by the AEC to synchronize far- and near-end streams. | |
102 // In some cases a system may introduce a delay which goes unreported by the | |
103 // OS, but which is known to the user. This method can be used to compensate | |
104 // for the unreported delay. | |
105 virtual void SetDelayOffsetMs(int offset) = 0; | |
106 virtual int DelayOffsetMs() = 0; | |
107 | |
108 // Modifies settings for the AEC designed for mobile devices (AECM). | |
109 virtual int SetAecmMode(AecmModes mode = kAecmSpeakerphone, | |
110 bool enableCNG = true) = 0; | |
111 | |
112 // Gets settings for the AECM. | |
113 virtual int GetAecmMode(AecmModes& mode, bool& enabledCNG) = 0; | |
114 | |
115 // Enables a high pass filter on the capture signal. This removes DC bias | |
116 // and low-frequency noise. Recommended to be enabled. | |
117 virtual int EnableHighPassFilter(bool enable) = 0; | |
118 virtual bool IsHighPassFilterEnabled() = 0; | |
119 | |
120 // Gets the VAD/DTX activity for the specified |channel|. | |
121 // The returned value is 1 if frames of audio contains speech | |
122 // and 0 if silence. The output is always 1 if VAD is disabled. | |
123 virtual int VoiceActivityIndicator(int channel) = 0; | |
124 | |
125 // Enables or disables the possibility to retrieve echo metrics and delay | |
126 // logging values during an active call. The metrics are only supported in | |
127 // AEC. | |
128 virtual int SetEcMetricsStatus(bool enable) = 0; | |
129 | |
130 // Gets the current EC metric status. | |
131 virtual int GetEcMetricsStatus(bool& enabled) = 0; | |
132 | |
133 // Gets the instantaneous echo level metrics. | |
134 virtual int GetEchoMetrics(int& ERL, int& ERLE, int& RERL, int& A_NLP) = 0; | |
135 | |
136 // Gets the EC internal |delay_median| and |delay_std| in ms between | |
137 // near-end and far-end. The metric |fraction_poor_delays| is the amount of | |
138 // delay values that potentially can break the EC. The values are aggregated | |
139 // over one second and the last updated metrics are returned. | |
140 virtual int GetEcDelayMetrics(int& delay_median, | |
141 int& delay_std, | |
142 float& fraction_poor_delays) = 0; | |
143 | |
144 // Enables recording of Audio Processing (AP) debugging information. | |
145 // The file can later be used for off-line analysis of the AP performance. | |
146 virtual int StartDebugRecording(const char* fileNameUTF8) = 0; | |
147 | |
148 // Same as above but sets and uses an existing file handle. Takes ownership | |
149 // of |file_handle| and passes it on to the audio processing module. | |
150 virtual int StartDebugRecording(FILE* file_handle) = 0; | |
151 | |
152 // Disables recording of AP debugging information. | |
153 virtual int StopDebugRecording() = 0; | |
154 | |
155 // Enables or disables detection of disturbing keyboard typing. | |
156 // An error notification will be given as a callback upon detection. | |
157 virtual int SetTypingDetectionStatus(bool enable) = 0; | |
158 | |
159 // Gets the current typing detection status. | |
160 virtual int GetTypingDetectionStatus(bool& enabled) = 0; | |
161 | |
162 // Reports the lower of: | |
163 // * Time in seconds since the last typing event. | |
164 // * Time in seconds since the typing detection was enabled. | |
165 // Returns error if typing detection is disabled. | |
166 virtual int TimeSinceLastTyping(int& seconds) = 0; | |
167 | |
168 // Optional setting of typing detection parameters | |
169 // Parameter with value == 0 will be ignored | |
170 // and left with default config. | |
171 // TODO(niklase) Remove default argument as soon as libJingle is updated! | |
172 virtual int SetTypingDetectionParameters(int timeWindow, | |
173 int costPerTyping, | |
174 int reportingThreshold, | |
175 int penaltyDecay, | |
176 int typeEventDelay = 0) = 0; | |
177 | |
178 // Swaps the capture-side left and right audio channels when enabled. It | |
179 // only has an effect when using a stereo send codec. The setting is | |
180 // persistent; it will be applied whenever a stereo send codec is enabled. | |
181 // | |
182 // The swap is applied only to the captured audio, and not mixed files. The | |
183 // swap will appear in file recordings and when accessing audio through the | |
184 // external media interface. | |
185 virtual void EnableStereoChannelSwapping(bool enable) = 0; | |
186 virtual bool IsStereoChannelSwappingEnabled() = 0; | |
187 | |
188 protected: | |
189 VoEAudioProcessing() {} | |
190 virtual ~VoEAudioProcessing() {} | |
191 }; | |
192 | |
193 } // namespace webrtc | |
194 | |
195 #endif // WEBRTC_VOICE_ENGINE_VOE_AUDIO_PROCESSING_H | |
OLD | NEW |