diff --git a/src/media/audio/audio-processing/audio_processor.h b/src/media/audio/audio-processing/audio_processor.h
index e89289bcd3a197f4fa271ec3b3b0c99dc8be0367..55a4c1171743ffa605762a1d297dd39b4a7bcab0 100644
--- a/src/media/audio/audio-processing/audio_processor.h
+++ b/src/media/audio/audio-processing/audio_processor.h
@@ -82,6 +82,11 @@ public:
*/
virtual void enableAutomaticGainControl(bool enabled) = 0;
+ /**
+ * @brief Set the status of voice activity detection
+ */
+ virtual void enableVoiceActivityDetection(bool enabled) = 0;
+
protected:
AudioFrameResizer playbackQueue_;
AudioFrameResizer recordQueue_;
diff --git a/src/media/audio/audio-processing/null_audio_processor.h b/src/media/audio/audio-processing/null_audio_processor.h
index 18c3288c498a5b67462a64a6742975718e84b8e4..2cff551c0f791f793c4715fa3fe5a343222d242c 100644
--- a/src/media/audio/audio-processing/null_audio_processor.h
+++ b/src/media/audio/audio-processing/null_audio_processor.h
@@ -35,6 +35,8 @@ public:
void enableNoiseSuppression(bool) override {};
void enableAutomaticGainControl(bool) override {};
+
+ void enableVoiceActivityDetection(bool) override {};
};
} // namespace jami
diff --git a/src/media/audio/audio-processing/speex.cpp b/src/media/audio/audio-processing/speex.cpp
index 8d4ed45ee62cf2f3f399938fabcf562e70733ac6..2040005b0d6eb2c9e75e13a71b7ca99e3451b94b 100644
--- a/src/media/audio/audio-processing/speex.cpp
+++ b/src/media/audio/audio-processing/speex.cpp
@@ -152,6 +152,21 @@ SpeexAudioProcessor::enableAutomaticGainControl(bool enabled)
}
}
+void
+SpeexAudioProcessor::enableVoiceActivityDetection(bool enabled)
+{
+ JAMI_DBG("[speex-dsp] enableVoiceActivityDetection %d", enabled);
+
+ shouldDetectVoice = enabled;
+
+ spx_int32_t speexSetValue = (spx_int32_t) enabled;
+ for (auto& channelPreprocessorState : preprocessorStates) {
+ speex_preprocess_ctl(channelPreprocessorState.get(),
+ SPEEX_PREPROCESS_SET_VAD,
+ &speexSetValue);
+ }
+}
+
std::shared_ptr<AudioFrame>
SpeexAudioProcessor::getProcessed()
{
@@ -212,7 +227,7 @@ SpeexAudioProcessor::getProcessed()
iProcBuffer.interleave((AudioSample*) processed->pointer()->data[0]);
// add stabilized voice activity to the AudioFrame
- processed->has_voice = getStabilizedVoiceActivity(overallVad);
+ processed->has_voice = shouldDetectVoice && getStabilizedVoiceActivity(overallVad);
return processed;
}
diff --git a/src/media/audio/audio-processing/speex.h b/src/media/audio/audio-processing/speex.h
index b68714047a76f338db8a61ea7e4282745f09b42e..80b64be14c2ae625f0f3586f47740de4ec7d5224 100644
--- a/src/media/audio/audio-processing/speex.h
+++ b/src/media/audio/audio-processing/speex.h
@@ -43,6 +43,7 @@ public:
void enableEchoCancel(bool enabled) override;
void enableNoiseSuppression(bool enabled) override;
void enableAutomaticGainControl(bool enabled) override;
+ void enableVoiceActivityDetection(bool enabled) override;
private:
using SpeexEchoStatePtr = std::unique_ptr<SpeexEchoState, void (*)(SpeexEchoState*)>;
@@ -59,5 +60,9 @@ private:
// if we should do echo cancellation
bool shouldAEC {false};
+
+ // if we should do voice activity detection
+ // preprocess_run returns 1 if vad is disabled, so we have to know whether or not to ignore it
+ bool shouldDetectVoice {false};
};
} // namespace jami
diff --git a/src/media/audio/audio-processing/webrtc.cpp b/src/media/audio/audio-processing/webrtc.cpp
index 519c6e6cc08f156938e143cdf0a4b0bbf548b0af..5f99a9eacdb20d34fa663cf5ecb7d01ce999934a 100644
--- a/src/media/audio/audio-processing/webrtc.cpp
+++ b/src/media/audio/audio-processing/webrtc.cpp
@@ -53,20 +53,6 @@ WebRTCAudioProcessor::WebRTCAudioProcessor(AudioFormat format, unsigned frameSiz
JAMI_ERR("[webrtc-ap] Error initialising audio processing module");
}
- // voice activity
- if (apm->voice_detection()->Enable(true) != webrtcNoError) {
- JAMI_ERR("[webrtc-ap] Error enabling voice detection");
- }
- // TODO: change likelihood?
- if (apm->voice_detection()->set_likelihood(webrtc::VoiceDetection::kVeryLowLikelihood)
- != webrtcNoError) {
- JAMI_ERR("[webrtc-ap] Error setting voice detection likelihood");
- }
- // asserted to be 10 in voice_detection_impl.cc
- if (apm->voice_detection()->set_frame_size_ms(10) != webrtcNoError) {
- JAMI_ERR("[webrtc-ap] Error setting voice detection frame size");
- }
-
JAMI_INFO("[webrtc-ap] Done initializing");
}
@@ -118,6 +104,23 @@ WebRTCAudioProcessor::enableEchoCancel(bool enabled)
}
}
+void
+WebRTCAudioProcessor::enableVoiceActivityDetection(bool enabled)
+{
+ JAMI_DBG("[webrtc-ap] enableVoiceActivityDetection %d", enabled);
+ if (apm->voice_detection()->Enable(enabled) != webrtcNoError) {
+ JAMI_ERR("[webrtc-ap] Error enabling voice activation detection");
+ }
+ if (apm->voice_detection()->set_likelihood(webrtc::VoiceDetection::kVeryLowLikelihood)
+ != webrtcNoError) {
+ JAMI_ERR("[webrtc-ap] Error setting voice detection likelihood");
+ }
+ // asserted to be 10 in voice_detection_impl.cc
+ if (apm->voice_detection()->set_frame_size_ms(10) != webrtcNoError) {
+ JAMI_ERR("[webrtc-ap] Error setting voice detection frame size");
+ }
+}
+
std::shared_ptr<AudioFrame>
WebRTCAudioProcessor::getProcessed()
{
@@ -193,7 +196,8 @@ WebRTCAudioProcessor::getProcessed()
format_.nb_channels);
iRecordBuffer_.interleave((AudioSample*) processed->pointer()->data[0]);
- processed->has_voice = getStabilizedVoiceActivity(apm->voice_detection()->stream_has_voice());
+ processed->has_voice = apm->voice_detection()->is_enabled()
+ && getStabilizedVoiceActivity(apm->voice_detection()->stream_has_voice());
return processed;
}
diff --git a/src/media/audio/audio-processing/webrtc.h b/src/media/audio/audio-processing/webrtc.h
index 5d0295c142a349bc81a4a0bf8ee41a7b5b07e9cb..bfb17dde5d8d8afbb67cf9f24dff9af78020b8c5 100644
--- a/src/media/audio/audio-processing/webrtc.h
+++ b/src/media/audio/audio-processing/webrtc.h
@@ -38,6 +38,7 @@ public:
void enableEchoCancel(bool enabled) override;
void enableNoiseSuppression(bool enabled) override;
void enableAutomaticGainControl(bool enabled) override;
+ void enableVoiceActivityDetection(bool enabled) override;
private:
std::unique_ptr<webrtc::AudioProcessing> apm;
diff --git a/src/media/audio/audiolayer.cpp b/src/media/audio/audiolayer.cpp
index b640999c88c8a113480d1816bdff453686effdc2..e45e740393743f98d0e16448aa091c571aec8a82 100644
--- a/src/media/audio/audiolayer.cpp
+++ b/src/media/audio/audiolayer.cpp
@@ -28,13 +28,12 @@
#include "tonecontrol.h"
#include "client/ring_signal.h"
-// TODO: decide which library to use/how to decide (compile time? runtime?)
+#include "audio-processing/null_audio_processor.h"
#if HAVE_WEBRTC_AP
#include "audio-processing/webrtc.h"
-#elif HAVE_SPEEXDSP
+#endif
+#if HAVE_SPEEXDSP
#include "audio-processing/speex.h"
-#else
-#include "audio-processing/null_audio_processor.h"
#endif
#include <ctime>
@@ -54,8 +53,16 @@ AudioLayer::AudioLayer(const AudioPreference& pref)
, urgentRingBuffer_("urgentRingBuffer_id", SIZEBUF, audioFormat_)
, resampler_(new Resampler)
, lastNotificationTime_()
+ , pref_(pref)
{
urgentRingBuffer_.createReadOffset(RingBufferPool::DEFAULT_ID);
+
+ JAMI_INFO("[audiolayer] AGC: %d, noiseReduce: %d, VAD: %d, echoCancel: %s, audioProcessor: %s",
+ pref_.isAGCEnabled(),
+ pref.getNoiseReduce(),
+ pref.getVadEnabled(),
+ pref.getEchoCanceller().c_str(),
+ pref.getAudioProcessor().c_str());
}
AudioLayer::~AudioLayer() {}
@@ -120,14 +127,27 @@ AudioLayer::recordChanged(bool started)
recordStarted_ = started;
}
+// helper function
+static inline bool
+shouldUseAudioProcessorEchoCancel(bool hasNativeAEC, const std::string& echoCancellerPref)
+{
+ return
+ // user doesn't care which and there is a system AEC
+ (echoCancellerPref == "auto" && !hasNativeAEC)
+ // use specifically wants audioProcessor
+ or (echoCancellerPref == "audioProcessor");
+}
+
void
-AudioLayer::setHasNativeAEC(bool hasEAC)
+AudioLayer::setHasNativeAEC(bool hasNativeAEC)
{
+ JAMI_INFO("[audiolayer] setHasNativeAEC: %d", hasNativeAEC);
std::lock_guard<std::mutex> lock(audioProcessorMutex);
- hasNativeAEC_ = hasEAC;
+ hasNativeAEC_ = hasNativeAEC;
// if we have a current audio processor, tell it to enable/disable its own AEC
if (audioProcessor) {
- audioProcessor->enableEchoCancel(!hasEAC);
+ audioProcessor->enableEchoCancel(
+ shouldUseAudioProcessorEchoCancel(hasNativeAEC, pref_.getEchoCanceller()));
}
}
@@ -145,14 +165,14 @@ AudioLayer::createAudioProcessor()
AudioFormat formatForProcessor {sample_rate, nb_channels};
-#if HAVE_SPEEXDSP && !HAVE_WEBRTC_AP
- // we are using speex
- // TODO: maybe force this to be equivalent to 20ms? as expected by speex
- auto frame_size = sample_rate / 50u;
-#else
- // we are using either webrtc-ap or null
- auto frame_size = sample_rate / 100u;
-#endif
+ unsigned int frame_size;
+ if (pref_.getAudioProcessor() == "speex") {
+ // TODO: maybe force this to be equivalent to 20ms? as expected by speex
+ frame_size = sample_rate / 50u;
+ } else {
+ frame_size = sample_rate / 100u;
+ }
+
JAMI_WARN("Input {%d Hz, %d channels}",
audioInputFormat_.sample_rate,
audioInputFormat_.nb_channels);
@@ -162,23 +182,41 @@ AudioLayer::createAudioProcessor()
nb_channels,
frame_size);
+ if (pref_.getAudioProcessor() == "webrtc") {
#if HAVE_WEBRTC_AP
- JAMI_INFO("[audiolayer] using webrtc audio processor");
- audioProcessor.reset(new WebRTCAudioProcessor(formatForProcessor, frame_size));
-#elif HAVE_SPEEXDSP
- JAMI_INFO("[audiolayer] using speex audio processor");
- audioProcessor.reset(new SpeexAudioProcessor(formatForProcessor, frame_size));
+ JAMI_WARN("[audiolayer] using WebRTCAudioProcessor");
+ audioProcessor.reset(new WebRTCAudioProcessor(formatForProcessor, frame_size));
+#else
+ JAMI_ERR("[audiolayer] audioProcessor preference is webrtc, but library not linked! "
+ "using NullAudioProcessor instead");
+ audioProcessor.reset(new NullAudioProcessor(formatForProcessor, frame_size));
+#endif
+ } else if (pref_.getAudioProcessor() == "speex") {
+#if HAVE_SPEEXDSP
+ JAMI_WARN("[audiolayer] using SpeexAudioProcessor");
+ audioProcessor.reset(new SpeexAudioProcessor(formatForProcessor, frame_size));
#else
- JAMI_INFO("[audiolayer] using null audio processor");
- audioProcessor.reset(new NullAudioProcessor(formatForProcessor, frame_size));
+ JAMI_ERR("[audiolayer] audioProcessor preference is speex, but library not linked! "
+ "using NullAudioProcessor instead");
+ audioProcessor.reset(new NullAudioProcessor(formatForProcessor, frame_size));
#endif
+ } else if (pref_.getAudioProcessor() == "null") {
+ JAMI_WARN("[audiolayer] using NullAudioProcessor");
+ audioProcessor.reset(new NullAudioProcessor(formatForProcessor, frame_size));
+ } else {
+ JAMI_ERR("[audiolayer] audioProcessor preference not recognized, using NullAudioProcessor "
+ "instead");
+ audioProcessor.reset(new NullAudioProcessor(formatForProcessor, frame_size));
+ }
+
+ audioProcessor->enableNoiseSuppression(pref_.getNoiseReduce());
+
+ audioProcessor->enableAutomaticGainControl(pref_.isAGCEnabled());
- audioProcessor->enableNoiseSuppression(true);
- // TODO: enable AGC?
- audioProcessor->enableAutomaticGainControl(false);
+ audioProcessor->enableEchoCancel(
+ shouldUseAudioProcessorEchoCancel(hasNativeAEC_, pref_.getEchoCanceller()));
- // can also be updated after creation via setHasNativeAEC
- audioProcessor->enableEchoCancel(!hasNativeAEC_);
+ audioProcessor->enableVoiceActivityDetection(pref_.getVadEnabled());
}
// must acquire lock beforehand
diff --git a/src/media/audio/audiolayer.h b/src/media/audio/audiolayer.h
index b1d252fdd3acdab76c8234e3da8a88827d5e29cc..c1b1a8cd54ad1ca4830779be54837f89bb56b42c 100644
--- a/src/media/audio/audiolayer.h
+++ b/src/media/audio/audiolayer.h
@@ -250,6 +250,9 @@ protected:
*/
double playbackGain_;
+ // audio processor preferences
+ const AudioPreference& pref_;
+
/**
* Buffers for audio processing
*/
diff --git a/src/media/audio/pulseaudio/audiostream.cpp b/src/media/audio/pulseaudio/audiostream.cpp
index 55676d0bb4568fd48b0954ea45ac1e094d345f4c..683a8c50120617619e64c3990025310acde94311 100644
--- a/src/media/audio/pulseaudio/audiostream.cpp
+++ b/src/media/audio/pulseaudio/audiostream.cpp
@@ -35,11 +35,13 @@ AudioStream::AudioStream(pa_context* c,
unsigned samplrate,
const PaDeviceInfos& infos,
bool ec,
- OnReady onReady, OnData onData)
+ OnReady onReady,
+ OnData onData)
: onReady_(std::move(onReady))
, onData_(std::move(onData))
, audiostream_(nullptr)
, mainloop_(m)
+ , audioType_(type)
{
pa_sample_spec sample_spec = {PA_SAMPLE_S16LE, // PA_SAMPLE_FLOAT32LE,
samplrate,
@@ -88,29 +90,37 @@ AudioStream::AudioStream(pa_context* c,
PA_STREAM_ADJUST_LATENCY | PA_STREAM_AUTO_TIMING_UPDATE | PA_STREAM_START_CORKED);
if (type == AudioDeviceType::PLAYBACK || type == AudioDeviceType::RINGTONE) {
- pa_stream_set_write_callback(audiostream_, [](pa_stream* /*s*/, size_t bytes, void* userdata) {
- static_cast<AudioStream*>(userdata)->onData_(bytes);
- }, this);
+ pa_stream_set_write_callback(
+ audiostream_,
+ [](pa_stream* /*s*/, size_t bytes, void* userdata) {
+ static_cast<AudioStream*>(userdata)->onData_(bytes);
+ },
+ this);
pa_stream_connect_playback(audiostream_,
- infos.name.empty() ? nullptr : infos.name.c_str(),
- &attributes,
- flags,
- nullptr,
- nullptr);
+ infos.name.empty() ? nullptr : infos.name.c_str(),
+ &attributes,
+ flags,
+ nullptr,
+ nullptr);
} else if (type == AudioDeviceType::CAPTURE) {
- pa_stream_set_read_callback(audiostream_, [](pa_stream* /*s*/, size_t bytes, void* userdata) {
- static_cast<AudioStream*>(userdata)->onData_(bytes);
- }, this);
+ pa_stream_set_read_callback(
+ audiostream_,
+ [](pa_stream* /*s*/, size_t bytes, void* userdata) {
+ static_cast<AudioStream*>(userdata)->onData_(bytes);
+ },
+ this);
pa_stream_connect_record(audiostream_,
- infos.name.empty() ? nullptr : infos.name.c_str(),
- &attributes,
- flags);
+ infos.name.empty() ? nullptr : infos.name.c_str(),
+ &attributes,
+ flags);
}
}
-void disconnectStream(pa_stream* s) {
+void
+disconnectStream(pa_stream* s)
+{
// make sure we don't get any further callback
pa_stream_set_write_callback(s, nullptr, nullptr);
pa_stream_set_read_callback(s, nullptr, nullptr);
@@ -121,7 +131,9 @@ void disconnectStream(pa_stream* s) {
pa_stream_set_started_callback(s, nullptr, nullptr);
}
-void destroyStream(pa_stream* s) {
+void
+destroyStream(pa_stream* s)
+{
pa_stream_disconnect(s);
pa_stream_set_state_callback(s, nullptr, nullptr);
disconnectStream(s);
@@ -137,6 +149,9 @@ void
AudioStream::start()
{
pa_stream_cork(audiostream_, 0, nullptr, nullptr);
+
+ // trigger echo cancel check
+ moved(audiostream_);
}
void
@@ -147,9 +162,8 @@ AudioStream::stop()
JAMI_DBG("Destroying stream with device %s", pa_stream_get_device_name(audiostream_));
if (pa_stream_get_state(audiostream_) == PA_STREAM_CREATING) {
disconnectStream(audiostream_);
- pa_stream_set_state_callback(audiostream_, [](pa_stream* s, void*){
- destroyStream(s);
- }, nullptr);
+ pa_stream_set_state_callback(
+ audiostream_, [](pa_stream* s, void*) { destroyStream(s); }, nullptr);
} else {
destroyStream(audiostream_);
}
@@ -160,13 +174,53 @@ void
AudioStream::moved(pa_stream* s)
{
audiostream_ = s;
- JAMI_DBG("Stream %d to %s", pa_stream_get_index(s), pa_stream_get_device_name(s));
+ JAMI_DBG("[audiostream] Stream moved: %d, %s",
+ pa_stream_get_index(s),
+ pa_stream_get_device_name(s));
+
+ if (audioType_ == AudioDeviceType::CAPTURE) {
+ // check for echo cancel
+ const char* name = pa_stream_get_device_name(s);
+ if (!name) {
+ JAMI_ERR("[audiostream] moved() unable to get audio stream device");
+ return;
+ }
+
+ pa_context* context = pa_stream_get_context(s);
+ auto* op = pa_context_get_source_info_by_name(
+ context,
+ name,
+ [](pa_context* /*c*/, const pa_source_info* i, int /*eol*/, void* userdata) {
+ AudioStream* thisPtr = (AudioStream*) userdata;
+ // this whole closure gets called twice by pulse for some reason
+ // the 2nd time, i is invalid
+ if (!i) {
+ // JAMI_ERR("[audiostream] source info not found for %s", realName);
+ return;
+ }
+
+ // string compare
+ bool usingEchoCancel = std::string_view(i->driver) == "module-echo-cancel.c";
+ JAMI_WARN("[audiostream] capture stream using pulse echo cancel module? %s (%s)",
+ usingEchoCancel ? "yes" : "no",
+ i->name);
+ if (!thisPtr) {
+ JAMI_ERR("[audiostream] AudioStream pointer became invalid during "
+ "pa_source_info_cb_t callback!");
+ return;
+ }
+ thisPtr->echoCancelCb(usingEchoCancel);
+ },
+ this);
+
+ pa_operation_unref(op);
+ }
}
void
AudioStream::stateChanged(pa_stream* s)
{
- //UNUSED char str[PA_SAMPLE_SPEC_SNPRINT_MAX];
+ // UNUSED char str[PA_SAMPLE_SPEC_SNPRINT_MAX];
switch (pa_stream_get_state(s)) {
case PA_STREAM_CREATING:
diff --git a/src/media/audio/pulseaudio/audiostream.h b/src/media/audio/pulseaudio/audiostream.h
index 779518cba5e1ddf3c582d9912fcf5c41427001be..6f90d3d12f564df9b606a5b45b689a4c11033fd1 100644
--- a/src/media/audio/pulseaudio/audiostream.h
+++ b/src/media/audio/pulseaudio/audiostream.h
@@ -52,7 +52,8 @@ public:
unsigned,
const PaDeviceInfos&,
bool,
- OnReady onReady, OnData onData);
+ OnReady onReady,
+ OnData onData);
~AudioStream();
@@ -88,6 +89,8 @@ public:
bool isReady();
+ void setEchoCancelCb(std::function<void(bool)>&& cb) { echoCancelCb = cb; }
+
private:
NON_COPYABLE(AudioStream);
@@ -109,6 +112,16 @@ private:
* A pointer to the opaque threaded main loop object
*/
pa_threaded_mainloop* mainloop_;
+
+ /**
+ * The type of this audio stream
+ */
+ AudioDeviceType audioType_;
+
+ /**
+ * Function called whenever the stream is moved and we check for an echo canceller
+ */
+ std::function<void(bool)> echoCancelCb;
};
} // namespace jami
diff --git a/src/media/audio/pulseaudio/pulselayer.cpp b/src/media/audio/pulseaudio/pulselayer.cpp
index 134aa7ebc60caee9660ac4147c2b96e6b1446f91..c23f423a5b301aa4322913cd6d154817c0374fd6 100644
--- a/src/media/audio/pulseaudio/pulselayer.cpp
+++ b/src/media/audio/pulseaudio/pulselayer.cpp
@@ -432,11 +432,12 @@ PulseLayer::startStream(AudioDeviceType type)
{
waitForDevices();
PulseMainLoopLock lock(mainloop_.get());
+ bool ec = preference_.getEchoCanceller() == "system"
+ || preference_.getEchoCanceller() == "auto";
// Create Streams
if (type == AudioDeviceType::PLAYBACK) {
if (auto dev_infos = getDeviceInfos(sinkList_, getPreferredPlaybackDevice())) {
- bool ec = preference_.getEchoCanceller() == "system";
createStream(playback_,
type,
*dev_infos,
@@ -451,8 +452,12 @@ PulseLayer::startStream(AudioDeviceType type)
false,
std::bind(&PulseLayer::ringtoneToSpeaker, this));
} else if (type == AudioDeviceType::CAPTURE) {
- if (auto dev_infos = getDeviceInfos(sourceList_, getPreferredCaptureDevice()))
- createStream(record_, type, *dev_infos, true, std::bind(&PulseLayer::readFromMic, this));
+ if (auto dev_infos = getDeviceInfos(sourceList_, getPreferredCaptureDevice())) {
+ createStream(record_, type, *dev_infos, ec, std::bind(&PulseLayer::readFromMic, this));
+
+ // whenever the stream is moved, it will call this cb
+ record_->setEchoCancelCb([this](bool echoCancel) { setHasNativeAEC(echoCancel); });
+ }
}
pa_threaded_mainloop_signal(mainloop_.get(), 0);
diff --git a/src/preferences.cpp b/src/preferences.cpp
index 3a7eb5cd0af04a4434086580574b0aa276ad8980..6e14b7955454899a350a6b584f8a537baefbee3d 100644
--- a/src/preferences.cpp
+++ b/src/preferences.cpp
@@ -117,11 +117,13 @@ static constexpr const char* RECORDPATH_KEY {"recordPath"};
static constexpr const char* ALWAYS_RECORDING_KEY {"alwaysRecording"};
static constexpr const char* VOLUMEMIC_KEY {"volumeMic"};
static constexpr const char* VOLUMESPKR_KEY {"volumeSpkr"};
-static constexpr const char* ECHO_CANCELLER {"echoCanceller"};
+static constexpr const char* AUDIO_PROCESSOR_KEY {"audioProcessor"};
static constexpr const char* NOISE_REDUCE_KEY {"noiseReduce"};
static constexpr const char* AGC_KEY {"automaticGainControl"};
static constexpr const char* CAPTURE_MUTED_KEY {"captureMuted"};
static constexpr const char* PLAYBACK_MUTED_KEY {"playbackMuted"};
+static constexpr const char* VAD_KEY {"voiceActivityDetection"};
+static constexpr const char* ECHO_CANCEL_KEY {"echoCancel"};
#ifdef ENABLE_VIDEO
// video preferences
@@ -284,9 +286,11 @@ AudioPreference::AudioPreference()
, alwaysRecording_(false)
, volumemic_(1.0)
, volumespkr_(1.0)
- , echoCanceller_("system")
+ , audioProcessor_("webrtc")
, denoise_(false)
, agcEnabled_(false)
+ , vadEnabled_(true)
+ , echoCanceller_("auto")
, captureMuted_(false)
, playbackMuted_(false)
{}
@@ -424,9 +428,7 @@ AudioPreference::serialize(YAML::Emitter& out) const
// common options
out << YAML::Key << ALWAYS_RECORDING_KEY << YAML::Value << alwaysRecording_;
out << YAML::Key << AUDIO_API_KEY << YAML::Value << audioApi_;
- out << YAML::Key << AGC_KEY << YAML::Value << agcEnabled_;
out << YAML::Key << CAPTURE_MUTED_KEY << YAML::Value << captureMuted_;
- out << YAML::Key << NOISE_REDUCE_KEY << YAML::Value << denoise_;
out << YAML::Key << PLAYBACK_MUTED_KEY << YAML::Value << playbackMuted_;
// pulse submap
@@ -447,7 +449,13 @@ AudioPreference::serialize(YAML::Emitter& out) const
out << YAML::Key << RECORDPATH_KEY << YAML::Value << recordpath_;
out << YAML::Key << VOLUMEMIC_KEY << YAML::Value << volumemic_;
out << YAML::Key << VOLUMESPKR_KEY << YAML::Value << volumespkr_;
- out << YAML::Key << ECHO_CANCELLER << YAML::Value << echoCanceller_;
+
+ // audio processor options, not in a submap
+ out << YAML::Key << AUDIO_PROCESSOR_KEY << YAML::Value << audioProcessor_;
+ out << YAML::Key << AGC_KEY << YAML::Value << agcEnabled_;
+ out << YAML::Key << VAD_KEY << YAML::Value << vadEnabled_;
+ out << YAML::Key << NOISE_REDUCE_KEY << YAML::Value << denoise_;
+ out << YAML::Key << ECHO_CANCEL_KEY << YAML::Value << echoCanceller_;
out << YAML::EndMap;
}
@@ -502,7 +510,9 @@ AudioPreference::unserialize(const YAML::Node& in)
parseValue(node, RECORDPATH_KEY, recordpath_);
parseValue(node, VOLUMEMIC_KEY, volumemic_);
parseValue(node, VOLUMESPKR_KEY, volumespkr_);
- parseValue(node, ECHO_CANCELLER, echoCanceller_);
+ parseValue(node, AUDIO_PROCESSOR_KEY, audioProcessor_);
+ parseValue(node, VAD_KEY, vadEnabled_);
+ parseValue(node, ECHO_CANCEL_KEY, echoCanceller_);
}
#ifdef ENABLE_VIDEO
diff --git a/src/preferences.h b/src/preferences.h
index 6ff9765c8da5e3323911ffe2a9cbc6a952046dad..2a2985f0ff7d5f51ef839057c8fdc38c47b2ef40 100644
--- a/src/preferences.h
+++ b/src/preferences.h
@@ -236,9 +236,17 @@ public:
void setPlaybackMuted(bool muted) { playbackMuted_ = muted; }
+ const std::string& getAudioProcessor() const { return audioProcessor_; }
+
+ void setAudioProcessor(const std::string& ap) { audioProcessor_ = ap; }
+
+ bool getVadEnabled() const { return vadEnabled_; }
+
+ void setVad(bool enable) { vadEnabled_ = enable; }
+
const std::string& getEchoCanceller() const { return echoCanceller_; }
- void setEchoCanceller(const std::string& ec) { echoCanceller_ = ec; }
+ void setEchoCancel(std::string& canceller) { echoCanceller_ = canceller; }
private:
std::string audioApi_;
@@ -265,10 +273,14 @@ private:
bool alwaysRecording_;
double volumemic_;
double volumespkr_;
- std::string echoCanceller_;
+ // audio processor preferences
+ std::string audioProcessor_;
bool denoise_;
bool agcEnabled_;
+ bool vadEnabled_;
+ std::string echoCanceller_;
+
bool captureMuted_;
bool playbackMuted_;
constexpr static const char* const CONFIG_LABEL = "audio";