Updated WebRTC APM

I'm now using the entire audio processing module from WebRTC as opposed to individual DSP algorithms pulled from there before. Seems to work better this way.

EchoCanceller.h

@@ -13,6 +13,11 @@
 #include "MediaStreamItf.h"
 #include "utils.h"
 
+namespace webrtc{
+	class AudioProcessing;
+	class AudioFrame;
+}
+
 namespace tgvoip{
 class EchoCanceller{
 
@@ -24,7 +29,7 @@ class EchoCanceller{
 	virtual void Stop();
 	void SpeakerOutCallback(unsigned char* data, size_t len);
 	void Enable(bool enabled);
-	void ProcessInput(unsigned char* data, unsigned char* out, size_t len);
+	void ProcessInput(int16_t* inOut, size_t numSamples);
 	void SetAECStrength(int strength);
 
 private:
@@ -33,24 +38,14 @@ class EchoCanceller{
 	bool enableNS;
 	bool isOn;
 #ifndef TGVOIP_NO_DSP
+	webrtc::AudioProcessing* apm=NULL;
+	webrtc::AudioFrame* audioFrame=NULL;
 	void RunBufferFarendThread();
 	bool didBufferFarend;
-	Mutex aecMutex;
-	void* aec;
-	void* splittingFilter; // webrtc::SplittingFilter
-	void* splittingFilterIn; // webrtc::IFChannelBuffer
-	void* splittingFilterOut; // webrtc::IFChannelBuffer
-	void* splittingFilterFarend; // webrtc::SplittingFilter
-	void* splittingFilterFarendIn; // webrtc::IFChannelBuffer
-	void* splittingFilterFarendOut; // webrtc::IFChannelBuffer
 	Thread* bufferFarendThread;
 	BlockingQueue<int16_t*>* farendQueue;
 	BufferPool* farendBufferPool;
 	bool running;
-	void* ns; // NsxHandle
-	void* agc;
-	int32_t agcMicLevel;
-	//int32_t outstandingFarendFrames=0;
 #endif
 };
 

MediaStreamItf.cpp

@@ -22,7 +22,7 @@ void MediaStreamItf::SetCallback(size_t (*f)(unsigned char *, size_t, void*), vo
 size_t MediaStreamItf::InvokeCallback(unsigned char *data, size_t length){
 	if(callback)
 		return (*callback)(data, length, callbackParam);
-	return NULL;
+	return 0;
 }
 
 AudioMixer::AudioMixer() : bufferPool(960*2, 16), processedQueue(16), semaphore(16, 0){

OpusEncoder.cpp

@@ -30,8 +30,8 @@ tgvoip::OpusEncoder::OpusEncoder(MediaStreamItf *source, bool needSecondary):que
 	complexity=10;
 	frameDuration=20;
 	levelMeter=NULL;
-	mediumCorrectionBitrate=ServerConfig::GetSharedInstance()->GetInt("audio_medium_fec_bitrate", 10000);
-	strongCorrectionBitrate=ServerConfig::GetSharedInstance()->GetInt("audio_strong_fec_bitrate", 8000);
+	mediumCorrectionBitrate=static_cast<uint32_t>(ServerConfig::GetSharedInstance()->GetInt("audio_medium_fec_bitrate", 10000));
+	strongCorrectionBitrate=static_cast<uint32_t>(ServerConfig::GetSharedInstance()->GetInt("audio_strong_fec_bitrate", 8000));
 	mediumCorrectionMultiplier=ServerConfig::GetSharedInstance()->GetDouble("audio_medium_fec_multiplier", 1.5);
 	strongCorrectionMultiplier=ServerConfig::GetSharedInstance()->GetDouble("audio_strong_fec_multiplier", 2.0);
 	secondaryEncoderEnabled=false;
@@ -80,15 +80,15 @@ void tgvoip::OpusEncoder::SetBitrate(uint32_t bitrate){
 	requestedBitrate=bitrate;
 }
 
-void tgvoip::OpusEncoder::Encode(unsigned char *data, size_t len){
+void tgvoip::OpusEncoder::Encode(int16_t* data, size_t len){
 	if(requestedBitrate!=currentBitrate){
 		opus_encoder_ctl(enc, OPUS_SET_BITRATE(requestedBitrate));
 		currentBitrate=requestedBitrate;
 		LOGV("opus_encoder: setting bitrate to %u", currentBitrate);
 	}
 	if(levelMeter)
-		levelMeter->Update(reinterpret_cast<int16_t *>(data), len/2);
-	int32_t r=opus_encode(enc, (int16_t*)data, len/2, buffer, 4096);
+		levelMeter->Update(data, len);
+	int32_t r=opus_encode(enc, data, static_cast<int>(len), buffer, 4096);
 	if(r<=0){
 		LOGE("Error encoding: %d", r);
 	}else if(r==1){
@@ -98,7 +98,7 @@ void tgvoip::OpusEncoder::Encode(unsigned char *data, size_t len){
 		int32_t secondaryLen=0;
 		unsigned char secondaryBuffer[128];
 		if(secondaryEncoderEnabled && secondaryEncoder){
-			secondaryLen=opus_encode(secondaryEncoder, (int16_t*)data, len/2, secondaryBuffer, sizeof(secondaryBuffer));
+			secondaryLen=opus_encode(secondaryEncoder, data, static_cast<int>(len), secondaryBuffer, sizeof(secondaryBuffer));
 			//LOGV("secondaryLen %d", secondaryLen);
 		}
 		InvokeCallback(buffer, (size_t)r, secondaryBuffer, (size_t)secondaryLen);
@@ -132,33 +132,30 @@ void tgvoip::OpusEncoder::SetEchoCanceller(EchoCanceller* aec){
 }
 
 void tgvoip::OpusEncoder::RunThread(){
-	unsigned char buf[960*2];
 	uint32_t bufferedCount=0;
 	uint32_t packetsPerFrame=frameDuration/20;
 	LOGV("starting encoder, packets per frame=%d", packetsPerFrame);
-	unsigned char* frame;
+	int16_t* frame;
 	if(packetsPerFrame>1)
-		frame=(unsigned char *) malloc(960*2*packetsPerFrame);
+		frame=(int16_t*) malloc(960*2*packetsPerFrame);
 	else
 		frame=NULL;
 	while(running){
-		unsigned char* packet=(unsigned char*)queue.GetBlocking();
+		int16_t* packet=(int16_t*)queue.GetBlocking();
 		if(packet){
 			if(echoCanceller)
-				echoCanceller->ProcessInput(packet, buf, 960*2);
-			else
-				memcpy(buf, packet, 960*2);
+				echoCanceller->ProcessInput(packet, 960);
 			if(packetsPerFrame==1){
-				Encode(buf, 960*2);
+				Encode(packet, 960);
 			}else{
-				memcpy(frame+(960*2*bufferedCount), buf, 960*2);
+				memcpy(frame+(960*bufferedCount), packet, 960*2);
 				bufferedCount++;
 				if(bufferedCount==packetsPerFrame){
-					Encode(frame, 960*2*packetsPerFrame);
+					Encode(frame, 960*packetsPerFrame);
 					bufferedCount=0;
 				}
 			}
-			bufferPool.Reuse(packet);
+			bufferPool.Reuse(reinterpret_cast<unsigned char *>(packet));
 		}
 	}
 	if(frame)

OpusEncoder.h

@@ -41,7 +41,7 @@ class OpusEncoder{
 private:
 	static size_t Callback(unsigned char* data, size_t len, void* param);
 	void RunThread();
-	void Encode(unsigned char* data, size_t len);
+	void Encode(int16_t* data, size_t len);
 	void InvokeCallback(unsigned char* data, size_t length, unsigned char* secondaryData, size_t secondaryLength);
 	MediaStreamItf* source;
 	::OpusEncoder* enc;

VoIPController.cpp

@@ -54,6 +54,7 @@ bool VoIPController::didInitWin32TimeScale = false;
 
 #ifdef __ANDROID__
 #include "os/android/JNIUtilities.h"
+#include "os/android/AudioInputAndroid.h"
 extern jclass jniUtilitiesClass;
 #endif
 
@@ -1042,6 +1043,20 @@ void VoIPController::InitializeAudio(){
 	audioIO=audio::AudioIO::Create();
 	audioInput=audioIO->GetInput();
 	audioOutput=audioIO->GetOutput();
+#ifdef __ANDROID__
+	audio::AudioInputAndroid* androidInput=dynamic_cast<audio::AudioInputAndroid*>(audioInput);
+	if(androidInput){
+		unsigned int effects=androidInput->GetEnabledEffects();
+		if(!(effects & audio::AudioInputAndroid::EFFECT_AEC)){
+			config.enableAEC=true;
+			LOGI("Forcing software AEC because built-in is not good");
+		}
+		if(!(effects & audio::AudioInputAndroid::EFFECT_NS)){
+			config.enableNS=true;
+			LOGI("Forcing software NS because built-in is not good");
+		}
+	}
+#endif
 	LOGI("AEC: %d NS: %d AGC: %d", config.enableAEC, config.enableNS, config.enableAGC);
 	echoCanceller=new EchoCanceller(config.enableAEC, config.enableNS, config.enableAGC);
 	encoder=new OpusEncoder(audioInput, true);
@@ -1056,7 +1071,7 @@ void VoIPController::InitializeAudio(){
 #endif
 
 	if(!audioOutput->IsInitialized()){
-		LOGE("Erorr initializing audio playback");
+		LOGE("Error initializing audio playback");
 		lastError=ERROR_AUDIO_IO;
 
 		SetState(STATE_FAILED);
@@ -1102,6 +1117,7 @@ void VoIPController::UpdateAudioOutputState(){
 			areAnyAudioStreamsEnabled=true;
 	}
 	if(audioOutput){
+		LOGV("New audio output state: %d", areAnyAudioStreamsEnabled);
 		if(audioOutput->IsPlaying()!=areAnyAudioStreamsEnabled){
 			if(areAnyAudioStreamsEnabled)
 				audioOutput->Start();

VoIPController.h

@@ -39,7 +39,7 @@
 #include "MessageThread.h"
 #include "utils.h"
 
-#define LIBTGVOIP_VERSION "2.3"
+#define LIBTGVOIP_VERSION "2.4"
 
 #ifdef _WIN32
 #undef GetCurrentTime

client/android/tg_voip_jni.cpp

@@ -499,6 +499,7 @@ extern "C" void tgvoipRegisterNatives(JNIEnv* env){
 		AudioInputAndroid::releaseMethod=env->GetMethodID(cls, "release", "()V");
 		AudioInputAndroid::startMethod=env->GetMethodID(cls, "start", "()Z");
 		AudioInputAndroid::stopMethod=env->GetMethodID(cls, "stop", "()V");
+		AudioInputAndroid::getEnabledEffectsMaskMethod=env->GetMethodID(cls, "getEnabledEffectsMask", "()I");
 
 		cls=env->FindClass(TGVOIP_PACKAGE_PATH "/AudioTrackJNI");
 		AudioOutputAndroid::jniClass=(jclass) env->NewGlobalRef(cls);

configure

@@ -1,6 +1,6 @@
 #! /bin/sh
 # Guess values for system-dependent variables and create Makefiles.
-# Generated by GNU Autoconf 2.69 for libtgvoip 2.2.4.
+# Generated by GNU Autoconf 2.69 for libtgvoip 2.4.
 #
 # Report bugs to <https://github.com/grishka/libtgvoip/issues>.
 #
@@ -590,8 +590,8 @@ MAKEFLAGS=
 # Identity of this package.
 PACKAGE_NAME='libtgvoip'
 PACKAGE_TARNAME='libtgvoip'
-PACKAGE_VERSION='2.2.4'
-PACKAGE_STRING='libtgvoip 2.2.4'
+PACKAGE_VERSION='2.4'
+PACKAGE_STRING='libtgvoip 2.4'
 PACKAGE_BUGREPORT='https://github.com/grishka/libtgvoip/issues'
 PACKAGE_URL=''
 
@@ -1360,7 +1360,7 @@ if test "$ac_init_help" = "long"; then
   # Omit some internal or obsolete options to make the list less imposing.
   # This message is too long to be a string in the A/UX 3.1 sh.
   cat <<_ACEOF
-\`configure' configures libtgvoip 2.2.4 to adapt to many kinds of systems.
+\`configure' configures libtgvoip 2.4 to adapt to many kinds of systems.
 
 Usage: $0 [OPTION]... [VAR=VALUE]...
 
@@ -1430,7 +1430,7 @@ fi
 
 if test -n "$ac_init_help"; then
   case $ac_init_help in
-     short | recursive ) echo "Configuration of libtgvoip 2.2.4:";;
+     short | recursive ) echo "Configuration of libtgvoip 2.4:";;
    esac
   cat <<\_ACEOF
 
@@ -1552,7 +1552,7 @@ fi
 test -n "$ac_init_help" && exit $ac_status
 if $ac_init_version; then
   cat <<\_ACEOF
-libtgvoip configure 2.2.4
+libtgvoip configure 2.4
 generated by GNU Autoconf 2.69
 
 Copyright (C) 2012 Free Software Foundation, Inc.
@@ -2264,7 +2264,7 @@ cat >config.log <<_ACEOF
 This file contains any messages produced by compilers while
 running configure, to aid debugging if configure makes a mistake.
 
-It was created by libtgvoip $as_me 2.2.4, which was
+It was created by libtgvoip $as_me 2.4, which was
 generated by GNU Autoconf 2.69.  Invocation command line was
 
   $ $0 $@
@@ -3130,7 +3130,7 @@ fi
 
 # Define the identity of the package.
  PACKAGE='libtgvoip'
- VERSION='2.2.4'
+ VERSION='2.4'
 
 
 cat >>confdefs.h <<_ACEOF
@@ -19133,7 +19133,7 @@ cat >>$CONFIG_STATUS <<\_ACEOF || ac_write_fail=1
 # report actual input values of CONFIG_FILES etc. instead of their
 # values after options handling.
 ac_log="
-This file was extended by libtgvoip $as_me 2.2.4, which was
+This file was extended by libtgvoip $as_me 2.4, which was
 generated by GNU Autoconf 2.69.  Invocation command line was
 
   CONFIG_FILES    = $CONFIG_FILES
@@ -19199,7 +19199,7 @@ _ACEOF
 cat >>$CONFIG_STATUS <<_ACEOF || ac_write_fail=1
 ac_cs_config="`$as_echo "$ac_configure_args" | sed 's/^ //; s/[\\""\`\$]/\\\\&/g'`"
 ac_cs_version="\\
-libtgvoip config.status 2.2.4
+libtgvoip config.status 2.4
 configured by $0, generated by GNU Autoconf 2.69,
   with options \\"\$ac_cs_config\\"
 

configure.ac

@@ -2,7 +2,7 @@
 # Process this file with autoconf to produce a configure script.
 
 AC_PREREQ([2.69])
-AC_INIT([libtgvoip], [2.2.4], [https://github.com/grishka/libtgvoip/issues])
+AC_INIT([libtgvoip], [2.4], [https://github.com/grishka/libtgvoip/issues])
 AC_CONFIG_SRCDIR([config.h.in])
 AC_CONFIG_HEADERS([config.h])
 AM_INIT_AUTOMAKE([subdir-objects])