Lotsa changes

interface/resources/qml/hifi/tts/TTS.qml

@@ -202,7 +202,6 @@ Rectangle {
 
         TextArea {
             id: messageToSpeak;
-            placeholderText: "<i>Message to Speak</i>";
             font.family: "Fira Sans SemiBold";
             font.pixelSize: 20;
             // Anchors
@@ -229,6 +228,17 @@ Rectangle {
                     event.accepted = true;
                 }
             }
+
+            HifiStylesUit.FiraSansRegular {
+                text: "<i>Input Text to Speak...</i>";
+                size: 20;
+                anchors.fill: parent;
+                anchors.topMargin: 4;
+                anchors.leftMargin: 4;
+                color: hifi.colors.lightGrayText;
+                visible: !parent.activeFocus && messageToSpeak.text === "";
+                verticalAlignment: Text.AlignTop;
+            }
         }
 
 		HifiControlsUit.Button {

interface/src/Application.cpp

@@ -1188,10 +1188,6 @@ Application::Application(int& argc, char** argv, QElapsedTimer& startupTimer, bo
             audioIO->handleRecordedAudioInput(frame->data);
         });
 
-        auto TTS = DependencyManager::get<TTSScriptingInterface>().data();
-        connect(TTS, &TTSScriptingInterface::ttsSampleCreated, audioIO, &AudioClient::handleTTSAudioInput);
-        connect(TTS, &TTSScriptingInterface::clearTTSBuffer, audioIO, &AudioClient::clearTTSBuffer);
-
         connect(audioIO, &AudioClient::inputReceived, [](const QByteArray& audio) {
             static auto recorder = DependencyManager::get<recording::Recorder>();
             if (recorder->isRecording()) {

interface/src/scripting/TTSScriptingInterface.cpp

@@ -37,6 +37,9 @@ TTSScriptingInterface::TTSScriptingInterface() {
     if (FAILED(hr)) {
         qDebug() << "Can't set default voice.";
     }
+
+    _lastSoundAudioInjectorUpdateTimer.setSingleShot(true);
+    connect(&_lastSoundAudioInjectorUpdateTimer, &QTimer::timeout, this, &TTSScriptingInterface::updateLastSoundAudioInjector);
 #endif
 }
 
@@ -58,38 +61,22 @@ private:
 };
 #endif
 
-void TTSScriptingInterface::testTone(const bool& alsoInject) {
+const std::chrono::milliseconds INJECTOR_INTERVAL_MS = std::chrono::milliseconds(100);
-    QByteArray byteArray(480000, 0);
+void TTSScriptingInterface::updateLastSoundAudioInjector() {
-    _lastSoundByteArray.resize(0);
+    if (_lastSoundAudioInjector) {
-    _lastSoundByteArray.resize(480000);
-
-    int32_t a = 0;
-    int16_t* samples = reinterpret_cast<int16_t*>(byteArray.data());
-    for (a = 0; a < 240000; a++) {
-        int16_t temp = (glm::sin(glm::radians((float)a))) * 32768;
-        samples[a] = temp;
-    }
-    emit ttsSampleCreated(_lastSoundByteArray, AudioConstants::NETWORK_FRAME_SAMPLES_PER_CHANNEL * 50, 96);
-
-    if (alsoInject) {
         AudioInjectorOptions options;
         options.position = DependencyManager::get<AvatarManager>()->getMyAvatarPosition();
-
+        _lastSoundAudioInjector->setOptions(options);
-        _lastSoundAudioInjector = AudioInjector::playSound(_lastSoundByteArray, options);
+        _lastSoundAudioInjectorUpdateTimer.start(INJECTOR_INTERVAL_MS);
     }
 }
 
-void TTSScriptingInterface::speakText(const QString& textToSpeak,
+void TTSScriptingInterface::speakText(const QString& textToSpeak) {
-                                      const int& newChunkSize,
-                                      const int& timerInterval,
-                                      const int& sampleRate,
-                                      const int& bitsPerSample,
-                                      const bool& alsoInject) {
 #ifdef WIN32
     WAVEFORMATEX fmt;
     fmt.wFormatTag = WAVE_FORMAT_PCM;
-    fmt.nSamplesPerSec = sampleRate;
+    fmt.nSamplesPerSec = 24000;
-    fmt.wBitsPerSample = bitsPerSample;
+    fmt.wBitsPerSample = 16;
     fmt.nChannels = 1;
     fmt.nBlockAlign = fmt.nChannels * fmt.wBitsPerSample / 8;
     fmt.nAvgBytesPerSec = fmt.nSamplesPerSec * fmt.nBlockAlign;
@@ -156,16 +143,17 @@ void TTSScriptingInterface::speakText(const QString& textToSpeak,
     _lastSoundByteArray.resize(0);
     _lastSoundByteArray.append(buf1, dwSize);
 
-    // Commented out because this doesn't work completely :)
-    // Obviously, commenting this out isn't fit for production, but it's fine for a test PR
-    //emit ttsSampleCreated(_lastSoundByteArray, newChunkSize, timerInterval);
-
-    if (alsoInject) {
     AudioInjectorOptions options;
     options.position = DependencyManager::get<AvatarManager>()->getMyAvatarPosition();
 
-        _lastSoundAudioInjector = AudioInjector::playSound(_lastSoundByteArray, options);
+    if (_lastSoundAudioInjector) {
+        _lastSoundAudioInjector->stop();
+        _lastSoundAudioInjectorUpdateTimer.stop();
     }
+
+    _lastSoundAudioInjector = AudioInjector::playSoundAndDelete(_lastSoundByteArray, options);
+
+    _lastSoundAudioInjectorUpdateTimer.start(INJECTOR_INTERVAL_MS);
 #else
     qDebug() << "Text-to-Speech isn't currently supported on non-Windows platforms.";
 #endif
@@ -174,7 +162,6 @@ void TTSScriptingInterface::speakText(const QString& textToSpeak,
 void TTSScriptingInterface::stopLastSpeech() {
     if (_lastSoundAudioInjector) {
         _lastSoundAudioInjector->stop();
+        _lastSoundAudioInjector = NULL;
     }
-
-    emit clearTTSBuffer();
 }

interface/src/scripting/TTSScriptingInterface.h

@@ -12,6 +12,7 @@
 #define hifi_SpeechScriptingInterface_h
 
 #include <QtCore/QObject>
+#include <QTimer>
 #include <DependencyManager.h>
 #ifdef WIN32
 #pragma warning(disable : 4996)
@@ -31,19 +32,9 @@ public:
     TTSScriptingInterface();
     ~TTSScriptingInterface();
 
-    Q_INVOKABLE void testTone(const bool& alsoInject = false);
+    Q_INVOKABLE void speakText(const QString& textToSpeak);
-    Q_INVOKABLE void speakText(const QString& textToSpeak,
-                               const int& newChunkSize = (AudioConstants::NETWORK_FRAME_SAMPLES_PER_CHANNEL * 50),
-                               const int& timerInterval = 96,
-                               const int& sampleRate = 24000,
-                               const int& bitsPerSample = 16,
-                               const bool& alsoInject = false);
     Q_INVOKABLE void stopLastSpeech();
 
-signals:
-    void ttsSampleCreated(QByteArray outputArray, const int& newChunkSize, const int& timerInterval);
-    void clearTTSBuffer();
-
 private:
 #ifdef WIN32
     class CComAutoInit {
@@ -90,6 +81,8 @@ private:
 
     QByteArray _lastSoundByteArray;
     AudioInjectorPointer _lastSoundAudioInjector;
+    QTimer _lastSoundAudioInjectorUpdateTimer;
+    void updateLastSoundAudioInjector();
 };
 
 #endif // hifi_SpeechScriptingInterface_h

libraries/audio-client/src/AudioClient.cpp

@@ -245,8 +245,6 @@ AudioClient::AudioClient() :
     packetReceiver.registerListener(PacketType::NoisyMute, this, "handleNoisyMutePacket");
     packetReceiver.registerListener(PacketType::MuteEnvironment, this, "handleMuteEnvironmentPacket");
     packetReceiver.registerListener(PacketType::SelectedAudioFormat, this, "handleSelectedAudioFormat");
-
-    connect(&_TTSTimer, &QTimer::timeout, this, &AudioClient::processTTSBuffer);
 }
 
 AudioClient::~AudioClient() {
@@ -1202,45 +1200,6 @@ int rawToWav(const char* rawData, const int& rawLength, const char* wavfn, long
     return 0;
 }
 
-void AudioClient::processTTSBuffer() {
-    Lock lock(_TTSMutex);
-    if (_TTSAudioBuffer.size() > 0) {
-        QByteArray part;
-        part.append(_TTSAudioBuffer.data(), _TTSChunkSize);
-        _TTSAudioBuffer.remove(0, _TTSChunkSize);
-        handleAudioInput(part);
-    } else {
-        _isProcessingTTS = false;
-        _TTSTimer.stop();
-    }
-}
-
-void AudioClient::handleTTSAudioInput(const QByteArray& audio, const int& newChunkSize, const int& timerInterval) {
-    _TTSChunkSize = newChunkSize;
-    _TTSAudioBuffer.append(audio);
-
-    handleLocalEchoAndReverb(_TTSAudioBuffer, 48000, 1);
-
-    //QString filename = QString::number(usecTimestampNow());
-    //QString path = PathUtils::getAppDataPath() + "Audio/" + filename + "-before.wav";
-    //rawToWav(_TTSAudioBuffer.data(), _TTSAudioBuffer.size(), path.toLocal8Bit(), 24000, 1);
-
-    //QByteArray temp;
-
-    _isProcessingTTS = true;
-    _TTSTimer.start(timerInterval);
-
-    //filename = QString::number(usecTimestampNow());
-    //path = PathUtils::getAppDataPath() + "Audio/" + filename + "-after.wav";
-    //rawToWav(temp.data(), temp.size(), path.toLocal8Bit(), 12000, 1);
-}
-
-void AudioClient::clearTTSBuffer() {
-    _TTSAudioBuffer.resize(0);
-    _isProcessingTTS = false;
-    _TTSTimer.stop();
-}
-
 void AudioClient::prepareLocalAudioInjectors(std::unique_ptr<Lock> localAudioLock) {
         bool doSynchronously = localAudioLock.operator bool();
         if (!localAudioLock) {

libraries/audio-client/src/AudioClient.h

@@ -197,11 +197,6 @@ public slots:
     void checkInputTimeout();
     void handleDummyAudioInput();
     void handleRecordedAudioInput(const QByteArray& audio);
-    void handleTTSAudioInput(const QByteArray& audio,
-                             const int& newChunkSize,
-                             const int& timerInterval);
-    void clearTTSBuffer();
-    void processTTSBuffer();
     void reset();
     void audioMixerKilled();