Propagate optional audio stream type to the creation of the audio resources
to enable the playback of TTS content on various stream types.
diff --git a/packages/TtsService/jni/android_tts_SynthProxy.cpp b/packages/TtsService/jni/android_tts_SynthProxy.cpp
index 1958ba9..4247483 100644
--- a/packages/TtsService/jni/android_tts_SynthProxy.cpp
+++ b/packages/TtsService/jni/android_tts_SynthProxy.cpp
@@ -33,6 +33,8 @@
#define DEFAULT_TTS_FORMAT AudioSystem::PCM_16_BIT
#define DEFAULT_TTS_NB_CHANNELS 1
#define DEFAULT_TTS_BUFFERSIZE 1024
+// TODO use the TTS stream type when available
+#define DEFAULT_TTS_STREAM_TYPE AudioSystem::MUSIC
#define USAGEMODE_PLAY_IMMEDIATELY 0
#define USAGEMODE_WRITE_TO_FILE 1
@@ -46,10 +48,12 @@
jmethodID synthProxyMethodPost;
};
+// structure to hold the data that is used each time the TTS engine has synthesized more data
struct afterSynthData_t {
jint jniStorage;
int usageMode;
FILE* outputFile;
+ AudioSystem::stream_type streamType;
};
// ----------------------------------------------------------------------------
@@ -62,6 +66,7 @@
jobject tts_ref;
TtsEngine* mNativeSynthInterface;
AudioTrack* mAudioOut;
+ AudioSystem::stream_type mStreamType;
uint32_t mSampleRate;
AudioSystem::audio_format mAudFormat;
int mNbChannels;
@@ -73,6 +78,7 @@
tts_ref = NULL;
mNativeSynthInterface = NULL;
mAudioOut = NULL;
+ mStreamType = DEFAULT_TTS_STREAM_TYPE;
mSampleRate = DEFAULT_TTS_RATE;
mAudFormat = DEFAULT_TTS_FORMAT;
mNbChannels = DEFAULT_TTS_NB_CHANNELS;
@@ -97,34 +103,33 @@
}
}
- void createAudioOut(uint32_t rate, AudioSystem::audio_format format,
- int channel) {
+ void createAudioOut(AudioSystem::stream_type streamType, uint32_t rate,
+ AudioSystem::audio_format format, int channel) {
mSampleRate = rate;
mAudFormat = format;
mNbChannels = channel;
- // TODO use the TTS stream type
- int streamType = AudioSystem::MUSIC;
+ mStreamType = streamType;
// retrieve system properties to ensure successful creation of the
// AudioTrack object for playback
int afSampleRate;
- if (AudioSystem::getOutputSamplingRate(&afSampleRate, streamType) != NO_ERROR) {
+ if (AudioSystem::getOutputSamplingRate(&afSampleRate, mStreamType) != NO_ERROR) {
afSampleRate = 44100;
}
int afFrameCount;
- if (AudioSystem::getOutputFrameCount(&afFrameCount, streamType) != NO_ERROR) {
+ if (AudioSystem::getOutputFrameCount(&afFrameCount, mStreamType) != NO_ERROR) {
afFrameCount = 2048;
}
uint32_t afLatency;
- if (AudioSystem::getOutputLatency(&afLatency, streamType) != NO_ERROR) {
+ if (AudioSystem::getOutputLatency(&afLatency, mStreamType) != NO_ERROR) {
afLatency = 500;
}
uint32_t minBufCount = afLatency / ((1000 * afFrameCount)/afSampleRate);
if (minBufCount < 2) minBufCount = 2;
int minFrameCount = (afFrameCount * rate * minBufCount)/afSampleRate;
- mAudioOut = new AudioTrack(streamType, rate, format, channel,
+ mAudioOut = new AudioTrack(mStreamType, rate, format, channel,
minFrameCount > 4096 ? minFrameCount : 4096,
0, 0, 0, 0); // not using an AudioTrack callback
@@ -142,21 +147,21 @@
// ----------------------------------------------------------------------------
-void prepAudioTrack(SynthProxyJniStorage* pJniData,
- uint32_t rate, AudioSystem::audio_format format, int channel)
-{
+void prepAudioTrack(SynthProxyJniStorage* pJniData, AudioSystem::stream_type streamType,
+ uint32_t rate, AudioSystem::audio_format format, int channel) {
// Don't bother creating a new audiotrack object if the current
- // object is already set.
+ // object is already initialized with the same audio parameters.
if ( pJniData->mAudioOut &&
(rate == pJniData->mSampleRate) &&
(format == pJniData->mAudFormat) &&
- (channel == pJniData->mNbChannels) ){
+ (channel == pJniData->mNbChannels) &&
+ (streamType == pJniData->mStreamType) ){
return;
}
if (pJniData->mAudioOut){
pJniData->killAudio();
}
- pJniData->createAudioOut(rate, format, channel);
+ pJniData->createAudioOut(streamType, rate, format, channel);
}
@@ -186,7 +191,7 @@
}
if (bufferSize > 0) {
- prepAudioTrack(pJniData, rate, format, channel);
+ prepAudioTrack(pJniData, pForAfter->streamType, rate, format, channel);
if (pJniData->mAudioOut) {
pJniData->mAudioOut->write(wav, bufferSize);
//LOGV("AudioTrack wrote: %d bytes", bufferSize);
@@ -241,7 +246,7 @@
SynthProxyJniStorage* pJniStorage = new SynthProxyJniStorage();
prepAudioTrack(pJniStorage,
- DEFAULT_TTS_RATE, DEFAULT_TTS_FORMAT, DEFAULT_TTS_NB_CHANNELS);
+ DEFAULT_TTS_STREAM_TYPE, DEFAULT_TTS_RATE, DEFAULT_TTS_FORMAT, DEFAULT_TTS_NB_CHANNELS);
const char *nativeSoLibNativeString =
env->GetStringUTFChars(nativeSoLib, 0);
@@ -526,7 +531,7 @@
static int
android_tts_SynthProxy_speak(JNIEnv *env, jobject thiz, jint jniData,
- jstring textJavaString)
+ jstring textJavaString, jint javaStreamType)
{
int result = TTS_FAILURE;
@@ -545,6 +550,7 @@
afterSynthData_t* pForAfter = new (afterSynthData_t);
pForAfter->jniStorage = jniData;
pForAfter->usageMode = USAGEMODE_PLAY_IMMEDIATELY;
+ pForAfter->streamType = (AudioSystem::stream_type) javaStreamType;
if (pSynthData->mNativeSynthInterface) {
const char *textNativeString = env->GetStringUTFChars(textJavaString, 0);
@@ -672,7 +678,7 @@
(void*)android_tts_SynthProxy_stop
},
{ "native_speak",
- "(ILjava/lang/String;)I",
+ "(ILjava/lang/String;I)I",
(void*)android_tts_SynthProxy_speak
},
{ "native_synthesizeToFile",
diff --git a/packages/TtsService/src/android/tts/SynthProxy.java b/packages/TtsService/src/android/tts/SynthProxy.java
index bb16b14..41ff92a 100755
--- a/packages/TtsService/src/android/tts/SynthProxy.java
+++ b/packages/TtsService/src/android/tts/SynthProxy.java
@@ -15,6 +15,8 @@
*/
package android.tts;
+import android.media.AudioManager;
+import android.media.AudioSystem;
import android.util.Log;
import java.lang.ref.WeakReference;
@@ -52,8 +54,13 @@
/**
* Synthesize speech and speak it directly using AudioTrack.
*/
- public int speak(String text) {
- return native_speak(mJniData, text);
+ public int speak(String text, int streamType) {
+ if ((streamType > -1) && (streamType < AudioSystem.getNumStreamTypes())) {
+ return native_speak(mJniData, text, streamType);
+ } else {
+ Log.e("SynthProxy", "Trying to speak with invalid stream type " + streamType);
+ return native_speak(mJniData, text, AudioManager.STREAM_MUSIC);
+ }
}
/**
@@ -156,7 +163,7 @@
private native final int native_stop(int jniData);
- private native final int native_speak(int jniData, String text);
+ private native final int native_speak(int jniData, String text, int streamType);
private native final int native_synthesizeToFile(int jniData, String text, String filename);
diff --git a/packages/TtsService/src/android/tts/TtsService.java b/packages/TtsService/src/android/tts/TtsService.java
index 60c30b2..949dfae 100755
--- a/packages/TtsService/src/android/tts/TtsService.java
+++ b/packages/TtsService/src/android/tts/TtsService.java
@@ -22,6 +22,7 @@
import android.content.SharedPreferences;
import android.content.pm.PackageManager;
import android.content.pm.PackageManager.NameNotFoundException;
+import android.media.AudioManager;
import android.media.MediaPlayer;
import android.media.MediaPlayer.OnCompletionListener;
import android.net.Uri;
@@ -109,6 +110,8 @@
private static final int MAX_SPEECH_ITEM_CHAR_LENGTH = 4000;
private static final int MAX_FILENAME_LENGTH = 250;
+ // TODO use the TTS stream type when available
+ private static final int DEFAULT_STREAM_TYPE = AudioManager.STREAM_MUSIC;
private static final String ACTION = "android.intent.action.START_TTS_SERVICE";
private static final String CATEGORY = "android.intent.category.TTS";
@@ -450,14 +453,15 @@
synth.start();
return;
}
+ int streamType = DEFAULT_STREAM_TYPE;
if (params != null){
String language = "";
String country = "";
String variant = "";
for (int i = 0; i < params.size() - 1; i = i + 2){
String param = params.get(i);
- if (param != null){
- if (param.equals(TextToSpeech.Engine.TTS_KEY_PARAM_RATE)){
+ if (param != null) {
+ if (param.equals(TextToSpeech.Engine.TTS_KEY_PARAM_RATE)) {
setSpeechRate("", Integer.parseInt(params.get(i+1)));
} else if (param.equals(TextToSpeech.Engine.TTS_KEY_PARAM_LANGUAGE)){
language = params.get(i+1);
@@ -465,6 +469,12 @@
country = params.get(i+1);
} else if (param.equals(TextToSpeech.Engine.TTS_KEY_PARAM_VARIANT)){
variant = params.get(i+1);
+ } else if (param.equals(TextToSpeech.Engine.TTS_KEY_PARAM_STREAM)) {
+ try {
+ streamType = Integer.parseInt(params.get(i + 1));
+ } catch (NumberFormatException e) {
+ streamType = DEFAULT_STREAM_TYPE;
+ }
}
}
}
@@ -472,7 +482,7 @@
setLanguage("", language, country, variant);
}
}
- nativeSynth.speak(text);
+ nativeSynth.speak(text, streamType);
} catch (InterruptedException e) {
Log.e("TTS speakInternalOnly", "tryLock interrupted");
e.printStackTrace();
@@ -651,8 +661,7 @@
// Utterance is part of the app calling the library
Context ctx;
try {
- ctx = this.createPackageContext(sr.mSourcePackageName,
- 0);
+ ctx = this.createPackageContext(sr.mSourcePackageName, 0);
} catch (NameNotFoundException e) {
e.printStackTrace();
mSpeechQueue.remove(0); // Remove it from the queue and
@@ -675,6 +684,7 @@
}
mPlayer.setOnCompletionListener(this);
try {
+ mPlayer.setAudioStreamType(getStreamTypeFromParams(currentSpeechItem.mParams));
mPlayer.start();
} catch (IllegalStateException e) {
mSpeechQueue.clear();
@@ -695,6 +705,24 @@
}
}
+ private int getStreamTypeFromParams(ArrayList<String> paramList) {
+ int streamType = DEFAULT_STREAM_TYPE;
+ if (paramList == null) {
+ return streamType;
+ }
+ for (int i = 0; i < paramList.size() - 1; i = i + 2) {
+ String param = paramList.get(i);
+ if ((param != null) && (param.equals(TextToSpeech.Engine.TTS_KEY_PARAM_STREAM))) {
+ try {
+ streamType = Integer.parseInt(paramList.get(i + 1));
+ } catch (NumberFormatException e) {
+ streamType = DEFAULT_STREAM_TYPE;
+ }
+ }
+ }
+ return streamType;
+ }
+
private void cleanUpPlayer() {
if (mPlayer != null) {
mPlayer.release();
