public void StartSpeechRecognition()
{
listeningLabel.Text = "Listening";
recognitionRequest = new SFSpeechAudioBufferRecognitionRequest();
audioEngine.Prepare();
NSError error;
audioEngine.StartAndReturnError(out error);
if (error != null)
{
Console.WriteLine(error.ToString());
return;
}
recognitionTask = speechRecognizer.GetRecognitionTask(recognitionRequest, (SFSpeechRecognitionResult result, NSError err) =>
{
if (err != null)
{
Console.WriteLine(err.ToString());
}
else
{
if (result.Final == true)
{
var results = result.BestTranscription.FormattedString;
if (results.ToLower() == "left")
{
CameraCommand(results);
}
listeningLabel.Text = result.BestTranscription.FormattedString;
}
}
});
}
public void StartVoiceRecognition()
{
try
{
RecognitionTask = SpeechRecognizer.GetRecognitionTask(LiveSpeechRequest, (SFSpeechRecognitionResult result, NSError err) => {
if (result == null)
{
CancelRecording();
return;
}
// Was there an error?
if (err != null)
{
CancelRecording();
return;
}
// Is this the final translation?
if (result != null && result.BestTranscription != null && result.BestTranscription.FormattedString != null)
{
Console.WriteLine("You said \"{0}\".", result.BestTranscription.FormattedString);
TextChanged(result.BestTranscription.FormattedString);
}
if (result.Final)
{
TextChanged(result.BestTranscription.FormattedString, true);
CancelRecording();
return;
}
});
}
catch (Exception ex)
{
Console.WriteLine(ex.Message);
}
}
public void StartRecording(Action <string> callback)
{
var audioSession = AVAudioSession.SharedInstance();
var error = audioSession.SetCategory(AVAudioSessionCategory.Record);
audioSession.SetMode(AVAudioSession.ModeMeasurement, out error);
error = audioSession.SetActive(true, AVAudioSessionSetActiveOptions.NotifyOthersOnDeactivation);
_speechRecognitionRequest = new SFSpeechAudioBufferRecognitionRequest();
var inputNode = _audioEngine.InputNode;
var recordingFormat = inputNode.GetBusOutputFormat(0);
inputNode.InstallTapOnBus(0, 1024, recordingFormat, (buffer, when) =>
{
_speechRecognitionRequest?.Append(buffer);
});
_audioEngine.Prepare();
_audioEngine.StartAndReturnError(out error);
_speechRecognitionTask = _speechRecognizer.GetRecognitionTask(_speechRecognitionRequest, (result, recognitionError) =>
{
if (result != null)
{
var transcribedText = result.BestTranscription.FormattedString;
callback?.Invoke(transcribedText);
}
});
}
// ============== Speech Recognition Functions ============
public void StartRecording()
{
Querybox.Placeholder = "Recording";
recognitionRequest = new SFSpeechAudioBufferRecognitionRequest();
audioEngine.Prepare();
NSError error;
audioEngine.StartAndReturnError(out error);
if (error != null)
{
Console.WriteLine(error.ToString());
return;
}
recognitionTask = speechRecognizer.GetRecognitionTask(recognitionRequest, (SFSpeechRecognitionResult result, NSError err) =>
{
if (err != null)
{
Console.WriteLine(err.ToString());
}
else
{
if (result.Final == true)
{
YouAskedLabel.Text = "You asked: " + result.BestTranscription.FormattedString;
final_query = result.BestTranscription.FormattedString;
}
}
});
}
public void StartVoiceRecognition()
{
RecognitionTask = SpeechRecognizer.GetRecognitionTask(LiveSpeechRequest, (SFSpeechRecognitionResult result, NSError err) =>
{
if (result == null)
{
CancelRecording();
return;
}
if (err != null)
{
CancelRecording();
return;
}
if (result != null && result.BestTranscription != null && result.BestTranscription.FormattedString != null)
{
TextChanged(result.BestTranscription.FormattedString);
}
if (result.Final)
{
TextChanged(result.BestTranscription.FormattedString, true);
CancelRecording();
return;
}
});
}
private void StartRecordingAndRecognizing()
{
_timer = NSTimer.CreateRepeatingScheduledTimer(5, delegate
{
DidFinishTalk();
});
_recognitionTask?.Cancel();
_recognitionTask = null;
var audioSession = AVAudioSession.SharedInstance();
NSError nsError;
nsError = audioSession.SetCategory(AVAudioSessionCategory.PlayAndRecord);
audioSession.SetMode(AVAudioSession.ModeDefault, out nsError);
nsError = audioSession.SetActive(true, AVAudioSessionSetActiveOptions.NotifyOthersOnDeactivation);
audioSession.OverrideOutputAudioPort(AVAudioSessionPortOverride.Speaker, out nsError);
_recognitionRequest = new SFSpeechAudioBufferRecognitionRequest();
var inputNode = _audioEngine.InputNode;
if (inputNode == null)
{
throw new Exception();
}
var recordingFormat = inputNode.GetBusOutputFormat(0);
inputNode.InstallTapOnBus(0, 1024, recordingFormat, (buffer, when) =>
{
_recognitionRequest?.Append(buffer);
});
_audioEngine.Prepare();
_audioEngine.StartAndReturnError(out nsError);
_recognitionTask = _speechRecognizer.GetRecognitionTask(_recognitionRequest, (result, error) =>
{
var isFinal = false;
if (result != null)
{
_recognizedString = result.BestTranscription.FormattedString;
//MessagingCenter.Send<ISpeechToText, string>(this, "STT", _recognizedString);
_timer.Invalidate();
_timer = null;
_timer = NSTimer.CreateRepeatingScheduledTimer(2, delegate
{
DidFinishTalk();
});
}
if (error != null || isFinal)
{
MessagingCenter.Send <ISpeechToText, string>(this, "STT", _recognizedString);
MessagingCenter.Send <ISpeechToText>(this, "Final");
StopRecordingAndRecognition(audioSession);
}
});
}
void StartRecording()
{
// Cancel the previous task if it's running.
recognitionTask?.Cancel();
recognitionTask = null;
var audioSession = AVAudioSession.SharedInstance();
NSError err;
err = audioSession.SetCategory(AVAudioSessionCategory.Record);
audioSession.SetMode(AVAudioSession.ModeMeasurement, out err);
err = audioSession.SetActive(true, AVAudioSessionSetActiveOptions.NotifyOthersOnDeactivation);
// Configure request so that results are returned before audio recording is finished
recognitionRequest = new SFSpeechAudioBufferRecognitionRequest {
ShouldReportPartialResults = true
};
var inputNode = audioEngine.InputNode;
if (inputNode == null)
{
throw new InvalidProgramException("Audio engine has no input node");
}
// A recognition task represents a speech recognition session.
// We keep a reference to the task so that it can be cancelled.
recognitionTask = speechRecognizer.GetRecognitionTask(recognitionRequest, (result, error) => {
var isFinal = false;
if (result != null)
{
textView.Text = result.BestTranscription.FormattedString;
isFinal = result.Final;
}
if (error != null || isFinal)
{
audioEngine.Stop();
inputNode.RemoveTapOnBus(0);
recognitionRequest = null;
recognitionTask = null;
recordButton.Enabled = true;
recordButton.SetTitle("Start Recording", UIControlState.Normal);
}
});
var recordingFormat = inputNode.GetBusOutputFormat(0);
inputNode.InstallTapOnBus(0, 1024, recordingFormat, (buffer, when) => {
recognitionRequest?.Append(buffer);
});
audioEngine.Prepare();
audioEngine.StartAndReturnError(out err);
textView.Text = "(Throw Darts!)";
textView2.Text = "";
}
public void StartSpeechRecognition()
{
StopSpeechRecognizer(false);
StopAudio();
_recognizerText = "";
AVAudioSession.SharedInstance().SetActive(true, AVAudioSessionSetActiveOptions.NotifyOthersOnDeactivation);
var request = new SFSpeechAudioBufferRecognitionRequest();
request.ShouldReportPartialResults = true;
var inputNode = engine.InputNode;
_task = _recognizer.GetRecognitionTask(request, (result, error) =>
{
if (result != null)
{
_recognizerText = result.BestTranscription.FormattedString;
if (_recognizerTimer == null)
{
_recognizerTimer = new Timer(RecognizeTimerElapsed, this, TimeSpan.FromSeconds(1), Timeout.InfiniteTimeSpan);
}
else
{
_recognizerTimer.Change(TimeSpan.FromSeconds(1), Timeout.InfiniteTimeSpan);
}
if (result.Final)
{
SpeechRecognitionResult?.Invoke(this, new SpeechRecognitionEventArgs(_recognizerText));
}
else
{
SpeechRecognitionPartialResult?.Invoke(this, new SpeechRecognitionEventArgs(_recognizerText));
}
}
if (error != null || result?.Final == true)
{
StopSpeechRecognizer(false);
}
});
var recordingFormat = inputNode.GetBusOutputFormat(0);
inputNode.InstallTapOnBus(0, 1024, recordingFormat, (AVAudioPcmBuffer buffer, AVAudioTime when) =>
{
request.Append(buffer);
});
engine.Prepare();
NSError engineError = null;
if (engine.StartAndReturnError(out engineError))
{
}
}
private void StopRecordingAndRecognition(AVAudioSession aVAudioSession = null)
{
if (_audioEngine.Running)
{
_audioEngine.Stop();
_audioEngine.InputNode.RemoveTapOnBus(0);
_recognitionTask?.Cancel();
_recognitionRequest.EndAudio();
_recognitionRequest = null;
_recognitionTask = null;
}
}
private void ResetSpeechToText()
{
_speechRecognizer = new SFSpeechRecognizer();
_speechAudioEngine = new AVAudioEngine();
_speechRecognitionRequest = new SFSpeechAudioBufferRecognitionRequest();
_speechRecognitionRequest.ShouldReportPartialResults = true;
_speechRecognitionRequest.TaskHint = SFSpeechRecognitionTaskHint.Search;
_speechRecognitionRequest.ContextualStrings = new string[] { "for", "the", "a", "an" };
_speechRecognitionTask = null;
}
public void StartRecording(out int warningStatus)
{
if (!IsAuthorized())
{
warningStatus = (int)Warning.AccessDenied;
return;
}
var node = AudioEngine.InputNode;
var recordingFormat = node.GetBusOutputFormat(0);
node.InstallTapOnBus(0, 1024, recordingFormat, (AVAudioPcmBuffer buffer, AVAudioTime when) => {
LiveSpeechRequest.Append(buffer);
});
AudioEngine.Prepare();
NSError error;
AudioEngine.StartAndReturnError(out error);
if (error != null)
{
Console.WriteLine(strings.speechStartRecordProblem);
warningStatus = (int)Warning.RecordProblem;
return;
}
// Play start sound
if (player.IsPlaying)
{
player.Stop();
}
player.Load("Sounds/siri_start.mp3");
player.Play();
RecognitionTask = SpeechRecognizer.GetRecognitionTask(LiveSpeechRequest, (SFSpeechRecognitionResult result, NSError err) =>
{
if (err != null)
{
Console.WriteLine(strings.speechRecordError);
viewController.ProcessSpeech(null);
}
else
{
if (result.Final)
{
viewController.ProcessSpeech(result.BestTranscription.FormattedString);
}
}
});
warningStatus = -1;
}
public SpeechToText()
{
_audioEngine = new AVAudioEngine();
_speechRecognizer = new SFSpeechRecognizer();
_speechRecognitionRequest = new SFSpeechAudioBufferRecognitionRequest();
_speechRecognitionTask = new SFSpeechRecognitionTask();
SFSpeechRecognizer.RequestAuthorization((SFSpeechRecognizerAuthorizationStatus status) =>
{
// We're going to assume that you've selected to authorise the request, otherwise,
// you're wasting your own time.
});
}
public void Start()
{
recognitionTask?.Cancel();
recognitionTask = null;
recognizedText = "";
var audioSession = AVAudioSession.SharedInstance();
NSError err;
err = audioSession.SetCategory(AVAudioSessionCategory.Record);
audioSession.SetMode(AVAudioSession.ModeMeasurement, out err);
err = audioSession.SetActive(true, AVAudioSessionSetActiveOptions.NotifyOthersOnDeactivation);
liveRequest = new SFSpeechAudioBufferRecognitionRequest {
ShouldReportPartialResults = true
};
var node = audioEngine.InputNode;
var recordingFormat = node.GetBusOutputFormat(0);
node.InstallTapOnBus(0, 1024, recordingFormat, (AVAudioPcmBuffer buffer, AVAudioTime when) => {
// Append buffer to recognition request
liveRequest.Append(buffer);
});
recognitionTask = speechRecognizer.GetRecognitionTask(liveRequest, (SFSpeechRecognitionResult result, NSError error) => {
var isFinal = false;
if (result != null)
{
recognizedText = result.BestTranscription.FormattedString;
OnSpeechRecognized(new SpeechRecognizedEvent(recognizedText));
isFinal = result.Final;
}
if (error != null || isFinal)
{
audioSession.SetCategory(AVAudioSessionCategory.Playback);
audioSession.SetMode(AVAudioSession.ModeDefault, out err);
node.RemoveTapOnBus(0);
audioEngine.Dispose();
liveRequest.Dispose();
recognitionTask.Dispose();
liveRequest = null;
recognitionTask = null;
}
});
audioEngine.Prepare();
audioEngine.StartAndReturnError(out err);
}
public void StartRecording()
{
// Setup audio session
var node = AudioEngine.InputNode;
node.RemoveTapOnBus(0);
var recordingFormat = node.GetBusOutputFormat(0);
node.InstallTapOnBus(0, 1024, recordingFormat, (AVAudioPcmBuffer buffer, AVAudioTime when) => {
// Append buffer to recognition request
LiveSpeechRequest.Append(buffer);
});
// Start recording
AudioEngine.Prepare();
NSError error;
AudioEngine.StartAndReturnError(out error);
// Did recording start?
if (error != null)
{
// Handle error and return
input = "Error: Recording init error!";
}
// Start recognition
RecognitionTask = SpeechRecognizer.GetRecognitionTask(LiveSpeechRequest, (SFSpeechRecognitionResult result, NSError err) => {
// Was there an error?
if (err != null)
{
// Handle error
input = "Error: Recognition task";
}
else
{
// Is this the final translation?
if (result.Final)
{
input = result.BestTranscription.ToString();
new UIAlertView("DONE", "You said \"{" + input + "}\".", null, "OK", null).Show();
/*var donecontroller = UIAlertController.Create("DONE","You said \"{" + input + "}\".", UIAlertControllerStyle.Alert);
* donecontroller.AddAction(UIAlertAction.Create("Ok", UIAlertActionStyle.Default, null));
* PresentViewController(donecontroller, true, null);*/
}
}
});
}
void StartRecording ()
{
// Cancel the previous task if it's running.
recognitionTask?.Cancel ();
recognitionTask = null;
var audioSession = AVAudioSession.SharedInstance ();
NSError err;
err = audioSession.SetCategory (AVAudioSessionCategory.Record);
audioSession.SetMode (AVAudioSession.ModeMeasurement, out err);
err = audioSession.SetActive (true, AVAudioSessionSetActiveOptions.NotifyOthersOnDeactivation);
// Configure request so that results are returned before audio recording is finished
recognitionRequest = new SFSpeechAudioBufferRecognitionRequest {
ShouldReportPartialResults = true
};
var inputNode = audioEngine.InputNode;
if (inputNode == null)
throw new InvalidProgramException ("Audio engine has no input node");
// A recognition task represents a speech recognition session.
// We keep a reference to the task so that it can be cancelled.
recognitionTask = speechRecognizer.GetRecognitionTask (recognitionRequest, (result, error) => {
var isFinal = false;
if (result != null) {
textView.Text = result.BestTranscription.FormattedString;
isFinal = result.Final;
}
if (error != null || isFinal) {
audioEngine.Stop ();
inputNode.RemoveTapOnBus (0);
recognitionRequest = null;
recognitionTask = null;
recordButton.Enabled = true;
recordButton.SetTitle ("Start Recording", UIControlState.Normal);
}
});
var recordingFormat = inputNode.GetBusOutputFormat (0);
inputNode.InstallTapOnBus (0, 1024, recordingFormat, (buffer, when) => {
recognitionRequest?.Append (buffer);
});
audioEngine.Prepare ();
audioEngine.StartAndReturnError (out err);
textView.Text = "(Go ahead, I'm listening)";
}
public void StopRecording()
{
try
{
_audioEngine.Stop();
_audioEngine.InputNode.RemoveTapOnBus(0);
_speechRecognitionTask?.Cancel();
_speechRecognitionRequest.EndAudio();
_speechRecognitionRequest = null;
_speechRecognitionTask = null;
}
catch (Exception ex)
{
}
}
private void StartSpeechRecognizer()
{
if (!recording)
{
speechRecognizer = new SFSpeechRecognizer();
node = audioEngine.InputNode;
var recordingFormat = node.GetBusOutputFormat(0);
liveSpeechRequest = new SFSpeechAudioBufferRecognitionRequest();
node.InstallTapOnBus(0, 1024, recordingFormat,
(AVAudioPcmBuffer buffer, AVAudioTime when) =>
{
liveSpeechRequest.Append(buffer);
});
recording = true;
audioEngine.Prepare();
audioEngine.StartAndReturnError(out NSError error);
if (error != null)
{
return;
}
Timer timer = new Timer(2000);
timer.Start();
timer.Elapsed += EndRecognition;
RecognitionTask = speechRecognizer.GetRecognitionTask(liveSpeechRequest,
(SFSpeechRecognitionResult result, NSError err) =>
{
if (err != null)
{
Recorded?.Invoke("");
return;
}
else
{
lastSpokenString = result.BestTranscription.FormattedString;
timer.Stop();
timer.Interval = 2000;
timer.Start();
}
});
}
else
{
Recorded?.Invoke("");
}
}
public string RecognizeVoice(string path, Editor editor)
{
SFSpeechRecognizer recognizer = new SFSpeechRecognizer(NSLocale.CurrentLocale);
// Is the default language supported?
if (recognizer == null)
{
// No, return to caller
return("recognizer is null");
}
// Is recognition available?
if (!recognizer.Available)
{
// No, return to caller
return("recognizer is not available");
}
isTranslatingDone = false;
NSUrl url = new NSUrl(path, false);
SFSpeechUrlRecognitionRequest request = new SFSpeechUrlRecognitionRequest(url);
SFSpeechRecognitionTask sptask = recognizer.GetRecognitionTask(request, (SFSpeechRecognitionResult result, NSError err) =>
{
if (err != null)
{
resultString = err.Description;
isTranslatingDone = true;
editor.Text = resultString;
}
else
{
if (result != null)
{
if (result.Final)
{
resultString = result.BestTranscription.FormattedString;
isTranslatingDone = true;
// call back??
editor.Text = resultString;
}
}
}
});
return("cannot recognize");
}
public void TranslateSpeachToText(FoodListFromSpeachViewModel model)
{
// Setup audio session
var node = AudioEngine.InputNode;
var recordingFormat = node.GetBusOutputFormat(0);
node.InstallTapOnBus(0, 1024, recordingFormat, (AVAudioPcmBuffer buffer, AVAudioTime when) => {
// Append buffer to recognition request
LiveSpeechRequest.Append(buffer);
});
// Start recording
AudioEngine.Prepare();
NSError error;
AudioEngine.StartAndReturnError(out error);
// Did recording start?
if (error != null)
{
// Handle error and return
model.FoodItemToAdd = "Error";
}
// Start recognition
RecognitionTask = SpeechRecognizer.GetRecognitionTask(LiveSpeechRequest,
(SFSpeechRecognitionResult result, NSError err) =>
{
// Was there an error?
if (err != null)
{
// Handle error
}
else
{
// Is this the final translation?
if (result.Final)
{
model.FoodItemToAdd = result.BestTranscription.FormattedString;
}
}
});
}
private async void StopSpeechRecognizer(bool waitForSpeech)
{
try
{
var task = _task;
_task = null;
if (task != null)
{
if (waitForSpeech)
{
await Task.Delay(1000);
}
task.Cancel();
engine.Stop();
engine.InputNode.RemoveTapOnBus(0);
}
}
catch (Exception e)
{
}
}
private void StopRecordingAndRecognition()
{
try
{
if (!_audioEngine.Running)
{
return;
}
_audioEngine.Stop();
_audioEngine.InputNode.RemoveTapOnBus(0);
_recognitionTask?.Cancel();
_recognitionRequest.EndAudio();
_recognitionRequest = null;
_recognitionTask = null;
}
catch (Exception ex)
{
Console.WriteLine(ex.Message);
}
}
void StartListening()
{
/* Setup audio session. */
var node = AudioEngine.InputNode;
var recordingFormat = node.GetBusOutputFormat(0);
node.InstallTapOnBus(0, 1024, recordingFormat, (AVAudioPcmBuffer buffer, AVAudioTime when) => {
/* Append buffer to recognition request. */
LiveSpeechRequest.Append(buffer);
});
/* Start recording */
AudioEngine.Prepare();
NSError error;
AudioEngine.StartAndReturnError(out error);
/* Did recording start? */
if (error != null)
{
/* Do nothing. */
return;
}
/* Start recognition. */
RecognitionTask = SpeechRecognizer.GetRecognitionTask(LiveSpeechRequest, (SFSpeechRecognitionResult result, NSError err) => {
/* Was there an error? */
if (err != null)
{
/* Do nothing. */ }
else
{
// Is this the final translation?
if (result.Final)
{
UserText.Text = result.BestTranscription.FormattedString;
}
}
});
}
static void StartRecording()
{
lock (Lock)
{
if (SpeechRecognizer == null)
{
SpeechRecognizer = new SFSpeechRecognizer();
LiveSpeechRequest = new SFSpeechAudioBufferRecognitionRequest();
}
var audioSession = AVAudioSession.SharedInstance();
audioSession.SetCategory(AVAudioSessionCategory.PlayAndRecord);
audioSession.SetMode(AVAudioSession.ModeDefault, out NSError error);
audioSession.OverrideOutputAudioPort(AVAudioSessionPortOverride.Speaker, out NSError speakerError);
audioSession.SetActive(true);
if (LogErrorAndStop(error) || LogErrorAndStop(speakerError))
{
return;
}
AudioEngine = new AVAudioEngine();
var node = AudioEngine.InputNode;
LiveSpeechRequest.ShouldReportPartialResults = true;
RecognitionTask = SpeechRecognizer.GetRecognitionTask(LiveSpeechRequest, (SFSpeechRecognitionResult result, NSError err) =>
{
if (LogErrorAndStop(err))
{
return;
}
var currentText = result.BestTranscription.FormattedString;
if (currentText.HasValue())
{
Listeners?.Invoke(currentText, result.Final);
}
if (IsContinuous)
{
Timer = new System.Timers.Timer(20000)
{
Enabled = true
};
Timer.Elapsed += (s, ev) =>
{
StopInstances();
StartRecording();
};
Timer.Start();
}
});
var recordingFormat = node.GetBusOutputFormat(0);
node.InstallTapOnBus(0, 1024, recordingFormat, (AVAudioPcmBuffer buffer, AVAudioTime when) =>
{
LiveSpeechRequest.Append(buffer);
});
if (AudioEngine == null)
{
Stop();
return;
}
AudioEngine?.Prepare();
AudioEngine?.StartAndReturnError(out error);
if (LogErrorAndStop(error))
{
return;
}
}
}
public IAsyncOperation <SpeechRecognitionResult> RecognizeAsync()
{
_initialSilenceTimeout = new Timer();
_initialSilenceTimeout.Interval = Math.Max(Timeouts.InitialSilenceTimeout.TotalMilliseconds, 5000);
_initialSilenceTimeout.Elapsed += OnTimeout;
_endSilenceTimeout = new Timer();
_endSilenceTimeout.Interval = Math.Max(Timeouts.EndSilenceTimeout.TotalMilliseconds, 150);
_endSilenceTimeout.Elapsed += OnTimeout;
// Cancel the previous task if it's running.
_recognitionTask?.Cancel();
_recognitionTask = null;
var audioSession = AVAudioSession.SharedInstance();
NSError err;
err = audioSession.SetCategory(AVAudioSessionCategory.Record);
audioSession.SetMode(AVAudioSession.ModeMeasurement, out err);
err = audioSession.SetActive(true, AVAudioSessionSetActiveOptions.NotifyOthersOnDeactivation);
// Configure request to get partial results
_recognitionRequest = new SFSpeechAudioBufferRecognitionRequest
{
ShouldReportPartialResults = true,
TaskHint = SFSpeechRecognitionTaskHint.Dictation
};
var inputNode = _audioEngine.InputNode;
if (inputNode == null)
{
throw new InvalidProgramException("Audio engine has no input node");
}
var tcs = new TaskCompletionSource <SpeechRecognitionResult>();
// Keep a reference to the task so that it can be cancelled.
_recognitionTask = _speechRecognizer.GetRecognitionTask(_recognitionRequest, (result, error) =>
{
var isFinal = false;
var bestMatch = default(SpeechRecognitionResult);
if (result != null)
{
_initialSilenceTimeout.Stop();
_endSilenceTimeout.Stop();
_endSilenceTimeout.Start();
bestMatch = new SpeechRecognitionResult()
{
Text = result.BestTranscription.FormattedString,
Alternates = result.Transcriptions?
.Select(t => new SpeechRecognitionResult()
{
Text = t.FormattedString
})
.ToList()
};
isFinal = result.Final;
OnHypothesisGenerated(bestMatch.Text);
}
if (error != null || isFinal)
{
_initialSilenceTimeout.Stop();
_endSilenceTimeout.Stop();
_audioEngine.Stop();
inputNode.RemoveTapOnBus(0);
inputNode.Reset();
audioSession = AVAudioSession.SharedInstance();
err = audioSession.SetCategory(AVAudioSessionCategory.Playback);
audioSession.SetMode(AVAudioSession.ModeDefault, out err);
err = audioSession.SetActive(false, AVAudioSessionSetActiveOptions.NotifyOthersOnDeactivation);
_recognitionTask = null;
OnStateChanged(SpeechRecognizerState.Idle);
if (bestMatch != null)
{
tcs.TrySetResult(bestMatch);
}
else
{
tcs.TrySetException(new Exception($"Error during speech recognition: {error.LocalizedDescription}"));
}
}
});
var recordingFormat = new AVAudioFormat(sampleRate: 44100, channels: 1);
inputNode.InstallTapOnBus(0, 1024, recordingFormat, (buffer, when) => {
_recognitionRequest?.Append(buffer);
});
_initialSilenceTimeout.Start();
_audioEngine.Prepare();
_audioEngine.StartAndReturnError(out err);
OnStateChanged(SpeechRecognizerState.Capturing);
return(tcs.Task.AsAsyncOperation());
}
public async Task <Tuple <string, Tuple <bool, CMPSpeechError> > > StartRecordingAsync()
{
if (IsRecording() == true)
{
return(new Tuple <string, Tuple <bool, CMPSpeechError> > (string.Empty, null));
}
var authorizationResult = await CheckAuthorizationAsync();
if (authorizationResult == null)
{
return(new Tuple <string, Tuple <bool, CMPSpeechError> >(string.Empty, null));
}
if (authorizationResult.Item1 == false)
{
return(new Tuple <string, Tuple <bool, CMPSpeechError> >(string.Empty, authorizationResult));
}
CMPSpeechError genericError = null;
var inputNode = _speechAudioEngine.InputNode;
if (inputNode == null)
{
var audioEngineError = new NSError(new NSString(string.Empty), nint.Parse(SpeechToTextErrorEnum.
eNoInputNode.ToString()));
genericError = PrepareAudioEngineError(audioEngineError);
ResetSpeechToText();
return(new Tuple <string, Tuple <bool, CMPSpeechError> >(string.Empty, new Tuple <bool, CMPSpeechError>
(false, genericError)));
}
Tuple <string, Tuple <bool, CMPSpeechError> > recognitionResult = null;
await Task.Run(() =>
{
try
{
_speechRecognitionTask = _speechRecognizer.GetRecognitionTask(_speechRecognitionRequest,
(SFSpeechRecognitionResult result,
NSError speechError) =>
{
if (speechError != null)
{
_speechAudioEngine.Stop();
genericError = new CMPSpeechError(speechError);
recognitionResult = new Tuple <string, Tuple <bool, CMPSpeechError> >(string.Empty,
new Tuple <bool, CMPSpeechError>(false,
genericError));
ResetSpeechToText();
_speechSemaphore.Release();
return;
}
if (result.Final == true)
{
_speechAudioEngine.Stop();
inputNode.RemoveTapOnBus(0);
recognitionResult = new Tuple <string, Tuple <bool, CMPSpeechError> >(result.BestTranscription.
FormattedString,
new Tuple <bool,
CMPSpeechError>(true,
null));
ResetSpeechToText();
_speechSemaphore.Release();
return;
}
});
var audioFormat = inputNode.GetBusOutputFormat(0);
inputNode.InstallTapOnBus(0, 2048, audioFormat, (AVAudioPcmBuffer buffer, AVAudioTime when) =>
{
var state = _speechRecognitionTask.State;
_speechRecognitionRequest.Append(buffer);
});
_speechAudioEngine.Prepare();
NSError audioEngineError = null;
bool couldStart = _speechAudioEngine.StartAndReturnError(out audioEngineError);
if (couldStart == false)
{
genericError = PrepareAudioEngineError(audioEngineError);
recognitionResult = new Tuple <string, Tuple <bool, CMPSpeechError> >(string.Empty,
new Tuple <bool, CMPSpeechError>(false,
genericError));
ResetSpeechToText();
_speechSemaphore.Release();
return;
}
}
catch (Exception exception)
{
Diagonostics.Debug.WriteLine(exception.Message);
ResetSpeechToText();
_speechSemaphore.Release();
}
});
await _speechSemaphore.WaitAsync();
return(recognitionResult);
}
private void StartRecordingAndRecognizing()
{
try
{
_timer = NSTimer.CreateRepeatingScheduledTimer(5, delegate
{
DidFinishTalk();
});
// Cancel the previous task if it's running.
_recognitionTask?.Cancel();
_recognitionTask = null;
var audioSession = AVAudioSession.SharedInstance();
var nsError = audioSession.SetCategory(AVAudioSessionCategory.PlayAndRecord);
audioSession.SetMode(AVAudioSession.ModeDefault, out nsError);
nsError = audioSession.SetActive(true, AVAudioSessionSetActiveOptions.NotifyOthersOnDeactivation);
audioSession.OverrideOutputAudioPort(AVAudioSessionPortOverride.Speaker, out nsError);
// Configure request so that results are returned before audio recording is finished
_recognitionRequest = new SFSpeechAudioBufferRecognitionRequest();
var inputNode = _audioEngine.InputNode;
if (inputNode == null)
{
throw new InvalidProgramException("Audio engine has no input node");
}
var recordingFormat = inputNode.GetBusOutputFormat(0);
inputNode.InstallTapOnBus(0, 1024, recordingFormat, (buffer, when) =>
{
_recognitionRequest?.Append(buffer);
});
_audioEngine.Prepare();
_audioEngine.StartAndReturnError(out nsError);
// A recognition task represents a speech recognition session.
// We keep a reference to the task so that it can be cancelled.
_recognitionTask = _speechRecognizer.GetRecognitionTask(_recognitionRequest, (result, error) =>
{
if (result != null)
{
var eventArg = new TextReceivedEventArg
{
Text = result.BestTranscription.FormattedString
};
OnTextReceived(eventArg);
_timer.Invalidate();
_timer = null;
_timer = NSTimer.CreateRepeatingScheduledTimer(2, delegate
{
DidFinishTalk();
});
}
if (error == null)
{
return;
}
OnStoppedListening();
StopRecordingAndRecognition();
});
}
catch (Exception ex)
{
Console.WriteLine(ex.Message);
}
}
void StartRecording()
{
if (recognitionTask != null)
{
recognitionTask.Cancel();
recognitionTask = null;
}
var audioSession = AVAudioSession.SharedInstance();
try
{
NSError err;
audioSession.SetCategory(AVAudioSessionCategory.PlayAndRecord);
audioSession.SetMode(AVAudioSession.ModeMeasurement, out err);
audioSession.SetActive(true);
}
catch (Exception ex)
{
var s = ex.ToString();
}
recognitionRequest = new SFSpeechAudioBufferRecognitionRequest();
var inputNode = audioEngine.InputNode;
recognitionRequest.ShouldReportPartialResults = true;
recognitionTask = speechRecognizer.GetRecognitionTask(recognitionRequest, (arg1, arg2) =>
{
var isFinal = false;
if (arg1 != null)
{
var inputtedText = arg1.BestTranscription.FormattedString;
englishText.Text = inputtedText;
var fromLanguage = LanguageCodes.English;
var toLangague = AppDelegate.CurrentLanguage.LanguageCode;
translator.TranslateText(inputtedText, fromLanguage, toLangague).ContinueWith(async(arg) =>
{
var translated = await arg;
InvokeOnMainThread(() => translatedText.Text = translated);
});
isFinal = arg1.Final;
}
if (arg2 != null || isFinal)
{
audioEngine.Stop();
inputNode.RemoveTapOnBus(0);
recognitionRequest = null;
recognitionTask = null;
askQuestion.Enabled = true;
}
});
var recordingFormat = inputNode.GetBusOutputFormat(0);
inputNode.InstallTapOnBus(0, 1024, recordingFormat, (buffer, when) =>
{
recognitionRequest?.Append(buffer);
});
audioEngine.Prepare();
try
{
NSError err = new NSError();
audioEngine.StartAndReturnError(out err);
englishText.Text = "OK, here we go!";
translatedText.Text = "";
}
catch (Exception) { }
}
// 音声認識の本処理
private void startRecognitionSession()
{
// 音声認識のパラメータ設定と認識開始。ここのパラメータはおまじない。
audioEngine.InputNode.InstallTapOnBus(
bus: 0,
bufferSize: 1024,
format: audioEngine.InputNode.GetBusOutputFormat(0),
tapBlock: (buffer, when) => { recognitionRequest?.Append(buffer); }
);
audioEngine?.Prepare();
NSError error = null;
audioEngine?.StartAndReturnError(out error);
if (error != null)
{
Console.WriteLine(error);
return;
}
try
{
if (recognitionTask?.State == SFSpeechRecognitionTaskState.Running)
{
// 音声認識が実行中に音声認識開始処理が呼び出された場合、実行中だった音声認識を中断する。
recognitionTask.Cancel();
}
recognitionTask = speechRecognizer.GetRecognitionTask(recognitionRequest,
(SFSpeechRecognitionResult result, NSError err) =>
{
if (result == null)
{
// iOS Simulator等、端末が音声認識に対応していない場合はここに入る。
StopRecognizing();
return;
}
if (err != null)
{
Console.WriteLine(err);
StopRecognizing();
return;
}
if ((result.BestTranscription != null) && (result.BestTranscription.FormattedString != null))
{
// 音声を認識できた場合、認識結果を更新する。
RecognizedText = result.BestTranscription.FormattedString;
}
if (result.Final)
{
// 音声が認識されなくなって時間が経ったら音声認識を打ち切る。
StopRecognizing();
return;
}
}
);
}
catch (Exception ex)
{
Console.WriteLine(ex.Message);
}
}
void StartRecording()
{
// Cancel the previous task if it's running.
recognitionTask?.Cancel();
recognitionTask = null;
var audioSession = AVAudioSession.SharedInstance();
NSError err;
err = audioSession.SetCategory(AVAudioSessionCategory.PlayAndRecord, AVAudioSessionCategoryOptions.DefaultToSpeaker);
audioSession.SetMode(AVAudioSession.ModeMeasurement, out err);
err = audioSession.SetActive(true, AVAudioSessionSetActiveOptions.NotifyOthersOnDeactivation);
// Configure request so that results are returned before audio recording is finished
recognitionRequest = new SFSpeechAudioBufferRecognitionRequest {
ShouldReportPartialResults = true,
};
var inputNode = audioEngine.InputNode;
if (inputNode == null)
{
throw new InvalidProgramException("Audio engine has no input node");
}
// A recognition task represents a speech recognition session.
// We keep a reference to the task so that it can be cancelled.
recognitionTask = speechRecognizer.GetRecognitionTask(recognitionRequest, (result, error) => {
var isFinal = false;
if (result != null)
{
speechIdleTimer.Stop();
speechIdleTimer.Start();
textView.Text = result.BestTranscription.FormattedString;
isFinal = result.Final;
}
if (error != null || isFinal)
{
if (result != null)
{
var intent = nlp.GetMatchingIntent(result.BestTranscription.FormattedString);
string resultText;
if (intent != null)
{
textView.Text += "\nAction is " + intent.Action + ".";
resultText = "Action is " + intent.Action + ". ";
if (intent.Parameters != null)
{
intent.Parameters.ForEach(p => {
resultText += "Parameter " + p.Key + " with values" + string.Join(",", p.Value) + ". ";
textView.Text += "\nParameter " + p.Key + " with values " + string.Join(",", p.Value) + ". ";
});
}
}
else
{
resultText = "Sorry, I did not get that.";
}
var su = new AVSpeechUtterance(resultText)
{
Rate = AVSpeechUtterance.MaximumSpeechRate / 2,
Voice = AVSpeechSynthesisVoice.FromLanguage("en-US"),
PitchMultiplier = 1.0f,
Volume = 1
};
ss.SpeakUtterance(su);
}
audioEngine.Stop();
inputNode.RemoveTapOnBus(0);
recognitionRequest = null;
recognitionTask = null;
recordButton.Enabled = true;
//recordButton.SetTitle ("Start Recording", UIControlState.Normal);
recordButton.Hidden = false;
recordStatus.Hidden = true;
speechIdleTimer.Stop();
}
});
var recordingFormat = inputNode.GetBusOutputFormat(0);
inputNode.InstallTapOnBus(0, 1024, recordingFormat, (buffer, when) => {
recognitionRequest?.Append(buffer);
});
audioEngine.Prepare();
audioEngine.StartAndReturnError(out err);
}
public void StartRecording(string voice, bool longTimeout = false)
{
if (!SpeechEnabled)
{
return;
}
// Setup audio session
LastResult = "";
AudioEngine = new AVAudioEngine();
NSLocale voiceLocale = NSLocale.FromLocaleIdentifier(voice);
SpeechRecognizer = new SFSpeechRecognizer(voiceLocale);
LiveSpeechRequest = new SFSpeechAudioBufferRecognitionRequest();
NSError error;
var audioSession = AVAudioSession.SharedInstance();
audioSession.SetCategory(AVAudioSessionCategory.Record);
audioSession.SetMode(AVAudioSession.ModeMeasurement, out error);
if (error != null)
{
OnSpeechError?.Invoke("Audio session error: " + error.ToString());
return;
}
audioSession.SetActive(true, AVAudioSessionSetActiveOptions.NotifyOthersOnDeactivation);
LiveSpeechRequest.ShouldReportPartialResults = true;
var node = AudioEngine.InputNode;
if (node == null)
{
OnSpeechError?.Invoke("Couldn't initialize Speech Input");
return;
}
RecognitionTask = SpeechRecognizer.GetRecognitionTask(LiveSpeechRequest, (SFSpeechRecognitionResult result, NSError err) => {
if (IsCancelled)
{
node.RemoveTapOnBus(0);
return;
}
if (err != null)
{
OnSpeechError?.Invoke(err.ToString());
}
else if (result != null)
{
LastResult = result.BestTranscription.FormattedString;
Console.WriteLine("You said: \"{0}\". Final: {1}",
LastResult, result.Final);
m_lastSpeech = DateTime.Now;
if (result.Final)// || !IsRecording) {
{
OnSpeechOK?.Invoke(LastResult);
}
}
if ((result != null && result.Final) || err != null || !IsRecording)
{
IsRecording = false;
//node.RemoveTapOnBus(0);
AudioEngine.Stop();
m_speechTimer.Close();
}
});
var recordingFormat = node.GetBusOutputFormat(0);
node.InstallTapOnBus(0, 1024, recordingFormat, (AVAudioPcmBuffer buffer, AVAudioTime when) => {
//Console.WriteLine("--> {0}: {1} {2}.{3}", buffer.FrameLength, when.HostTime);
// Append buffer to recognition request
LiveSpeechRequest.Append(buffer);
});
// Start recording
AudioEngine.Prepare();
AudioEngine.StartAndReturnError(out error);
if (error != null)
{
OnSpeechError?.Invoke("Speech init error: " + error.ToString());
IsRecording = false;
return;
}
IsRecording = true;
IsCancelled = false;
LastResult = "";
m_lastSpeech = DateTime.MaxValue;
m_startSpeech = DateTime.Now;
m_timeout = longTimeout ? m_phraseTimeout : m_wordTimeout;
m_speechTimer = new System.Timers.Timer(250);
m_speechTimer.AutoReset = true;
m_speechTimer.Elapsed += (sender, e) => {
CheckRecording();
};
m_speechTimer.Start();
}
