/// <summary>
/// Extract the speech-to-text result info from the next response JSON in the queue.
/// </summary>
void ProcessNextResponseJSON()
{
// Create a JSON object from the next string in the queue and process the speech-to-text result.
var responseJSON = new JSONObject(m_ResponseJSONsQueue.Dequeue(), int.MaxValue);
SmartLogger.Log(DebugFlags.GoogleStreamingSpeechToText, responseJSON.ToString());
string errorText = GoogleSpeechToTextResponseJSONParser.GetErrorFromResponseJSON(responseJSON);
if (errorText != null)
{
if (m_OnError != null)
{
m_OnError(errorText);
}
}
JSONObject resultsJSON = responseJSON.GetField(Constants.GoogleResponseJSONResultsFieldKey);
if (resultsJSON != null && resultsJSON.Count > 0)
{
JSONObject resultJSON = resultsJSON[0];
SpeechToTextResult textResult = GoogleSpeechToTextResponseJSONParser.GetTextResultFromResultJSON(resultJSON);
bool isFinal = false;
resultJSON.GetField(out isFinal, Constants.GoogleResponseJSONResultIsFinalFieldKey, isFinal);
textResult.IsFinal = isFinal;
SmartLogger.Log(DebugFlags.GoogleStreamingSpeechToText, "processing result - isFinal = " + isFinal);
if (m_OnTextResult != null)
{
m_OnTextResult(textResult);
}
m_LastResult = textResult;
}
}
/// <summary>
/// Returns a speech-to-text result object based on information in the result JSON.
/// </summary>
/// <param name="resultJSON">Google speech-to-text result JSON object</param>
/// <returns>Speech-to-text result object</returns>
public static SpeechToTextResult GetTextResultFromResultJSON(JSONObject resultJSON)
{
SpeechToTextResult textResult = null;
JSONObject alternatives = resultJSON.GetField(Constants.GoogleResponseJSONAlternativesFieldKey);
if (alternatives != null)
{
textResult = new SpeechToTextResult();
textResult.TextAlternatives = new TextAlternative[alternatives.Count];
for (int i = 0; i < textResult.TextAlternatives.Length; ++i)
{
var alternative = new GoogleTextAlternative();
string text = "";
float confidence = 0;
alternatives[i].GetField(out text, Constants.GoogleResponseJSONAlternativeTranscriptFieldKey, text);
alternatives[i].GetField(out confidence, Constants.GoogleResponseJSONAlternativeConfidenceFieldKey, confidence);
alternative.Text = text;
alternative.Confidence = confidence;
textResult.TextAlternatives[i] = alternative;
}
}
if (textResult == null || textResult.TextAlternatives == null || textResult.TextAlternatives.Length == 0)
{
textResult = GetDefaultGoogleSpeechToTextResult();
}
return(textResult);
}
/// <summary>
/// Function that is called when an interim text result is received.
/// </summary>
/// <param name="text">The interim text result</param>
void OnDictationHypothesis(string text)
{
SmartLogger.LogFormat(DebugFlags.WindowsSpeechToText, "Dictation hypothesis: {0}", text);
m_LastResult = new SpeechToTextResult(text, false);
if (m_OnTextResult != null)
{
m_OnTextResult(m_LastResult);
}
}
/// <summary>
/// Function that is called when a final text result is received.
/// </summary>
/// <param name="text">The final text result</param>
/// <param name="confidence">Confidence level of the text result</param>
void OnDictationResult(string text, ConfidenceLevel confidence)
{
SmartLogger.LogFormat(DebugFlags.WindowsSpeechToText, "Dictation result: {0}", text);
m_LastResult = new SpeechToTextResult(text, true);
if (m_OnTextResult != null)
{
m_OnTextResult(m_LastResult);
}
}
/// <summary>
/// Starts recording audio if the service is not already recording.
/// </summary>
/// <returns>Whether the service successfully started recording</returns>
public override bool StartRecording()
{
if (base.StartRecording())
{
m_LastResult = new SpeechToTextResult("", false);
m_DictationRecognizer.Start();
return(true);
}
return(false);
}
/// <summary>
/// Starts recording audio if the service is not already recording.
/// </summary>
/// <returns>Whether the service successfully started recording</returns>
public override bool StartRecording()
{
if (base.StartRecording())
{
m_LastResult = new SpeechToTextResult("", false);
StartCoroutine(RecordAudio());
StartCoroutine(StreamAudioAndListenForResponses());
return(true);
}
return(false);
}
/// <summary>
/// Returns a speech-to-text result with a single empty Google text alternative.
/// </summary>
/// <returns>Default Google speech-to-text result object</returns>
public static SpeechToTextResult GetDefaultGoogleSpeechToTextResult()
{
var textResult = new SpeechToTextResult();
textResult.TextAlternatives = new TextAlternative[1];
var alternative = new GoogleTextAlternative();
alternative.Text = "";
alternative.Confidence = 0;
textResult.TextAlternatives[0] = alternative;
return(textResult);
}
/// <summary>
/// Populates and returns a SpeechToTextResult object from a given Watson SpeechResult object.
/// </summary>
/// <param name="watsonResult">Watson SpeechResult object</param>
/// <returns>A SpeechToTextResult object</returns>
public SpeechToTextResult CreateSpeechToTextResult(SpeechResult watsonResult)
{
var textResult = new SpeechToTextResult();
textResult.IsFinal = watsonResult.Final;
textResult.TextAlternatives = new TextAlternative[watsonResult.Alternatives.Length];
for (int i = 0; i < textResult.TextAlternatives.Length; ++i)
{
SpeechAlt watsonAlternative = watsonResult.Alternatives[i];
var alternative = new WatsonTextAlternative();
alternative.Text = watsonAlternative.Transcript;
alternative.Confidence = (float)watsonAlternative.Confidence;
alternative.TimeStamps = watsonAlternative.Timestamps;
alternative.WordConfidenceValues = watsonAlternative.WordConfidence;
textResult.TextAlternatives[i] = alternative;
}
return(textResult);
}
