/// <summary>
/// Cleans up the microphone resources and the stream and unregisters from MediaCapture events
/// </summary>
/// <returns>true if successful</returns>
public async System.Threading.Tasks.Task <bool> CleanupRecording()
{
if (isRecordingInitialized)
{
// If a recording is in progress during cleanup, stop it to save the recording
if (isRecording)
{
await StopRecording();
}
isRecordingInitialized = false;
}
if (mediaCapture != null)
{
mediaCapture.RecordLimitationExceeded -= mediaCapture_RecordLimitationExceeded;
mediaCapture.Failed -= mediaCapture_Failed;
mediaCapture.Dispose();
mediaCapture = null;
}
if (STTStream != null)
{
STTStream.BufferReady -= STTStream_BufferReady;
STTStream.AudioLevel -= STTStream_AudioLevel;
STTStream.Dispose();
STTStream = null;
}
return(true);
}
public static SpeechToTextMainStream Create(ulong maxSizeInBytes = 0, uint thresholdDuration = 0, uint thresholdLevel = 0)
{
SpeechToTextMainStream stts = null;
try
{
stts = new SpeechToTextMainStream(maxSizeInBytes, thresholdDuration, thresholdLevel);
}
catch (Exception ex)
{
System.Diagnostics.Debug.WriteLine("Exception while creating SpeechToTextMainStream: " + ex.Message);
}
return(stts);
}
/// <summary>
/// StartRecording method
/// Start to record audio using the microphone.
/// The audio stream in stored in memory with no limit of size.
/// </summary>
/// <param name="MaxStreamSizeInBytes">
/// This parameter defines the max size of the buffer in memory. When the size of the buffer is over this limit, the
/// client create another stream and remove the previouw stream.
/// By default the value is 0, in that case the audio stream in stored in memory with no limit of size.
/// </param>
/// <param name="ThresholdDuration">
/// The duration in milliseconds for the calculation of the average audio level.
/// With this parameter you define the period during which the average level is measured.
/// If the value is 0, no buffer will be sent to Cognitive Services.
/// </param>
/// <param name="ThresholdLevel">
/// The minimum audio level average necessary to trigger the recording,
/// it's a value between 0 and 65535. You can tune this value after several microphone tests.
/// If the value is 0, no buffer will be sent to Cognitive Services.
/// </param>
/// <return>return true if successful.
/// </return>
public async System.Threading.Tasks.Task <bool> StartContinuousRecording(ulong MaxStreamSizeInBytes, UInt16 ThresholdDuration, UInt16 ThresholdLevel)
{
thresholdDuration = ThresholdDuration;
thresholdLevel = ThresholdLevel;
bool bResult = false;
maxStreamSizeInBytes = MaxStreamSizeInBytes;
if (isRecordingInitialized != true)
{
await InitializeRecording();
}
if (STTStream != null)
{
STTStream.BufferReady -= STTStream_BufferReady;
STTStream.AudioLevel -= STTStream_AudioLevel;
STTStream.Dispose();
STTStream = null;
}
STTStream = SpeechToTextMainStream.Create(maxStreamSizeInBytes, thresholdDuration, thresholdLevel);
STTStream.AudioLevel += STTStream_AudioLevel;
STTStream.BufferReady += STTStream_BufferReady;
if ((STTStream != null) && (isRecordingInitialized == true))
{
try
{
Windows.Media.MediaProperties.MediaEncodingProfile MEP = Windows.Media.MediaProperties.MediaEncodingProfile.CreateWav(Windows.Media.MediaProperties.AudioEncodingQuality.Auto);
if (MEP != null)
{
if (MEP.Audio != null)
{
uint framerate = 16000;
uint bitsPerSample = 16;
uint numChannels = 1;
uint bytespersecond = 32000;
MEP.Audio.Properties[WAVAttributes.MF_MT_AUDIO_SAMPLES_PER_SECOND] = framerate;
MEP.Audio.Properties[WAVAttributes.MF_MT_AUDIO_NUM_CHANNELS] = numChannels;
MEP.Audio.Properties[WAVAttributes.MF_MT_AUDIO_BITS_PER_SAMPLE] = bitsPerSample;
MEP.Audio.Properties[WAVAttributes.MF_MT_AUDIO_AVG_BYTES_PER_SECOND] = bytespersecond;
foreach (var Property in MEP.Audio.Properties)
{
System.Diagnostics.Debug.WriteLine("Property: " + Property.Key.ToString());
System.Diagnostics.Debug.WriteLine("Value: " + Property.Value.ToString());
if (Property.Key == new Guid("5faeeae7-0290-4c31-9e8a-c534f68d9dba"))
{
framerate = (uint)Property.Value;
}
if (Property.Key == new Guid("f2deb57f-40fa-4764-aa33-ed4f2d1ff669"))
{
bitsPerSample = (uint)Property.Value;
}
if (Property.Key == new Guid("37e48bf5-645e-4c5b-89de-ada9e29b696a"))
{
numChannels = (uint)Property.Value;
}
}
}
if (MEP.Container != null)
{
foreach (var Property in MEP.Container.Properties)
{
System.Diagnostics.Debug.WriteLine("Property: " + Property.Key.ToString());
System.Diagnostics.Debug.WriteLine("Value: " + Property.Value.ToString());
}
}
}
await mediaCapture.StartRecordToStreamAsync(MEP, STTStream);
bResult = true;
isRecording = true;
System.Diagnostics.Debug.WriteLine("Recording in audio stream...");
}
catch (Exception e)
{
System.Diagnostics.Debug.WriteLine("Exception while recording in audio stream:" + e.Message);
}
}
return(bResult);
}
/// <summary>
/// SendStorageFile method
/// </summary>
/// <param name="wavFile">StorageFile associated with the audio file which
/// will be sent to the SpeechToText Services.
/// </param>
/// <param name="locale">language associated with the current buffer/recording.
/// for instance en-US, fr-FR, pt-BR, ...
/// </param>
/// <return>The result of the SpeechToText REST API.
/// </return>
public async System.Threading.Tasks.Task <SpeechToTextResponse> SendStorageFile(Windows.Storage.StorageFile wavFile, string locale)
{
SpeechToTextResponse r = null;
int loop = 1;
while (loop-- > 0)
{
try
{
string os = "Windows" + SpeechToText.SystemInformation.SystemVersion;
string deviceid = "b2c95ede-97eb-4c88-81e4-80f32d6aee54";
string speechUrl = SpeechUrl + "?scenarios=ulm&appid=D4D52672-91D7-4C74-8AD8-42B1D98141A5&version=3.0&device.os=" + os + "&locale=" + locale + "&format=json&requestid=" + Guid.NewGuid().ToString() + "&instanceid=" + deviceid + "&result.profanitymarkup=1&maxnbest=3";
Windows.Web.Http.HttpClient hc = new Windows.Web.Http.HttpClient();
hc.DefaultRequestHeaders.TryAppendWithoutValidation("Authorization", Token);
hc.DefaultRequestHeaders.TryAppendWithoutValidation("ContentType", "audio/wav; codec=\"audio/pcm\"; samplerate=16000");
Windows.Web.Http.HttpResponseMessage hrm = null;
Windows.Storage.StorageFile file = wavFile;
if (file != null)
{
using (var fileStream = await file.OpenAsync(Windows.Storage.FileAccessMode.Read))
{
if (STTStream != null)
{
STTStream.AudioLevel -= STTStream_AudioLevel;
STTStream.Dispose();
STTStream = null;
}
STTStream = SpeechToTextMainStream.Create();
if (STTStream != null)
{
byte[] byteArray = new byte[fileStream.Size];
fileStream.ReadAsync(byteArray.AsBuffer(), (uint)fileStream.Size, Windows.Storage.Streams.InputStreamOptions.Partial).AsTask().Wait();
STTStream.WriteAsync(byteArray.AsBuffer()).AsTask().Wait();
Windows.Web.Http.HttpStreamContent content = new Windows.Web.Http.HttpStreamContent(STTStream.AsStream().AsInputStream());
content.Headers.ContentLength = STTStream.GetLength();
System.Diagnostics.Debug.WriteLine("REST API Post Content Length: " + content.Headers.ContentLength.ToString() + " bytes");
System.Threading.CancellationTokenSource cts = new System.Threading.CancellationTokenSource();
IProgress <Windows.Web.Http.HttpProgress> progress = new Progress <Windows.Web.Http.HttpProgress>(ProgressHandler);
hrm = await hc.PostAsync(new Uri(speechUrl), content).AsTask(cts.Token, progress);
}
}
}
if (hrm != null)
{
switch (hrm.StatusCode)
{
case Windows.Web.Http.HttpStatusCode.Ok:
var b = await hrm.Content.ReadAsBufferAsync();
string result = System.Text.UTF8Encoding.UTF8.GetString(b.ToArray());
if (!string.IsNullOrEmpty(result))
{
r = new SpeechToTextResponse(result);
}
break;
case Windows.Web.Http.HttpStatusCode.Forbidden:
string token = await RenewToken();
if (string.IsNullOrEmpty(token))
{
loop++;
}
break;
default:
int code = (int)hrm.StatusCode;
string HttpError = "Http Response Error: " + code.ToString() + " reason: " + hrm.ReasonPhrase.ToString();
System.Diagnostics.Debug.WriteLine(HttpError);
r = new SpeechToTextResponse(string.Empty, HttpError);
break;
}
}
}
catch (Exception ex)
{
System.Diagnostics.Debug.WriteLine("Exception while sending the audio file:" + ex.Message);
}
}
return(r);
}
