protected async Task <string> GetAmazonSynthSpeech(string text, TTSVoice voicePreference, TTSPitch pitchPreference, string filename = null)
{
AmazonSynthesizeSpeechRequest synthesisRequest = voicePreference.GetAmazonTTSSpeechRequest();
synthesisRequest.TextType = TextType.Ssml;
synthesisRequest.Text = PrepareAmazonSSML(text, pitchPreference);
// Perform the Text-to-Speech request, passing the text input
// with the selected voice parameters and audio file type
AmazonSynthesizeSpeechResponse synthesisResponse = await amazonClient.SynthesizeSpeechAsync(synthesisRequest);
// Write the binary AudioContent of the response to file.
string filepath;
if (string.IsNullOrWhiteSpace(filename))
{
filepath = Path.Combine(TTSFilesPath, $"{Guid.NewGuid()}.mp3");
}
else
{
filepath = Path.Combine(TTSFilesPath, $"{filename}.mp3");
}
using (Stream file = new FileStream(filepath, FileMode.Create))
{
await synthesisResponse.AudioStream.CopyToAsync(file);
await file.FlushAsync();
file.Close();
}
return(filepath);
}
public override void DownloadItem(PhraseItem item, string folder)
{
String SSMLText = String.Format(@"
<speak>
{0}
</speak >
", item.Phrase.Replace("&", "&"));
try
{
//new Task(() =>
//{
Amazon.Polly.Model.SynthesizeSpeechRequest ssr = new Amazon.Polly.Model.SynthesizeSpeechRequest();
ssr.TextType = TextType.Ssml;
ssr.Text = SSMLText;
ssr.VoiceId = polly.DescribeVoices(new Amazon.Polly.Model.DescribeVoicesRequest()).Voices.Find(n => n.Name == SelectedVoice.Name).Id;
ssr.OutputFormat = OutputFormat.Mp3;
using (FileStream output = File.Create(String.Format("{0}\\mp3\\{1}\\{2}.mp3", folder, item.Folder, item.FileName)))
{
polly.SynthesizeSpeech(ssr).AudioStream.CopyTo(output);
}
ConvertToWav(item, folder, false, new String[] { Name, SelectedVoice.Name, SelectedDiscreteSpeed, SelectedDiscreteVolume });
//}).Start();
}
catch (Exception Ex)
{
Logger.Log(Ex.ToString());
item.DownloadComplete = false;
}
}
public override void Play(PhraseItem item)
{
if (!disabled)
{
Amazon.Polly.Model.SynthesizeSpeechRequest ssr = new Amazon.Polly.Model.SynthesizeSpeechRequest();
ssr.Text = item.Phrase;
ssr.VoiceId = polly.DescribeVoices(new Amazon.Polly.Model.DescribeVoicesRequest()).Voices.Find(n => n.Name == SelectedVoice.Name).Id;
ssr.OutputFormat = OutputFormat.Mp3;
byte[] mp3;
using (MemoryStream output = new MemoryStream())
{
polly.SynthesizeSpeech(ssr).AudioStream.CopyTo(output);
mp3 = output.ToArray();
}
MainWindow.PlayAudioStream(mp3);
}
}
static void Main(string[] args)
{
if (args.Count() == 0)
{
Console.WriteLine("Basic text to speach converter");
Console.WriteLine("Expecting text in the command line parameter");
Environment.Exit(0);
}
Console.WriteLine("Creating Amazon Polly client");
Amazon.Polly.AmazonPollyClient cl = new Amazon.Polly.AmazonPollyClient(RegionEndpoint.USWest2);
Amazon.Polly.Model.SynthesizeSpeechRequest req = new Amazon.Polly.Model.SynthesizeSpeechRequest();
req.Text = String.Join(" ", args);
req.VoiceId = Amazon.Polly.VoiceId.Salli;
req.OutputFormat = Amazon.Polly.OutputFormat.Mp3;
req.SampleRate = "8000";
req.TextType = Amazon.Polly.TextType.Text;
Console.WriteLine("Sending Amazon Polly request: " + req.Text);
Amazon.Polly.Model.SynthesizeSpeechResponse resp = cl.SynthesizeSpeech(req);
MemoryStream local_stream = new MemoryStream();
resp.AudioStream.CopyTo(local_stream);
local_stream.Position = 0;
Console.WriteLine("Got mp3 stream, lenght: " + local_stream.Length.ToString());
NAudio.Wave.Mp3FileReader reader = new NAudio.Wave.Mp3FileReader(local_stream);
NAudio.Wave.WaveStream wave_stream = NAudio.Wave.WaveFormatConversionStream.CreatePcmStream(reader);
NAudio.Wave.BlockAlignReductionStream ba_stream = new NAudio.Wave.BlockAlignReductionStream(wave_stream);
NAudio.Wave.WaveOut wout = new NAudio.Wave.WaveOut();
Console.Write("Playing stream...");
wout.Init(ba_stream);
wout.Play();
while (wout.PlaybackState == NAudio.Wave.PlaybackState.Playing)
{
Thread.Sleep(100);
}
Console.WriteLine("..Done");
}
private IEnumerator speak(Model.Wrapper wrapper, bool isNative)
{
//Debug.Log("Speak: " + wrapper);
connectToAmazon();
if (client == null)
{
Debug.LogWarning("'client' is null! Did you add the correct Cognito credentials?");
}
else
{
if (wrapper == null)
{
Debug.LogWarning("'wrapper' is null!");
}
else
{
if (string.IsNullOrEmpty(wrapper.Text))
{
Debug.LogWarning("'wrapper.Text' is null or empty!");
}
else
{
if (!Util.Helper.isInternetAvailable)
{
string errorMessage = "Internet is not available - can't use AWS Polly right now!";
Debug.LogError(errorMessage);
onErrorInfo(wrapper, errorMessage);
}
else
{
yield return(null); //return to the main process (uid)
string voiceCulture = getVoiceCulture(wrapper);
string voiceName = getVoiceName(wrapper);
System.Text.StringBuilder sbXML = new System.Text.StringBuilder();
//SSML
sbXML.Append("<?xml version=\"1.0\" encoding=\"UTF-8\" ?>");
sbXML.Append("<speak version=\"1.0\" xmlns=\"http://www.w3.org/2001/10/synthesis\" xml:lang=\"");
sbXML.Append(voiceCulture);
sbXML.Append("\"");
//sbXML.Append (" xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\"");
//sbXML.Append (" xsi:schemaLocation=\"http://www.w3.org/2001/10/synthesis http://www.w3.org/TR/speech-synthesis/synthesis.xsd\"");
sbXML.Append(">");
sbXML.Append(prepareProsody(wrapper));
sbXML.Append("</speak>");
Amazon.Polly.Model.SynthesizeSpeechRequest synthesizeSpeechPresignRequest = new Amazon.Polly.Model.SynthesizeSpeechRequest();
synthesizeSpeechPresignRequest.Text = sbXML.ToString();
synthesizeSpeechPresignRequest.TextType = Amazon.Polly.TextType.Ssml;
synthesizeSpeechPresignRequest.VoiceId = voiceName;
synthesizeSpeechPresignRequest.OutputFormat = Amazon.Polly.OutputFormat.Ogg_vorbis;
silence = false;
if (!isNative)
{
onSpeakAudioGenerationStart(wrapper);
}
string outputFile = getOutputFile(wrapper.Uid);
bool success = false;
bool requestFinished = false;
client.SynthesizeSpeechAsync(synthesizeSpeechPresignRequest, (responseObject) =>
{
if (responseObject.Exception == null)
{
using (System.IO.FileStream fileStream = System.IO.File.Create(outputFile))
{
copyTo(responseObject.Response.AudioStream, fileStream);
}
success = true;
}
else
{
string errorMessage = "Could not speak the text: " + wrapper + System.Environment.NewLine + "Error: " + responseObject.Exception;
Debug.LogError(errorMessage);
onErrorInfo(wrapper, errorMessage);
}
requestFinished = true;
});
do
{
yield return(null);
} while (!requestFinished);
if (success)
{
yield return(playAudioFile(wrapper, Util.Constants.PREFIX_FILE + outputFile, outputFile, AudioFileType, isNative));
}
}
}
}
}
}
public static Amazon.Polly.Model.SynthesizeSpeechRequest GetAmazonTTSSpeechRequest(this TTSVoice voice)
{
Amazon.Polly.Model.SynthesizeSpeechRequest synthesisRequest = new Amazon.Polly.Model.SynthesizeSpeechRequest
{
OutputFormat = OutputFormat.Mp3,
Engine = Engine.Standard,
LexiconNames = awsLexicons
};
switch (voice)
{
case TTSVoice.en_AU_Standard_A:
case TTSVoice.en_AU_Standard_B:
case TTSVoice.en_AU_Standard_C:
case TTSVoice.en_AU_Standard_D:
case TTSVoice.en_IN_Standard_A:
case TTSVoice.en_IN_Standard_B:
case TTSVoice.en_IN_Standard_C:
case TTSVoice.en_IN_Standard_D:
case TTSVoice.en_GB_Standard_A:
case TTSVoice.en_GB_Standard_B:
case TTSVoice.en_GB_Standard_C:
case TTSVoice.en_GB_Standard_D:
case TTSVoice.en_GB_Standard_F:
case TTSVoice.en_US_Standard_B:
case TTSVoice.en_US_Standard_C:
case TTSVoice.en_US_Standard_D:
case TTSVoice.en_US_Standard_E:
case TTSVoice.en_US_Standard_G:
case TTSVoice.en_US_Standard_H:
case TTSVoice.en_US_Standard_I:
case TTSVoice.en_US_Standard_J:
BGC.Debug.LogError($"Tried to get Amazon VoiceId from Google TTS Voice {voice}");
goto case TTSVoice.en_GB_Brian;
case TTSVoice.en_AU_Nicole:
synthesisRequest.VoiceId = VoiceId.Nicole;
break;
//Olivia is unsupported
case TTSVoice.en_AU_Olivia:
synthesisRequest.VoiceId = VoiceId.Nicole;
break;
case TTSVoice.en_AU_Russell:
synthesisRequest.VoiceId = VoiceId.Russell;
break;
case TTSVoice.en_GB_Amy:
synthesisRequest.VoiceId = VoiceId.Amy;
break;
case TTSVoice.en_GB_Emma:
synthesisRequest.VoiceId = VoiceId.Emma;
break;
case TTSVoice.en_GB_Brian:
synthesisRequest.VoiceId = VoiceId.Brian;
break;
case TTSVoice.en_IN_Aditi:
synthesisRequest.VoiceId = VoiceId.Aditi;
synthesisRequest.LanguageCode = LanguageCode.EnIN;
break;
case TTSVoice.en_IN_Raveena:
synthesisRequest.VoiceId = VoiceId.Raveena;
synthesisRequest.LanguageCode = LanguageCode.EnIN;
break;
case TTSVoice.en_US_Ivy:
synthesisRequest.VoiceId = VoiceId.Ivy;
break;
case TTSVoice.en_US_Joanna:
synthesisRequest.VoiceId = VoiceId.Joanna;
break;
case TTSVoice.en_US_Kendra:
synthesisRequest.VoiceId = VoiceId.Kendra;
break;
case TTSVoice.en_US_Kimberly:
synthesisRequest.VoiceId = VoiceId.Kimberly;
break;
case TTSVoice.en_US_Salli:
synthesisRequest.VoiceId = VoiceId.Salli;
break;
case TTSVoice.en_US_Joey:
synthesisRequest.VoiceId = VoiceId.Joey;
break;
case TTSVoice.en_US_Justin:
synthesisRequest.VoiceId = VoiceId.Justin;
break;
case TTSVoice.en_US_Matthew:
synthesisRequest.VoiceId = VoiceId.Matthew;
break;
case TTSVoice.en_GB_WLS_Geraint:
synthesisRequest.VoiceId = VoiceId.Geraint;
break;
case TTSVoice.Unassigned: goto case TTSVoice.en_US_Joanna;
default:
BGC.Debug.LogError($"TTS Voice not supported {voice}");
goto case TTSVoice.Unassigned;
}
return(synthesisRequest);
}
