public PortAudioPlaybackHandle(AudioStream baseStream, PortAudioDevice dev)
{
BaseStream = new BufferedAudioStream(baseStream, baseStream.Format.SampleRate * baseStream.Format.BytesPerFrame);
PortAudioSampleFormat.PortAudioNumberFormat numberFormat;
switch (baseStream.Format.DataFormat.NumberFormat)
{
case NumberFormat.FloatingPoint:
numberFormat = PortAudioSampleFormat.PortAudioNumberFormat.FloatingPoint;
break;
case NumberFormat.Signed:
numberFormat = PortAudioSampleFormat.PortAudioNumberFormat.Signed;
break;
case NumberFormat.Unsigned:
numberFormat = PortAudioSampleFormat.PortAudioNumberFormat.Unsigned;
break;
default: throw new ArgumentException();
}
_pump = new PortAudioDevicePump(dev, baseStream.Format.Channels,
new PortAudioSampleFormat(numberFormat, baseStream.Format.DataFormat.BytesPerSample),
dev.DefaultHighOutputLatency, baseStream.Format.SampleRate, DataCallback);
_pump.StreamFinished += PumpOnStreamFinished;
}
public async Task PlayAudioStream(Stream audioStream, CancellationToken token = default)
{
try
{
Playing?.Invoke(this, true);
await semaphore.WaitAsync(token);
using var pump = new PortAudioDevicePump(OutputDevice, 2, outputFormat, OutputDevice.DefaultLowOutputLatency, 48000, ReadCallback);
using var handle = new ManualResetEventSlim(false);
pump.StreamFinished += FinishedHandler;
pump.Start();
handle.Wait();
pump.StreamFinished -= FinishedHandler;
int ReadCallback(byte[] buffer, int offset, int count) => audioStream.Read(buffer, offset, count);
void FinishedHandler(object sender, EventArgs eventArgs) => handle.Set();
}
catch (OperationCanceledException)
{
}
finally
{
if (semaphore.CurrentCount == 0)
{
semaphore.Release();
}
Playing?.Invoke(this, false);
}
}
/// <summary>
/// Starts the media capturing/source devices.
/// </summary>
public override async Task Start()
{
if (!_isStarted)
{
await base.Start();
var apiType = PortAudioHostApiType.DirectSound;
if (Environment.OSVersion.Platform == PlatformID.Unix)
{
apiType = PortAudioHostApiType.Alsa;
}
_portAudioOutputDevice = PortAudioHostApi.SupportedHostApis.Where(x => x.HostApiType == apiType).First().DefaultOutputDevice;
_outputDevicePump = new PortAudioDevicePump(_portAudioOutputDevice, AUDIO_CHANNEL_COUNT,
new PortAudioSampleFormat(PortAudioSampleFormat.PortAudioNumberFormat.Signed, AUDIO_BYTES_PER_SAMPLE),
TimeSpan.FromMilliseconds(SAMPLING_PERIOD_MILLISECONDS), AUDIO_SAMPLING_RATE, ReadAudioDataCalback);
_portAudioInputDevice = PortAudioHostApi.SupportedHostApis.Where(x => x.HostApiType == apiType).First().DefaultInputDevice;
_inputDevicePump = new PortAudioDevicePump(_portAudioInputDevice, AUDIO_CHANNEL_COUNT,
new PortAudioSampleFormat(PortAudioSampleFormat.PortAudioNumberFormat.Signed, AUDIO_BYTES_PER_SAMPLE),
TimeSpan.FromMilliseconds(SAMPLING_PERIOD_MILLISECONDS), AUDIO_SAMPLING_RATE, WriteDataCallback);
_outputDevicePump.Start();
_inputDevicePump.Start();
}
}
#pragma warning restore CS0067
/// <summary>
/// Creates a new basic RTP session that captures and renders audio to/from the default system devices.
/// </summary>
public PortAudioEndPoint(IAudioEncoder audioEncoder)
{
_audioEncoder = audioEncoder;
_audioFormatManager = new MediaFormatManager <AudioFormat>(_audioEncoder.SupportedFormats);
var apiType = PortAudioHostApiType.DirectSound;
if (Environment.OSVersion.Platform == PlatformID.Unix)
{
apiType = PortAudioHostApiType.Alsa;
}
_portAudioOutputDevice = PortAudioHostApi.SupportedHostApis.Where(x => x.HostApiType == apiType).First().DefaultOutputDevice;
_outputDevicePump = new PortAudioDevicePump(_portAudioOutputDevice, AUDIO_CHANNEL_COUNT,
new PortAudioSampleFormat(PortAudioSampleFormat.PortAudioNumberFormat.Signed, AUDIO_BYTES_PER_SAMPLE),
TimeSpan.FromMilliseconds(SAMPLING_PERIOD_MILLISECONDS), AUDIO_SAMPLING_RATE, ReadAudioDataCalback);
_portAudioInputDevice = PortAudioHostApi.SupportedHostApis.Where(x => x.HostApiType == apiType).First().DefaultInputDevice;
_inputDevicePump = new PortAudioDevicePump(_portAudioInputDevice, AUDIO_CHANNEL_COUNT,
new PortAudioSampleFormat(PortAudioSampleFormat.PortAudioNumberFormat.Signed, AUDIO_BYTES_PER_SAMPLE),
TimeSpan.FromMilliseconds(SAMPLING_PERIOD_MILLISECONDS), AUDIO_SAMPLING_RATE, WriteDataCallback);
}
async Task Run()
{
using (var cts = new CancellationTokenSource())
{
Console.InputEncoding = Encoding.Unicode;
Console.OutputEncoding = Encoding.Unicode;
Console.CancelKeyPress += (s, e) =>
{
e.Cancel = true;
cts.Cancel();
};
try
{
using var input = new StreamReader(Console.OpenStandardInput(), Encoding.Unicode);
using var api = PickBestApi(input, PortAudioHostApi.SupportedHostApis, cts.Token);
using var device = GetOutputDevice(input, api.Devices, cts.Token);
Console.Write("Saisir le code de langue voulu (défaut : fr-FR): ");
if (!ReadLine(input, cts.Token, out string language))
{
return;
}
if (string.IsNullOrWhiteSpace(language))
{
language = "fr-FR";
}
Console.WriteLine("Choisir le type de voix (femme/homme):");
Console.WriteLine("[0] - Non spécifié");
Console.WriteLine("[1] - Homme");
Console.WriteLine("[2] - Femme");
Console.WriteLine("[2] - Neutre");
if (!ReadLine(input, cts.Token, out string voiceGenderStr) || !int.TryParse(voiceGenderStr, out int voiceGenderInt))
{
return;
}
var voiceGender = (SsmlVoiceGender)voiceGenderInt;
Console.WriteLine($"Prêt à parler sur {device.Name}");
while (!cts.IsCancellationRequested)
{
Console.Write("->");
// Handle user input
if (!ReadLine(input, cts.Token, out string line) || line is null)
{
return;
}
Console.WriteLine("Récupération du son pour le texte : " + line);
Console.WriteLine("Son récupéré, lecture...");
using var audioStream = await TextToAudioStreamAsync(line, language, voiceGender, cts.Token);
using var pump = new PortAudioDevicePump(
device,
2,
new PortAudioSampleFormat(PortAudioSampleFormat.PortAudioNumberFormat.Signed, 2),
device.DefaultLowOutputLatency,
48000,
(buffer, offset, count) => audioStream.Read(buffer, offset, count));
using var handle = new ManualResetEventSlim(false);
pump.StreamFinished += FinishedHandler;
pump.Start();
handle.Wait();
pump.StreamFinished -= FinishedHandler;
void FinishedHandler(object sender, EventArgs eventArgs) => handle.Set();
Console.WriteLine("Lecture terminée");
}
}
catch (OperationCanceledException)
{
}
}
async Task <Stream> TextToAudioStreamAsync(string text, string language, SsmlVoiceGender voiceGender, CancellationToken token)
{
var request = new SynthesizeSpeechRequest
{
AudioConfig = new AudioConfig
{
AudioEncoding = AudioEncoding.OggOpus,
},
Input = new SynthesisInput
{
Text = text
},
Voice = new VoiceSelectionParams
{
LanguageCode = language,
SsmlGender = voiceGender,
},
};
var response = await ttsClient.SynthesizeSpeechAsync(request, token);
using (var opusStream = new MemoryStream())
{
response.AudioContent.WriteTo(opusStream);
opusStream.Position = 0;
var opusDecoder = new OpusDecoder(48000, 2);
var oggIn = new OpusOggReadStream(opusDecoder, opusStream);
var pcmStream = new MemoryStream();
while (oggIn.HasNextPacket)
{
short[] packet = oggIn.DecodeNextPacket();
if (packet != null)
{
for (int i = 0; i < packet.Length; i++)
{
var bytes = BitConverter.GetBytes(packet[i]);
pcmStream.Write(bytes, 0, bytes.Length);
}
}
}
pcmStream.Position = 0;
return(pcmStream);
}
}
}
