public IWaveProvider Resample(IWaveProvider inputProvider)
{
var output = inputProvider;
if (_inFormat.Encoding == WaveFormatEncoding.IeeeFloat)
{
//This needs to change if the Discord sound format changes
output = new WaveFloatTo16Provider(inputProvider);
_log.Debug("Resampling from IeeeFloat using FloatTo16Provider");
}
else if (_inFormat.BitsPerSample != SoundSettings.DiscordFormat.BitsPerSample)
{
output = new WaveFormatConversionProvider(new WaveFormat(_inFormat.SampleRate, SoundSettings.DiscordFormat.BitsPerSample, _inFormat.Channels), output);
_log.Debug($"Resampling from {_inFormat.BitsPerSample} to {SoundSettings.DiscordFormat.BitsPerSample} bit using FormatConversionProvider");
}
if (_inFormat.Channels != SoundSettings.DiscordFormat.Channels)
{
output = new WaveFormatConversionProvider(new WaveFormat(_inFormat.SampleRate, SoundSettings.DiscordFormat.BitsPerSample, SoundSettings.DiscordFormat.Channels), output);
_log.Debug($"Resampling from {_inFormat.Channels} to {SoundSettings.DiscordFormat.Channels} using FormatConversionProbider");
}
if (_inFormat.SampleRate != SoundSettings.DiscordFormat.SampleRate)
{
output = new WaveFormatConversionProvider(SoundSettings.DiscordFormat, output);
_log.Debug($"Resampling from {_inFormat.SampleRate} to {SoundSettings.DiscordFormat.SampleRate} using FormatConversionProbider");
}
return(output);
}
public void shutDown()
{
_waveInStream.StopRecording();
_waveIn16Stream = null;
_waveTemp = null;
_waveInStream.Dispose();
_waveInStream = null;
}
public Speakers()
{
_recording = false;
_soundServer = new UDPServer(0);
_waveInStream = new WasapiLoopbackCapture();
_waveTemp = new WaveInProvider(_waveInStream);
_waveIn16Stream = new WaveFloatTo16Provider(_waveTemp);
//Console.Out.WriteLine("Sample rate: " + _waveInStream.WaveFormat.SampleRate);
/*Console.Out.WriteLine("\nEnconding: " + _waveInStream.WaveFormat.Encoding);
* Console.Out.WriteLine("Sample rate: " + _waveInStream.WaveFormat.SampleRate);
* Console.Out.WriteLine("Bits per sample: " + _waveInStream.WaveFormat.BitsPerSample);
* Console.Out.WriteLine("Channels: " + _waveInStream.WaveFormat.Channels);*/
}
static void Main(string[] args)
{
var fileName = @"C:\Users\milkitic\Downloads\HuΣeR Vs. SYUNN feat.いちか - 狂水一華.mp3";
var obj = new AudioDataHelper(fileName);
var data = obj.GetData(out var waveFormat);
var memoryStream = new MemoryStream(data);
var waveStream = new RawSourceWaveStream(memoryStream, waveFormat);
var p = new WaveFloatTo16Provider(waveStream);
WaveFileWriter.CreateWaveFile("a.wav", p);
var reader = new Mp3FileReaderBase(fileName, format => new DmoMp3FrameDecompressor(format));
WaveFileWriter.CreateWaveFile("b.wav", reader);
}
private void initAudioRecorder()
{
if (audio != null)
{
audio.DataAvailable -= Audio_DataAvailable;
audio.RecordingStopped -= Audio_RecordingStopped;
audio.Dispose();
}
if (blankplayer != null)
{
blankplayer.Dispose();
}
audio = new WasapiLoopbackCapture(device);
sourceFormat = audio.WaveFormat;
if (sourceProvider == null)
{
sourceProvider = new BufferedWaveProvider(sourceFormat);
sourceProvider.ReadFully = false;
wfto16prov = new WaveFloatTo16Provider(sourceProvider);
monovolumeprovider = new StereoToMonoProvider16(wfto16prov);
formatconv = new WaveFormatConversionProvider(new WaveFormat(24000, 16, 1), monovolumeprovider);
}
text_encoding.Text = sourceFormat.Encoding.ToString();
//var client = device.AudioClient.AudioRenderClient;
blankplayer = new WasapiOut(device, AudioClientShareMode.Shared, false, 0);
silence = new SilenceProvider(sourceFormat).ToSampleProvider();
AudioDevice_Text.ForeColor = Color.Black;
try
{
blankplayer.Init(silence);
}
catch
{
AudioDevice_Text.ForeColor = Color.Red;
}
audio.DataAvailable += Audio_DataAvailable;
audio.RecordingStopped += Audio_RecordingStopped;
AudioMonitor = device.AudioMeterInformation;
}
public static void ApplyAutoTune(string fileToProcess, string tempFile, AutoTuneSettings autotuneSettings)
{
using (WaveFileReader waveFileReader = new WaveFileReader(fileToProcess))
{
IWaveProvider source = new Wave16ToFloatProvider(waveFileReader);
IWaveProvider sourceProvider = new AutoTuneWaveProvider(source, autotuneSettings);
IWaveProvider waveProvider = new WaveFloatTo16Provider(sourceProvider);
using (WaveFileWriter waveFileWriter = new WaveFileWriter(tempFile, waveProvider.WaveFormat))
{
byte[] array = new byte[8192];
int num;
do
{
num = waveProvider.Read(array, 0, array.Length);
waveFileWriter.Write(array, 0, num);
}while (num != 0 && waveFileWriter.Length < waveFileReader.Length);
}
}
}
public static void ApplyAutoTune(string fileToProcess, string tempFile, AutoTuneSettings autotuneSettings)
{
using (var reader = new WaveFileReader(fileToProcess))
{
IWaveProvider stream32 = new Wave16ToFloatProvider(reader);
IWaveProvider streamEffect = new AutoTuneWaveProvider(stream32, autotuneSettings);
IWaveProvider stream16 = new WaveFloatTo16Provider(streamEffect);
using (var converted = new WaveFileWriter(tempFile, stream16.WaveFormat))
{
// buffer length needs to be a power of 2 for FFT to work nicely
// however, make the buffer too long and pitches aren't detected fast enough
// successful buffer sizes: 8192, 4096, 2048, 1024
// (some pitch detection algorithms need at least 2048)
var buffer = new byte[8192];
int bytesRead;
do
{
bytesRead = stream16.Read(buffer, 0, buffer.Length);
converted.Write(buffer, 0, bytesRead);
} while (bytesRead != 0 && converted.Length < reader.Length);
}
}
}
public AudioPassthrough(HotkeyAudioDevice input, HotkeyAudioDevice output, AutoTuneSettings autoTuneSettings)
{
LastPauseOrStart = new DateTime(2000, 1, 1);
_autoTuneSettings = autoTuneSettings;
_waveSource = new WaveIn();
WaveSource.DeviceNumber = input.AudioDeviceId;
//WaveSource.WaveFormat = new WaveFormat();
WaveSource.WaveFormat = WaveFormat.CreateIeeeFloatWaveFormat(44100, 1);
WaveSource.RecordingStopped += WaveSource_RecordingStopped;
WaveSource.DataAvailable += WaveSource_DataAvailable;
//WaveSource.BufferMilliseconds = 300;
//_waveInProvider = new JoeWaveInProvider(WaveSource);
//StereoToMonoProvider16 stereoToMonoprovider = new StereoToMonoProvider16(_waveInProvider);
//_wave16ToFloatProvider = new Wave16ToFloatProvider(stereoToMonoprovider);
//var test3 = ;
MemoryStream = new MemoryStream(new byte[8192]);
var rawSourceStream = new RawSourceWaveStream(MemoryStream, WaveFormat.CreateIeeeFloatWaveFormat(44100, 1));
//_autoTuner = new AutoTuneWaveProvider(_wave16ToFloatProvider, AutoTuneSettings);
_autoTuner = new AutoTuneWaveProvider(rawSourceStream, AutoTuneSettings);
_waveFloatTo16Provider = new WaveFloatTo16Provider(_autoTuner);
_monoToStereoProvider = new MonoToStereoProvider16(_waveFloatTo16Provider);
//_monoToStereoProvider = new MonoToStereoProvider16(_autoTuner);
_waveOutput = new WaveOutEvent();
WaveOutput.DeviceNumber = output.AudioDeviceId;
WaveOutput.PlaybackStopped += WaveOutput_PlaybackStopped;
WaveOutput.Init(_monoToStereoProvider);
//JoeWaveOutEvent joeWaveOut = new JoeWaveOutEvent();
//joeWaveOut.DeviceNumber = output.AudioDeviceId;
//joeWaveOut.DesiredLatency = 329;
//joeWaveOut.NumberOfBuffers = 7;
//joeWaveOut.Init(_monoToStereoProvider);
//DirectSoundDeviceInfo foundDeviceInfo = null;
//foreach (DirectSoundDeviceInfo deviceInfo in DirectSoundOut.Devices)
//{
// if (deviceInfo.Description.Contains(output.AudioDeviceName))
// {
// foundDeviceInfo = deviceInfo;
// break;
// }
//}
//if (foundDeviceInfo != null)
//{
// DirectSoundOutput = new DirectSoundOut(foundDeviceInfo.Guid, 300);
// DirectSoundOutput.PlaybackStopped += DirectSoundOutput_PlaybackStopped;
// DirectSoundOutput.Init(_monoToStereoProvider);
//}
}
/// <summary>
/// Cast the byte stream into a float buffer and launch the pitch detection,
/// the smooth maker and raise events in case if the standard deviation is too high.
/// </summary>
/// <param name="buffer">Byte buffer corresponding to the raw signal.</param>
/// <param name="bytesRecorded">Number of frames in the buffer.</param>
public void pitchComputePeak(byte[] buffer, int bytesRecorded)
{
Stream stream = new MemoryStream(buffer);
var reader = new RawSourceWaveStream(stream, recordingFormat);
IWaveProvider stream32 = new Wave16ToFloatProvider(reader);
PitchWaveProvider streamEffect = new PitchWaveProvider(stream32, this);
IWaveProvider stream16 = new WaveFloatTo16Provider(streamEffect);
var buffert = new byte[1024];
int bytesRead;
do
{
bytesRead = stream16.Read(buffert, 0, buffert.Length);
} while (bytesRead != 0);
reader.Close();
stream.Close();
// Calculate the volume of the voice
long totalSquare = 0;
for (int i = 0; i < buffer.Length; i += 2)
{
short sample = (short)(buffer[i] | (buffer[i + 1] << 8));
totalSquare += sample * sample;
}
long meanSquare = 2 * totalSquare / buffert.Length;
double rms = Math.Sqrt(meanSquare);
double volume = rms / 32768.0;
// volume is between 0.0 and 1.0
// I used the average of 8 values to have a better result than just with the volume value
tabVol[0] = tabVol[1];
tabVol[1] = tabVol[2];
tabVol[2] = tabVol[3];
tabVol[3] = tabVol[4];
tabVol[4] = tabVol[5];
tabVol[5] = tabVol[6];
tabVol[6] = tabVol[7];
tabVol[7] = volume;
volume = tabVol.Average();
// To compare easily the volume value I use an integer
volume *= 10;
volume = Convert.ToInt32(volume);
//wiggle = the WIGGLE_SIZE(= 170) last elenents from pitchList
if (pitchList.Count > WIGGLE_SIZE)
{
for (i = 0; i < WIGGLE_SIZE; i++)
{
wiggle[i] = pitchList[pitchList.Count - (WIGGLE_SIZE + 1) + i];
}
}
int newValues = pitchList.Count - oldPitchListSize;
oldPitchListSize = pitchList.Count;
for (i = newValues; i > 0; i--)
{
DrawingSheetAvatarViewModel.backgroundXMLVoiceRecordingEventStream?.Invoke(this, pitchList[pitchList.Count - 1 - i]);
}
this.PitchSmoothing();
if (canSendEvent)
{
lock (Lock)
{
double sd1 = StdDevTShort();
double sd2 = StdDevLong();
double sd = StdDev();
// each time the function is called we shift the values
tab[0] = tab[1];
tab[1] = tab[2];
tab[2] = tab[3];
tab[3] = tab[4];
tab[4] = tab[5];
tab[5] = tab[6];
tab[6] = tab[7];
tab[7] = sd2;
// If the rising is over, permit to don't count a peak as a rising up and down in a row
if (tab[0] == 0 || tab[tab.Length - 1] == 0 || (tab[0] <= beginningValue + 5 && tab[0] >= beginningValue - 5 && advancementCounter > 2))
{
state = 0;
beginningValue = 0;
advancementCounter = 0;
}
else
{
// If the curve is flat
if (state == 0)
{
// Calculate the wanted index
int indMax = localMaximum(tab);
int indMin = localMinimum(tab);
// If there is a big rising up
if (tab[indMax] > tab[0] + peakThreshold)
{
state = 1;
beginningValue = tab[0];
advancementCounter = 0;
}
// Or if there is a big rising down
else if (tab[indMin] < tab[0] - peakThreshold)
{
state = -1;
beginningValue = tab[0];
advancementCounter = 0;
}
else
{
beginningValue = 0;
advancementCounter = 0;
}
}
// If the curve is growing up or down
else
{
double average = 0.0;
// Calculate the average of the seven last value
for (int cpt = 1; cpt < tab.Length - 1; cpt++)
{
average += tab[cpt];
}
average /= tab.Length - 1;
// The curve is flat, so it reinitialize all
if (average > tab[0] - 3 && average < tab[0] + 3)
{
beginningValue = 0;
advancementCounter = 0;
state = 0;
}
// The counter advance
else
{
advancementCounter++;
}
}
}
// All the feedback files of the different tests
//using (System.IO.StreamWriter file = new System.IO.StreamWriter(@"C:\Users\Public\TestFolder\test.txt", true))
//{
// file.WriteLine("frame n" + k + " -- state : " + state);
//}
//using (System.IO.StreamWriter file = new System.IO.StreamWriter(@"C:\Users\Public\TestFolder\testShort.txt", true))
//{
// file.WriteLine(sd1);
//}
//using (System.IO.StreamWriter file = new System.IO.StreamWriter(@"C:\Users\Public\TestFolder\testLong.txt", true))
//{
// file.WriteLine(sd2);
//}
//using (System.IO.StreamWriter file = new System.IO.StreamWriter(@"C:\Users\Public\TestFolder\testVol.txt", true))
//{
// file.WriteLine("frame n" + k + " -- state : " + volume);
//}
// For testing the efficient of the peak detection we used the boring event to display the feedback.
// ASu has now changed this to use a peak event
if (volume >= 2) //volumethreshold)
{
if (state == 1) //&& !this.sent)
{
PeakEvent(this, new InstantFeedback("Rising Tone"));
//this.sent = true;
}
if (state == -1) //&& !this.sent)
{
PeakEvent(this, new InstantFeedback("Falling Tone"));
//this.sent = true;
}
}
k++;
i++;
}
}
}
/// <summary>
/// Cast the byte stream into a float buffer and launch the pitch detection,
/// the smooth maker and raise events in case if the standard deviation is too high.
/// </summary>
/// <param name="buffer">Byte buffer corresponding to the raw signal.</param>
/// <param name="bytesRecorded">Number of frames in the buffer.</param>
public void pitchCompute(byte[] buffer, int bytesRecorded)
{
Stream stream = new MemoryStream(buffer);
var reader = new RawSourceWaveStream(stream, recordingFormat);
IWaveProvider stream32 = new Wave16ToFloatProvider(reader);
PitchWaveProvider streamEffect = new PitchWaveProvider(stream32, this);
IWaveProvider stream16 = new WaveFloatTo16Provider(streamEffect);
var buffert = new byte[1024];
int bytesRead;
do
{
bytesRead = stream16.Read(buffert, 0, buffert.Length);
} while (bytesRead != 0);
reader.Close();
stream.Close();
this.PitchSmoothing();
if (canSendEvent && !MainWindow.main.audioProvider.replayMode)
{
lock (Lock)
{
double sd1 = StdDevTShort();
double sd2 = StdDevLong();
double sd = StdDev();
//Console.WriteLine("SD =" + sd + " Threshiold " + Threshold + "sd1 = " + sd1);
if ((sd <= Threshold && sd != 0.0) && pitchList.Last() > 0 && !this.sent)
{
BoringEvent(this, new LongFeedback("", true));
this.sent = true;
}
if ((sd > Threshold || sd == 0.0) && this.sent)
{
BoringEvent(this, new LongFeedback("", false));
this.sent = false;
}
/* There is something wrong with the way the above code. The tooBoringText displays at the beginning of a boring episode
* and then disappears after a few seconds. Then it flashes up again at the end of a boring episode for a few seconds
* I tried to model the code below on the emotion recognition code which also uses long feedback and responds very quickly
* to a change in facial expression.
*/
// Too much variation
/* if (sd > ThresholdVariation && !this.sent && pitchList.Last() > 0)
* {
* BoringEvent(this, new LongFeedback(tooMuchVariationText, true));
* this.sent = true;
* }
* if (sd <= ThresholdVariation || sd == 0.0 && this.sent)
* {
* BoringEvent(this, new LongFeedback(tooMuchVariationText, false));
* this.sent = false;
* }
*/
i++;
}
}
}
