private async void openToolStripMenuItem_Click(object sender, EventArgs e)
{
var ofd = new OpenFileDialog();
if (ofd.ShowDialog() != DialogResult.OK)
{
return;
}
using (var stream = new FileStream(ofd.FileName, FileMode.Open))
{
var waveFile = new WaveFile(stream, true);
_signal = waveFile[Channels.Left];
}
var sr = _signal.SamplingRate;
var barkbands = FilterBanks.BarkBands(16, 512, sr, 100 /*Hz*/, 6500 /*Hz*/, overlap: false);
var barkbank = FilterBanks.Triangular(512, sr, barkbands);
var mfccExtractor = new MfccExtractor(_signal.SamplingRate, 13,
//filterbankSize: 40,
//lowFreq: 100,
//highFreq: 4200,
//lifterSize: 22,
preEmphasis: 0.97,
//filterbank: barkbank,
window: WindowTypes.Hamming);
_mfccVectors = mfccExtractor.ComputeFrom(_signal);
//FeaturePostProcessing.NormalizeMean(_mfccVectors); // optional
//FeaturePostProcessing.AddDeltas(_mfccVectors);
FillFeaturesList(_mfccVectors, mfccExtractor.FeatureDescriptions);
mfccListView.Items[0].Selected = true;
melFilterBankPanel.Groups = mfccExtractor.FilterBank;
mfccPanel.Line = _mfccVectors[0].Features;
using (var csvFile = new FileStream("mfccs.csv", FileMode.Create))
{
var header = mfccExtractor.FeatureDescriptions;
//.Concat(mfccExtractor.DeltaFeatureDescriptions)
//.Concat(mfccExtractor.DeltaDeltaFeatureDescriptions);
var serializer = new CsvFeatureSerializer(_mfccVectors, header);
await serializer.SerializeAsync(csvFile);
}
}
private async void openToolStripMenuItem_Click(object sender, EventArgs e)
{
var ofd = new OpenFileDialog();
if (ofd.ShowDialog() != DialogResult.OK)
{
return;
}
using (var stream = new FileStream(ofd.FileName, FileMode.Open))
{
var waveFile = new WaveFile(stream, true);
_signal = waveFile[Channels.Left];
}
var mfccExtractor = new MfccExtractor(13,
//frameSize: 0.03125,
//hopSize: 0.015625,
melFilterbankSize: 20,
//lowFreq: 100,
//highFreq: 4200,
//lifterSize: 22,
preEmphasis: 0.95,
window: WindowTypes.Hamming);
_mfccVectors = mfccExtractor.ComputeFrom(_signal);
//FeaturePostProcessing.NormalizeMean(_mfccVectors);
//FeaturePostProcessing.AddDeltas(_mfccVectors);
FillFeaturesList(_mfccVectors, mfccExtractor.FeatureDescriptions);
mfccListView.Items[0].Selected = true;
melFilterBankPanel.Groups = mfccExtractor.FilterBank;
mfccPanel.Line = _mfccVectors[0].Features;
using (var csvFile = new FileStream("mfccs.csv", FileMode.Create))
{
var header = mfccExtractor.FeatureDescriptions
.Concat(mfccExtractor.DeltaFeatureDescriptions)
.Concat(mfccExtractor.DeltaDeltaFeatureDescriptions);
var serializer = new CsvFeatureSerializer(_mfccVectors, header);
await serializer.SerializeAsync(csvFile);
}
}
public void TestOnlineFeatureExtractor()
{
var mfccOptions = new MfccOptions
{
SamplingRate = 8000,
FeatureCount = 5,
FrameSize = 256,
HopSize = 50,
FilterBankSize = 8
};
var signal = new WhiteNoiseBuilder().OfLength(1000).Build();
var mfccExtractor = new MfccExtractor(mfccOptions);
var mfccVectors = mfccExtractor.ComputeFrom(signal);
var onlineMfccExtractor = new OnlineFeatureExtractor(new MfccExtractor(mfccOptions));
var onlineMfccVectors = new List <float[]>();
var i = 0;
while (i < signal.Length)
{
// emulating online blocks with different sizes:
var size = (i + 1) * 15;
var block = signal.Samples.Skip(i).Take(size).ToArray();
var newVectors = onlineMfccExtractor.ComputeFrom(block);
onlineMfccVectors.AddRange(newVectors);
i += size;
}
var diff = mfccVectors.Zip(onlineMfccVectors, (e, o) => e.Zip(o, (f1, f2) => f1 - f2).Sum());
Assert.That(diff, Is.All.EqualTo(0).Within(1e-7f));
}
private void buttonCompute_Click(object sender, EventArgs e)
{
var filterCount = int.Parse(textBoxSize.Text);
var samplingRate = _signal.SamplingRate;
var fftSize = int.Parse(textBoxFftSize.Text);
var lowFreq = float.Parse(textBoxLowFreq.Text);
var highFreq = float.Parse(textBoxHighFreq.Text);
Tuple <double, double, double>[] bands;
float[][] filterbank = null;
VtlnWarper vtln = null;
if (checkBoxVtln.Checked)
{
var alpha = float.Parse(textBoxVtlnAlpha.Text);
var vtlnLow = float.Parse(textBoxVtlnLow.Text);
var vtlnHigh = float.Parse(textBoxVtlnHigh.Text);
vtln = new VtlnWarper(alpha, lowFreq, highFreq, vtlnLow, vtlnHigh);
}
switch (comboBoxFilterbank.Text)
{
case "Mel":
bands = FilterBanks.MelBands(filterCount, fftSize, samplingRate, lowFreq, highFreq, checkBoxOverlap.Checked);
break;
case "Mel Slaney":
bands = FilterBanks.MelBandsSlaney(filterCount, fftSize, samplingRate, lowFreq, highFreq, checkBoxOverlap.Checked);
filterbank = FilterBanks.MelBankSlaney(filterCount, fftSize, samplingRate, lowFreq, highFreq, checkBoxNormalize.Checked, vtln);
break;
case "Bark":
bands = FilterBanks.BarkBands(filterCount, fftSize, samplingRate, lowFreq, highFreq, checkBoxOverlap.Checked);
break;
case "Bark Slaney":
bands = FilterBanks.BarkBandsSlaney(filterCount, fftSize, samplingRate, lowFreq, highFreq, checkBoxOverlap.Checked);
filterbank = FilterBanks.BarkBankSlaney(filterCount, fftSize, samplingRate, lowFreq, highFreq);
break;
case "Critical bands":
bands = FilterBanks.CriticalBands(filterCount, fftSize, samplingRate, lowFreq, highFreq);
break;
case "Octave bands":
bands = FilterBanks.OctaveBands(filterCount, fftSize, samplingRate, lowFreq, highFreq, checkBoxOverlap.Checked);
break;
case "ERB":
bands = null;
filterbank = FilterBanks.Erb(filterCount, fftSize, samplingRate, lowFreq, highFreq);
break;
default:
bands = FilterBanks.HerzBands(filterCount, fftSize, samplingRate, lowFreq, highFreq, checkBoxOverlap.Checked);
break;
}
if (bands != null && filterbank == null)
{
switch (comboBoxShape.Text)
{
case "Triangular":
filterbank = FilterBanks.Triangular(fftSize, samplingRate, bands, vtln, Utils.Scale.HerzToMel);
break;
case "Trapezoidal":
filterbank = FilterBanks.Trapezoidal(fftSize, samplingRate, bands, vtln);
break;
case "BiQuad":
filterbank = FilterBanks.BiQuad(fftSize, samplingRate, bands);
break;
default:
filterbank = FilterBanks.Rectangular(fftSize, samplingRate, bands, vtln);
break;
}
if (checkBoxNormalize.Checked)
{
FilterBanks.Normalize(filterCount, bands, filterbank);
}
}
var spectrumType = (SpectrumType)comboBoxSpectrum.SelectedIndex;
var nonLinearity = (NonLinearityType)comboBoxNonLinearity.SelectedIndex;
var logFloor = float.Parse(textBoxLogFloor.Text);
var mfccExtractor = new MfccExtractor(//samplingRate, 13, 0.025, 0.01,
samplingRate, 13, 512.0 / samplingRate, 0.01,
filterbank: filterbank,
//filterbankSize: 26,
//highFreq: 8000,
//preEmphasis: 0.97,
//lifterSize: 22,
//includeEnergy: true,
spectrumType: spectrumType,
nonLinearity: nonLinearity,
dctType: comboBoxDct.Text,
window: WindowTypes.Hamming,
logFloor: logFloor);
_mfccVectors = mfccExtractor.ComputeFrom(_signal);
//_mfccVectors = mfccExtractor.ComputeFrom(_signal * 32768);
//var mfccVectorsP = mfccExtractor.ParallelComputeFrom(_signal * 32768);
//for (var i = 0; i < _mfccVectors.Count; i++)
//{
// for (var j = 0; j < _mfccVectors[i].Features.Length; j++)
// {
// if (Math.Abs(_mfccVectors[i].Features[j] - mfccVectorsP[i].Features[j]) > 1e-32f)
// {
// MessageBox.Show($"Nope: {i} - {j}");
// return;
// }
// if (Math.Abs(_mfccVectors[i].TimePosition - mfccVectorsP[i].TimePosition) > 1e-32f)
// {
// MessageBox.Show($"Time: {i} - {j}");
// return;
// }
// }
//}
//FeaturePostProcessing.NormalizeMean(_mfccVectors); // optional (but REQUIRED for PNCC!)
//FeaturePostProcessing.AddDeltas(_mfccVectors);
var header = mfccExtractor.FeatureDescriptions;
//.Concat(mfccExtractor.DeltaFeatureDescriptions)
//.Concat(mfccExtractor.DeltaDeltaFeatureDescriptions);
FillFeaturesList(_mfccVectors, header);
mfccListView.Items[0].Selected = true;
melFilterBankPanel.Groups = mfccExtractor.FilterBank;
mfccPanel.Line = _mfccVectors[0].Features;
}
public void extractFeatures()
{
//NWaves
//Initial setup
if (_filePath != null)
{
DiscreteSignal signal;
// load
var mfcc_no = 24;
var samplingRate = 44100;
var mfccOptions = new MfccOptions
{
SamplingRate = samplingRate,
FeatureCount = mfcc_no,
FrameDuration = 0.025 /*sec*/,
HopDuration = 0.010 /*sec*/,
PreEmphasis = 0.97,
Window = WindowTypes.Hamming
};
var opts = new MultiFeatureOptions
{
SamplingRate = samplingRate,
FrameDuration = 0.025,
HopDuration = 0.010
};
var tdExtractor = new TimeDomainFeaturesExtractor(opts);
var mfccExtractor = new MfccExtractor(mfccOptions);
// Read from file.
featureString = String.Empty;
featureString = $"green,";
//MFCC
var avg_vec_mfcc = new List <float>(mfcc_no + 1);
//TD Features
var avg_vec_td = new List <float>(4);
//Spectral features
var avg_vec_spect = new List <float>(10);
for (var i = 0; i < mfcc_no; i++)
{
avg_vec_mfcc.Add(0f);
}
for (var i = 0; i < 4; i++)
{
avg_vec_td.Add(0f);
}
for (var i = 0; i < 10; i++)
{
avg_vec_spect.Add(0f);
}
string specFeatures = String.Empty;
Console.WriteLine($"{tag} Reading from file");
using (var stream = new FileStream(_filePath, FileMode.Open))
{
var waveFile = new WaveFile(stream);
signal = waveFile[channel : Channels.Left];
////Compute MFCC
float[] mfvfuck = new float[25];
var sig_sam = signal.Samples;
mfccVectors = mfccExtractor.ComputeFrom(sig_sam);
var fftSize = 1024;
tdVectors = tdExtractor.ComputeFrom(signal.Samples);
var fft = new Fft(fftSize);
var resolution = (float)samplingRate / fftSize;
var frequencies = Enumerable.Range(0, fftSize / 2 + 1)
.Select(f => f * resolution)
.ToArray();
var spectrum = new Fft(fftSize).MagnitudeSpectrum(signal).Samples;
var centroid = Spectral.Centroid(spectrum, frequencies);
var spread = Spectral.Spread(spectrum, frequencies);
var flatness = Spectral.Flatness(spectrum, 0);
var noiseness = Spectral.Noiseness(spectrum, frequencies, 3000);
var rolloff = Spectral.Rolloff(spectrum, frequencies, 0.85f);
var crest = Spectral.Crest(spectrum);
var decrease = Spectral.Decrease(spectrum);
var entropy = Spectral.Entropy(spectrum);
specFeatures = $"{centroid},{spread},{flatness},{noiseness},{rolloff},{crest},{decrease},{entropy}";
//}
Console.WriteLine($"{tag} All features ready");
for (int calibC = 0; calibC < mfccVectors.Count; calibC += (mfccVectors.Count / duration) - 1)
{
featureString = String.Empty;
var tmp = new ModelInput();
for (var i = 0; i < mfcc_no; i++)
{
avg_vec_mfcc[i] = mfccVectors[calibC][i];
}
for (var i = 0; i < 4; i++)
{
avg_vec_td[i] = tdVectors[calibC][i];
}
featureString += String.Join(",", avg_vec_mfcc);
featureString += ",";
featureString += String.Join(",", avg_vec_td);
featureString += ",";
featureString += specFeatures;
Console.WriteLine($"{tag} Feature String ready {featureString}");
if (File.Exists(Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData), "temp")))
{
File.Delete(Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData), "temp"));
File.WriteAllText(Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData), "temp"), featureString);
}
else
{
File.WriteAllText(Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData), "temp"), featureString);
}
MLContext mLContext = new MLContext();
string fileName = Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData), "temp");
IDataView dataView = mLContext.Data.LoadFromTextFile <ModelInput>(
path: fileName,
hasHeader: false,
separatorChar: ',',
allowQuoting: true,
allowSparse: false);
// Use first line of dataset as model input
// You can replace this with new test data (hardcoded or from end-user application)
ModelInput sampleForPrediction = mLContext.Data.CreateEnumerable <ModelInput>(dataView, false)
.First();
ModelOutput opm = ConsumeModel.Predict(sampleForPrediction);
featureTimeList.Add(opm.Score);
Console.WriteLine($"{tag} Feature vs time list ready");
}
//Console.WriteLine($"{tag} MFCC: {mfccVectors.Count}");
//Console.WriteLine($"{tag} TD: {tdVectors.Count}");
//Console.WriteLine($"{tag} featureTimeArray: {featureTimeList.Count} {featureString}");
}
}
}
async public void extractFeatures(string _filepath, StorageFile sf)
{
op = new float[10];
tdVectors = new List <float[]>();
mfccVectors = new List <float[]>();
featureTimeList = new List <float[]>();
//NWaves
FilePath = _filepath;
PredictedLabel = "Ready!.";
//player.Load(GetStreamFromFile(FilePath));
//player.Play();
mMedia.Source = MediaSource.CreateFromStorageFile(sf);
bool test = player.IsPlaying;
mMedia.AutoPlay = true;
MusicProperties properties = await sf.Properties.GetMusicPropertiesAsync();
TimeSpan myTrackDuration = properties.Duration;
duration = Convert.ToInt32(myTrackDuration.TotalSeconds);
if (FilePath != null)
{
DiscreteSignal signal;
// load
var mfcc_no = 24;
var samplingRate = 44100;
var mfccOptions = new MfccOptions
{
SamplingRate = samplingRate,
FeatureCount = mfcc_no,
FrameDuration = 0.025 /*sec*/,
HopDuration = 0.010 /*sec*/,
PreEmphasis = 0.97,
Window = WindowTypes.Hamming
};
var opts = new MultiFeatureOptions
{
SamplingRate = samplingRate,
FrameDuration = 0.025,
HopDuration = 0.010
};
var tdExtractor = new TimeDomainFeaturesExtractor(opts);
var mfccExtractor = new MfccExtractor(mfccOptions);
// Read from file.
featureString = String.Empty;
featureString = $"green,";
//MFCC
var mfccList = new List <List <double> >();
var tdList = new List <List <double> >();
//MFCC
//TD Features
//Spectral features
for (var i = 0; i < mfcc_no; i++)
{
mfccList.Add(new List <double>());
}
for (var i = 0; i < 4; i++)
{
tdList.Add(new List <double>());
}
string specFeatures = String.Empty;
Console.WriteLine($"{tag} Reading from file");
using (var stream = new FileStream(FilePath, FileMode.Open))
{
var waveFile = new WaveFile(stream);
signal = waveFile[channel : Channels.Left];
////Compute MFCC
float[] mfvfuck = new float[25];
var sig_sam = signal.Samples;
mfccVectors = mfccExtractor.ComputeFrom(sig_sam);
var fftSize = 1024;
tdVectors = tdExtractor.ComputeFrom(signal.Samples);
var fft = new Fft(fftSize);
var resolution = (float)samplingRate / fftSize;
var frequencies = Enumerable.Range(0, fftSize / 2 + 1)
.Select(f => f * resolution)
.ToArray();
var spectrum = new Fft(fftSize).MagnitudeSpectrum(signal).Samples;
var centroid = Spectral.Centroid(spectrum, frequencies);
var spread = Spectral.Spread(spectrum, frequencies);
var flatness = Spectral.Flatness(spectrum, 0);
var noiseness = Spectral.Noiseness(spectrum, frequencies, 3000);
var rolloff = Spectral.Rolloff(spectrum, frequencies, 0.85f);
var crest = Spectral.Crest(spectrum);
var decrease = Spectral.Decrease(spectrum);
var entropy = Spectral.Entropy(spectrum);
specFeatures = $"{centroid},{spread},{flatness},{noiseness},{rolloff},{crest},{decrease},{entropy}";
//}
Console.WriteLine($"{tag} All features ready");
for (int calibC = 0; calibC < mfccVectors.Count;)
{
featureString = String.Empty;
var tmp = new ModelInput();
for (var j = 0; j < (mfccVectors.Count / duration) - 1 && calibC < mfccVectors.Count; j++)
{
for (var i = 0; i < mfcc_no; i++)
{
mfccList[i].Add(mfccVectors[calibC][i]);
}
for (var i = 0; i < 4; i++)
{
tdList[i].Add(tdVectors[calibC][i]);
}
calibC += 1;
}
var mfcc_statistics = new List <double>();
for (var i = 0; i < mfcc_no; i++)
{
//preheader += m + "_mean";
//preheader += m + "_min";
//preheader += m + "_var";
//preheader += m + "_sd";
//preheader += m + "_med";
//preheader += m + "_lq";
//preheader += m + "_uq";
//preheader += m + "_skew";
//preheader += m + "_kurt";
mfcc_statistics.Add(Statistics.Mean(mfccList[i]));
mfcc_statistics.Add(Statistics.Minimum(mfccList[i]));
mfcc_statistics.Add(Statistics.Variance(mfccList[i]));
mfcc_statistics.Add(Statistics.StandardDeviation(mfccList[i]));
mfcc_statistics.Add(Statistics.Median(mfccList[i]));
mfcc_statistics.Add(Statistics.LowerQuartile(mfccList[i]));
mfcc_statistics.Add(Statistics.UpperQuartile(mfccList[i]));
mfcc_statistics.Add(Statistics.Skewness(mfccList[i]));
mfcc_statistics.Add(Statistics.Kurtosis(mfccList[i]));
}
var td_statistics = new List <double>();
for (var i = 0; i < 4; i++)
{
td_statistics.Add(Statistics.Mean(tdList[i]));
td_statistics.Add(Statistics.Minimum(tdList[i]));
td_statistics.Add(Statistics.Variance(tdList[i]));
td_statistics.Add(Statistics.StandardDeviation(tdList[i]));
td_statistics.Add(Statistics.Median(tdList[i]));
td_statistics.Add(Statistics.LowerQuartile(tdList[i]));
td_statistics.Add(Statistics.UpperQuartile(tdList[i]));
td_statistics.Add(Statistics.Skewness(tdList[i]));
td_statistics.Add(Statistics.Kurtosis(tdList[i]));
}
// Write MFCCs
featureString += String.Join(",", mfcc_statistics);
featureString += ",";
featureString += String.Join(",", td_statistics);
//Write Spectral features as well
featureString += ",";
featureString += specFeatures;
Console.WriteLine($"{tag} Feature String ready {featureString}");
if (File.Exists(Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData), "temp")))
{
File.Delete(Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData), "temp"));
File.WriteAllText(Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData), "temp"), featureString);
}
else
{
File.WriteAllText(Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData), "temp"), featureString);
}
MLContext mLContext = new MLContext();
string fileName = Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.LocalApplicationData), "temp");
IDataView dataView = mLContext.Data.LoadFromTextFile <ModelInput>(
path: fileName,
hasHeader: false,
separatorChar: ',',
allowQuoting: true,
allowSparse: false);
// Use first line of dataset as model input
// You can replace this with new test data (hardcoded or from end-user application)
ModelInput sampleForPrediction = mLContext.Data.CreateEnumerable <ModelInput>(dataView, false)
.First();
ModelOutput opm = ConsumeModel.Predict(sampleForPrediction);
featureTimeList.Add(opm.Score);
Console.WriteLine($"{tag} Feature vs time list ready");
}
//Console.WriteLine($"{tag} MFCC: {mfccVectors.Count}");
//Console.WriteLine($"{tag} TD: {tdVectors.Count}");
//Console.WriteLine($"{tag} featureTimeArray: {featureTimeList.Count} {featureString}");
}
}
playAudio();
}
static void Main(string[] args)
{
DiscreteSignal signal;
// load
var mfcc_no = 24;
var samplingRate = 16000;
var mfccOptions = new MfccOptions
{
SamplingRate = samplingRate,
FeatureCount = mfcc_no,
FrameDuration = 0.025 /*sec*/,
HopDuration = 0.010 /*sec*/,
PreEmphasis = 0.97,
Window = WindowTypes.Hamming
};
var opts = new MultiFeatureOptions
{
SamplingRate = samplingRate,
FrameDuration = 0.025,
HopDuration = 0.010
};
var tdExtractor = new TimeDomainFeaturesExtractor(opts);
var mfccExtractor = new MfccExtractor(mfccOptions);
var folders = Directory.GetDirectories(Path.Combine(Environment.CurrentDirectory, "Dataset"));
Console.WriteLine($"Started!");
using (var writer = File.CreateText(Path.Combine(Environment.CurrentDirectory, "Data.csv")))
{
//Write header
var main_header = "genre,";
main_header += String.Join(",", mfccExtractor.FeatureDescriptions);
main_header += ",";
main_header += String.Join(",", tdExtractor.FeatureDescriptions);
main_header += ",centroid,spread,flatness,noiseness,roloff,crest,decrease,spectral_entropy";
writer.WriteLine(main_header);
string feature_string = String.Empty;
foreach (var folder in folders)
{
var f_name = new DirectoryInfo(folder).Name;
var files = Directory.GetFiles(Path.Combine(Environment.CurrentDirectory, "Dataset", folder));
//Write the genre label here
Console.WriteLine($"{f_name}");
foreach (var filename in files)
{
feature_string = String.Empty;
feature_string = $"{f_name},";
//MFCC
var avg_vec_mfcc = new List <float>(mfcc_no + 1);
//TD Features
var avg_vec_td = new List <float>(4);
//Spectral features
var avg_vec_spect = new List <float>(10);
for (var i = 0; i < mfcc_no; i++)
{
avg_vec_mfcc.Add(0f);
}
for (var i = 0; i < 4; i++)
{
avg_vec_td.Add(0f);
}
for (var i = 0; i < 10; i++)
{
avg_vec_spect.Add(0f);
}
string specFeatures = String.Empty;
using (var stream = new FileStream(Path.Combine(Environment.CurrentDirectory, "Dataset", filename), FileMode.Open))
{
var waveFile = new WaveFile(stream);
signal = waveFile[Channels.Average];
//Compute MFCC
tdVectors = tdExtractor.ComputeFrom(signal);
mfccVectors = mfccExtractor.ComputeFrom(signal);
var fftSize = 1024;
var fft = new Fft(fftSize);
var resolution = (float)samplingRate / fftSize;
var frequencies = Enumerable.Range(0, fftSize / 2 + 1)
.Select(f => f * resolution)
.ToArray();
var spectrum = new Fft(fftSize).MagnitudeSpectrum(signal).Samples;
var centroid = Spectral.Centroid(spectrum, frequencies);
var spread = Spectral.Spread(spectrum, frequencies);
var flatness = Spectral.Flatness(spectrum, 0);
var noiseness = Spectral.Noiseness(spectrum, frequencies, 3000);
var rolloff = Spectral.Rolloff(spectrum, frequencies, 0.85f);
var crest = Spectral.Crest(spectrum);
var decrease = Spectral.Decrease(spectrum);
var entropy = Spectral.Entropy(spectrum);
specFeatures = $"{centroid},{spread},{flatness},{noiseness},{rolloff},{crest},{decrease},{entropy}";
}
//Write label here TODO
foreach (var inst in mfccVectors)
{
for (var i = 0; i < mfcc_no; i++)
{
avg_vec_mfcc[i] += inst[i];
}
}
foreach (var inst in tdVectors)
{
for (var i = 0; i < 4; i++)
{
avg_vec_td[i] += inst[i];
}
}
for (var i = 0; i < mfcc_no; i++)
{
avg_vec_mfcc[i] /= mfccVectors.Count;
}
for (var i = 0; i < 4; i++)
{
avg_vec_td[i] /= tdVectors.Count;
}
// Write MFCCs
feature_string += String.Join(",", avg_vec_mfcc);
feature_string += ",";
feature_string += String.Join(",", avg_vec_td);
//Write Spectral features as well
feature_string += ",";
feature_string += specFeatures;
writer.WriteLine(feature_string);
var file_name = new DirectoryInfo(filename).Name;
Console.WriteLine($"{file_name}");
}
}
}
Console.WriteLine($"DONE");
Console.ReadLine();
}
