Calculating the FFT in C#

admin

With NAudio (light)

example project:spectrograph

using NAudio.Wave; // for sound card access

using NAudio.Dsp; // for FastFourierTransform

// prepare the complex data which will be FFT'd

Complex[] fft_buffer = new Complex[fft_size];

for (int i=0; i < fft_size; i++)

{

    fft_buffer.X = (float)(unanalyzed_values* FastFourierTransform.HammingWindow(i, fft_size));

    fft_buffer.Y = 0;

}

// perform the FFT

FastFourierTransform.FFT(true, (int)Math.Log(fft_size, 2.0), fft_buffer);

// a list with FFT values

List<double> new_data = new List<double>();

for (int i = 0; i < spec_data[spec_data.Count - 1].Count; i++)

{

    // should this besqrt(X^2+Y^2)?

    double val;

    val = (double)fft_buffer.X+ (double)fft_buffer.Y;

    val = Math.Abs(val);

    if (checkBox1.Checked)val = Math.Log(val);

    new_data.Add(val);

}

With Accord (heavy)

Here is aminimal-case example how to convert an array of doubles into the frequencydomain using a Fast Fouriertransformation in C# (Visual Studio Community 2017). It usesthe Accord .NET library. Add areference to the Assembly Framework System.Numerics. Use NuGet toinstall Accord.Audio. example project:microphone FFT

using System.Numerics;

public double[] FFT(double[] data)

{

    int nPoints = data.Length; // whatever wemeasure must be a power of 2

    for (int i = 0; i < data.Length; i++) data = Math.Sin(i); // fill it withsome data

    double[] fft = new double[nPoints]; // this is wherewe will store the output (fft)

    Complex[] fftComplex = new Complex[nPoints]; // the FFTfunction requires complex format

    for (int i = 0; i < data.Length; i++)

        fftComplex = new Complex(data, 0.0); // make itcomplex format

    Accord.Math.FourierTransform.FFT(fftComplex, Accord.Math.FourierTransform.Direction.Forward);

    for (int i = 0; i < data.Length; i++)

        fft = fftComplex.Magnitude;// back to double

    return fft;

    //todo: this couldbe much faster by reusing variables

}