public void DynamicParameters(List <double> d, WindowBase window,
double fs, double f0, double fmin, double fmax, bool remove_dc, out double snr, out double sndr, out double enob)
{
double[] dd = new double[d.Count];
double dc_offset = 0;
if (remove_dc)
{
dc_offset = ArrayMath.Sum(d.ToArray()) / d.Count;
}
if (dc_offset != 0)
{
for (int i = 0; i < d.Count; i++)
{
dd[i] = d[i] - dc_offset;
}
}
Complex[] spec = DoFFT(dd, window);
snr = SignalNoise(spec, dd.Length, fs, f0, 6, 3, fmin, fmax);
sndr = SignalNoise(spec, dd.Length, fs, f0, 0, 3, fmin, fmax);
enob = (sndr - 1.76) / 6.02;
}
public double[] PowerSpectralDensity(double[] d, WindowBase window, double fs, double maxSignalAmpl)
{
Complex[] spec = DoFFT(d, window);
double[] y = new double[spec.Length];
double max = Double.MinValue;
for (int i = 0; i < spec.Length; i++)
{
y[i] = 20 * Math.Log10(spec[i].Abs);
if (y[i] > max)
{
max = y[i];
}
}
if (double.NaN.CompareTo(maxSignalAmpl) == 0)
{
ArrayMath.Add(ref y, -max);
}
else
{
max = 20 * Math.Log10(maxSignalAmpl / 2);
ArrayMath.Add(ref y, -max);
}
return(y);
}
public List <ChartData> Plot(List <double> d)
{
if (!visible)
{
return(new List <ChartData>());
}
dc_offset = 0;
if (remove_dc)
{
dc_offset = ArrayMath.Sum(d.ToArray()) / d.Count;
}
FFT fft = new FFT();
ChartData cd;
List <double> dd = new List <double>();
foreach (double da in d)
{
dd.Add(da - dc_offset);
}
List <ChartData> cds = new List <ChartData>();
fft.PowerSpectralDensity(dd.ToArray(), out cd, new Hanning(), fs, maxamplitude);
cd.Title = this.Title;
cd.ShowFullNameLegend = false;
cds.Add(cd);
return(cds);
}
/*
* The point on the Y - axis where the line Y = A + BX intercepts it is given by the equation:
* A = [(SUM(Y))*(SUM(SQUARE(X))) - (SUM(X))*(SUM(X*Y))] / [N*SUM(SQUARE(X)) - SQUARE((SUM(X)))].
*
*
*
* The slope of the line Y = A + BX is given by the equation:
* B = N*SUM(X*Y) - (SUM(X))*(SUM(Y)) / [N*SUM(SQUARE(X)) - SQUARE((SUM(X))]
*
*
*
* The correlation coefficient of the line Y = A + BX where 0 means no correlation and 1 means perfect
* correlation is given by the equation:
* R = N*SUM(X*Y) - (SUM(X))*(SUM(Y)) / SQRT[N*SUM(SQUARE(X)) - SQUARE(SUM(X))]*SQRT[N*SUM(SQUARE(Y)) - SQUARE((SUM(Y)))]
*/
public void Coefficients(out double a, out double b, out double r)
{
double x2sum = ArrayMath.Sum(ArrayMath.Pow(X, 2));
double xsum2 = Math.Pow(ArrayMath.Sum(X), 2);
double ysum = ArrayMath.Sum(Y);
double y2sum = ArrayMath.Sum(ArrayMath.Pow(Y, 2));
double ysum2 = Math.Pow(ArrayMath.Sum(Y), 2);
double xsum = ArrayMath.Sum(X);
double xysum = ArrayMath.Sum(ArrayMath.Multiply(X, Y));
a = (ysum * x2sum - xsum * xysum) / (N * x2sum - xsum2);
b = (N * xysum - xsum * ysum) / (N * x2sum - xsum2);
r = (N * xysum - xsum * ysum) / (Math.Sqrt(N * x2sum - xsum2) * Math.Sqrt(N * y2sum - ysum2));
}
public Complex[] DoFFT(double[] d, WindowBase window)
{
double M = d.Length;
double[] dwindow = window.Calc((int)M);
ArrayMath.Multiply(ref d, ref dwindow);
Complex[] fft = Run(d);
Complex[] spec = new Complex[fft.Length / 2];
for (int i = 1; i < fft.Length / 2 + 1; i++)
{
spec[i - 1].Real = fft[i].Real * 2 / M;
spec[i - 1].Imag = fft[i].Imag * 2 / M;
}
return(spec);
}
public LinearRegression(double[] X, double[] Y)
{
N = ArrayMath.MakeEqualLength(ref X, ref Y);
this.X = X;
this.Y = Y;
}
