# FFT in the real world

## Main Question or Discussion Point

Hi
so I have a working FFT. I tested it but inputting sin waves in to it.
will not i am inputting sound waves from a Mic and it is not looking to do...
example:
I am using to youtube vid
it outputs B7 (Musical note) b′′′′ Four-lined 3951.066 Frequency
and I run my program this is what I am getting 3951.066

I am getting more action around 41951, 42112, 20062, and 24035

then
3951.066

so is there anything I am add to my FFT that would help ?

oh I am sampling at 44100 Hz
and my N = 88200

I also tried this one

and I got a higher magnitude from 424.5 and 435.5
then
from 528
do you think there too much noise coming from the computer speakers ?

Last edited by a moderator:

Related Electrical Engineering News on Phys.org
DrClaude
Mentor

jim mcnamara
Mentor

public polar1[] FFT(polar1[] x)
{
int N2 = x.Length;
polar1[] X = new polar1[N2];
if (N2 == 1)
{
return x;
}
polar1[] odd = new polar1[N2 / 2];
polar1[] even = new polar1[N2 / 2];
polar1[] Y_Odd = new polar1[N2 / 2];
polar1[] Y_Even = new polar1[N2 / 2];

for (int t = 0; t < N2 / 2; t++)
{
even[t].img = x[t * 2].img;
even[t].real = x[t * 2].real;
odd[t].img = x[(t * 2) + 1].img;
odd[t].real = x[(t * 2) + 1].real;
}
Y_Even = FFT(even);
Y_Odd = FFT(odd);
polar1 temp4;

for (int k = 0; k < (N2 / 2); k++)
{
temp4 = Complex1(k, N2);
X[k].real = Y_Even[k].real + (Y_Odd[k].real * temp4.real);
X[k + (N2 / 2)].real = Y_Even[k].real - (Y_Odd[k].real * temp4.real);
X[k].img = Y_Even[k].img + (Y_Odd[k].real * temp4.img);
X[k + (N2 / 2)].img = Y_Even[k].img - (Y_Odd[k].real * temp4.img);
}

return X;
}

public polar1 Complex1(int K, int N3)
{
polar1 temp;
double temp1;
temp1 = (2D * Math.PI *K) / N3;
temp.real = Math.Cos(temp1);
temp.img = Math.Sin(temp1);
return temp;

}

public struct polar1
{
public double real;
public double img;

};

DrClaude

have you had time to look at the code ?
what do you think?
that is the same code i use in my other thread

DrClaude
Mentor
I'm sorry, but I don't have the patience to read through an implementation of a FT. If you just need the result, then use one of the many wonderful implementations available (GSL, FFTW, etc.). If you are doing it to learn about numerical methods, then roll up your sleves and debug the code as if there was no tomorrow. You will probably not learn as much if someone else just points out the error.

that is cool.. but I do think it is working