0

I'm doing a pitch detection using a combination of an ACF and AMDF.
First I was using ACF in the time domain like this:

Get a buffer of 2048 samples
Window it (Hamming window)
sum=Sum(Buffer[i]*Buffer[i+lag]) for all i < 2048 - lag
acf = sum / 2048

And repeat the last 2 steps for all lags to be considered. (actually doing interpolation for non-integer lags)

Now I found that you can use FFT to calculate the ACF:

Get a buffer of 2048 samples
Window it (Hamming window)
fftBuf=fft(buffer)
buffer[i]=real(fftBuf[i])^2+imag(fftBuf[i])^2
fftBuf=fft(buffer) //ifft=fft for real signals
acfBuf = real(fftBuf) / 2048

Then actBuf[lag] is the ACF value at that lag.

I expected that the results will be the same or at least similar. But they are not. E.g for a 65.4Hz Sine wave (note C2) I get ~0.2 for a the corresponding lag of 674.25 using the time-domain approach and ~536.795 using the fft.

What did I miss? Or isn't both the same?

Flamefire
  • 5,313
  • 3
  • 35
  • 70
  • Did you pad your signal before taking the FFT? If not you would get the cyclic autocorrelation. – SleuthEye Apr 27 '14 at 00:49
  • AKF=Autocorrelation function, AMDF=Average Mean Difference Function Do you have any information on padding (why, how much...)? I thought, I won't need it as I already window the input. – Flamefire Apr 27 '14 at 06:40
  • Ok tried this: Window the 2048 samples and append 2048 zeroes. Result is now around twice the result in the time-domain. Is this due to the padding by a factor of 2? Besides that: Using 4096 samples instead of 2048 is quite some overhead. Any better solution with less overhead? – Flamefire Apr 27 '14 at 08:17

0 Answers0