Hi all, where I can find a library or a tool for performing cross-correlation by FFT, or an easy understandable (for a biologist !) description of the implementation ?
Gianni
Dear Gianni
I suggest you use MATLAB software and use the following function:
r = xcorr(x,y);
This function returns the cross-correlation of two discrete-time sequences, x and y. Cross-correlation measures the similarity between x and shifted (lagged) copies of y as a function of the lag. If x and y have different lengths, the function appends zeros at the end of the shorter vector so it has the same length, N, as the other.
For further information go to MATALB help and search for (xcorr) function.
I hope that it is the beginning to you to solve your problem.
With my best wishes.
thanks for the answer. However, I'm willing to implement the cross-corr by explicitly using the complex FFT on two real signals and the manipulation to get the cross-corr by IFFT, this to be then able to perform the cross-correlation only in a given bandwidth. What I'm searching for is hwo to perfomr the requied operations in the complex domain.
You can do the calculations in the frequecy domain. It is explained in detail here.: https://dsp.stackexchange.com/questions/22877/intuitive-explanation-of-cross-correlation-in-frequency-domain
Best,
In words:
1) Put real signals in complex arrays --> real value as real part and set imaginary part to zero
2) compute complex FFT of both complex arrays
3) multiplication of complex FFT_1 and conjugate complex FFT_2
4) compute inverse FFT of multiplication
fx = fft(x,2*N); % 2N point FFT of x
fy = fft(y,2*N); % 2N point FFT of y
r_xy = ifft(fx.*conj(fy)) % r = [r0,r1,...,r_N-1,0,r_1-N,...,r_-1]
Just wrote the equivalent of Dr. Oleg I. Berngardt's mathematical expression for quick reference for new visitors! ;)
- Badges
- Science topic