The Fast Fourier Transform explained (#000)

Digital Signal Processing
Digital Signal Processing
52,349
0
Published 2013-10-15
The Fast Fourier Transform uses a simple trick: divide the time series in odd/even sequences and perform DFTs on them. Then repeat this until we have two point DFTs which are just X(0) = x(0) + x(1), X(1) = x(0) - x(1). Then recombine the results and we have computed the DFT much quicker.
All Comments (17)
  • @thebigVLOG
    Best FFT explanation video on YouTube, fact.
  • @iSolarSunrise
    This is the BEST intuitive explanation I have found on FFT! Keep these videos coming! Also, I really enjoy your biosignal videos too. Keep them coming as well :)
  • @21jon
    You, good sir, are my hero! I had trouble understanding the fundamentals of FFT until I watch this. Cheers :)
  • @BoZhaoengineering
    hi, Thank you for your interpretation of FFT. I am a structure engineer working in wind energy section. But signal processing is one of my favorite subject.
  • @moyeen
    Great Explanation :) Thank you very much .
  • @coffeecuppepsi
    please help me understand. if i sample in a signal and perform FFT, now i want to know the real coefficients and the imaginary coefficients.. i.e. in a practical sense i want to build an impedance tester, the real part will give me R and the imaginary denotes the L or C component.
  • @andyw732
    Thanks for the video! I still don't fully understand it though, for example: at 11:55 I don't get why it's (N/2)² for the even. For me the inner loop is N/2, and the k still goes from 0..N-1 so it's N²/2. Same for the uneven gives N² . Could you elaborate?