I have to apply a window to an IFFT'ed signal (based on LTE standards, simulated for 1 symbol - 100 resource blocks - 1200 subcarriers). I have a signal which I have simulated till the IFFT. Now I have to apply a window function to it to analyse the Out of band emissions. How do I decide the window function length and then how do I apply it? e.g Hamming/Hanning/Blackman etc.


If allowed, I would view the signal in the frequency domain (2048 pt FFT for a single symbol, Fs=30.72Mhz).

Anything outside the central 1200 carriers in the symbol is "noise" and shouldn't be there.

You can filter the noise by setting the outer FFT bin amplitudes to 0, then convert back to the time-domain or perform other signal processing as desired.

Edit: that may not really help your cause, since analysis and demodulation is performed on the center 1200 carriers only... Setting outer carriers to 0 won't change the carriers/frequency content that matters... So maybe I'm missing your point?

  • $\begingroup$ I applied the window function (Hann(order 2048)) to the IFFT signal (doing a point wise multiplication) and got the windowed signal. The side lobes reduce considerably however the BER is quite high 6.3e-02. Is this normal or should I try something else as well $\endgroup$ – Fam1234 Dec 11 '18 at 18:48
  • $\begingroup$ That could be due to poor framing (not exactly aligned in time-domain to start of symbol). Otherwise, are you applying the reference signal (RS) for correction? $\endgroup$ – bornruffians Dec 12 '18 at 16:25
  • $\begingroup$ The length of FFT bin is 2048 and I applied Hann/Hamm of length 2048 so I guess alignment is not an issue. As far as applying RS goes, I am to study the impact of windows and filters on PAPR and OOB so RS correction is not part of my task as of now. $\endgroup$ – Fam1234 Dec 13 '18 at 16:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.