Comments on: Interpreting the output of fft() operation in Matlab http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/ Signal Processing for Communication Fri, 10 Feb 2012 01:03:15 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: Krishna Sankar http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-91708 Krishna Sankar Mon, 02 Jan 2012 23:43:34 +0000 http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-91708 @saru: Hmm... in OFDM, we wish to send a group of complex sinewaves with each sine having different weights, and the frequency of the sine waves are integer multiplies of each other. Does it sound similar to what fft/ifft does? @saru: Hmm… in OFDM, we wish to send a group of complex sinewaves with each sine having different weights, and the frequency of the sine waves are integer multiplies of each other. Does it sound similar to what fft/ifft does?

]]> By: Krishna Sankar http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-90173 Krishna Sankar Thu, 29 Dec 2011 00:44:05 +0000 http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-90173 @HariJo: To get a real output from the ifft, make the input conjugate symmetric. For example, playing with the Matlab code provided in http://www.dsplog.com/2008/06/10/ofdm-bpsk-bit-error/ clear all nFFT = 64; % fft size nDSC = 52; % number of data subcarriers nBitPerSym = 52; % number of bits per OFDM symbol (same as the number of subcarriers for BPSK) nSym = 1; % number of symbols % Transmitter ipBit = rand(1,nBitPerSym*nSym) > 0.5; % random 1's and 0's ipMod = 2*ipBit-1; % BPSK modulation 0 --> -1, 1 --> +1 ipMod = reshape(ipMod,nBitPerSym,nSym).'; % grouping into multiple symbolsa xF_symm = [zeros(nSym,6) ipMod(:,[1:nBitPerSym/2]) zeros(nSym,1) ipMod(:,[nBitPerSym/2:-1:1]) zeros(nSym,5)] ; xt_symm = (nFFT/sqrt(nDSC))*ifft(fftshift(xF_symm.')).'; @HariJo: To get a real output from the ifft, make the input conjugate symmetric. For example, playing with the Matlab code provided in
http://www.dsplog.com/2008/06/10/ofdm-bpsk-bit-error/

clear all
nFFT = 64; % fft size
nDSC = 52; % number of data subcarriers
nBitPerSym = 52; % number of bits per OFDM symbol (same as the number of subcarriers for BPSK)
nSym = 1; % number of symbols
% Transmitter
ipBit = rand(1,nBitPerSym*nSym) > 0.5; % random 1′s and 0′s
ipMod = 2*ipBit-1; % BPSK modulation 0 –> -1, 1 –> +1
ipMod = reshape(ipMod,nBitPerSym,nSym).’; % grouping into multiple symbolsa
xF_symm = [zeros(nSym,6) ipMod(:,[1:nBitPerSym/2]) zeros(nSym,1) ipMod(:,[nBitPerSym/2:-1:1]) zeros(nSym,5)] ;
xt_symm = (nFFT/sqrt(nDSC))*ifft(fftshift(xF_symm.’)).’;

]]> By: HariJo http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-90043 HariJo Wed, 28 Dec 2011 14:14:55 +0000 http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-90043 Hi Krishna, I am trying to generate a real unipolar ofdm signal using 16 qam, 52 subcarriers and fft-size 64 to begin with. The way to generate unipolar ofdm is to modulate only odd subcarriers and the way to have real output from IFFT is to make the input vector have Hermitian symmetry. However, i am confused with the indices that matlab uses to designate subcarriers and the actual subcarriers. I want to simulate the BER performance of real unipolar OFDM in awgn channel. Can you please help? thanks. Hi Krishna,

I am trying to generate a real unipolar ofdm signal using 16 qam, 52 subcarriers and fft-size 64 to begin with. The way to generate unipolar ofdm is to modulate only odd subcarriers and the way to have real output from IFFT is to make the input vector have Hermitian symmetry. However, i am confused with the indices that matlab uses to designate subcarriers and the actual subcarriers. I want to simulate the BER performance of real unipolar OFDM in awgn channel. Can you please help? thanks.

]]> By: saru http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-87806 saru Thu, 22 Dec 2011 04:41:02 +0000 http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-87806 hi, i wanna know why fft/ifft is important in ofdm .does it play any signifinace role in ofdm hi,
i wanna know why fft/ifft is important in ofdm .does it play any signifinace role in ofdm

]]> By: Krishna Sankar http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-56457 Krishna Sankar Thu, 26 May 2011 00:48:15 +0000 http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-56457 @samba: In 802.11, the specification allows for a long training field (LTF) which is known a priori at the receiver. The model is Y = HX + N, where X is known at the receiver N is the noise, H is the channel Y is the received signal The noisy estimate of channel is H' = Y/X @samba: In 802.11, the specification allows for a long training field (LTF) which is known a priori at the receiver.
The model is Y = HX + N,
where X is known at the receiver
N is the noise,
H is the channel
Y is the received signal

The noisy estimate of channel is H’ = Y/X

]]> By: Krishna Sankar http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-56227 Krishna Sankar Tue, 24 May 2011 00:42:16 +0000 http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-56227 @teddy: FFT size and subcarrier number need not be the same. The text you mentioned means that out of 2048 available subcarriers, only 1201 are used. @teddy: FFT size and subcarrier number need not be the same. The text you mentioned means that out of 2048 available subcarriers, only 1201 are used.

]]> By: samba http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-53507 samba Sat, 23 Apr 2011 01:49:36 +0000 http://www.dsplog.com/2007/06/17/interpreting-the-output-of-fft-operation-in-matlab/#comment-53507 Hi sir can you explain how we do the channel estimation in ofdm 802.11 like in order to find the timing sync using cross correlation and auto correlation In some papers they used something like cross correlation of received signal with short preamble and divided it by auto correlation of short preamble , can you explain what exactly happens with this ? Hi sir

can you explain how we do the channel estimation in ofdm 802.11
like in order to find the timing sync using cross correlation and auto correlation

In some papers they used something like
cross correlation of received signal with short preamble and divided it by auto correlation of short preamble , can you explain what exactly happens with this ?

]]>