|
Message
From: terkko at cc.hut.fi<terkko@c...>
Date: Thu Jan 22 11:02:43 CET 2004
Subject: Odp: [oc] FFT for COFDM systems
On Thu, 22 Jan 2004 10:13:35 +0100 Jerzy G <furia1024@w...> wrote:> Hello
>
> > I'm working on a small parallel FFT core (either 8 or 16 point
> FFT).
> > The idea is that this core will be surrounded by a lot of
> memory blocks
> > in an FPGA and data will be routed into and out of it by some
> other
> > logic. This will be a building block for a 8K FFT or
> whatever. Anyway,
> > is COFDM transmition systems there is an I and a Q stream of
> symbols. I
> > know that the I and Q are refered to real and imaginary, but
> do I need a
> > complex FFT or two real FFT's?
>
> Complex FFT of course,
> if s(t) = q(t)+i(t) then S(w) != Q(w) + I(w);
>
> I think so.
>
> Jerzy Gbur
I have understood, that in radio technology it is quite common to send data on
"I" and "Q" portions of the carrier. The carrier can be only real or have also
the imaginary part and imaginary differs pi/2 from real as a vector... as
sin/cos won't correlate with same frequency signal with 90 degrees phase
difference, it is possible to use such carriers to convey totally different
information.
So allthough the names come from complex mathematics, these signals might not
have nothing to do with each other (could be even different tv-channels).
I do not remember how is the case in DVB and such, but "O" in COFDM comes from
"orthogonal", might hint that they are dependent...
Check the standard; if investigating DVB, as it would seem, go to
"http://portal.etsi.org", thereon broadcast and dvb...
hopefully you find out what you need...
Teemu Erkko
|
 |
Browse
