| Author |
Message |
lucy
Guest
|
 Posted:
Sun Dec 19, 2004 2:58 am Post subject:
DFT zero padding problem -- finding odd-indexed spectrum? |
|
|
Hi all,
Suppose I have a N-point DFT F[m] of signal f[n],
and now I zero-padded my N point signal f[0]...f[N-1] to 2N points by adding
N trailing zeros... and define
new signal g[n]=
f[n], for n=0, to n=N-1;
0, for n=N to n=2N-1;
Now taking 2N DFT of g[n] and get G[m]...
I can easily find
G[m]=F[m/2] for m=even...
but can you find G[m] in terms of F[something] for m=odd?
Thanks a lot |
|
| Back to top |
|
 |
Bob Cain
Guest
|
| Posted:
Sun Dec 19, 2004 7:57 am Post subject:
Re: DFT zero padding problem -- finding odd-indexed spectrum |
|
|
lucy wrote:
| Quote: | Hi all,
Suppose I have a N-point DFT F[m] of signal f[n],
and now I zero-padded my N point signal f[0]...f[N-1] to 2N points by adding
N trailing zeros... and define
new signal g[n]=
f[n], for n=0, to n=N-1;
0, for n=N to n=2N-1;
Now taking 2N DFT of g[n] and get G[m]...
I can easily find
G[m]=F[m/2] for m=even...
|
Don't you mean F[m]=G[m/2] ?
| Quote: | but can you find G[m] in terms of F[something] for m=odd?
|
Sure, F[m]=G[(m-1)/2] for m=odd
I know that's not what you mean but I'm not sure what you do
mean.
Bob
--
"Things should be described as simply as possible, but no
simpler."
A. Einstein |
|
| Back to top |
|
|
Rick Lyons
Guest
|
| Posted:
Mon Dec 20, 2004 11:55 pm Post subject:
Re: DFT zero padding problem -- finding odd-indexed spectrum |
|
|
On Sat, 18 Dec 2004 19:06:13 -0800, Bob Cain
<arcane@arcanemethods.com> wrote:
| Quote: |
lucy wrote:
Hi all,
Suppose I have a N-point DFT F[m] of signal f[n],
and now I zero-padded my N point signal f[0]...f[N-1] to 2N points by adding
N trailing zeros... and define
new signal g[n]=
f[n], for n=0, to n=N-1;
0, for n=N to n=2N-1;
Now taking 2N DFT of g[n] and get G[m]...
I can easily find
G[m]=F[m/2] for m=even...
Don't you mean F[m]=G[m/2] ?
but can you find G[m] in terms of F[something] for m=odd?
Sure, F[m]=G[(m-1)/2] for m=odd
I know that's not what you mean but I'm not sure what you do
mean.
Bob
|
Hi Bob,
yep, lucy's question is not
very well "worded". If lucy has computed
G(m) with a 2N-point DFT, then lucy doesn't
have to "find" G(m).
I'll bet what lucy means is: "How can we
compute G(m) if all we have is F(m)?"
Am I correct lucy?
[-Rick-] |
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in