| Author |
Message |
Curl
Guest
|
 Posted:
Tue Dec 14, 2004 8:05 pm Post subject:
Slight Modification of the FFT algorithm |
|
|
Hello All,
Suppose I have a non linear system with four inputs x1,x2,x3,x4 and
one output y²
x1(n) --> --> y1²(n)-------\
x2(n) --> (H) --> y2²(n)------ +-----> y²(n) =
y1²(n)+y2²(n)+y3²(n)+y²4(n)
x3(n) --> --> y3²(n)------/
x4(n) --> --> y4²(n)-----/
I'm only interested in the energy of signals.
sum(y²(n)) = sum(Y²(k)) where Y²(k) = |FFT(y(n))|² (I
voluntary forget the usual N scaling factor)
I could compute: Y²(k) = Y²1(k) + Y²2(k)+ Y²3(k) + Y²4(k)
In this Case I need to compute 4 FFTs : FFT(x1(n)) ,....,
FFT(x4(n))
Due to Real-Time Constraints, i'm looking the way to transform my FFT
algorithm, or to correctly interleave the datas
(not sure its possible) to get this operation
(x1(n), x2(n), x3(n), x4(n)) -----> Y²(k)
Thank you for your comments |
|
| Back to top |
|
 |
Curl
Guest
|
| Posted:
Tue Dec 14, 2004 8:07 pm Post subject:
Re: Slight Modification of the FFT algorithm |
|
|
Curl <Mr.Bilou@microsoft.fr> a tapoté :
| Quote: |
y²(n) = y1²(n)+y2²(n)+y3²(n)+y²4(n)
|
|
|
| Back to top |
|
|
Fred Marshall
Guest
|
| Posted:
Tue Dec 14, 2004 10:15 pm Post subject:
Re: Slight Modification of the FFT algorithm |
|
|
"Curl" <Mr.Bilou@microsoft.fr> wrote in message
news:41bf00c3$0$3815$636a15ce@news.free.fr...
| Quote: | Hello All,
Suppose I have a non linear system with four inputs x1,x2,x3,x4 and
one output y²
x1(n) --> --> y1²(n)-------\
x2(n) --> (H) --> y2²(n)------ +-----> y²(n) =
y1²(n)+y2²(n)+y3²(n)+y²4(n)
x3(n) --> --> y3²(n)------/
x4(n) --> --> y4²(n)-----/
I'm only interested in the energy of signals.
sum(y²(n)) = sum(Y²(k)) where Y²(k) = |FFT(y(n))|² (I
voluntary forget the usual N scaling factor)
I could compute: Y²(k) = Y²1(k) + Y²2(k)+ Y²3(k) + Y²4(k)
In this Case I need to compute 4 FFTs : FFT(x1(n)) ,....,
FFT(x4(n))
Due to Real-Time Constraints, i'm looking the way to transform my FFT
algorithm, or to correctly interleave the datas
(not sure its possible) to get this operation
(x1(n), x2(n), x3(n), x4(n)) -----> Y²(k)
|
Let's see if I can translate your requirement:
You have a composite signal in y²(n).
What's not clear to me is how it can be that y²(n) is always equal to the
sum of the squares of y1, y2....
That seems most unusual and begs the definition of y(n).
How is it that you could compute the FFT of various outputs when, in fact,
you say there is only *one* output? That makes no sense to me.
Why is it that you can't compute the FFT of y(n) or y²(n)?
I think there's more going on here than you've told us.
Fred |
|
| Back to top |
|
|
Curl
Guest
|
| Posted:
Wed Dec 15, 2004 10:06 pm Post subject:
Re: Slight Modification of the FFT algorithm |
|
|
Fred Marshall <fmarshallx@remove_the_x.acm.org> a tapoté :
| Let's see if I can translate your requirement:
<COUIC>
Sorry I did not have time to give you further precisions today.
I will draw something and scan it this evening or tomorrow.
Thank you |
|
| Back to top |
|
|
Jerry Avins
Guest
|
| Posted:
Wed Dec 15, 2004 10:20 pm Post subject:
Re: Slight Modification of the FFT algorithm |
|
|
Curl wrote:
| Quote: |
Fred Marshall <fmarshallx@remove_the_x.acm.org> a tapoté :
| Let's see if I can translate your requirement:
COUIC
Sorry I did not have time to give you further precisions today.
I will draw something and scan it this evening or tomorrow.
Thank you
|
Images shouldn't be posted in most news groups; many services reject
them. Give it a URL somewhere and post that. (There are binary
newsgroups, but many ISPs don't support them and not everyone who can
subscribes.) If you email your image to me -- no more than an MB please
-- I'll put it on my site.
Jerry
--
Engineering is the art of making what you want from things you can get.
ŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻ |
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in